The Details Of Water Intakes Their Effect On The Turbidity Of Water.(1).pdf

THE DETAILS OF WATER INTAKES THEIR EFFECT ON THE TURBIDITY OF WATER

by

Robert J. Lindsay

A thesis submitted to the Faculty of Graduate Studies and Research in partial fulfilment of the requirements for the degree of Master of Engineering.

Department of Civil Engineering McGill University Montreal August 1960

TABlE OF CONTENTS Page

. . . . . . . . . . . . . . . . . . . . . . . . . . . . ••

Preface • CHAPTER I General Considerations. • • • • o o o • • • • o • • • Limita ti ons of Investigation. • • • • • • • • • • • o Classification of Intakes • • • • • • • • • • • • • •

o

•

o

•

•

•

o

•

o

•

•

•

o

•

o

•

o

•

•

•

•

CHAPTER II - "The Common Factors" The Common Qualities • • • • • • • • • • • • • • • • • • • • • • • • The Common Problems • • • • • • • • • • • • • • • • • 1. The Problem.. of Position Requirements for Reliability of Operation. • • • • • • • • Requirements for Efficient Operation • • • • • • • • • • • Procedures in solving the Position Problem • • • • • • • • 2. The Problem of Structural Form Tower Intakes • • • • • • • • • • • • • • • • • • • • Submerged Crib Intakes • • • • • • • • • • • • • • • • • • Simple Pipe Intakes. • • • • • • • • • • • • • • • • • • • Shore Intakes • • • • • • • • • • • • • • • • • • • • • • Special Intakes. • • • • • • • • • • • • • • • • • • • • • Requirements for Econo~ • • • • • • • • • • • • • • • • • Requirements for Reliability of Operation • • • • • • (a) Foundations • • • • • • • • • • • • • • • • • • • • (b) External Forces - Loading • • • • • • • • • • • • • (c) Protective Works • • • • • • • • • • • • • • • • • • Requirements for Efficient Operation • • • • • • • • • • • (a) Material used to weigh down Intakes • • • • • • • • (b) Racks and Screens • • • • • • • • • • • • • • • • • (c) Inlet Ports • • • • • • • • • • • • • • • • • • • • (d) Intake Velocities • • • • • • • • • • • • • • • • • (e) Shape of Inlet Port • • • • • • • • • • • • • • • • (f) Position of Inlet Port • • • • • • • • • • • • • • • (g) The Auxiliary Structures • • • • • • • • • o • • • o The Problem of Installation Large Intakes. • • • • • • Smaller Intakes. • • • • • • o • • • • • • • • • o • • o • Shore Intakes. • • • • • • • • • • • • • • • • • • • • • • River Intakes - Auxiliary Structures • • • • • • • • • • • 4. The Problem of Maintenance Back-Flushing the Intake • • • o • • • 0

•

...

.......

...... ..

CHAPTER III - "The Special Conditions

0

•

7 8

9 10 12 15 16 17 18 19 21 22 22 24 30 31 32 33 34 35 37 45 48 56 57 58 59

62

lee and Variable Stage"

Section I - lee • • o • • • o o o o o o o o • • o • • • • o • o o • Classification of lee. • • • • • • o • • • • • • o o o • o • • Frazil lee - Scope of the Problem. • • • • • • • • • • • • • • Countermeasures Available. o • o • o • • • • • • • o • • • (a) Effect on Position • • • • o • • • • • • • • • • • • •

i

1 4 4

66 67 69 71 72

Page

74

(b) Effect on Structural Form ••• • • (c) Effect on Installation. • • • • • (d) Effect on Maintenance • • • • • • • • • Effect of lee - Conclusion. • • • • • • • • Section II - Rivera and Streams of Variable Stage. • Effect in General • • • • • • • • • • • • • Effect on Position. • • • • • • • • • •••••••••• Effect on Structural Form • • • • • Effect on Installation. • • • • • • • • • • • • • • • • • • Effect on Maintenance • • • • • • • • • • • • • • •

. .. .. ..

78 84 84 84

... ... .. . . . ... ..

