Cdnc85082enc_001.pdf

:

'ëuX io. /ti. Commission of the European Communities

Fire-safe steel construction practical design

Commission of the European Communities

Fire­safe steel construction : practical design Proceedings of the international conference held in Luxembourg, 11 and 12 April 1984 This conference was organized by the Commission of the European Communities in conjunction with : Beratungsstelle für Stahlverwendung, Düsseldorf Centre Belge­Luxembourgeois d'Information de l'Acier, Bruxelles Centro Italiano Sviluppo Impieghi Acciaio, Milano Constructional Steel Research and Development Organizations, Croydon Office Technique pour l'Utilisation de l'Acier, Paris Stichting Staalcentrum Nederland, Rotterdam and The European Convention for Constructional Steelwork

Directorate­General I r A K L Π ■',. Science, Research and Developmerjt­ 1985

NC

CL

J

^. Hicifiofh.

­ E Ü R 10116 EN

Published by the COMMISSION OF THE EUROPEAN COMMUNITIES Directorate-General Information Market and Innovation Bâtiment Jean Monnet LUXEMBOURG

LEGAL NOTICE Neither the Commission of the European Communities nor any person acting on behalf of the Commission is responsible for the use which might be made of the following information

This publication is also available in the following languages: DE ISBN 92-825-5717-0 FR ISBN 92-825-5719-7

Cataloguing data can be found at the end of this publication

Luxembourg: Office for Official Publications of the European Communities, 1985 ISBN 92-825-5718-9

Catalogue number: CD-NC-85-082-EN-C

© ECSC-EEC-EAEC, Brussels · Luxembourg, 1985 Printed in Belgium

CONTENTS OPENING

SESSION

Chairman : P.R.V. EVANS, Head of Division, Directorate-General for Science, Research and Development, Division "Technical research (steel)", Commission of the European Communities, Brussels Belgium Introductory address H. TENT, Director, Directorate-General for Science, Research and Development, Commission of the European Communities, Brussels, Belgium

3

Trends in design methods for structural fire safety J. WITTEVEEN, Director Institute TNO for Building Materials and Building Structures, DELFT, The Netherlands

11

Fire development and safety requirements in construction W. KLINGSCH, Bergishe University of Wuppertal, F.R. Germany

26

ECSC contribution in the field of fire-safety of steel structures M. DESCUDE, Industrial idviser, president of Executive Committee on properties and behaviour in service of light structures, Paris, France

42

Summary of discussions Rapporteur : J. FERRON

52

S E S S I ON

I

:

DESIGN

METHODS

Chairman : G. Th. WUPPERMANN, Geschäftsführender Gesellschafter Th. Wuppermann, GmbH, Leverkusen, F.R. Germany Presentation of european recommendations for the fire safety of steel structures J. KRUPPA, Head of fire service at the Metal Structures Technical Industrial Centre and Member of Commission 3 of the European Convention on Metal Structures, Puteaux, France

-III-

58

Examples of calculation of fire resistance of the steel members according to the european recommendations of ECCS TC3 S. BRYL, Research and Development Department, Geilinger Ltd., Winterthur, Switzerland Fire safety, design of composite columns J.B. SCHLEI CH, Department Manager, ARBED-Research, Esch/Alzette, Luxembourg Fire engineering design of composite concrete slabs with profi lied steel sheet L. TWI LT, I nstitute TNO for Building Materials and Building Structures, Delft, The Netherlands

73

89

128

Fire resistance of composite steel decks, floors and beams H.B. WALKER, Head of Advisory Services, Constructional Steel Research and Development Organisations, Croydon, United Kingdom

144

Report on Session I Rapporteur: H. WITTE

161

S E S S I O N II

:

P R A C TI C A L ASPECTS OF IMPLEMENTING SAFETY

Chairman : TI EZZI , Capo della Direzione Generale Antincendio & Protezione Civile, Ministero dell'I nterno, Rome, Italy Practical solutions by architects Κ. SCHUWIRTH, Architect Bureau Schuwirth 6 Erman, Hannover, F.R. Germany Practical solutions by architects. Practical aspects of implementing safety G. ABBADO, Architect, I NSO SpA, E.N.I . Group - Florence, Italy

164

175

A consultant's view of steel structures M. LAW, Technical Director, Ove Arup Partnership, London, United Kingdom

194

How to reduce the cost of fire safety L. FRUI TET, I ngénieur-Conseil de l'Office Technique pour l'Utilisation de l'Acier, Ecole des Beaux-Arts, Paris, France

204

-IV-

Industrial buildings - fire losses and appropriate protective measures J. THOR, Swedish Institute of Steel Construction, Stockholm, Sweden Report on Session II

219

235

Rapporteur: G.M.E. COOKE S E S S I O N

III

:

Chairman :

F U T U R E

P R O S P E C T S

P. BORCHGRAEVE, Centre Belgo-Luxembourgeois d'Information de l ' A c i e r , Brussels, Belgium

Application on the computer to model structural f i r e endurance

240

D.C. JEANES, Senior Engineer, American Iron and Steel I n s t i t u t e , Washington, USA Computer aided f i r e resistance for steel and composite structures

261

J.C. DOTREPPE, Senior Research Associate, J.M. FRANSSEN, Research Assistant, National Fund for Scientific Research, Liège, Belgium; J.B. SCHLEICH, Department Manager, ARBED, Esch/Alzette, Luxembourg Requirements of f i r e resistance based on actual f i r e s (Swedish approach)

279

0. PETTERSSON, Division of Building Fire Safety and Technology, Lund I n s t i t u t e of Technology, Sweden A probability based f i r e safety concept

294

M. KERSKEN-BRADLEY, I n s t i t u t für Bautechnik, B e r l i n , F-.R. Germany Report on Session I I I

307

Rapporteur: L. TWILT

CLOSING

SESSION

Conference conclusions

312

P. BORCHGRAEVE, Centre Belgo-Luxembourgeois d'Information de l ' A c i e r , Brussels, Belgium

L I S T

OF

P A R T I C I P A N T S

-V-

319

OPE NING

SESSION

Introductory address Trends in design methods for structural f i r e safety Fire development and safety requirements in construction ECSC contribution in the field of fire-safety of steel structures

Summary of discussions

INTRODUCTORY ADDRESS. H. TENT, Director, General Science direction, of Research and Development. Ladies and Gentlemen, An important objective of the Community is always to ensure and develop the competitivity of our industries, particularly by a steady effort in the field of industrial research, on both a national and a Community scale. In fact, the development and reactivation of the so-called conventional industries are two of the objectives that the Commission has Included in its research programme for science and technology in future years. This desire to ensure a research strategy is thus materialised by the initiation of vast programmes such s the "ESPRIT" programme for the technology of information and the "BRITE" programme on basic technology and the development of new technology. We must not, however, lose sight of the fact that the first initiative of a community programme of applied research dates back some 30 years, the date of the first ECSC research programme. It is interesting to note that the theme of this programme was exactly the same as the one you are to discuss in the next two days, i.e. fire safety of steel structures. I must stress that the actions of ECSC Steel research do not only act as catalysts in research and development, but also:

- enable a comparison and association of the research methods used in the Community -develop the actions and exchanges between scientists and technologists -reduce the duplication of effort -give added credibility and international significance to the results. This is particularly true in the field of fire prevention with which we are concerned to-day and which has undergone considerable development during the last few years. However, if we master the technique, we must also develop a better understanding, a common language between the originators, designers and those in charge of fire prevention. We must show them that viable methods exist, including those of information. When the title of the conference - T i r e Safety of Steel Structures = practical conception" was announced, three questions sprang to my mind: questions to which, I hope, the answers or at least further enlightenment will be provided in the course of the next two days: -firstly -next -finally

Why this conference, and why now? For whom this conference? How to attain the objectives that we, the interested parties, propose?

I. WHY THIS CONFERENCE 1. Because this sector of construction is a key sector of the economy and social life of the Community. In fact, if one considers the overall scope of all interested business organisations in the construction industry, whether for supply of products, tools etc. or for actual construction, the order of magnitude of the number employed in the Community is probably over 10 million, which represents at least lot of the working population! The economic and social implications are therefore enormous.

Furthermore, in comparison with its world-wide competitors LU.S.A., Japan), the European construction industry is well situated for export business. This means that exports are vital for the professions involved, and that we know how. to design structures, components, ensembles, attractive and competitive arrangements. B ut we must not rest on our laurels - only a constant effort is rewarding! 2.Because steel construction in its widest sense, i.e. considering structures in all forms of utilisation of steel = shells (roofing) finishing and completion of buildings, represents 12.5% of steel consumption in the Community. This is a far from negligible percentage for our steel industry. The products involved are essentially: - hot-rolled sections, beams, - cold-rolled sections and various shapes, - heavy and medium plates, -thin sheet, mainly clad for corrosion resistance, in the form of ρre-painted, galvanised or plastic coated sheet. In all cases the products incorporate the most recent developments in regard to grade of material, quality, anti—corrosion properties, related to aesthetic and decorative qualities such as their surface condition (relief effect) and colour. These properties should permit the architect to innovate and cause the creation of a better future. 3. Because the safety of buildings, which includes the preservation of people and goods, and more particularly the fire safety, is a subject that concerns us deeply. He know that the task of designers is often extremely difficult when they are confronted with the often rigid attitude of control organisations and the different interpretation of rules.

That is why we must know the methods that, at the present time, enable us in a practical and viable manner and to questions raised in the study and resistant steel buildings.

and recommendations to tackle these problems to provide answers construction of fire-

This change i n attitude, which should lead to re-consideration of the spirit of rules rather than their blind application, has become imperative. 4.Because it is now that we must think of tomorrow. In fact any crisis contains the seeds of a new idea. This is why careful attention should be paid to the slightest symptoms, to the minutest details that can indicate, arouse and assist the reactivation of these seeds and create the environment favourable for their development. The indicators of activity are favourable and suggest a slight increase in activity. Thus for building and civil engineering this index was: 96.9 for the 4th quarter of 1983, 84.7 for the 1st quarter of 1984, and it should reach lo2 by the 2nd quarter of 1984. Our role is thus to create a network of techniques, and technology, in various orgainisations, vitalised by responsible men conscious of their common task. That is why you are here, and I thank you for attending. II

FOR WHOM.

To whom should we address ourselves during this meeting, and also by whom will these efforts be made?

We must maintain a dialogue with, all parties concerned: architects, consulting engineers, construction engineers, controlling authorities, fire-fighting organisations, insurers and investors. Thus the planning offices lack information for joint planning with both metal constructors and producers. It is necessary to engender an atmosphere of intercommunication between tne various parties to tackle and solve the problems in a conerent manner. In this process, the relations with safety organisations, fire fighters are well experienced. Their experience in relation to fire development and ways of controlling it enables us to evaluate our own ideas. Thus the ideas have been well developed oyer the years. It has been difficult to make the transition between the concept of "standard fire" and that of "real fire". Our effort is now directed to the determination of protection of structures: is this always necessary, in what cases and how? These questions are on the agenda and you will doubtless provide elements of the answers. In fact, the idea or fire safety of metal structures is (dare I say?) as much a question of mentality as of technique. The notion of safety snould actually appear not only at the design stage, but also in the use and exploitation of a building; the plans must be observed, the methods of control maintained, whatever the structure. That is why, to promote the use of steel, experts must be convinced that the intelligent use of steel can only reinforce safety factors.

III. HOW The various organisations concerned have made considerable efforts to convince thier speakers- EEC in the case of ECSC steel activities, the European Convention for Metallic Construction in the case of work on EUROCODES, and also by the Steel Information Centres which are in daily contact with the interested professions and who provide the indispensible link between knowledge and its application. «-· These various actions are coordinated at community level in such a way as to create the conditions necessary for circulation of products and opening of markets, as a result of a constant modification of codes (EUROCODES), of methods and techniques for laboratory testing and of production methods (in the case of ECSC), also of the methods of application (this is one of the tasks of the Information Centre). -For this, the priorities have been defined in the case of our activities: -to consider national regulations with a view to their harmonization and their adaptation to the results of research, thanks to a more precise definition of the range of safety requirements and of structural performance. -to provide suitable methods and calculation tables and suitably modified guidelines to the responsible personnel. This common and coherent approach to the problems raised by Fire Safety of Steel Structures is absolutely essential. It is, in fact, one of the important roles delegated to participants in meetings such as to-day's. -It is by direct exchange that the best understanding and appreciation of opinions can be achieved.

So, in this appreciation of the concept of "Fire Safety", the ideas have changed considerably- empiricism is a thing of the past. The present-day methods of calculation by computer, controlled and invalidated where appropriate by experience, enable a realistic approach to the conditions of fire-behaviour of metal structures and an evaluation of the performance of steel structures taking account of the new technologies and techniques for production, fabrication and application. The steels currently available are the result of extremely advanced research work, which enables them to satisfy the most stringent conditions of use. Our interest today relates to a highly specialised environment, but one which must be carefully considered, that of fire. We shall see that nothing can be said to show that steel is not, in this context, a safe material. This action of safety, to which it is convenient to add that of "cost" as there is always a financial effect, has been thebasis of research and development work in the field of metal construction; work whose continuity expresses tne will to explain how to use steel in an optimal manner in association with other materials. It is not, of course, desired to achieve a systematic replacement of one material by another, but to use steel advantageously in cases where its intrinsic qualities may best be exploited. It is interesting to note that these projects were undertaken with prudent logic, and with a tenacious desire to convince architects, decision makers and insurers that steel, by virtue of its favourable performance/cost ratio, can be proved as a highly competitive material in the building industry.

We have the opportunity to be helped in our task by eminent experts from the Community, and also from Switzerland, Sweden, and the U.S.A., and who will share with you their approach to the questions raised by "Fire Safety". The way in which these countries have tackled the problem of regulations will be very interesting and very instructive. It is worth mentioning that in North America 80% of multistorey buildings have steel frames, whereas the corresponding figure for Europe is only 28%! What a potential market for our steel industry! Ladies and Gentlemen, I do not think I need remind you of the importance that we attach to the lessons that this conference will bring to us; a conference on which rests, as with all efforts concerned with steel, the beneficial spirit of the promoters of ECSC and particularly here in Luxembourg. I am sure that you have the "Sacred Fire" for the success of this conference and that your deliberations will be fruitful. I wish you a successful activity during these two days.

-10-

TRENDS IN DESIGN METHODS FOR STRUCTURAL FIRE SAFETY

J. Witteveen Director Institute TNO for Building Materials and Building Structures, Rijswijk, The Netherlands Professor of Structural Mechanics Delft University of Technology, The Netherlands

Summary During the last decade there has been an Important progress in analytical modelling of fire exposure and in the development of probabilistic methods of fire risk assessment. As a result, the required structural fire protection can be assessed in a rational way in combination with active preventative measures, such as early detection and sprinklers. Analytical methods have also been developed for the determination of the load bearing capacity of elements and structures at elevated temperatures as an alternative to the standard fire resistance test. However, the choice and use of new design methods will be greatly influenced by the present rigid building regulations and the shared competence between authorities, responsible for the requirements and the designer for proving compliance. The author concludes that major progress in the implementation of new design concepts can only be achieved with reformed building regulations and change in competence. Broad cooperation among all concerned is needed and it is recommended to direct research to policy related programs rather than to the present physically oriented programs.

-11-

1. INTRODUCTION Fire prevention measures and suppression in general serve both social and monetary interest simultaneously. The overall objective is an optimum return on investment in fire precautions in terms of lives and property saved. It is important to appreciate that fire protection by structural fire resistance alone does not generally assure adequate reduction in material damage and personal risk. Apart from escape routes and control of combustible materials, essential measures to be considered are sprinklers to avoid flash-over and fire groth as well as partitions to limit fire spread. For a given budget the optimum level of fire protection generally is provided by a combination of active measures, such as early detection and sprinklers and passive measures, provided by the building structure itself. When, for reasons of live and property safety, sprinklers are installed, it can be argued that the fire resistance of structural elements can be reduced. However, in the building regulations the required level of fire protection is expressed in one single parameter "required fire resistance time". As a consequence the present regulations emphasize on structural fire protection and do not provide means to balance use of alternative protective measures against reduction in structural fire protection to meet the same level of safety (keyword: equivalency or tradeoff). During the last few years one can observe a changing attitude to existing regulations and codes, and attempts are being made to achieve flexible solutions with greater economy and a defined and more uniform safety. The main components of such improved regulations and design methods for structural fire safety are: - improved heat exposure models; - improved structural response models, including analytical models as an alternative to the standard fire resistance test; - a probabilistic design including a methodology by which the required fire protection can be assessed in a rational way, in combination with active preventative measures such as early detection and sprinklers. The effect of suppression by fire brigades can also be dealt with. One consequence of applying such improved methods is a change in competence and responsibility between public agencies responsible for the

-12-

requirements and designers for proving compliance. The paper deals with a review of Improved design methods for structural fire safety, which have been developed during the last decade and are now becoming operational for practical application and incorporation in the building codes. The nature of this paper is conceptional rather than giving operational solutions for particular design situations. For the latter it is referred to the relevant literature.

2. THE PRESENT CLASSIFICATION SYSTEM Fires affect the structural performance of buildings, because they change the physical and mechanical properties of materials of construction. As a consequence a fire engineering design system needs to quantify the fire exposure on one hand and the effects of that exposure on structural behaviour on the other hand. Internationally, the generally accepted method for the design of load bearing structural elements under fire action is based on a classification system. The system is characterized by shared competence between public agencies responsible for the requirements and designers for proving compliance (Fig. I).

Γ Structural application

1

structural element

Requirement by public agencies required fire duration t,.,

building code

compliance by designers

1

standard fire resistance test

fire resistance

YES

END

NO

_l Fig. I. Fire engineering design based on a classification system related to the standard fire resistance test, characterized by shared competence between public agencies and designers. Requirement A fire exposure according to ISO 834, with a required time of fire duration t„., stipulated in buildig regulations and codes for the structural application in question

usually expressed in multiples of 30

minutes.

­13­

Compliance A standard fire resistance test according to ISO 834 by which the fire resistance time t,

of the structural element In question is determined

experimentally - usually classified in multiples of 30 minutes (1)· As an alternative to the standard fire resistance test, in some countries analytical methods are accepted. The design implies a proof that the structural element has a fire resistance t, , which meets the required time of fire duration t . fd Although the classification system has been in use for over half a century, it has some serious weaknessess. These weaknessess apply to both components of the design procedure and can be summarized as follows: Requirements The rise of temperature as a function of time according to ISO 834 and the fire duration are a rough approximation of the real gas-temperature time curve of a fully developed compartment fire. The required time of fire duration is generally related, not only to the estimated fire exposure, but is also differentiated with respect to safety considerations relevant for the building in question. This usually leads to a required time of fire duration, which is more severe than the actual fire exposure. The estimated fire exposure and the safety considerations are intermingled inextricably. As a result, in situations not covered by the building regulations, the required time of fire duration is often a matter of dispute between authorities and designers (2). Moreover, no basis exist for trade-off between reduction in structural fire protection and alternative measures such as compartmentatlon and sprinklers.

Compliance The specification of the fire resistance test according to ISO 834 is insufficient in several aspects, such as heat-flow characteristics of furnaces, material properties and imperfections of the specimen, temperature distribution along members and restraint conditions. The structural element to be tested has to be modelled with respect to actual conditions expected in the structure. Deviations from conditions in the actual structure are forced by limited dimensions of furnaces, idealized characteristics of the loading device and unsufficiently defined support conditions during the test (3,4). Fig. II shows some results of a

-14-

correlation test series on composite columns carried out in different laboratories (5). It appears that a considerable and random difference in results exist.

ψ 2M­2C0­S.3

«ϊκιϊΐ ■ iLuuC test laboratory

ψ 150­150­5

«ΕM£ a k. u υ

Fig. II. Some results of fire resistance tests on identical concrete filled hollow steel sections obtained in various test laboratories (5).

The deficiencies of the present classification system have certainly stimulated the development of rational methods of fire risk assessment and analytical modelling of thermal actions as well as structural response, which potentially give possibilities to achieve solutions with greater economy and a defined and more uniform safety. Horeover, It is recognized that, following probabilistic design procedures in other fields of design for accidental events, structural fire engineering design should be probability based. In contrast to the present classification system, probabilistic design includes a methodology by which all relevant factors, such as safety considerations from both the human and economic point of view, probability of flash­over, uncertainties in fire exposure and structural response, the effect of structural fire protection, fire brigade actions, early detection and sprinklers can be dealt with systematically.

­15­

3. CONCEPTS IN STRUCTURAL FIRE ENGINEERING DESIGN Generally a structural fire engineering design Includes two main elements, corresponding to the two components as described in Chapter 2, i.e. requirement and compliance (6, 7 ) . Both components can have different levels of schematization. Requirement A heat exposure model H, for the determination of the rise of temperature as a function of time. Basicly three types of heat exposure models may be identified with respect to the type of thermal exposure. The listing starts with the heat exposure model, presently used in most building regulations. (Hj) A rise of temperature as a function of time according to ISO 834. The duration of the temperature rise is equal to the "required time of fire duration", expressed in building regulations and codes for the particular use of the building or fire compartment. (H-) A rise of temperature as a function of time according to ISO 834. The duration of the temperature rise is equal to the "equivalent time of fire exposure", a quantity which relates a non-standard or natural fire exposure to the standard temperature-time curve (10, 11, 12). (Ho) A rise of temperature as a function of time characterized by an analytical determination of the gas-temperature time curve of a fully developed compartment fire (12, 13). The heat exposure model H is supplemented in a probabilistic way by factors such as: - the probability of flash over; - the effect of early detection, the reliability of sprinklers etc. to be considered as trade-off for structural fire protection. - the occupancy and importance of the building; - the height and volume of the building and the size of the exposed area; - the availability of escape routes and rescue facilities; - the consequence of violating a limit state.

-16-

Compliance A structural response model S, for the determination of the heat transfer to and within the structure and the ultimate load bearing capacity of the structure. The structural response model may be experimental or analytical. The design implies a proof that the structure or structural member, under a defined load and subjected to the specified heat exposure, fulfils certain functional requirements, expressed by the limit states with respect to load bearing capacity, thermal insulation, fire Integrity (1, 8, 9). Basicly, three types of structural response models may be identified with respect to the type of structural system. The listing starts with the structural response model, presently used in most building codes. (S.) The load bearing structure is decomposed in single members with simplified restraint conditions such as beams and columns. The model can be either experimental -standard fire resistance test- or analytical (1, 8, 9). (Sn) The load bearing structure is decomposed in sub-assemblies, such as beam-column systems. Although the model can occasionally be experimental -standard fire resistance test- an analytical approach will be prevalent (8, 9 ) . (S,) The load bearing structure, such as a building frame or a floor slab system is analysed as a whole. The model is only suitable for an analytical design, assuming fire exposure throughout the structure or only within an Individual compartment.

4. TOWARDS NEW PROBABILITY BASED DESIGN METHODS THROUGH COMBINATIONS OF HEAT EXPOSURE MODELS AND STRUCTURAL RESPONSE MODELS In the table of Figure III the heat exposure models and structural response models are combined in a matrix ln^sequence of improved schematization, but consequently also with increased complexity in practical application. In principle each element in the matrix represents a particular design procedure. The matrix therefore can be considered as a classification system for methods of structural fire engineering design. As mentioned before, safety considerations from both the human and

-17-

s,

Structural

Elements

\Model

s2 Sub­assembly

s3 Structures

« * *

1

Heat ^v DM S—^~Λ Exposure ^v Model \ ^ Λ ft λ

1111

Tf

J*»

I SO­834 Τ

H,

Li

i n i or calculation

calcul a t ion occasional t*st

oitførvncf ­n schMnallzallon bocontfs loo lorojp

IM) or calculation

calculation occasional tost

calcutalion

««o

τ

1 SO­834

H2

unpractical

ted

τ

H3

compor trant Fir·

Ar

calculation

calculation

occasional

coJculalion occasional and let nmoreh

tfe ■ rfqutrMi tmr- of lir. duration t^cj a «quivokml limp ol fir«* vipMur»

Fig. Ill· Matrix of heat exposure models and structural response models in sequence of improved schématisation. economic point of view as well as assessment of frequency such as probability of flash­over and effect of sprinklers are taken into account within the improved heat exposure models (see Chapter 4 ) . The design method H, ­ S. and occasionally H. ­ S. with experimental verification of the fire resistance, corresponds to a vast majority of national building codes (see Chapter 2). In many countries Improved methods based on heat exposure models H_ and H« (10, 11, 12, 13) have occasionally been used, but, except in Sweden, they are not yet automatically accepted as methods which satisfy the requirements of the building regulations. In contrast to the acceptance of Improved heat exposure models there is a growing acceptance of design methods H S and H S2 with an analytical verification of the fire resistance. In several countries these methods are now being used as an alterntive to the standard fire resistance test. Recently the European Convention for Constructional Steelwork (ECCS) and the Comité Euro­International du Beton (CEB) completed Recommendations providing reference documents for national codes of practice (8, 9 ) . These Recommendations apply to design methods based on heat exposure models H.

­18­

and H. and structural response models S. and S.. In principle a differentiated fire engineering design allows to make a problem oriented choice of a combination of heat exposure models and structural models, taking Into account reduction In structural fire protection when alternative active measures are employed. Relevant factors essential for the practical application of the respective design methods, summarized in Figure III, are (see for a detailed discussion (6): - The rule should be to provide a sensible relation In the levels of advancement of both models. Consequently the combination H1 - S_ and H, Sj, cannot be considered as design methods for general application. - From an operational point of view the complexity of the design procedures is of Importance. For the structural model S3 a computer is required for carrying out the structural analysis. Hence the combinations H

- S, and

Ho - S, are not to be considered operational design procedures for everyday practice. - As in a structural fire engineering design the fire resistance test will still be used for the years to come, design methods for general application should comprise an experimental as well as an analytical verification. Heat exposure model H- generally cannot be combined with an experimental structural model. - Within the four design procedures, being combinations of heat exposure models H. and H. and structural models S. and S,, there is an option for either an experimental or an analytical verification of the fire resistance. Both should be made compatible in order to render the same degree of reliability (14). - Finally, the choice of a design procedure will also be influenced by the present shared competence in structural fire protection between public agencies responsible for the requirements (i.e. the heat exposure model) and the designer for proving compliance (i.e. the structural response model). This will be discussed more in detail in Chapter 5. Generally, the design criterion in a fire engineering design requires that no limit state is reached during the fire exposure. For a load bearing structure, the design criterion implies that the minimum value of the load bearing capacity

( R (rO during the fire exposure shall meet the load

-19-

effect on the structure (S) i.e.: min {R ( t J - S > 0

(4.1)

In this formula the design criterion is adapted to design methods based on a natural fire, i.e. heat exposure model H,. For design methods based on the standard temperature-time curve i.e. heat exposure models Hţ and H_, the design criterion is expressed in a time domain, e.g.:

tfr ­ t£d > 0

(4.2)

where tf_ is the time in which the limit state of the structural element is reached, i.e. the fire resistance of the structural element; t,, is the required fire duration specified in the building regulations (heat exposure model H.) or calculated on the basis of heat exposure model H~. In the design methods based on heat exposure model H2

an

d Η,, the

following probabilistic aspects should be considered (heat exposure model Hi implicitly includes these aspects). ­ Intrinsic randomness of design parameters and properties. ­ Model uncertainties of the analytical models for the heat exposure and the structural response. ­ A ssessment of frequency, such as the probability of occurrence of a large fire, the effect of early detection, the reliability of sprinklers. ­ Safety considerations from both the human and economic point of view such as, the height, volume and occupancy of the building, the availability of escape routes and rescue facilities as well as the consequence of violating a limit state. Introducing these sources in a probabilistic manner into the design means that they must be expressed in numerical values. The level of the probabilistic analysis may well be limited to a semi­probabilistic approach, in which the aspects mentioned above are clustered and expressed in partial factors and characteristic values are used for action and response effects. For the design method H 2 ~ S„ with an analytical structural model, this probabilistic design format reads (6, 7, 10, 11):

ή"Ύη1 \2\^>0

<4·3>

-20-

The structural response model represents the first term of the equation and the heat exposure model the second term (see also Fig. IV). analytically determined fire resistance time of a sub-assembly equivalent time of fire exposure for the fire load and the fire compartment in question Ύ*

- partial factor taking into account intrinsic randomness of design parameters and material properties at elevated temperatures, uncertainty in loads and load combinations, as well as uncertainty in the analytical structural response model partial factor taking Into account the uncertainty in specifying the fire load, ventilation characteristics of the fire compartment and the thermal properties of the enclosure, as well as uncertainty in the heat exposure model

'nl T

n2

partial factor taking into account the assessment of frequency partial factor taking into account the safety considerations The partial factors γ follow from statistical data and socio-economic

optimization supplemented by engineering judgement (10, 11). The design can be simplified by using unified γ factors for certain classes of buildings, such as appartment buildings, schools, offices etc. Finally it should be emphasized that a transition from a purely deterministic classification system to probability based methods of design, including analytical design methods as an alternative to the standard fire resistance test, requires improvement and extension of the concepts outlined, as well as extensive calibration to existing code requirements (6, 7, 10, 11, 14, 15, 16). properties fire compartment

fire load

1

£

equivalent time of fire exposure (t t )

uncertainties in heat exposure!/,)

assessment of frequency ( 7nl )

safety considerations ( T ^ )

design equivalent time of fire exposure

Fig. IV. Heat exposure model Ho including assessment of frequency, safety considerations and uncertainties in the heat exposure model.

-21-

5. LEGAL IMPLICATIONS ASSOCIATED WITH THE USE OF NEW DESIGN METHODS In contrast to the numerous innovations in the building industry, including new fire fighting and protection systems, it can be observed that the requirements for structural fire protection remained practically unchanged over the years (17). Requirements are legally fixed in the building regulations in terms of required fire resistance time, instead of in required safety levels. The latter would give the possibility to a more flexible compliance by structural fire protection measures in combination with alternative active measures. Moreover, the choice and use of new design procedures will be greatly influenced by the present shared competence in structural fire protection between public agencies responsible for the requirements (i.e. the heat exposure model) and the designer for proving compliance (i.e. the structural response model). The present legal situation is illustrated in Fig. V, in which the single parameter "required fire resistance time" is the key factor and practically the only way of communication between authorities and designers. Within this system the designer is only concerned with choosing structural members of the required fire resistance grade.

requirement

authorities

building regulations

required fire resistance time

compliance

desi gners

design codes fire test

Fig. V. The present legal situation with the single parameter "required fire resistance time" as the only way of communication between public agencies responsible for requirements and designers for proving compliance. As stated before, a differentiated fire engineering design allows to make a problem oriented choice of a combination of heat exposure models and structural response models, taking into consideration reduction in

-22-

structural fire resistance when alternative active measures are employed. This matter becomes increasingly important, because there is a growing use of automatic detection and extinguishing systems in industrial as well as in public buildings. Bearing in mind the present legal situation with shared competence and rigid building regulations, major progress in the implementation of new design concepts for structural fire protection can only be achieved with reformed building regulations (18, 19). With retention of the responsibility of the authorities to set general safety levels required, this involves an increasing freedom and responsibility of the designer for a practical design situation, in particular when a mix of active measures and structural fire protection is employed (19). Finally it can be observed that nationally as well as internationally, research programs have mainly been directed to the physical aspects of fire safety, i.e. heat exposure models and structural response models. Practically no research has been performed into the manner in which requirements are specified as well as to the various ways of complying with these requirements. Therefore, it is recommended that there should be a change from the present physically oriented research programs, including those sponsered by the European Community, towards policy related programs. These should include studies on functional requirements, based on specified fire safety objectives, allowing for equivalency of different design solutions, as well as studies on the legal implications of the use of new concepts of structural fire engineering design. Broad cooperation among all concerned is needed.

ACKNOWLEDGEMENT International cooperation on the development of new concepts for structural fire engineering design takes place in the Fire Committee of the Conseil International du Bâtiment (CIB/W14) (7). The author is grateful for the stimulating discussions and contributions in this committee, which certainly have Influenced the contents of this paper.

-23-

REFERENCES 1. ISO: Fire Resistance Tests - Elements of Building Construction.International Standard 834, 1975. 2. Witteveen, J.: Policies for Fire Safety. Proceedings of CIB-TNOSymposium 'Fire Safety in Buildings', Amsterdam 1977. CIB-Proceedings No 48. 3. Pettersson, 0., and S.E. Magnusson,: Fire Tests Methods, Background, Philosophy, Trends and Future Needs. Doc. Gen. Oil NORDTEST, Stockholm 1977. 4. Witteveen, J. and Twilt, L.: A Critical View on the Results of Standard Fire Resistance Tests on Steel Columns. Fire Safety Journal, Vol. 4, No. 4, 1982. 5. Grandjean, G., Grimault, J.P. and Petit, L. : Determination de la durée au feu des profile remplis de béton. Convention 7210 SA/3/302, Commission des Communautés Européennes Recherche Technique Acier, Bruxelles 1980. 6. Witteveen, J., A Systematic Approach Towards Improved Methods of Structural Fire Engineering Design. Proceedings 6th International Fire Protection Seminar, organized by VFDB, Karlsruhe 1982. 7. CIB/W14: A conceptional Approach towards a Probability Based Design Guide on Structural Fire Safety.: Fire Safety Journal, Vol. 6, no 1, 1983. 8. ECCS: European Recommendations for the Fire Safety of Steel Structures. Elseviers Scientific Publishing Company, 1983. A summary is published in: Witteveen, J.: Steelstructures exposed to the standard fire, an introduction to the European recommendations. ASCE Spring Convention, New York, 1981, Preprint 81-035. 9. CEB: Design of Concrete Structures for Fire Resistance. 1982.

-24-

10. Bub, H. et al: Baulicher Brandschutz. Institut für Bautechnik, Berlin 1979. 11. DIN 18230 Entwurf: Baulicher Brandschutz Im Industriebau. August 1978. 12. Pettersson, 0., Magnusson, S.E., and Thor. J.,: Fire Engineering Design of Steel Structures. Swedish Institute of Steel Construction, Publication 50, Stockholm 1976.

13. Law, M.: Design Guide for Fire Safety of Bare Exterior Structural Steel. Ove Arup & Partners, London 1977. 14. Pettersson, 0., and Witteveen,J.: On the Fire Resistance of Structural Steel Elements, derived from Standard Fire Resistance Tests or by Calculation. Fire Safety Journal 2 (1979/1980).

15. Brozzetti, J., Law, M., Pettersson, 0. and Witteveen, J.: Safety Concept and design for fire resistance of steel structures. IABSE­ SURVEY S­22/83, Zürich 1983.

16. Brozzetti, J., Law, Μ., Pettersson. 0. and Witteveen J.: Fire protection of steel structures. Examples of application, IABSE Proceedings, P­61/83, Zürich 1983.

17. Ehm, H.: Brandschutzanforderungen im Wandel wirtschaftlicher Randbedingungen. Brandverhalten von Stahl und Stahlverbund­ konstruktionen, Statusseminar 1983, Studiengesellschaft für Anwendungstechnik von Eisen und Stahl, Köln, Verlag TUV Rheinland 1983. 18. Behets, J., Law, M., Study of research into the behaviour of structural steel elements exposed to fire. Centro Belgo­Luxembourgeous d'Information de l'Acier, Brussels and Ove Arup & Partners, London, 1981.

19. Witteveen, J.: Neue Wege in baulichen Brandschutz; eine Synthese zwischen Anforderung und Nachweis, Brandverhalten von Stahl und Stahl­ verbundkonstruktionen, Statusseminar 1983, Studiengesellschaft für Anwendungstechnik von Eisen und Stahl, Köln, Verlag TUV Rheinland 1983.

­25­

FIRE DEVELOPMENT AND SAFETY REQUIREMENTS IN CCMbTKUL'l'iqN

Prof. Dr. I n g . Wolfram KLINCSCT Berçishe

ttiiversity

o f Wuppertal

SYNOPSIS

The development of a fire can b e Influenced by careful attention to the physical laws on which fire development depends. It is possible to influence architectural and planning aspects, and features of building utilisation and static or constructive building development. Both intensity and duration of fires can become controllable, and injuries and damage minimised. A prerequisite for this is the application of integrated fireengineering. By using such procedures, the safety requirements can be considered alongside economic aspects. There are many suitable methods of fire protection suitable for use in steel structures, to give a high degree of fire safety economically. Together with traditional methods of fire protection, section insulation offers new developments in regard to the inclusion of whole building analysis and safety theory. The advantage of these developments is particularly apparent in that many building requirements cam be met simultaneously. Fire safety is no longer an isolated procedure, but is integrated in the overall building design.

-26-

1. INTRODUCTION Fires involving personal injury or large fires causing substantial damage to property are inevitably at the centre of public interest. Both the necessity and uncertainty of adequate safety precautions are highlighted in such cases. In regard to injury statistics, many less spectacular fires have a significant effect. The objective of fire engineering should thus be to make use of all available data. The progress of a destructive fire and its effects are considered by many dsigners and users of buildings to be not capable of prediction. The development of a fire and its effects on the supporting structure are, however subject to physical laws which enable a relatively accurate prediction to be made. It thus becomes possible to use numerical simulation of the physical factors to predict the rate of development and the intensity of a fire in quantitative terms and to estimate the damage likely to occur to buildings. Variations in building design, from the architectural design stage to the production of the support structure can thus be introduced in a process of optimisation of fire protection. 2.

FIRE DEVELOPMENT

The majority of destructive fires begin with a small localised fire, which is insignificant in regard to its effects on the supporting structure and which can be easily resisted in normal circumstances. Only if these primary fire—fighting measures, active as in the use of fire-extinguishers or passive as a result of local conditions, can the development of a large fire occur. During the course of a small localised fire, the temperature effect on the surroundings is relatively small. However, the hot gases which rise, and the flammable gases produced by pyrolysis but not ignited, collect in localised areas

-27-

such as the interior vertical effect of a effect, particularly During the life of a

of rooms or workshops. To this developing fire is added a horizontal the component of radiant heat. relatively low-risk localised small

fire the risk builds up increasingly. A critical condition may then arise, e.g. the attainment of ingition temperature of material in the roof area or adjacent storage areas, leading to spontaneous spreading of the former small fire over a large area. This spontaneous transition from a small fire to a full scale fire is termed "flashover". The features of a small localised fire are termed "preflashover" and those of a full scale fire as "post-flashover", Fig. 1 illustrates the course of fire development. It is in the post-flashover full scale fire that damage to buildings occurs. The main features of such a fire are shown in fig. 1, the main elements being a rapid fire development phase, the attainment of a peak temperature for a period that varies, and the decay of the fire.

|TfCl -*·

«*

-post flashover

pr« floshover-

"Flaahpoint

t (mm )

small local fire

.full fire

Fig. 1 : Characteristic phases of fire development

-28-

The time-temperature characteristics of a natural fire are mainly controlled by two parameters- fire load and ventilation. The term "fire load" describes the nature and quantity of flammable material enclosed within the fire zone, and its distribution and fire behaviour. The term "ventilation" describes the availability of air (oxygen), in the fire zone including geometrical ratios. In comparison with both the above parameters, the thermal properties of the material in the surrounding structure are .of secondary importance. However, consideration of these parameters is necessary for a realistic appraisal of fire behaviour, particularly in the construction of steel buildings as explained in section 3. The term "fire load*1 describes the total energy evolved on complete combustion. Fire load thus depends on the nature and quantity of combustible material. In practice, the fire load is related to the surface area considered, e.g. to the base area A. On the basis of extensive international statistical evaluation of fire load distribution, the fire load value can be determined for various applications such as schools, offices, hospitals, hotels, dwellings etc. As fire development is dependent on both main parameters, fire load and ventilation, various fire characteristics can be developed although there is one constant parameter. Fig. 2 shows various fire behaviour patterns for constant values of fire load q. These significant differences in time-temperature development are entirely due to variations in ventilation conditions V. With adequate ventilation sufficient oxygen is available for optimum combustion, and the maximum temperature and duration of the full fire will thus be dependent on fire load. With restricted ventilación, the high temperatures cannot be attained and heat output is reduced, with a corresponding increase in the duration of the fire. In this type of fire the characteristic features are determined by the ventilation. As shown in fig. 2, various behaviour patterns are possible for

-29-

' q ■ const V.A/h/A,

mi

^v

/ /

r

¿>

V1> V|>V,

\ V,

*

Fig. 2 : Ventilation­controlled fire. Constant fire­load, q .

\

vs t (atotl

Fig. 3 : Fire­load controlled fire.

Ventilation V constant.

Igyl­MJ/i 1 TfCI

[kg/«')

A/h/A t «157­10'V

Fig. 4 : ISO standard fires and natural fires. (V ­ constant ; after (9))

­30­

different values of ventilation: the transition from a fire load controlled fire to a ventilation controlled fire shows an abrupt change from an intensive fire to a longer lasting fire at lower temperatures. Fig. 3 shows the change in fire characteristics for different fires at constant ventilation. It is clear that all the fires investigated behave in a similar way, at least qualitatively, in that the ventilation enables optimum combustion of the maximum fire load. With smaller fire loads, i.e. lower energy liberation, the intensive mixing of cold fresh air inhibits the development of higher temperature peaks and leads to a rapid consumption of the available fire load. Figs. 2 and 3 show that only the two main parameters, fire load g and ventilation v, cause a wide variation in the expected behaviour of a natural fire. As the classification of fire behaviour of structural components and the definition of protective measures cannot be achieved in a universally acceptable form, the international regulations were formulated on the basis of a standard fire. In fig. 4 the ISO standard fire is compared with various fire load controlled natural fires. Extremely high fire loads can cause transient temperature peaks exceeding the value for the ISO fire. For fire loads normally encountered in building structures and the corresponding ventilation ratios, this effect is normally only expected in he first few minutes of a fire; ventilation controlled fires are normally below the ISO curve for normal fire loads.

3. SAFETY REQUIREMENTS. The design of a building incorporating fire protection technology is primarily linked to two criteria:

-31-

-protection of personnel -protection of property. In is by of

and

the case of protection of personnel, the first requirement to guarantee the rapid evacuation in case of fire, the provision of safe escape routes. For the protection property the prime need is for a fire restrictor.

The common factor between the two protection aspects is the feasibilty of a rapid and accurate attack by extinguisher. The various effects of fire on personal injury and damage to property may be broadly classified (with some overlap) as shown in fig. 5. In the case of property damage it is also possible to differentiate between damage occurring during the fire itself and latent effects. Personal Injury Oxygen starvation Toxic gases Smoke Heat Corrosive gases Fig. 5.

X X X X X

Property Damage

X X X

Assignment of fire effects.

Personal injuries in fires are not normally caused by collapse of buildings, but result from the effects of smoke, toxic gases and oxygen starvation. Preventive measures against this can be taken at the design stage, including the requirement for smokefree escape routes and the selection of suitable building materials. In the case of structural damage, the immediate effects result from the action of high temperatures on the building.

-32-

By reduction of material properties, increasing exposure to the heat of the fire causes reduction of load bearing capacity and increased distortion. The distortion caused by heat can lead to stresses between two connecting structural members and thus makes further demands on the already reduced load bearing capacity. Increasing deformation of members can also jeopardise the closing of rooms and thus the requirement to form a fire boundary: similar consequences arise from excessive surface heating of structural members away from the fire area, and here there is danger of renewed ignition. Failure of structural members is expected in the case of fire if the effective load bearing capacity is exceeded. The safety margin between the stresses under working conditions and the "cold" loading limit determines the time taken to reach the critical temperature at which the load bearing capacity is exceeded due to the effects of heat. The indication of a critical temperature is thus linked to the design of a permissible safety rating. Fig. 6 shows the example of variation of normal strength bending moment load bearing capacity of a steel beam with increased heating.

♦ NU(T)/NU|TO

0

0.1 0.2 0,3 0A

0,5 0,6 0.7 0 Í

0,9 1,0

Fig. 6 : Variation of load bearing capacity of rolled steel sections

­33­

If the fire protection requirements of a building are restricted to load bearing capacity, then there will b e a distinct difference in behaviour depending on fire characteristics. Fig. 7 shows the variation between results of a fire load controlled natural fixe (qL and the standard ISO fire. The load bearing capacity R of a structural member changes with heating in relation to the rise in temperature T. Thus for steel structures it is accepted that in the decaying stages of the fire the properties will, to a first approximation, b e restored. This can lead to a full restoration of load bearing capacity after cooling, but it must b e remembered that there may b e some restriction on serviceability due to the effects of residual distortion. The failure of a structural member can occur during a fire if the load bearing capacity

Fig. 7 : Interaction between - fire development (T) - member load bearina capacity (R) and - building behaviour (F)

-34-

MO 120 — JO0­

ρ

F»0

— eo 60

40

FM

.20 "H>

20 30

50 TO »0 150 200 300 500 700 — U/A Im­'l

Fig. β : Influence of isulation on cross section heating behaviour, variation of failure time t as function of load depending failure temperature. (

I J unprotected

Θ

Insulated (25 itm vermitecta)

R becomes lower than the value of the load F.

For the

ISO fire the failure of members is always expected, whereas it is possible for a natural fire to die down before the critical condition is reached. In complex support structures the constraining stresses can lead to an increase or a decrease in the planned stress ratings. The failure time can be reduced or may b e extended indefinitely

­35­

The smoke gases originating from a fire can,in addition to their toxic effects, have detrimental effects on materials. For steel structures the effects of the corrosive gas component are particularly important: they are frequently produced by thermal decomposition of decorating materials. As a rule, this type of damage causes no immediate reduction in load bearing capacity, but remains as a cause for later concern. Restoration procedures in the form of passivation are feasible in principle, but the cost of the damage remains comparable with that of member failure. In this case the ventilation ratios in buildings during a fire can only have a limited preventive effect, and the effect on damage is similarly limited. In regard to safety requirements this type of fire-related damage is subjet to control only in the planning stage, as the use of suitable materials and the provision of smoke control can b e arranged. The safety requirements of a building in relation to load bearing capacity can thus be affected by: -size of load, -static system, -probable fire development. The traditional fire safety requirement for structural members and buildings assume that breakdown will occur after a certain time of exposure to fire. This follows from the use of ISO fire development as a b a s i s . The transition to natural fires can lead to a new concept of safety requirements: object protection. The safety requirements given below guarantee that no building collapse will occur in the most critical fires to be expected, although the restoration to full serviceability may be restricted.

-36-

These object protection requirements can be of considerable significance for specific structural components or for buildings with outstanding function. The realisation of broader safety requirements in these buildings assumes the integrated consideration of all components concerned and is normally only attainable by iterative optimisation of the design of the whole supporting structure. The technical safety requirements for buildings can be divided into 3 groups, for which the next higher group enables a realistic calculation of safety standards, normally linked to economic design, -single component calculation for ISO fires, -whole-building calculation for natural fire development, -object protection planning. 4. INTERACTION BL'fWhlM PIBE OCCURRENCE AND BUHJING BEHAVIOUR The collapse of a building or component due to fire is always attributable to direct or indirect heating of structural members. Direct effects accrue from reduction in strength of materials, with their effects on load bearing capacity. Damage due to the effects of corrosive gases must be considered in this category. Indirect fire damage results from an increase in stresses in building components as a result of prevention of free thermal expansion: damage caused in this way can also affect members that are not directly affected by fire and thus have not been subjected to the effects of high temperature. For the direct effects of fire in buildings, 'failure time and failure temperature can be related. As the majority of sections used in steel construction have low cross sectional m a s s , i.e. have a large section factor U/A, no significant temperature gradients are established and one can refer to isothermal cross sections in buildings. The collapse temperature is dependent on the loading of structural members. Thus, in addition to the customary equivalent safety value, the possible change of loading conditions during fires due to constraint also has its -37-

effect on collapse temperature, which, can be positive or negative. (See fig. 7.) Fig. β clarifies this relationship for protected and unprotected steel sections. The change of collapse temperature is consistent with the change in load. From the shape of the curves it is apparent that for unprotected members a reduction in load usually causes only a negligible increase in collapse time, at least for the case of an ISO fire. On the other hand, insulation of members can have a considerable effect. In practice, the transition from single members to members simply connected, e.g. continuous beams spanning two areas, is accompanied by a redistribution of shear forces which has advantages in regard to the fire rating. When the building safety analysis relates to natural fire development instead of to the ISO fire, a wider range of parameters is observed in accordance with the above-mentioned effects on collapse time. These parameters mainly affect the rate of fire development. The dominant effects of ventilation and fire load have already been demonstrated in point 1. Reference has already been made to the effects of thermal properties of the materials used on the change of temperature in a fire area. The material properties of the structural members surrounding a fire area can affect the fire development in a fire zone by heat transfer. Good insulating materials retard the heat flow through the enclosing walls and thus cause a rise in fire space temperature provided the material itself has adequate fire resistance. It is, however, true that many inferior heat-checking materials, or materials with inadequate fire resistance, are in common use. In the past, the significance of the fire stability of wall materials, or of the insulation materials used, has not received adequate attention. This can lead to an unrealistic calculation, as it gives an excessive temperature. The full consideration of this factor leads to a far more progressive consideration of the interaction between fire development and building behaviour. -38-

As in the case of the ISO fixe it is taken for granted that temperature rise is almost inevitable, it is still normal to consider only the irreversible effects of fire on the structure when using traditional methods of analysis. The transition to natural fires broadens this method of examination 'to an interaction betwen fire occurrence and the buiding structure. With this method of calculation it is, however, still assumed that the structure of a building remains unchanged or that the change is expressed by the collapse temperature. In fire protection technology planning of complete structures this can lead to an unrealistic calculation which acts as a deterrent to optimisation of building planning. The collapse of individual members of a complete structure may not have a controlling effect on the safety of buildings, but could equally have a significant effect on fire development by changing the ventilation conditions. Only the consequent calculation of this reaction of building behaviour to fire development leads to a realistic description of the interaction between fire behaviour and building behaviour. The respective steps in continuous fire engineering are described in U , 2, 3..). 5. SUMMRKy Conventional calculations on ISO fires do not show a correlation between fire occurrence and building behaviour. This method of calculation undoubtedly has the advantage of a relatively simple application and also provides a basis for comparability and reproducibility in fire investigation and subsequent classification. This advantage is at least partially lost on transferring to natural fires. The procedures for calculation are expensive and the results obtained are only valid for the specific case involved, and cannot be used for general application. On the other hand, a more economic building design can emerge from such realistic calculations without reduction in safety standards. Application of the method opens up various possibilities for arriving'at an optimised solution. There are possible effects in regard to both planning and construction. -39-

Both fire load and ventilation can b e influenced at the planning stage. Ventilation can usually b e modified by room geometry, and in industry supplementary measures such as smoke- and heat blocking equipment, fire curtains etc.. The fire load can b e varied according to the building materials used for the load bearing members and the stock involved. The constructive possibilities of affecting fire behaviour involve static, structural and probabilistic components. The choice of building material, or combinations of materials, and the choice of construction method and static system enable the desired effect of achieving a practicable failure time. The arrangement of a specified breaking point can thus be an effective criterion for varying the ventilation method in case of fire and for controlling the failure mechanism, e.g. by fire walls in industrial premises. New structural member developments offer additional possibilities for decorative uses of steel and for guaranteeing load bearing capacity for definite fire times. New developments in this field also include compund building methods and water cooled components (4 , 5, 6, 7, 8 ) . Fire development and fire behaviour of a building are not factors that are incapable of being influenced, and even the probability of a full scale fire can be influenced. The effect of fire can b e included in the original design, and the probable extent of damage calculated. Optimisation of fire protection techniques leads to economic solutions and can integrate completely the various requirements relating to use, planning and construction. The instrumentation available enables a comprehensive evaluation and interpretation of the objectives and requirements of fire safety technology.

-40-

6. LITERATURE: 1.

Petterson, 0: "Fire exposure", Manual on the fire safety of structures (Chapter 2), ECSC-T3 (in preparation).

2.

Witteveen, J.:"new procedures of building fire-safety" Status seminar, Cologne, 1983.

3.

Hönig, 0., et al.,"Safety analysis of fire safety requirements for buildings. Interaction between fire occurrence and building behaviour. Forschungsvorhaben BMFT Bau 6004, Studiengesellschaft P86-3.4/3.5.

4.

Witte, H. "Water cooling for the fire protection of buildings. "Acier-stahl-steel, 4/1981."

5. . Hönig, 0 et al."Fire-safety of steel clumns by water filling and circulation. Status Seminar, Cologne, 19 83. 6.

Schleich, J.B. et al., "A new technology for fire proof steel construction." Acier-stahl-steel, 3/1983.

7.

Kordina, K., W. Klingsch.: "Fire resistance of composite columns and solid steel columns." Acier-stahl-steel, 2/1984.

8.

Kordina, K., W. Klingsch.: "Fire hehaviour of compund columns and solid columns." Forschingsbericht P35/EGKS 7210-SA 1-108. Düsseldorf, 1984.

9.

Arnault, P., et al. "Experimental report on tests with natural fires made in the small plant, Maizieres-les-Metz." Document CECM-3/73-11-F, 1973.

-41-

ECSC CONTRIBUTION IN THE FIELD OF FIRE-SAFETY OF STEEL STRUCTURES.

M. DESCUDE, Industrial adviser, president of Executive Committee on properties and behaviour in service of light structures. In recent years, 3 million ECU's have been devoted to research on the behaviour of steels used for structures subjected to incendiary conditions. Positive results were obtained and practical methods of application developed. The safety regulations and insurance conditions should take account of these favourable observations. 1.MOTIVATION OF RESEARCH. Since it was first discovered, fire has been uppermost in man's preoccupation, because of the dangers it presents to people and property. The development of city life has reinforced this fear, to a time when the configuration of towns and the material used in their construction were responsible for the destruction of several of them: the fires of Constantinople, London and New York are recorded in History. Such disasters do not occur nowadays, but even localised fires can have very serious consequences. . During the last ten y e a r s , thirty fires have been reported world-wide, involving the loss of 2000 human lives and concerning very different buildings: large stores, leisure centres, residential and office property and hospital buildings. The protection of both people and property against fire thus remains an ongoing problem, not only on humanitarian and economic grounds, but also because of the psychological impact of the media in regard to such events.

-42-

This naturally led to the formulation of very conservative regulations for urban development and construction. From the legal aspect, the concept of responsibility has also been particularly stressed, and the institution in several countries of construction insurance independent of fire insurance is an example. It may thus be said that the technical solutions involved in the design and construction of buildings should be adapted to the optimisation of safety conditions. The diversity of materials, and possibilities of combination, the architectural needs and the cost factors also present a number of requirements that can only be resolved by an in-depth knowledge of the fire behaviour of materials, in regard to their properties and service conditions. This knowledge is also indispensible in combatting the conservatism resulting from tradition, commercial preoccupation, or fear of responsibility.

Steel, a modern material with potential for progress in the building industry, is frequently confronted by these problems. Metallurgists and metal constructors have therefore undertaken research on the subject over the last few y e a r s . The research, whether individual or collective, national or international, on the fire-safety of structures, has demonstrated their efficacy thanks to ECSC. The human, industrial and economic aspects of the research objectives, their effect on the regulations and technical barriers, led the Commission to coordinate, with financial assistance, the projects of major importance. Article 5s of the Treaty of Paris, in which the general provisions cover the encouragement of research, enables development of application of material within its terms of reference.