CHAPTER IV

86 86

89 99 99

"Illustrations of Some Typical Intakes"

Foreword • • • . . • • • • • • • • • • • • • • • • • • • • • • • • • 101

Tower Intake Fig. 19 • • • • • • • • • • • • • • • • • • Intake in a Bay - Fig. 2o(a) • • • • • • • • • • • • • • • • • • • • • •••••••••••• River Intake- Fig. 20(b). • • • • • • • • • • • • Submerged Crib Intake - Fig. 21. • • • Shore-Intake with Dry-Well - Fig. 22 • • • • • • • • • • • • • • • • CHAPTER V - "The Experimental Investigation - Scope, Theory, Object and Limits" General Scope. • • • • • • • • •••••••••• Terms of Reference • • • • • • • • • • • • • • • • • • • Specifie Assumptions • • • • • • • • • • • • • The Hypothe sis • • • • • • • • • • • • • • • • • • • • • • • • • Discrete Particle Theory • • • • • • • • • • • • • • • • • • • • • • • ••••• Experimental Objectives - Summary. • • • • • • • • • Limits of Investigation. • • • • • • • • ••••••••••• CHAPTER VI "The Experiment, Apparatus, Procedure and Observations" Apparatus • • • • • • • • • • • • • • • • • • • Additional Equipment ••• • • • • • • • • • Procedure • • • • • • • • • • • • • • Observations • • • • • • • • • • • • CHAPTER VII - "The Experimental Conclusions" Evaluation of Resul ts. • • • • • • ••••••••••••••• The First Objective •••. • • • • • • • • • • • • • • • • • • • • • • The Second Objective • • • • • • • • • • • • • • • • • • • • • • • • Curve Fitting - Method of Least Squares • • • • • • • • • • • • • • • Conclusions. • • • • • • • • • • • • • • • ••• Summary of Conclusions • • • • • • • • • • • • • • • • • • • • ••• Importance of Conclusions. • • • • • • • • • • • • • • • • • • • • • Future Investigations. • • • • • • • • • • • • • • • • • • • • • • • APPENDIX A Bib1iography

. . .. .. .. .. .. .. .. .. .. . . . ... . . . . . . . . .. .. .. . . .. .. ..

APPENDIX B Foreword • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • PART I - Experimental Conditions Estab1ished by Pre1iminary Investigation and Study • • • • • • • • • • • • • • • • • • PART II - Calculations Preliminary to the Design of Some Elements of the Experimental Apparatus. • • • • • • • • • ••• PART III - Report on the Performance of Some Elements of the Experimental Appara tus. • • • • • • • • • • • • • • • • • - ii -

102 103 103 104 105

106 107 111 113 114 115 116 119 125 127 131 156 156 159 166 175

177 178 179

i ii v xxiii

PREFACE

This Thesis is primarily concerned with Water Intakes.

These

structures, which are an elementary part of any self-sufficient watersupply system, are necessary wherever the source of raw-water is a stream, river, lake or impounding reservoir.

In comparison to the more complex

units of a water-supply system, such as pumping stations and treatment plants, the water intake has been the subject of very little research. Even a summary reading of sorne of the published reference material on water intakes will serve to establish this statement.

One might conclude,

with some justification, that the detailed design of water intakes is 11

based on certain proven

rules-of-thumb 11 and

11

conventional 11 practice,

rather than on established principle and experimental research; that expediency and not experiment has been responsible for the details of water intakes.

&

When on considera that the very operation, efficient or

otherwise, of all other elements of a water-supply system depends primarily on the functioning of the intake, this imbalance of interest is hardly justifiable in theory or practice. However illogical, this is the situation confronting the investigation under consideration.

For this reason, the Bibliography appended

to the thesis (Appendix A) has been divided into three parts, separating the pertinent and relevant material from that which is general and vague. The first three Chapters of the thesis represent an attempt by the author to correlate and analyse the problems encountered in the positioning, design, and operation of water intakes.

The remaining Chapters

present a report on the experimental investigation of one of these problems, which is related to the positioning of the mouth of the intake

in relation to the direction of the flowing current. In an effort to present information which would be applicable to a wide range of

installation~

which would reflect the practice favoured

on the North American continent and which would be in keeping with modern constructional methods, the examples and references cited in the text have been limited to those which appear in technical journals published subsequent to 1930, or to textbooks of recognized authority. tion, particular emphasis has been placed on water intakes

In addithat would

be suitable for the smaller urban communities with populations of under fifty thousand persons, without excluding at the same time, the necessary and valuable consideration of the larger intakes which serve the metropolitan areas. The author is indebted to his Director of Research, Professor Andrejs Pakalnins, Department of Civil Engineering and Applièd Mechanics, McGill University, for his encouragement and helpful supervision in all phases of this research; to Professor V.

w.

G. Wilson, also of the

Department of Civil Engineering, for his advice on the assembly of the test flume and for the assistance of his staff in the erection of the apparatus; to Mr. Wellington Chen, B•Eng. (Civil) '59, for his assistance during the experimental period in the repetitive work of weighing samples.

Robert J. Lindsay

Montreal, Quebec August 1960

CHAPTER I General Considerations.

In 1957, there were some eight hundred

municipalities in Canada with populations in excess of one thousand persona.

In all but a few of these, water for domestic purposes was

supplied by a publicly owned utility which maintained the distribution system required.

Ten per cent of the municipalities were dependent on

adjacent cities for their water supply.

Among the remaining self-

sufficient ninety per cent, the sources of raw-water were as follows (*38): Rivera, streams and creeks •• • • • • • • • • 34% Lakes and impounding reservoirs. • • • • 30% Total - Surface Water Sources 6~ =

.....

Springs, Wells and Infiltration Galleries •• • • • 36% Total - Ground Water Sources = 36% It may be concluded therefore that two-thirds of the municipalities in Canada today are utilizing some form of water intake as an integral part of their water-supply system.

The water intakes constructed by

industrial concerna for their own use, are not included in the statistics above; were they to be included, the proportion between surface and ground water sources of supply would be approximately three to one. These findings are in accord with the general rule that in any locality, .where ground water and surface water resources are equally accessible to the consumer, and where the quantity of water required is so large asto make the quality of secondary importance,an economical

*

Bibliography Reference

2

choice will generally favour the surface supply.

While the rule tends

to oversimplifY the problem by its very premise, the validity of its conclusion is evident in both industrial and municipal practice. Large users of water; railroads, breweries, sugar and oil refineries, pulp and paper mills, locate on riparian property in an effort to ensure for themselves not only the economical advantage of marine transportation for their manufactured products, but also an abundant and cheap supply of raw-water for their industrial processes. The demands on the municipal supply are also increasing.

The

widespread use of labour-saving household appliances, the trend to air conditioned homes and business premises, the increasing demand for public recreational facilities which require water being expended, and the general overall rise in population density of the cities and

town~all,

point to an increase in the number of gallons per capita, per day, (gcd) that must be supplied.

In 1940, Turneaure and Russell concluded that

water consumption in the United States ranged between 45 gcd and 175 gcd

(u.s.),

with a probable average of llO gcd

(u.s.). In

1947, Mr. Steel

had already revised the average value upward to 135 gcd

(u.s.).

Modern

practice favours the design of filtration plants for industrial towns on the basis of an average demand of 150 gcd, recommending as well that fully metered distribution systems be introduced in the event that consumption begins to outstrip plant capacity before the desing-period of twenty years has elapsed. The evidence then clearly indicates an upswing in water consumption. Among the resulting effects, the following are important to the investigation under consideration:(a) Water Intakes will increase in number and geographical

3 distribution.

Municipalities, presently supplying water

from ground water resources which are taxed to maximum capacity, will be compelled to revaluate the advisability of dual supply.

Hivers and streams hitherto considered

as undesirable sources of water because of pollution or chemical content will be re-examined in the light of economical necessity.

An increase in the cost of water

treatment is to be preferred to no water at all.

This

increase in oost of water from the municipal supply may well priee it beyond the economie reach of sorne of the smaller industries which, up to now, have made use of such water because it was less costly than providing and maintaining their own intake.

For these, especially where

water treatment of the calibre required for domestic purposes is an unnecessary refinement, the economies of the situation may well be reversed. (b)

The reliability and efficiency of the water in.tà.ke will be the subject of greater interest and experimentation. The increase in the number of installations forecasted in (a) above will produce a widening interest in the problems of intakes.

A larger and more critical group of

investigators will become concerned in the search for solutions.

It is to be hoped that such interest will

serve to reinforce the general premise of this thesisf namely, that water treatment is not the exclusive domain of the filtration plant, and that partial treatment can, and should, begin at the intake proper.

4

Limitations of Investigation.

This thesis concerna itself primarily

with the intake structure itself - the water intake proper - as opposed to the intake system which normally includes the intake proper, the conduit between the intake and the shore, and the shore-well or suction well. The inter-relation of these three elements which are common to most intake systems is shown in Fig. 1, Page 5.

The function of the in-

take proper is to admit raw-water of desirable quality, and in the required quantity, to the intake system.

The conduit between the intake

and the shore serves to convey this raw-water to the suction inlets of low-lift pumps located in, or immediately above, the suction-well. From the suction-well, the raw-water is pumped into another conduit which carries it to the filtration or treatment plant.