-43-

Since 1966, three million ECU's have been devoted solely to research on fire, with the financial support of ECSC, independent of regional or private sponsorship. The main objectives of these programmes include: -the establishment of fire testing stations, to provide the community with research materials and to define the European criteria for appreciation. -general enquiries permitting the analysis of causes and actual effects of fires and classification of the research opportunities. -study of static behaviour of structures and simple or composite structural elements subjected to different types of fire. -examination of service conditions and methods of preservation, enabling steel to present optimum fire-resistance capacity. -the establishment, and communication to constructors, of recommendations and current methods of calculation for the revision and unification of European regulations. The research has been made possible by collaboration beween Universities, laboratories and specialist institutes, witn the help of numerous international experts. The investigations have not been of a theoretical, closed circuit nature, but have related to actual problems, allowing the establishment of dialogue between the relevant participants; administrators, insurance, safety services, standards organisations and research organisations. Lastly, the economic balance of the behaviour of structures exposed to fire has been a major consideration for the investigators. Before embaarking on a discussion of the technical details of the main research projects undertaken, I would like to thank all those who have contributed to their success: the Commission, Steel and Building Industries and, of course, those who have been involved in the day to day execution of complex programmes.

2.

ROLE OF TESTING STATIONS.

The intensity of a fire depends on the nature and quantity of the combustible materials at its source. It also depends on the speed of release of heat, which depends in turn on the ventilation factor. The resistance capacity of structures subjected to fire is consequently a función not only of these elements but also of their individual properties: mechanical and static properties, structural system and service load. The evaluation of fire risks, precise formulation of their conditions and the systematic study of the behaviour of metal structures in fires must be considered in specialised testing stations. By its financial contribution to the testing stations at Brunswick, Gand and Maiziere les M e t z , the Community has enabled several tests to be m a d e , forming a scientific base for practical knowledge and essential data: the effect of higher loads, establishment of coefficients of combustibility, determination of curves of temperature rise against duration of exposure. The behaviour of several elements and structures has been examined as a function of natural or standardised fire conditions. 3.

GENERAL INVESTIGATIONS.

3.1. State of knowledge and regulations regarding the behaviour of steel in fire.

-45-

This investigation w a s particularly important for the orientation of research still to b e done, and actions to be undertaken: it was also justified by the desire of metal constructors and metallurgists to: -improve their knowledge of the requirements of regulations relating to the protection of steel against fire -remove all ambiguity concerning the actual capacity of steel in relation to that of other materials -evaluate the necessary conditions for its ultimate protection. This task led to a comparison study of the situation in Western Europe, Japan, U.S.A. and Canada. The resiOts of this investigation are as follows: -regulations concerning firs material and insurance conditions are excessively conservative in Viestern Europe and do not take sufficient note of the variability of risks. These risks, and the consequences of fire, cannot b e generalised, in regard t o , for example, a swimming p o o l , an old peoples' home or a warehouse for storing products of varying combustibility. -steel, doubtless because it has only recently been introduced to the construction industry in Europe, is the object of unwarranted prejudice. The risks incurred are in fact much lower than for materials that are unstable at high temperature, with toxic products of combustion, with an unfavourable topography of premises or the absence of prevention and emergency measures. There is thus a tendency fgr oyer-dimensioning of structural elements in steel, or for exaggerated protection requirements whose cost loads the cost of the structure. -It was finally established that the majority of these spurious problems could be resolved by the development of contact with the authorities and builders, on the basis of results of work already completed. The priority accorded by the enquiry to the provision of documentation facilitating the application of these techniques is relevant in this context.

-46-

3.2. Statistical analysis of actual fire conditions in industrial buildings. This investigation arises from the conclusions of the previous project and is at present in progress. It concerns collaboration with the authorities and fire-fighting and demolition organisations, to study in situ actual fires in ground-floor industrial and commercial buildings, to establish a relationship between fire behaviour and construction, and the losses resulting from the fire. The analysis requires objective criteria, uniform for all materials : direct cost of building (internal and external), indirect costs arising from loss of use, duration of stability. It is hoped that the results of this investigation will enable the establishment of more realistic safety regulations for this type of building. The participation of the authorities and representatives of all branches of the building industry in this comprehensive analysis will reinforce, by its undoubted objectivity, the credibility of the results. 4.BEHAVIOUR OF STEEL MATERIALS IN FIRE.

LiKe all other materials, steel is subject to a reduction in its load-bearing capacity above a certain temperature. This results in increasing deformation for a given load. Several researches have been undertaken to determine the precise conditions of this change, to establish realistic rules for practical calculations, to show that steel completely satisfies the safety requirements in the presence of fire. The investigations concerned a wide range of elements and structures stressed in different ways and consider the possible use of complementary protection elements.

-47-

4.1.

Structural elements.

One particular project covered the fire resistance of metal pillars as a function of the various parameters likely to affect their behaviour: sudden loading, rate of reheating, conditions of dilatation and loading. The position of columns relative to the seat of the fire was considered, a reduction in the degree of heating having the effect of extending the fire resistance time. In regard to beams., their behaviour was determined firstly as a function of different building systems, iso-static and hyperstatic, and secondly by testing the effect of protection by experimental coatings or concrete cladding. Comparative research was also undertaken on the behaviour of different types of pillar: hot-rolled or forged, cold formed hollow or open sections. The effect was also considered of the combined effect with concrete, reinforced or otherwise, used either as an internal core for hollow and open sections, or as an external cladding for h o t rolled sections. In connection with claddings, research was also undertaken on the strength of thin cold-rolled sections protected by various types of fire-resistant material , applied direct or in plate form. Finally, a current project concerns the study of protection of certain sections by special applied coatings, allowing reconciliation of the increase in durability and the economic advantages resulting from reduction of the thickness of applied coating. 4.2. Structures. Research on the fire behaviour of metal structures has naturally examined their behaviour in relation to both small scale models or full size structures.

-48-

Model studies included the case of gantries, with or without wind-bracing. An example of the full scale work relates to a tetrahedral structure with a tubular profile, with protection by a suspended fire screen. A particularly original experiment concerned the IRSID building at Düsseldorf. This three storey building, comprising a steel space frame, is actually specially designed to improve its fire resistance. The exterior columns are in the form of continuous hollow sections which form a circulation channel for cooling water, with provision for regulation.

Internal columns

are protected by asbestos cement coatings. floors rest on steel beams.

The concrete

Finally, the elements of

the facade are slightly behind the external columns. Tests have been made on the efficacy of the system, using a fire chamber specially designed for the purpose. The results were highly informative. 5.CALCULATION OP FIRE RESISTANCE. The various research projects udertaken on the behaviour of steel under fire have revealed a good correlation between the results of theoretical studies and practical tests. This agreement h a s , in many cases, enabled the development, or modification, of mathematical models designed for simulation, analysis or prediction. Thus, in several countries of the Community, information programmes have been initiated to facilitate the practical application of recommendations

and building regulations.

Another project aims to establish an information programme for calculation of the fire-resistance of buildings incorporating steel and steel-concrete structures.

-49-

6. RESEARCH AND DEVELOPMENT. The value of the objectives, the extent and diversity of research that has been taken to a successful conclusion, the pursuit of investigations for evaluation demonstrate the extent of the contributions of the Community, the steel industry and the steel building industry in the field of safety of steel structures. At the present time there is extensive knowledge available of the causes and consequences of fires, of the effect of fire on the behaviour of material and its durability, and on the possible methods of effective and economic protection. Some lessons may also be learned: -the favourable behaviour of steel structures subjected to fire goes well beyond the times required by the safety regulations for material properties. -It will also be possible to ensure that these regulations take into account the results obtained from scientific and practical research. -the notions of cost and risk of fires are more a function of the environment, content of buildings and prevention measures in use rather that the nature of the structures themselves, in that their capacity has been conveniently calculated. The insurance regulations should therefore be related to more selective conceptions of risk, with a possibly more equitable appreciation of the potential of the material.

-50-

From the overall balance point of view, steel Is competitive with other materials subjected to fire. On the technical front, it has some specific advantages and it may be perfectly integrated in composite systems from the safety aspect. On the economic front, its cost is favourable so long as it is not loaded by superfluous protection. It may thus be said that the application of steel in the building industry is viable, both in regard to its potential for prevention and simplification of the problems arising from damage to buildings. The time is therefore ripe to demonstrate to these who have been scornful of steel that it is a safe material with predictable behaviour and service conditions in the presence of fire. This Conference will provide an admirable opportunity to demonstrate the future potential of steel, by discussing methods of calculation, provision of data and circulation of appropriate safety documents.

-51-

OPENDE SESSION Siimiary of discussions Rapporteur: Jean FERRON P r i n c i p a l EEC A d m i n i s t r a t o r R e s p o n s i b l e f o r " P r o p e r i e s and S e r v i c e b e h a v i o u r of s t e e l s " . In welcoming delegates to the Conference, M r . Evans expressed regret concerning the recent sudden death of M r . Carpena, technical secretary of the European Convention for Metallic Construction, which w a s felt by a l l . Mr. Carpena worked actively with the EEC, particularly in the field of EUROCODES and t h e use of steel in construction.

This first session, devoted to the opening addresses, enabled the objectives of the Conference to b e defined, i.e.: to present, with the objective of sharing, methods and recommendations that enable at the present time a practical and viable programme for the study and construction of buildings with fire resistant steel structures, resulting from satisfactory solutions in relation to economy, architecture and safety. To this end, the speakers referred to the scientific knowledge, techniques and technology acquired in the course of development in the field of fire safety applied to steel structures. The need for positive collaboration between the authorities responsible for regulations, safety and fire fighting on the one hand, and insurers, investors, designers and constructors on the other, w a s particularly stressed. Cost aspects were also discussed, as w a s the consideration of the notion of "real fire" in the evaluation of risk, taking note of the specific conditions of use of buildings.

-52-

Mr.Kruppa- CTICM, Paris. Fire safety extends over a global study of the problem and it is for this reason that w e support the position of Professor Klingsch regarding, for example, the use of roof apertures for evacuation of heat and fumes. As Monsieur Descude also mentioned in his paper, an investigation of fires in industrial buildings is at present in progress. Preliminary analysis suggests that for buildings with metal structures having coverings with very low fire resistance the structure itself behaves w e l l . In his paper, Professor Klingsch also made reference to fires in industrial situations. We have made several different tests with a large quantity of expanded polystyrene, in a situation having a large volume (10 000m3. We have then established tnat the temperatures attained are very low (J50°C) compared with those referred to in the ISO curve, thus not endangering the stability of the structure. Commander Muller, Paris Fire Brigade. I would like to comment on Monsieur Witteveen 1 s paper. If the fire resistance time is the only criterion used by designers, then it is difficult to appreciate and to determine an acceptable safety level for each type of occupation of buildings. I would add that whilst there are obvious attractions in using dynamic safety measures (detectors, sprinklers), we must not overlook the fact that such systems are not always 100% reliable, due, for example, to lack of maintenance or budget considerations. In short, w e are in favour of: -the provision of static compensatory equipment, such as large size outlets, -the limitation of surface of the compartments, -the increase in height of levels, -the limitation of fire load.

-53-

Professor Witteveen - TVO - Delft. The preceeding remarks amply illustrate that the important point is that of communication between the authorities and the designer. More dialogue is required between the parties concerned. But with the present rigid regulations based on a single parameter this dialogue cannot be established. More flexibility is required and w e must strive to achieve it by convincing the authorities of the validity of our tests. These ideas w i l l b e developed during the course of the next two d a y s . M r . Meskens - a Brussels Architect. I have listened with interest to M r . Tent's paper and I should like to stress the fact that in regard to fire safety it is Man w h o introduced fire to his h o m e , for light, heat and comfort, without undue regard for the risks involved. Some famous buildings have been destroyed by fire. That is why it is necessary to pursue our efforts in regard to safety development with the use of all modern means at our disposal, such as micro-processors to enable early detection of the parameters that can encourage fire. Ing. De Martino - Nuova Italsider - Genova. Fire behaviour is an element that enters into the overall design of a building. All the components contribute to safety - structure, floors, w a l l s , partitions and thus ensure a uniform stability with time. Mr. Favre - Regional Insurance Establishment, Berne, Switzerland. In Switzerland it is the regional establishments that determine the conditions to be observed for fire safety. That is made by a general approach with the help of the fire prevention services, the Swiss Society of Engineers and Architects and the groups working on the respective

-54-

materials, wood, concrete and steel. Into tnls global approach the Idea of evaluation of fire risk has been Introduced, balancing active and passive measures. The rule will be to determine a fire load and dimensions of apertures that determine the risk (equivalent temperature) rather than to fix a time of 30 or 90 minutes. Mr. Kruppa -CTICM - France. It has also been established that the present regulation system based on a unique and standardised type of fire is very restrictive and limits architectural expression and at the same time increases the cost of the building. A study group has recently been formed at the instigation of the Minister of the Interior, to study how it will be possible to introduce real fires into the regulation system. It is a long task, but it has been established that a slow evolution is in progress in the direction of a probabilistic approach to fire. Mr. Hammer -Federation of Insurers. The primary objectives in the different countries are admittedly quite conservative. But by actual tests one can make a realistic appraisal of risk;in regard to steel in particular the most recent information has been used. If metal construction should contribute, in the future, to the reduction of risk, the primary directives will certainly incorporate this, depending of course on the type of utilisation of the building concerned. Mr. Bonqard - Stahlbau-Verband - Cologne. It is important to differentiate between the risk to the shell, the building, and the risk due to the contents = a good example is that of a single level building. The premiums must take account of this.

-55-

Mr. Hammer. That is actually a subject what will be discussed during this Conference,

particularly in Session I I .

Dr. Kersken-Bradley -Building Institute - Berlin. It should be noted that many current standards are still based on ideas of an age when the state of knowledge regarding design and construction of buildings and the development of fires was very limited. It is necessary to continue to improve them with the help of results now available - involving, for example, the ductility of steel.

-56-

SESSION

I

:

DESIGN

METHODS

Presentation of european recommendations for the f i r e safety of steel structures Examples of calculation of f i r e resistance of the steel members according to the european recommendations of ECCS TC3 Fire safety, design of composite columns Fire engineering design of composite concrete slabs with profilled steel sheet Fire resistance of composite steel decks, floors and beams Report on Session I

-57-

PRESENTATION OF EUROPEAN RECCMMfcNDATIONS POR THE FIRE SAFETSf OF STEEL STRUCTURES D r . J . KRUPPA

Head of fire service at the Metal Structures Technical Industrial Centre and Member of Commission 3 of the European Convention on Metal Structures. INTRODUCTION. The European recommendations for calculation of the resistance of steel structural elements exposed to standardised fires (1) have been compiled by Commission 3 "Fire Safety of Steel structures" of the European Convention on Metal Structures. They are the result of several years of study and research undertaken in various European laboratories ( 2 , 3 , 4 ) .

The primary objective of the recommendations is t o provide the user of steel structures in building with a simple tool for calculation that w i l l enable h i m to justify the duration of stability of his structures in fire, and t o prove that they comply with regulations. In comparison with Laboratory tests, this procedure is far more rapid, cheaper, and can take precise account of the various factors affecting fire resistance. A s a result, solutions may b e proposed that are optimal in regard to their compliance with safety requirements and minimisation of building costs. Witn these recommendations the builder no longer submits, he acts. A further objective is to demonstrate that the extent of current knowledge of the behaviour of steel structures to

-58-

fire is such that it is possible to offer safety levels comparable in all respects, or even superior, to those achieved by traditional structures.

In fact, the stability of metal frames in fire

is such that it is easy to obtain a fire resistance of 2 hours, or even 4 hours with standard fires.

In most cases., standards and national fire safety regulations relate to standard fires, and thus have generally satisfactory safety coefficients. The European recommendations are thus voluntarily limited to the presentation of a mathematical model having the sole objective of achieving results identical to those that would be obtained by testing structural elements in a furnace. The present state of knowledge , which w i l l be discussed in the final session of this symposium/ offers the possibility of studying the overall behaviour of a steel building subjected to any type of fire. When the regulations have been extended to methods of risk analysis enabling a consideration of the overall safety of a building, it will b e easy to modify the European recommendations to provide a method of appreciation of the stability of the whole. Regarding the present state of the European regulations, a working manual (5) h a s been prepared by Commission 3.

PRINCIPAL HYPOTHESES. The hypotheses on which the European Recommendatopns are based are: -fire resistance of individual structural elements (beam or column) subjected to an ambient temperature rise defined by standard ISO R 834 ( 6 ) . -in each element uniform temperature is assumed over

-59-

the entire section and length. -creep, the effect of which becomes significant above 400°C, implicitly included in relations behaviour with- temperature

linking mechanical

(fig. 1 ) .

20 "C

(i)

Fig. 1.

Stress-strain curves for mild steel at various temperatures. The last two hypotheses have made it possible to split the calculation of fire stability of a steel structural element into two independent parts: -calculation of the temperature attained by the element after a given time of exposure to a standardised fire, -calculation of the critical temperature, i.e. the limiting temperature above which the element is in danger of being unable to support the loads applied and thus of collapsing. Comparison of the two values thus determined indicates whether or not the structure will have the required stability Fig. 2 represents the change in temperature of an element with time. At 60 min., the temperature is of the order

-60-

of 55ûeC. If the critical temperature is 500°C, the element will not have a stability period of 1 hour under standardised fire conditions, and will be in danger of collapsing after about 54 minutes. If the critical temperature is 600 C, the element will have a stabilityof 1 hour, as there will only be danger of collapse after 65 minutes. It is in fact desirable that the temperature attained at the stability time required by the regulations or standards should be equal to the critical temperature. The European recommendations enable this objective to be achieved.

600

/Λ

550 500

α ε

I

/

400

1

ι I

1 1 1

200

1 1

, 20

40

ii­ ii 54

60 65

Time (Min) .

Fig. 2. Change in temperature of a structural element with time: comparison with critical temperature. HEATING OF STRUCTURAL ELEMENTS. The change in temperature in a metal element subjected to fire is a function of: (a) conditions of heat transfer by radiation and convection (ltt 3.2.1)*: from numerous tests in furnaces an expression * The relevant paragraph of manual or recommendations is given

­61­

has been found t o estimate the flux transmitted per unit of surface (Q) :

where © t = furnace temperature; θ = surface temperature of element.

(b)­the ratio between the heated contour of the element, by which the heat exchange occurs between the ambient medium and the steel section being heated. This ratio is termed the massivity factor ( F / V ) . It plays an important role, and it is obvious that the three sections in fig. 3 will not have similar heating characteristics.

> f

I

J V.

HEM 300

HEA 200

F/V = 60 m­1 Pig. 3.

F/V = 211 m­1

IPE 100 F/V = 389 m­1

Steel sections with different massivities.

The increase in temperature ΔΦβ during a time interval ¿ t is given by (1, 3.3.1) :

1

Á9

S ­ Q .

—

At c

s's

where

Cg

= specific heat of steel

ps

= density of steel.

­62­

With this formula, experimentally verified, it is possible to calculate the temperatures attained after k h and *¡ h exposure to standardised fire as a function of massivity factor (fig. 4) (5 ­chapter V )

These data may b e represented in a different form, indicating the time required for the steel to reach a given temperature, fig. 5., or to provide an approximate mathematical expression (5, chapter III) / F \ *0,6 t ­ 0,54 (*s ­ 50) Í — )

[min]

900 G"

30 Bin

800

(Ρ V *i

a

o 3 «

500

I

400

i­

300 200 100

«0

80

120

160

200

240

280

MASSIVITY FA CTOR

Fig. 4.

320

360

­ C"1)

Temperature attained by sections of different massivity after 15 and 30 min. in standard fire.

­63­

80 c •H

E ­ 60

S 20

250 F/V (m­1) 50 100 150 200 Fig. 5. Time required for steel sections of different massivity to reach temperatures of between 500 and 600°C in standard fire.

TABLE I Temperatures attained after different exposure times to standard fire by steel sections of different massivity protected by material of thermal charcteristic : JL

Ai di/ Λ, « 0,10

= io m2 °C w­1

/ : F1/V (■

massivity factor

Time

10

20

30

50

100

150

200

250

_1

)

300

350

400

(«In)

α

20

20

20

20

20

20

20

20

20

20

20

15

32

43

54

76

126

172

212

249

283

313

341

30

48

74

99

146

247

327

393

447

493

531

564

45

65

107

146

216

353

454

529

588

634

671

701

60

83

140

192

281

445

555

633

690

733

767

793

75

102

173

237

342

523

637

714

768

807

837

860

90

120

206

279

398

591

705

779

829

865

891

910

105

139

238

320

449

649

762

832

878

910

933

949

120

15B

268

359

496

700

810

876

919

947

967

981

­64­

(c) -of trie tilermal p rotection that can be used to reduce the rate of heating of steel elements. The efficacy of this p rotection is a function of the material thickness (d^) , its thermal conductivity (Λί) and its volumetric heat ( C i , P i ) . T h e increase in temperature is then defined by (1 * 3.5.2):

3, 5

where

dţ

1 Cs

P%

c

i di Ί

F

1

V

1 ♦Í

Δθ

ί

Γ

1

F

1

2. C s P$ V

and F. = internai surface of protection.

This expression may b e simplified in the case of protection materials for which it is possible to neglect the amount or heat that they absorb (ci,#i = 0 ) . In practice, the European recommendations and their working manual give numerous tables, graphs and mathematical expressions giving values for the temperature attained by metallic elements (Table 1 and Fig. 6) (1, * 3.4.2) and (.5, chapter I I I ) :

d, » ·

« . « , . , ­ M , | J_J

»·"

M

Further, according to the arrangements specific to each country, it is possible to produce graphs specific to each protection product

(fig.7) (5, chapter I V ) .

­65­

200

100

0.001

0.002

0.003

0.004

0.005

0.006

0.007

Fig. 6 Graphical representation of tables indicating steel temperature as a function of massivity factor and properties of protection material. PRODUCT : X STABILITY TIME : 60 MIN.

Fig. 7 Typical graphs for direct determination of thickness of protection to be used as a function of required fire stability, massivity factor and permissible temperature limit for the structural element

g

700

40

80

120

160

200

240

280

320 360

F/V ( · · '

-66-

)

TEMPERATURE FOR COLLAPSE. The determination of the critical temperature of a steel structural element, i.e. the temperature at which it will collapse, depends on the theory of calculation of the extreme conditions. A direct relation has been established, for a structural element, between the ratio of applied load (Q*) to the cold crippling load (0 ) and the decrease in yield point of tne steel (oy) with temperature (θ), valid for all uniformly heated elements (1, # 5 . 3 . 2 . ) :

Qp

<*v (20 °C) 'y

The European recommendations having been fixed on the basis of laboratory measurements, a multiplication factor Κ must be introduced to cater for the increase in mechanical strength due to uneven temperature distribution in tests and to the use of steels with properties superior to those stated, parameters implicit in the results of official tests. The critical temperature is then obtained from: KQ* Ψ m

QP and the change in yield point with temperature is shown in fig. 8 (1. # 2 . 3 . 3 ) . Thus, for ψ and

for

=0.b7

ψ =0.18

the critical temperature is 450°C the critical temperature is 650°C.

It is thus established that the collapse temperature of a steel structural element is not a constant value but depends on the loading conditions and the strength of the element. The determination of ψ is made from the applied loads and the application of formulae or use of tables:

-67-

1.0

"y.« " y . 20

0,8

0.6

0.4

y. β

0.2

108

'y.20

200

Fig. 8.

400

«00

800

(i a ­

­

β/îooo) 440

1000

Steel temperature (°C) Change in yield point of mild steel with temperature

For elements in compression, such as columns, the crippling load is calculated from:

1 ♦*( * ­0,2 ) + 3 2

I./!

,2

Ncr ­ A . "y 2 0

Λ

when

> 0,2

and from Nc

r­ * ■

"yzo

Λ « 0.2

when

the various coefficients being defined in

(1*5.4.2.)

Knowing N*, the applied load on the element, we obtain:

and with K=0.85,

N* ψ « 0,85

Ncr This model has been tested by numerous experimental results (fig.9) (7 & 8 ) .

­68­

'

120 Fig.9.

160

"cr (experimental) (N/mm2) Fire stability of steel columns. Theoretical v. experi­ mental results (7).

For elements in bending such as beams, a coefficient of hyperstaticity ( θ ) , which is generally greater than unity for hyperstatically designed elements, under service loads and in the elastic region, must be introduced. The expression then becomes : ,. f · σ where a = maximum applied stress f = section factor. Values of Κ are given in Table 2. TABLE 2.

Values of coefficient k for beams. LOADING LEVEL

"¡J^

0.2

0,3

0,4

0.5

0,6

0.7

0,8

0,9

1,00

0,80

0,83

0,86

0,88

0,90

0,93

0,95

0,98

1,00

Beans with degree if hyperstaticity

0.60

0,65

0,70

0,76

0,80

0,85

0,90

0,95

1,00

Beams with degree of liyperstatlclty

0.40

0,48

0.55

0.65

0,70

0,78

0,85

0,93

1,00

ISOSTATIC BEA MS

­69­

Depending on the level of loading and the type of element, the critical temperatures for elements designed elastically. may have the values indicated in Table 3. Table 3. Critical temperature as a function of loading level and type of element, assuming section factor f = 1.15 and coefficients Κ as in Table 2.

α

Static β

■yatea

IS0STAT1C

hyper­ 1

/.,

«ty "

b.3

0.4

0.5

0.6

0.7

­

­

625

590

550

505

450

1,12

1

675

625

690

550

510

1.00

>ι

700

640

595

555

500

ι.47

1

710

660

625

595

565

740

680

630

600

560

H»PERST*TIC •L 4 ι

i­U *

i

Ϊ

1 6 stilus |UiUiJ|

1.33

The correlation between theoretical and experimental results has been verified on many occasions. For example, fig. 10 shows the good correlation between a method of calculation very similar to that in the recommendations and tests made in Germany, Japan and France ( 9 ) . In regard to assemblies with horizontal and vertical elements, tests (10) and investigations after disasters have shown that these zones have a fire resistance greater than that of the elements assembled. This is due to a greater concentration of material , which reduces the temperature of these zones and to a general over­designing of combined systems. Except for some precautions in use ( 1 * 4 . 5 ) there is thus no particular need to study the fire resistance of steel assemblies.

­70­

o

β

fri

χ

uniform variable temperature: i s o s t a t i c beams o Japanese t e s t s + German t e s t s . .French t e s t s , hyperstatic beams

800 »C Experimental temperature

Fig. 10

(°C)

Theoretical v. experimental collapse temperature for isostatic and hyoerstatic beams. (9)

Several European countries have already adopted rules for calculation very similar to these recommendations, or will soon do so. (Netherlands, Switzerland, Germany, Britain, France). Some countries have even

more sophisticated regulations,

referring to several possible fires (Sweden and Denmark) or calculation using probabilistic methods (Germany).

­71­

REFERENCES 1 ­ ECCS ­ Technical Committee 3 ­ Fire safety of steel structures. European recommendations for the fire safety of steel structures. Calculation of the fire resistance of load bearing elements and structural assemblies exposed to the standard fire. ELSEVIER ­ 1983 ­ NETHERLANDS 2 ­ Communauté européenne

du Charbon et de l'Acier.

Recherche sur la tenue au feu des constructions métalliques. Doc. EÜR 5180 F ­ Août 1974. 3 ­ European Convention for Constructional Steelwork Fire safety in constructional steelwork. Doc. CECM ­ III ­ 74 ­ 2 E ­ 1974. 4 ­ International Seminar on steel and composite elements DELFT ­ 6,7 November 1980. Fire safety journal volume 4 ­ Ne 4 ­ 1981/1982. 5 ­ ECCS ­ Technical Committee 3. Manual on the European recommendations for the fire safety of steel structures ­ 1984. 6 ­ Organisation Internationale de Normalisation Recommandation ISO R 834. 1968 ­ F. 7 ­ M. VANDAMME et J. JANSS. Buckling of axially loaded steel columns in fire conditions IABSE Proceedings ­ Ρ 43/81 ­ August 1981. 8 ­ J. KRUPPA Calcul des températures critiques des structures en acier. Revue Construction métallique n" 3 ­ 1976. 9 ­ J. KRUPPA Résistance au feu des structures métalliques en température non homogène Thèse présentée â l'INSA de Rennes ­ Juin 1977. . 10 ­ J. KRUPPA Résistance au feu des assemblages par boulons haute résistance. CTICM ­ doc. n° 1013­1 ­ Juin 1976.

­72­

EXAMPLES OF CALCULATION OF FIRE RESISTANCE OF THE STEEL MEMBERS ACCORDING TO THE EUROPEAN RECOMMENDATIONS OF ECCS TC3

S. BRYL Research and Development Department, Geilinger Ltd, Switzerland Summary The European Recommendations for the Fire Safety of Steel Structures concentrate on the analytical determination of the fire resistance of load bearing steel elements as an alternative for the standard fire restistance test. The calculation of the fire resistance consists of the computation of the rise of the temperature in the steel member under the influence of the standard fire conditions and of the calculation of the critical temperature of this element. The fire resistance is the time necessary for the steel element to reach its critical temperature. According to the initial data given, the calculation can be conducted in three different ways: 1. Given: Steel section, insulation, loading. Asked: Fire resistance. Example: Column with light weight insulation. Example: Column with heavy, moist insulation. 2. Given: Steel section, loading, required fire resistance. Asked: Type and thickness of insulation. Example: Continuous beam. Box type cladding. 3. Given: Steel section, insulation, required fire resistance. Asked: Admissible loading under fire conditions. Example: Unprotected steel column. (14 Fig.)

-73-

1. INTRODUCTION The European Recommendation for the Fire Safety of Steel Structures concentrate on the analytical determination of the fire resistance of load bearing steel elements as an alternative for the standard fire resistance test. The calculation of the fire resistance consists of the computation of the rise of the temperature in the steel member un­ der the influence of the standard fire conditions and of the calcula­ tion of the critical temperature of this element. The fire resistance is the time necessary for the steel element to reach its critical tem­ perature (Fig. 1) "c ISO sv kNDARD FRE

* ^

*­♦„

no

» ι

700

500

UJ K Ψ-

τ ÍMPER ATURE

300

/ ţ

<

STEEL

oc

*i»tlF/Vj ^i.di.tv)

Η UJ

ϊ »

ty 100

■J 0

^ 30

M

M 1IME

FIRE RESISTA NCE Ir

Fig. 1

Γ

­74­

120 t

MM

Following data are necessary for such calculations (Fig. 2 ) : ­ properties of steel at high temperatures as the yield stress σ γ and specific heat c , ­ structural behaviour of the steel element f(S), mostly beam or col­ umn, ­ loading of the element during the fire test

κ·ρ

­ the shape factor of the cross section F/V, ­ the thickness and the thermal conductivity of the insulation

\./d.,

­ density p. and the moisture content p. of the insulation.

By

F/V.C,

r.P*

*¡/d¡

f(s)

Si.Pi

OD

Fig. 2. According to the initial data given, the calculations can be conducted in three different ways, to prove that: Case 1: the time necessary to heat the steel element to the critical temperature is longer then the required fire restistance (υ.

«cr)

Case_2: the steel temperature υ

reached after the fire duration equal

to the required fire resistance time is lower then the critical temperature ° s S °cr Case 3: the loading of the element during the fire test is lower then the loading which would cause, at the maximum temperature reached at the required fire duration, the failure of the structural element: K.P

s

Ρ υ,cr

­75­

■c

ι^­τ *

^

0 0 SIA NOA RO F W t

2. CALCULATION CASE 1 Given: Steel section

CASE 1 OVEN : S U E L SECTION INSULA TION LOA DING ASKED: FIRE RESISTA NCE

«M

Insulation Loading Asked: Fire resistance Course of calculation:

Gy.T.P.»

900

1) Load level κΡ/ K P / P U and the critical temperature

υ

2) Temperature rise in the steel sec­ tion

V

υ

s υ read the tine t . cr r 3) For υ. and t = fire resistance. r


"V

t=? f

IME

00 I

MM

Fig. 3

Steel column with light weigth

­ = l ·

ISO STA NDA RD FIRE * ­ * _ · US. log ( I t ·11

insulation (Fig. 4) Column HEB 300. Steel grade 235 with σ

= 235 N/mm2.

Loading during fire Ρ = 1700 kN Heigth of column L. = 350 cm. Insulation: Sprayed mineral

¿cr=517'C

wool, thickness d. = 2 cm thermal conductivity λ. = 0,10 W/m.K, d^Xj = 0,2 m2.K/W Step 1 Steel section A = 149 cm i

2

= 7,58 cm.

Under the assumption, that the slenderness ratio of the column in fire test is: λ

= 0,7xLt/iy = 32,3

we receive for European buckling

Fig. 4

curve C:

­76­

buckling stresso

= 217 N/mm and the buckling load Ρ = 3233 kN. cr u With the correction factor for fire tests of K = 0,85 the load level is:

K.P /P

= 0,85.1700/3233 = 0,447

From the condition, that the relationship of the yield stresses σ

/σ

should not be smaller then the load level:

°y.A = ι + 767 In(υ /1750)

K.P/Pu = 0,447

s

we can easily find the critical

&y,¿/Ey

temperature, which is:

7. 100

u„ „„ = 517 °C s,cr For this purpose one can use diagrams as in Fig. 5 | 11 or

80

tables as published in the Appendix A | 2| . 60 Step 2

O.J7

The temperature increase of a lightly insulated steel element

40

during a time interval A t can

■

IUI li \

be calculated: Δυ = — s c

Κ

(υ t ­Ug). At

where:

20

Fig, 5 100

Κ = — —α„+α —

\

0,2( 11

+ d/λ· ι Λι

300

I

m wy,

'

1

For different values of F/V and d./λ. the increase of temperature has been calculated and the resulting values of temperature are tabulated in the Appendix Β |2|. From these tables it is also possible to con­ In our case we receive: d i A i = 0,2 m2.K/W For HEB 300: F = 1.73 m2/rn F/V = 116 m"

and

V = 0,0149 m3/m

1

­77­

700 'C

500

mm

struct diagrams as in Fig. 6.

,. co co

Through interpolation between F/V = 100 and 150 and for the critical temperature of 517 °C * we find, that this temperature will be reached after 111 minutes. That is, the fire resistance of bUU the column is:

d j A i = 0 2m*.K/W F/V

* ( Γ = ί17·

/

S

S 150 116 ι

ΙΙΗ)

t = 111 minutes r For a practical range of critical temperatures, 400 to 600°C and a constant λ-value, the time t ' r can be very easily calculated from the formula |2|: t = 40. (υ

140)

■m

t =111 30

0,77

Fig é

In our example: t

= 40.(517 - 140).(0,2/116)°·77 = 112 minutes

2.2.EXAMPLE2

Column as in Example 1 but with box-type cladding with moist insulation. Insulation: Thickness d. = 2 cm thermal conductivity X i =0,2 W/m.K diAi = 0,1 mZ.K/W specific heat c 1.7 kJ/kg.Κ density ρ BOO kg/m3 moisture content p. = 20 % Fig. 7

-78-

60

90

120 t

Step 1 Load level and critical temperature. Aa in Example 1: κΡ/Ρ

u

= 0,447

υ

= 517 °C

cr

Step 2 Temperature increase. 2 Area of inner surface of the insulation F r 4x0,30 = 1,2 m /m V = 0,0149 m 3 /m

Volume of steel

Section factor F/V = 81 m The section factor should be modified, because the heat capacity of the insulation c..p ,d..F = 1,7x800x1,2x0,02 = 32,7 kJ/m.K is greater then the half of the heat capacity of the steel section 0,5.c .ρ .V = = 0,5x0,52x7850x0,0149 = 30,4 kJ/m.K The modified section factor will be:

»W.mod

c .ρ 8c.K S.p . .F. .d. = 64

F/V. C

s°s+

1

1

2.V

1

F

1

^ v m o d = 6* m"

d¡/X¡ ïO.lm'.K/W

For the modified section factor of 64, the d.A .­value of 0,10 and the critical temperature of 517°C

500

we can find from Fig. 8 the neces­ sary time and at the same time the fire resistance: t

= 105 minutes r

Quite the same value can be re­ ceived using the simplified

tr :105Min

equation: = 40x(517­140)x(0,10/64)°' 77

t t

= 104 minutes

Fig. 8 3

r

The evaporation of moisture will slightly prolong this time by about: P 1 .p 1 .d i .d i 20x800x0^02' fc ,2 = 6 minutes v = ~ 5x0,20 5

^i

The fire resistance of the column is

t

r

­79­

+ t

ν

= 111 minutes,

3. CALCULATION CASE 2

ISO STANDARO FRE 4­l 0
Given: steel section, loading required fire resist­ ance. Asked: necessary insulation. Course of calculation: 1) Load level and the critical temperature

υ

2) Necessary d./X.­value.

Fig. 9 3.1._EXAMPLE 3

μ=τ­

Continuous steel beam IPE 300 Steel grade 235. σ

= 235 N/mm2

Span L = 600 cm. Loading at fire conditions ρ = 36 kN/m. Required fire resistance: 90 minutes. Step 1 Plastic moment of IPE 300: M = Ζ.σ = 628.23.5 = 147 kNm Ρ y Ultimate loading of a middle span of the continuous beam: ρ = 16.M /L 2 = 16xl47/62 K u ρ p u = 65,3 kN/m Correction factor for statical­ ly indetermined beam with two or more redundancies:

Fig. 10

­80­

κ= 0,25.(1 + 3.p/pu) = 0,25χ(1 + 3x36/65.3) Load level: κ.ρ/ρ

0.66

= 0.66.36/65.3 = 0.363

Critical temperature (Fig. 5) υ

σ

/σ y,υ' y

= 558 "C

Step 2 Section factor for IPE 300 with box­type cladding, one side against fire screened: F = 0,15 + 2.0,30 = 0,75 mVm V = 0,00538 m3/m

­1

F/V = 140 m'

We must now find such a value of d./λ­, which after 90 minutes of the standard fire, results in a steel temperature of 558 °C. From the tabu­ lated values of the rise of temperatures |2| one can construct a dia­ gram (Fig. 11) giving an inter­

*

dependence between the tempera­ ture of steel, the d./λ.­values

\1 V \ 1

and the section factor F/V. Interpolation for F/V = 140 gives the searched value as

FIRE RESISTA NCE 901V in

, s\ y *cr= 558"

V\

F/V di/\i

= 0,17 »IbO 100

In order to obtain a fire re­ sistance of 90 minutes the heat

01

transfer coefficient of the in­ sulation should be not greater then λ/d.

ι ι

1/0.17 = 6 W/m .K

02

0.5

d¡/Xj =0.17m*.K/W X¡ = 0.1 d¡> 0.017 m Fig. 11

For instance, the thickness of a mineral spray with λ. = 0.10 W/m.Κ should be: min d 1 = 0,168.0,1 = 0,017 m = 17 mm Instead of using the diagram the necessary value of d./λ. can be ap­ proximatly calculated with following equation: d./λ. = 0,0083. ^ . [t/(Us ­ 140)] 1 ' 3 For values of the example above: o y ^ = 0,00B3xl40x [90/(558­140)] Χ ' 3 = 0,16 m2.K/W min d, = 0,16x0,1 = 16 mm

­81­

d/λ;

4. CALCULATION CASE 3 Given: steel section, insulation, required fire resist­ ance, Asked: admissible loading at fire conditions. Course of calculations: 1) Steel temperature υ

after

the fire duration equal to the required fire resistance 2) Yield stress of steel at the temperature υ . 3) Admissible loading for tem­ perature υ ■ Fig. 12

4.1 . EXAMPLE 4

ISO SIANDAJtD FRE f * * ; " * I·«« (■>·■!

Unprotected steel column. Steel column, made from solid V«

round section D = 240 mm. Steel grade 355, with the guaranteed yield stress of σ = 355 N/mmZ. Length of the column L = 400 cm. Required fire resistance 60 minutes. Step 1 4/D ­1 F/V = 4/0,24 = 16,7 m" Section factor F/V

Through interpolation (Fig. 14) between F/V = 15 and 20 the steel temperature after

F i g . 13

­82­

60 minutes of standard fire can be read υ = 636 °C s

600

Step 2 The corresponding yield stress

°y,u /tJ y σ

u

=

°* 2 0 1

°y,u

=

°-201x355

= 72 N/mm2

Step 3 Buckling load at normal temperature: .2 452 cm

i = 6 cm

0.7x400 Fig. 14

λ= 47. For buckling curve C: 276 N/mm* and the buckling load Ρ = 27.6x452 = 12470 kN cr g u Buckling load at the temperature of 636 °C: Ρ = Ρ J} /O = 12470x0,201 = 2500 kN υ,υ u yfj y ' Correction factor for fire test: κ= 0,85 Admissible load under fire conditions: Ρ υ

= Ρ l< u,u

- 2500/0,85 = 2950 kN

The loading of the column under fire conditions should not be greater then 2950 kN.

5. REFERENCES |1| ECCS-Technical Committee 3. European Recommendations for Fire Safety of Steel Structures. Elsevier Scientific Publishing Co, Amsterdam 1983. |2| ECCS-Technical Committee 3. Manual on the European Recommendations for the Fire Safety of Steel Structures. (To be published before the Conference in Luxembourg).

-83-

APPENDIX A. YIELD STRESS OF STEEL AT HIGH TEMPERA TURES υ

°C

°.. yi u .y°.. y

300 0,778 301 0,777 302 0,776 303 0,775 304 0,774 305 0,772 306 0,771 307 0,770 308 0,769 309 0,768 310 0,766 311 0,765 312 0,764 313 0,763 314 0,762 315 0,761 316 0,759 317 0,758 318 .0,757 319 0,756 320 0,754 321 0,753 322 0,752 323 0,751 324 0,750 325 0,748 326 0,747 327 0,746 328 0,745 329 0,743 330 0,742 331 0,741 332 0,740 333 0,738 334 0,737 335 0,736 336 0,735 337 0,733 338 0,732 339 0,731 340 0,729 341 0,728 342 0,727 343 0,726 344 0,734

u σ„ ,,/σ„ o c y>» y

υ σ, υ,,/σ„ oc ^ y

345 346 347 348 349

0,723 0,722 0,720 0,719 0,718

390 391 392 393 394

0,661 0,660 0,658 0,657 0,655

350 351 352 353 354

0,716 0,715 0,714 0,713 0,711

395 396 397 398 399

0,654 0,653 0,651 0,650 0,648

355 356 357 358 359

0,710 0,709 0,707 0,706 0,705

400 401 402 403 404

0,647 0,645 0,644 0,642 0,641

360 361 362 363 364

0,703 0,702 0,700 0,699 0,698

405 406 407 408 409

0,639 0,638 0,636 0,635 0,633

365 366 367 368 369

0,696 0,695 0,694 0,692 0,691

410 411 412 413 414

0,632 0,630 0,629 0,627 0,626

370 371 372 373 374

0,690 0,688 0,687 0,685 0,684

415 416 417 418 419

0,624 0,622 0,621 0,619 0,618

375 376 377 378 379

0,683 0,681 0,680 0,678 0,677

420 421 422 423 424

0,616 0,615 0,613 0,612 0,610

380 381 382 383 384

0,676 0,674 0,673 0,671 0,670

425 426 427 428 429

0,608 0,607 0,605 0,604 0,602

3B5 386 387 388 389

0,668 0,667 0,666 0,664 0,663

430 431 432 433 434

0,601 0,599 0,597 0,596 0,594

­84­

υ σ„ ,,/ow

°C 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479

y»u

y

0,593 0,591 0,589 0,588 0,586 0,584 0,583 0,581 0,580 0,578 0,576 0,775 0,573 0,571 0,570 0,568 0,566 0,565 0,563 0,561 0,560 0,558 0,556 0,555 0,553 0,551 0,549 0,548 0,546 0,544 0,543 0,541 0,539 0,537 0,536 0,534 0,532 0,530 0,529 0,527 0,525 0,523 0,522 0,520 0,518

υ

°c 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524

σ.. ,,/σ. y»u y 0,516 0,514 0,513 0,511 0,509 0,507 0,505 0,504 0,502 0,500 0,498 0,496 0,494 0,493 0,491 0,489 0,487 0,485 0,483 0,482 0,480 0,478 0,476 0,474 0,472 0,470 0,468 0,466 0,465 0,463 0,461 0,459 0,457 0,455 0,453 0,451 0,449 0,447 0,445 0,443 0,441 0,439 0,437 0,435 0,433

APPENDIX A. υ Oy^j/Oy

υ

ν».*,

Ou/a

°c

°c 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569

Continued

0,431 0,429 0,428 0,426 0,424 0,422 0,420 0,417 0,415 0,413 0,411 0,409 0,407 0,405 0,403 0,401 0,399 0,397 0,395 0,393 0,391 0,389 0,387 0,385 0,383 0,380 0,378 0,376 0,374 0,372 0,370 0,368 0,366 0,364 0,361 0,359 0,357 0,355 0,353 0,351 0,348 0,346 0,344 0,342 0,340

570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614

615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659

0,337 0,335 0,333 0,331 0,329 0,326 0,324 0,322 0,320 0,318 0,315 0,313 0,311 0,308 0,306 0,304 0,302 0,299 0,297 0,295 0,292 0,290 0,288 0,286 0,283 0,281 0,279 0,276 0,274 0,272 0,270 0,268 0,265 0,263 0,261 0,259 0,256 0,254 0,252 0,250 0,248 0,246 0,244 0,242 0,240

­85­

0,238 0,236 0,234 0,232 0,230 0,228 0,226 0,224 0,222 0,221 0,219 0,217 0,215 0,214 0,212 0,210 0,209 0,207 0,205 0,204 0,202 0,201 0,199 0,197 0,196 0,194 0,193 0,191 0,190 0,188 0,187 0,186 0,184 0,183 0,181 0,180 0,179 0,177 0,176 0,175 0,173 0,172 0,171 0,169 0,168

υ σ

660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700

/θ

0,167 0,166 0,164 0,163 0,162 0,161 0,160 0,158 0,157 0,156 0,155 0,154 0,153 0,152 0,150 0,149 0,148 0,147 0,146 0,145 0,144 0,143 0,142 0,141 0,140 0,139 0,138 0,137 0,136 0,135 0,134 0,133 0,132 0,131 0,130 0,129 0,128 0,127 0,126 0,126 0,125

APPENDIX Β. MEAN TEMPERATURES IN PROTECTED STEEL MEMBERS ISO STANDARD FIRE. INITIAL TEMPERATURE 20 "C. SECTION FACTOR F/l/ m" 1

TIME Min

30

40

50

60

70

80

90

100

150

200

250

300

400

500

38 61 86 111 136 161 186 210

40 66 93 121 149 177 203 230

42 71 101 131 162 191 221 249

53 94 137 180 221 260 298 334

63 117 171 223 273 320 364 405

73 139 203 264 320 372 420 465

83 159 232 300 362 418 470 517

103 198 286 365 434 496 551 601

121 234 333 420 494 559 616 666

41 69 98 127 156 185 213 241

44 74 107 139 171 203 233 263

46 80 115 151 186 219 252 284

59 108 157 206 252 296 338 377

71 134 196 255 310 361 409 453

83 159 232 299 361 418 469 516

95 182 265 339 406 466 520 569

117 226 324 409 482 547 603 653

46 80 115 150 185 219 251 283

49 87 126 164 202 239 274 308

52 94 136 178 219 258 296 332

68 127 186 242 295 345 391 434

83 158 231 298 360 416 468 515

97 187 271 347 415 476 531 580

111 214 308 391 464 527 583 633

138 264 373 465 543 609 666 716

54 97 142 186 228 269 308 345

58 106 155 203 249 293 335 374

62 115 168 220 269 316 360 401

82 157 229 296 358 414 466 513

101 195 282 360 430 492 547 597

119 230 329 415 490 555 612 662

136 262 371 463 540 607 664 714

169 320 442 540 620 6B 7 743 792

2

INSULATION d./λ. = 0,60 m .K/W 15 30 45 60 75 90 105 120

31 46 62 78 95 112 129 146

33 51 70 89 109 129 149 168

36 56 78 100 123 145 168 189

INSULATION d./λ. = 0,50 m2.K/W 15 30 45 60 75 90 105 120

33 51 70 89 109 129 148 168

INSULATION

d^X.

15 30 45 60 75 90 105 120

36 57 79 102 125 148 171 193

38 63 88 115 141 167 192 217

= 0,40 m2.K/W

36 58 81 105 129 152 176 199

40 65 93 120 148 175 202 228

43 73 104 135 167 197 227 256

INSULATION d./λ. = 0,30 m2.K/W 15 30 45 60 75 90 105 120

33 50 69 88 108 127 147 166

37 60 84 109 134 159 183 207

41 70 99 129 159 189 218 246

46 79 114 149 183 217 250 281

50 88 128 168 206 244 280 314

-86-

APPENDIX Β. Continued. SECTION FACTOR F/V m" 1

TIME Min

30

40

50

60

70

80

90

100

150

200

250

300

2

INSULATION d-Λ- = 0,20 m .K/W 15 30 45 60 75 90 105 120

39 64 90 118 145 172 198 224

45 78 112 147 181 214 247 278

51 91 133 174 215 253 291 326

57 104 153 200 246 290 331 370

63 117 172 225 276 323 368 409

69 130 191 249 303 354 402 445

74 142 208 271 329 383 433 478

80 154 225 292 354 410 461 508

108 209 302 385 457 521 577 627

134 258 366 458 536 603 660 711

158 302 421 518 598 666 723 772

181 341 467 568 648 715 771 818

Min

10

20

30

40

50

60

70

80

90

100

150

200

250

86 168 247 320 386 446 500 550

97 189 276 355 426 489 545 595

107 209 304 388 462 527 584 635

117 228 329 418 494 560 619 670

126 247 353 445 523 591 649 700

172 327 454 555 637 705 762 810

212 393 529 633 714 779 832 876

249 447 588 690 768 829 878 919

131 259 373 469 551 621 681 733

147 289 411 512 595 665 725 776

163 317 445 549 633 703 761 811

178 343 476 582 666 734 791 839

193 368 505 611 694 762 817 864

259 467 612 716 793 853 901 940

315 540 683 780 850 902 943 977

148 293 417 519 604 675 735 786

166 326 457 563 648 718 776 826

184 356 493 600 685 753 810 857

201 384 525 632 716 783 838 883

217 409 553 660 743 808 861 904

290 512 658 759 832 888 932 967

INSULATION d ^ 15 30 45 60 75 90 105 120

= 0,10 m2.K/W 54 99 146 192 237 279 320 359

65 123 182 239 292 343 390 434

76 146 216 281 342 398 449 496

INSULATION d.A· = 0,05 m2.K/W 15 30 45 60 75 90 105 120

40 68 98 129 160 191 222 251

60 112 167 221 273 322 368 411

INSULATION d ^ 15 30 45 60 75 90 105 120

44 76 112 148 184 220 254 288

66 128 191 252 311 365 416 463

78 153 228 298 363 423 477 527

96 191 282 364 438 503 561 613

114 226 330 421 499 568 628 680

= 0,04 m2.KW 88 175 260 338 409 473 531 583

109 218 319 409 488 557 617 670

129 257 371 469 552 622 683 735

-87-

APPENDIX Β. Continued. SECTION FACTOR F/V m"1

TIME Min

10

15

20

25

30

40

50

60

70

80

90

100

126 254 369 468 553 625 686 739

149 298 426 531 618 689 750 801

172 338 474 583 669 740 798 847

193 374 516 626 711 780 835 881

213 406 553 662 746 812 865 908

232 436 584 693 774 838 888 929

251 463 612 719 798 859 907 946

151 306 440 550 639 713 774 825

180 357 501 614 702 773 831 878

206 402 552 665 751 818 871 915

231 441 594 705 788 851 901 942

255 476 630 738 817 877 924 961

277 507 661 766 B41 897 941 975

299 535 687 788 860 913 954 986

194 393 551 671 762 832 887 931

230 451 615 731 815 878 927 965

263 294 322 349 373 501 543 579 609 635 664 703 734 759 780 775 808 832 851 866 853 879 898 912 922 909 930 945 955 963 953 969 981 989 994 987 1000 1009 1015 1020

INSUALTION d./λ. = 0,03 n2.K/W 15 30 45 60 75 90 105 120

48 88 131 175 218 260 300 339

62 120 180 239 296 349 399 446

INSULATION d./k. 15 30 45 60 75 90 105 120

56 107 161 216 269 320 369 415

73 146 221 293 361 424 481 534

75 150 224 296 362 423 479 530

88 178 265 346 420 486 545 598

101 204 303 392 470 539 600 654

= 0,02 m2.K/W 89 182 275 360 437 506 568 623

106 217 323 418 501 573 637 692

121 249 366 468 554 628 692 746

INSULATION d.A · = 0,01 m2.K/W 15 30 45 60 75 90 105 120

68 139 215 289 360 426 487 542

91 191 292 387 472 547 614 673

113 239 360 467 559 637 705 762

135 283 418 532 627 706 771 826

WITHOUT CLADDING d.A-

155 323 469 587 682 758 820 871 =0

1 1

15 30 45 60 75 90 105 120

95 215 343 467 580 678 762 833

131 296 460 603 721 814 886 942

164 196 226 282 333 378 418 454 486 514 367 429 484 572 637 685 721 747 767 781 552 625 682 760 808 836 854 865 872 877 700 767 814 869 897 911 919 924 928 930 809 864 898 934 949 957 961 964 966 967 889 930 954 976 985 990 993 995 996 997 946 977 993 1007 1013 1016 1018 1020 1021 1022 989 1011 1021 1031 1036 1038 1040 1041 1042 1043

-88-

FIRE SAFETY, DESIGN OF COMPOSITE COLUMNS

J. Β. SCHLEICH Department Manager ARBED-Recherches LUXEMBOURG

Summary

The four basic types of composite columns are presented.These columns, alike by the combination of the two materials STEEL and CONCRETE, differ by individual behaviour under fire conditions. Together they offer a whole set of very interesting technical and practical solutions. The theoretical basis of the fire resistance calculation of composite columns, their experimental support as well as the available practical design tools are exposed.

"This taper is dedicated to the memory of Dr. Augusto CARPEM, who suddenl y passed away from us the 25th March 1984. As General Secretary of the European Convention for Constructional l Stee work, he significantl y contributed to the success of this l ast conference."

-89-

INTRODUCTION From view point of performance under fire conditions four different types of composite columns should be considered (Fig. 1): a) b) c) d)

rolled Η­profiles encased in concrete rolled Η­profiles concreted between flanges concrete filled circular or rectangular hollow sections with and without reinforcing bars steel core columns embeded in concrete with outer circular square steel sheet.

a)

b)

a?)

V " > ^'Ί«i * s ^ ^ O lM'. M

* sSVN1 \>N>. ' s ^\Ν\Ί \­ vXSH 1­X >\:

l·^:

Γ7\ s^ ν1· 1 1 \ ^N ^ J

L NX1

\1 Λ

c·

ί ¡Jj 1

η FIG. 1.

or

\v

M

V

J

\ \^ "J ^^^\

Typical cross sections of composite columns

One of the great advantages of the composite columns are their constant outside dimensions in multi­storey buildings. By varying the thickness of the steel section, the material qualities of steel and concrete, the percentage of reinforcement, the cross section of the column may be adapted to the increasing load without changing significantly the outer dimensions. Every type of composite columns has its typical advantages and its proper range of use. The oldest type of composite column is the steel section encased in con­ crete. Its advantages are: High possible load level in fire conditions and high carrying capacity not only for axial loads, but also for bending moments (1). The second type, the profiles concreted between the flanges, can bear considerable axial loads and high bending moments. The shuttering is strongly reduced and, if the section is concreted before erection, even totally superfluous. Other advantages are: good resistance against the mechanical damage without any reinforcement of corners, conventional steel connections between columns and steel beams even when columns are concreted before erection (2). Both first types of composite columns furthermore have an excellent resistance to earthquakes. It should be underlined that these two types of composite sections are also used as beams.

­90­

The carrying capacity of the concrete filled hollow sections in case of fire depends on the load level, but this capacity can be increased by using reinforcement and high values of concrete compressive strength. The use of reinforcing bars allows to consider bending moments or email load eccentricities under fire conditions (3). The steel core column is a further development of the concrete filled hollow section, but with the main part of the load carrying steel cross section protected against fire by layers of concrete. However, the columns with steel core should be used as axial loaded members with only small eccentricities (4). Both last types of composite columns have a considerable advantage compared with the reinforced concrete columns. The most severe drawback of reinforced concrete members is the danger of spalding, by which the reinforcing bars will be layed free and the expected fire resistance considerably reduced. In the last two composite column types the ouside casing will prevent the spalding and the core will remain better protected against the fire. The range of use of the previous given composite column types is very wide. They finally cover a spectrum of axial loads extending from 100 kN to 20000 kN for the fire class F 120. Higher carrying capacities can be obtained by using special built up composite column types like those given in fig. 2 and 3.

FIG. 2 F 120 built up composite column, based on rolled Η-profiles, of 0.9 m diameter, with the service load of 45 000 kN.

FIG. 3 AF 240 built up composite column, based on rolled Η-profiles of 980 mm χ 1058 mm cross section, with the service load of 90 000 kN.

-91-

BASIC PRINCIPLES OF TEMPERATÜRE AND LOAD BEARING CA LCULA TIONS

2.1. General load bearing behaviour under fire conditions Temperature gradient over cross section of a composite structural member leads to different heating rates of its single components. In connection with temperature dependent decrease of material strength, fire resistance will depend on cross section design. Differences in cross section type, involve differences in load bearing behaviour in fire case independent of load bearing capacity at normal temperature. In fig. 4 the principle of change of component's ultimate plastic load is plotted against fire time t. External steel directly exposed to a fire will loose very rapidly its plastic load bearing capacity, because of its high heating rate and consequently its high strength reduction. That part of plastic load bearing capacity, which comes from an external steel shell, must then be carried by the concrete core. Contrary to this characteristic, in cross sections with centrically imbeded steel profils, load will be rearranged from concrete to profil. In this case average of concrete strength over cross section will decrease more rapidly than strength of the imbeded steel profil (23). To the second basic type belong cross­sections of type a and d of figure 1, whereas concrete filled hollow sections (type c) must be attributed to the first basic type. Rolled profiles concreted between flanges (type b ) , form a mixed type with directly exposed steel of the flanges and insulated steel of the web.

Internal Steel (hot rolled Mettons, reinforcement,...)

FIG. 4 Ultimate plastic load of the cross­section components in function of fire time t.

Fir* Tim· t­ 2.2. Temperature distribution As the heat capacity of the concrete part of a composite column is not at all unimportant, the general differential equation of heat must be solved g e n e r a l transient d i f f e r e n t itemperature a l e q u a t i o n field. of weak in order to get the accurate ιccurate transient temperature field.

[1]

^J-f^t

«

­92­

where /{(θ): ,¿.(9): ƒ : c5:

Thermal conductivity W/mK Specific heat Wh/kgK kg/mJ Density Fire temperature as a function of time t (in minutes) according for example ISO-standard: 20 + 345 log10 (8 t + 1) θ

The numerical methods using finite differences or finite elements are the most common and unfailing ways to solve this differential heat conduction problem. Whereas the finite difference methods are all easy to use with one dimensional problems of simple structures, they become difficult with several dimensional problems. The formulation of boundary conditions may also raise problems. Therefore the more generally applicable programs use the method of finite elements, which is applicable to even complicated structures and to all types of boundary conditions. Typical networks for finite elements analysis are presented in figures 5 and 6. It is essential to know, that the heat (Q) is transferred from the hot gases of the furnace or fire to the surfaces of a column by the two mechanisms of convection (Q ) and radiation (Q ) . Q ■ Qc

+

Qr ·

[2]

" ^

Together with the furnace or fire temperature Θ,, the column surface temperature θ (Κ), the coefficient 0( of convection heat transfer and the resultant emissivity of steel ¿ or/and concrete, determine the heating up of the column.

­ífte f ­o e ) +
[3] w/

2 m m

4 tf κ ·

For practical use, values of Of » 25 W/ 2 and ¿ » o,4 to 0,7 lead to surface temperatures which are in good accordance with test results. For cross sections with outer steel sheet the resultant emissivity can be reduced to
FIG. 5

1XS

« lie on

Typical FEM­network for steel profil concreted between flanges. Steel sections (profil and reinforcing bar) have been hatched, whereas the numbers given represent the temperatures (°C) computed in nodes after 90 minutes of ISO­standard fire exposure (6). The finite elements THERNL, with 8 nodes, are Isoparametric thus allowing the simulation of any shape.

­93­

FIG. 6 Typical network for the computation with finite elements. The numbers presented indicate the temperatures after 90 minutes of ISO-Standard fire exposure. ■ The circular core has been replaced by a quadratic one of equal area (5), which is sufficient for this type of composite column.

Internal stress diagram due to the unequal temperature field in a composite AF cross section, after 120 minutes of ISO fire (12) 2.3. Load bearing capacity of composite columns under fire conditions As seen before, the transient temperature field, created under fire conditions in a composite cross section, is highly differential. Thus the simplified proceedings (11) used with sufficient precision in order to analyse protected or unprotected steel elements, are no longer applicable. Indeed first of all the unequal temperature field, through the temperature

-94-

depending, non linear, thermal and mechanical material properties, creates a rather inhomogeneous material properties distribution within the cross eection. Finally the unequal temperature field accompanied by an unequal free thermal strain field, through the plane cross­section assumption, leads to strong internal stresses (fig. 7). These however affect undoubtedly the load bearing capacity of composite construction elements. For these reasons the calculation of the ultimate buckling load Ν of composite columns under fire conditions is only possible, without* any restrictions as to the geometry of the cross­sections, the building structural system, the load combination etc. if a numerical model is used allowing an exact thermal and mechanical system analysis (6, 13, 14, 15, 15'. 16). However the numerical complexity is quickly increasing with precision of such an analysis. Therefore simplified calculation been developped. These models allowing a quick, every day use of the ultimate buckling load of composite columns, for a given field and a given cross section, are based generally on

the growing models have calculation application

­ a transient thermal field analysis assigning, for a given fire class, a mean temperature to any finite element in case of a sufficiently fine mesh. This temperature gives the respective material parameters of the element. ­ the calculation of the ultimate plastic load of the total composite cross section which is obtained by summing up the plastic limit loads of all finite elements of the discretized section. This procedure represents the so­called summation method.

Ν Ρ Θ ­ j£{À;*f o

+¿i*Se,*

>

W

As however the thermal, internal stress distribution, increased by the axial column load, leads to heavily crushed border elements whereas the core elements are less loaded, border elements should not be considered or at least their contribution in the total plastic load should be strongly reduced. This fact is taken into account by the balanced summation method for which

Ν

xA x

Ρ.Θ '¿Vi

i ^i,»

+

Îi *Aj *<E. e > I«

with o f frf 1 ­ the calculation of the ultimate buckling load Ν

for a given fire

class, by using the ultimate plastic load previously calculated and following the guidlines for ambient temperature (17, 18, 19) N

cr,e " * * Ν Ρ , Θ

I«

Among those simplified calculation models, the summation method should only be used for short columns which are more or less centrically loaded. The balanced summation method may be applied to columns with moderate slenderness and small eccentricity of load, if based on real fire test calibration.

­95­

Based on the aforementioned calculation methods i.e. transient thermal field and exact system analysis, summation method and balanced summation method calibrated on real fire tests, practical design tools like tables, graphs and diagrams have been established enabling architects and engineers a quick, safe and economic calculation of the ultimate buckling loads for composite columns. Thus in practice 3D, non linear, transient FEM-programs have not to be used. They remain research tools which of cause are needed in order to develop our knowledge on composite construction elements and to assure the firm basis to those practical design tools which will be explained hereafter. 3.

FIRE RESISTANCE OF COMPOSITE COLUMNS WITH ROLLED PROFILES ENCASED IN CONCRETE

3.1. Description Columns of hot rolled profiles totally encased in concrete have been used in practical engineering for a long time. In these cases concrete served as fire protection of the load bearing steel columns, which continued to be designed as steel elements. The concrete only produced the effect of thermal insulation. Recent development created the type of composite columns taking advantage of the concrete's own load bearing capacity. Thus, besides the well-known fire resistance, a considerable improvement of load bearing capacity is achieved. Columns of this type attain a much better load bearing capacity than columns of reinforced concrete or pure steel (1). The cross section of this composite column consists of an Η-shaped rolled profile arranged centrically within a reinforced section. According to the design regulations for a reinforced cross-section, the latter is provided with longitudinal reinforcements at least in the section corners and with stirrups (Fig. 8). FIG. 8 Composite column section with Η-profile encased in concrete, including rebars and stirrups

-A-

•B-

The construction of closed stirrups of form A can be replaced by 2 half-stirrups of form B, provided that the hook (u) is long enough to tie completely the 2 half-stirrups. The reinforcement can either be concentrated in the corners, distributed proportionally along the section sides or be symmetrically arranged on tension side and compression side in case of additional bending moments. 3.2. Method of calculation and experimental basis Within an European research project, extensive experimental and theoretic-numerical work has been done in order to ascertain the fire behaviour of these columns (20). The knowledge of the dependence on parameters, gained by experimental investigations has been generalized by numerical analysis (9, 14, 15, 21, 22). The comparison between failure times ascertained either numerically or by experiment is given in figure 9. The data show a divergence of less than 10 X which may partially be a

-96-

result of geometrical imperfections. The different heating behavioure of the individual component parts, i.e. re-bars, concrete and profile section are shown in fig. 10. As a consequence of the earlier heating of the reinforcing bars in comparison to the steel section, increase of the percentage of reinforcement in order to improve considerably the load bearing capacity at normal temperature gives way, in case of fire, to an earlier failure of columns. Indeed the total load of the failing outer component must be transferred to the efficient inner one. In case the influence of the profile section prevails, load bearing capacity of the column is improved in case of fire because of considerably slower heating of the steel component. t* [mini

FIG. 9

X.

160

y

Ultimate failure times of experiments (ttt) compared to numerically predicted failure times ( t£), for 18 tested composite columns with encased rolled profiles.

I

/ m

/ s

//y

/

y^/

s

/

y

%/

/ / ·/ • ·/· bU

—^TC^

t" (mini

0 60

90

120

150

160

^ FIG. 10 The different heating behaviours of individual component parts 1) reinforcing bars 2) profile flange edge 3) profile web-flange junction 4) Center part of profile web.

Concrete should not be the dominant component, because it has an earlier heating and loses strength earlier than the profile section. An increase in

-97-

the profile cover d. intensifies the effect of insulation and thus improves failure time. This Increase in profile cover means, however, increase in concrete as load bearing component and, at the same time, decreasing part of profile section in relation to the load bearing capacity. Since concrete has a higher heating rate than the profile, an increasing portion of load has to be redistributed from concrete to profile. This procedure cannot be recommended to obtain optimum fire resistant qualities of cross sections of this type of composite columns. With equal loading and equal failure time, outer dimensions of cross sections can be considerably reduced, if the part of reinforced concrete is reduced and the profile section becomes the predominant component of the cross section. Composite columns with hot rolled steel sections totally embedded in concrete, which have been optimized in this way, not only attain a considerably higher load bearing capacity than columns of reinforced concrete with comparable dimensions (1,23) but additionally have higher fire resistance times. In figure 11 it is proved that, in comparison to reinforced concrete columns, the required fire resistance can be obtained with much smaller cross section dimensions. FIG. 11 Minimum dimension "a" of concrete columns and composite columns with encased rolled profiles, in function of the required fire class F 30 to F 180,

FAILURE TIME

( u I mini ·

^ R einforced Concrete Columns ' occloDIN 4102 (BSI420/500R ) I-Composite Columns , usingmax HE - M Profils (SI 52)

Increase in column length involves a decrease of failure time due to slenderness. Figure 12 shows, however, that even in case of total design load (t)ø~ 100 Ζ) with column slenderness of normal multi-storey buildings, where columns can be supposed to have at least one fixed end in fire conditions (Euler-case 3 ) , still a minimum of 90 minutes failure time will be attained. Furthermore, figure 12 proves the validity of this statement, in spite of smaller outer dimensions, up to a failure time of 120 minutes in case of cross sections with high fire resistant qualities, for instance by using HEM-profiles. The heavier HD profiles will of course be even more efficient (42). Systematic investigations have shown that load eccentricities have no negative effects on column fire behaviour, if bending moments caused by this eccentricity have been considered in the design for service conditions. This favourable behaviour can be assumed up to an eccentricity

-98-

of e^0,5.a, i.e. normal force acting at the cross section edge. In case of a normal force applied outside the cross section, the eccentricity e^0,5a should act according to the main axis of inertia of the hot rolled steel profile. FIG. 12 Failure times of composite columns with encased profiles, in function of column lengths.

0

3

A

5

3.3. Practical design The following recommendations (23) guarantee, besides optimum behaviour in case of fire, advantages under service conditions and for technical production. l

u

H)

Ρ Γ 0 f i l e

HE ■1*

roBsrks

1 *7

HD

«ï

FIG. 13

M

>look

>so

>

Ilo 31ο

ht

M00I

têC

>

ï6o 36ο

7J

MOOI

MO

>

Ilo 31ο

97

>100λ

>so

>

He ato

13* 167

<*000 JO

<4500 *0

MOOI

MO

>]00B

MO

> doc ■too

MOOI

>so

MIO

MOOI

MO

Μ00Μ

>4 0

MOOI

>S0


M20M

MO

>1I0N

MC

IDI

~2t0 MIO

130 .

to coacrot»

­μ < 3 t foi

Ita

* fO'

aad

L

> 4.S

.)

e

14 Β

MIO *31C

. . 100 . . 1*9

«400C

>I2M

>SC

>3tc . "3tc

<·000

MIOH

MC

> aie 31o

. . ì&e

> 36c " 3(c

. . 191

130

accordlag rolofoxc«6 coloana

Minimum steel profiles assuring a given fire resistance class without load reduction, in function of the column length Lo (E3). (see also (59)).

•7

'aio Mte

M40 'ito MOO

WWW:, w/W* τ

1>

>40

"310 <*000

Www?

­for

t^MJO'i ρ < 1 «

> *fC . " hoc . U T

Design of cross section: ­ The minimum cover d. of the hot rolled steel section has to be 40 mm. ­ In order to obtain optimum fire resistant qualities with minimized

­99­

cross section dimensions, thick hot rolled steel sections are of advantage i.e. HEM, HD profiles and the like. It is recommended to use high strength steel for the rolled shapes. Reinforcement (A »tf­ ) should be minimized. Γ

y,r

Λ

Concrete of higher quality is not advantageous (if possible, limit/a to 35 N/mm 2 ). » If composite columns are supported according to Euler­case 3, that means one end fixed and one end hinged, they attain in normal multi­storey buildings, without reduction of load ( /*J© ■= 100 %) a minimum fire resistance of 90 minutes; the fire resistance of 120 minutes is given when the rolled profile is clearly the predominant load bearing component (see fig. 13). 3.4. Constructional details (24) *

The hot rolled steel section has to be embeded evenly into the concrete; vertical concretinR is advantageous. The concrete has to be compacted by vibration. The maximum grain size should be limited to 16

*

Spacers should be arranged at a sufficient amount and size in order to guarantee the centric position of the hot rolled steel profile within the cross section, during concreting. The cross section has to be strengthened by stirrups. If high bending moments are introduced in the beam­column connection, the number of stirrups has to be encreased. Proportional introduction of load into the two components, hot rolled steel section and reinforced concrete, has to be ensured for instance by using top and base plates. Load introduction between these plates into the steel section of the column, is possible by means of a welded butt strap, further transfer of the load to the reinforced concrete can be obtained through shear studs welded to the web of the profile or the like. Composite columns with encased rolled profiles are qualified for préfabrication as well as for multi­storey elements.

*

*

*

■ à

LUENEBURG, Administrative building, W.Germany FIG. 15: Cross section and pers­ pective of column­beam connection.

FIG. 14: Composite columns with rolled profiles encased in concrete.

­100­

FIG. 16 et 17: Composite columns before and after concreting, used in connection with flat slabs - LUXEMBOURG, office building LE FOYER.

-101-

4.

FIRE RESISTANCE OF COMPOSITE COLUMNS WITH ROLLED PROFILES CONCRETED BETWEEN THE FLANGES

4.1. Description This column type, a component of the so called AF 30/120 composite construction system, has been developped by ARB ED, Luxembourg in collaboration with Prof. Dr. Ing. 0. JUNGBLUTH of the Technical University, Darmstadt in Western Germany (25, 2, 26, 27). This .system, right away available for columns, beams and their connections, can be designed for any fire resistance time of 30, 60, 90, or 120 minutes (28, 29, 30, 31, 32) and has been installed at some ten buildings in several European countries (33, 34, 35). The main characteristic of the system consists in using rolled H-profiles and filling the spaces between the flanges with concrete. The exterior faces of the steel flanges remain visible. The concrete contains longitudinal reinforcing bars which contribute to support loads. Stirrups or shear studs are welded to the web of the beam in order to ensure solidarity of the reinforced concrete with the steel profile at normal service (18) and under fire conditions (fig. 18). FIG. 18 Composite AF 30/120 column section with main components: H-profile, re-bars, stirrups welded to web and concrete between flanges.

4.2. Method of calculation The exact method of calculation is based on the principles exposed in chapter 2. The transient thermal field analysis of a composite AF 30/120 cross section ( 2, 6, 36), and the numerical calculation of the ultimate buckling load Ν - under fire conditions of the same AF 30/120 column (12, 13) have shown a very good agreement with the measured temperatures and the buckling load during the corresponding fire test. However for everyday use, a simplified calculation model has been developed for centrically loaded AF 30/120 columns. This model established according to the balanced summation method explained in chapter 2.3, has been calibrated on a series of AF 30/120 test columns (28, 37).

-102-

According to this method, the cross section of the analysed composite column is divided into various parts (see figure 19). It is granted that the mechanical properties of the aforementioned parts vary according to their average temperatures, which are known in function of the fire exposure time (t) and of the section size. When calculating the ultimate load bearing capacity, the whole section is supposed yielded according to its properties reduced in function of the average attained temperatures. FIG. 19 Reduced composite cross section depending on the fire exposure time. The static properties of the four individual parts are given by: w - I , moment of inertia of reduced w web area A r

fl, moment

of Inertia of fl flanges with area A'

- I

- I , moment of inertia of reinforcing bars with area A rb - I , moment of inertia of reduced concrete area A r By comparing the various test results, showing the temperature evolution in function of time and cross section shape, to the well known variation of material properties in function of temperature (38, 39, 40, 41), the following relationships could be set up.

4-4

The neglected outer parts (h ) of the web and the reduced yield point (o _ ) of the remaining central web are given by ry ,tf h w

f (t, h, e)

s

f (t, h)

ffry.e

t — fire exposure time h - depth of steel shape e » thickness of flanges

The mechanical properties, yield point and Young modulus, of the flanges will be strongly reduced by following relations: <7

fl ry,e

f (t, F/V) F/V - section factor (m

E

fl

)

f (t, F/V)

r.e

A similar reduction is applied to the mechanical properties of the reinforcing bars:

rb ary.e " E*

f (t

'

d - concrete covering of reinforcing bars

d)

- f (t, d)

-103-

The concrete layers (Sb) directly exposed to fire or in contact with the hot flanges­ (Fig. 19) are neglected. The remaining part of the concrete section has reduced compressive strength and Young modulus: Γ.β

J

f

(t

'

F/V

>

c E

r 8

­ f (t, F/V)

Thus the total ultimate plastic load (N „) of the composite section is P»" given by adding up the individual plastic loads of the web, the flanges, the reinforcing bars and the concrete. » «w ~.v ^ «fl —.fl ^ «rb rb , .c Λ c Ν ­­ A . O" Q + A .Q" α + Α .fl­ „ + A . « „ ρ,θ r ry.ö ry,θ ^ ry,β r pr,θ The effective rigidity being El

eff,e » <E20­C­lW>+ (Er!e ^

^

Ο

­ ^ ( E ^ ­ l '

> ·

the Euler buckling load should be N

2 ^Eleff θ^ *Lcr " c o l u n m b u c k l i n 8 length cr and the equivalent slenderness ratio is calculated by E 9 *

This slenderness ratio "X used with European Buckling Curve C (19), gives reduction factor d£ (17) to be applied to the total plastic load in order to obtain the ultimate, or critical buckling load N

cr,e

­*·

Ν

Ρ .Θ

4.3. Experimental basis Practical tests of fire endurance on AF composite columns have been carried out in the Fire Resistance Laboratory of Brunswick Technical University under the leadership of prof. Dr.­Ing.Dr.­Ing. E.h. Karl Kordina. These tests consisted in heating up the ambient gas temperature, around the previously loaded column, according to the ISO standard curve. The moment at which the loaded column collapses under these fire conditions gives the respective ultimate failure time. The main characteristics of the tested column specimens are: ­

­ ­

rolled Η­profiles according to European and American Standards Í42), yield point of profile steel given by ţ7~ « 235 N/mm. to 355 N/mm , yield point of longitudinal re­bars C~ ■ 420 N/mm ; normally 4 6 16 to 4 i 24 were installed, stirrups or studs welded to the profile's web, compressive strength of concrete p ~ 35 N/ 2 to 45 N/ 2. ƒ

-104-

ΙΗΠΙ

ΠΙΠΙ

More detailed test conditions and test results are given in fig. 20 and in the respective available technical certificates (28). FIG. 20 TECHNICAL COLUMN COLUMN ULTIMATE FA ILURE AXIAL TEST COLUMN Centrically loaded AF CERTIFICATES TUIE (■in.) TIST NUMBER PROFILE LENGTH END BRUNSWICK CONDITION LOAD (kJ!) L(m) columns tested under COHPUTED KEASURED fire conditions in HP 36oxl32 123 22oo 1*1 7715oR 3.8 Brunswick. 97 3σοο HP )6oxI32 112 HP 24oxS7

7oo

81

86

HP 24ox57

6oo

99

92

16oo

(β

7«, 5

12 So

97

98

325o

So

76

2ioo

91

88

HP 3oox86 HP

3oox86

HP

36oxl74

HP

36oxl74

HE 32oA A

lo

HE 32oA A

11

HP 31ox79

12

HP 31ox79

3.8

λ 3,68

5,71

2o9o

86

16oo

lo2

94

US

127,5

116

125

Τ

1

7oo 7oo

8o64*

79,5

831oo9

831ol6

It can be seen from fig. 20 and 21 that the ultimate failure times predicted by the simplified calculation method are very close the measured failure times, except for test number 7, in which case important concrete spalding occured and therefore gave a smaller test failure time.

FIG. 21 Comparison of fire endurances of A F columns, measured from tests and computed according to ARBED's "Reduced Composite Cross Section" method of S 4.2., for real steel and concrete qualities (see also fig. 20).

4.4. Practical design tools * A catalogue of composite columns for nearly all the series of European and American rolled Η­profiles has been set up on the basis of the calculation method described in paragraph 4.2., so enabling architects and engineers a quick analysis of AF 30/120 composite columns. This catalogue (29, 30) contains for the different rolled H­sections: ­ the allowable service load Ν of the rolled shape at ambient temperature ­ the allowable service load Ν of the composite column at ambient temperature ­ the ultimate buckling loads Ν „ of the composite column for the fire classes F 30, F 60, F 90 and F 120.

­105­

Following parameters have been considered: ­ ­ ­ ­

the column lengths L from 2.50 m to 4.50 m the concrete qualities Β 25, Β 35, Β 45 according to DIN 1045 the steel grades ST 37, ST 52 according to DIN 17100 the longitudinal reinforcing bars of quality BST 420/500 according to DIN 488 up to 3 Ζ of the concrete area ­ the creep factor of concrete, ratio between the permanent load and the total load applied to the composite section, supposed equal to 50 Z. A selection of this catalogue, containing a total of some 50 000 design values, is given in fig. 22. Calculation of allowable service loads Ν for composite AF 30/120 columns at ambient temperature, according to German standard DIN 18806, is based on a safety factor of 1.7 and on the assumption that the buckling length (L ) should be taken equal to the column length (L) from floor to floor, thus covering the most unfavourable sinusoidal instability case over several levels. But the calculation of the ultimate buckling loads Ν

­

of composite AF

30/120 columns in fire condition is based on a safety factor of 1,0. Furthermore as the real fire normally occurs on one level, a stabilizing effect is created through the strong rigidity decrease of the directly heated column, whereas the rigidity of the same column on the upper and lower levels remains more or less constant. Therefore, the directly heated column on one level, behaves as if its ends were almost fixed. Thus the given ultimate buckling loads Ν „ are based on the assumption that the buckling length (Lcr) is equal to 85 Ζ of the column length (L) from floor to floor. Buckling has always been presumed around the weak axis Z. * Concurrently, a set of diagrams has been drawn up, which allow a very fast design. Indeed each diagram contains the whole set of one profile ­ type from 240 to 1000 mm height. For a given fire class, load to be supported and column length, each diagram gives at once an adequate composite cross section including geometry and qualities of all the steel and concrete components (see fig. 23 and 24).

* Furthermore, CAD programs for the design of AF columns will be available on different hardware configurations as HP 41 CV, HP 85, VAX etc. These programs allow to consider any steel or concrete qualities given by national standards, provided the application field covered by the initial catalogue is not outwalked.

­106­

B E I ­ A S T U N Q S T ABEL­I­El* UERKSTOFFE

FUER VERBUNDSTLIETZEr >­9β­12β

RROFILSTRHL ST 37 BEKHRUHOSSTRHL »ST 42Β/386

PROFILREIIC BETONGUETE

HP ■ 13

KDIECHERZEUOENOE LA ST

BERECHMUNGS­ FAELLE

ZULRESSIOE ZENTRISCHE LftST IN KN BEI EINER FREIEN STUETZENLREHGE VON 3.23 2. SB 4. 38 4. 23 4.88 173 1 SB 3. M 273

STRHLPROFIL VERBUNOPROFIL

2411 4313 4»S6 3747 2886 192B

2366 4463 4926 3724 27»7 1M7

2336 4418 4Θ92 3697 2767 1892

2323 433B 4B36 3669 2745 1B73

2288 4286 4B16 3637 2728 1B37

2231 4217 4772 3683 2694 1838

2211 4143 4726 3367 2666 1817

2178 4869 4677 3328 2636 1793

2128 3991 4623 3487 2664 1772

F12B

2716 4B2B 3188 3693 2919 2B1B

26B6 4767 3138 3871 2968 2BB4

2633 47B9 3116 3B43 28ΒΘ 1M9

2616 4646 3878 3613 2837 1972

2378 4379 3037 1764 2B32 1933

2S3C 4386 4993 3749 2883 1933

2493 443B 4946 3712 2777 1912

2447 4331 4896 3673 2747 1Θ89

2466 4268 4842 3632 2715 1B66

12 β 18

STRM. PROFIL VERBUNOPROFIL F3B Fã« FM F12B

3822 3127 34 BB 4843 3834 211»

2989 3B72 3376 4821 3813 21B3

2932 5Θ11 3342 3994 2994 2BB8

2913 4943 33B3 3964 2971 2B7B

2878 4874 3262 3932 2946 2831

2823 47M 3216 3697 2*1» 2831

2777 4719 3168 383» 2698 2869

2727 4613 3117 3828 283» 1986

2673 4546 5B62 3778 2627 1961

U ( U

STRHLPROFIL VERBUHDPROFIL F3B FSB FM F12B

3273 3388 5*2« 41B3 3148 228»

3238 3324 3393 4161 3121 2193

3208 3261 3357 4133 3188 2179

3138 3194 331» 4183 3877 2161

3114 3121 3476 4871 3831 2142

1866 3843 3438 4833 3824 2121

3816 4961 33ΘΒ 399Θ 2993 2899

2963 4873 3328 3957 2964 2873

2969 4786 5272 3*15 2931 2836

12 · IB

STRHL PROFIL VERBUM3PROFIL F36 F66 FM F12S

3383 3692 SBS6 4338 3238 2312

3343 3633 3823 4313 3239 2297

3384 336B 3787 428« 3217 2281

143» 3497 3747 4233 3194 2262

3411 3421 3784 4222 3168 2243

336B 3348 3657 4186 3140 2221

3383 3255 3687 4148 3111 219»

3248 3163 3353 4187 3879 2175

3189 5671 5496 4064 3843 2149

BEUEHRUMl

12 β IS

Fie

F6B FM F12B STRHLPROFIL VERBUNDPROFIL F3B F68

FM

FIG.22: Catalogue for AF composite columns; this selection shows the design values (kN) for several rolled profiles HP 400 χ 144 to HP 400x213 (i.e. HP 14"xl6"x97 to 14"xl6"xl43), with steel quality St 37 ( Q"" ­ 235 N/__2) and concrete quality Β 35 ( A­ 35 N/ 2 ) .

L (m)­ FIG 24:F 120 ultimate buckling loads of American wide flange shapes W10"xl0"to W40"x 18"

FIG 23: F 90 ultimate buckling loads of HEAA European sections. « Τy ­ 355 N/mm2;/J­ 45 N/ ι mm2)

(^"355,,/ππη2^"45

­107­

N/

mm 2 )

4.5. Constructional details It should be noted that a construction designed according to the AF 30/120 process, still has the characteristics of a steel structure. The elements are prefabricated in the shop or on site and are erected and assembled according to methods similar to those of traditional steel construction. Steel- or AF-beams can be bolted to vertical steel plates welded to the AF-columns (fig. 25 and 26).

FIG.25: AF90 column with gussets for floor beam connection-COLOGNE, office building TradeARBED, W. Germany.

FIG.26: AF beams bolted on AF column - DELMENHORST, office building Magnus Muller, W. Germany.

But composite AF columns are also used in connection with flat slabs. The column is delivered with the steel shear head already welded (fig. 27 and 28). Besides the AF column can be prefabricated in one piece over several levels, re-bars being adapted to the changing axial loads, so that the building's construction speed will be accelerated. Concreting in horizontal position before erection, asks for no shuttering (fig. 28) whereas concreting an AF column in vertical position after erection is always possible. In this last case the shuttering will be strongly simplified, thanks to the very straight edges of the profile's flanges (fig. 27). Proportional introduction of load into the two components hot rolled steel profile and reinforced concrete must be guaranteed either by using top and base plate, either by welding locally shear studs to the profile's web or the like. The outer dimensions of this composite profile are strongly decreased compared to the dimensions of a traditional steel or concrete column. It should be noted that the column of fig. 28 supporting a service load of 7300 kN for the fire class F 120, has a cross section of 1000 mm to 300 mm. Furthermore AF columns have an appealing surface, since the directly painted steel flanges, together with the chosen concrete texture may create a most successful architectural feature (34). Besides their high resistance to impacts, due to the bare steel flanges, is a big advantage in case of industrial and public buildings.

-108-

FIG.27: AF 90 columns with steel shear heads after erection - GENEVA, Commercial Center Jumbo Vernier, Switzerland.

FIG.28: AF 120 column during horizontal concreting before erection - LUXEMBOURG, Social Assurance EVI.

4.6. Future developments Several research programs are going on for the moment. First of all different column-beam connection types are analysed under fire conditions (43). Furthermore a general numerical model for steel and composite structures (16) is developped. It should allow an exact thermal and mechanical system analysis without any restrictions as to the geometry of the cross-sections, the building structural system, the load combination etc. This more general fire approach is requested as the simplified calculation models are limited to the application fields covered by fire tests. Accordingly we should be able to analyse special types of composite cross sections (fig.29) and to solve the M/N interaction problem for composite columns under fire conditions (44). Fire retardant paint

FIG.29:

Special types of AF composite cross sections able to support axial loads and bending moments according to YYor/and ZZ axes.

-109-

5. FIRE RESISTANCE OF COMPOSITE COLUMNS WITH CONCRETE FILLED HOLLOW SECTIONS 5.1. Description Developments in the field of concrete filled hollow sections have been carried on steadily during the last ten years, for as well normal service conditions [ (18), (45) to (52) ] as fire conditions (23). The fire behaviour of columns with concrete filled hollow sections has been thoroughly studied in several countries, especially in France and Western Germany, where several research programmes were carried out between 1972 and 1983 (53, 54, 20). Theoretical and experimental studies resulted in working out computer simulation programmes of fire tests for composite columns of this type, among others (21, 55, 56, 15). Practical design tools like tables and diagrams are available allowing a quick design of this column type for the fire classes F30, F60, F90 and F 120 (58, 59, 3,).

^7T. '•'/""/s

FIG. 30 Traditional cross sections of composite columns with concrete filled hollow sections.

*//*> V? ///· '/// //'///, ¿////A /// '/'/ Zxu. As shown in fig. 30, different cross-section types are used depending on circular or square hollow shapes, and concrete with or without reinforcing bars. 5.2. Method of calculation * The purpose of the theoretical study of COMETUBE (53, 55) was to find the relationship between the imposed load and the ultimate failure time under fire conditions. It was based on the following assumptions concerning concrete filled steel hollow sections: - equare or circular hollow sections, - the hollow sections are warm or cold finished welded profiles, - section size between 140 and 400 mm, - filling concrete made of current aggregates, - either reinforced or non reinforced concrete, - the percentage of reinforcement varies from 0 up to 3,5 Z, - simulation of the heating process according to the ISO standard curve, - columns with no external protection, - subjected to a centric load without end moment.

-110-

For the numerical simulation of the fire stability of these composite columns, the thermal and mechanical properties of both materials ­ steel and concrete ­ have been defined at both ambient and high temperatures. The temperature field accuring in the cross section of this composite column is calculated on the basis of the finite difference method and following assumptions: ­ thickness and thermal conductivity of steel Involve a zero temperature gradient in steel. Heat resistance between steel and concrete is zero, and concrete is considered as homogeneous material. ­ there is no heat transfer along the vertical axis of the columns, which means that the problem is two­dimensional. ­ only radiation transmission was taken into account for the heat flow transmitted to the column. ­ free water influence was simulated by considering a rapid increase in the concrete's specific heat. For square shapes, the cross section is divided in a square grid, each square having a side length positioned diagonally; for reasons of sym­ metry, only 1/8 of the whole section, the triangle ABC (figure 31) must be analysed.

FIG. 31 Typical FDM­network for a square cross­section with n»6.

The half halt .sidt side a of the concrete core is divided into η equal parts so that a η . J> . f 2. To give stable conditions to the numerical calculation, the value of Jß should be 1 to 2 cm Since the temperature field of the composite column cross­section can be calculated for any instant, the ultimate failure time of the corresponding column can be determined (see also S 2.3). The medhod of calculation proposed is a generalisation for hot conditions of the method drawn up by GU1AUX and JANSS (45,47) for the calculation at ambient temperature of concrete filled hollow sections subjected to centric load. The method has, however, been modified to take account of load eccentricity (56).

­111­

* Following important conclusions could be drawn from the numerical calculations and tests of the German research program (20, 14, IS). ­ It is necessary on one side to have a minimum reinforcement in order to obtain sufficient fire resistance time. On the other hand, however, only slightly better fire behaviour can be expected by a very high percentage of reinforcement (see fig. 32).

FIG. 32 Influence of reinforcement on the failure time of con­ crete filled hollow sections.

ji=0.0V.

>J«1.0·/.

μ«3.0°/·

­ It has been observed that for cross section size •X'240 mm, a decrease of load by at least 50 % is necessary to reach a failure time higher than 60 minutes. For higher load levels, larger cross sections do not bring any considerable advantage in spite of delayed heating. This is due to the fact that a still significant part of the load at ambient temperature is supported by the hollow section itself, which leads to an early critical overloading of the concrete core under fire conditions (see fig. 33).

η » 0.7 Q6 0.5 0.45 0.4

60

90

FIG. 33 Influence of load level ^¡ and cross section size "a" on failure time.

120

VERSAGENSZEIT I I m n l (F­KLASSE)

­112­

5.3. Experimental basis * A large number of concrete filled hollow sections has been tested by COMETUBE under fire conditions and centric loads (53). About 79 composite columns containing circular or square profiles with and without re­bars were tested according to ISO 834 standard curve. Various parameters like end conditions, load level, thickness of hollow section, percentage of reinforcement and slenderness have been studied. The minimum section tested was the square 140 mm profile, whereas the biggest sections tested were the square 350 mm profile and the circular 406,4 mm profile. The large number of test measurements available enabled a good comparison between theoretical calculations and experimental results, thereby ensuring a high degree of reliability in the proposed method of calculation (fig.34).

FIG. 34 Ultimate failure times of ex­ periments (Texp) compared to calculated failure times (Tcalc) for 79 tested concrete filled hollow section columns under centric loads (53).

30

tO

10

10

100

130

HO

IH

IH

3O0

Furthermore 28 fire tests on loaded columns with small eccentricities (57) show good agreement with the calculation method. * The German research program of the S.A .E.S. (20) deals with 24 tested, concrete filled hollow section columns with and without load eccentrici­ ties. Failure times measured are very close the numerically calculated values (14, 15) as shown in fig. 35. t ; [mini

FIG. 35 Failure times of tests (t ) u compared to calculated failure times (tţi) , for 24 tested co­

/

y/}

t'­t"

/

/7

lumns with concrete filled

** A

hollow sections (20).

**As r

A

t" 'υ

Omini

­113­

5.4. Practical Design Tools For everyday use practical design tools, deduced directly from computer calculations have been established. Design values are given in tablee (58) and diagrams (3, 53) for 16 standardized square and circular hollow sections. The diagrams (see fig.36 and 37) are of most practical use as every graph gives for one required fire resistance time and one specified hollow section, the ultimate centrical buckling load Ν . (kN) in function of the column buckling length L (m) the reinforcement percentage 0,5; 1 or 1,5 Ζ the concrete compressive strength 30; 40 or 50 N/ 2

f

90

I 2DD

Ν.«. (KN)

|T­

60mn|

^ v

\

O»? 1

χ**

^

­—'Ζ,

s

­v V N

FIG. 36 Ultimate centrical buckling load of concrete filled hollow section column 200 ζ 200 χ 5 mm, for the fire class F 60.

^

•X ^ ^ \ Ν

§5^

too

t­^

'

1

o

Nu t (KN)

1

457.0

S

SS I J y

1

—'— ­ ~

|Τ.

Ί

120mi

S

■ ' ­<s M 0.S—1\ <β · . · — » κ

^. ­ ­

j _

Γ1­

­1­ J ­­ L ­

­L.

■4 rh ΓΗ ΓΗ Γ4 ι

u <·)

­114­

FIG. 37 ultimate centrical buckling load of concrete filled hollow section column 0 457 mm, for the fire class F 120.

The possibility of taking small load eccentricities or small moments into account is foreseen in monograph η S of CIDECT (47), which deals with the calculation at ambient temperature of concrete filled hollow sections. This is done by applying a reduction factor "Ä" to the ultimate centrical buckling load obtained from graphs. This reduction factor "o(" is given in fig. 38, in function of the 1./b and e/b ratios.

FIG. 38 Reduction factor et in case of load eccentricity: e: load eccentricity(mm) 1, : buckling length (mm) b: profile size (mm)

The reduction factor to be applied to the ultimate centrical buckling load under fire conditions in case of small eccentricity is given by the factor (see fig. 39) tK1 ­0Í. CŒFFtCtCNT

imtOAATEUÑ

FIG. 39 C facto factor function of m reinforcement percentage.

DC ALPHA

__j_ — —

■

— ■

-

>

"

A

ƒ

/

y

ß»

mulm ft arm■ fum

Consequently the diagrams (3,53) given for centrical loads under fire conditions can be used in case of eccentricities by considering that (56, 58): centrical,

«1

Νult.ö

Νservice

?

„centrical w 0r N

ult.e

¿

load

service load

"οζ

­115­

5.5. Constructional details The wall thickness of hollow sections is to be minimized for the fire design of this composite column type. Steel of higher grades should not be chosen. Columns are to be supported basically with the full surface area of their head and footplates. Loads coming from concrete floors (fig. 41 and 42) are transmitted to concrete filled hollow section columns through steel shear heads. Steel floorbeams are connected to the concrete filled hollow section through gussets welded according to fig. 40.

FIG. 40 Plate welded to hollow section for beam-column bolt connection (see also figure 43 and 44).

Steam escape holes are to be bored in the top and bottom zones of columns. Stirrups are to be arranged inside the hollow section, in order to fix the re-bars and guarantee their concrete cover. Stirrups do not have any other statical function. Aggregates of an adequate graded mesh curve are to be used for core concrete. The value of water to cement ratio is to be selected as low as possible. It is advantageous to use concrete liquifier. Columns are to be filled up with concrete in a vertical position. Concrete-filling is to be carried out in layers or continuously at a slow rate (see fig. 44). In order to achieve sufficient compression and to avoid gaps, the application of shuttering vibrators is recommended, which are to be installed transversely to the column axis.

-116-

FIG. 41: Residence building based on F90 concrete filled hollow section columns (300x300 mm) - BORDEAUX MERIGNAC, France.

FIG. 42: Hollow sections filled with concrete, supporting concrete floor - STRASBOURG, Centre de formation des PTT, France.

HÉHiii SÍSII

FIG. 43: Steel frame with F9o concrete filled hollow section columns (220x220) - BIELEFELD, administrative building, W. Germany.

-117-

FIG. 44: Steel frame detail with opening in column for concrete filling - Bielefeld.

6. STEEL CORE COLUMNS 6.1. Description This type of composite column has been developed by Geilinger Ltd. Switzerland. The column cross section, according to fig. 45, consists in a steel core embeded in concrete with outer circular or rectangular steel sheet. Contrary to other steel core column types (62), the concrete of this core column is strengthened by no reinforcing bars.

FIG. 45 Circular or square steel core columns.

6.2. Method of calculation At ambient temperatures the core column is calculated as a composite section. The combined action of sheeting, concrete and core is provided, even if the load is conducted only into the core of the column. For the calculation of the temperature distribution during the fire, the TASEF­2 Program was used (4,5). Typical results of such calculations are presented in Fig. 46. 30 MINUTES

60 MINUTES

1

90 MINUTES

Ml

m m

|

Β. *t

m

» »» » * « ft.I"1 va ■

□J L_

2

io

(

Nn η

t

ik *. Β». » t

β

tl β«

FIG. 46: Temperature development in core column 250 χ 250/140 after 30, 60 and 90 minutes of ISO fire.

­118­

With the known temperature distribution and the material properties, the load bearing capacity of the heated column is calculated according to the summation method given in chapter 2.3 (see fig. 47).

*-X<

*i

FIG. 47

f "t777^

Ή;«χ4 # ;4, Γ-Ή;$>"7·;^

'-*. Τ

Computation of the buck­ ling load at elevated temperatures.

--?<*V<Ço> + ^ ( / s V/3 c e >

Ultimate plastic load

:N„ Vo

Euler buckling load

: N E 9 ­ffr 4 *,·^»* ^¿(ΔΑ^Σ^

Equivalent slenderness ratio: 1

x¿2l

■ u'N _/N„„'

ST* "

The reduction d€of the ultimate plastic load due to buckling is given by the European buckling curve C. The ultimate buckling load at elevated temperatures follows by:

Νcr,β ■ae- »,ρ,β Cross section elements with temperatures greater than 600°C are not taken into account. In the above formulas: ΔΑ , ΔΑ s c

section elements of steel and concrete coordinate of the center of gravity of an element buckling length of the column

The results of computations for some typical sections of the composite columns with steel core are presented in fig.48. The values are given for (Γ ­ ­ 215 N / 2 , Λ ­ 30 N/ 2 and a co Lumn length of L ­ 3000 mm. y,¿o mm /*c mm

Section

300.300 300.300 0 300 t 300 200.200 200.200 0 220 t 220

Core

RND RND RND RND RND RND RND RND

160 180 160 180 120 100 120 100

Sheeting

6 6 7,1 7,1 6 6 5,9 5,9

mm mm mm mm mm mm mm mm

Buckling load in kN „ *) F60 F90 20°C ' F30 6280 7290 6450 7450 3370 2770 3470 2890

4370 5420 3970 4890 2020 1490 1970 1460

3930 4930 3740 4450 1630 1140 1560 1300

3580 4510 3150 3270 1270

870 1050

950

*) To obtain the service load at ambient temperature, the values should be devided by the safety factor.

­119­

FIG. 48 Ultimate buck­ ling loads of core Columns. Under fire con ditions L ­ 0,7 χ L.

6.3. Experimental basis The method of calculation was verified by three whole scale tests: two of them were carried out at the Swiss Federal Research Station in Dübendorf at ambient temperatures (63), one was a fire test at the Technical University of Braunschweig, West Germany (64). The most important data of these tests are given in figure 49. FIG. 49: Whole Scale Tests on Steel Core Columns Section

Core

Sheeting Length

150.150 240.240

RND 75 RND160

2 4

3000 3000

300.300

RND160

6

3700

Conditions

20°C 20°C P-2510 kN M-12,5 kNm

Results calcumeasured lated Ρ -1536 Pcr-4572 cr 90 min.

Remarks

1540 kN 5360 kN 128 min.

ISO F i r e Test

The fire test was run with a column load of 2510 kN and a measured eccentricity of 5 mm (M ­ 12,5 kNm). Under this loading a fire resistance of 90 minutes was predicted. The test has shown, that the fire resistance is greater, because reaching 128 minutes. During the test the temperatures in concrete and on the steel core were measured. The comparison between the measured and calculated temperatures are presented in Fig. 50. ^ — ^ ­ ^

me measured UhJUNl

woo Standard Fire (150) HO Point

HO

·"" "^

" /

_—■■*

100 II

.7

_.·»*

·.·

' "" ^^——"

­­O.

f^ ­­~"7

FIG. 50 Comparison between cal­ culated and measured temperatures (64). The measured temperatures are below the calculated ones.

s.w.i.n

Fire Duration Min.

It is obvious, that the calculated temperatures were reached with a delay of some 30 minutes, i.e. the calculation of temperatures is conservative. The failure occured as the temperatures reached the predicted ones. Thus, the method of calculation is on the safe side, which is desirable, as the material properties at high temperatures may considerably scatter.

­120­

6.4. Constructional details The sheeting of the columns can be manufactured from round or square hollow sections (Fig. 51) or from cold formed steel sheet.In the last case two C-formed parts will normally be used (Fig. 52).

FIG. 51: Core column with tube sheeting

FIG. 52: Core column with cold formed sheeting

In both cases there should be local connections between steel core and the sheeting. If the sheeting is made from hollow sections, evaporation holes should be provided, as in other concreted hollow sections. The composite columns with core are mostly used as centrically loaded columns in connection with flat slabs. The column is delivered with a steel shear head already welded and put upon shuttering of the slab or held in position with special devices (Fig. 53). The core columns have the smallest dimensions in comparison to any other type of column (Fig. 54). Furthermore, they have an appealing surface, since the steel sheeting may be painted directly. Their resistance to impacts may be advantageous in case of industrial and public buildings.

-121-

råt,

FIG. 53: Erection of F 90 steel core columns (65) WINTERTHUR, Commercial center SISKA, Switzerland. Max. column load 2000 kN.

CHUR, Parochial building Titthof, Switzerland. Max. column load 2100 kN.

FIG. 54: Two composite F90 columns with exactly the same height and loading: left -H section embeded in concrete, right steel core column. WINTERTHUR, Industrial building Weilenmann Ltd., Switzerland. Max. column load 4000 kN.

-122-

7. CONCLUSIONS

Composite columns are modern and slender construction elements with high load bearing capacities. They combine numerous advantages relating to design and practical construction and are able to fulfil given fire resistance requirements.

If the design of composite columns include fire engineering, the best possible technical and economical solutions are available. Indeed among others the load bearing materials remain visible because doing without any outer surface insulation, the fire resistance is unaffected by age, and the smallest possible cross section size is obtained.

Of course composite construction asks for more competency than mere steel construction. This apparent disadvantage is however largely compensated on one side by the aforementioned practical design tools and existing computerized calculation methods. On the other side the close association of steel and concrete inside composite elements, undoubtedly leads to an improved competitiveness and consequently to a larger use of steel in buildings.

P.S.: Thanks are due to S. BRYL, research manager at Geilinger Ltd., Winterthur, J.F. GRIMAULT, engineer at Come tube, Paris, and W. KLINGSCH, professor at Wuppertal University. All these gentlemen are members of the TECHNICAL WORKING GROUP 3.2., a sub-committee of ECCS-TC3. 'By their own scientific and diversified work done in the fields of different composite column types, the author, chairman of TWG 3.2, was able to elaborate the present, general state of the art report on the FIRE SAFETY OF COMPOSITE COLUMNS.

-123-

BIBLIOGRAPHY (1) Klingeen,.W. ­ Tragverhalten von Verbundstützen aus einbetonierten Walzprofilen im Brandfall ­ Thyssen Technische Information, Duisburg, 1980. (2)

Schleich, J.B., Lahoda, E., Lickes, J.F. ­ TradeARBED; Une Nouvelle Technologie dans la Construction en Acier Résistant au Feu ­ Séminaire Nations Unies, Turin, Juin 1982.

(3) Grimault, J.F. ­ Calcul à l'Incendie des Profils Creux remplis de Bé­ ton; Manuel Pratique ­ Notice 1092 CSFTA; Chambre Syndicale des Fabri­ cants de Tubes en Acier, Paris, juin 1982. (4) Β ryl, S. und Keller, Β. ­ Über die Berechnung des Feuerwiderstandes von VerbundstUtzen mit Stahlkern ­ Schweizer Ingenieur und Architekt, Nr. 40, 1982 (5) Wicketroem, U. ­ TASEF/2, A Computer Frogram for Temperature Analysis of Structures Exposed to Fire ­ Lund Institute of Technology, Rep no 79,2 ­ 1979. (6)

SA MCEF ­ Système d'A nalyse de Milieux Continus par Eléments Finis ­ Laboratoire de Techniques Aéronautiques et Spatiales, LTAS, Université de LIEGE, 1982.

(7) Bathe, K.J. ­ ADINAT, A Finite Element Frogram for Automatic Dynamic Incremental Nonlinear Analysis of Temperatures ­ MIT Rep.82448,5­1977. (8)

Iding, Bresler, Nizamudin ­ FIRES / T 3, A Computer Program for the Fire Response of Structures, Thermal, 3dimensional ­ Report No UCB/FRG 77/15, University of California, Berkeley, 1977.

(9) Klingsch, W., Rudolph, K. ­ FIRES/T (BS) 7 revised version of "A Com­ puter Program for the Fire Response of Structures (2D) by Bresler"­SFB 148, Technical University of Braunschweig, W. Germany, 1978. (10) Rudolph, R., Müller, R. ­ ALGOL/Computerprogramm zur Berechnung zwei­ dimensionaler instationärer Temperaturverteilungen mit Anwendungen aus dem Brand­ und Wärmeschutz ­ BAM, Forschungsbericht 74, Berlin, 1980. (11) ECCS, Technical Committee 3 ­ European Recommendations for the Fire Safety of Steel Structures ­ Elsevier Scientific Publishing Company, Amsterdam, 1983. (12) Charlier, R. ­ Analyse de la charge critique d'une colonne mixte AF 30/120 par le Programme FLAMB 15 ­ Rapport interne, Service de Mé­ canique des Structures, Université de Liège, juin 1983. (13) FLAMB 15­Programme d'Instabilité de Poutres­Colonnes élasto­plastiques soumises à compression et flexion ­ Rapport interne, Service de Méca­ nique .des Matériaux et Stabilité des constructions, Université de Liège, 1983. (14) Klingsch, W. ­ KSTTR, Computer program for load bearing analysis of steel, reinforced concrete and composite columns in fire case (physi­ cal and geometrical non linear) ­ SFB 148, Technical University Braun­ schweig, 1975. (15) Quast, U., Hass, R., Rudolph, K. ­ STABA/F, a Computer Programme for the Determination of Load­bearing and Deformation Behaviour of Uni­axial Structural Elements under Fire A ction ­ Technical University Braunschweig, March 1984.

­124­

(15.')Jeanes, D.C. ­ FASBUS II, American Experience with Fire Endurance Computer Modelling of Steel Framed Floors, A ISC ­ International Conference "Fire Safe Steel Construction: Practical Design", Luxembourg, April 1984. (16) Dotreppe, J.Cl., Franssen, J.M., Schleich J.B. ­ Computer A ided Fire Resistance for Steel and Composite Structures, ECSC Research 7210­SA/502 ­ International Conference "Fire Safe Steel Construction: Practical Design", Luxembourg, April 1984. (17) ECCS ­ Composite Structures ­ The Construction Press, London and New York, 1981 (18) Deutsches Institut für Normung ­ DIN 18806, Verbundstützen, März 1984. (19) ECCS ­ European Recommendations for Steel Construction ­ 1978. (20) Kordina, Κ., Klingsch, W.­ "Brandverhalten von Stahlstützen im Verbund mit Beton und von massiven Stahlstützen ohne Beton"­ Forschungsbericht Ρ 35, Studiengesellschaft für A nwendungstechnik von Eisen und Stahl e.V., Düsseldorf, 1983. (21) Klingsch, W. ­ Traglastberechnung instationär belasteter schlanker Stahlbetondruckglieder mittels zwei­ und dreidimensionaler Diskreti­ sierung ­ Schriftenreihe des IBMB, TU Braunschweig, Heft 33, 1975. (22) Klingsch, W. ­Traglastanalyse brandbeanspruchter tragender Bauteile ­Kordina Festschrift, Forschungsbeiträge für die Baupraxis ­ Berlin, München, Düsseldorf; Wilhelm Ernst & Sohn, 1979. (23) Klingsch, W. ­ "Grundlagen der brandschutztechnischen A uslegung und Beurteilung von Verbundstützen" ­ Bauphysik; Wilhelm Ernst & Sohn, Berlin, H. 4., 1981. (24) Klingsch, W., Witte, H., ­ "Anwendung von Verbundstützen aus einbeto­ nierten Walzprofilen bei einem grossen Verwaltungsgebäude in Lüneburg" ­ Zeitschrift Acier­Stahl­Steel, Nr. 3, 1983. (25) Jungbluth, 0., Feyereisen, H. ­ Verbundkonstruktionen mit erhöhter Feuerwiderstandsdauer ­ International Conference "A Challenge for Steel", Luxembourg, September 1980. (26) Jungbluth, 0. ­ Untersuchungen zum Beulverhalten von Η­Profilen mit ausbetonierten Kammern ­ Gutachten in Nr. 98/83, Technische Hochschule Darmstadt, 29.1.83. (27) ARBED Forschungsprogramm ­ 4. Versuchsreihe; Überlange Stützen, Dübel­ verbund, Laschen­ und Knaggenanschluss von Träger an Stütze, April 1983. (28) Kordina, K., Wesche, J., Walter, R. und Hass, R. ­ Amtliche Material­ prüfungsanstalt für das Bauwesen, T.U. Braunschweig ­ Untersuchungs­ berichte und Prüfungszeugnisse Nr. 77150 R, 80341, 80644, 831009, 831016, 831025, 831032. (29) Kordina, K., Klingsch, W. ­ Institut für Baustoffe, Massivbau und Brandschutz, TU Braunschweig ­ Gutachtliche Stellungnahme Nr. 827585, 22.11.1982. (30) Jungbluth, 0., Hahn, S. ­ Traglastenkatalog für ARBED AF 30/120 Ver­ bundstützen auf Walzträgerbasis, 1984.