Detailed dis-

cussion of the functions of the conduit or suction-well, important though they may be, are beyond the scope of this investigation; these elements will be considered only in so far as they affect, or are affected by, the details of the water intake itself. Classification of Intakes.

In the interests of brevity, most

authors do not discuss the water intake as a separate structure unless it is very large, preferring to stress its interdependant role in the system of raw-water conveyance.

Furthermore, in an effort to impart

general principles, most authors classify water intakes in accordance with the character of the supplying stream or body of water.

Thus

Turneaure and Russell, Steel, Lischer and Hartung, subdivide intakes i nto; River Intakes, Lake Intakes and Intakes from Impounding Reservoirs, ,fo

Intakes on Va r iable Streams (*1).

However authori t a tive the wri ter, this

method of class ification inevitably encounters a dilemma; either to stress

5

6 the factors common to all analogous installations and omit the peculiar details, or to emphasize and explain the details which make the installation unique and narrow the range of application.

For example, a

river of variable stage in the southeastern United States presents a problem of high turbidity when the river is at flood stage; a river of comparable size in Canade is relatively free from turbidity at high stage, but frazil ice formation at low winter flows constitutes the real difficulty.

The intake structure designed in each instance will have

some common traits, but equally important and critical dissimilarities will exist. In an effort to reconcile the two schools of thought and to correlate the information supplied by both, the classification of intakes in accordance with the character of the supplying source will be abandoned in favour of a more analytical approach.

The classification

substituted will endeavour to divide the problems encountered in the positioning, structural design, installation and operation of water intakes into two categorie5; namely, those which are common to all (Chapter II), and those which are of special importance (Chapter III). In both instances, solutions will be

p ~~ed

in the briefest of terms.

This brevity should not be construed as an implication that the solutiens advanced are either facile or foolproof. as a last generality, that such is not the case.

The author acknowledges,

7

CHAPTER II "The Common Factors" The Common Qualities.

All intakes, irrespective of the character

of the supplying source, must possess certain common attributes or qualities of operation.

These qualities are expressed in a two-fold

re quiremen t: Reliability of Operation, - the prime requisite of all intakes; by comparison, all other considerations are secondary in importance.

This simple statement requires no further elabora-

tion beyond that of mentioning that where intakes have failed in this regard, the remedies are expensive and extreme including in a few cases, the dynamiting of the intake conduit! More commonly, the capacity of the intake is sharply reduced and the resulting inadequacy of the supply depletes the reserves of water required for fire fighting and the maintenance of pressure. In theory, a dual supply is the answer.

In practice, as

Mr. McDonald confirms (*3), economical considerations make this solution prohibitive except for the largest metropolitan areas. Efficiency of Operation.

Under this heading are grouped a

number of requirements; some refer to the quality of the rawwater admitted, such as the exclusion of highly turbid and polluted waters, while others refer to the attributes of the intake itself, such as ease of maintenance and accessibility.

8

These requirements are not all absolute and alternatives do exist; in any given case, they may make conflicting demands on the design; in all cases, they are subordinate to reliability of operation. In the discussion which follows immediately, the relative importance of these qualities will be stressed and the material which relates to each of the topical problems presented will be divided accordingly. The Common Problems.

The design of any water intake requires the

answers to four questions; namely,

I

i)

Where should the intake be placed?

ii)

What form should the intake assume?

iii)

How should the intake be installed?

iv)

How will the intake be maintained? The Problem of Position.

A judicious positioning of the

intake can effect a reduction in turbidity and the exclusion of the more heavily polluted water of the surface supply.

This

highly desirable goal can only be attained by an investigation, which will be termed "location-survey" for simplicity, and which, even in its most limited form requires: a)

Qualitative analyses (Bacteriological, Chemical and Physical) of numerous samples of the supplying source.

b)

Quantitative measurements of the velocity of all prevailing currents and, in the case where the supply is derived from a stream or brook, the determination . of the discharge may be required as well.

9

c)

Field surveys, in order to establish the relative positions of currents, sample-points, sewer-outfalls and topographical details of importance.

d)

Evaluation of wind affect on the qualitative investigation mentioned under (a) above.

This procedure,

which is most important in the case of lake intakes, should provide, as a minimum requirement, a comparison of water quality under the conditions where: i) ii) iii)

There is no wind at all. The prevailing wind is at normal velocity. The contrary wind is at maximum velocity.

Such an investigation or 11 location-survey 11 is an expensive proposition.