­125­

(31) Favre, J.P., Zumbühl, R. - Gutachtliche Stellungnahme für eine "Technische Auskunft" der VKF bezüglich AF 30/120 Stützen - Bern und Zürich, 30.9.83. (32) VKF - Technische Auskunft Nr. 4078 über ARBED AF 30/120 Verbundstützen - Brandschutzregister der Vereinigung Kantonaler Feuerversicherungen VKF, Bern, 1984. (33) Schleich, J.B., Hutmacher, H., Lahoda E., Lickes, J.P. - Breitflanschprofile mit ausbetonierten Kammern, ein wirtschaftliches und erprobtes Verbundsystem für die Brandbeanspruchung F3Û bis 120 - Statusseminar, Köln, 6.9.1983. (34) Schleich, J.B., Hutmacher, H., Lahoda, E., Lickes, J.P. - A new technology in fireproof steel construction - Revue Acier/Stahl/Steel, Nr. 3, 1983. (35) Schleich, J.B. - Le Système Mixte AF 30/120, une réponse pratique aux exigences de résistance au feu des constructions en acier - Conférence â l'ICOM, EPFL, Lausanne - 11.1.84. (36) Charlier, R. - Analyse Thermique Transitoire AF 90 - Rapport Interne, Service de Mécanique des Structures, Université de Liège, mai 1982. (37) Jungbluth, 0., - Optimierte Verbundbauteiie-Stahlbau Stahlbau Verlags GmbH, Köln, 1982.

Handbuch

1,

(38) Kreith, F., Black, W.Z. - Basic Heat Transfer - Harper & Row, Publishers, Mew Tork, 1980. (39) Dotreppe, J.Cl. - Thèse d'agrégation, Méthodes Numériques pour la Simulation du Comportement au Feu des Structures en Acier et en Béton Armé - Université de Liège, 1980. (40) Schneider, U. - Behaviour of Concrete at High Temperatures.Deutscher Ausschuss für Stahlbeton - Verlag Wilhelm Ernst & Sohn, Berlin,1982. (41) Anderberg, Y. - Behaviour of Steel at High Temperatures - Rilem Committee 44, PHT, February 1983. (42) ARBED-Rolling Programme Structural Shapes: European and American Wide Flange Shapes, 1984. (43) Jungbluth, 0., Heddrich, R.- AIF Forschungsprogramm, Trag- und Verformungsverhalten von Schraubenverbindungen bei brandgeschützten Stahlund bei feuerwiderstandsfähigen Verbundprofilkonstruktionen unter Brandeinwirkung - Darmstadt, 1983/84. (44) Klingsch, W., Nowak, R. - Verbundstützen; Interaktionsbeziehungen für Kaltbemessung - Forschungsbericht, Lehrstuhl für Baustofftechnologie und Brandschutz, Bergische Universität Wuppertal, 1983. (45) Guiaux, P. and Janas, J. - Comportement au flambement des tubes en acier remplis de béton - CRIF, MT 65, novembre 1970. (46) Roik, K., Bergmann, R., Bode, H., Hagenknecht, G.- Tragfähigkeit von ausbetonierten Hohlprofilstützen aus Baustahl - Ruhr Universität Bochum, TWM Heft 75-4, 1975. (47) CIDECT - MONOGRAPHY N° 5, Parts 1 and 2: Calculation of columns in concrete filled hollow sections - Paris, 1977. (48) Dowling, P.J., Janss, J., Virdi, K.S. - The design of composite steelconcrete columns - Introductory report to 2nd International Colloquium on Stability, Liège, 1977.

-126-

(49) Hasahide Tornii, Kenji Sakino ­ Inelastic behaviour of concrete filled square steel tubular beam columns ­ USA /JA PA N Seminar on Composite Structures and Mixed Structural Systems, 1980. (50) Bode, H. ­Developments in concrete filled tubular columns ­ USA/JAPAN Seminar on Composite Structures and Mixed Structural Systems, 1980. (51) Klingsch, W., Uürker, K.G. ­ Les constructions avec poteaux en profils creux remplis de béton, en A llemagne ­ CIDECT Proceedings, Nancy, 1981. (52) Beratungsstelle für Stahlvervendung ­ Betongefüllte Stahlhohlprofil­ stUtzen ­ Merkblatt 167, 1981. (53) Grandjean, G., Grimault, J.P. and Petit, L. ­ Détermination de la durée au feu des profils creux remplis de béton ­ ECSC Final Report, N e EUR 7171 RF (CIDECT document 15 Β/80­10), 1980. (54) Hikaru Saito, Hideki Uesugi ­ Fire resistance of concrete filled steel columns ­ USA /JA PA N Seminar on Composite Structures and Mixed Structural Systems, 1980. (55) Grandjean, G. and Lelong, C. ­ Tenue au feu des profils creux carrés remplis de béton ­ Annales de l'ITBTP, N" 347, Février 1977. (56) Grimault, J.P. ­ CIDECT Document 15B­15C­83/6­FIRE STA BILITY 0F CONCRETE FILLED HOLLOW SECTION COLUMNS ­ New proposal for taking into account the influence of small eccentricities or small moments ­ Cannes, 30.5.83. (57) Grimault, J.P. ­Stabilité au feu des poteaux en profils creux remplis de béton ­ Annales de l'Institut Technique du Bâtiment et des Travaux Publics, n° 423, mars­avril 1984. (58) Grimault, J.P. ­Dimensionnement des poteaux en profils creux remplis de béton ­ Notice N° 2029 of COMETUBE, avril 1981. (59) Quast, U., Rudolp, K. ­ Baupraktische Bemessungshilfen für den Brand­ sicherheitsnachweis von Verbundstützen ­ Forschungsbericht zum Vorha­ ben Ρ 86/2.3/BMFT/SAES,1984. (60) Bailly, R. ­ Construction Mixte Tube Béton, 4 m e solution, CMTB 4­ Notice 1105 C S F Τ A, Paris, octobre 1982. (61) Klingsch, H., Würker, K.G. ­ Hohlprofil­Verbundstützen, Sichtbarer Stahl für feuerwiderstandsfähige Konstruktionen ­ DBZ, Bertelsmann, H.H., 1982. (62) Krapfenbauer, R. ­ Die Verwendung von Kernblockstützen beim Neubau des Allgemeinen Krankenhauses in Wien ­ Bauingenieur 51, 1976. (63) EMPA Test Nr. 42573 ­ Statische Druckversuche an zwei Geilinger Bau­ stützen ­ EMPA, Dübendorf, 1980. (64) Amtliche Materialprüfanstalt für das Bauwesen, Braunschweig ­ Unter­ suchungsbericht Nr. 83058, 1983. (65) Brandsichere Geilinger­Stahlstützen mit hoher Tragfähigkeit ­Geilinger Ltd, Winterthur, 1982.

­127­

FIRE ENGINEERING DESIGN OF COMPOSITE CONCRETE SLABS WITH PROFILLED STEEL SHEET

by L. Twilt Institute TNO for Building Materials and Building Structures Delft, The Netherlands SUMMARY The behaviour of fire exposed composite concrete slabs with profiled steel sheet is discussed. On basis of experiments carried out in various European fire test laboratories, it is concluded that the load bearing capacity of this type of composite system is sufficient for a fire resistance of at least 30 minutes, also if no specific means of fire protection are applied. Condition is that the design at room temperature complies with an approved method. To check the other fire resistance criteria (insulation, integrity) a simple verification rule is presented. If the required fire resistance is over 30 minutes, additional means of fire protection may be necessary. In this respect the following possibilities are briefly discussed: - additional reinforcement - insulating coating - suspended ceilings It follows from this discussion that - for various reasons preference is for additional steel reinforcement to increase the fire resistance of composite slabs. This concept is therefore further evaluated. The evaluation includes general fire engineering considerations as well as the presentation of practical calculation rules for minimum slab tickness and additional steel reinforcement to meet given fire resistance requirements. These rules apply to normal weight concrete and lead to conservative solutions.

-128-

1.

INTRODUCTION The composite concrete slab with profiled steel sheet is the type of

composite system most frequently found in buildings today. The fire resistance of composite slabs is significant, even if no additional fire safety precautions are taken. If necessary the fire resistance can be increased to practically any desired load level by simple and reliable means. Until recently, however, the structural fire engineering design of composite slabs could only be based on fire resistance tests. This procedure is time consuming and expensive, and sometimes gives rise to anomalies due to variation in test results. Consequently, there is a strong need for a practical design method, by which the fire resistance of composite slabs can be determined analytically and this should lead to more uniform levels of safety. Furthermore it leads to a more simple and systematic design procedure, thus stimulating the use of composite slabs. Such a design method, established by Technical Committee 3 of the European Convention for Constructional Steelwork and derived for normal weight concrete, is available now (1) and will be reviewd.

2.

CRITERIA

FOR FIRE RESISTANCE

Fire resistance is determined under standard fire conditions, characterized by the so­called standard gas­temperature­time curve. This curve is shown in Fig. 1.

rcoc-r #>° » o o ­

-Γ "Js d. T

200

-fr time l (min.) Fig. 1 : The standard fire curve Composite steel­concrete slabs have both a load bearing and a separating function, and the following criteria for fire resistance shall therefore be taken into account: (2).

­129­

(1)

LOAD BEARING CAPACITY : Resistance to collapse or excessive deflection

(2)

INSULATION

: Limitation of the temperature increase on the

(3)

INTEGRITY

: Ability of the slab to resist penetration of

under structural loading unexposed side of the slab flames or hot gases through the formation of cracks and openings. The time taken to fail any of these 3 criteria is taken as the fire rating of the slab, even though failure under other criteria may not occur until much later. It is common practice to determine the fire resistance by means of standard fire resistance tests. During such tests, the test specimen is exposed to the standard fire on the underside while loaded with a load calculated to produce the normal maximum working stresses in the floor construction. Similar assumptions are adopted when using the analytical approach. For a proper verification of the performance criteria in an analytical approach however, some additionals assumptions are necessary. The criterion of load bearing capacity requires that the slab shall not cease to perform the load bearing function for which it was constructed. During tests, collapse of the slab is avoided so as to prevent damage to the furnace and other apparatus. This is achieved by limiting an excessive deflection. From a functional point of view, the failure condition is prefered. Since this publication deals with an analytical approach rather than with experiments, the load bearing criterion is used here. In order to fulfill the insulation criterion, the temperature rise of the unexposed side should not exceed 180*C at any point and the average should not exceed 140*C. This criterion, given in (2), is applied in most national standards. Because of the profiled shape of the slab, care must be taken when checking that the insulation criterion is satisfied. Theoretically the temperature at the unexposed side will vary at a function of the place at which the temperature is measured. Tests show however that in practical cases the temperature differences are small. In theory it is also possible that passage of heat

through joints may result in a non

uniform temperature distribution at the unexposed side. However, composite steel concrete slabs are normally manufactured in situ and this

-130-

complication does not arise. In this publication therefore a uniform temperature distribution at the unexposed side is assumed. A temperature increase of 140"C at this side is taken as the limiting insulation criterion. Integrity is a measure of the ability of the construction to resist the passage of flames and hot gases through cracks. For composite steel concrete floors the integrity criterion is not difficult to fulfill. The main reason is that, as mentioned before, the floor slab is cast in situ. This means that joints are adequately sealed. Any cracks which may occur in the concrete during fire exposure are unimportant because the steel sheet will prevent penetration by the flames and hot gases. Therefore it is assumed here that if the insulation criterion is fulfilled, then the integrity criterion is also fulfilled. 3.

BEHAVIOUR OF FIRE EXPOSED COMPOSITE SLABS WITHOUT SPECIFIC MEANS OF FIRE PROTECTIONS The steel sheet of a composite slab is normally designed to transmit

the tensile stresses due to positive bending moments when at ambient temperature. When exposed to fire, the temperature of the steel sheet will increase, and consequently the mechanical properties such as yield stress and elastic modulus will decrease. At a certain temperature, which is dependent on the load level and the statical system, the steel sheet is no longer able to transmit the applied tensile force and as a result the slab fails the criterion for load bearing capacity. Such a failure may, dependent on the thickness of the slab, be preceded by failure under the insulation criterion. Table I gives results of fire resistance tests on composite concrete slabs with profiled steel sheet, conducted in various European fire test laboratories. No additional means of fire protection was present. The tests cover a practical range of application. It is seen that In all cases the fire resistance is governed by the criterion for load bearing capacity and is over 30 minutes. The considerable scatter is caused by differences in design assumptions at room temperature and in the support conditions. All tests, except fort Test 1 and 10, were conducted on simply supported slabs.

-131-

Such a static system obivously constitutes relatively unfavourable conditions» since no beneficial moment redistribution or catenary force can occur as It often does in practice· 1) te v >

UM

1M4

F i n n t l e u a e F i n fMlataM· Ol tari·· nsAMM^ Ima Mena« Mptcltr

r'iöfc ­A*

ι

π

{'h

τφ·

­»—Η» Μ3Μ/2 (MCMtl^)

■t;

f.J

_rz rz_3> Q»

r'fe

T^?

t'fc

■ΛJ imn·)

inumi

•Τ

ItaMl

è'fe

T4fMftU

­| "" | ■"u·1 g

f

f»t

rtί

M.U7U (ΠΜΜ)

<*>

1»

1

l'k

4

f'fe ; '" ; on M.JJW

"î =

LT

¿•fe ţflC 3BB , Β ^

i'fe

ad UM· (MCUtlM«)

¿un

)■ ;,B;

I^Q^y^ t

t'fc -MS

f-

I) dense concrete unless otherwise stated. Table I Fire resistance of composite concrete slabs with profiled s t e e l sheet without additional means of fire protection.

-132-

Table I therefore suggests the following general practical design rule: "The fire resistance of composite concrete slabs with profiled steel sheet without additional means of fire protection is at least 30 minutes when assessed under the criterion for load bearing capacity". The application of this rule should obviously be restricted to those cases in which the design at room temperature is based on an approved method. The European Recommendations for the Design of Composite Floors with Profiled Steel Sheet (3), could be taken as a reference. It is also necessary to check that the insulation criterion is fulfilled. The simple verification rules presented in chapter 5 can be used for this purpose. As a direct consequence of the above design rule, an explicit analysis of the load bearing capacity of fire exposed composite concrete steel slabs is only necessary for requirements over 30 minutes. Additional means of fire protection may then be necessary. 4.

ADDITIONAL MEANS OF FIRE PROTECTION The following means of additional fire protection can be distinguished: - additional reinforcement - Insulating coatings - suspended ceilings Additional reinforcement In a composite slab some steel reinforcement (say: 100 mm /m width of

slab) is normally included to control shrinkage and creep of the concrete. This reinforcement may be placed directly upon the steel sheet. When no coating or suspended ceiling is used, the steel sheet is directly exposed to fire. As a result, the temperature increase in the steel sheet and in the reinforcement can be expected to be approximately the same. Consequently, the beneficial effect of the shrinkage reinforcement on the fire resistance may only be marginal. However, additional reinforcement placed in the centre of the ribs may significantly contribute to the fire resistance. The same holds for the top reinforcement over Intermediate supports of continuous slabs. Fig. 2 illustrates the various possibilities.

-133-

1 - additional bottom reinforcement in the ribs 2 - top reinforcement used over supports in continuous slabs 3 - reinforcement against shrinkage Fig. 2 Reinforcement in a composite concrete slab with profiled steel sheet.

H - structural height u - concrete cover (axial)

Coatings Insulating coatings may be necessary when extremely high fire resistance ratings are required and/or when deflections have to be severely limited under fire exposure. Sprayed coatings (e.g. based on mineral fibres or vermiculite) are directly applied to the surface of the steel sheet. In order to achieve good adhesion, the steel surface should be properly cleaned to remove dirt and grease. Fire protecting boards can also be used (e.g. based on vermiculite, gypsum, fibre). These are then directly adhered or mechanically fixed to the ribs of the steel sheet. As with sprayed coatings, a thorough cleaning is necessary to ensure good adhesion. Special attention should be paid to the type of adhesive used and to adequate connection of boards under fire conditions. These aspects must be verified by experimental evidence. Only a relatively small thickness of insulation is necessary to achieve a considerable fire resistance (A). Nevertheless, the application of fire protecting coating will involve considerable extra cost.

Suspended ceilings A suspended ceiling functions as a heat shield for the structural components above and thus can contribute to the fire resistance of the floor assembly. In the cavity over the fire protecting ceiling a timetemperature curve which is less severe than the standard fire curve, can be taken into account, subject to the condition that there is only a limited amount of combustible material in the (unventilated) cavity. The extent to which the standard fire curve is thus reduced, will depend on the quality of the celling and on the floor above it. The behaviour of a suspended ceiling during fire, however, is critical since it depends very much on good detailing, workmanship and maintenance. These aspects have to be verified experimentally and fire resistance tests are therefore essential.

-134-

Summarizing, it can be concluded that additional steel reinforcement is a simple, reliable and economic device to increase the fire resistance of composite concrete steel slabs. Moreover, the assessment of the effect of additional reinforcement is open to a theoretical analysis. The reliability of the two other means of fire protection, i.e. coatings and suspended ceilings, is much more critical, being highly dependent on factors such as detailing and workmanship. Experimental verification is deemed to be necessary in these cases.

5.

CALCULATION RULES FOR MINIMUM SLAB

THICKNESS

The insulation criterion of fire resistance is fulfilled if the average temperature increase at the unexposed side of the slab exceeds 140 *C. See 2. This requires a sufficient effective thickness of slab, which will depend on the period of fire resistance required. B ased on experimental data, the following, conservative, rules for effective slab thickness can be given ( 5 ) .

Required fire Minimum effective resistance Hin thickness h_ mm

30 60 90 120

60 70 80 100

Equation for effective thickness

h

e

= h

h

1, + 1, 1 + 7- ■ TT

Restrictions

for h2/h.>1.5 ht >50 mm r^'-HT-'J

Ί

Ί

Table II : Effective thickness of a composite concrete slab with profiled steel sheet as function of the fire resistance time.

As is seen from the equation for h e , the effective thickness corresponds to an arithmetical average of the thickness which takes account of the profiled shape of the slab. The calculation rule applies to standard fire exposure and normal weight concrete.

-135-

6.

CALCULATION RULES FOR ADDITIONAL

REINFORCEMENT

6.1 Failure conditions The load bearing capacity may be analysed on the basis of elementary plastic theory (limit state design). For various statical systems the failure conditions can then easily be formulated

Statical system

CU E •t- tu

(Table I I I ) .

Failure condition

Mu J *q ■ L V 8

* * M u3

c

f/L

q >_8 . My

M W­»

0.5 . Mu < q . L'/8

■Φ—w M u.»

q > (8M+

+ 4M"

)/L2

Ό ε

ω ω υ > i­ ■r­ O +■» 4 ­

m c: σι­ι­ υ
Mu j

■«

(Λ O O 4­ CX C

M

H;

W

M a» ■

»

+ Hu < q . L'/8

q > 8 . (M*

M,u»

Si ­ι­ i­

M u.»

Mu

+ M~

)/LZ

4 q · LV8

■

q > 8 . Mu

Mu>=0

/L'

Table III : Failure conditions for slabs In Table III the following notation is used: M

, M

­ absolute value of the positive and negative plastic bending moment respectively at the end of the required period of standard fire exposure

q .

­ load on the slab to be accounted for during fire

L

­ span of the slab.

To evaluate the failure conditions it is necessary to quantify the plastic moments M

, and Μ ,. Typical stress distributions over the cross uJ uJ

section are represented in Fig. 3 and 4 respectively. The following simplifying assumptions are made:

­136­

OB <¡c20*)

+-M

Z j = e y j.A r

=

u.j

Z

S-Z

F i g . 3 : P o s i t i v e p l a s t i c marnent

Z = Dj M^»Z.z OAcc¿*) Fig. A : Negative plastic moment M ~ U Q General: - the tensile strength of concrete does not contribute to the load bearing capacity at elevated temperatures and thus may be ignored, - the steel sheet does not contribute to the load bearing capacity at elevated temperatures and thus may be ignored. For the positive plastic moment: - the ultimate strength of concrete in the compression zone is not influenced by temperature and the room temperature values may be taken. - the effective yield stress of the additional reinforcement is affected by temperature For the negative plastic moment: - in calculations, the profiled concrete slab may be replaced by a slab with a uniform thickness equal to the effective thickness h

in

accordance with Table II. - the ultimate strength of concrete in the compression zone (exposed side) is affected by the temperature. - the effective yield stress of the reinforcement (unexposed side) is not influenced by temperature and room temperature values may be used. *) The factor 0.8 is introduced to correct for the assumed full plastic stress distribution in the concrete compression zone. In the ultimate state a non uniform stress distribution will occur, due to the limited capacity of the concrete to accept deformation.

-137-

The load on the slab to be accounted for during fire (­ q) follows from 0.85 χ q

where q

Is the load to be used In fire test and chosen In

accordance with ISO 834 (2). The reduction Is motivated since, due to the various simplifying assumptions, the calculation rule give ­ compared with the amount of fire resistance measured in a fire resistance test ­ a conservative result. The actual value of the reduction factor (i.e. 0,85) is based on comparative calculations. See also (6).

On basis of the above mentioned assumptions and using the information on temperature distribution and mechanical properties of steel and concrete at elevated temperature as presented under 6.3 and 6.4 respectively, the evaluation of the failure conditions can proceed in a similiar way as for conventional reinforced concrete slabs under ambient temperature conditions. First however, additional design consideration should be discussed.

6.2 Design considerations In statically indeterminate slabs a redistribution of moments will occur during the period of fire exposure. This phenomenon will be discussed for the continuous slab presented in Fig. 5 a . The moment distribution at room temperature under working load conditions is shown in Fig. 5 . .uniformly diatri bute d lood q

(a) Statical system (b) Moment distribution at room temperature

(c) Moment distribution in an initial stage of fire exposure

(d) Moment distribution at failure

¿* 2 Mu*

W2

M u.»

­X

MuA.r Fig. 5 Moment redistribution in a continuous slab during fire exposure.

­138­

Directly after commencement of the fire exposure, a steep temperature gradient will be attained over the height of the slab due to the relatively low thermal conductivity of the concrete. Consequently, additional negative bending moments will develop, which relieve the positive moment in the mid span region but increase the negative moment at the supports. Since, in the first stage of fire exposure, the value of the full plastic moment at the supports will not be affected significantly by temperature, a moment distribution as presented in Fig. 5 C will tend to occur (conservative assumption). The additional negative reinforcement should then be designed to cope with such a moment distribution. This means that the additional reinforcement at the supports shall be extended at least over a distance L' (cf Fig. 5 C ) . The anchorage length should be determined in accordance with room temperature design. For other statical systems, other minimum lengths will apply. When heating continues, both the (possitive) plastic moment capacity at mldspan and the (negative) plastic moment capacity at the supports will decrease, finally leading to failure. The moment distribution at failure is presented in Fig. 5

and corresponds to the

relevant condition given in Table III. To arrive at such a moment distribution, sufficient rotation capacity is necessary, especially at the supports. The amount of the negative reinforcement and its ductility are then of crucial importance. The present state of knowledge does not allow the formulation of criteria, specifically derived for fire circumstances. In the room temperature design, however, a limit is normally set for both the maximum and the minimum amount of negative reinforcement in order to guarantee adequate rotation capacity at the supports. When these rules are obeyed, it will be assumed here that under fire conditions the necessary moment redistribution is also possible. Also the ductility of the reinforcement steel should meet room temperature specifications. 6.3 Temperature distribution Concrete The temperature distribution in the concrete slab is assumed to be independent of the effective thickness h

and can be derived, for various

times of the standard fire exposure and normal weight concrete, from Table IV (7).

-139-

Depth Χ

Ţeaperacure in U C after a fire dura­ ción (min) of:

9B

60

S

705

10

90

120

XA

642 738

15

581

681

754

20

525

627

697

25

469

571

642

30

421

519

591

35

374

473

542

40

327

428

493

45

289

387

454

50 55

250 345 415 200 294 369

60

271

"■»e ", Γ"

he

/

/

J= ΓΛ \

/

^_

342

Table IV Temperature distribution in the concrete. Additional reinforcement The temperature of the additional reinforcement depends on the position of the reinforcement bars and the shape of the steel sheet profile. Both factors can be represented by the coefficient γ which is given by following equation (Fig. 6): (5)

L + 1 ­ +_2_ 1

^

^

The distances u,, u. and u, shall be taken in mm.

Fig. 6

Ϊ^ΓΖ

Calculation of the coefficient γ.

The temperature of the reinforcement bars can be calculated using the following equations: ­

fire duration

60 min:

1175

90 min: & ­ 1285 s fire duration 120 min: tì' » 1370

fire duration

350.Ύ _< 810 'C

(Ύ <. 3,3)

350.γ <^ 880 'C

<Τ <.3,6)

350.γ < 930 *C

(Ύ < 3,8)

8

­140­

6.4 Mechanical properties at elevated temperature Concrete The following, approximate relation between the ultimate compressive strength σ a, and the temperature *

u­ < 200 'C; c 200 < *· < 700 'C; — c — θ­ > 700 'C; c

of concrete may be used ( 8 ) .

θ' ­ 200

"eft­ " °cfr (1 ­ 0,8

~5ÕÕ~) '

α

* °c20

°c*-°

where σ .­ ­ compression strength of concrete at ambient temperature.

The strength of the compressive zone of a slab (Fig. 7) with a width of 1000 mm can be calculated as follows: i

*c — \ l Fig. 7

ff

c*

ιI

\ \

k

—

S* i ι

e

^

,i

Resultant compressive force in the concrete compressive zone.

D,. » 0.8 Σ(σ „.dx.1000) ­ 0.8 a ,. . Z(a.dx.lOOO) ­ 0.8 σ „_ . A with * cir c20 c20 cr e ­ (Σ(σ ö,.dx.l000).x)/Dd.

The reduced compressive area A

in mm

and the position of the compressive

force e in mm are given in Table V.

Steel The following approximate relation between the effective yield stress σ ­, and the temperature θ" of the additional reinforcement may be used: ( 8 ) . 6­ < 250 *C O j . « 1.0 σ ,„ — yo y20 a­ _ 2 S 0 250 'C < & < 650 'C a », ­ σ Q . (1 ­ 0.6 . — * ) "„*. " "„in · d ~ 0.8 . — 7 T T — ) where: a _­ ■ yield stress at room temperature.

­141­

for hot rolled bar for cold drawn bar

Total depth

Values of A

of compressive zone

•

Table V

7.

e

A

cr xlO" 2 ¿

-

and e after a fire duration in min. 90 120

60 A

cr xlO ­ 2

¡i

■

a ■B

A

cr xlO­2 «BZ

a

™

10

12.1

7.5

15

29.1

10.4

20

50.9

13.5

13.7

17.5

25

77.1

16.5

31.8

20.3

12.4

22.5

30

107.5

19.6

54.2

23.3

29.0

25.4

33

141.7

22.7

80.5

26.3

49.7

28.3

40

179.6

25.9

110.4

29.3

74.2

31.4

45

221.0

29.0

143.8

32.4

102.3

34.4

SO

265.4

32.1

180.5

35.5

133.5

37.5

55

313.4

33.2

220.9

38.6

168.2

40.6

60

363.4

38.3

264.2

41.7

205.7

43.7

65

309.6

44.7

245.4

46.7

70

357.0

47.8

287.2

49.7

75

405.9

50.7

331.4

52.8

80

455.9

53.7

376.7

55.7

Strength of the concrete compressive zone at elevated temperature.

CONCLUSIONS It follows from the above discussions that an explicit analysis of the

load bearing capacity of fire exposed concrete slabs with profiled steel sheet is not necessary, if the required fire resistance is not over 30 minutes and the room temperature design complies with an approved method. For a check on the other fire resistance criteria (insulation, integrity) the simple verification rule for minimum slab thickness as presented in chapter 5 may be used. When the required fire resistance is over 30 minutes, additional fire protection may be necessary. In this respect, preference appears to be on additional reinforcement. The practical calculation rules presented in chapter 6 allow for a simple verification whether the amount of additional reinforcement is sufficient to meet given requirements of fire resistance. The method of assessment is derived for normal weight concrete and gives conservative solutions.

­142­

ACKNOWLEDGEMENT This contribution is based on the Technical Note "Calculation of the Fire Resistance of Composite Concrete Slabs with Profiled Steel Sheet Exposed to the Standard Fire", prepared within Technical Committee 3 of the European Convention for Constructional Steelwork. The author is grateful for the stimulating discussions in this committee. More particulary he thanks S. Bryl and J. Kruppa for their important contributions. REFERENCES (1)

'Calculation of the Fire Resistance of Compsite Concrete Slabs with Profiled Steel Sheet Exposed to the Standard Fire'. European Convention for Constructional Steelwork, Committee T3, 1984.

(2)

'Fire resistance teste - Elements of building constructions'. International Standard ISO 834, first 1975.

(3)

'European Recommendation for the Design of Composite Floors with Profiled Steel Sheet'. European Convention for Constructional Steelwork, Committee Til. Constrado, London 1975.

(4) Muess, H. 'Brandverhalten von bekleideten Stahlbauteilen'. Stahlbau-Verlage-GmbH, Köln 1978. (5)

Kruppa, J. Echauffement des plancers béton â bac acier soumis ä l'incendie conventional, CTICM-Paris, 1983.

(6)

Pettersson, 0. and Witteveen, J. 'On the critical temperatures of steel elements derived for conventional fire resistance tests and from calculations'. Fire Safety Journal 2, 1979/1980, Elsevier Sequoia SA, Lausanne.

(7)

'FIP/CEB Report on Methods of Assessment of the Fire Resistance of Concrete Structural Members'. FIP Commission on the Fire Resistance of Prestreseed Concrete Structures, 1978.

(8)

Comité Euro-International du Beton. 'Design of concrete structures for fire resistance'. (First draft of an appendix to the CEB/FIP Model Code for Concrete Structures), Bulletin 145, Paris 1982.

-143-

FIRE RESISTANCE OF COMPOSITE STEEL DECKS, FLOORS AND BEAMS H.B. WALKER, M.Univ. C.Eng. M.I.C.E. M.I.Mech.E. M.R.Ae.S. Constructional Steel Research and Development Organisation Croydon, Surrey, United Kingdom

Summary There has been a considerable upsurge in the United Kingdom in the use of steel for the frames and floors of multi-storey buildings, largely brought about by the relatively lower prices for steel and the faster construction times that can be achieved. An important aspect is the use of profiled steel decks and beams acting compositely together. Propping during construction is not required, thus allowing following trades to commence work without delay. The performance of this floor system in fire has been a critical factor, and this has now been established using relatively simple methods based on an extension of current reinforced concrete design methods. The advantages of using lightweight concrete are discussed. Calculation methods approved by National Authorities based upon Limit State conditions are given. Partial factors are listed as well as tables for the thickness of concrete, temperature distribution and material strengths at elevated temperatures. Beams acting compositely with the floor via insltu welded connectors are discussed and the different stress paterne for composite and non-composite beams are explained. A computer model has been evolved and the program output is presented graphically. Recommendations for the protection of composite beams are given.

-144-

1.

INTRODUCTION Over recent years there has been a considerable upsurge in the U.K.

in the use of steel for the frames and floors of multi-storey buildings. This has come about for a number of reasons, principally a change in the relative costs between steel and concrete construction, the faster construction times that can be obtained with steel, problems with the Integrity of concrete construction and the relatively high interest rates chargeable on the capital required for construction. The Constructional Steel Research and Development Organisation, or CONSTRADO as it is generally known in the U.K., has been very active in this field, and has investigated several design aspects in depth.

There

has been an active liaison with property developers and an extensive study has been made of the costs and charges involved in constructing floors and frames as well as for the complete building. The fire resistance of steel-framed structures is always a matter of concern to architects and engineers, and CONSTRADO has developed considerable expertise in this field, and is able to demonstrate that this functional aspect is simply achieved and is dealt with in a straightforward manner. 2.

FLOOR SYSTEMS

Floors can be provided for steel-framed buildings in a number of different ways.

In the past ineitu concrete was very often used but this

requires formwork and props and can be a relatively slow process. Precast concrete slabs is another method, but they have a number of problems and require concrete screed topping with steel reinforcement, to give a level floor which is capable of carrying the lateral shear loads from wind or from earthquakes. Whilst in the U.K. a number of Important buildings have been constructed using precast floor slabs, the trend now is towards the use of profiled steel decks with concrete topping power-floated to a final finish.

This method is proving to be very fast and economic, and has an

inherent simplicity which appeals to the architect, engineer and construction manager. The profiled steel deck is used unpropped with supporting beam centres between 2.4m to 3.6m centres.

The steel deck is placed, and to

prevent movement due to wind, lightly fixed to the supporting beams by means of self-drilling and tapping screws or by puddle welding, and then

-145-

shear connectors are insitu welded through the deck on to the beam flange to enable the floor slab to act composltely with the support beams.

Light

steel reinforcement mesh Is laid upon the deck and then the concrete topping Is placed, being power-floated to a final finish, see Figure 1.

Figure 1.

Arrangement of Floor

An important point to remember when considering the deck systems is that the unpropped steel deck system is the only one which allows the floor to be power-floated to a final flat finish during the Initial construction stage. Precast concrete slabs must be screeded, and Insitu concrete will deflect when the formwork is removed, necessitating a second operation. Lightweight concrete is very often used in preference to normal weight concrete as it has several distinct advantages.

Lightweight

concrete has better performance in fire than normal weight concrete, and this enables the depth of the floor to be reduced.

This reduction in

thickness, coupled with reduced weight, means that the resultant concrete component of the floor is nearly half the weight when compared with normal weight concrete.

This allows a thinner gauge steel deck to be used, and

also reduces the load on the supporting beams, and then through the

-146-

collons, reducing their size as well as having an effect on the foundations.

The profile and thickness of the deck is determined by the

span and the weight of the wet concrete plus equipment in the construction stage.

Steel deck which has indentations in the sloping webs gives full

composite action independent of whether there is a chemical bond between the concrete and the steel deck or not. 3.

U.K. NATIONAL BUILDING REGULATIONS & STANDARDS The National Building Regulations enable the architect to

determine the fire rating required for the building in terms of size of the building and the use to which it has to be put.

The materials that

are used in the building then have to meet the testing requirements set out in British Standard BS.476.

Alternatively, other methods may be used

to satisfy the Authorities, but they have to be backed up by calculations and data.

For composite floors and beams Part 8 of BS.476 has to be

complied with.

This requires that when the element or sub-assembly is

tested in the furnace to the standard time/temperature curve, three criteria must be satisfied. 3.1

These are:-

Stability This is the ability to support the load whilst the deflection is

limited to span/30 and the ability to support the load 24 hours after the test. This limiting deflection of span/30 is expected to be modified. In the Draft European Code EEC 1202 a rate of change of deflection is specified.

This rate is such that the actual failure under test

conditions and the notional failure according to the code will occur very closely to each other.

It is expected that this criteria will be adopted

in a forthcoming revision to the British Standard. 3.2

Integrity This is the ability to resist the passage of flame and hot gases

and is ensured with composite steel decks by the combined action of the diaphragm formed by the steel sheet and the mesh reinforced concrete. 3.3

Insulation This is the ability to resist the conduction of heat, and normally

means that the temperature on the top surface of the composite floor is

-147-

limited to an average of 140°C with no Individual value of more than 180°C.

This Is normally ensured by the provision of an adequate

thickness of concrete. It Is worth noting that In a composite steel deck floor the fire reinforcement normally has much greater concrete cover than Is obtained In a reinforced concrete slab.

This means that the fire reinforcement is at

a comparatively low temperature and only suffers a minimal loss of strength.

For Instance, for one hour's fire resistance the fire

reinforcement temperature may be no more than 250°C and for two hours, 370°C. 4.

CALCULATION OF FIRE RESISTANCE FOR COMPOSITE FLOOR The methods given below which have been developed by Constrado are

based upon standard techniques for the design of steel reinforced concrete and make use of existing British Standards and established publications by the Institution of Structural Engineers.

In extending R.C. design methods

at ambient temperatures to predict performances at elevated tempertures no new theoretical concepts have been Introduced. The calculation method compares favourably with actual fire tests carried out in the standard European manner without external restraints. For a calculation method to be usable by the Engineer it must be accepted by the Building Authorities when designs are submitted for Approval and because this method is only an extension of existing methods this acceptance has been readily given. 4.1

Design Concept A composite steel deck floor Is unique in being one of the few

elements in a building which is very largely designed on the basis of the required fire resistance.

The steel deck profile Itself is selected on

the basis of its ability to carry the weight of the wet concrete and equipment during the concrete placing operation.

The thickness of the

topping depends upon the thermal resistance of the concrete and the maximum temperatures allowed on the top surface.

With these conditions

met, and with the shear resisting indentations In the deck, there is usually a high degree of imposed load-carrying capability, normally much bigger than Is actually required by the building specification. For office buildings it is normal to have suspended ceiling systems and, whilst sometimes the celling systems can be made fire

-148-

resistant, it is usual for only the beams and column tops to be sprayed with fire-resistant materials.

It would be extremely expensive if it were

necessary to spray the soffit of the composite floor, an operation which would also make it extremely difficult to subsequently fix the normal suspension wires for carrying services and ceilings. In order to meet the requirements of the Building Authorities it is necessary to provide a method of calculating the strength of a composite floor in fire. When a composite floor is tested in a furnace it is usual for the steel deck to expand and separate away from the concrete. With the trapezoidal-type profiles this tends to happen over the entire surface, but with dovetail-type profiles the dovetail remains embedded in the concrete and may contribute to the bending resistance of the floor. However, the deck is usually discounted and the concrete, together with the anti-crack/fire reinforcement mesh, is then considered as a reinforced concrete slab and calculations are carried out in a manner appropriate to that form of construction.

The deck, however, plays another important

part in contributing to the fire resistance of the floor in that it acts as a diaphragm preventing the passage of flame and hot gases, as well as acting as a shield reducing the flow of heat into the concrete and helping to control and contain the concrete spalling.

As time and temperature

continue to increase the reinforcement will yield and the floor will hang as a catenary with the reinforcement assisted to some extent by the steel deck acting in tension supporting the loads. This catenary condition is known to exist within the boundaries of the slab and will sustain higher loads than the calculation method based purely on bending strength currently being used and can be considered as providing an additional safety factor. 4.2

Fire reinforcement The arrangement of reinforcement within the concrete requires

careful consideration both from the structural and economic standpoints. Some arrangements of the mesh are illustrated in Figures 2a, 2b and 2c. Mesh can easily be obtained from the suppliers to match the pitch of the steel deck at little extra cost.

-149-

-υ—υ—υ—Ό—πTYPICAL CROSS SECTION Figure 2a. Simply supported design

r j

Overlapping Mesh

j

ENLARGED PLAN OVER SUPPORT lote: The transverse wires are discontinued allowing the longitudinal wires to mesh together and lay side by side. Overlapping Mesh

TYPICAL CROSS SECTI ON

\Z7

\ SECTION AT INTERNAL SUPPORT Figure 2b.

I SECT ON AT MID-SPAN

Continuous design using draped mesh

-150-

TZ!Z7—ţ-^J. TYPICAL CROSS SECTION Figure 2c.

4.3

Continuous design using 2 layers of reinforcement

Limit State Principles Design for fire is based upon ultimate Limit State principles, the

floor being considered in bending, either as a simply-supported or as a continuous element. The following partial factors which are taken from "Design and Detailing of Concrete Structures for Fire Resistance", published by the Institution of Structural Engineers, are suggested (see Ref.l).

4.3.1

4.3.2

4.4

Loads Dead load

Yfd -

1.05

Imposed load

Tfi -

1.00

Materials Reinforcement

Y

Concrete

ymr.

mr -

1.00 1.30

Design Bending Moments f o r Continuous Construction As the d e s i g n i s based upon u l t i m a t e c o n d i t i o n s r e d i s t r i b u t i o n o f

moments may be assumed.

The bending moment diagram f o r any i n t e r n a l span

i s shown i n F i g . 3 a .

-151-

MF

\

^

w

/. '/

MH MS

Figure 3a

The only condition is that:M

H

+ M

S — MF> t'ie ^ ree bending moment.

Where: MT,

• Free bending moment per metre width " £ (Yfd.Fd + Yfi-Wi) 8 -

Hogging moment of resistance per metre width, kNm Sagging moment of resistance per metre width, kNm

m

Beam centres (- floor span), m Total dead load, kN/m2 Imposed load , kN/m

The bending moment diagram for any end span is:

Figure 3b. As the ratio of Mjj to Mg is important, and this is a more complex situation to analyse than that of Internal spans, a solution may be found graphically or use may be made of the following approximate formula for which within the stated range the maximum error is not greater than 2%. (See Ref 2 Appendix). O.AlMjj + 1.05MS >.

MF

For 0.35 <_ MJJ <_ 3.3

-152-

4.5

Concrete Thickness The rise of temperature on the upper surface of the composite

floor at the end of the fire resistance period is limited to an average of 140°C with no local reading being more than 180°C, and this determines the required thickness of the concrete topping.

Lightweight concrete has

considerable advantages requiring on average only two thirds of the thickness of normal weight concrete.

The reduced volume

can be placed

quicker and, combined with its lower density, means that the total load on the floor is reduced.

Concrete thicknesses based upon relative humidity

in the concrete of not greater than 75% are given in Tables 1 & 2.

\ .

Minimum insulation thickness (including non-combustible screeds)

^ V4 hour

1 hour

3 hours

4 hours

Normal w i . concrete

65

90

105

115

135

150

Light wi. concrete

55

65

75

85

115

130

Fire resistance

Table 1.

1 H hours 2 hours

Minimum I n s u l a t i o n t h i c k n e s s o f c o n c r e t e for t r a p e z o i d a l decks ( t h i c k n e s s i n mm)

J7

Π

Minimum insulation thickness (including non-combustible screeds)

SL % hour

1 hour

3 hours

4 hours

Normal wt. concrete

90

90

110

125

ISO

170

Light wt. concrete

90

90

105

115

135

150

Fire resistance

1 % hours 2 hours

Table 2. Minimum insulation thickness of concrete for re-entrant profile decks (depth in mm, equals overall slab depth)

Distribution of Temperature Throughout a Floor Slab Temperature varies depending upon the distance between the point under consideration and the surface exposed to the fire measured normal to the surface of the steel deck (see fig.4).

U Figure 4.

\L

Measurement of depth of profile into concrete

-153-

The temperatures given in Table 3 are based upon data given in reference 1, and are slightly conservative when compared with results from actual fire tests. Rr· resistance (hours) Depth into slab

10 20 30 40 50 Θ0 70 80 90 100

NW 470 340 250 180 140 110 90 80 70 ΘΟ

H LW 460 330 280 200 180 130 80 80 40 40

NW 650 530 420 330 250 200 170 140 120 100

1 LW 620 480 380 290 220 170 130 80 70 80

NW 790 650 540 430 370 310 280 220 180 160

IM LW 720 580 460 360 280 230 170 130 100 80

NW 880 720 810 510 440 370 320 270 240 210

2 LW 770 640 530 430 340 280 220 180 150 140

3 N W LW

• •

700 800 520 460 410 360 320 280

• 740 630 520 430 380 320 270 230 190

NW

• •

770 670 600 540 480 430 380 380

4 LW

· ·

700 800 510 440 380 320 280 270

N W ­ Normal weight concrete LW ­ Lightweight concrete Temperature in deg. C " * " indicates a temperature greater than 800°C

Table 3·

A.7

Temperature distribution through a concrete slab.

Material Strengths The strengths of steel reinforcement and concrete both reduce as

temperature is increased and the relative values can be obtained by multiplying the ambient temperature value by the Factor Kr given in Table 4 below: For design at elevated temperatures the following stresses may be used.

Reinforcement: Design strength, p r

» Jjiir

Concrete : Design strength, p_ L

­

0.67 _ —

f

v

'cu'^r

Where: Reinforcement yield strength f

­

Characteristic concrete cube strength

H^

­

Factor from Table 4

0.67 ­

Effective average stress factor for concrete (see reference 1)

­154­

Light' wt. HY or MS' NormaP Cone. Reinf. w t Cone. Up to 300°C no reduction 1.00 1.00 1.00 1.00 1.00 0.91 1.00 0.91 0.81 1.00 0.82 0.72 1.00 0.73 0.62 0.90 0.64 0.53 0.80 0.43 0.55 0.70 0.34 0.46 0.60 0.37 0.24

Tamp °C

300 350 400 450 500 550 600 650 700

Table 4.

1. M.S. reinforcement to BS4449: 1978 H.Y. reinforcement to BS4449: 1978 Mesh to BS44S3: 1969 2. Concrete to CP110: 1972

Kr strength variation factor

Method of Calculation As Limit State Methods are being used, calculations are carried out to determine the maximum hogging and sagging moments which resist the total moments applied to the composite slab by the factored loads· Firstly, an estimate is made of the likely size of the wire, and the maximum force in the fire reinforcement is determined· Force in reforcement FR

No. wires χ Area χ Κr-.Yr.

This must be balanced by an equal and opposite force in the concrete. 0.67 'f .K A cu r c

0.67 f cu .K r Ac

-

Concrete area in compression

Yr

-

Reinforcement yield strength

fcu

-

Characteristic concrete cube strength

^

-

Factor from Table 4

0.67 -

Effective average stress factor for concrete

Ymr

-

Material factor for reinforcement strength

Ymc

-

Material factor for concrete strength

The effective area of concrete will vary dependant on the case being considered as indicated in Figure 4a and 4b.

-155-

, fc

1* Figure 4a.

Section resisting

4b.

hogging m oneri t

Section resisting sagging moment

When the effective area, and hence d c has been calculated the lever arm h can be determined·

The maximum hogging and sagging moments produced by

the slab can then be calculated.

These may then be used in the formulae

given in 4.4 above. 5.

COMPOSITE FLOORS AND BEAMS For the type of construction shown in Figure 1. it is usual

practice to fire-protect the floor supporting beams by mineral spray where a suspended celling will be used or by dry boards If the beams are to be exposed.

However, the steel beam acting compositely Figure 5a at room

temperature has a different stress pattern to a similar beam acting noncompositely Figure 5b and the performance of the beam at elevated temperature will depend upon which design condition is under consideration.

This is generally recognised by the North American Fire

Test Laboratories where the amount of protection on beams tested compositely is also accepted for the same beams acting non-compositely. However the amount of protection on beams tested non-compositely is not accepted for the composite case.

It is important to understand why this

difference occurs.

i

i

I Figure 5a.

Composite beam

5b.

Non Composite beam

In the design of beams acting compositely at room temperatures the beam, firstly has to carry the dead and construction loads, at which stage

-156-

it acts non-compositely.

After the concrete has cured and attained its

strength the composite beam is formed and is capable of resisting the imposed loads.

The supporting beam is, therefore, initially stressed as a

non-composite beam and then, subsequently, has additional stresses when acting ccmpositely·

The maximum allowable stress for the combined loading

can be as high as 90Z of the allowable yield stress for the material. Conversely, a beam designed to act non-compositely will, at working load, have a stress of about 65Z of the allowable yield stress.

It will be seen

that, as the temperature of the beam increases, the allowable stress in the lower flange of the beam reaches a critical temperature sooner for a composite beam than for a non-composite beam. Consideration also needs to be given to the effective width of the concrete top flange of the composite beam.

For normal design purposes at

room temperature an effective width of approximately one fifth of the span of the beam may be assumed.

In fire the composite floors spanning between

the support beams will deflect and the concrete may crack near the supports. The amount of cracking will depend upon the degree of reinforcement which is used over the supports.

This will effectively

reduce the width of the concrete flange of the composite beam although, as the concrete is unlikely to be operating at very high stress levels, may not have a great effect on the moment of resistance of the composite beam at elevated temperatures. To investigate the stress pattern in composite and non-composite beams for a range of temperature levels, and accounting for the various factors mentioned above, an experimental computer model has been evolved. The program takes account of the design condition, the temperature gradient across the section, the variation in steel strength with temperature and, for composite beams the reducing width of the concrete compression flange, but the possible effect of differential expansion has not been included.

For any temperature distribution a linear variation of

strain is calculated at which the section can resist the applied moment. The program output is presented graphically in Figure 6, where the stress distribution across a composite beam and a non composite beam are shown for the beginning, intermediate and failure stages of a notional fire test.

In both cases the thickness of the fire protection was the same.

The composite beam reached the runaway condition at a bottom flange temperature of S00°C whereas the non composite beam was stable up to 550°C.

The differing stress patterns between the two beam systems can

-157-

Tension

li

Compression

Tension

Compression

Boom Temperature

Intermediate etage Bottom flange 370°C

Imminent failure

Bottom flange 550 C

Non composite beam

Unpropped composite beam

Figure 6.

Stress distributions during fire tests

-158-

clearly be seen and the reduced time rating obtained for the composite beam shows that, for similar fire ratings, a beam acting compositely requires a greater thickness of fire protection than one acting noncompositely. As It is not practical to carry out individual calculations for each composite beam application, it has been concluded that the amount of additional fire protection may be assessed assuming that the lower flange of the steel section does not rise above 500°C compared with 550°C which is used for normal assessment purposes.

For the effect of this on

any given fire protection system the manufacture should be consulted but, in general terms, it will mean an increase in thickness of protection of about 10Z.

In lieu of any information from the protection manufacturers,

the next higher tabulated fire-protected period may be used.

This means

one hour becomes one and half hours; one and a half hours becomes two hours; two hours becomes three hours, and three hours becomes four hours. It must be borne In mind that in a real fire the performance of the composite floor system will be somewhat different to that obtained under furnace conditions, and the suggested increases In protection are, probably, conservative.

However, they will ensure a high degree of

safety, but may well be modified in the light of further investigation and research. 6.

CONCLUSION By the application of established conventional design methods

coupled with well documented data for fire tests reliable calculations can be performed on unprotected composite floors. Composite beams present a more complex problem but the recommended increase in thickness of the fire protection should ensure more than adequate safety. There is a research program currently under way In the U.K. broadly based upon the methods outlined above which when completed should enable an even more economic solution to be given. For the longer term however, research work Is needed to establish the effect of varying degrees of restraint which operate on a floor slab in a real building.

This will enable the amount of fire reinforcement to

be considerably reduced and make far faster construction times. In the meantime the Architect and Engineer can specify composite floors and beams with the knowledge, at least in the U.K. that their performance in Fire can be safely predicted.

-159-

References Design and detailing of concrete structures for fire resistance. Institution of Structural Engineers London, 1978. Design recommendations for composite floors and beams using steel decks. Section 1

Structural

Section 2

Fire R e s i s t a n c e

Constrado London, 1983.

-160-

REPORT CM SESSION I ; DESIGN METHODS Chairman : G.Th. HUPPERMANN Reporter : H. WITTE Research was reported from Sweden and the northern European countries on calculations of steel temperatures, involving the development of mathematical and physical models. For the mathematical models, the finite element method was used, while the physical principles were based on experiments. In Denmark, experiments were being carried out with insulating materials made from mineral wool, calcium silicate panels and gypsum, the resultant proposal featuring temperature-dependent values for the conduction of heat in the insulating material. The average of the ambient fire temperature and the temperature of the steel was recommended as a reference temperature. From Germany it was reported that the effect of the thickness of the insulating material as fire protection was exponential rather than linear and the European Regulations were well on the safe side. The authors of the European Regulations pointed out that they wanted to develop simple formulae for practical use. It was always possible to work out more accurate calculations. The European Regulations did not introduce any restrictions. The same applied to the discussion on European buckling stress curves which had been confined initially to the "C" curve. The a, b and d curves could also be used if appropriate evidence were available* A few questions were raised on the corrective factor æ . This was used to calibrate the calculations, in order to achieve the same level of safety as in the laboratory tests carried ou hitherto. Limiting factors were, for example, variations in yield points and the uneven distribution of temperature over the cross-section of a beam. On product developments : With sheet section composite steel decks i t was pointed out that i t was very important to know precisely the temperature changes in the deck cross-section (vertical) and in the sheet (horizontal). Limiting factors were, for example, the shape of the section, the humidity content and

-161-

accidental air pockets. Full-scale experiments were expensive, which was why several variants had been tested at the same time in an experiment carried out in Great Britain. The results of experiments carried out in Western Europe and the USA had also been evaluated and had proved to be very varied. Other variants were reported from Germany, including the possibility of considering additional reinforcement purely as a means of fire protection reinforcement, and another case where the steel sheet could act as lost casing in a fire.

Sumnary : After the discussion on a few details, Professor Witteveen referred to the aim of the work and of the conference and said that the papers had set out simple and practical methods of calculation. The European Recommendations for the fire safety of steel structures were engendering new concepts: in order to assess fire safety, calculations were being made and used instead of the results of tests, as previously. Calculations were not possible in every case, but where they were possible, they should be used in place of experiments. The authorities were asked to accept these calculations in future. The discussion should not become bogged down in matters of detail and the researchers should not be more precise than was required by the nature of the problem. Dr. Wuppermann summarized the findings of the discussions with an appeal for the European Recommendations to be used as a means of eliminating the national and regional differences in fire protection regulations. It had been shown that buildings could be constructed safely of steel which, as a construction material, had other advantages, such as its flexibility and recycling. Thanks were extended to the Technical Committee 3 to BCCS for its work and to those who had presented papers and joined in the discussions.

-162-

SESSION

II

:

PRACTICAL ASPECTS OF IMPLEMENTING SAFETY

Practical solutions by architects Practical solutions by architects. Practical aspects of implementing safety A consultant's view of steel structures How to reduce the cost of fire safety Industrial buildings - fire losses and appropriate protective measures Report on Session I I

-163-

PRACTICAL SOLUTIONS BY ARCHITECTS Klaus Schuwirth, Dipl.-Ing. Rathnaustr. 12 3000 Hannover 1.0

General

1.1. The conflict between social reality and the built-up environment. 1.2

"TECHNICAL ARCHITECTURE"Desirable objective or picture of fear? (explanation by transparency)

1.3

Establishment of technical architecture by examples (Architects and their w o r k ) . Technical architecture and building materials.

1.4

2.0 Examples from the practice of Schuwirth & Erman, Hanover. 2.1 "HOSEG" office. 2.1.1 The site, the user, his wishes and requirements. 2.1.2. The design and logical development of construction and materials (Steel and fire-safety). 2.1.3 The approval phase (authorities) and materialisation. 2.1.4. First experiences of the user of the building. 2.1.5. The reaction of public opinion. 2.2 Large laundry "BOCO" 2.2.1 Special problems of industrial buildings -economic (low building costs and operating expenses) -simple but attractive image -minimum building time(low cost of finance and uninterrupted progress of building) -maximum flexibility (continuous production changes to accommodate market requirements) 2.2.2 The plan and its materialisation (re building method, fire safety, construction. 2.2.3 The result -Image -turnover -environment -public opinion.