However, the alternative, water treatment of highly turbid and

polluted raw-water, is even more so.

Modern practice favours the

elimination of the cause rather than remedying the effect.

Furthermore,

in any given situation, for reasons which are quite apart from water quality, there is a feasible limit to the number of positions which the intake may occupy. fine the

11

These restrictions, enumerated below, serve to con-

location-survey" to its proper role, and the resulting cost

to a reasonable sum. Reliability of Operation - requires that intakes be placed away from: 1.

Navigation channels in rivers and lakes.

2.

All high-velocity channels carrying floating or water logged debris, or ice (cf. Chap. III - Sec. 1).

3. Areas in which shoals or bars are liable to formas a result of a sudden change in the velocity or direction of currents. 4.

Shallow reaches of a river or stream where limited depth and

10

velocity of flow combine to produce bed-load; i.e., the movement of granular particles of significant size, bottom-hugging debris, along the river bed.

5.

Areas where the water-stage and seasonal variations may normally be expected to produce any of the above conditions.

6.

Areas where the operation or maintenance of presently existing intakes, submerged conduits or buried cables, might be the cause of mutual interference.

1.

Any location where the intake conduit, due to the relative position of intake and suction-well, would have to be excessively long or undergo severa! changes in direction.

The scarcity of

riparian property, suitable as a site for the suction-well, may present special problems in this regard. 8.

Any location where the river or lake bed presents foundation difficulties of an undesirable or uncertain nature.

All of the foregoing applies with equal force to industrial water intakes which, from the viewpoint of reliability, are similar to intakes for municipal purposes. Efficiency of Operation.

In addition to the restrictions enumerated

above, efficiency of operation requires that intakes be excluded from: 1.

Areas which are adjacent to sewer-outfalls.

In the case of

rivers and streams, the intake should be placed well upstream of such discharees.

For lake installations, there is no

equivalently simple rule.

However, it should be noted that

a common experience is to find wind and current uniting to confine pollution to the more shoreward areas.

An

intake

placed behind such natural barriers may be comparatively

11 safe from the inroads of polluted water.

As a general rule,

lake intakes are placed at considerable distances from the shore (*13, *14, *15). In point of fact, the simple rule quoted above in reference to river intakes, grossly oversimplifies the actual conditions encountered in practice.

For a number of reasons, none of which

will be discussed here, water intakes must often be placed dawnstream from sewer-outfalls.

The real problem may revolve around

the distance that exista between the two.

Admittedly, the

natural process of self-purification of streams by reaeration is significant when the distances can be measured in tens-of-miles, and the elapsed time-of-flow in days; however, in many cases, either the distance simply cannot be of such magnitude, or the accumulated pollution-load is so great that the effect of reaeration over a score of miles and several days is of little benefit.

Under such conditions, the character of the problem of

intake positioning is drastically altered to a choice between the lasser of many evils.

These regrettable facts, not often

stressed in textbooks, are becoming increasingly representative of the typical design conditions. 2.

Areas adjacent to the inlet or outlet of a tributary stream or river.

These locations are particularly prone to high turbidities

and concentrations of pollution.

3.

Shallow areas where the fetch of the wind will produce waves of such height and force as to stir up bottom silt and mud.

4.

Areas where pollution by industrial wastes produce difficult treatment problems.

Particularly undesirable are the phenols

12

and creosols which impart strong tastes to the water (*3). 5.

Harbour areas, public beaches and areas receiving discharges of surface drainage, where surface contamination and pollution may be expected.

6.

All areas where the prevalent winds, or contrary winds of maximum velocity, or variations in the velocity or direction of currents produce marked increases in turbidity or pollution.

The requirements enumerated immediately above do not, in many instances, apply to intakes for industrial purposes.

The raw-water

requirements demanded by the particular industry will determine, in each case, the applicability of the foregoing restrictions. Solution of the Position Problem.

The solution to the difficulty of

positioning an intake has already been inferred in the discussion of the problem itself.

However, in the interests of completeness, an

ampliative summary is included here. A typical procedure would be as follows:1.

Topographical maps of the general area, at a scale of one inch to the mile or better, are studied in order to obtain an overall concept of the character of the supplying source, the areas of t- ·

probably pollution and the broad limits of the anticipated "location-survey." 2.

A topographical and hydrographical map of the area within which the intake must be positioned is then compiled at a suitable scale incorporating all available information of value.

Since

most off-shore intakes are located in public domains, Governmental departments and agencies must be consulted at this time in arder

13 to comply with the provisions of legislation enacted to protect public waters.