-164-

2.3. NORCONHAUS- Office at Berckhusenst. 150 2.3.1 The specific problem of speculative real estate. 2.3.2 The planning and its basis -Contractor (requirements and financial arrangements) -Architect (ideas and perseverance) -Eingineers (construction and materialisation). -Authorities (.Building and approval regulations) 2.3.3. The special problem of multi-storey steel buildings with visible construction. -fire safety -approval and permission for special procedures -performance and allocation of orders -administration and coordination -discussions between tenant and contractor. 2.3.4. The result - comments. 3.O. Future prospects for various projects. -Hyatt Hanover (steel building and monument) -Organisation for Music and Communication, Hanover (high building and hall) -Exhibition pavilion for Krupp exhibition ground at Hanover (steel building as exponent of movement and dynamics in industrial building.

-165-

Ladies and Gentlemen,

I would like to begin my short

contribution with some words of the Viennese architect Otto Wagner, a contemporary and colleague of Sigmund Freud: "All modern creations must be consistent with-. the new materials and requirements of the present time. They must illustrate our own better democratic, self-conscious technical and economic achievements and endure the practical non-stop strain of man's thoughts- that is obvious". This statement, although it was made 86 years ago, has not lost its topicality. The attitude of mind that it expresses is still as necessary as ever. The future of architecture as we understand it is closely linked with the future of our technical development, and will only retain its present status if it succeeds in using this development to its advantage. EXAMPLE 1. You will be surprised if I tell you that the contracting director of our company is one of the leading conservationists of ancient monuments is Lower Saxony, and is also a lecturer on sacred building of the middle ages at the University of Gottingen. He commissioned our office to design and build a house for him that must clearly show the history of the time of its construction and should have no offensive pseudo-romantic features. As the site was thickly wooded and our instructions were to avoid changing the surroundings, the building was designed in prefabricated sections that could be transported between the trees to the building site, where a mobile, flexible container was erected. The building is about a metre from the surrounding woodland and appears to have sprung up naturally from the wild. It is understandable that the neighbours and local inhabitants, who were provoked during the building stage, became irritated and tagged the building as "UFO", "Container" and "Spaceship" . A design of this kind for a h ouse that is suitable for living purposes and compatible with the desires of those who wish to integrate modern technology into their life—

-166-

style, forces the architect to use compatible .materials and methods.

The building is a steel skeleton with

heat blocking elements, covered on the inside with textile material. In front of these an anodised aluminium facade is suspended, painted a reddish aubergine colour, which forms an attractive contrast to the seasonal changes in the green of the woods and produces fascinating reflections as a result of the light and shade effects of the surroundings. Fire protection of the building is conventional, with reinforced concrete floors and cladding of the interior of the frame by fire protection plates. Because of the double shell structure, the fire regulations were restricted to a 30 minute fire stability. I need hardly stress that the building met with strong opposition from the licensing authorities in the first instance, and also provoked active discussion and opposition from the local residents. This was welcomed by both the contracting director and ourselves, as we believed that provocation of the environment by contemporary architectural innovation was overdue, and only in this way can progress in this field be brought about. The building was the start of our association with structural steelwork, which has continued up to the present time, in which fire safety features have been developed in a series of buildings. EXAMPLE 2. Our firm received an order to design an office building, suitable for transportation to third world countries or the near and far east, which provides an example of modern design and construction for "Mobile Real Estate". The building comprises two full storeys and a loft, and aims to represent the field of activity of the user in

-167-

that it can be placed on an unprepared site in the same way as a large industrial container, raised about 1.5m from the ground with an internal support system on a raster 6X I4jn, and external wind bracing in the form of a network, of steel ropes spanning the building. The silvered plate glass facade covers most pf the building and ensures that it merges into the surroundings. The otherwise plain facade forms an almost frivolous transparent shell for the office function, freely interrupted by stepping and terracing of the facade elements in the region of the stairwell. The external staircase is like a gangway with a filigree special steel space frame. It symbolises that this kind of building concept no longer justifies the requirement to build for perpetuity, but the design can be such that rapidly changing requirements can easily be accommodated. Within the building, fire protection is by flame protection, the ceilings being of a reinforced concrete construction. The forces of the wind braces are transmitted by special steel brackets to the internal ceiling supports, and are the nucleus of development. Building time is only 6 months.

EXAMPLE 3. In 1983 we were commissioned to develop a marketing centre for dressings on behalf of an international organisation. In this case a span of 30 X 18 m was involved, with a requirement for flexibility of the internal storage system. At the same time good daylight was called for and a dustfree atmosphere was required. For delivery, a number of gates was provided for lorries and a platform for nightly deliveries by rail from the main works. As the building was in any case fitted with sprinklers because of the high fire load of the materials being stored, the authorities waived the requirement for protection of the roof covering and roof structure. Only the inner supports were built of steel sections and lined with reinforced concrete. Here the possibilities of a combined

-168-

support were used and the large load span of the roof surface obtained by lattice supports, which were vertical In the glazed light sheds; in front of the building glass canopies were cantilevered over the delivery and service areas. The office and social area was arranged centrally between the ramps, with a full glass frontage. The building clearly shows that with the help of steel buildings there is the possibility of maximum flexibility due to wide spans that can be achieved by contemporary filigree construction, but still provides an imposing structure for large surface industrial buildings without substantially increasing the cost. By prefabricating and assembly, building time can be reduced to less than 6 months.

EXAMPLE 4 Another client is the Electrolux company, which owns a number of large laundries throughout the world, and who required a new building for a new site in Hanover. In industrial building, the socalled level building method is frequently adopted as an ostensibly simple and economic solution, and unfortunately in many cases the structures are built over, with serious consequences in regard to collapse. We have shown that it is quite feasible to produce improved and durable designs without neglecting cost aspects. The building comprises a main shop, 8 m high, 65 m. long with a 45 m span.

The main structure consists of four

lattice supports spanning the width of the building, from which the roof purlins are suspended. In the interior there is a two-storey U-shaped building, which encloses the main laundry area. The ceiling is installed as a sectional structure, which remains safe in all regions, and is protected by flame-resistant paint. For fire protection of the outer fascia, the authorities waived the usual

-169-

exhaustive testing, as the heat checking of the roof was effective and non-combustible, and the distance of the fascia supports from the building guaranteed the flame protection of the inner shell of the facade. The ceiling covering is a linked structure of metal elements. The curtained facade is of aluminium pillar frame construction, divided by blue enamel and plate glass. The under-ceilings are of white painted steel frames double glazed with sparkling safety glass.

The colouring and shape of

the building clearly show that it is a place of work, for the

cleaning of textiles.

Light structures with.

wide spans are. undoubtedly attainable by steel buildings and there is an added advantage

in the possibilities

of short building times and alleviation of transport and removal problems.

The project was started in July

19 83 and completed in December of that year.

EXAMPLE 5. . It has been demonstrated that by contemporaty building construction it is possible to use the advantages of steel structures with very small building materials. The building concerned was for a firm supplying dental products. In the building, a central control system was installed, with automatic feed. For this, a support-free room was required, 30m X 30 m, to accommodate the office and social area. The building structure was of two tubular steel pylons outside the building, from which the roof purlins for the central section were suspended by ropes. Fire protection of the external pylons, which remain untreated, was achieved by the fireproof inner shell of the building facade. Protection of the tension ropes is achieved by the insulating material of the roof covering. The office section is also protected by tubular steel supports, connected to compound supports and, together with the fire protection, meet the static regulations.

-170-

The outside of the building is of self-coloured oxidised aluminium. This type of design is not only artistically exciting, and compatible with our social standing, but also provides onlookers with the opportunity of learning and understanding the static design and efficiency and thus serves our desire to bring the environment together with technical buildings and to accept the often strong suspicion based on ignorance. EXAMPLE 6 This office and administration building shows very futuristic elements in regard to steel construction and fire protection methods, and should make clear where our office sees the short term use and development potential of the material "Steel in building construction". The director of a Hamburg real estate company built a spectacular building for rental to national and international organisations. The unusual structure and design were intended to demonstrate the possibilities of new structures and designs. The structure is conventional, with a reinforced concrete sectional basement and a ground floor steel structure 2 with a surface area of 1 200m and four upper storeys 2 with a surface area of 6 50m . The steel roof supports for the upper floors are about 2 m in front of the glass frontage., and attached to a freely visible framework over the roof, spanning the whole width of the building. The weight is then transferred to four pylon supports. The stability of the building ia guaranteed by a reinforced concrete staircase tower. The ceiling construction is a futuristic mixture in the form of trapezoidal steel sections joined together, serving as enclosures and also reinforcing the area. The support structure is of beams with welded dowels and plates. The combination of materials

-171-

enables a particularly economical celling construction and an exceptionally rapid method of construction. Similarly, the removal of heat from the concrete in case of fire allows an unprotected view of the ceiling surface in a fire-safe structure. A completely new method of building is provided by the fire safety system for exterior unclad steel sections (hanging struts and pylons). The cross section is a round steel tube with an inner tube for water circulation. In case of fire, water from a tank is circulated rapidly for at least 90 minutes. By this means, the costly and unsightly cladding for fire protection can be avoided. Only the visible framework sections, which are created in the form of an artistic space frame, are normally coated with plate elements. This integrated fire protection system was developed for the research group HKW" -Hflnug, Klingsch and Witte, Wiesbaden, and was described by Prof. Klingsch in his paper yesterday. The development of this project by our company clearly demonstrated to us that collaboration with national and international specialists in the construction field should be further intensified in regard to fire safety techniques, to enable a short term exchange of information as only in this way can the further development of structural steelwork be kept abreast of general developments. Whilst approval for individual buildings of this kind is still required, from the main licensing office in Berlin, we hope that in the near future normal licensing procedure will be accepted and that horizontal elements will be protected by an integrated water cooling system. Despite its lavish facade and building quality, the building is only slightly more expensive than a conventional structure, and has the advantage of an extremely short building

-172-

time of only 11 months. This can be explained by the simplification of inspection procedures which permits uninterrupted operation. The model photograph displayed shows some alternative features that may be expected in the near future. The project described here is a competitor's design for an exhibition pavilion for a world company in the field of industrial technology. It demonstrates the state of present day technology and thus completely reveals the range of current techniques, with the objective of offering a hitherto unknown wealth of use and experiencepotential, with a highly artistic design. The building itself comprises two filigree lattice spaceframes with internal water cooling, which lead, at a distance and a height of 35 m, to a vertically rotating steel wheel track, in which three tubular steel and glass structures are located and which run in turn, satellite fashion, in a channel inside the display hall, using a magnetic system. Whilst the curious visitors to the exhibition are transported through the pavilion· on a moving pavement or escalator, the specialist visitor can visit the surrounding body of the hall, uninterrupted by exhibition personnel. All this, combined with the height of rotation, adds up to an exceptionally interesting view of the exhibition, and the design is made clear inside the building by light reflections and light changes. The overall effect is enhanced by supporting the main body of the building on stilts over a shallow tank of water, which is illuminated with changing colours to give numerous special attractions and effects, in addition to he continuous movement of the objects. This kind of project offers the possibility of applying new techniques to specific qualities of the building concerned, including modern steel and industrial technology, and water technology.

-173-

Ladies and Gentlemen, I would like to thank you for your attention, and to conclude with a comment of our eminent colleague Walter Gropius at the 1911 Engineering Congress: "The range of pictures shown does not pretend to be complete, but nevertheless proof is perhaps offered that an industrial prefabricated building need not always be merely a necessary evil, but can be a reflectiion of the best influences of our time".

-174-

PRACTICAL SOLUTIONS BY ARCHITECTS Practical aspects of implementing safety Gabriele Abbado, Architect INSO SpA, E.N.I. Group - Florence, Italy

1. INTRODUCTION The phenomenon of fire is made up of a large number of interconnected variables, each of which can contribute to determining whether a fire can break out, how it develops and what are its consequences. Previous speakers have discussed the topics of standards and regulations, design and research, while economic aspects will be tackled in subsequent papers. The aim of this paper, on the other hand, is to identify the architectural variables which characterize the design, construction and use of buildings incorporating steel structures, viewed from the fire safety standpoint. Its scope does not therefore include assessment of a large number of factors, such as human reactions (e.g. fear and panic) the general "active safety" measures taken with respect to the risks of fire, for example the automatic alarm and firefighting systems and the rapidity with which the emergency services arrive on the scene, and the conditions in which fire breaks out, for example as a result of short circuits, arson, terrorism or vandalism, or the improper use of cooking or heating appliances. Neither are the rules governing the behaviour of users taken into consideration, even though such rules are connected

with the general problems of

the correct design for maintenance purposes. A no ther topic excluded from this presentation is the reaction to fire of the secondary structures and non load-bearing components of the building. 2. ENVIRONMENTAL CONDITIONS If a fire prevention system is to be valid from a technical standpoint, it must not just satisfy general requirements relating to the safety of

-175-

buildings during the design and construction phases and in building management. Such requirements govern merely the correct layout and proper design of access points, lifts, staircases and escape routes, as well as early fire detection and alarm systems and the use of materials with good fireendurance properties, such as strength, reaction to fire and the production of non-toxic fumes. There are, however, other aspects connected with town planning which have to be considered before the specifically architectural parameters and make it possible to set more clearly in perspective which prevention system should be preferred. The density of the built-up areas envisaged in the town plan determines the height and population density of the buildings in relation to the building styles allowed in each zone. In general, the economic forces exploiting the centres of urban areas tend to increase the population and building density in city centres, with the resulting pyramid-shaped progression in building height, in direct proportion to the cost of land. This state of affairs, which has prevailed for the last 40 years, is challenged by the sound theories of Lewis Mumford and Giuseppe De Finetti, which demonstrate that low buildings are more suitable for expensive areas since they use the available space more efficiently. It is sufficient to bear in mind the amount of space occupied in a muLtistorey office or appartment block by the lifts and service shafts and, conversely, the efficient use of space on the middle or low floors of commercial premises. Even if the speed of the lifts i,s increased geometrically, the maximum height of a building, which must not cause the evacuation timetoexceed 5 minutes, cannot economically exceed 300 m, while Wright's design for a one-mile-high skyscraper remains a useless dream. The indirect costs in terms of energy, transport, safety, management and

-176-

maintenance caused by the congestion of the habitat cancel out the initial benefit. Moreover, the greatest saving in space is obtained through horizontal building design, not only for commercial premises but also in the case of service buildings, offices, health facilities and educational premises. In this context, the conditions in which the environmental factors enter into account are of interest as regards the variables of temperature, humidity and wind, in addition to the size, span and location of the compartments and rooms in comparison with the main load-bearing structure. 3. DISCUSSION In order to economically ensure that steel structures display adequate fire safety the simplest method is to avoid having to protect them with any fire-resistant material, by reducing the fire loads to extremely low levels (lower than 20 kg/m ) .

This is achieved in various ways; the commonest method is to divide the building into homogeneous compartments, so as to obtain safe external and internal areas which allow the building to be evacuated only partially in the event of a fire. The fire load of each homogeneous compartment should not be above class 15, while the building management arrangements should include checks to ensure that dangerous substances and materials, which could jeopardize the theoretical calculation, do not accumulate therein. In the example of the hospital, each part of the two-storey building is served by staircases and lifts. The staircases are smoke-proof and each area is in contact with the outside or with inner courtyards. The ward area, which is composed of four compartments, is separated from the health services area by a multi-level gangway, which is built with an independent structure and in which fire can be sealed off by automatic compartmentali zat i on.

-177-

This system is used in Italy in applications of the Oxford Method, for hospital structures designed by working groups: at Casalmaggiore Cremona the head of the group and coordinator is Giulio C. Daolio and at Ostia Lido in Rome, the architects are E. Monaco and A. Martini. 3¿2_0yTER_STRUCTURES The use of steel in buildings, although well-accepted in anti-seismic structures, still gives rise to doubts concerning its reaction to fire. This is due partly to the lack of a sound body of regulations in the industrial field and also the scarcity of both general and specific information. If the materials which are to be relied on for the safety of the buildings must have good fire-resistance properties and well-known thermal properties, it is difficult to see what other material than steel has been characterized with greater precision as regards its behaviour at different temperatures, whoever is carrying out the tests. It is therefore necessary to analyse the factors that could militate against the more widespread use of steel in construction. It is also necessary to describe all the situations in which steel is the only material capable of providing proper solutions to the present-day problems created by current architectural styles (for example canopies and roofing for stations and service stations). Moreover, such architectural styles are also of interest to the fire insurance business, which is endeavouring to encourage building systems that involve less exposure to fire. One of the most successful applications of structures that are external to the building is without doubt the administrative centre of John Deere and Co. at Moline (Illinois) which was designed by the architect Eero Saarinen. This architectural concept gave birth to other styles, such as the one adopted for the Hotel Jolly at Villa Borghese, Rome, built in 1972 on a design by the architects Vincenzo and Edoardo Monaco. The outer structure does not require any fire protection, since it stands clear of the outer wall.

-178-

The variables deriving from the distance between the outer ualL and the outer structures have already been taken into consideration in two 5 9 publications by Arnault, Ehm and Kruppa and Mrs M. Law . The inner parts of the structures are clad with panels of compressed mineral wool which act as flame barriers and heat insulation, by separating the internal environment from the thermal bridge constituted by the IPE 300 girders.

Integration with various other functions of the building components providing fire protection constitutes another economically attractive solution to the problem of fire safety. Fire protection can be integrated with the following functions: heat insulation, sound absorption, soundproofing and corrosion protection of the steel. By way of an example, vertical structures are protected by intumescent paints, which are applied on top of a coat of anti-rust primer. The Liceo Classico Rinaldini at Ancona was damaged by arson in 1978: the fire lasted for 70 minutes, reaching high temperatures that exceeded 600°C and destroyed the aluminium window-frames. The school is a three-storey structure designed by architects P. Castelli and L. Cristini and is highly earthquake-resistant, like all the buildings erected by the Firm INSO at Ancona. The fire was lit on the first floor, in the headmaster's study, which is located at one corner of the building where the load-bearing columns, had they yielded, would have caused the entire building to collapse. The intumescent paint prevented the collapse temperature from being reached. The structures retained their integrity to such an extent that, after cleaning, they were re-coated with the same paint, which has been tested under fire for up to 120 minutes at an officially-approved testing centre by the Milan company Protect. Other fires took place in 1981 and 1983 in steel structures coated with heat-insulating materials by way of fire protection (sprayed mineral wools supplied by the firm Davidson SpA, Milan).

-179-

' '

A short circuit caused a fire at the Ristorante Motta at Bologna Cantagallo, on the A1 motorway. The fire persisted some two hours after the arrival of the fire brigade, damaging the superstructures. The basic structures are still in existence today, and the present Autogrill was rebuilt on them, after they had been re-calculated on the basis of a design drawn up by Mr Martinez of the Milan Polytechnic. Part of the upper structures were demolished in order to allow the architectural appearance to be modified, but the original main HSA and AE girders in the bridge over the motorway were maintained and covered with another similar f ire.protection material. In February 1984, another fire damaged the two uppermost storeys of the Chemistry Faculty of the Milan State University. The steel structures, which were covered with Limpet fibres, withstood a fire that lasted two hours.

3¿4_SPECIAL_STRUCTyRES After the construction of the Pittsburg headquarters of the US Co. in 1969 with water-irrigated structures, other similar structures have been erected in Georgia, Idaho and California. In Europe, the headquarters of the German Association of Steel Constructors in Düsseldorf was built on this principle . The only drawback is cost, since water-irrigated pillars offer virtually unlimited fire resistance. One way of off-setting the higher initial costs of irrigated structures is to integrate such fire protection with a solar energy collection and storage system. Cladding with slabs of extruded polycarbonate causes the greenhouse effect to take place and provides heat insulation for the entire column/beam system.

In order to avoid excessive fluid pressures, the system can be designed for a three-storey building or for multiples of three stories. Each set of storeys has its own expansion and degassing tank.

-180-

The variables deriving from the distance between the outer wall and the outer structures have already been taken into consideration in two 5 9 publications by Arnault, Ehm and Kruppa and Mrs M. Law . The inner parts of the structures are clad with panels of compressed mineral wool which act as flame barriers and heat insulation, by separating the internal environment from the thermal bridge constituted by the IPE 300 girders. 3¿3_CLAD_STR¡JCTyRES Integration with various other functions of the building components providing fire protection constitutes another economically attractive solution to the problem of fire safety. Fire protection can be integrated with the following functions: heat insulation, sound absorption, soundproofing and corrosion protection of the steel. By way of an example, vertical structures are protected by intumescent paints, which are applied on top of a coat of anti-rust primer. The Liceo Classico Rinaldini at Ancona was damaged by arson in 1978: the fire lasted for 70 minutes, reaching high temperatures that exceeded 600°C and destroyed the aluminium window-frames. The school is a three-storey structure designed by architects P. Castelli and L. Cristini and is highly earthquake-resistant, like all the buildings erected by the Firm INS0 at Ancona. The fire was lit on the first floor, in the headmaster's study, which is located at one corner of the building where the load-bearing columns, had they yielded, would have caused the entire building to collapse. The intumescent paint prevented the collapse temperature from being reached. The structures retained their integrity to such an extent that, after cleaning, they were re-coated with the same paint, which has been tested under fire for up to 120 minutes at an officially-approved testing centre by the Milan company Protect. Other fires took place in 1981 and 1983 in steel structures coated with heat-insulating materials by way of fire protection (sprayed mineral wools supplied by the firm Davidson SpA, Milan).

-179-

' '

A short circuit caused a fire at the Ristorante Motta at Bologna CantagalLo, on the A1 motorway. The fire persisted some two hours after the arrival of the fire brigade, damaging the superstructures. The basic structures are still in existence today, and the present Autogrill was rebuilt on them, after they had been re-calculated on the basis of a design drawn up by Mr Martinez of the Milan Polytechnic. Part of the upper structures were demolished in order to allow the architectural appearance to be modified, but the original main HSA and AE girders in the bridge over the motorway were maintained and covered with another similar fire.protection material. In February 1984, another fire damaged the two uppermost storeys of the Chemistry Faculty of the Milan State University. The steel structures, which were covered with Limpet fibres, withstood a fire that lasted two hours. 3¿*_SPECIAL_STRyCTURES After the construction of the Pittsburg headquarters of the US Co. in 1969 with water-irrigated structures, other similar structures have been erected in Georgia, Idaho and California. In Europe, the headquarters of the German Association of Steel Constructors in Düsseldorf was built on this principle . The only drawback is cost, since water-irrigated pillars offer virtually unlimited fire resistance. One way of off-setting the higher initial costs of irrigated structures is to integrate such fire protection with a solar energy collection and storage system. Cladding with slabs of extruded polycarbonate causes the greenhouse effect to take place and provides heat insulation for the entire column/beam system.

In order to avoid excessive fluid pressures, the system can be designed for a three-storey building or for multiples of three stories. Each set of storeys has its own expansion and degassing tank.

-180-

The problem to be solved is the chemical composition of the anti-freeze mixture which must not corrode the steel.

In the case of irrigated steel structures integrated with a solar energy collection and storage system, an economically viable cost/benefit ratio can be achieved, and this modern architectural style could take a share of the market. It. CONCLUSION It has been demonstrated for more than 20 years

that in steel structures

with a fire load of less than 20 kg/m , the steel does not have to be coated or protected and that even a fire load of 25 kg/m

is not hazardous.

In the case of structures belonging to higher fire classes, a wide range of f ire-protection systems can be used which have demonstrated optimum fire-endurance properties in actual fires. The European market, on which this conference is focussing, should attempt to use steel to a greater extent in the construction industry, in order to help overcome the crisis in the European steel industry. The official authorities of the EEC Member States should bring themselves into line with Japan and the United States as regards regulations, and in particular those relating to anti-seismic structures. As far

as the construction of buildings in earthquake areas is concerned,

a more widespread use of steel in all types of architecture should be encouraged, as in Japan. European regulations are paving the way for the harmonization of national 18-28 requirements relating to the application of ISO standards in national laws and in Community provisions into national regulations.

-181-

which have been incorporated

BIBLIOGRAPHY

1. Curt F. Kollbrunner: "L'attuale e futura costruzione ad ossatura metallica". Alta Autorità della Comunità Europea del Carbone e dell' Acciaio, Congresso Acciaio 1964: "I progressi nelle costruzioni in acciaio", Lussemburgo, 28-30 Ottobre 1964. 2. George E. Danforth: "Costruzioni in Acciaio", Alta Autorità della Comunità Europea del Carbone e dell'Acciaio, Congresso Acciaio 1964: "I progressi nelle costruzioni in acciaio", Lussemburgo, 28-30 Ottobre 1964 3. L.F. Donato - L. Sampaolesi: "Gli acciai e la sicurezza delle costruzioni" Ed. ITALSIDER, Genova 1971. 4. Consiglio Nazionale delle Ricerche: "Principi per una normativa tecnica sulla sicurezza contro il fuoco dei fabbricati con struttura di acciaio", Bollettino Ufficiale CNR, Norme Tecniche, Anno 7°, n" 87, Roma 1973. 5. P. Arnault, H. Ehm, J. Kruppa: "Evoluzione delle temperature nelle colonne esterne sottoposte a incendi", Documentation CECH 3-7/4/7F. Centre Technique de la Construction Métallique, Puteaux, 1974. 6.. S. Cuomo: "Elementi di resistenza al fuoco delle strutture d'acciaio e loro protezione", Liguori editore, Napoli 1975. 7. Mommertz, Pethier, Weineck: "Protezione antincendio delle colonne di acciaio con raffreddamento ad acqua", comunicazione n° 13 dell'Istituto di Ricerche della Associazione Tedesca del Costruttori in Acciaio, Düsseldorf. 8. I. Tiezzi: "La sicurezza dell'incendio nell'edilizia, alti fabbricati", EPC Edizioni di Protezione Civile, Roma 1976. 9. M. Law: "Design Guide for Fire Safety of Bare Exterior Structural Steel, 1° Theory and Validation, 2° State of Art", OveArupi Partners, London 1977. 10. A. Cascarino: "La funzione della rivelazione e segnalazione degli incendi nella prevenzione incendi" Antincendio

-182-

e Protezione Civile n° 4 Aprile 1977.

11. I. Tiezzi: "I costi della protezione antincendio" EPC Edizioni di Protezione Civile, Roma 1978.

12. F. Hart, W. Henn, H. Sonntag: "A rchitettura A cciaio, Edifici Civili", Deutscher Stahlbau­Verband, Italsider, Genova 1979.

13. G. Nava, Ν. Dalumi, G. Scotti: "Comportamento al fuoco e compatibilità di smalti di varia natura, quali finiture per cicli di vernici intumescenti contro il fuoco", rivista A cciaio/CISIA , n° 5, 1981.

14. G. Nava, Ν. Dalumi: "Comportamento al fuoco delle antiruggini quali fondi per cicli di vernici intumescenti contro il fuoco", Rivista Acciaio/CISIA, n° 1, 1981. 15. N. Dalumi, P. Setti: "Comportamento al fuoco delle strutture protette con materiali intumescenti", conferenza "Giornate Italiane della Costruzione in Acciaio del C.T.Α.", Perugia 1983.

16. I. Tiezzi: "La nuova legge di prevenzione incendio in Italia", Conferenza tenuta al Congresso della A ssociazione Europea degli Ufficiali Professio­ nisti del Vigili del Fuoco (European A ssociation of Professional Fire Brigade Officers), Milano 1983.

17. S. Cuomo, G. De Martino: "La sicurezza contro l'incendio degli edifici a strutture in acciaio", Monografia n° 6 della Ricerca "Il comportamento delle strutture portanti in acciaio". Ricerca Nuova Italsider ­ Comunità Europea, Genova 1983.

18. Circolare n° 12 del 17/5/80 del Ministero dell'Interno: "Reazione al fuoco dei materiali impiegati nell'edilizia, Specifiche e modalità di prove e classificazione".

19. Circolare n° 91 del H/9/61 del Ministero dell'Interno: "Norme di sicurezza per la protezione contro il fuoco dei fabbricati a struttura in acciaio destinati ad uso civile".

20. Decreto Presidenziale 29/7/83 n° 577.

21. Decreto Ministeriale 30/11/83 del Ministero dell'Interno.

­183­

22. A ustria: "Norma ONORM B 3800".

23. Francia: "Decreti e circolari raccolti in ­ Sécurité contre les incendies brochure η. 1011 du Journal Officiel de la République Française".

24. Germania: "Deutsche Normen DIN 4102 ­ Caratteristiche di comportamento al fuoco di materiali e componenti per l'edilizia". 25. Gran Bretagne: "Norm British Standard BS 476 ­ Fire Tests on building materials and structures". 26. Scandinavia: "Norme Nordiska Kommittén for Byqqmertãnnuelser ­ Metodi NKB Nordtest fire numerati da NTF 101 a NTF 109".

27. Stati Uniti d'A merica: "Regolamentazione federale norme A NSI/A STM D­E".

28. Svizzera: "Regolamentazione federale standard SNV 520183/2 ­ Uso dei materiali combustibili nell'edilizia. Metodi di prova".

r 1B4­

ILLUSTRATIONS

Fig. 1.

Casalmaggiore-Cremona Hospital G. Daolio, Engineer

Fig. 2.

Ostia-Lido Hospital, Romp E. Monaco and A. Martini, Architects

-185-

FIG. 3.

FIG. 4.

Jolly's Hotel,Rome E. and V. Monaco, Architects

Liceo Classico Rinaldini, Ancona P. Castelli and L. Cristini, Architects

-186-

FIG. 5.

FIG. 6.

Liceo Classico Rinaldini, Ancona P. Castelli and L. Cristini, Architects

Liceo Classico Rinaldini, Ancona P. Castelli and L. Cristini, Architects

-187-

FIG. 7. Liceo Classico Rinaldini, Ancona, P. Castelli and L. Cristini, Architects

FIG. 8. Liceo Classico Rinaldini, Ancona, P. Castelli and L. Cristini, Architects

-188-

■**

FIG. 9.

­

«*i*.

Liceo Classico Rinaldini, Ancona, P. Castelli and L. Cristini, Architects

FIG. 10. Autogrill Motta, Bologna Cantagallo, Mr Martinez, Engineer

­189­

FIG. 11. Autogrill Motta, Bologna Cantagallo, Mr Martinez, Engineer

FIG. 12. Autogrill Motta, Bologna Cantagallo, Mr Martinez, Engineer

-190-

FIG. 13 Autogrill Motta, Bologna Cantagallo Mr Martinez, Engineer

FIG. IA Autogrill Motta, Bologna Cantagallo Mr Martinez, Engineer

-191-

FIG. 15. Autogrill Motta, Bologna Cantagallo, Mr Martinez, Engineer 11

FIG. 16. Laboratories at the Chemistry Faculty of the Milan State University

-192-

FIG. 17. Laboratories at the Chemistry Faculty of the Milan State University

-193-

A CONSULTANT'S VIEW OF STEEL STRUCTURES

M. LAW Ove Arup Partnership, London

Summary Calculation methods and design manuals give scope for better use of traditional methods of fire protection, the development of new methods and the definition of fire exposure and structural behaviour in fire. They are ideally suited for international application. Particular topics of interest include calculation of fire resistance for steel with cladding and for composite steel and concrete decks and columns. Water-filled hollow sections, identification of the circumstances where unprotected steel may be used and the calculation of external fire exposure can all lead to economies and may allow the steel structure to be expressed. Case studies of recent steel framed buildings illustrate the points discussed.

-194-

1. INTRODUCTION The conventional methods of satisfying the requirements of building regulations for the fire protection of structural steelwork are straightforward and well accepted.

New protection methods and new ways of using

calculations for the design of traditional methods can also be straightforward; acceptance is being gained more readily than in the past and such acceptance is eased by the provision of design codes and manuals.

This

paper discusses these recent developments in relation to a number of steel-framed framed buildings. 2. Cladding of columns and beams A most welcome development in this field is the acceptance of calculation methods for the determination of the cladding thickness needed to limit the temperature rise of the structural steel to a specified 'critical temperature'

. It is thus possible to determine the protec-

tion needed for a wide range of section sizes without being required to suffer the expense and delay of many standard fire resistance tests. As a recent example, this approach has been adopted for the Hong Kong and Shanghai Bank. The critical temperature itself can also be calculated

;

it is not necessarily the often-quoted 550°C, since its value depends on the stress, mode of failure and type of steel.

Calculation of the

critical temperature is essential for building elements which, because of furnace limitations, cannot be subjected to the standard test; examples are the internal trusses of Centre Pompidou, Paris, and the hangers of the Central Bank, Dublin.

Such calculation approaches should help to

break down the barriers between countries, because they can take into account national variations in, for example, load levels and material properties.

3. Profiled steel and concrete decks Ribbed steel decks acting compositely with concrete are a popular form of construction but unless the steel is protected from fire then, for periods of fire resistance exceeding 30 minutes, it must be assumed that it has no structural function.

The favoured solution is to incor-

porate steel reinforcement, for fire purposes only, in the concrete. Most test data have been on proprietary systems and the results have not always been easy to generalise.

It has however been possible to use

-195-

guidance issued for concrete structures and fire the Water Research Centre, Swindon.

(2)

, as for example, in

Lightweight aggregate concrete can

offer advantages and both lightweight and normal weight concrete are dealt with in a guide which has just been published

. This guide gives

no test results but an opportunity for comparison will become available when the ECCS manual is published. 4. Composite steel and concrete columns In the UK, for fire resistance periods of up to 2 hours (which covers most buildings) it is usual to have an all-steel or all-concrete structural solution, although there has been some interest in concretefilled hollow steel sections, particularly if this leads to a reduction in column diameter.

As with the steel/concrete decks, for fire resis-

tance periods exceeding 30 minutes it is assumed the steel has no structural function.

Where appropriate,

the concrete can incorporate

fire reinforcement, as for example in the Water Research Centre. 5. Water-filled hollow sections An attraction of water-filling, as a method of keeping the steel cool during a fire, is that the steel itself is exposed to view.

One

early example is the US Steel Headquarters in Pittsburgh, where the structure of weathering steel is displayed externally.

Other examples

of external water-filled steel are the Centre Pompidou and Bush Lane House, London (in stainless steel to avoid maintenance).

In both these

examples the external structure maximised the floor area available.

With

adequate water storage a hollow section can be of smaller diameter than a concrete column of the same fire resistance and, once again, floor area is maximised. However, the anticorrosion and antifreeze additives can be expensive for low-rise buildings and this is one reason for the relatively few buildings which use water-cooled steel. A design manual related (4) to the standard fire existence test is available but for external columns it is preferable to calculate the fire exposure according to their location and the fire load of the building as discussed below. 6. Unprotected steel elements Where it can be demonstrated that public safety is not at risk, structural fire protection need not be required.

-196-

An example is the Royal

Exchange Theatre, Manchester.

In most countries, the structure of a

single storey building needs no fire protection, provided there is negligible risk of fire spread to adjacent buildings, and likewise the roof of any building may be unprotected.

An example is the Garden

Festival Building, Liverpool. When the fire load is low it may be possible to show that if a fire occurred a steel structure would not reach its critical temperature. This is already accepted for car parks, where there is a low fire load and a low probability of fire spread from one car to another.

Other low

fire load, low fire risk buildings which might also be in unprotected steel are transport terminals and sports stadiums; these usually contain large circulation areas and little in the way of fire load. is at Ibrox Park, Glasgow.

An example

It would be very useful to have general

guidance on structural requirements for these types of building. It is also accepted that structure external to the building is likely to reach a lower temperature than the internal structure; based on a technical study

design manuals have been published

which

show how to calculate external fire exposure and structural temperatures. This method was useful in the design of the Central Bank, for example. 7. Case studies The buildings mentioned below have been chosen to illustrate some of the aspects discussed above. (9) Bush Lane House, London, UK Architects and engineers:

Arup Associates

This building provides 8 office floors, above a plant room at first floor level, the ground floor area being left free for use by London Transport.

In order to provide maximum floor area within planning and

other constraints, the floor loads are transmitted by an external steel lattice to the lift core and three columns.

To avoid corrosion and

maintenance problems the lattice is exposed stainless steel and is waterfilled for fire protection. The patterns of water flow, maximum potential steel temperature and the amount of water storage were all established by calculation, as a standard fire resistance test could not be carried out. Central Bank Offices, Dublin, Eire Architects: Stephenson Gibney and Associates Engineers:

Ove Arup S Partners, Dublin

-197-

This eight-storey block of offices has floors supported on twin reinforced concrete cores and 12 external hanger points round the perimeter. The hangers are of Macalloy steel bars and for architectural reasons it was essential to avoid very thick cladding.

This was achieved by calcul-

ating the external fire exposure and demonstrating that the critical temperature would not be attained with the thin layer of cladding provided.

The critical temperature itself was established by an analysis

of the mode of failure and taking into account the steel characteristics, since a standard fire resistance test was not possible. Centre Pompidou, Paris, France Architects: Engineers:

Piano S Rogers Ove Arup s Partners

This six-storey building, which is an arts centre, has a steel structure rising above a concrete sub-structure.

The main lattice

girders, most of which are internal, have cladding to give 2h fire resistance, but as it was not practical to subject them to a standard fire resistance test the critical steel' temperature was established by calculation.

Much of the rest of the structure is exposed externally.

Calculations showed that some elements, by virtue of their distance from the windows (7.6m) did not need fire protection.

Others are protected by

fire resistant shields on sprinklers, and the main columns, at 1.6m distance, are water filled. Digital Equipment Company, Reading, UK Architects and engineers:

(12)

Arup Associates

This steel-framed building combines a single storey manufacturing area and a two-storey office area within a steel-clad envelope.

Under

Building Regulations, only the wall separating these two uses and the floor of the upper storey offices are required to have fire resistance, the rest of the structure being unprotected steelwork.

The structure has

a number of 7.2m square bays which form a series of roof pyramids supported by a central column. The upper level of the offices is open to the floor below and overlooks the central 'street'.

The external walls

and the roof are of galvanised steel insulated sandwich panels. Engineering Research Centre, Water Research Centre, „ . . (13) (14) Swindon, UK Architects:

Architects Design Partnership

Engineers:

Roughton and Fenton

-198-

This two-storey building accommodates research laboratories, offices, an experimental test hall and various communal facilities.

The

client required a maximum floor area within the budget and the ability to adapt and rearrange the working areas. To this end a component cladding panel system hung on a steel frame was used and most of the columns were placed outside the building envelope.

The standard of fire resistance is

1 hour. The columns are 457mm diameter circular hollow sections, concrete filled.

The original intention was to use water filling for the fire

protection but the cost of the additives was considered to be too great. The columns carrying axial loads only are filled with mass concrete while the columns required to carry substantial bending moments are filled with reinforced concrete designed as a column according to British Standard (2) . The

CP110 with a partial safety factor of 1.05 for accidental damage

external columns have the same standard of fire resistance as the internal ones and do not rely on any protection from the external wall panels. The first floor slab is a concrete deck with bare ribbed steel acting as permanent formwork and fire reinforcement being contained in the concrete.

The steel beams are protected with mineral spray.

Gateway Two, UK Architect and engineers: Arup Associates Gateway Two is an office building with 5 storeys of offices built round a central atrium.

The galleries in the atrium have a profiled

steel deck acting as permanent formwork for the concrete slab so that they could be built without scaffolding. carries the galleries, lifts and roof.

An exposed steel structure Pneumatically operated roof vents

are controlled to respond to wind and temperature conditions during normal use and are also linked automatically to smoke detectors so that they act as smoke vents in the event of fire.

These features of the

atrium were accepted by the authorities on the understanding that the atrium would be used for circulation only during normal office hours; some social uses are also permitted in the evenings and weekends when the offices are empty. The Govan Stand, Ibrox Park, Glasgow, UK Architect: Τ M Miller and Partners Engineers:

Thorburn and Partners

-199-

This stand, at Glasgow Rangers Football Club, is a two-tier steel and concrete construction.

Columns within the structure were protected

with lightweight blocks but horizontal members were left unprotected. The latter were estimated to have 40 min fire resistance, on the basis of results of fire resistance tests, the values of the section factors for the steel beams and an assessment of the actual structural behaviour during a fire.

By measuring fire loads and ventilation an 'equivalent (18) was calculated to be 32 min. The saving in

time of fire duration'

costs of conventional fire protection was estimated to be over £40,000. Liverpool Festival Building, Liverpool, UK Architects and engineers:

Arup Associates

This building is designed for the International Garden Festival 1984 and is essentially a steel-framed vault 140m long by 60m wide with polycarbonate cladding.

At each end of the vault there is a 30m diameter

half-dome, aluminium clad, with pressed steel baked enamel internal lining panels.

Being a single storey building, the structure does not

need fire resistance.

The steelwork can therefore be left exposed.

The

building will be converted after the Festival into a Leisure Centre, the concrete structure of which will be contained within the glazed steel vault. The Royal Exchange Theatre, Manchester, UK Architects:

Levitt Bernstein Associates

Engineers:

Ove Arup S Partners

(19)

This open-stage auditorium, stands within the Great Hall of the Manchester Royal Exchange.

There is a stage and seating for 450 at the

level of the Exchange floor and seating for a further 300 people is provided by two galleries suspended from tubular steel trusses. A fire engineering appraisal of smoke generation and crowd movements established that people could escape readily and therefore the theatre remains an open structure without fire cladding. References 1.

ECCS.

"European Recommendations for the Fire Safety of Steel

Structures.

Calculation of the Fire Resistance of Load Bearing

Elements and Structural Assemblies Exposed to the Standard Fire". European Convention for Constructional Steelwork. Amsterdam, 1983.

-200-

Elsevier,

2.

Design and detailing of concrete structures for fire resistance. Interim guidance by a Joint Committee of the Institution of Structural Engineers and The Concrete Society.

The Institution of

Structural Engineers, London, 1978. 3.

Newman, G.M. and Walker, H.B.

Steel framed multi-storey buildings.

Design recommendations for composite floors and beams using steel decks. 4.

Section 2.

Bond, G.V.L.

Fire resistance.

Constrado, Croydon, 1983.

"Water cooled hollow columns".

Constrado, Croydon,

1975. 5.

Ove Arup s Partners. structural steel.

"Design guide for fire safety of bare exterior

Technical reports". American Iron and Steel .

Institute/Constrado, January 1977. 6.

Law, Margaret, "Fire safety of external building elements - the design approach".

Engineering Journal, American Institute of Steel

Construction, Second Quarter, 1978, pp 59-74. 7.

American Iron and Steel Institute.

"Firesafe structural steel. A

8.

Law, Margaret and O'Brien, Turlogh. "Fire safety of bare external

design guide". Washington D.C., 1979. structural steel". Constrado, Croydon, 1981. 9.

Eatherley, M.J. "The design and construction of Bush Lane House". The Structural Engineer, February 1977, No. 2, Volume 55, pp 75-85.

10.

McSweeney, M.F. "New HQ for Central Bank".

Irish Engineers, Vol. 31,

No. 2, February 1978, pp 3,5,7-8. 11.

Ahm, P.B. et al.

"Design and Construction of the Centre National

d'Art et de Culture Georges Pompidou".

Proc Instn Civ Engrs. Part 1,

1976, 66, Nov. 557-593. 12.

The Architects Journal, 4 May 1983, pp 51-66.

13.

The Architects Journal, 24 February 1982, pp 47-65.

14.

Fenton, R.F. Personal communication.

15.

The Architects Journal, 3 August 1983, pp 26-34.

16.

Smith, Ian. Fire engineering and the design of sports stadia. Building with Steel, Vol. 9, No. 2, December 1982, pp 10-13.

17.

Framed in Steel 6.

the Ibrox Stadium Redevelopment.

British Steel

Corporation, November 1981. 18.

ECCS.

Fire safety in constructional steelwork.

CECM III-74-2E,

1974, Chapter II. 19.

Morreau, P. and Baldock, N.

"Royal Exchange Theatre, Manchester".

The Structural Engineer, July 1978, No. 7, Vol. 56A, pp 189-197.

-201-

CENTRAL BANK OFFICES

BUSH LANE HOUSE

CENTRE POMPIDOU

DIGITAL EQUIPMENT COMPANY

(SIDE VIEW)

-202-

GATEWAY TWO

LIVERPOOL FESTIVAL BUILDING (MODEL)

ROYAL EXCHANGE THEATRE

-203-

HOW TO REDUCE THE COST OF FIRE SAFETY

L. FRUITET Inaénieur-Conseil de l'Office Technique pour l ' U t i l i s a t i o n de l'Acier

Summary

The protection of structural frameworks against fire too often escalates their costs. These costs can be markedly reduced by a careful analysis of the risks, and the structural behaviour, by designs more suitable for security and, when judged necessary by different means of protection. This economy can be carried as far as the elimination of all special protection against fire, for all or part of the framework of a building. One must also consider the costs of maintenance, upkeep and the conditions for repairs and alterations.

-204-

1. TOTAL OR PARTIAL ELIMINATION OF PROTECTION It is too commonly believed that all steel structural elements incur danger in case of fire, if they are not specially protected. There is no foundation for this belief, in numerous cases i.he absence of protection does not constitute a danger to the occupants. Either collapse of the structure is not feared, or any collapse cannot reach people in the fire area itself or in other parts of the burning building. 1.1. When no stability to fire requirement is stated Regulations concerning the security of people exclude for some buildings, even when the public are admitted, the necessity for fire stability of the framework. This is generally the case for single storey buildings and for certain others with a few floors, for which evacuation can by very rapid and easy access for help. This disposition avoids costly protection, but is matched by conditions elsewhere which reinforce the security arrangements, requiring for example the incombustability of the framework material (which is given without expense by steel structures) or the existance of automatic fire detection. Another favourable element particularly interests us : the visability of the roof trusses from the ground (for single-storey buildings) or the top floor. This is in effect a very favourable arrangement for the rapid and efficient intervention of safety personnel and also gives large economies. However, it is interesting to question the reasons for these arrangements, since they lead in certain cases to the possibility of structural collapse. It is necessary in effect to consider that the high temperatures reached by building materials (higher than 500°C for steel) at the moment of loss of strength bringing collapse, which compare to even higher temperatures in the areas containing them, are completely incompatible with the possibility of there being living persons present. It is not to these people that the falling structure can be a danger but the consequences of its collapse to other parts of the building may be occupied or contain safety personnel.

-205-

This observation leads us to an analysis of the risks, which consists of imagining the existance of a fire in a given area of the building and of calculating the forseeable consequences of this fire, in particular where it concerns the eventual collapse of the structure. This procedure leads us to the distinction between principal and secondary frameworks. 1.2. Secondary framework We designate as "secondary" (from the point of view of stability to fire) the elements of the frame, who's destruction will not entail that of other elements of the structure, situated outside the zone of the fire. Figure 1 gives a schematic example of the distinction between principal and secondary elements of a framework. This sketch shows a cross section of a building, composed of two parts : - multi-storey section where the portal frame construction gives transverse stability, and comprising discontinuous mezzanine floors. - a single storey lean to building. The building is divided into separate compartments by fire resistant walls. One can imagine therefore that a fire can develope in any one of these compartments without reaching the other compartments under a reasonable time lapse. The elements of the framework of the burning compartment will be considered as secondary as they will not affect the stability of the other compartments. It is again, for example for the framework of the mezzanine floors whos collapse would not endanger the general stability. So far as the lean-to section is concerned we will consider 2 cases : 1st) In the first case, the framework (posts and beams) of this leanto do not form part of the stability of the whole structure, (for example, the fixings of the posts to the beams are articulated). These elements are therefore secondary.

-206-

2nd) In the second case, the fixings of the beams of the lean-to into the exterior posts are 'cast-in' (rigid joints), forming semi-portals which form part of the stability of the entire building. They are then "principal elements" of the frame. The figure indicates the principal elements in full lines and the secondary by fine lines, in both parts of the sketch. The analysis can next be refinded by calculation, in the second case above, for example one can check if, according to the basis of calculation assumed in case of fire, the forces applied to the building make the framework of the lean-to necessary for the general stability. Should the framework of the multi-storey building be sufficient to ensure stability, the elements of the lean-to can be qualified as "secondary" as in the first case of figure 1. A general method of analysis discriminating between principal and secondary elements can be as follows : 1st) Section the building into compartments (separated by fire resistant barriers) or into 'zones' (for buildings with large surface areas) in each of which a fire can be contained. 2nd) From the elements of each compartment or zone select those which in collapsing, can bring down the elements of other compartments or zones of the building, or the fire barriers. - either by rule of thumb (example, the columns of a multy-storey building, the bracing between several zones); - or by calculation as detailed in the above example. At the end of such an analysis it can be decided to finally protect only the principal elements. These elements are generally the most massive, which makes protection easier and cheaper, the protected surface area being less, relative to the weight of steel.

-207-

For example, at the School of Architecture, Nantes, France (Fig.2) the cost of protection by paint initially envisaged for the whole of the steel structure, would have cost 15% of the price of the structure itself, the cost was reduced to 8% by analysis of the principal elements, only these being protected (5). In the case of buildings of large surface area, it is particularly interesting, having divided the surface into zones (even not isolated by fire barriers), to ensure independent stability of these zones, in such a way that a fire localised in one zone does not risk bringing collapse to unaffected zones, like a pack of cards (domino effect). The so called "Autostable" systems, comprising modules independent one from the other, are much more certain and justify the absence of any special protection. 1.3. Principal elements of the structure, where the stability to fire can be judged sufficient without special protection. It is known that unprotected steel structural elements only rarely give a 30 minutes stability to fire. One can conclude that it is necessary to protect all principal elements whenever there is a requirement of more than 15 minutes stability to fire. It is appropriate however to examine the true significance of these degrees of stability to fire. Let us recall first that it is completely erroneous to compare degrees of stability to fire expressed in time (hours or minutes), with a real evacuation time or access for help. The times determined by tests on structural elements in a laboratory furnace can be very much lower or very much higher than the exposure time of the same element in a building affected by fire. This is proved by laboratory tests of normal fires, even though these are simplified in comparison to conditions in a real fire. The degree of stability to fire only constitutes a comparison scale, between different solutions, as to the resistance to high temperatures.

-208-

One of course chooses, between two acceptable solutions, the one offering the greatest resistance, i.e. the one with the highest degree of stability. But is it not right to reduce the protection to the point where they become useless, that is to say when destruction is not feared. As is will shown by natural fire tests, in which are accurately measured the incendiary charge (calorific potential per square metre, or cubic metre) the ventilation coefficient taking into account the degree of permeability of the walls to heat, one always achieves a maximum temperature, as much in the furnace atmosphere as on.the structural element itself. Whilst ever the maximum temperature of the element is lower than the critical temperature 0 M 0_ , destruction cannot be achieved, regardless of the degree of stability. Structural steel e"'ements sited in conditions identical to or better than a given locality present no risk of failure and all suplimentary protection is useless. This observation is particularly useful for areas of low calorific potential (swimming pools, gymnasiums, school class rooms etc.) One has shown (2) that large volumes offer very mach more favourable conditions than do small volumes simulated in laboratory furnaces, all other conditions being identical. In these large volumes there is rarely any risk of attaining the critical temperature of steel. There are two recent examples of these analysis methods which have avoided all surface protection from the steel framework, in^buildings intended for the use of large numbers of the public. The first example is that of a multi-sports hall at Meriadec, near to Bordeaux (France) (Fig. 3) (3). To meet the regulations all the four 60 M span portal beams must give al 1/2 hour fire stability and the elements of the roof structure 1/2 hour rating. The estimation of the calorific potential within the volume of the hall 2 gave the following values, in Kg of wood per M .

-209-

- Minimun potential (used a skating rink) = 8.5 - Maximum potential (used a multi-sports) = 12.4 Comparison of critical temperatures (calculated for the 49 most popular elements) with the maximum temperatures reached in a natural fire test with 15 kgs of wood per square metre, showed that destruction could not be achieved for any of the unprotected structural elements. From this it is concluded that the steel structure of a building can ensure their working throughout the life of a fire without special protection. The saving in cost realised was 800,000 FF. Nevertheless, it was necessary to increase the size of several beams to attain the required critical temperatures, which lead to an additional steel weight of 1,800 kg (on the total of 528,200 Kg for the building). The second example concerns a sports hall of Bercy, Paris (Fig.4). The tri-dimensional roof structure of crossed beams was left without surface fire protection, whereas strict application of the regulations requires a protection corresponding to 1 1/2 hours (90 Mins.) (4) fire resistence. It will be noticed that in these two examples, the margin of safety obtained is still very large, because of the large volumes and the wide distances between a possible fire and the principal structures. Exterior structures can also be left without obvious special protection, or have only partial protection on the faces exposed to the facade. Figure 5 shows the solution adopted for the-exterior structure of a UNESCO building in Paris, very h]gh office building, for which the general regulations require a stability to fire of 2 hours. A series of systematic tests, carried out since by CITICM (France) for the CECA, will without doubt, allow still further reductions in those partial protections. They show, in a general way, that the critical temperatures of unprotected steels cannot be attained whilst ever the columns are not sited directly in front of openings and close to the facade. Finally, the steelwork of a structure can remain exposed in many composite constructions :

-210-

- Columns with hollow profiles filled with concrete, with or without bracing; - Composite columns steel/concrete; - Composite and flagged floors. Most often the stability to fire so obtained does not require any r special protection, to the extent that steel/concrete is used for every day protection, without the occurance of a fire. It is seen from the examples above that in numerous cases the steel framework, even the principal elements, can be used without special protection, or with reduced protection without risk. Which must give considerable reduction in costs, at the same time ensuring maximum protection to the parts of the construction where it is justified. We must now examine the most economical solutions for the elements which must be protected. 2. ECONOMIC PROTECTION The application of surface insolators to structural elements enables the degrees of stability desired to be obtained in the most economical manner; the least costly is the spraying of mineral fibres, of vermiculite cement or of plaster (see table fig. 6 ) . Protection by sheets or shells of similar materials fixed mechanically or by adhesion facilitates surface alterations more easily. It must not be forgotten that ease of alteration is an important advantage for metallic constructions and fire protection must conserve this asset. Suspended ceilings or partitions, used for insolating can constitute sufficient barrier to protect the steel elements against fire, in return for some cheap precautions.(Flame resistant walls, or even simple incombustable screens can suffice). It is always advisable to avoit the chimney effect and horizontal connections between compartments, produced by double walls run for some

-211-

distance, which have caused in several circumstances particularly serious accidents, by propogating asphyxiating fumes and fire itself across entire buildings (Pailleron college, Paris; GEAI dwellings, Rouen, France). Frequent divisions by simple incombustable walls should be sufficient to avoid these propagations and make buildings, much, safer. The calculation methods for heating of steel in accordence with a given law of temperature variation (see 2.1 and 2.2.) enables the thickness of protection to be modified, and therefore its costs, by adapting it as closely as possible to the particular conditions of the construction elements and of their use in each building, without resorting to costly tests. Supplying water to hollow steel sections is an interestina solution which can be economical if the architectural designs incorporate this idea from the outset. The architecture of a building must not be governed by security to fire, but it must integrate this problem from original conception, in order to obtain the most economic solutions. Adaptations after the event are often difficult and costly. Water sprays, possibly using sprinkler systems although still little used, can often be very interesting. The spray heads must be arranged to reach correctly the zones containing the elements to be protected, or those elements themselves. Intumescent paints, the price of which per square metre is rather high, often allows one to avoid aestetic coverings, without finishing up by abusing their use for all the buildings. 3. UPKEEP AND MAINTENANCE OF PROTECTION The premise of global cost requires taking into account in the economics of a project, besides the construction costs, those of the upkeep and maintenance during the life span of the work, and also the alterations which may be needed, including the cost of demolition. Fire protection must be selected, taking into account this idea, as it concerns its own upkeep

-212-

and maintenance, those of the elements it protects and the facility for alterations that we have already mentioned. Some protections are more fragile than others. Nevertheless, it is not appropriate to exagerate these problems as has been done for example with intumescent paints. It is not necessary, contrary to what has been said, to periodically recoat completely. An enquiry by CITICM (France) has shown that, for all cases studied, the intumescence was as good, if not better several years after application as at the beginning. Only wear of degredation due to humidity are to be feared and in those cases the repairs are almost always limited to small affected areas. Possible alterations of the building during its useful life pose the problem of the adaption of the protection initially selected to the new conditions. All changes to the use of an area or entire building necessitate a re-study of the fire protection system, and not only the fire stability of the structure. In these cases, dismountable or easily transformable protection gives reductions in the cost of alterations. It is not therefore always the cheapest protection at the time of construction which will in the end be the most economic. Revisions to the security system will be eased by the existance of a "Safety Register" containing the particular details of the fire stability of the structure and the conditions which caused their use. It appears to us essential, for the security itself, that the occupants of the building are perfectly informed as to these provisions and bear in mind any constrainte on usage they entail. Without doubt this care would often let us avoid the grave consequences of many fires, without it being necessary to put up the costs of buildings for an illusionary absolute security in every condition imaginable, often to the detriment of an easy and correct use.

-213-

4. CONCLUSION Taking into account, from the birth of the design of a building, security to fire and the judicious use of systems of stability for steel structures often gives a considerable lowering in the costs required for this security. In numerous cases, favourable solutions can be found by the Architects and Engineers, in economic and perfectly safe conditions, provided that the application of general regulations to the particular situation do not blindingly clash, but are matched by accurate analyses of the risks and bear in mind the results of much careful research which has been carried out over the last few years. It is to be hoped in particular that the requirements of the degree of stability to fire are revised to take account more precisely of the true construction conditions and the different areas, as well as their use. This would certainly lead to an appreciable reduction in protection costs and at the same time better real security, the protection being better adapted to the most effective places.

BIBLIOGRAPHIE (1) CTICM - C. AIMONE - CAT - J. KRUPPA - G. LAMBOLEY Stabilité au feu des charpentes métalliques. Matériaux de protection. (2) Essais d'incendies naturels dans un grand volume à la Villette (Paris, France) J. KRUPPA 1983 (3) OTUA - Revue L'Acier pour Construire No. 82-1 (4) OTUA - Revue L'Acier pour Construire No. 83-4 (5) OTUA - Revue L'Acier pour Construire No. 77-1

-214-

^Kl>^

^7<1X^

1st case

2nd case

Fia. 1.

VUE AXONOMÉTRIQUE D'ENSEMBLE

ECOI£ D'ARCHITECTURE DE NANTES Fiq. 2.

-215-

Plan of half building : Gangways Scale : 1/500 2mm = 1m

Plan of half building : roof Scale : 1/500 2mm = 1m

Fig.

3.

Mériadek I c e - r i n k , Bordeaux

-216-

Diagrammatic general plan - structure

Scale : 1/1000 1mm = 1m

! Fig. A.

Diagrammatic section A-A Scale 1/1000 1mm = 1m

Bercy Indoor Stadium I

I

Paris 77.00 m 126.00 m

-217-

I

1

,

Fig. 5. UNESCO VI PARIS

PROTECTION PRODUCTS

(a)

COST IN FF per m' (Aug 1982)

1/2 h

42 to 83

1 h

260

1/2 h

210

STABILITY

INTUMESCENT PAINTS interior exterior SPRAYED PRODUCTS (b)

SHEET PRODUCTS

h h h h

38 43 45 45

to to to to

66 100 220 250

1/2 h

91 109 132 205

to to to to

150 150 186 325

1/2 1 1 1/2 2

1 h 1 1/2 h 2 h

(a) The stability is given from standard tests on working to the maximum cold load to French design. (b) The costs are very variable according to the surface finish required. Fig. 6.

-218-

INDUSTRIAL BUILDINGS - FIRE LOSSES AND APPROPRIATE PROTECTIVE MEASURES Dr. J. Thor Swedish Institute of Steel Construction

Sunmary Industrial fires can be estimated to account for two thirds or more of the total fire lose. Therefore it is very important that the parameters that are decisive for fire spread and large fire losses in industrial buildings are recognized so that appropriate fire protective measures can be described. A traditional fire protective measure luce high fire resistance of the load bearing structure can not be expected to reduce the fire losses significantly. This conclusion is based on results from comprehensive fire investigations of industrial fires. Also results from some full scale tests in a large fire compartment centered with computer calculated results leads to the same conclusion. A technique for cost benefit analysis of different fire protective measures is also demonstrated by same examples.

-219-

1.

INTRODUCTION The total number of fires In Sweden per year has been estimated to be

in the order of 30 000. Only a snail number of these fires (a few hundred) becane so called large fires, defined as fires with a total loss of at least SEK 300 000 (50 000 US Ş). However, these few fires account for the majority of the total fire loss. Mareover most of these large fires are industrial fires. An estimation has shown that industrial fires account for about two thirds or more of the total fire loss (fig 1 ) . In rough terms, the same proportion may probably be expected in most industrial countries. If we wish to reduce the total fire loss we must learn more about in­ dustrial fires and those parameters that are decisive for fire spread and large losses In industrial buildings. With this knowledge we can then des­ cribe the appropriate protective measures. Fire engineering design has traditionally been concerned with passive fire protective measures where measures such as high fire resistance of the load bearing structure has been given high priority and relied upon. However, the possibility of reducing the fire loss in industrial buildings through high fire resistance of the structure is somewhat debatable.

2.

FIRE INVESTIGATIONS

2.1 Investigations in Sweden Fire statistics can be used for many purposes. The possibility of using the fire statistics of to­day to determine the influence on the spread of fire and the fire loss on a single parameter, for instance the fire re­ sistance of the load bearing structures, is however very limited. The rea­ son for this is that the spread of fire and the fire loss often depends not only on a single parameter but on a lot of different parameters, many of which are statistically connected and dependent on each other, for In­ stance the type of activity, the type of building, the area of building,, the alarm system, the sprinklers etc. In order to draw any reliable con­ clusion about a single parameter from statistics only, data must be collec­ ted down to a very detailed level and comprise a very large number of fires. This probably means that data have to be collected over a great number of years. During this time there will for instance be changes in industrial planning, changes in building design, changes in building materials and combinations of materials. When enough data have been collected the value of the data may therefore be limited due to all the changes that have taken

­220­

place In the meantime. to other way of learning more fron industrial fires is by thorough in­ investigation and analysis of real fires. Such an investigation and ana­ lysis was carried out by the Swedish Institute of Steel Construction for all industrial fires in single storey buildings in Sweden during 1975 with a loss of SEK 200 000 or more (1). There were 69 such fires, for each of these fires data were collected concerning the: . . . . . . .

Type of occupancy in the building Type and amount of fire load Design of the building and its structures Presence of sprinklers, fire alarm and fire ventilation Origin and spread of fire Detection and fighting of fire Extent and cost of damage to the building and its contents

From the analysis that followed it was found that, on average, the loss of contents accounted for about 50 %, the loss of profit due to stop in production for about 30 % and the loss of building for about 20 % of the total loss (fig 2 ) . Due to the traditionally high priority given to passive fire pro­ tection and the design of building the fires were divided into the follow­ ing four categories according to the type of building:

C S Τ O

= = = =

Concrete structure and roof of conrete or aerated concrete Steel structure and roof of steel sheet Halls of masonry and timber roof structure Others, which can be combinations of the other types or semidetached buildings of different types

number of fires 23 14 14

18 69 As can be seen, there are more fires in pure concrete buildings than in pure steel buildings, which does not mean that concrete itself gives rise to more fires than steel. The explanation in the greater number of fires in concrete buildings is primarily due to the larger number of pure concrete buildings compared with pure steel buildings within the stock of the existing industrial buildings in Sweden. However, of more interest than the number is the cost of fire damage.

­221­

The average direct fire loes for these four building categories is shown in fig 3. Ihe loss is divided into loss of building and contents res­ pectively. Starting with diagram a) it can be seen that the category Τ ­ which means a building with a very combustible roof structure ­ amazingly enough represents the lowest average loss. Ihe explanation is that this type of building in average is very small compared with the other types. Hence the average loss will be relatively small in spite of the fact that the loss often is total. Even if the fire load is low and the original fire is small the very combustible roof structure often gives cause to a rapid fire spread and a total loss. This fact is reflected in diagram c) giving the average loss referred to the area of the primarily damaged room or rooms when the largest fires of the analysis have been excluded. Ihe building structures with the highest fire resistance will nor­ mally be found in category C but also to a relatively large extent in cate­ gory 0. According to fig 3 differences in fire resistance seem to be of no significance for the average loss, this is not too surprising taking into consideration that the value of the building itself normally accounts for a minor part of the total value. An explanation to the very high ave­ rage loss in category 0 is that these objects often consist of very big complexes of semi­detached buildings where possibilities have existed for the fire to spread between the different buildings. After the investigation of the fires of 1975, similar investigations have been carried out for 1976 and 1977 by the Swedish Fire Protection Association. Ihe same tendency as concern the different type of buildings could be noticed (fig 4 ) . Ihe analysis of the industrial fires further showed that the fire load had a great significance for the fire loss. In the investigations the fires were divided into three groups according to the fire load. Ihe fire load was defined as high, medium or low. In all cases where the fire load was high the loss was total, independent of the type of building or type of structure. The number of fires and the corresponding loss within the three fire load groups can be seen in table I. High fire load accounted for 20 % of the fires but for 60 % of the losses. Low fire load accounted for 50 % of the fires but for only 10 % of the losses. The reason for the large fire loss within the fire load category "high" is primarily due to the rapid fire spread and flashover. Often as not flashover occurs a few minutes after the outbreak of the fire. The total amount of energy released after

­222­

flashøver in an industrial building of ordinary size is of such magnitude that no fire­brigade can extinguish it. Table I. Average number in percentage and the corresponding loss of fires in single storey industrial buildings in Sweden which had high, medium and low fire load respectively Fire load

Number

Medium

% 20 30

Low

50

High

Loss

% 60 30 10

Summing up the Swedish analysis it could be concluded that the type of load bearing structure had no significance on the fire loss as long as the buildings were mainly built of incombustible materials which did not contribute to fire spread. More specifically, there was no difference in fire loss between steel framed buildings and concrete buildings. The fire load on the other hand snowed a very great significance on the fire loss. The same is true for large complexes of semi­detached buildings where fire spread very often occured between the different buildings. 2.2 Investigations in other countries In Finland an investigation of industrial fires has been carried out as a diplom thesis at the Tempere university of Technology (4). The in­ vestigation comprises all fires in Finland in single storey industrial buil­ dings with a loss of more than 250 000 FIM (40 000 US Ş) during the year 1975­1980 insured by three of the largest insurance companies in Finland. The number of fires was 95. In table II the average loss is given for different building types and fire load.

­223­

Table II. Survey of fire lose in industrial buildings in Finland during 1975-1980 (4)

Fire load

Building type

Number of fires

Average loss Loes of building and contents

Loss of building

IO 6 FIM

FIM/m

10

2.90

1930

1.37

910

3.17

1420

2.46

1110

Timber

3 2 9

9.29

3270

3.35

1180

Others

8

9.99

5530

2.69

1490

1.37

1330

0.92

1.38

1410

0.84

Timber

5 4 12

1.07

2390

0.43

890 860 950

Other

18

1.88

2490

0.69

910

< 100 MJ/m2

Steel

0.64

(tonerete

0.64

670 760

0.52

(Low)

Timber

2 5 12

1.16

2130

0.44

540 360 810

Others

15

1.28

1420

0.62

690

> 400 MJ/m2

Steel

(High)

Ooncrete

100-400 MJ/m2 Steel (Medium)

Ooncrete

FIM

0.30

FIM/m

In table III the number of fires and the corresponding loss within the three fire load groups are sunned up. The table should be compared with table I concerning the Swedish investigations. Table III. Average nunber in percentage and the corresponding loss of fires in single storey industrial buildings in Finland uhich had high, medium and low fire load respectively Fire load

Number

Loss

%

%

Medium

23 41

Low

36

64 22 14

High

-224-

Similar conclusions as from the Swedish investigations can be drawn from the Finnish investigation i.e. there is no significant difference in fire loss between steel frame buildings and concrete buildings. Ihe fire load on the other hand has a great influence on the fire loss. Also in Norway a survey of large fires recently has been carried out. Again similar conclusions as above could be drawn. For the time being there are two comprehensive investigations of in­ dustrial fires going on in Europe within the frame of BCSC. One is carried out in France by CTICM and one is carried out in the Netherlands by ΊΝΟ. Results from these two investigations can be expected during 1984.

3. FIRE TESTS AND THEORETICAL CALCULATIONS Normally there is a great difference between a fire in a small fire compartment for instance a hotel room, and a fire in a large fire compart­ ment, for instance a large open single storey industrial building. In the first case flashover will occur shortly after the outbreak of fire. In order that the fire shall be confined within the compartment of fire origin the surrounding structures must fulfil certain requirements. A fire engi­ neering design of these structures based on the stage after flashover there­ fore is relevant. Mostly this stage is described by the standard fire curve. As long as the fire is confined within the compartment of fire origin, the possibility of extinguishing the fire also is good. In the case of fire in the large fire compartment the conditions, however, is quite different. In the early stage of fire or as long as the fire is small compared with the size of the compartment the standard fire curve does not describe the situation accurately. The gas temperatures' caused by the fire is much lower due to the mixing of the combustion pro­ ducts by cold air entrained into the hot plume rising from the fire. The problem is not the temperatures but the production of smoke. The smoke even from a small fire may rapidly fill the compartment and damage the goods and make the fighting of fire difficult if the building is not equipped with a sufficient fire ventilation. The fire must be extinguished during its early stage, either by sprinklers or by the fire brigade. If not the fire will grow if there is enough of fire load in the compartment or if for instance the roof is com­ posed of combustible materials in such a way that the roof can contribute to fire spread. The fire growth will sooner or later, depending on the rate of fire growth, result in flashover of the whole compartment. The total

­225­

energy output will then be of such magnitude that the fire brigade can not extinguish the fire until all the combustible materials are consumed. The loss of the contents and most often also of the building will be total. This demonstrates that it is irrelevant to base the requirements of the structures in a single storey industrial building on the traditional con­ cept of standard fire. Before flashover has occured the temperature is low and most type of structures can withstand its effect. After flashover has occurred the fire can not be extinguished until all the combustible mate­ rials are consumed and the loss is then total. This can be seen as a sim­ plified physical explanation of the main results from the fire investiga­ tions of industrial buildings, i.e. there is no significant difference in fire loss between different types of structure as long as they do not con­ tribute to fire spread. In france some interesting full scale fire tests were carried out last summer in an outranged hangar building. Ihe main object was to study the temperature effect on steel structures by local fires in the building (5). Ihe area of the fire compartment was 28 χ 39 m that is about 1 100 m . The height of the compartment was 9.5 m (fig 5 ) . Five tests were carried out and the fire load varied between 2 000­4 000 kg of wood. Ihe fire load 2 was distributed on an area between 39 and 150 m . That means a local fire 2 load density of 450­850 HJ/m . As a comparison it can be mentioned that 2 the mean fire load in office buildings is in the order of 500 to 600 MJ/m . 2 Ihe fire ventilation in the tests varied between 11 and 22 m . A lot of measurements were made of the gas temperatures and of the temperatures in the steel structures. Also observations were made of the smoke production and the height of the clear layer above the floor. In none of the tests there were any high steel temperatures recorded. Normally the maximum steel temperatures were in the order of 100­300 C, that is well below the critical temperature of steel. A computer proyiaii has been developed in Sweden which make it possib­ le to calculate the gas temperatures and smoke filling from local fires in large compartment (6). In data for the calculation are for instance the size of the fire compartment, the fire ventilation, the size or the growth of the fire. Ihe computer program has been used to compare the results from the French full scale tests. Ihe agreement between the observed and measured smoke production and temperatures and the calculated results is good. An

­226­

example can be seen in fig 6 for test number 1 vaiere the calculated height of the clear layer above the floor is given as a function of time. At the test it was observed that the height of the clear layer was at least 3 m which is also demonstrated by photographs fron the test. Measured and calculated gas temperture for the same test is given in fig 7 as a function of time. Ihe temperature is measured close to the ceiling above the centre of fire. The fire load ves 2 000 kg wood distributed 2 on an area of about 60 m . In the same figure also the measured steel temperature in a roof beam above the centre of fire can be seen. Ihe computer progrjn makes it possible to simulate different fire situations in large fire compartment, for instance different rates of fire growth and its effect on smoke filling, temperatures and time to flashover can be studied. Ihe rate of fire growth is defined by the doubling time. A doubling time of 2 minutes for example means that the fire size is doubled every second minutes. With a high rate of fire growth, which today can be expected in many industrial activities, calculations show that flashover may occur within only a few minutes (7). This again is in agreement with the experience from the fire investigations (1), (2), (3). 4.

POST BENEFIT ANALYSIS OF DIFFERENT FIRE PROTECTIVE MEASURES Of great interest for industrial buildings is to compare the cost

and benefit of different fire protective measures. Let us therefore study some different protective measures for an industrial building according to fig 8 (3). Alternative 1 means no fire protective measures at all. Alternative 2 means one of the following protective measures. 2 a)

Sprinklers are installed, which are expected to extinguish a fire in a very early stage

2 b) A partition is built, which is expected to cut the fire loss in half in case of large fire 2 c) Ore whole building is built with high fire resistance in such a way that it is expected that the building without too much of repairs can be used after a large fire (a rather unrealistic assumption)

-227-

The following definitions are nade ρ ρ ρ ρ Β C

= Probability that a large fire will occur during the life time of the building in case of no fire protective measures = Probability of success of sprinklers according to 2 a) = Probability of success of the partition according to 2 b) » Probability of success of the fire resistance according to 2 c) » Value of building = Value of contents

As concerns the loss in case of fire the following is assumed. for a total damage, which is supposed to occur for a large fire in alternative 1 and in alternative 2 if the protective measures fail, the loss can be expressed as B+C. If alternative 2 a is successful the loss is supposed to be 1 % of the total value, that is (B+C)/100. This loss corresponds for instance to 2 smoke damage. The 1 % loss will for a building of about 2 000 m be in the order of 100 000 SEK. If alternative 2 b is successful the loss after a large fire is supposed to correspond to half the total value, that is (B+C)/2. If alternative 2 c is successful the loss after a large fire is supposed to be B/10+C. The 10 % loss of the building corresponds to the cost of cleaning an repainting the building. The above assumed losses are the losses given that a fire occurs and given that the protective measures succeed or fail. If we multiply these losses with the corresponding probabilities we will get the expected losses (8). A sumnary of the probabilities, the losses and expected losses for the different alternatives can be found in table IV.

­228­

Table IV. Sirmary of probabilities, losses and expected losses for the different alternatives according to fig 8 (3) Alternative

1 2 a) success failure

Probability

Loss

Expected loss

Po

wc

P 0 (BK:)

(B4C)/100

P O P S (BK:)/IOO

PoPs

P0(l­P8)(BtC)

(1

Po ­Ps> (B+C)/2

2 b) success failure

P0(l­Pp)(B+C)

Po(1­Pp> 2 c) success failure

POPR

pj? (B+O/2

B/10+C

P O P R (B/IOK:)

P0(1­PR)(B+C)

PO(1"PR)

The expected loss for alternative 2 a, 2 b and 2 c respectively in relation to the expected loss for alternative 1 according to fig 8 can be seen in fig 9. The figure is based on the assumption that the probability of success for each protective measure is 90 %, that is ρ = ρ = ρ = 0.9. S

Ρ R

Under the above mentioned assumptions the expected loss in case of sprinklers is reduced to about 11 % and in case of a partition to about 55 % of the expected loss in case of no fire protective measure at all. In the alternative with a fire resistant building the expected loss reduc­ tion depends on the value of contents C to the value of building B. for a C to B ratio of 6 the expected loss is reduced to about 90 % of the ex­ pected loss in case of no fire protective measure at all. The sprinkler installation gives the biggest reduction of the ex­ pected loss. However, it is not evident that this alternative always is the best one. The costs of the fire protective measures as well as the pro­ bability of a large fire most also be taken into consideration. A resonable level as a basis for discussion of the cost of a protective measure is that the measure will not cost more than the expected loss is reduced. By pre­ ference the cost should be less than the expected loss reduction. This is illustrated in fig 10. At the vertical axis is given the maximum acceptable cost for sprinklers S, for a partition Ρ and for the fire resistance R, in order that the cost will not exceed the expected loss reduction accord­ ing to the assumptions in the example above. The costs are given in rela­ tion to building cost Β and the probability ρ that a large fire will occur

­229­

during the life time of the building. If for instance the value of contents is 5 times the value of buil­ ding and we estimate the probability of success of the sprinkler installa­ tion to 90 %, that is ρ = 0.9 fig 10 gives S/Bpo^5 If the probability that a large fire will occur during the life time of the building is estimated to 10 %, that is ρ « 0.1 we will get S<» 0.5 Β The result implies that the cost of a sprinkler installation S in this case could be about half that of the building cost without exceeding the expected loss reduction. If the probability of success for the partition or for the fire resistance is estimated to 70 % fig 10 gives under the same assumption as above P«*0.2 Β and R*¿0.05 Β By the demonstrated technique it is thus possible to make at least rough estimations of the cost to benefit of different fire protective mea­ sures for industrial buildings. Again similar conclusions can be drawn as from the fire investiga­ tions, for instance an increase of the fire resistance of the load bearing structure can not be expected to reduce the fire loss very much. The oppo­ site, however, may be true for sprinklers and partitions.

­230­

(1)

Sedin, G and Thor, J: Basic information from an investigation of in­ dustrial fires. Swedish Institute of Steel Construction, Publication 61, 1978

(2)

Thor, J and Sedin, G: Some results from an analysis of industrial fires in Sweden. Swedish Institute of Steel Construction, Publication 56, 1977

(3)

Thor, J and Sedin, G: Fire risk evaluation and cost benefit of fire protective measures in industrial buildings. Swedish Institute of Steel Construction, Publication 64, 1979

(4)

Private communication with P. Mäkeläinen, Finland

(5)

Kruppa, J and Lambohey, G: Contribution a 1'etude des incendies dans les bâtiments de grand volume realises en construction métallique. CTICM, Sept 1983

(6)

Hägglund, B: Simulating the smoke filling in single enclosures. FOA Report C20­513­D6, Oct 1983, Stockholm

(7)

Hägglund, B: Hazardous Conditions in Single Enclosures Subjected to Fire ­ a parameter study. FOA D6 42H, Nov 1983, Stockholm

(8)

Baldwin, R and Thomas, Ρ Η: Passive and active fire protection ­ the optimum combination. Fire Research Station, Fire Research Note No 963, London 1973

­231­

tZD Not Industry E333 Industry all loss>200000 SwCf.

F

m.

23

SwCr/tn»

îff.

14

18

b,

8lHoss>200000Sw.Cr

SwCryhrr"

ν'*

SL 67

ω ro

68

69

70

71

72

m

i

73

74

75

23

76 yesr

14

14

IB

Fig. 1 : The percentages of the annual large direct fire Fig. 3 : The direct fire loss of all large fires in Sweden losses in Sweden in industry and other sectors respecţi­ in single storey buildings during 1975 divided with respect vely during a period of ten years (1) to type of building C,S,T and 0 respectively. The loss is given as the average loss within each category. The number of objects can be seen under each colmnnBTOLoss of buildings 1 Loss of contents

M.Sw.Cr.

8H Fig. 2 : The average losses of the building, the con­ tents and the loss of pro­ fit due to stop in produc­ tion for all the analysed fires (2)

5"

1975

ffl

23 M V» 18

1976

SÖQ

25 18 5 20

1977

SÖQ

11 7 2 IS

Fig. 4 : Estimation of the average fire loss for all large fires in single 8torey industrial buildings during 1975, 1976 and 1977 respectively within the four building categories C,S,T and 0. The number of objects can be seen under each column

Plan

Height (m)

>/////////////////////////'//////////////////////////.

E

10

15

20

25

30

ι

28m

^

Time (min)

Fig. 6 : Calculated smoke f i l l i n g and height of clear layer above floor in test number 1

Fire load 2000­4000 kg wood Fire area 39­150 m2 measured gas temp calculated gas temp measured steel temp

»

"Ν. Section

60

Fig. 5 : Full scale fire tests in France (5)

Time (min)

Fig. 7 : Measured and calculated temperatures

1

I

I

?a

IÆ.

I

?h

I

2c

Β

i

% 100·

I H

...

pyO.9

/' 50

­­PÌ.­0.9

Ί'

­ft­05 ι >

5

Lié: 8

Different fire protective measures 1 ■ No measures at all 2a ■ Sprinklers 2b » A partition 2c ­ High fire resistance of the whole building

Fig. 9 : Expected loss for the fire protective measures 2a, 2b and 2c according to fig. 8 in relation to the expected loss in case of no fire protective measure at all. P_, Ρ is the proba­ b Ρ bility of success for the sprinklers, the partition and the fire resistance respectively. Β " value of building. C ­ value of contents.

10C/B

R/Bp0 Resistance /R.­0.9 ­Q5

τ—i—i—ι—ι—ι—ι—ι—η

5 Fig. 10 : The and not Β »

B C/

maximum acceptable cost for sprinklers S, for a partition Ρ for fire resistance R respectively in order that the cost will exceed the expected loss reduction. value of building, C ­ value of contents ρ ­ probability that

a large fire will occur duting the life time of the building p_, p_, p„ ­ probability of success of the sprinklers, the partition and the fire resistance respectively.

­234­

REPORT ON SESSION II : PRACTICAL ASPECTS OF IMPLEMENTING SAFETY Chairman : Ing. Italiano TIEZZI Reporter : G.M.E. COOKE Dr J Kruppa I would like to provide some information about four recent French fire teste carried out in a large building. In one test 2 tonnes of wood were spread over 39 m of floor area, corresponding to a high fire load density of 30 kg/m . Flames 4 to 5 m high produced combustion gas temperatures of 900°C at 5 m height whereas temperatures in steel beams 9Ì m above floor level reached only 35O C. In another test the contents of a modern building, which often involves a mixture of synthetic and cell— ulosic combustibles, were simulated using 3 m of expanded polystyrene slabs together with 1 tonne of wood to give a fire load density of 26 kg/m . In this test, flames reached the roof at 9Ì m but only £or a short period producing combustion gas temperatures of up to 1100 C but only 300 C in the steel members, The test results indicated that uninsulated steel members could be safely used in large buildings of low fire load density such as museums and theatres, but the tests should only be regarded as the beginning of a study of the effect of fires in large buildings. Prof V ELingsch I have the impression from the lectures by Miss Law and Dr Abbado that water cooling in England and Italy is very expensive. In Germany we are also faced with low temperatures, eg. - 30 C, but we are able to achieve an economic system by eliminating interconnecting pipework and this was illustrated in Mr Scnuwirth's lecture which described two recent German projects in which costs of the water cooling system were not at all critical. Miss H Law The economics depend upon the particular design of the building, and the new method of water oooling described by Prof ELingsch is an advance on existing methods. The system used in the Bush Lane office project I described was competitive with conventional encasement but for 2 storey buildings such as the Water Research Centre it was found to be expensive. Inventive engineers can find better ways of providing fire protection, and the use of rational design manuals stimulates designers to provide safer and more cost effective buildings than in the past. Sr G Abbado In response to Professor ELingsch I would Bay that we need to know the life span of the fire protection system so that the costs can be correctly amortized. Mr E Schuwlrth Making a comparison of costs for water cooling and conventional encasement is difficult. In the Hanover project, difficulty of encasing the steel trusses and suspension rods on site meant that using water filled sections reduced the costs substantially.

-235-

Dr H Witte It was interesting to hear Mr Abbado say that solar energy can be gained and stored in water filled external eteel members. There are two German projects in which water filling is used both for fire protection and heating. Work by Blume in Berlin has shown that water cooling can be used to help reduce the costs of air conditioning. Mr Roux Too much emphasis has been placed on the amalgamation of commercial and industrial risks of fire and other types of risk. The heat potential of contents in these buildings is very specific and this means that some generalisations are invalid. The use of an industrial building can change during ite lifetime and this presents a problem at design stage because the fire resistance of the structure or the rating of the sprinkler system may be adequate initially but inadequate at some latex stage. It has been suggested that the provision of heat and smoke outlets meanΒ that less money need be spent on fire resistance methods, but can we rely on these outlets working in all fire conditions? We must be cautious about accep­ ting calculations which allow a reduction in one part of the fire precau­ tions (eg reduced fire resistance or compartmentation) when another part of the fire precautions is added (eg. automatic sprinklers). Insurers have always defended the use of sprinkler installations for impe fling fires but we are very concerned about changes in use in the building which can give rise to rapidly developing fires which may not be controlled by the sprinkler installation. Mr Sette Italian research on intumescent paints, to be published shortly, has shown that the amount of fire protection achieved varies and depends on the type of fire exposure and the thermal inertia of the underlying metal. More research is needed to correlate the performance of intumescents under the ISO 834 exposure and in real fires. Mr L Pruitet Mr Roux said that fire protecting systems should be adaptable to allow for changes in the use of the building, and I agree. Building users must be made aware of safety limitations of their building because the alternative approach of providing large safety factors to allow for unforeseen future uses of the building is not economically viable or sensible. On a separate point I would like to say that one cannot compare durations of exposure in the standard fire resistance test and the real fire. Miss M Law We have just heard that a real industrial fire is very different from a standard fire and most experts would agree. What we really need is a fire engineering approach in which the fire behaviour is calculated, based on experimental data, and the structure is then designed to suit. However, regulations are based on the standard fire, and only in industrial build­ ings do major structural failures occur. For these types of building a fire engineering approach should be used, but the regulation authorities believe only in the standard fire.

­236­

Dr J Kruppa I would like to support Miss Law. Yes, the present regulations are a problem in that they state means of achieving goals which axe not defined. If we can adequately define the objectives then it should be possible for the engineer or fire expert to adopt the best solution. However we must have an objective method of deriving the best solution and it is here that risk assessment has a part to play. Mr Τ Giddings British Steel Corporation has been interested for many years in the use of intumescents, particularly thin coating eystems, for providing fire protec­ tion. In my experience the limit is 1 hour fire resistance and I was therefore interested to hear Dr Abbado mention a system which has approval in Italy for 120 min. Could he provide information on the size of steel protected and the thickness of system used? Dr G Abbado lhe intumescent paint used in the Ancona building provides a fire resis­ tance of 1 hour, whereas the test certificate is for 2 hours, the diff­ erence being the thickness used. I do not have the information requested by Mr Giddings but I can provide a report at a later stage. Prof V Klingsch I would like to dampen optimism over the international exchange and accep­ tance of fire test results. In my lecture yesterday I showed that one may get 30 min in °ne test laboratory and 100 min in another for identical specimens. I aleo showed a paint which basically gave 30 min fire resis­ tance but with marginal alterations could give over 100 minutes: massivity of the steel profile was an important factor. Mr Demartino I do not think all the problems associated with thin soft fire protecting coatings have been solved. Such coatings ehould also provide corrosion protection, thermal insulation, sound deadening and resistance to noise transmission, and, of course, be economical. Mr F Borchgraeve In concluding thie session I would like to say that in considering the design of steel framed buildings we have to find ways of integrating fire safety into all buildings ­ Industrial buildings, dwellings, theatres for example. Biis means considering all aspects of safety from both View­ pointe ­ the viewpoint of the designer on the one hand and the approving authority on the other. We' must consider what happens during the life of the building and what the user wants of his building before we can safely design the fire precautions.

­237­

SESSION

I I I

:

FUTURE

PROSPECTS

Application on the computer to model structural f i r e endurance Computer aided f i r e resistance for steel and composite structures Requirements of f i r e resistance based on actual f i r e s (Swedish approach) A probability based f i r e safety concept Report on Session I I I

-239-

APPLICATION OF THE COMPUTER TO MODEL STRUCTURAL FIRE ENDURANCE David C. Jeanes, P.E. Senior Engineer American. Iron and Steel Institute 1000 16th Street, N.W. Washington, D.C., United States

Summary The rational approach for the design of building structures to resist the effects of fire has developed significantly over the past decade. Work sponsored by many research and industry organizations have resulted in the development of analytical methods to predict fire growth and development, heat transfer to and through the structural frame, and structural response of the building. An evaluation of the published work is now being conducted to develop the appropriate methodology for an engineering design of structural fire endurance. As part of this process two computer models. FIRES-T3 and FASBUS II, have been applied to predict heat transfer and structural response of steel framed floor systems, respectively. An evaluation of these models has demonstrated their ability to predict the extent of the fire exposure and the corresponding performance of the structural assembly.

-240-

INTRODUCTION The analysis of structural fire resistance is a complicated process because of the many variables involved. These variables include fire growth and duration, temperature distribution in the structural elements, Interaction between the building components, changes in material properties, and the influence of loads on the structural system. For this reason the building codes and regulation in the U.S. have relied on standardized test methods (1) to specify fire endurance requirements. Fire endurance times are assigned by the building codes for various portions of the assembly, depending on its relative significance to the overall structural stability.The primary objective of the test methods is to determine the length of time that a structural assembly will withstand exposure to the test conditions. While this approach provides a reasonably simple solution to an otherwise complex problem, it does not provide the designer with a prediction of actual structural performance. Until recently the designer has not played a part in accessing structural fire endurance requirements. The structural design would be made independant of any consideration of the thermal effects of the fire. Fire protection would then be added on to the completed assembly in acccordance with the established test ratings. With costs of the fire proofing representing as much as 20% of the cost structural frame and with attempts to define structural conditions in the test furnance the engineer is becoming increasingly more concerned with the proper design for fire endurance. A more realistic fire endurance analysis can be made based on established engineering principles. Using appropriate computer models, this approach has become Increasingly more practical. With the development of this technology the designer Is better able to evaluate the influence structural response on the performance of supported utilities and systems the effect of compartment

-241-

size and "real" fires on exposure severity, and the potential damage to unexposed portion of the assembly. DEVELOPING A METHODOLOGY The development of a method for the rational determination of structural fire endurance 1s a complex process.To evaluate all aspects of the problem the solution needs to consider three distinct components: the fire exposure, the transfer of heat from the fire to the structure, and the response of the structure. The solution Is not only complicated by the many variables defining each of these components but by the expertise necessary to assess each one. Combustion chemistry tells us the way fires grow. Thermodynamics explains how heat is transferred from the fire to the structure. Metallurgy defines the effects of high temperatures on the properties of the structural steel. Statistical methods help identify the probable risk. The building authorities specify the level of acceptable performance. A proper design method needs to account for the combined effect of all the prescribed conditions. In order to develop the engineering methodology a program was initiated at Worcester Polytechnic Institute (Worcester, Massachusetts) (2). The initial objective of this program was to establish a systematic approach which defines the Interrelationship associated with each of the three components of the fire problem and identifies the many design parameters. Once this was done a survey was Initiated to examine the state-of-the-art technology now becoming available to the designer. Computer models which represent the most significant work, have been identified Figure 1. Those models selected to address various aspects of the solution evaluated are now being evaluated to verify their acceptability, Figure 1. In general, 1t presently appears that the solution to each component of the problem, fire growth, heat transfer and

-242-

structural response, can be reliably solved independently of each other. As the development of this process continues certain of the computer models will be used to conduct sensitivity analyses. From these analyses the key design parameters can then be identified. Statistical methods will then be applied to evaluate the probability of certain exposure conditions developing and the Influence of the these conditions in combination with other loads on the structure. Once these studies have been completed it Is anticipated that a significant simplification In the analysis will be realized resulting In a design method for fire endurance Integrated as part of the basic engineering calculations for steel structures. Several computer models under study as part of this developing methodology provide for the evaluation of heat transfer and structural response. The use and application of two models as "tools" for evaluating structural fire endurance of steel framed floor systems are discussed In the following sections.

HEAT TRANSFER MODEL The ability of a building to remain stable during exposure to a fire has for a long time been equated to temperature rise in the exposed structural elements. This approach Is based on the fact that the mechanical properties of the structural materials are reduced as the temperature of the material is raised to some critical level. The changes 1n material properties most significant to structural performance are: yield strength, modulus of elasticity and coefficient of thermal expansion. The critical level 1s generally defined as the temperature at which the yield strength of the material Is reduced until It nearly equals the design strength and therefore reduces the factor of safety to near unity. However, using temperatures as Input, a structural analysis can be made to more accurately predict performance.

-243-

The fire endurance test has been run repeatedly over the past years for various size members, types of fire protection materials and thicknesses of application. From this data base certain systems and materials have demonstrated consistantly reliable performance. By characterizing the properties of these "proven" materials numerical techniques for solving the heat transfer problem can be approached with reasonable confidence. The factors influencing the heating of a structural member include: the thermal properties of the materials, the surface area exposed to the fire, and the Intensity and duration of the fire. Each of these factors, inherently present In a fire test, must be specifically defined in modeling the heat transfer. FIRES-T3 Model The computer model (3), FIRES-T3 (Fire REsponse of Structures - Thermal 3_ Dimensional Version) is a three dimensional finite element heat transfer program.

It 1s suitable for use in evaluating the temperature distribution

history through solids of composite materials such as fire protected structural steel and reinforced concrete. A limitation of the present version however, is that it cannot model heat transfer through cavities 1n the assembly. The model allows for consideration of the nonlinear characteristics of the thermal properties of the materials and the heat transfer from the fire environment. The solution technique requires an iterative intergration process within each time step throughout the exposure period. Accordingly, the program user must exercise judgement as to the appropriateness of the solution as the analysis progresses. The principal factors influencing its effective use are the layout of the finite element mesh and the selection of the time-step size. Both factors need to be dimensioned so that sufficient detail is available in the region

-244-

and over the time period for which the thermal analysis can be expected to be most sensitive. From a user point of view, the FIRES-T3 model allows for consideration of the following design parameters, Figure 2: 1. Material Properties - the thermal properties (thermal conductivity and specific heat) and density of materials are considered with respect to their change in value at elevated temperatures.

(Effects

of internal heat generation can also be considered). 2. Fire Environment - the time-temperature history of the heated environment is considered by specifically defining the temperature at each time step during the solution. Therefore, the fire exposure curve can take any form (ie. constant temperature, linear change, El 19 curve or natural burning). 3. Heat Transfer - the heat transfer process due to the fire exposure is modeled as convection and radiation in the fire boundary and as conduction through the member. The emissivlty of the flame and surface, view factor, and surface absorption are considered In calculating radiation effects. Convection Is modeled using a convection factor and power of convection. Conductivity is computed using the appropriate material properties. 4. Geometry - the shape and size of the structural element can be considered In one-, two-, or three dimensions. This is accomplished by drawing a mesh representing the shape and arrangement of materials of the element and describing this arrangement in terms of the coordinates of each of the nodal points in the mesh. The output of the FIRES-T3 analysis provides a listing of the calculated temperature at each node, the average temperature of each element in the mesh, and a summary of the test conditions at that point in the analysis.

-245-

In order to evaluate the accuracy of the FIRES-T3 program, the program was first used to model assemblies for which actual test data was available. This approach allowed for confidence to be established in the model without a need to be able to specifically understand all modelling techniques used. Predictions were made of the heat transfer through steel beams with direct-applied fire protection material from assemblies tested at Ohio State University, Underwriters Laboratories and the U.S. National Bureau of Standards. The modeling was done using the nodal mesh Illustrated in Figure 3. The results demonstrated favorable agreement between the predicted and recorded average section temperatures and the temperature profiles through the sections, Figures 4. As a result of the satisfactory agreement demonstrated by this modelling, a series of analysis were conducted in order to develop data useful as design aids. This was done by analyzing different size steel beams with direct applied fire protection thicknesses of 1/2, 1, and 1 1/2 inches. The beams were selected to cover a range in W/D values from 0.5 to 2.5. The fire exposure used in the analysis was the ASTM El 19 time-temperature curve over a four hour period. The results of this series of analysis have been compiled and presented as "Fire Endurance Time versus W/D", Figure 5. This general form of the data utilizes the W/D characteristic of the beam as the basic design parameter. The data presented is based on the average section temperature for the 1000F (538C) criteria. STRUCTURAL RESPONSE MODEL The structural fire endurance of a building system Is a measure of its ability to resist collapse during exposure to a fire. The approaches used to make this measurement range from the use of standardized laboratory tests to the application of engineering methods. In either method a certain level of damage is acceptable provided it does not result in the collapse of any part

-246-

of the assembly or contribute to the spread of the fire. The influence of a building fire on the structural steel frame is not often significant until or unless the fire becomes fully developed, therefore, the period of fire growth 1s not usually considered as part of the structural fire endurance time period. The analysis of the structure exposed to a fire can be accomplished using the established principles of engineering mechanics applied 1n conventional design practices. The analysis, however, needs to recognize the continuingly changing properties of the materials at elevated temperatures. Those properties which are most significant to structural performance are: yield strength, modulus of elasticity, and coefficient of thermal expansion. Studies have been made to characterize the changes in these properties with temperature. Utilizing this data structural fire endurance can be determined by repetitive calculations. Because of the ability of the computer to quickly solve these dedious types of problems, modelling techniques have been developed making it possible to provide this kinds of analysis for steel framed floor systems. FASBUS II Computer Model The computer model, FASBUS II (FIRE Analysis of Steel Building Systems) is a structural analysis program specifically designed to analyze the fire endurance of steel framed floor systems, Figure 6. The model utilizes the finite element method where beam elements and non-conforming triangular plate bending elements are used to represent the frame and slab, respectively, Figure 7. The incremental solution used by the model provides for consideration of changes In temperature, with corresponding changes in material properties, throughout the exposure period. Using an Iterative process the model determines the displacements necessary to bring the structure to a point static equal i bri urn under the loads and conditions Imposed.

-247-

As with any engineering analysis, the designer must have a basic understanding of the problem being solved and the solution techniques applied. Accordingly, the user of the FASBUS II computer program should have a basic knowledge of structural mechanics, an understanding of the modeling techniques and a familiarity with both methods of building construction and the thermal effects of a fire. With this background the user will be able to more accurately define the physical characteristics of the problem and express them in terms identifiable to the computer model. From a user point of view, the model provides for consideration of the following design parameters: 1. Geometry of Structural Elements - In addition to the layout of the framing members, detailed description of the structural elements is permitted which Includes shape and placement of the steel beam sections, deck profile and reinforcement locations. 2. Material Properties - Non-linear changes 1n the yield strength, modules of elasticity, and coefficient of thermal expansion with respect to material temperature are inputted directly (material models within the program allow consideration of the elastic/plastic character of steel and cracking or crushing of concrete). 3. Loads and Restraint - Provision is made for the direct Input of point loads (any direction) and uniform vertical loads acting on the floor system. The resistance of structural elements connected to the assembly, such as columns and braces, can be modeled. 4. Time-Temperature Exposre - The shape of the temperature profile with respect to time of exposure for up to five groups of elements In the model can be specified. Such

-248-

profiles are based on either measured or calculated data which reflect the nature of the fire exposure being considered. The results of.a successful analysis provide the designer with predictions on deflections and rotations across the floor system and stress and strain conditions within the structural members. Because the engineering approach represented by the computer model 1s a significant departure from the laboratory test methods contained In the building codes, a substantial evaluation of the computer model was necessary. The analysis of data collected from a large scale test program conducted at the U. S. National Bureau of Standards provided the basis for this evaluation (6). The test program provided for the measurement of the response of a structural system representative of actual building construction, when a portion of that assembly 1s exposed to fire. This was accomplished by recording vertical and horizontal deflections of the frame and floor slab and temperatures on the exposed structural components. The actual test assembly consisted of a two story-four bay structural steel frame with a concrete and steel deck floor slab Figure B. A total of three tests were conducted on the assembly which Included both controlled exposure fires (ASTM El 19) and a free burning "real" fire. Evaluations of the fire exposure conditions recorded during each of the tests have been made using the finite element mesh illustrated In Figure 9. Comparisons made between the record and predicted performance for each set of test conditions demonstrated good agreement for both the deflected shape of the floor assembly and level of damage to the concrete slab and steel frame. These comparisons for a 90 minute exposure to the ASTM E119 time-temperature curve are Illustrated in Figure 10.

-249-

APPLICATION OF THE MODELS Once the computer models have been developed and adequately validated they can be considered as engineering "tools" in the design of the structural elements. The computer models, FIRES-T3 and FASBUS II, are more specifically analytical programs than design programs. As such, the models are used to evaluate a certain set of exposure conditions of a particular structural assembly. Using the results of the analysis the designer can then determine the acceptability of the predicted performance. The scope of the analysis must first be determined. This Includes identifying the assembly or portion thereof to be modeled, the temperature conditions of the exposure, the distribution of live loads (or load combinations) and the types of materials and construction represented. The structural assembly to be analyzed must then be redefined in the form of the element types included 1n the model. The size of the elements Is determined by the dimensions of the "nodal mesh" Into which the assembly Is divided. The spadai position and dimensions of the elements, defined by the nodal coordinate system therefore, should match that of the actual assembly. The application of the analytical techniques for determining structural fire endurancce are permitted within the U.S. building codes under the general provisions for "alternates" to the prescribed code requirements. To exercise the provisions, the designer must produce evidence sufficient to satisfy the Interests of the responsible building official. The specific requirements for a particular case will therefore vary according to the level of Interest and expertise of the individual of f i cal reviewing the analysis. Typical of the acceptance of most new design approaches to the engineering requirements of a structure Is the need for a project significant enough to warrant the interest of the designer. The following 1s a brief description of a successful analysis, using the two computer models described

-250-

above, which resulted in In establishing a change to the fire protection requirements on portions of a structural frame. Evaluations of An Office Building The first application of the analytical approach represented by the FIRES-T3 and FASBUS II computer nodels was first made on a 42 story office building located on the West coast of the U.S. (7). The interest of the designer developed when he considered the exccessive f1 reproofing requirements placed on the large spandrel beams of the building frame. The specific beams, which were designed to carry earthquake forces, provided substantial bracing to the structure through moment connections at the columns. Despite the structural conditions, fire endurance requirements specified by the code to be based on a standard fire test rating of a simply supported beam significantly smaller in size than those in the structure. In addition, the code required a three hour rating on columns and these beams, because of their function in stabilizing the columns. The building designer was only Interested in evaluating the spandrel beams, thereby limiting the analysis to the assembly shown in Figure 11. Using a direct applied fireproofing thickness of 3/4 inch (1.9 cm) a FIRES-T3 analysis was conducted on each of the spandrel beams of size W33xll8 and larger. The results of the analysis after 1-1/2 hours and 3 hours of exposure for the W33xll8 beam predicted a high point temperature of 1490F (810C) and an average section temperature of 1300F (705C), Figure 12. Using the predicted temperature history conditions a structural analysis was made using FASBUS II. The modelling considered only the gravity loads supported by the beam thereby ignoring the higher design stress levels only considered to occur during an earthquake. The influence of the frame columns were modeled as equivalent stiffnesses applied to the ends of the beam. From the results of the analysis the vertical deflection and elongation of the beam

-251-

could be examined over the course of the exposure period, Figure 13. In addition, the stress levels across the beam section could be evaluated to determine the development of plastic material conditions. The combination of the heat transfer analysis and the structural response modeling provided evidence satisfactory to the building official. As a result the thickness of the fire protection material on all the spandrel beams equal to or larger than the beam analyzed were reduced to 3/4 inch (1.9 cm). This limited analysis of the structures fire endurance resulted in a savings of over $250,000. SUMMARY

The development of an engineering method for calculating structural fire endurance of steel buildings is now under development In the U.S.

Based on a

study of these design parameters, a systematic approach has been defined which Identifies the various components of the design problem and their Interrelationship.

Examination of the state-of-the-art technology available

for addressing each of the design parameters is now being accomplished. Because of tne complexity of the problem computer models are commonly required. Two computer models, FIRES-T3 and FASBUS I I , have been developed to predict heat transfer and structural response, respectively.

As a result of a

substantial evaluation of each of these models their validity to accurately predict structural fire endurance of typical steel framed floor constructions has been established.

The application of these models to actual building

constructions has demonstrated their value as an engineering "tool". With the continued development of a engineer solution these and other computer models are being used to identify and evaluate each of the parameters.

The most significant of these parameters will be used to

developed a concise and optimum design.

I t Is anticipated that the design of

structures to resist the effects of a building f i r e will eventually become a routine part of the structural design of the building frame.

-252-

REFERENCES

1 ) "Standard Methods of Fire Tests of Building Constructions and Materials", ASTM El 19, Annual Book of Standards, Part 18, pp. 941-967, American Society for Testing and Materials, 1982. 2) Fitzgerald, R.W., Development of an Engineering Method to Calculate the Fire Resistance of Structural Steel Frames, Status Report to American Iron and Steel Institute, Worcester Polytechnic Institute, December 1982. 3)

Iding, R.J., Bresler, B. and Nizamuddin, Z., "FIRES-T3, a Computer Program for the Fire Response of Structures-Thermal", Report No. UCB FRG77-15, University of California, Berkeley, October 1977.

4) Chiapetta, R.L. et al, "The Effect of Fire Temperatures on Buildings with Steel Frames", Final Report IITRI Project J8095, Chicago, ILL., April 1972. 5)

Iding, R.H. and Bresler, B., "Effect of F1re Exposure on Steel Framed Buildings", Report to American Iron and Steel Institute, WJE No. 78124, Wiss, Janney, Elstner and Associates, Inc., Emeryville, CA., March 1982,

6) Jeanes, David C , "Predicting Fire Endurance of Steel Structures", Preprint 82-033, American Society of Civil Engineers, ASCE Convention, Nevada, April 16-30, 1982. 7) Bresler, B., Iding, R., Amin, J., and Laws J., "Evaluation of Fire Proofing Requirements for a High-Rise Steel Building" paper presented at the AISC National Engineering Conference, Memphis, TN., April 1983.

FIRE GROWTH

E-119 ÍUHNACE COMPUTER M O D U S HAOVIG (DENMARK)

HEAT TRANSFER

STRUCTURAL RESPONSE

UNPROTECTED

COMPUTER MODEL

STEEL;

FASBU5 II

INSULATED STEEL;

AIMLYIICAL MC IMOD

MEMBRANE POST · FLASHOVER F WE

PROTECTED

VCIIiriCAllON

niLiAUiuir ANALYSIS

DESIGN METHOD

IlltOHCIICAl STUDIES

STEEL COMPUTER MODELS:

COMPUI ER IfMJOELS HARVARD

FIRES T3

exPUOMCNIA!

TASCF 2

IESI

COMHF H APPROK

IO

COMPF II

PflE-H ASHOVEÍ1 FlKC COMPUI Ell M O W I S SMI I I I IOIUO SIA1E) HARVARD

Fig. 1. DEVELOPING A DESIGN METHOD FOR FIRE ENDURANCE OF STEEL THE IDENTIFICATION OF SOME APPROPRIATE COMPUTER PROGRAMS.

-253-

Mat'I Prop. Fire Expo. |-

Computer

Model Temperatures

FIRES T3

Heat Tran, ρ Geometry

FIGURE 2

The I n p u t / O u t p u t Data C h a r a c t e r i s t i c s of Computer Model

the FIRES-T3

MESH FOR BEAM THERMAL ANALYSI S

FIGURE 3

F i n i t e Element Model Used f o r FIRES-T3 A n a l y s i s of S t e e l Beams with D i r e c t Applied F i r e P r o t e c t i o n

-254-

1500

BEAM:

W12

χ 27

F ι REPROOF I N G :

(W/D = 7/8"

0.63)

T H I C K (rcNCKOTE)

1000 ­

500

— O

FIRES­T3

0

F i g . 4.

3.0

TIME, Hours

Comparison of FIRES-T3 p r e d i c t e d temperature with recorded t e s t data

-Γ 0.0

­ ι — 2.0

1.0

0.0

PREDICTION

TEST DA TA

-

0.5

-r-

1.5

1.0

¡a

W/D of B:

Fig. 5. Fire protection thicknesses for steel beams based on average section temperature of 1000F

­255­

­ ι 2.5

TEMPERATURE PROFILES MATERIAL PROPERTIES GEOMETRY STRUC/ELEM

STRESSES COMPUTER ANALYSIS (FASBUS II)

STRAINS DEFLECTIONS ROTATIONS

LOADS AND RESTRAINTS FIGURE 6

a.

The Input/Output Data Characteristics of the FASBUS II Computer Model.

Beam Element

b.

Slab Element

FIGURE 7

Types of Finite Elements Used in the FASBUS II Computer Model.

-256-

STEEL DECK ROOF

TRUCTURAL STEEL FRAME T I E ANGLES

CONCRETE/STEEL DECK FLUOR SLAB

INDIVIDUAL FOOTI

FIGURE 8

FIRE COMPARTMENT

S t r u c t u r a l F i r e E n d u r a n c e T e s t Frame a t t h e U . S . N a t i o n a l B u r e a u of S t a n d a r d s (NBS).

wv = LOCATION SPRINGS TO SIMULATE COLUMNS

FIGURE 9

F i n i t e E l e m e n t Mesh Used t o Model NBS T e s t

-257-

Frame.

♦ TEST DA TA • FA SI1US DA TA ( J T S . P I N N E D ) o FA SBUS DA TA ( J T S . F I X E D ) o

FIG. 10a ' Vertical deflection a t the' c e n t e r of the t e s t bay during loo course of the

the test.

TIME, MINUTES

+ TEST DA TA • FA SBIIS DA TA • FA SBIIS DA TA

( J T S . PINNED) ( J T S . FIXED)

FIG. 10b Vertical deflection across c e n t e r of t e s t bay f l o o r s l a b (a 90 m i n . , El 1 9 ) .

+ TEST DA TA

· FA SBUS DA TA

( J T S . PINNED) FIG. Lateral

10c

deflection

I of test floor frame

]

t£^­r_ ­258­

(a 90 min., E119) .

CRACK DATA: ■ FASBUS (STRESS! 4 PULSE ECHO DAT —­

FIG. lOd

VISUA L (SHRINKAGE, AND STRESS)

Comparison of recorded

slab

crack data (after test) .

Figure

10.

Comparison of FA SBUS

II predictions with test data.

ufa ι ;­­­ _■?_. .--ytiìr*

¿φ.

­FLOOK. LEVEL

I

I

I / /I

ι

I T

\ \ \ \

\ ■tør

-ΐΊ-otx. Levee &EUM/-^n

FIG. 11 .

Üííu

ABOVE-^Ά

I I

ι 41

_i_L +_L \ \

\1 \

u

Portion of 42 story office building using FIRES­T3 and FA SBUS II.

­259­

analyzed

Srs

LOCATION

'AVG

w/d-1.21 3 HOURS

T

l

980*F

1480-F

T

2

990"F

U90*F

Τ

3

990*F

U40-F

Τ

4

790T

1190T

5 Τ 6

• 630*F

990T

°0"F

200'F

870*F

1300"F

Τ

(^

W3ixiia ík HOURS

STEEL

(810 O

3/4" Monokote FireproofIng

FIG. 12.

Temperature predictions in W33xU8 spandrel beam using FIRES­T3.

I20

2¿>

iSo

iea

EXPOSURE TIME, MINUTES FIG. 13.

Structural response of W33xll8 spandrel beam predicted using FASBUS II.

­260­

COMPUTER AIDED FIRE RESISTANCE FOR STEEL AND COMPOSITE STRUCTURES.

3.C. DOTREPPE J.M. FRANSSEN Senior Research Associate Research Assistant National Fund For Scientific Research (Belgium) Department of Civil Engineering University of Liège, Belgium

3.B. SCHLEICH Department Manager ARBED-Recherches Luxembourg

SUMMARY. In order to improve the evaluation oí the fire resistance of steel and composite structures an E.C.S.C, research has been introduced. One part of this research consists in developing a numerical model for the analysis of these types of structures in a fire environment. This model is based on the finite element method using beam elements with subdivision of the cross section in a rectangular mesh. The structure submitted to increasing temperatures is analyzed step-by-step using the Newton-Raphson process. A comparison between theoretical and experimental results is made for a composite beam. A further calibration of this numerical model will be achieved by the end of 1984, when new practical test results are available for columns, beams and frames, according to the aforementioned research program.

-261-

1. INTRODUCTION. The standard fire resistance test according to ISO 834 has been used quite intensively to determine the fire resistance of structural elements. Nevertheless in its present form the test procedure has several shortcomings, for instance concerning the heating and restraint characteristics. This last point may be considered as the main weakness of - the standard test, since the structural response is highly dependent on the conditions of restraint due to the building system and the end conditions. Therefore the need for analytical predictions of thermal and structural responses has grown more and more intensively. During the last decade there has been important progress in the development of analytical methods for the calculation of the behaviour of structures under fire conditions. In several countries the practical evaluation of the fire resistance can now be made through simple methods of calculation. This type of method is already available for almost all steel elements, but it is not yet applicable to all composite and concrete elements. This is due to the fact that the concept of critical temperature cannot be applied to all cases. It is then necessary to use tables and empirical relationships based on tests and experience. Though all these methods are very useful for the designer, the element will probably behave differently in a real structure if a fire occurs. It has been found that very often the protection of this element appears exaggerated. To improve the prediction of fire resistance, it is necessary to have a very powerful numerical tool, i.e. a computer code able to simulate the real behaviour of the structural element in a fire environment. This type of analysis should lead to an improvement of the competitivity of steel and composite constructions. Therefore a research called REFAO/CAFIR (7) has been introduced by ARBED and accepted by the ECSC authorities. This research contains an experimental part and a theoretical one. The tests will be executed in various European laboratories, but the experimental part of the research will not be discussed in the paper.

-262-

The theoretical part is realized in the Department of Bridges Structural Engineering of the University of Liège. The code is based on computer program presented in (5). I t must be developed in such a way it can be applied to all types of protected and unprotected steel composite (steel-concrete) structures.

and the that and

2. RECENT DEVELOPMENTS CONCERNI NG THE BEHAVI OUR OF COMPOSI TE STRUCTURES UNDER FI RE CONDI TI ONS. Considerable progress has been achieved recently in the field of composite construction elements and their individual behaviour under fire conditions. Composite beams (1, 8) as well as different composite column types (16) have been tested in several countries. Simplified calculation models, based on real fire test calibration, have been developed enabling architects and engineers to perform a quick analysis of composite columns for instance. Even if these methods are limited to given application fields depending on tests, they however permit right now a practical design in many situations. A very typical example of these fire test calibrations is the "reduced composite cross section" method of ARBED presented in (14). This method has been developed for the so-called AF-columns, i.e. rolled profiles, concreted between the flanges (see fig.l). I t should be underlined that the fire endurances measured (3) and those computed according to this simplified calculation method are in good agreement (see fig.2). Besides a catalogue (12) and diagrams for AF 30/120 columns based on rolled Η-profiles have been established. These most practical design tools give immediately the adequate composite section (including rolled Η-shape, concrete and reinforcing bars) in function of the axial load, the column length and the required time of fire exposure (see fig.3). In spite of the undoubtedly very high practical value of these simplified calculation models, it must be recognized, however, that problems arise as soon as the real situation is outside the application field covered by tests. So it will be difficult to find the adequate answer for very long or slender columns under fire conditions or to analyze special types of composite cross sections (see fig.4). Another important problem is the M/N interaction behaviour of composite columns (see fig.3) for which solutions are given (6, 13) at ambient temperature. No answer is available for this interaction

-263-

Method of 'reduced composite cross section' for AF 30/120 columns

*·

FIG. 2. Comparison of fire endurances measured from tests and computed according to the simplified calculation method, for AF columns (3) Ncr g(MNl

10 r

90

""^T.*>

ni

Ll<") FIG. 3.

Ultimate buckling loads of ARBED HEAA composite sections after

90 min. of exposure to iso fire jCT. 235 N/mm;ß=45N/nrwfi )

-264-

Fire retardant point

I2

Fire r e t a r d a n t paint

FIG, k·

\

Special types of composite cross sections able to support axial loads and bending moments according to yy or/and zz axes N/Npl

FIG. 5. M/N interaction diagram for AF composite columns at ambient temperature

-I

COMPRESSION H 12) WEB

1

0,2

ideally plastic with strain limitation 1 1 1 1 1 1 0,4

0,6

0,8

\\\\ VM/kfeiy 1—*—I U IjO W

CRACKED CONCRETE ÍS)

FIG. 6. Internal stress diagram due to the temperature field in a composite AF section, after 120 min. of ISO fire according to (10)

-265-

behaviour at high temperature, in which case oí course the simplified model approach is unsuited. Besides it should be noted that the simplified calculation models do not cover the effects of the internal thermal stress field of a composite cross section. However the highly differential temperature field, created under fire conditions in a composite cross section, leads to strong internal stresses (fig.6) which undoubtedly affects the load bearing capacity. For these reasons a numerical model is needed in order to allow an exact thermal and mechanical system analysis without any restrictions as to the geometry of the cross-sections, the building structural system, the load combinations, etc.

3. NUMERICAL PROCEDURE FOR THE ANALYSIS OF STEEL AND COMPOSITE STRUCTURES UNDER FIRE CONDITIONS. 3.1. Basic knowledge for the theoretical analysis. Before the development of the fire the element is submitted to external loads corresponding to the situation existing in a real building. Therefore the structural behaviour of the element at ambient temperature under static loads must be analyzed. This is rather simple for steel beams where the material is assumed to be perfectly linear elastic. Some complications arise for columns where second order effects must be taken into account. For composite or concrete structures the problem is much more complicated since the stress-strain characteristics of concrete are no longer linear elastic and cracks appear for small tension stresses. Therefore a step-by-step analysis taking into ' account material and geometrical non linearities is used as a general procedure for this first part. To analyze

the

structure

during

the development

in

evaluation

of

the fire

two

distinct problems must be solved : -

a

thermal

problem

consisting

the

-

distribution in the element ; a mechanical problem consisting in the evaluation of the behaviour due to the temperature increase calculated hereabove.

-266-

of

the

temperature structural

To

solve

these

problems

analytically

it

is necessary to collect

data

about thermal and mechanical properties of the materials used, i.e. steel and concrete. Furthermore, due to

the high temperatures reached, the variations

of temperature affect significantly the properties of these materials and this must be taken into account in the numerical model. The

thermal

properties,

i.e.

the

thermal

capacity c, the density ρ , the thermal diffusivity

conductivity λ ,

the

heat

a = λ/c ρ and the thermal

strains will not be discussed here. The models adopted here are essentially the same as those described in (5). Concerning mechanical properties simplified methods for the evaluation of

the

fire

endurance

characteristics,

i.e.

require

ultimate

only

strength

the in

determination tension

of

the

and compression,

classical yielding

stress and modulus of elasticity. I n a step-by-step numerical procedure these characteristics are not sufficient and information concerning the instantaneous stress-strain relation, creep and relaxation are necessary. Experimental relaxation mainly

of

true

investigations

concrete for

and steel

elements

where

(9)

(15)

have

a

show

that

thermal

non negligible

compression

has a

creep

influence.

significant

and

This is

effect

like

columns, while in elements where bending is determinant like in beams and slabs it has been proved (5) that creep does not influence very much the failure mechanism and the fire endurance. Creep and relaxation models have been proposed for both materials (9) (15). These types of models should be introduced in the program in the near future. Up to now these effects are taken into account by adopting "smooth" stress-strain diagrams for concrete (figure 7.a) and steel (figures 7.b and c). 3.2. Temperature distribution in the element. The

first

problem

to be solved is the modeling of the

environment

created by a fire. The variation of external temperature is usually given by the standard temperature-time

curve defined in I SO 834, but other types of

equations can of course be introduced in the program. The density

of heat flow

transmitted to the element is a problem

quite involved. Usually it is divided in a convection part and a radiation part

-267-

300t 400 .500 .000 J 700 .eoo 20 β 4 as·«* 4 50­1Ö4

α) concrete

■β

2­IO­

iff 3

b) structural and hot­rolled reinforcing steels

Fig. 7

2­10­3

c ) cold­worked reinforcing steels

: Models presently adopted for the variation with temperature of the stress­strain diagram of the materials

"T b(i*i)

|Q3T Q

m!

!<»I

b(D

4­­Í­ I J ??. —i — «(M)

Fig. 8

• (I)

bO­D

■ (IO) ­

: Heat balance between adjacent elements

­268­

and is written as follows * = h (Τ e

- Τ) + σ ο

ε es

(Τr* - Τ, Λ) e

Τ

:

surface temperature of the element

Τ

:

temperature of the fire environment

h

:

(1)

coefficient of convection

°" : o e :

Stefan-Boltzmann constant resultant emissivity factor between the environment and the surface of the element.

The coefficient dependent

of convection and resultant emissivity are temperature

and are influenced by many

parameters. Nevertheless

one of

the

authors has shown (5) by numerical experimentation that constant values can be adopted for most cases. In order

to calculate

the distribution

of temperature in the elements

the equation of heat conduction must be solved. In the case of fire problems the thermal numerical

properties

methods

differences

of steel and concrete are temperature dependent and

must

equations

is

be

used.

obtained

In

by

this

approach

expressing

the

a heat

system balance

of

finite

between

adjacent rectangular elements (figure 8). The following possibilities can be considered in the program : - composite section ; - variation with temperature of the thermal properties ; - holes in the cross section ; - evaporation of free water.

3.3. Solution strategy for the thermomechanical analysis of the structure. In considered

the as

finite

element

an

assemblage

method a conventional of

structural

engineering structure

elements

interconnected

at

is a

discrete number of nodal points (fig.9). If

the

force-displacement

relationships

for

the individual elements are

known i t is possible to derive the properties and study the behaviour of the assembled structure.

-269-

'λν/t

VT??­

Element 1 2 nodes 1 2 ifrs> ­O

3 3 O

4 4 O

t

5 O

O

O

­o—o—o—o

o­ wfe

Figure 9 : Discretization of the structural element

'Λ Λ / Λ 'X/> Λ Λ > « xyx/ « / / <*/ /y; »

2Z2ZZ2ZZZZ2ZZZZZZZ

i Figure 10 : Discretization of the cross section for a composite element

­270­

The basic equation of the method can be written : {F e > = (K) . {u} {F }

:

(K)

:

(2)

vector of nodal forces applied to the structure structure stiffness matrix ; depends upon geometrical and material properties of the elements

{u}

:

vector of nodal displacements.

After solving the system (2) and determining the nodal displacements, the displacements at any point within each element can be defined as a column vector {f} : {f} =

(N) i u > e

(3)

in which (N ) are in general functions of position {u}

represents a listing of nodal displacements for a particular element.

With displacements known at all points within the element the strains at any point can be determined. These will always result in a relationship which can be written in matrix notation as : {e}

=

(B) . {f}

Using the appropriate material properties the stresses as functions of strains.

(4)

{σ} can be calculated

Because of material and geometrical non-linearities, an iterative approach is essential. Before fire occurs the loads are applied step-by-step and when the external temperature increases the time is divided into time increments At. Stress-strain relations in the materials are non-linear and moreover they vary with temperature. Since it is also desired to take large displacements into account the stiffness matrix has to be actualized at each step of the loading and at each time increment during the development of the fire. In the problem to be solved the materials are subjected to initial strains due to temperature changes ( e J and to creep effects ( ε ) ; at the o

cr

present time creep effects are not yet taken into account in the model. Thus the stresses will be caused by the difference between the total s'trains ( e.) derived from the nodal displacements and the initial strains :

-271-

ε σ ( ε ) = oier ■ "Γ -- ε„ θ -" ς.,) % σ

(5)

When the internal nodal forces {F. } are calculated by integrating the internal stresses (5) and compared with the applied nodal loads {F ], it can be observed that equilibrium is not reached. Thus, at every stage, the difference between the internal forces and the applied ' loads is determined at all nodes of the structure. These unbalanced residual forces are then redistributed throughout the structure to restore equilibrium. This combined with the actualization of the stiffness matrix gives rise to the Newton-Raphson process. Successive iterations take the form : (r) {AF

,

e\

,

= (K\

.(r)

(r)

·

{Au}

i

(6)

.(r)

IK L

(r) <*Fe>i

tk

:

structure stiffness matrix updated at the beginning of the r iteration in the i increment taking into account the changes in material and geometrical properties, unbalanced residual nodal forces.

The main originality of this program lies in the discretization of the cross section, which is divided into subslices forming a rectangular mesh (fig. 10). The discretization of the cross section is chosen in order to be the same as in the thermal analysis. Therefore the temperatures, strains and stresses can vary from one subslice to another. Thus the integrals appearing in the equations and the properties of the cross sections are computed in a discretized way.

4. COM PARISON BETWEEN THEORETICAL AND EXPERIM ENTAL RESULTS FOR A COM POSITE BEAM . To demonstrate the accuracy of the numerical results which can be obtained from the described procedure, a composite Τ beam has been analyzed and the theoretical results compared with test results obtained at the Technical University of Braunschweig (3). The loading and heating system is presented in figure 11. The beam is loaded and heated symmetrically. The thermal program is applied according to the ISO R 834 Recommendations. The dimensions of the cross section and the

-272-

FIG.

11.

LOA D ING AND

HEATING SYSTEM

u

(i),l»,fl,ffi,g),fi*ffl,g)|

i&hr 1

2

3

ί,

VI FIG.

12.

ELEMENT AND

5

6

7

β

! S Q ΖΖΖΖΖΖ2Ϊ

¡9

Ί I

SECTION D IVISION

! rrA//ts/i/r7t?7.

­273­

•c

s'

/ ',

/ // /

. •C

•C

S-*

A

•00

/

MO

¿..­'

no

1 t' ■00

/

MO

100

/,

t

tl· j0

' 40

I0

0D

«B

1B

> 9

10

ID

M

I

go

no

tIO

»*«

/

too t l · η)

1

J

t 10

Μ

bl

I)

FIG. 13­

I

ki

ini

V Ut

rip

fi

I

Ι,

100'

«

s

Λ

/;

100

•s

•ΛΛ ir*

BO'

I

u

­­

MO

'/

«00

liI

»0

s

/ ' //

I0

α c)

ιη

τ M

TEST PROGRAMM

TEMPERATURES CURVES

t (min.)

TESTS PROGRAM FIG.

l U . D EFLECTION CURVE

­274­

mimi

M

·

reinforcement symmetry division

arrangement

only one half in

subslices

are

indicated

in

figure

11.

Because

of

the

of the cross section has to be considered for the

and

only

one

half

of

the

length

of

the

beam

is

subdivided in 8 finite elements (see figure 12). Figures 13.a and b show the temperature increase in the steel profile. There

is

a

good agreement

between

theoretical

and experimental

results,

though the resultant emissivity factor of steel seems to have been chosen a little

low.

The

accuracy

of

the

numerical

results

is very

good for

the

temperature increase in the reinforcing bars (figure 13.c). In a simply supported composite beam submitted to a fire

test, the

steep thermal gradient on the cross section produces large deflections even at the

beginning

of

the

test

when

the

stiffness

properties

of

the

materials

remain unchanged. Figure 14 shows that the numerical procedure can simulate this

behaviour

(taking

into

account

the

underestimation

of

the

emissivity

factor). Some

numerical

problems

have

still

to be solved in order

that

the

simulation can be carried on for the whole fire test duration. This should be done in the very near future.

5. CONCLUSIONS. A numerical procedure for the analysis of the structural behaviour of steel and composite structures under fire conditions has been presented. It is based on the finite element method using beam elements with subdivision of the cross section in a rectangular mesh. The structure submitted to increasing temperatures is analyzed step-by-step using the NEWTON-RAPHSON process. A comparison between theoretical and experimental results has been made for a composite

beam.

experimental

There

results,

but

is

a

good

agreement

some parameters

have

between

theoretical

and

still to be calibrated and

some transformations have still to be made in order to ensure convergence when the structure is submitted to increasing temperatures. The development of this numerical tool will lead to an improvement of the

prediction

behaviour

of

of steel

fire

resistance

and

composite

intended to analyze the influence

and

to

a

better

understanding

of

constructions under fire conditions. It

the is

of several factors such as real fires and

-275-

special fire conditions, second order effects, creep and relaxation, thermal restraint and deflections. With this knowledge it is planned to show that very often the protection of these types of elements could be reduced, or that a given steel or composite structure behaves better under real fire conditions than assumed up to now.

R E F E R E N C E S (1)

ARNAULT, P., EHM, H., et KRUPPA, 3. Résistance au Feu des Poutres Mixtes (Isostatiques et Hyperstatiques). CTICM, Paris, février 1976.

(2)

FREY, F. L'Analyse Statique Non Linéaire des Structures par la Méthode des Eléments Finis et son Application à la Construction Métallique. Thèse de Doctorat, Laboratoire de Mécanique des Matériaux et de Statique des Constructions, Université de Liège, 197S.

(3)

KORDINA, K., WESCHE, J., WALTER, R., und HASS, R. Amtliche Materialprüfanstalt für das Bauwesen. T.U. Braunschweig Untersuchungsberichte und Prüfungszeugnisse Nr. 77150R, 80341, 80644, 831009, 831016, 831025, 831032.

(4)

SCHNOBRICH, W.C. Analysis of Reinforced Concrete Structures by using the Finite Element Method The Solution of Non Linear Equations. Corso di Perfezionamento per le Costruzioni in Cemento Armato, Politecnico di Milano, 1978.

(5)

DOTREPPE, 3.C. Méthodes Numériques pour la Simulation du Comportement au Feu des Structures en Acier et en Béton Armé. Thèse d'Agrégation de l'Enseignement Supérieur, Université de Liège, 1980.

(6)

MASAHIDE TOMII, and KEN3I SAKINO. Inelastic Behaviour of Concrete Filled Square Steel Tubular Beam-Columns. Proceedings of the USA-Japan Seminar on Composite Structures and Mixed Structural Systems, 1980.

-276-

(7)

DOTREPPE, 3.C., FRANSSEN, 3.M., et SCHLEICH, J.B.. Analyse de la Résistance au Feu des Structures en Acier et Mixtes Acier-Béton, Assistée par Ordinateur (REFAO/CAFIR). Recherche CCE 7210-SA/502, Rapports techniques N°l, 2 et 3, ARBED-Recherches, Luxembourg, 1982-198'».

(8)

HERSCHELMANN, F. Untersuchungen Ober konstruktive Massnahmen zur Verbesserung des Feuerwiderstandes von Stahl-Verbundträgern. Bericht, Institut für Baustoffe, Massivbau und Brandschutz, T.U. Braunschweig, April 19S2.

(9)

ANDERBERG, Y. Behaviour of Steel at High Temperatures. RILEM Committee 44-PHT, 1983.

(10)

CHARL1ER, R. Analyse de la Charge Critique d'une Colonne Mixte AF 30/120 par le Programme FLAMB 15. Rapport interne, Service de Mécanique des Structures, Université de Liège, juin 1983.

(11)

E.C.C.S. European Recommendations for the Fire Safety of Steel Structures. ECCS Technical Committee 3 - Fire Safety of Steel Structures, Elsevier, Amsterdam, 1983.

(12)

3UNGBLUTH, O., und HAHN, 3. Traglastenkatalog für ARBED Walzträgerbasis, 1983.

(13)

AF

30/120

-

Verbundstatzen

auf

KLINGSCH, W., und NOWAK, R. Verbundstützen - Interaktionsbeziehungen für Kaltbemessung. Forschungsbericht, Lehrstuhl für Baustofftechnologie und Brandschutz, Bergische Universität Wuppertal, 1983.

(14)

SCHLEICH, 3.B., LAHODA, E., LICKES, 3.P., and HUTMACHER, H. A New Technology in Fireproof Steel Construction. International Review ACIER-STAHL-STEEL, Nr.3, 1983.

-277-

(15)

SCHNEIDER, U. Behaviour oí 44-PHT, 1983.

(16)

Concrete

at

High

Temperatures.

RILE M Committee

SCHLEICH, J.B. Dimensionnement des Colonnes Mixtes. Conférence Internationale "Sécurité au Feu des Constructions en Acier : Conception Pratique", C.C.E., Luxembourg, avril 1984.

-278-

REQUIREMENTS OF FIRE RESISTANCE BASED ON ACTUAL FIRES (SWEDISH APPROACH)

O. PETTERSSON, Prof. Dr. Division of Building Fire Safety and Technology, Lund Institute of Technology, Sweden

Summary A rational, analytical approach to a fire engineering design of load bearing steel structures is described. The method of design is directly based on the natural compartment fire concept and on strictly defined functional requirements and performance criteria. The method is permitted to be generally applied in Sweden, as one alternative, since more than ten years. For facilitating the practical application, a comprehensive design basis has been worked out in the form of diagrams and tables for a direct determination of the maximum steel temperature during the relevant compartment fire and the corresponding design load bearing capacity of the fire exposed structure. The design basis is presented in a manual which has been given type approval for practical use by the National Swedish Board of Physical Planning and Building. The design procedure has recently been further developed in order to arrive at a design method in regard to fire exposure which is in principal agreement with modern loading and safety philosophy for the non-fire state.

-279-

1. INTRODUCTION In Sweden, an analytical design of fire exposed load bearing structures and partitions is officially approved for a general practical application, as one alternative, since more than 10 years (1). The design is directly based on the thermal characteristics of the fully developed compartment fire as a function of the fire load and the geometrical, ventilation and thermal properties of the compartment. For facilitating the practical application of the design method, diagrams and tables have been systematically produced and published in the form of manuals (2), (3), giving directly, on one hand, the temperature state of the fire exposed structure, on the other, a transfer of this information to the corresponding load bearing capacity. In its latest form, the design method is probability based. 2. LIMIT STATE CONDITION Generally, the design criterion in a structural fire design requires that no limit state is reached during the relevant fire exposure. Depending on the type of application, one, two or all of the following limit state conditions apply: * limit state with respect to load bearing capacity, * limit state with respect to insulation, * limit state with respect to integrity. For a load bearing structure, the design criterion implies that the minimum value of the load bearing capacity R(t) during the fire exposure shall meet the load effect on the structure S, i.e. min[R(t)] - S > 0

[2.1]

For a separating structure, the design criterion with respect to insulation reads T.. -max[T (t)] > 0 lim s

[2.2]

where T. . is the maximum temperature of the unexposed side of the structure, acceptable as concerns the requirement to prevent a fire spread from the fire compartment to an adjacent compartment. T (t) is the highest temperature on the unexposed side of the structure at time t of the fire process. The supplementary limit state condition regarding the integrity func-

-280-

tion has to be proved experimentally, when decisive. 3. PHYSICAL MODEL. FIRE EXPOSURE The physical model for the fire design is shown summarily in Fig. 1 for a load bearing structure. The design starts by a determination of the fire exposure, given by the gas temperature­time curve of the fully deve­ loped compartment fire. In the individual application, the fire exposure then can be obtained either by heat and mass balance calculations for the fire compartment or directly from a systematized design basis of the type exemplified by Fig. 2 (1). The fire load density q, the ventilation of the fire compartment expressed by the opening factor Α\ΛΓ/Α

and the thermal pro­

perties of the structures enclosing the compartment are the decisive fac­ tors.

FIRE LOAD DENSITY FIRE EXPOSURE FIRE COMPARTMENT

STRUCTURAL DA TA

TEMPERATURE STA TE

_·»

MINIMUM LOA D BEA RIN6 CAPACITY R m

LOAD EFFECT AT FIRE S

Figure 1 : Physical model for an analytical fire engineering design of load bearing structures, based directly on the exposure characteristics of the fully developed compartment fire

The gas temperature­time curves in Fig. 2 apply to a fire compartment with specified thermal properties of its surrounding structures, fire com­ partment type A. The surrounding material roughly corresponds to an average of brick, concrete and aerated concrete. Fire compartments with enclosing structures of deviating thermal properties can be transferred to fire com­ partment type A by using fictitious values of the fire load density q, and the opening factor (ΑνΊΓ/Α ) , according to the approximate formulae

­281­

QiSOO MJ/m1

Figure 2: Gas temperature­time curves for a fully developed compartment fire as a function of the fire load density q and the opening factor A\Zh"/A.. A is the total opening area, h is a weighted mean value of the height of the openings based on their size, and At is the total internal surrounding area of the compartment, including openings. Fire compartment type A (1)­(3)

qf ­ Kfq ; (Av/n"/At)f ­ KfAv/h"/At

[3.1]

In (1) ­ (3), the coefficient K. is given for different types of fire com­ partments defined by their surrounding structures.

The fire load density q is defined by the relationship -¿- Σμ m H A v v v

(MJ­m

[3.2]

)

where m » total mass of combustible material ν (kg), Η ­ net calorific value ° '

ν

of combustible material ν (MJ­kg

V

) and μ ­a fraction between 0 and 1,

giving the real degree of combustion for each individual component of the fire load. The gas temperature­time curves according to Fig. 2 are applicable to fire compartments of a size representative of dwellings, ordinary offices, schools, hospitals, hotels, and libraries. For fire compartments with a

­282­

very large volume - for instance, industrial buildings and sports halls the curves, and the corresponding heat and mass balance equations behind the curves, give an unsatisfactory description of the real fire exposure. At present, there is no

validated design basis available for the determi-

nation of the fire exposure in compartments with a very large volume. Returning to the physical model, as shown in Fig. 1, in the next step, the fire exposure is transferred analytically to transient temperature fields in the exposed structure and then a determination is carried out of the time variation of the load bearing capacity R(t). A comparison between the minimum value R of R(t) during the relevant m fire process and the load effect at fire S decides whether the structure can fulfil its required load bearing function or not during the fire, as specified by the limit state condition according to Eq. [2.1]. For a separating structure, the physical model gives the transient temperature state, defining the maximum value, max[T (t)], of the highest temperature on the unexposed side of the structure during the relevant fire exposure. The corresponding limit state condition follows Eq. [2.2], as concerns the required function of insulation. The limit state condition with respect to the integrity function has to be proved experimentally, when required. 4. PROBABILITY BASED DESIGN For the probabilistic model to be integrated with the physical model, different levels of ambition can be distinguished: * an accurate evaluation of the failure probability, using multi-dimensional integration or Monte Carlo simulation, * an approximate evaluation of the failure probability, based on first order reliability methods (FORM), and * a practical design format calculation, based on partial safety factors and taking into account characteristic values for action effects and response capacities. For practical purposes, an accurate evaluation of failure probability is not possible. Also, the FORM approximations are too cumbersome for everyday design and the more simplified practical design formats have to be used.

-283-

Fig. 3 illustrates a practical design format calculation for a fire exposed load bearing structure (4)­ (6). From the design fire load density q, and the geometrical, ventilation and thermal characteristics of the fire compartment, the design fire exposure is determined either by energy and mass balance calculations or from a systematized design basis. Together with the structural design data, the design thermal properties and the de­ sign mechanical strength of the structural materials, the design fire ex­ posure provides the design temperature state and the related design load bearing capacity R, for the lowest value of the load bearing capacity during the relevant fire process. τ FIRE EXTINGUISH­ MENT, F I R E F I G H ­ TING CH A R A CTE­ RISTICS

f\ 1

F I R E COMPA RTMENT CHARACTERISTICS

JESIGN F I R E EXPOSURE r­t

DESIGN MECHANICAL STRENGTH t

DESIGN THERMA L PROPERTIES

Τ

if V

1

1 DESIGN TEMPERATURE STATE

f DESIGN F I R E LOA O DENSITY

Η,,ΙΤΙ.Μ^Τ),..

DESIGN LOA O BEARING CAPACITY

FETTGFLDTU— FFECT AT FIRE „■S(G d .O d i

»

■«(»ο1·Η<Ι2···Ι 1

STRUCTURAL DESIGN DATA

Figure 3: Procedure for a practical design format calculation of a load bearing structure, exposed to a natural compartment fire

The design format condition to be proved is

R

[4.1]

d"Sd­°

V

where S, is the design load effect at fire. Depending on the type of prac­ tical application, the condition has to be verified for either the complete fire process or a limited part of it, determined by the time necessary for the fire brigade to attack the fire under the most severe conditions or by the design evacuation time for the building.

The probabilistic influences are considered by specifying characteris­ tic values and related partial safety factors for the fire load density,

­284­

such structural design data as imperfections, the thermal properties, the mechanical strength and the loading.

The functional requirements to be laid down for the fire design must be differentiated with respect to such aspects as the occupancy, the height and volume of the building, and the importance of the structure or struc­ tural member to the overall stability of the building. This can be done by, for instance, a system of safety classes with allocated failure probabili­ ties, affecting the design strength. The effect of the probability of occurrence of a postflashover compartment fire, the fire brigade actions and an installed fire extinguishment system, if any, can be accounted for principally in the same way. An alternative solution is to include these influences in the determination of the design fire load density and the de­ sign fire exposure, as indicated in Fig. 3. This latter way is chosen in the Swedish probability based design method by dividing the structures or structural members into categories with a related differentiation of the design fire load density and the length of the fire process, to be consi­ dered in the design. The presence of an approved sprinkler system then is taken into account in a very simplified way by a transfer of the structure or structural member to the next lower category.

5. TEMPERATURE DISTRIBUTION IN STRUCTURAL STEEL ELEMENTS AT FIRE EXPOSURE For a fire exposed, uninsulated steel structure, the energy balance equation gives the following formula for a determination of the steel tem­ perature­time curve Τ ­ t (Fig. 4)

ΔΤ

F ­ — 5 — . ­i ( T ­ T ) At s p e V t s s ps s

(°C)

[5.1]

where ΔΤ

­ change of steel temperature ( C) during time step At (s),

α

­ coefficient of heat transfer at fire exposed surface of structure (Wm"2­0C_1),

ρ s c F 8

V T

_3 ­ density of steel material (7850 kg­m ) , . . ■ specific heat of steel material (J­kg · C ) , » fire exposed surface of steel structure per unit length (m), 2 » volume of steel structure per unit length (m ) , ­ gas temperature ( C) within fire compartment at time t (s).

­285­

Figure 4: Fire exposed, uninsulated steel structure. T t » gas temperature within fire compartment, T 8 ­ steel temperature at time t Eq. [5.1] presupposes that the steel temperature T_ is uniformly dis­ tributed over the cross section of the structure at any time t.

The coefficient of heat transfer α can be calculated from the approxi­ mate formula

α ­ 23 +

5.77e_ r Τ +273 , Τ +273 4 ν ) 4 ­ (ν ­ 8 100 ' 100

­Ττs LΙ| χ Τr ­ t t

­) ]

(W.m­2.V1)

[5.2]

8 *■

giving an accuracy which is sufficent for ordinary practical purposes, ε is the resultant emissivity which for practical applications can be chosen according to the following table, giving values which are generally on the safe side.

1. Column, fire exposed on all sides 2. Column, outside a facade 3. Floor structure, composed of steel beams with a concrete slab on the lower flange of the beams 4. Steel beams with a floor slab on the upper flange of the beams 4a Beams of I cross section with width/height ^ 0.5 4b Beams of I cross section with width/height < 0.5 4c Beams of box cross section and trusses

More accurate values of the resultant emissivity ε

can be determined

for alternative 4 ­ steel beams with a floor slab, supported on the upper flange of the beams ­ from the diagrams of Fig. 5 and 6, applicable to floor structures with the flames completely below the steel beams and reach­

­286­

ing the slab, respectively. For the emissivity of the flames e , the value 0.85 is to be inserted, if not any other value can be proved to be more correct.

_,—'/­

0.5­

0.5

1.0 "/n

0Λ

1.0

B h

/

Es-Ebj-0.1

\—r

XT

X Ceiling or flam«»

«

e

Figure 5: Resultant emissivity ε for steel beams with a floor slab, sup­ ported on the upper flange of the beams. Flames completely below the steel beams. EJ,J ­ emissivity of the slab, ε 8 » emissivity of the steel beams, ct ■ ­ emissivity of the flames. I cross section, box cross section

» » » ·

Ì17)

■wffwm

%¡)í

it 1

' ' ι ' ' ■ ' ι .». 0.5 '.o lyh

Figure 6: Resultant emissivity ε Γ for steel beams of I cross section with a floor slab, supported on the upper flange of the beams. Flames reaching the slab. et ­ emissivity of the flames

­287­

At a given gas temperature­time curve T ­t of the fire compartment, the steel temperature T

can be directly calculated from Eqs. [5.1] and

[5.2] with regard taken to the temperature dependence of c and a. Such ps computations have been carried out in a systematized way, giving design tables as published in (2), (3). From such tables, the maximum steel tempe­ rature T

during a complete compartment fire can be determined directly

as a function of the fictitious fire load density qf, the fictitious open­ ing factor (A \Zh~/A ),, the F /V

ratio and the resultant emis s ivi ty ε . The

values are connected to gas temperature characteristics according to Fig. 2. Similarly, for a fire exposed, insulated steel structure, a simplified energy balance equation gives the following formula for a direct determina­ tion of the steel temperature­time curve Τ ­ t (Fig. 7)

nT

s ­
1

V

β ps s

(T

t­ V "

(

[5.3]

°C)

with the additional quantities A. ­ interior jacket surface area of insulation per unit length (m), d. » thickness of insulation (m), λ. « thermal conductivity of insulating material (W­m

· C ).

—Π

Figure 7: Fire exposed, insulated steel structure. T t » gas temperature with­ in fire compartment, T 8 ­ steel temperature at time t Eq. [5.3] presupposes that the steel temperature T

is uniformly dis­

tributed over the cross section of the structure at any time t, that the temperature gradient is linear and the heating contribution negligible for the insulation, and that the heat transfer is one­dimensional.

­288­

Computations, originating from Eqs. [5.2] and [5.3], provide a system­ atized design basis for a practical fire design. Such a design basis is pub­ lished in (2), (3) in the form of tables, giving the maximum steel tempera­ ture Τ

during a complete compartment fire for varying values of the

fictitious fire load density q­, the fictitious opening factor (Αν/ΓΓ/Α ) , , the structural parameter A./V , and the insulation parameter d./λ.. The values are connected to gas temperature characteristics according to Fig. 2.

For a specific insulating material, systematized design diagrams or tables can be computed very accurately with regard to the temperature de­ pendence of the thermal properties of the steel as well as the insulating material. The influence of an initial moisture content and of a disinte­ gration of the insulating material can be considered, too. Practically, such a determination can be carried out over a numerical data processing by com­ puters on the basis of a finite difference or a finite element method. A great number of design tables, computed according to such an accurate pro­ cedure, are presented in (2).

6. LOAD BEARING CAPACITY OF STEEL STRUCTURES AT FIRE EXPOSURE By applying the design tables, referred to in the previous chapter, the maximum steel temperature Τ

can be determined comparatively quick­

ly for an uninsulated or insulated steel structure, exposed to a compart­ ment fire with gas temperature­time characteristics according to Fig. 2. The corresponding design load­bearing capacity of the structure then is obtained by design diagrams of the type exemplified in Fig. 8 and 9 (2), (3).

Fig. 8 and 9 give the design load bearing capacity (Μ

,Ρ

,q

) of

fire exposed beams of constant I cross section at different types of load­ ing and support conditions, as a function of the steel beam temperature T . The design curves in Fig. 8 apply to a slow rate of heating ­ assumed to be I* C­min

, followed by a cooling with a rate of 1.33 C­min

­ and Fig. 9

gives the correction Δβ of the load bearing capacity coefficient 6 due to a more rapid rate of heating. In the formulae for the load bearing capacity σ 'yield stress of steel material at room temperature (MPa), L ■ span of 8 . 3 beam (m), W ­ elastic modulus of beam cross section (m ) . The design curves in Fig. 8 and 9 have been determined on the basis of

­289­

©

Θ O-—5) 1— t

ιι ι ιι ιι i m q

® tUl

1

„8o.W

U!L· ,*<>!.*

Mcr = ßo s W

ω airrrnxDq Ι­ ο«

..'_....ι

©

© ΙΟ.Ί11.Ι.1ΤΠΊ

rr

L­­.1 12 ot W

L!

©

'«■Ρ i■■ ο

ΙϋΟ

200

300

ί­

(00

Figure 8: Coefficient β for determination of critical load (Μ,.Γ, P c r , q c r ) for fire exposed beams of I cross sec­ tion at different types of loading and support conditions, as a function of the steel beam temperature T s . The curves have been calculated for a slow rate of heating of 4 0 C ­ m i n _ 1 and a subsequent cooling, assumed to be one third of the rate of heating ( 2 ) , (3)

0= 100°C min·'

c= 20°C minr1

¿50

500

Figure 9: Increase Δβ of coefficient Β, determined according to Fig. 8, for a rate of heating a 2 4°C­min~', as a function of the steel beam tempera­ ture T s (2), (3)

the deformation curve of the fire exposed beams calculated by an analytical model, presented in (7), which takes into account the softly rounded shape of the stress­strain curve of steel at elevated temperatures as well as the influence of creep strain, noticeable at temperatures in excess of about 450°C.

For a structural fire design of columns, unrestrained or partly re­ strained to a longitudinal expansion during the fire exposure, reference is made to (2).

7. CONCLUDING REMA RKS Compared with the conventional fire engineering design, based on clas­ sification and results of standard fire resistance tests, the presented analytical design procedure has a more logical structure, based on well­de­ fined functional requirements and performance criteria. Of the ensuing ad­ vantages, the following are seen to be the main ones: 1. More consistent safety levels. 2. Better

economy. The cost of structural fire protection is, as a rule,

hard to itemize and the cost­saving consequences have been quantified only in a few cases. Rough estimates indicate that while the cost for conventional structural fire protection may exceed 30 per cent of the cost for the steel frame material, the corresponding percentage may be as low as 10 with the design procedure based on analytical modelling, see

­291­

Fig. 10. This figure is based on the assumption that the advantages are fully exploited of integrating the design of the structural steel fire protection into the overall design process (inner and outer walls are used as fire protection whenever possible, concrete floor slabs are placed on the lower flange of the girders, inherently providing a small­ er area to insulate, etc).

Finally, it is recognized that the design system presented is not homo­ geneous due to a varying level of the present basis of knowledge for the different design steps. Naturally, this can be put forward as a criticism of the system. However, such a remark is not essential. Instead, this fact ought to be used as an important guide on how to systematize a future re­ search work for enabling a successive improvement of the system.

Fire protection according to integrated, rational design method

Fire protection according to conventional standard design method

Figure 10: Costs for fire protection

REFERENCES (1)

National Swedish Board of Physical Planning and Building, "Brandtek­ nisk dimensionering (Fire Engineering Design). Comments on SBN (Swe­ dish Building Code)", No. 1976:1.

(2)

Pettersson, 0., Magnusson, S.E., and Thor, J., "Fire Engineering De­ sign of Steel Structures", Swedish Institute of Steel Construction, Publication No. 50, Stockholm, 1976 (Swedish edition 1974).

(3)

Pettersson, 0., and ödeen, Κ., "Brandteknisk dimensionering av bygg­ nadskonstruktioner ­ principer, underlag, exempel (Fire Engineering Design of Building Structures ­ Principles, Design Basis, Examples)", Liber förlag, Stockholm, 1978.

(4)

Magnusson, S.E., and Pettersson, 0., "Rational Design Methodology for

­292­

Fire Exposed Load Bearing Structures", Fire Safety Journal 3, 1980/81. (5)

Pettersson, 0., "Reliability Based Design of Fire Exposed Concrete Structures", Contemporary European Concrete Research, Stockholm 1981.

(6)

CIB W14, "A Conceptual Approach Towards o Probability Based Design Guide on Structural Fire Safety". Report of CIB W14 Workshop "Structural Fire Safety", Fire Safety Journal 6, 1983.

(7)

Thor, J., "Deformations and Critical Loads of Steel Beams Under Fire Exposure Conditions", National Swedish Building Research, Document D16:1973, Stockholm.

-293-

A PROBABILITY BASED FIRE SAFETY CONCEPT M. KERSKEN-BRADLEY, Dr.-Ing. Institut für Bautechnik, Berlin

Summary A probability based safety concept provides the framework for the model code/design guide on structural fire design which is under preparation in the CIB. It has been successfully applied to a standard for the assessment of industrial buildings in the Federal Republic of Germany (Vornorm DIN 18230, 1982). The following contribution briefly outlines the main components of the safety concept and identifies those features which are of special interest for steel constructions.

-294-

1.

INTRODUCTION

The traditional procedure in structural fire design is based on a classification system: - On the one hand a required time of fire resistance in terms of a standard fire exposure is stipulated in building regulations or codes - usually expressed as required fire resistance classes in multiples of 30 minutes. - On the other hand structural elements are graded by determining their fire resistance time in standard test conditions - which is usually expressed by reference to fire resistance classes in mupltiples of 30 minutes. This classification system is very convenient, especially if well-prepared catalogues of graded structural members are available. It may provide a reasonable design for buildings and occupancies for which sufficient experience concerning the fire risk has been gained and for types of construction which are not very sensitive to rough grading criteria. But the design may become questionable from an economic or safety point of view in cases where the heat exposure, the structural response, the associated uncertainties and fire risks differ substantially from the average situation covered by the regulations. Moreover, types of construction which are sensitive to rough grading criteria suffer disadvantages in competition with other building materials (clearly steel and timber vs. concrete)which in many cases may not be the actual intention of the regulatory body with respect to their fire safety objectives. However, in the specification of the required fire resistance physical aspects, safety considerations and reserviceability aspects are lumped together and thus provide no guidance as to the actually intended level of safety. This may give rise to difficulties in the assessment of particular projects and may likewise impede any attempt to harmonize principles and rules in fire design. A probability based safety concept may be considered as a useful aid to cope with these difficulties without necessarily complicating the design procedure. Generally, probabilistic concepts are intended for deriving design rules for practical

-295-

application, thus the designer does not have to be concerned about the détails of probabilistic modelling. But as for the application of any design rules, some background knowledge on the implication and limits of the rules employed should be available. Alternatively, a direct probabilistic analysis on the basis of such a concept may be attempted for the assessment of the (structural) fire risk of a particular project, but this approach will definitely be confined to exceptional cases. Since this safety concept has been introduced in various publications (e.g. (1) to (4)) and will also be issued as a CIB Model Code (Design Guide) in the near future, this contribution only identifies the main components and emphasizes those features of the concept which are of special interest for steel constructions. 2. 2.1

COMPONENTS OF A PROBABILITY BASED SAFETY CONCEPT Specification of Objectives

A clear specification of the general fire safety objectives is a prerequisite for the development of a consistent safety concept. The objectives generally comprise the limitation of - life risk - neighbouring property risk and may also include a limitation of the - directly exposed property risk (building and/or contents). Whilst the public concern with regard to the first and second objective is straightforward, the competence for the third objective - in particular as concerns the protection of buildings of no special cultural/societal significance - is usually not clearly established and thus gives rise to problems. 2.2

Measures for Fire Risk Control

An important feature of the safety concept is that the various measures for fire risk control are considered with regard to their risk-reducing contribution. They include

-296-

. reducing the frequency of fire occurrence . control of fire (smoke and flames) at an initial stage . ensuring a safe evacuation of people . providing for safe and efficient operation conditions for fire brigades . preventing fire spread beyond a certain area . avoiding structural failure or limiting structural damage. Structural fire design ­ to which this concept refers ­ is only concerned with the dimensioning and detailing of structures and their individual members. These provisions will merely contrib­ ute to the prevention of fire spread through structural barri­ ers and to the avoidance or limitation of structural failure or damage. It is important to note that design only refers to fires which fail to be controlled at an initial stage and which are severe enough to cause structural damage. 2.3

Levels of Structural Fire Safety ­ Reliability Requirements ­

The level of structural fire safety (structural reliability) to be provided by design should be governed by i. the possible hazards (consequences) in the case of fire ­ depending on the type of building, the location and occupancy of the building and fire compartment ii. the risk­reducing contribution of dimensioning and detail­ ing the structure iii. the estimated frequency of severe fires ­ depending on the occupancy and size of the fire compartment iv. the possibly reduced frequency of severe fires due to special measures of fire risk control as are detecting and alarm devices,sprinkler systems, plant fire brigades. A useful presentation of reliability requirements is by ref­ erence to failure probabilities with regard to the attainment of specified limit states (cf. 2 . 5 ) . Items i. and ii. can be quantified by tolerable failure probabilities Ρ(fail), applying to a specified reference period. Ρ (fail) will generally range from 10 to 10 depending on the number of people endangered

­297­

by structural failure and possibly on economic losses due to failure. A ny numerical value for Ρ (fail), however, requires checking on a national basis ­ presumably by calibration to generally acknowledged fire design solutions. Items iii. and iv. refer to the probability of occurrence of severe fires which can be modelled as Ρ(severe fire) = Ρ (fire) · Ρ(severe fire Ifire)

(1)

wherein the various terms are briefly discussed in sec. 2.7. The failure probability decisive for design is the corre­ sponding probability conditioned by the occurrence of a severe fire Ρ(fail¡severe fire) = Ρ(fail)/P(severe fire)

(2)

Egu. (2) consistently reflects the fact that, if the fire hazard and/or the risk reducing contribution of dimensioning is con­ sidered low and/or the occurrence of severe fires is sufficient­ ly rare, then a fairly low level of structural reliability to be provided by design will suffice. It follows that in various cases structural requirements with regard to member design may even be dispensible. 2.4

Structural Requirements and Criteria

A consistent safety concept calls for requirements which are formulated in a functional manner, i.e. by reference to the ex­ pected performance of structures in the case of fire. Thus, structures and structural members may be required to adequately ­ sustain all relevant actions ­ perform as fire barriers during the relevant fire exposure. In addition ­ further limitations on structural damage after fire exposure may be stipulated ­ depending on the objec­ tives. For the purpose of design verification the requirements are expressed in terms of limit states, e.g. limit states with respect to ­ the load bearing capacity

­298­

(strength, stability, ductility) - the separating function (thermal insulation and integrity) possibly supplemented by limit states with respect to - reserviceability or repairability using appropriate models for describing the heat exposure and the structural response. The issue of adequately fulfilling the requirements is followed by designing for specified levels of reliability (cf. 2.3). Design may refer to the entire fire process or only to a limited part of it. Moreover, design may refer - either to conditions specific for a given project and fire compartment (individual assessment) - or to conditions representative for certain types of buildings and occupancy of fire compartments. 2.5

Models - Limit State Condition A limit state is generally expressed as a function of vari-

ablesX, denoted as basic variables g ( X r X 2 , ..., X n ) = 0

(3)

In fire design the following variables are taken into account, either implicitly or explicitly- depending on the method of assessment: - within the heat exposure model: . fire load density and . combustion behaviour of the fire load . geometrical parameters and . ventilation characteristics and . thermal properties of the fire compartment - within the structural response model . geometrical parameters of the structure and . the structural system . thermal and mechanical properties of the structural components . loads.

-299-

The variables ­ in particular within the heat exposure model ­ may be definea to describe the conditions specific for a particular fire compartment or representative for certain types of fire compartments. Egu. (3) can be expressed in the time domain, e.g. in terms of the fire resistance time t­ = t,(X) and the equivalent time of fire exposure tfi = t (X) (cf.(1, 2, 4)) tf ­ te = 0

(4a)

or in the temperature domain, e.g. in terms of a critical or ultimate temperature T = Τ (X) and a resulting maximum temper­ ature due to the relevant fire process Τ = max|T(t, X) Τ

­ Τ = 0

(4b)

or in the mechanical strength domain in terms of the minimum resistance capacity during the relevant fire process R = min JR(t, X)I and the corresponding action effects S = S(Χ) R ­ S = 0

(4c)

The relevant domain depends on the assessment method employed which in turn is governed by the need or option to include ex­ perimental models or not. For the current available assessment methods reference is made to previous contributions to this conference (e.g. (5), (6)). 2.6

Evaluation of Failure Probabilities

The failure probability introduced in sec. 2.3 is defined as the probability that a specified limit state is exceeded during the relevant fire process. For evaluating this probability the basic variables are explicitly treated as random variables by describing them in terms of distribution functions (type of function and distribution parameters) possibly by reference to stochastic processes. The probability to be calculated is P(fail|severe fire) = P(g(X) * 0)

(5)

and can be evaluated e.g. by first­order­reliability­methods (7), introducing the so­called safety index B. For particular limit state conditions and types of distribution functions

­300­

evaluation of the probability of equ. (5) requires only few elementary steps of calculation. Since generally the data base for the various variables is more than modest, specification of the distribution functions requires assistance by engineering judgement. Moreover, the models for describing limit state conditions in fire exposure are associated with considerable uncertainties, even if fairly sophisticated models are employed. These uncertainties may be considered by increasing the total variance of the state func­ tion (e.g. by increased coefficients of variation of the basic variables) ­ if not taken into account in the specification/ calibration of the tolerable failure probability. It is emphasized that an evaluation of failure probabilities in only required for assisting decisions on safety factors within code or standard committees and is not intended as an aid for practical dimensioning ­ apart from very exceptional cases where a probabilistic analysis may be considered neces­ sary or worthwhile. For the purpose of deriving safety factors, the resistance capacity in the limit state condition is determined such that the failure probability according to equ. (5) meets the toler­ able probability identified by equ. (2). This design solution thus corresponds to a specific set of design values for the basic variables considered. Repeating the procedure for various applications (e.g. different structural components, modes of loading, etc.) may render different sets of design values. It is then the task of the code or standard committee to specify characteristic (nominal) values and appropriate safety factors for a certain domain of application. 2.7

Models ­ Fire Frequency

Modelling the fire frequency within a fire compartment com­ prises models describing ­ the probability of fire occurrence, Ρ(fire), referring to fires of arbitrary extent ­ the probability that a(n) (initial) fire may develop into a

­301­

fire severe enough to cause structural damage, Ρ(severe fire I fire). The first probability primarily depends on the occupancy and size of the fire compartment. As concerns the size effect, theoretical modelling suggests a proportional increase of fire occurrences with increasing area which according to (8, 9) holds for office buildings, but may be too conservative for other occupancies. In (9) both terms of the probability, p(fire) = ρ · f(A)

(6a)

i.e. not only the probability ρ per unit floor area but also the functional dependency on the floor area f(A) are found to be different for different occupancies. Further investigations, in particular for non­industrial occupancies, may be necessary to support the presently available data base. The probability that, given an initial fire, a severe fire may develop, is for simplicity (and in accordance with the pre­ vailing model assumption for the heat exposure) also referred to as flash­over­probability. In the first instance this prob­ ability depends on the physical conditions in the fire compart­ ment (type and distribution of fire loads, size of fire com­ partment, ventilation conditions); these influences, however, are not yet sufficiently assessable to allow practical conclu­ sions. In addition, this probability is governed by the effi­ ciency of fire fighting at an early stage, which in turn de­ pends e.g. on the presence of detecting and alarm devices, sprinkler systems and on the possible employment of special plant fire brigades. A fairly extensive data base is available for sprinkler systems indicating a reduced flash­over­probabil­ ity up to two orders of magnitude (evaluated in terms of a successful or not successful operation of the system (10)). The data base with regard to the other measures is rather poor and assessment is basically by extrapolation from the sprinkler efficiency. Up to a certain extent the simultaneous employment of various measures can be considered by Ρ(severe fire fire) = p. · p_ ■ p, ...

­302­

(6b)

wherein the various probabilities p, describe the reduced flash­over­probability on behalf of measure i. However, due to a certain dependency among the effectiveness of the various measures, there are some limitations to egu. (6b). In (1 to 4) numerical values for the different probability terms are sug­ gested based on available data and supplemented by judgement. As more information becomes available, these figures may be up­ dated. Concluding, the probability for a severe fire ­ governing the reliability to be provided by design via equ. (2) ­ can be assessed by ρ (severe fire) = ρ · f (A) · p. · p_ · ρ, ...

2.8

(7)

Practical Application

Practical application is facilitated by the specification of (partial) safety factors ¡for the method of assessment and limit state considered. For convenience, safety factors may be established for an average level of reliability, i.e. average safety requirements and average frequency of fire. Adaption to different levels of reliability can be considered by differen­ tiation factors Τ

tn = r n1 ■ ί η2 wherein 0 n1 adapts safety differing from average in ţ__ adapts safety differing from average in (As an example: a reduced ­ 10~ 2 ­ results in T n 2 *

(β)

factors to levels of reliability view of a safety differentiation and factors to levels of reliability view of different fire frequencies. frequency of two orders of magnitude 0.6.)

The German Standard (Vornorm) DIN 18230 applying to the assessment of industrial buildings is an example for a prac­ tical application of the safety concept. It uses the equivalent time of fire exposure (t ) as an improved assessment measure for fire compartments with regard to the fire resistance time of the structure to be provided. In this interpretation t is applied independent of the type of material and construction

­303­

and for all limit states. It results In the simple design rule: requ. t f = t e · ţ with wherein

t g. : c :

w tf

Y

■ ·ζη

= c · w · q­ fire load density in |MJ/m{| conversion factor accounting for the

thermal properties of the fire compartment enclosure : ventilation factor : fire resistance which may be determined by testing, by analytical evaluation ­ or by :

ţ" :

reference to catalogues safety factor accounting for average reliability requirements differentiation factor accounting for different safety classes and active protection measures.

3.

EVALUATION AND DISCUSSION

Since steel constructions are rather sensitive to rough grading criteria, the benefit of a differentiated assessment is straightforward. However, it has to be clearly stated that at present it is difficult to pursue such an assessment within the framework of the existing building codes and regulations in the majority of countries. An important assessment calls for modi­ fied structural fire safety requirements ­ comparable to re­ quirements with respect to other accidental hazards (earth­ quake, impact). The most interesting feature of the safety concept is pre­ sumably the possibility for a reduced fire resistance in view of operational or active measures for fire risk control. Pre­ suming that the need for a certain fire resistance clearly re­ sults from fire safety objectives of public concern, the antic­ ipated trade­off has to be guided by some limitations: 1. The long­term efficiency and reliability of active protection measures has to be ensured, in principle by adequate maintenance and inspection of alarm and detecting

­304­

2.

3.

4.

5.

devices, sprinkler systems and of equipment and force of private fire brigades. An economic evaluation should consider building and installation costs including maintenance, inspection and operation costs throughout the intended service life of the structure - but also the expected reduction of losses due to a timely fire control. The possibly increased risk if fire occurs in conjunction with or as a consequence of other hazards (explosions, warfare, earthquake) has to be acknowledged and accepted. Some agreement with the public fire brigades as to the extent of " their assignment if, nevertheless, a severe fire occurs, may be necessary. It may be sensible to exclude particular fire barriers from trade-off (applying an occupancy-independent design concept).

The other important aspect for steel construction is that the safety concept - in its limit state formulation - allows for calculation models as well as experimental models. Thus, for assessment methods referring to the notion of a fire resistance time, this property may be determined analytically (cf. the European Recommendations) or by testing (cf. ISO 834) - including possible reference to catalogues. By physical and statistical considerations compatibility between either procedure with regard to the level of structural reliability can easily be established. 4.

REFERENCES

(1)

Bub, H. et al., Grundlagen zur Festlegung von Sicherheitsanforderungen im baulichen Brandschutz, Beuth Verlag Berlin-Köln, 1979 (new edition in preparation)

(2)

Bub, H., Hosser, D., Kersken-Bradley, M., Schneider, U. Eine Auslegungssystematik fUr den baulichen Brandschutz, Braba Heft 4, Erich Schmidt Verlag, Berlin 1982

-305-

(3)

Kersken-Bradley, M. A Safety Concept for Structural Fire Design, vfdb. Proceedings of the 6th International Fire Protection Seminar, Karlsruhe, 1982

(4)

A Conceptual Approach towards a Probability Based Design Guide on Structural Fire Safety, CIB W14 Workshop "Structural Fire Safety", Fire Safety Journal, Elsevier Sequoia S.A., 1983

(5)

Witteveen, J. Trends in Design Methods for Structural Fire Safety - Session I (1.2)

(6)

Pettersson, 0. Requirements of Fire Resistance based on Actual Fires - Session IV (4.3)

(7)

Rackwitz, R., FieBler B. Structural Reliability under Combined Random Sequences , Comp. & Structures 9, 1978, pp. 484-494

(8)

Wiggs, R. BOMA International Office Building Fire Survey, Skyscraper Management, 58 (6), 1973

(9)

Rutstein, R., Clarke, M. The Probability of Fire in Different Sectors of Industry, Fire Surveyor (Feb.), 1979, pp. 20-23

(10)

Verband der Sachversicherer, Jahresbericht 1979/80, Abteilung Schadenverhütung und Technik.

-306-

REPORT ON SESSION III : FUTURE PROSPECTS Chairman : P. BORCHGRAEVE Reporter : L. TWILT

From the illustrations given by Mr. Jeanes, it follows that there is an excellent agreement between experimental data and calculation results for beams and slabs. Speaker asks whether the computation model also allows to assess, in an operational way, the instability phenomena, such as buckling of columns or frame instability. The question is asked, because the structural instability under fire conditions, for example influenced by axial restraint, is a matter of discussion in some European countries at the moment. D.C. Jeanes (réf.: response to Mr. Twilt) The computer model, does not recognize the actual performance of columns. It only recognizes the effect of columns as far as they influence the performance of the floor system assembly at the point of contact. There are certain economic reasons for this limitation. The requirements to protect the floor sytem assembly are by far dominant. Columns ineventably are to be covered up for esthetic reasons etc. The additional efforts to arrive at the required fire resistance are therefore - with regard to the situation in the US - not significant. As far as extrapolating or perhaps extending the used element model contained in the FASBUS programme is concerned, there is however the possibility to use it as a tool for assessing the stability of columns. The most direct way is the use of the knowledge of the displacement of the assembly at the place of column location and using that displacement for a separate analysis to determine instability of the column. If one wants to look at the column specificly, this can be evaluated in a somewhat approximate way, using the beam element, applying an axial load to that element and defining the temperature profile accross the element. It is noted, however, that the beam element is speciflcly designed to model beams with gravity loads and assumes laterally fixed conditions. One could use it then, with that consideration in mind, to analyse stability over the weak axis. J. Kruppa, CTICM, France (réf.: general remark regarding presentation by Dr. Kersken-Bradley) Speaker stresses the importance of the approach presented by Dr. Kersken-Bradley. The approach is of vital interest for consulting offices which are dealing with the whole range of metalic constructions, since it is much closer to reality than the structural fire safety concepts commonly used today.

-307-

W. Klingeeh, University of Wuppertal, BDR (réf.: general remark on practical impact congress) Speaker refers to the information presented in the course of the congress which, as he sees it, clearly indicates the significant progress made during the last years trying to increase the safety levels and, at the same time, ensuring economic constructions. Speaker would like to hear whether insurance will accept the new instruments which are put foreward and asks whether it can therefore be expected to see some changes in the fire insurance premium. Mr. Kersken-Bradley, Inst, für Bautechnik, BRD (réf.: intervention by Prof. Klingsch) Although not a representative of an insurance company, speaker briefly comments on the intervention by Prof. Klingsch, by noting that insurance risk is only to a certain extent made up by the structural risk. If, for example, from a point of view of structural fire safety a fairly high factor for sprinklers can be introduced, insurance companies on the other hand may argue that sprinklers also cause damage, especially to the contents of a building. So, as far as the insurance of contents is concerned, the advantage of a sprinkler may not be quite as large as it is for the structural reliability. However, this only will be a question of having different factors. The basic approach, in speakers opinion, should be the same. J.P. Favre, Gebäudeversicherung des Kantons Bern, Switzerland (réf.: intervention by Prof. Klingsch) Speaker is of the opinion that the probabilistic approach, presented by Dr. Kersken-Bradley is a correct approach and is bound to eventually being used for the calculation of the premium and the risk. Be points out however that this opinion does not necessarily hold for the whole range of insurance, since speakers professional interest is limited to building insurance, within a monopolitic company. As far as the competition amongst insurance companies is concerned, it is very difficult to link directly between the risk certain buildings pose and a consensus between buildings cathegories. A short term solution, which would allow to take into account the new approach (with a better assessment of the risk and thus also of the level of the premium), is therefore, at the moment, very difficult to envisage. Such a solution will take at least, say, 10 years. On the other hand, it is noted that in certain kantons in Switzerland a system of risk assessment is used which is basicly not too far from the one presented by Dr. Kersken-Bradley. This situation however is typical for Switzerland and by no means representative for Europe.

-308-

Unindentifled speaker (réf.: intervention by Mr. Favre) Speaker is of the opinion that one has to look at the insurance companies as suppliers and that therefore their conditions ought to be discussed step by step. It is important, in this context, to stop the racism, racism which is often shown when insurance companies are confronted with steel constructions. A rational method for risk assessment would be of vital importance in such a discussion. A. Lickes, Offizier Kommandant Feuerwehr Stadt Luxembourg (réf.: presentation by Dr. Kersken-Bradley) Speaker regrets that the stand point of the fire brigade is poorly represented during this conference and advises that, at forth coming congresses of this type, fire brigade officials are invited to give their point of view as a speaker. Further to the probabilistic approach presented by Dr. Kersken-Bradley, it is noted that, in the opinion of the speaker, this cannot be accepted. It is not only a question of insurance policy, it covers also the safety of the people, which is to be guaranteed by the fire brigade. In speakers opinion emphasis should be on prevention. This means, for example, that reduction of fire resistance, when a sprinkler or an alarm system is installated, is not accepted. There should be a certain safety level with respect to fire resistance, lrrispective of the other applied fire safety measures, which are to be considered as additional. An exception to this rule might be the use of fire (smoke) ventilation. Speaker is prepared to give his point of view in writing and to communicate this to the organizers of this conference. Mr. Kerken-Bradley, Int. für Bautechnik, BRD (réf.: intervention by Mr. Lickes) In response to the remarks by Mr. Lickes, speaker draws the attention to the following: - DIN 18230, which provides a probabilistic approach similar to the one presented this afternoon, is prepared in close cooperation with the various parties involved, amongst which also the fire brigade. - Risk analysis has become a common approach now-a-days, the main feature being that not only the hazard or the maximum possible outcome is decisive, but also the probability that is associated to this event. Both aspects are equally relevant. - Approaches of the kind, presented ealier by the speaker, may Indeed require a revision of the present fire brigade rules or regulations, which in some countries require that the fire brigade should go into a building regardless what the state of the building still is. Trade off between passive and active fire prevention measures would have, of course, an Impact on the way of fire fighting in the future.

-309-

J. Roret, Syndicat de la Construction Métallique, France (réf.: intervention by Mr. Lickes) The statement is made that, so far, no cases are known in which people died as a direct result of the collapse of a steel structure in fire. M. Law, Ove Arup Partnership, UK (réf.: intervention by Mr. Lickes) The risk analysis, described by Dr. Kersken-Bradley, is Just a way of helping to make the best decision we can, e.g. on the way we tackle safety in buildings, the way we spend our money, and the return we get for it. The fire brigades themselves do not have unlimited resources. How do they decide how many firemen they send to the fire, what type of hoses they need, etc., etc. Anyway, they have to make decisions and they have to say what is reasonable under given circumstances. And that is what the approach, described by Dr. Kersken-Bradley intends to be: a way of trying to measure what we are doing. J. Kruppa, CTICM, France (réf.: intervention by Mr. Lickes) It is stated by the speaker that the aim of a probabilistic analysis is not to have less safety than we have got so far. On the contrary, we want to make sure that we have a sufficient and - perhaps - a more balanced degree of safety.

-310-

CLOSING

SESSION

Conference conclusions

-311-

CONFERENCE CONCLUSIONS. P. BORCHGRAEVE. Director, Belge—Luxembourg Steel Information Centre, Brussels Chairman of the Programme Committee.

Ladies and Gentlemen,

It is my duty, as president of

the programme committee for this international conference, to make a preliminary rapid resume of the proceedings. I would like first of all to stress that the organisation and success of such a conference depends on the active collaboration of a number of organisations and individuals, who have provided their ability and dynamism, and also their capacity to take initiatives, their capacity to understand and listen, their good sense and their capacity to lead. The planning, preparation and organisation of this conference were made possible by the joint efforts of several parties: l.The Commission of the European Communities, -via the division in charge of ECSC-Steel technical research, under the auspices of the General Direction XII "Science, research and Development": -via the Scientific and Technical Communication Division under General Direction XIII "Market of Information and Innovation"; -via the Luxembourg services charged with the material organisation of the conference. 2. The European convention for Metallic Construction, and in particular its Technical Committee No. 3, charged with the problem of fire safety.

-312-

3. The Steel Information Centres of the countries of the Community, metallurgical organisations of their countries authorised to provide information on and to promote the use of ferrous products. The similarity of objectives of all these parties and the complementary nature of their respective vocations have found their dynamic expression in the Programme committee. You will allow me also to thank those responsible for services in the Commission in Brusselis and Luxembourg who have been charged with the organisation, and the people who have assisted them, such as the members of the Programme Committee in charge of the design and., continuityof the programme. It is also appropriate to thank the experts from the USA, Sweden and Switzerland for their reports on developments and knowledge in their own countries. Our gratitude is also expressed to the various speakers, to the session Chairmen and to you, Ladies and Gentlemen, who, by your interest and participation in discussions, have demonstrated your interest and your acceptance of the theme of the conference. Finally, our interpreters have enabled us to break the language barrier and improve our comprehension. The objectives of the conference were two-fold: 1) To demonstrate that the present state of knowledge of the fire behaviour of steel structures is such that it is possible to offer safety levels quite comparable with those attainable in more traditional buildings:

-313-

2) to encourage objective progress in the fire safety of steel structures by the use of models and other methods of calculation. I think it may be said that the tone of these two days was such that they could encourage a better dialogue, a better mutual understanding and that they will be a basis for a common approach and positive collaboration between all parties concerned. The last lä years have seen a phase certainly indispensibleof deepening knowledge concerning the phenomna of fire in relation to steel. You will have appreciated that we have now reached a new stage which could prove more decisive. Three types of observation may be made: 1) First observation: The information required for the rational design of a structure from the fire safety aspect cannot be provided alone by the results of standardised tests on which codes and regulations are still based. The analytical prediction of thermal response, structural behaviour and fire risks becomes even more necessary. We have seen that the practical evaluation of fire resistance of steel structures can now be determined by simple calculations based on well established engineering principles. This type of more realistic analysis contributes to a more economical design and thus to a better competitivity of steel structures and mixed reinforced concrete structures. Recourse to the computer gives an eminently practical and useful dimension to the design of fire resistant steel structures. In this way, it is possible to determine on site the likely effects of a fire.

-314-

The designers, preventive organisations and insurers have therefore now at their disposal -and this must be stressed- new methods of approach that are dynamic and viable and provide quite simple solutions to complex problems 2. Second observation: Very serious enquiries have been made in the Scandinavian countries concerning the financial consequences of fires. I recall that such investigations

are not at present

in progress in France or the Netherlands. The results show that the propagation of fire and the losses due to fire depend on other things besides the fire resistance of the supporting structure. A multitude of different parameters, often in correlation with each other, have an effect. These include, for example, the type of activity and type of building, the active safety measures etc.,... It has been possible to demonstrate that there is no difference in fire losses between steel structures and concrete buildings. It is therefore desirable that insurance companies should take note of the new information available. The consideration of fresh criteria, in place of the presumptions previously used, should give them the means of establishing a tariff rating that is not discriminatory. 3. Third observation. The results of the major and continuous efforts made to study the phenomena, to improve and complete the knowledge via research, tests and experiments of all kinds, enquiries, application of developments and innovations- these results, I say, must be brought to the attention of all the organisations concerned.

-315-

To be sure, our Steel Information Centres, the National organisations for metal construction, The European Convention for metal structures and the ECSC itself, by its considerable financial support of research and development and to-day by the organisation of this conference, all these lead to a dissemination of knowledge by their information and promotion programmes. Their work is important. The impact of it is not, however, always proportional to the investment. It is therefore necessary, in this context, to recall two major principles: a) The information should be concise, clear and orderly and presented in an understandable language. There is no place for scientific jargon. The reader looks primarily for a guiding clue, a simple working tool. b) The information cannot be divorced from technico-economic aspects, the evident base for choice criteria used by decision makers and planners. The steel information centres are directly called to this specific but important task. In this context, they plan to prepare jointly, in the case of the Promotion Committee meeting under the auspices of EUROFER, a small brochure designed to provide this guiding clue not only to practical people but also to private and public decision makers and to investors, often betrayed by siren songs and smoke screens (without intending any play on w o r d s ) . The ECSC, in its role as catalyst that M. Tent mentioned in his introductory address, could usefully bring its blessing and its power to bear on this project, which seems to us to be of high priority. Ladies and Gentlemen, M. Tent mentioned that steel construction in general(i.e. steel used not only in the structure but also in secondary operations and finishing) represents on average about 12.5% of total steel utilisation in the European Community. -316-

It is the determination of the steel and construction industries of our countries to gamble on the future in regard to the improvement and development of their products and activities in steel construction. In the difficult circumstances of which you are aware, these two industries not only manage their problems but also project the base of their revival in the case of a greater satisfaction of the requirements of users. The fire safety of steel structures plays an important role in this strategy. Your participation in this conference has been yery stimulating to the organisers. When safety of people and property is involved, the same preoccupation is common to us all. We hope that this international meeting will have enabled us to make some progress together in this direction.

-317-

LIST

OF

AARNOUDSE, A. Staal­ & Betonkonstrukteur Dow Chemical Nederland B.V. Postbus 48 NL ­ 4530 AA TERNEUZEN ABBADO, G. Architetto INSO S.p.A. Via F. Matteucci, 2 I ­ 50127 FIRENZE ADAM, L. Ingénieur Trade Arbed Belgium S.A . 74, rue de Trêves Β ­ 1040 BRUXELLES ADOLPHS, W. Dipl.­Ing. Thyssen AG Ingenieurabteilung Kaiser Wilhelm Strasse 100 D ­ 4100 DUISBURG 11 ANCILLOTTI, P. Comandante Vigili del Fuoco Via Messina 37 I ­ MILANO ANDERBERG, Y. Lund Institute of Technology Div. of Building Fire Safety & Technology P.O. Box 725 S ­ 220 07 LUND ANDERSEN, N. Dantest Amager Boulevard 108 DK ­ 2300 KØBENHAVN S APPLEYARD, R. Directeur général Commission des Communautés européennes ­ D.G. Marché de l'information et innovation 200, rue de l a Loi B ­ 1049 BRUXELLES

PA R T I C I PA N T S

ARNAULT, P. Directeur CTICM Station d'essais 20, rue Jean Jaurès F ­ 92807 PUTEA UX

AUREAU, P.P. (France) c/o CTICM 20, rue Jean Jaurès F ­ 92807 PUTEAUX BAEHRE, R. Professor Lehrstuhl für Stahl und Leichtmetallbau Universität Karlsruhe Kaiserstr. 12 D ­ 7500 KARLSRUHE 1 BARTLE, P. Superintending Civil Engineer Department of Environment Room B 146 Romney House GB ­ LONDON SUI BARTELS, D. Civiel Ingenieur Hoogovens Groep B.V. NB CVT. 2H.13 Postbus 10.000 • NL ­ 1970 CA IJMUIDEN BATS, J.O. T.H. Eindhoven Den Dolech 2 NL ­ 5612 AZ EINDHOVEN BAUER, E.M. Geschäftsführer des Oesterrei chi sehen Stahlbauverb andes Larochegasse 28 A ­ 1130 WIEN

BAUMANN, H.J. Dipl.­Ing. ΕΤΗ Meto­Bau A G. CH ­ 5303 WUERENLINGEN BECKER, W. Dipl .­Ing./Bauingenieur BASF A ktiengesellschaft Aweta Brandschutztechnik D ­ 6700 LUDWIGSHAFEN/RHEIN

­319­

BEHETS, J.F. Conseiller Centre belgo­luxembourgeois d'Information de l'Acier (CBLIA ) 47, rue Montoyer B ­ 1040 BRUXELLES BIJL, C. L. ir. Staalcentrum Nederland Saturns plein 45 NL ­ ROTTERDA M

BELTRAMI, M. Ingegnere Fiat Engineering SpA Via Belfiore 23 I ­ 10125 TORINO BENNETTS, I. Research engineer BHP Melbourne Research Laboratories P.O. Box 264, Clayton Australia ­ VICTORIA

BICHEL, F. Directeur Constructions Métalliques Bichei Sari Zone I n d u s t r i e l l e Bredewé L ­ 1250 SENNINGERBERG BIRSCHEIDT, H. Ingénieur Service d'Incendie Ministère de l ' I n t é r i e u r 62, rue Principale L ­ 7450 LINTGEN

BOCK, H. Assistent der Geschäftsführung Greschbach Industrie GmbH & Co Postfach 43 Ol 80 D ­ 7500 KARLSRUHE 41

3168, MELBOURNE

BERENBAK, J . C i v i l Eng. Techn. Univ. Delft Hollandia Kloos N.V. Mariënwaard 37 NL ­ 2904 SE CA P. A .D. YSSEL

BOGAERT, W. Inspecteur­Generaal Ministerie van Openbare Werken

Wetstraat 155 (Residence Pal ace) B ­ 1040 BRUSSEL BONGARD, W. Dr.­Ing. Geschäftsführer Deutscher Stahlbau­Verband DSTV Ebertplatz 1 D ­ 5000 KOELN 1

BERGMANN, V. Dipl.­Ing. Deutscher Stahlbau­Verband DSTV Ebertplatz 1 D ­ 5000 KOELN 1

BERNARD, A. Ingénieur ARB ED 255, route d'Arlon L ­ 1150 LUXEMBOURG BERTRAND, J. Ingénieur Civil des Constructions Université de Liège Institut du génie Civil 6, quai Banning B ­ 4000 LIEGE BESSON, F. Ingénieur

BORCHGRAEVE, P. Directeur Centre belgo­luxembourgeois d'Information de l'Acier (CBLIA ) 47, rue Montoyer B ­ 1040 BRUXELLES BORDIN, Α. Chef du Bureau d'Etudes C.F.E.M. 6, boulevard Henri S e l l i e r

F ­ 92150 SURESNES BOUCHART, L. Chef des travaux Paul Wurth S.A . 32, rue d'A lsace L ­ 1122 LUXEMBOURG BOUE', P. Prof. D r . ­ I n g . Bauing. (Stahlbau) L e o ­ T o l s t o i ­ S t r . 19 D ­ 6100 DA RMSTA DT 13

Syndicat de l a Construction Métallique de France 20, rue Jean­Jaurès F ­ 92807 PUTEA UX

­320­

CHIESA, C. Centro Italiano Sviluppo Impieghi Acciaio ­ CISIA Piazza Vel asea 8 I ­ 20122 MILA NO

BOUILLETTE, J.­P. Ingénieur O.T.U.A. 5 b i s , rue de Madrid F ­ 75008 PA RIS

CLAEYS, R. BOUVY, I . Burgel i j k Ingenieur D i p l . Ingénieur R i j k s u n i v e r s i t e i t Gent, Laboratorium Trade A rbed S.A . voor A anwending der Brandstoffen Abteilung technische Beratung ­ DA T S en Warmteoverdracht I, av. des Terres Rouges 4 1 , Sint­Pietersnieuwstraat L ­ 4330 ESCH­SUR­A LZETTE Β ­ 9000 GENT

BRAAS, J. Ingénieur en Chef Paul Wurth S.A. 32, rue d'A lsace L ­ 1122 LUXEMBOURG BRAGARD, A. Ingénieur en Chef Centre de Recherches Métallurgiques Abbaye du Val Benoît II, rue Ernest Solvay Β ­ 4000 LIEGE

COLBRIDGE, G.B. Constrado NLA Tower, 12 Addiscombe Road GB ­ CROYDON CR9 3JH CONZEMIUS, J.­P. Architecte ARBED S.A . Ave de la Liberté L ­ 2920 LUXEMBOURG

BREUER, M. Inspecteur Luxcontrol asbl B.P. 28 L ­ 1050 DOMMELDA NGE

COOKE, G. M.E. Chartered Civil Engineer Fire Research Station of the Building Research Establishment Melrose A venue GB ­ BOREHAMWOOD, Herts. CORDA, F. Commission des Communautés européennes, D.G. Science, recherche et développement 200, rue de la Loi Β ­ 1049 BRUXELLES

BRYL, S. Geilinger Zentrale Forschung und Entwicklung AG CH ­ 8401 WINTERTHUR

CULER, L. Ingénieur­Conseil et Président Fire­Control A SBL 46, av. des V i l l a s Β ­ 1060 BRUXELLES

BRESCIANI, Secrétariat d'Etat a l'Environnement ­ c/o CTICM 20, rue Jean Jaurès F ­ 92807 PUTEA UX

OAHM, V. Officier­commandant adj. Service d'Incendie de l a V i l l e de Luxembourg 50, route d'A rlon L ­ 1140 LUXEMBOURG

CARPENA, A . Secrétaire général · CECM­ECCS­EKS 326,av. Louise, Bte 52

Β ­ 1050 BRUXELLES CAVELIUS, F. Ingénieur/CTICM

DANKERT, H.J. Dipl.­Ing. Deutscher Stahlbau­Verband DSTV Ebertplatz 1 D ­ 5000 KOELN 1

Station d'Essais au Feu Domaine de l'IRSID

F ­ 57210 MA IZIERES­LES­METZ

-321-

DOTREPPE, J.C. Maître de recherches Université de Liège Institut du Génie Civil 6, quai Banning Β ­ 4000 LIEGE

DAUBENFELD, J. Dipl. Ing. Ing.­Büro 116, rue Emile Metz L ­ 2149 LUXEMBOURG DE MACEDO, J.­F. Ingénieur Β Plus Développement 122, ch. de Sourmiou F ­ 13009 MA RSEILLE DE MARTINO, G. Ingegnere Nuova Italsider Via Corsica 4 I ­ 16128 GENOVA DE VRIES W.A.C. Dutch Steel Centre Del kant 7 NL ­ 5311 GA MEREN DEBACKER, Ph. E.Α. Ingenieur Ministerie van Openbare Werken Regie der Gebouwen Wetstraat 155 Β ­ 1040 BRUSSEL DEL PLA TO, S. Architetto Centro Comune di Ricerca delle C E . Divisione Infrastrutture Casella Postale 1 I ­ 21020 ISPRA

DRION, A .^ Attaché technique Service d'Incendie 4, rue de la Tannerie Β ­ 4890 MALMEDY

DUFRANE, G. Commission des Communautés européennes Secretar. Comité. Cons. CECA Bâtiment Jean Monnet L ­ 2920 LUXEMBOURG DURAND, Y. Ingénieur Civil Faculté Polytechnique de Mons 9, rue de Houdain Β ­ 7000 MONS DUTAILLY, L. Commission des Communautés européennes ­ D.G. Emploi, affaires sociales et éducation Bâtiment Jean Monnet L ­ 2920 LUXEMBOURG DUV AL, J.R. Responsable des Services Généraux Solmer F ­ 13776 FOS SUR MER CEDEX

DESCUDE', M. Conseiller Industriel 49, rue des Batignolles F ­ 75017 PA RIS

ELLER, H. Dipl.­Ing. Deutscher Stahlbau­Verb and DSTV Ebertplatz 1 D ­ 5000 KOELN 1

DEWALS, R. Ingenieur Acomal N.V. Hanswijkvaart, 10 Β ­ 2800 MECHELEN

ERMAN, E. Architekt Klaus Schuwirth & Eroi Erman Rathenaustr. 12 D ­ 3000 HANNOVER 1

DINNEQUIN, P. Ingenieur TP Etablissement Public Pare Villette 211, avenue Jean Jaurès F ­ 75019 PARIS

ESMEYER, H. Chef de Sécurité Commission des Communautés européennes Direction Générale "Personnel et Administration L ­ 2920 LUXEMBOURG

­322­

EVANS, P. Commission des Communautés européennes ­ D.G. Science, recherche et développement 200, rue de la Loi Β ­ 1049 BRUXELLES EVENEPOEL, H. Hoofdingenieur ­ directeur Ministerie van Openbare Werken Regie der Gebouwen Wetstraat 155 Β ­ 1040 BRUSSEL FAVRE, J.P. Dipl. Bau.Ing Eth/Sia Gebäudeversicherung des Kantons Bern Viktoriaplatz 25 CH ­ 13000 BERN 25 Postfach FENTON, R. Consulting Engineer Roughton and Fenton 51 Broad Street UK ­ BRISTOL BSl 2EJ, Avon FERRON, J. Commission des Communautés européennes ­ D.G. Science, recherche et développement 200, rue de la Loi Β ­ 1049 BRUXELLES FIOC, Ministère de l'Industrie c/o CTICM 20, rue Jean Jaurès F ­ 92807 PUTEAUX Cedex

FRANSSEN, J.M. Ingénieur C i v i l Fonds national de la Recherche Scientifique 6 , quai Banning Β ­ 4000 LIEGE FRUITET, L. OTUA 5bis rue de Madrid F ­ 75008 PA RIS

FUNHOFF, A. Beton ­ Staalkonstrukteur Bouwer Woningtoezicht Eindhoven Tromplaan 122 NL ­ 6004 ER WEERT GALLINA, G. Ricercatore ICITE CNR Via Lombardia 49 I ­ 20098 S. GIULIANO (MI) GAVRAY, J.­P. Ingénieur­architecte Architecture & Vie 42, rue des Houblonnières Β ­ 4020 LIEGE GERINGER, U. Président du Comité Exécutif de la Convention Européenne de la Construction Métallique GERINGER STA HLBA U Postfach 988 CH ­ 8401 WINTERTHUR

FLAMENT, J . P . Chef du service Sous­Trait ance C.F.E.M.

GELBMANN, Α. Geschäftsführer SYSTEM­Stahlbau Ganglgutstr. 84 A ­ 4060 TRA UN

6, boulevard Henri Sellier F ­ 92150 SURESNES FOURNEAU, X. Journaliste Confédération nationale de la Construction 34­42, rue du Lombard Β ­ 1000 BRUXELLES

GELUK, J.J. Civiel Ingenieur NACO B.V., Netherlands Airport Consultants Jan Van Nassaustraat 115 NL ­ 2596 BS DEN HA A G

FRANCK, N. Directeur A djoint Association des Compagnies d'Assurances 14, rue des Foyers L ­ 1537 LUXEMBOURG

GIBB, J.M. Commission des Communautés européennes ­ D.G. Marché de l'information et innovation Bâtiment Jean Monnet L ­ 2920 LUXEMBOURG

­323­

GIODINGS, T.W. Engineer ­ British Steel Corporation Tubes Division Technical Centre GB ­ CORBY, Northants NN7 IUA

HASS, R. Akad. Rat TU Braunschweig ­ Institut für Baustoffe, Massivbau und Brandschutz Beethovenstr. 52 D ­ 3300 BRA UNSCHWEIG

GLADISCHEFSKI, H. Dipl.­Ing. Beratungsstelle für Stahlverwendung Kasernenstr. 36 D ­ 4000 DUESSELDORF 1

HELDENSTEIN, J. Ing. Conseil, Expert Bureau d'Etudes Heldenstein 3, rue du Fort Reinsheim L ­ 2419 LUXEMBOURG HEVERS, H. Directeur Evers Staalconstr. Hillegom B.V. Horst ten Daal laan 5 NL ­ 2181 GP HILLEGOM

GLESENER, J. Ingénieur dipi Schroeder & Associés Ingénieurs­conseils 8, rue des Girondins L ­ 1626 LUXEMBOURG GOLAY, A. Executive Director International A ssociation for Bridge and Structural Engineering ΕΤΗ ­ Hönggerberg CH ­ 8093 ZUERICH GOSSELIN, J. Secrétaire général Centre belgo­luxembourgeois d'Information de l'A cier 47, rue Montoyer B ­ 1040 BRUXELLES

HILZENSAUER, V. Chemiker Isovolta Osterr. Isolierstoffwerke No rd land str. 1 A ­ 3300 A MSTETTEN HOENIG, 0. Dr.­Ing. Beratender Ing. für Wärme­Energietechnik Höhenblick 24 D ­ 3300 BRA UNSCHWEIG HOFFEND, F. Dipl.­Ing. TU Braunschweig ­ Institut für Baustoffe, Massivbau und Brandschutz Beethovenstr. 52 D ­ 3300 BRA UNSCHWEIG

GRIMAULT, J.P. Ingénieur COMETUBE 5, rue Maurice Ravel F ­ 92300 LEVA LLOIS GRUMBACH, M. Ingénieur IRSID 185, rue Président Roosevelt F ­ 78105 ST­GERMAIN­EN­LAYE CEDEX

HOFMANN, A. Ingénieur Centre Commun de Recherche des Communautés européennes Casella Postale 1 I ­ 21020 ISPRA (Va)

HAMMER, H. Dipl.­Ing. Verband der Sachversicherer e.V. Ri eh 1er Str. 36 D ­ 5000 KOELN 1

HOKARI, M. Structual engineer Nippon Steel Corporation 6­3 Otemachi 2­Chome Chiyoda­ku Japan ­ TOKYO 100

­324­

HOLENWEG, H.R. Dipl .­Ing. ΕΤΗ Beratender Ingenieur Wettsteinstrasse 75 CH ­ 8332 RUSSIKON

JANSS, J. Ingénieur C.R.I.F. 6, quai Banning B ­ 4000 LIEGE JERABEK, H. Techn. Angestellter, Dipl.­Ing. Voest­Alpine Hebag Oberiaaerstrasse 294 A ­ 1232 WIEN

HOLLERICH, J. Ingénieur Bâtiments publics 28, rue Jean l'A veugle L ­ LUXEMBOURG

JOHNSON, K. Research Investigator

HOMMEL, C. Sous­chef de service Arbed S.A . Division de Differdange L ­ DIFFERDA NGE

British Steel Swinden Laboratories Moorgate GB ­ ROTHERHA M,· South Yorks

HONGISTO, R.T. Product Manager Partek Finland SF ­ 08680 MUIJA LA

JONES, T. Commission des Communautés européennes ­ D.G. Marché de l'information et innovation Bâtiment Jean Monnet L ­ 2920 LUXEMBOURG

HOURLAY, P. Ingénieur Ministère de l ' I n t é r i e u r 9 , quai de l a Batte B ­ 4000 LIEGE

JUNCKER, J.P. Chef de service adjoint Arbed S.A . Division de Differdange L ­ DIFFERDANGE

HUBER, R. Responsable Se.Promotion Sacilor Cedex 34 F ­ 92072 PARIS LA DEFENSE

KAN ERVA , J . Assistant Manager Rautaruukki Oy P.O. Box 217 SF ­ 90101 OULU

HULPIAUX, W. Verzekeringsfirma Gerling­Konzern Belgie N.V. 270­272, Tervurenlaan, bus 19 B ­ 1150 BRUSSEL

KERSKEN­BRADLEY, M. Dr.­Ing. I n s t i t u t für Bautechnik Reichpietschufer 72­76 D ­ 1000 BERLIN 30

HUNDT, W. Dipl .­Ing. Klöckner­Werke AG Mannstaedt­Werke Louis­Mannstaedt­Str. 76 D ­ 5210 TROISDORF

KLINGSCH, W. Bergische Universität Wuppertal Pauluskirchstr. 7 D ­ 5600 WUPPERTAL 2

INHA, T. M.Sc. Tampere University of Technology P.O. Box 527 SF ­ 33101 TAMPERE

KNEIPE, L. O f f i c i e r Sapeur­Pompiers Service d'Incendie Agglomération de Bruxelles 1 1 , ave de 1'Héliport B ­ 1000 BRUXELLES

­325­

KRAUSE, R. Dr.­Ing.'/Abteilungslelt G + H Montage Westendstr. 17 D ­ 6700 LU/HAFEN KRUPPA, J. Centre Technique Industrial de la Construction Métallique 20, rue Jean Jaurès F ­ 92807 PUTEAUX CEDEX KUHN, H. Dipl. Baving. Eth Vereinigung Kantonaler Feuerversicherungen Bundesgasse 20 CH ­ 3001 BERN LACHER, G. Prof. Dr.­Ing. Universität Hannover Institut für Stahlbau Callinstr. 32 D ­ 3000 HANNOVER 1 LAFFINEUR, R. Conseiller NEUF ­ Périodique Bimestriel d'Architecture 84A, rue du Merlo, bte 1 Β ­ 1180 BRUXELLES LATHAM, D.J. Principal Investigator British Steel Corporation Sheffield Laboratories Swinden House GB ­ Moorgate, ROTHERHAM S60 3AR LAW, M. Technical Director Ove Arup Partnership 13 Fitzroy street GB ­ LONDON WIP 6BQ LEFEVRE, P. Directeur général Ponts & Chaussées Ministère des Travaux Publics Résidence Palace 155, rue de la Loi Β ­ 1040 BRUXELLES LEGROS, M. Oficier pompier Service Régional d'Incendie 77, rue de Homvent Β ­ 4500 LIEGE (Jupille)

­326­

LEHMANN, R. Dipl.­Ing. Forschungs­ und Materialprüfungsanstalt Baden­Württemberg Otto­Graf­Institut Pfaffenwal dring 4 D ­ 7000 STUTTGA RT 80

LEJEUNE, S. Directeur Université Libre de Bruxelles 50, av. F.D. Roosevelt, Bte 182 Β ­ 1050 BRUXELLES LICKESj J.P. Ingénieur technicien Arbed Recherches 66, rue de Luxembourg L ­ 4002 ESCH­SUR­A LZETTE LICKES, A. Offi zi er­Kommand ant Feuerwehr Stadt Luxemburg Arlonerstrasse 52 L ­ 1140 LUXEMBURG LINSTER, R. Commission des Communautés européennes ­ D.G. Personnel et A dministration Β Ρ 1907 L ­ 2 9 2 0 LUXEMBOURG MAEKELAEINEN, P. Dr. Techn. The A cademy of Finland c/o Helsinki University of Technology Rakentajanaukio 4 SF ­ 02150 ESPOO 15 MAEKELBERG, S. Ingénieur Civil Ministère de l'Intérieur 1 , Leuvense weg Β ­ 1000 BRUSSEL

MAUER, G. Ingénieur dipi. Schroeder & Associés Ingénieurs­conseiIs 8, rue des Girondins L ­ 1626 LUXEMBOURG

MELCHIORRE, Α. Ingegnere CISIA ­ Centro Italiano Sviluppo Impieghi A cciaio Piazza Vel asea 8 I ­ 20122 MILANO

NIEMIMAA, R. Dipl.­Ing. Ins.tsto Reino Niemimaa Mustikkatie 15 SF ­ 03100 NUMMELA

MERCEREAU, C. Assemblée Pieni ère des Sociétés d'Assurances I.R.D. 11, rue Pillet­Will F ­ 75009 PA RIS

OOMS, A. Directeur Constructiewerkplaats en Machinefabriek Bailey Ν.V. Lekdijk 266 a NL ­ 2957 ZG NIEUW LEKKERLA ND

MESKENS, J. Architekt­Professor Hoger St. Lukasinstituut Paleizenstraat 70 Β ­ 1030 BRUSSEL

ØSTBY, Α. Overingeniør Statens brannninspeksjon Postboks 498 Sentrum Ν ­ OSLO 1

MOECHER, P. Ingénieur technicien

PASSEROTTI, C. Progettista ITALIMPIANTI Div. Impianti Industrie Leggere Viale Liegi 33 I ­ 00198 ROMA

Bureau d'Etudes

Secotechnique 2, rue des Sapins L ­ 2513 SENNINGERBERG MOSER, K.M. Geschäftsführer Brandverhütungsstelle für Oö Staatlich autor. Prüfanstalt Petzoldstrasse 45 A ­ 4020 LINZ MOTTA, Α. Ricercatore CNR ­ ICITE Via Lombardia 49

PECHÓN, P. Président du CIDECT Comité Intern, pour le Develop, et l'Etude de la Construction Tubulaire 5, rue Maurice Ravel

F ­ 92300 LEVA LLOIS­PERRET

I ­ 20098 S. GIULIANO MILA NESE

POUPLOT, Ministère de l'Urbanisme et du Logement ­ c/o CTICM 20, rue Jean Jaurès F ­ 92807 PUTEAUX

MOUTY, J. Directeur Technique Cometube 5, rue Maurice Ravel F ­ 92300 LEVA LLOIS

PRADER, H. Ing. Geschäftsführer Tiroler Landesstelle für Brandverhütung Wilhelm­Greil­Strasse 12/11 A ­ 6020 INNSBRUCK (A )

MULLER, Commandant BSP (France) c/o CTICM 20, rue Jean Jaurès F ­ 92807 PUTEA UX

PRUEFER, H. Commission des Communautés européennes ­ D.G. Science, recherche et développement 200, rue de la Loi B ­ 1049 BRUXELLES

NERAD, L. Stahlbauingenieur VOEST­ALPINE AG Floragasse 7 A ­ 1040 WIEN

­327­

QUAST, U. Professor ' Technische Universität Braunschweig Beethovenstr. 52 D ­ 3300 BRA UNSCHWEIG RAHIER, J. Ingénieur

RORET, J. Premier Vice Président Syndicat de la Construction Métallique de France 20, rue Jean­Jaurès F ­ 92807 PUTEAUX ROTONDO', P.P. Commission des Communautés européennes ­ D.G. Marché de l'Information et Innovation Bâtiment Jean Monnet L ­ 2920 LUXEMBOURG

Service Régional d'Incendie 56, rue Laiwisse Β ­ 4571 St A NDRE

RAUTAKORPI, E. Dipl.­Ing.

ROUX, Comité Européen des Assurances 11, rue Pillet­Will F ­ 75009 PARIS RUBERT, A. Dr.­Ing. Fored. Krupp GmbH Krupp Forschungsinstitut Münchener Str. 100 D ­ 4300 ESSEN

Insinööritoimlsto Te­Em Oy Kauppakatu 7A 12 SF ­ 33200 TA MPERE

REMACLE, Η. Ing. PPL CH SV Institut National du Logement 10, Boulevard St Lazare Β ­ 1030 BRUXELES REUTER, C. Ingénieur principal Arbed S.A. Division de Differdange L ­ DIFFERDANGE

RUTISHAUSER, E. Architekt HTL Gebäudeversicherung Kt. Zürich Abteilung Feuerpolizei Kurvenstr. 31 CH ­ 8090 ZUERICH SAARIKOSKI, H. Civil engineer Finnish Centre for Radiation and Nuclear Safety PL 268 SF ­ 00101 HELSINKI 10

RINDORF, K. Civil engineer Odense Teknikum Niels Bohrs Alle 1 DK ­ 5230 ODENSE M ROBINSON, J. Macket Development Manager British Steel Corporation B.S.C. Section P.O. Box 24, Steel House GB ­ REDCAR, Cleveland TS10 5QW

SANDBERG, P. Managing Director The Finnish Constructional Steelwork A ssociation P.L. 623 SF ­ 00101 HELSINKI 10 SANDMAN, T. Civil Engineer Swedish Institute of Steel Construction Drottning Kristinas Väg 48 S ­ 11428 STOCKHOLM

ROBUSTELLI, L. Ingenieur SETIEM 44, rue de Cronstadt F ­ 75015 PA RIS RONDAL, J. Chargé de cours Université de Liège Institut du Génie Civil 6, quai Banning Β ­ 4000 LIEGE

SBARRA, J. Chef de service adjoint Arbed S.A. Division de Differdange L ­ DIFFERDANGE

­328­

SCHECH, H. Architekt BdB Thyssen Edelstahlwerke AG Bereich Neubau Oberschlesienstrasse 16 D ­ 4150 KREFELD SCHEIDWEILER, G.­ Inspecteur Ministère de l'Intérieur Service d'incendie 4, rue Kleischter L ­ 8078 BERTRA NGE SCHLEICH, J.B. Ingénieur diplômé Arbed Recherches 66, rue de Luxembourg L ­ 4002 ESCH­SUR­A LZETTE

SIPONEN, 0. Dipl.­Ing. Insinööritoimisto Ossi Siponen Ky Köydenpunojankatu 10B SF ­ 00180 HELSINKI 18 SMITH, C.I. Fire Engineering Consultant Ian Smith (Consultancy) LTD Cleveland House, 34 Borouhbridge Road GB ­ NORTHALLERTON, North Yorkshire DL7 8BE SMITHER, Government Scientist Biotechnology Unit­ Laboratory of the Government Cornwall House, Stamford Street UK ­ LONDON SEI 9NQ

SCHORDERET, E. Dipl.­Ing. Eth Zschokke ­ Hartmann AG CH ­ 5200 BRUGG

STEFFENS, J.G. Chef du Service Incendie Service d'Incendie Malmedy 31, place A lbert I Β ­ 4890 MALMEDY

SCHROEDER, P. Ing. dipi. ­ Directeur technique Schroeder & Associés Ingénieurs­conseils 8, rue des Girondins L ­ 1626 LUXEMBOURG

STEFFENS, J.A. Directeur ­ Gérant S.A. Travhydro Division Socotube 9, av. du Progrès Β ­ 4430 ALLEUR (A ns) STRIER, 0. Geschäftsführer Ott Mo 11 er GmbH Dornbacherstrasse 115/6 A ­ 1170 WIEN TANNINEN, T. Civil Engineer The Finnish Construction Steel Work A ssociation P.O. Box 623 SF ­ 00101 HELSINKI 10

SCHUWIRTH, K. Architekt Architekten Schuwirth & Erman Rathenaustrasse 12 D ­ 3000 HANNOVER 1 SERPILLO, D. Ingegnere Nuova Italsider S.p.A . Via Corsica, 4 I ­ 16128 GENOVA SIMON, N. Assurances "Le Foyer" 6, rue A. Borchette L ­ LUXEMBOURG

TASIC, S. Dipl. Ing. Institut za metalne konstrukcije Mencingerjeva 7 YU ­ LJUBLJA NA

SIOKOLA, W. Univ.­Assistent Universität Wien Inst, für Stahlbau Karlsplatz 13 A ­ 1040 WIEN

TENT, H. Commission des Communautés européennes ­ D.G. Science, recherche et développement 200, rue de la Loi Β ­ 1049 BRUXELLES

­329­

TEPHANY, H. Ministère de l'Intérieur et de la Décentralisation / Direction de la Sécurité Civile 20, rue Jean Jaurès F ­ 92807 PUTEA UX THOR, J . Stal byggnadsinstitutet Drottning Kristinas väg 48 S ­ 11428 STOCKHOLM

VAN DEN EYNDE, G. Verzekeringsfirma Gerling­Konzern België Ν.V. 270­272, Tervurenlaan, bus 19 B ­ 1150 BRUSSEL VERKEYN, A . Inspecteur­Generaal van Bruggen en Wegen Min. Openbare Werken Vossenlaan 28 B ­ 1980 TERVUREN

VERWAAL, J. Bouwkundige PUEM N.V. Postbus 110 NL ­ 3500 AC UTRECHT

TINCANI, M. Capo Pompieri CCR ­ ISPRA Via E. Fermi, 54 I ­ 21020 ISPRA (VA RESE)

VIGNOLI, G. Ingegnere TENSOTECI SrL Via Volta 16 I ­ 20093 C0L0GN0 MONZESE MILA NO

TODINI, G. Commission des Communautés européennes ­ O.G. Emploi, affaires sociales et éducation Bâtiment Jean Monnet L ­ 2920 LUXEMBOURG

VISEU, J. Ingénieur civil, Directeur Siderurgia Nacional E.P. Rua Braamcamp, 7 Ρ ­ 1297 LISBOA Codex WAESPE, W. Architekt

TOURNAYE, J. Ingénieur Baudin­Chateauneuf B.P. 19 F ­ 45110 CHA TEA UNEUF/Loire

BVD­Brand­Verhütungs ­ Dienst Nüschelerstr. 45 CH ­ 8001 ZUERICH

TOVAR DE LEMOS, A. Professor Istituto Superior Técnico CMEST Av. Rovisco Pais Ρ ­ 1096 LISBOA Cedex

WAHL, L. CTICM 20, rue Jean Jaurès F ­ 92310 PUTEA UX

TWILT, L. IBBC­TNO Postbus 49 NL ­ 2600 A A DELFT

WALKER, H.B. Constrado Addiscombe Road 12 GB ­ CROYDON CR9 3JH

VALLERA. Ingénieur­Conseil 52­54, rue du Docteur Blanche F ­ 75016 PA RIS

WALLIN, L. Swedish Institute of Steel Construction Drottning Kristinas väg 48 S ­ 114 28 STOCKHOLM

VAN BOCXLA ER, L. Reporter

WARDEN 1ER, J. Sen. Leet. Delft Univ. of Technology Stevinweg 1 NL ­ 2628 CN DELFT

Weekblad "Het Bouwbedrijf" Lombardstraat 34­42 B ­ 1000 BRUSSEL

­330­

WEBER, F·. Technicien Lobjois & Weber S.A . Bureau de Constructions Métalliques Passage de Montriond 4 CH ­ 1006 LA USA NNE WEBER, L. Ingénieur diplomé Trade ARBED L ­ OUDELANGE WISNIESWSKY, G.K. Geschäftsführer Studiengesellschaft für Anwendungstechnik von Eisen und Stahl E.V. Kasernenstrasse 36 0 ­ 4000 DUESSELDORF WITTE, H. Dr.­Ing. Instruct­Ingenieur ­ GmbH Liebenaustr. IIA D ­ 6200 WIESBADEN WITTEVEEN, J. Professor TNO Institut für Baumaterialien und Stahlkonstruktion Postbus 49 NL ­ DELFT

WUERKER, K.G. Dipl.­Ing. Mannesmann Röhrenwerke Postfach 1104 D ­ 4000 DUESSELDORF 1 WUPPERMANN, G.T. Dr.­Ing. Theodor Wuppermann GmbH Friedrichstr. 38 D ­ 5090 LEVERKUSEN 1 WYSS, U. Dipl.­Ing. Schweizerische Zentralstelle für Stahlbau Seefeldstr. 25 CH ­ 8034 ZUERICH ZUCKA, J. Conseiller Centre bel go­luxembourgeois d'Information de l'A cier 47, rue Montoyer Β ­ 1040 BRUXELLES

­331­

European Communities — Commission EUR 10116 — Fire­safe steel construction: practical design Luxembourg: Office for Official Publications of the European Communities 1985 — V, 331 pp., 163 fig., 16 tab., 24 ph. — 16.2 χ 22.9 cm Technical steel research series DE, EN, FR ISBN 92­825­5718­9 Catalogue number: CD­NC­85­082­EN­C Price (excluding VAT) in Luxembourg: ECU 26.58 BFR 1200 IRL 19.10 UKL 15.20 USD 21

Over the last few years, considerable progress has been achieved in the development of design methods for the study of the fire safety of buildings. This rational or analytical approach (fire engineering) now provides a more and more oper­ ational means for assessing the behaviour of steel structures exposed to fire. These favourable developments have their basis in a large number of international research projects to which cooperation at a European level has made considerable contribution. The Conference provided information on the methods and recommendations which en­ able a practical and reliable approach to be made to the design and construction of the buil­ dings with fire­resistant steel structures and to the search for solutions which can meet economic, architectural and safety requirements. It was mainly of interest to architects, engineers, construction and consultant, legislative bodies, firemen, insurance underwriters and investors.

Salg og abonnement - Verkauf und Abonnement - Πωλήσας και συνδρομές ­ Sales and subscriptions Vente et abonnements ° Vendita e abbonamenti ­ Verkoop en abonnementen BELGIQUE/BELGIË

IRELAND

M o n i t e u r belge / Belgisch Staatsblad

G o v e r n m e n t Publications Sales Office

Mundi­Prensa Libros. S.A .

Rue de Louvain 4 0 ­ 4 2 / Leuvensestraat 4 0 ­ 4 2 1 0 0 0 Bruxelles / 1 0 0 0 Brussel

Sun A lliance House Molesworth Street Dublin 2 Tel. 7 1 0 3 0 9

Castello 3 7 E ­ 2 B 0 0 1 Madrid Tel. (91) 2 7 6 0 2 5 3 ­ 2 7 5 4 6 5 5 Telex 4 9 3 7 0 ­ M P L I ­ E

or by post Stationery Office St Martin's House Waterloo Road Dublin 4 Tel. 6 8 9 0 6 6

PORTUGAL

Tél. 512 00 26 CCP/Postrekening 0 0 0 ­ 2 0 0 5 5 0 2 ­ 2 7 Sous­dópots / A gentschappen: Librairie européenne / Europese Boekhandel Rue de la Loi 2 4 4 / Wetstraat 2 4 4 1 0 4 0 Bruxelles / 1 0 4 0 Brussel CREDOC Rue de la Montagne 3 4 / Bergstraat 3 4 Bte 11 / Bus 11 1 0 0 0 Bruxelles / 1 0 0 0 Brussel

DANMARK Schultz Forlag Møntergade 2 1 1 1 1 6 København Κ Tlf: (01) 12 11 9 5 Girokonto 2 0 0 11 9 5

BR DEUTSCHLA ND

Livraria Bartrand, s.a.r.l. ITALIA

Rua João de Deus Venda Nova Amadora Tél 9 7 4 5 7 1 Telex 1 2 7 0 9 ­ L I T R A N ­ P

Licosa Spa Via Lamarmora, 4 5 Casella postale 5 5 2 5 0 121 Firenze Tel. 5 7 9 7 5 1 Telex 5 7 0 4 6 6 LICOSA CCP 3 4 3 5 0 9

I

Subagente: Librerìa scientifica Lucio de Biasio ­ A EIOU Via Meravigli. 16 2 0 1 2 3 Milano Tel. 8 0 7 6 7 9

GRAND­DUCHÉ DE LUXEMBOURG

Verlag Bundesanzeiger

Office des publications officielles des C o m m u n a u t é s européennes

Breite Straße Postfach 10 8 0 0 6 5 0 0 0 Köln 1 Tel. (02 2 1 ) 2 0 2 9 ­ 0 Fernschreiber: ANZEIGER BONN 8 8 8 2 5 9 5

5. rue du Commerce L ­ 2 9 8 5 Luxembourg Tél. 4 9 0 0 8 1 ­ 4 9 0 1 9 1 Télex PUBOF ­ Lu 1 3 2 2 CCP 1 9 1 9 0 ­ 8 1 CC bancaire BIL 8 ­ 1 0 9 / 6 0 0 3 / 2 0 0 Messageries Paul Kraus

GREECE

1 1 . rue Christophe Plantin L ­ 2 3 3 9 Luxembourg Tél. 4 8 2 1 3 1

Télex 2515 G.C. Eleftheroudakis SA International Bookstore 4 Nikis Street Athens ( 1 2 6 ) Tel. 3 2 2 6 3 2 3 Telex 2 1 9 4 1 0 ELEF

CCP 4 9 2 4 2 ­ 6 3

Sub­agent for Northern Greece:

Christoffel Plantijnstraat Postbus 2 0 0 1 4 2 5 0 0 EA 's­Gravenhage Tel. ( 0 7 0 ) 7 8 9 9 11

M o l h o ' s Bookstore The Business Bookshop 1 0 Tsimiski Street Thessaloniki Tel. 2 7 5 2 7 1 Telex 4 1 2 8 8 5 LIMO

SCHWEIZ / SUISSE / SVIZZERA Librairie Pavot 6. rue Grenus 1211 Genève Tél 31 89 50 CCP 12­236

UNITED STA TES OF A MERICA European C o m m u n i t y Information Service 2 1 0 0 M Street. N W Suite 7 0 7 Washington. DC 2 0 0 3 7 Tel. (202) 8 6 2 9 5 0 0

NEDERLAND Staatsdrukkerij­ en uitgeversbedrijf

UNITED KINGDOM

CANADA Renouf Publishing Co.. Ltd 6 1 Sparks Street (Mall) Ottawa Ontario K1P 5 A 6 Tel. Toll Free 1 ( 8 0 0 ) 2 6 7 4 1 6 4 Ottawa Region ( 6 1 3 ) 2 3 8 8 9 8 5 ­ 6

H M Stationery Office

FRANCE Service de vente en France des publications des Communautés européennes Journal officiel 2 6 . rue Desaix 7 5 7 3 2 Paris Cedex 15 Tél. (1) 5 7 8 6 1 3 9

H M S O Publications Centre 51 Nine Elms Lane London S W S 5DR Tel. 0 1 ­ 2 1 1 3 9 3 5

JAPAN

Sub­agent: Alan A rmstrong & A ssociates Ltd

Kinokuniya C o m p a n y Ltd

72 Park Road London N W 1 4 S H Tel. 0 1 ­ 7 2 3 3 9 0 2 Telex 2 9 7 6 3 5 A A A LTD G

1 7­7 Shinjuku 3­Chome Shiniuku­ku Tokyo 1 6 0 ­ 9 1 Tel. (03) 3 5 4 0 1 3 1

NOTICE TO THE READER All scientific and technical reports published by the Commission of the European Communities are announced in the monthly periodical 'euro abstracts'. For subscription (1 year: BFR 2 800) please write to the address below.

ISBN ^E-ñ5S-S710-T Price (excluding VAT) in Luxembourg ECU 26.58 BFR 1200 IRL19.10 UKL15.20 oj/, U p

OFFICE FOR OFFICIAL PUBLICATIONS OF THE EUROPEAN COMMUNITIES

L—2985 Luxembourg

USD21