These same departments and agencies are very

often a fund of information obtainable nowhere else.

Valuable

data on water-stage, discharge, currents, water-quality and hydrographie details can be obtained in this way. 3.

Consulting the compiled map, a number of areas can be automatically eliminated from further consideration because of the requirements of reliability or efficiency outlined previously. The remaining areas are subjected to careful scrutiny; field inspections and ground surveys are employed to complete the data required and to make sure that all salient points, such as sewer-outfalls, irrigation ditches, industrial intakes and docking areas are included in the shore surveys; hydrographie surveys of limited scope are instituted in order to verify that the information obtained from other sources is generally correct and up-to-date.

4.

The "location-survey" of the remaining areas can now be considered in detail.

Sample-points must be selected with care, bear-

ing in mind that samples must be taken at depths equivalent to those of the future intake-ports; that sampling must be performed at regular intervals over a period of time, one year being the minimum requirement in order to establish seasonal variations; that the measurement of current velocity, wind velocity and ice thickness can, and should, be obtained concurrently with the water sample.

5. The results of the "location-survey" should be studied as they become available.

In this way, trends may be discovered which

14 will permit the further elimination of some areas, the readjustment of sample points, and the closer scrutiny of the leading contenders.

11

Thus, for instance, a

coliform-density11 ma.p may

be the end-product of the "location-survey" (*11).

Here, the

average value of the number of coliform bacteria reported at each sample-point is ma.rked on the ma.p and "contours" of equal "coliform-densi ty 11 drawn thereon. used to produce contours of equal

A similar procedure may be 11

average-turbidity."

The

selection of the most suitable position for the intake is then merely a question of the economical choice between equally reliable and efficient alternatives. Before concluding the discussion on the problem of the positioning of intakes, the following remarks would appear to be pertinent to the value of "location-surveys":1.

A location-survey recently concluded for the municipal intake of Wayne County, Michigan required the analysis of over 5000 samples.

The results proved conclusively that a source of raw-

water of relatively good quality can be discovered in an otherwise highly polluted surface supply, whi ch in this ca se was the Detroit River with an inshore "coliform-density" of 50,000 (Most Probable Number of Coliforms, abbreviated as M.P.N.).

The sanitary survey pla ced the intake in a n area

where the "coliform-density 11 was a .mere 100 to 300 M.P.N., that is~ in water with a quality 150 times better! 2.

(*11)

It must be remembered that the data collected by a "locations urvey" for the purpose under dis cussi on is al so of i nes timable

15 value in the design of the filtration plant which will receive the raw-water for treatment.

Indeed, the design of the treat-

ment plant cannot be undertaken without this information. may be concluded therefore, that

11

It

location-surveys" in some

form or another must be resorted to

befo~e

progress can be

made in many problems related to municipal water supply. II The Problem of Structural Form.

Once the preferable position for the

intake has been determined, the designer can turn attention to the problem of structural form and the evaluation of the factors which will govern this facet of the investigation.

In all likelihood, one or more

of the common types of intake structures will be considered and the relative advantages and disadvantages weighed.

A more detailed descrip-

tion of each type, with accompanying drawing, will be found in Chapter IV.

For the moment, only a summary description will be given.

In the

order of their probable oost, the common types of intake structures might be reviewed as follows: TOWER INTAKES:

In general, this form of intake is a large and often

massive structure seated on the bottom of a lake or river with a superstructure which extends some distance above extreme high water level.

In some cases, this superstructure serves to house the

personnel (*19) as well as the equipment required to maintain the intake, which at times may have to be located several miles from shore in order to obtain sufficient depth or quality of water. Tower intakes possess the great advantage that any unexpected blockage or operational problem can be immediately counteracted, in addition to which must be added the further asset that, within limits,

16 water quality can be governed by selective use of inlet ports located at different depths and in various positions.

Typical examples of

such installations are the intakes for: Wayne County, Michigan, u.s. (*31) • • • • City of Chicago, Illinois, u.s •••••• City of Cleveland, Ohio, U.S.(*l9) • • • • City of St. Louis, Missouri, u.s ••••• SUBMERGED CRIB INTAKES:

Lake Michigan Lake Michigan Lake Erie Mississippi and Missouri Rivera

These structures are positioned on the

bottom of the lake or river and being entirely submerged, often to depths of forty feet or more, are inaccessible for all practica1 purposes.

The requirements of reliability must be "built-in" to

such installations, since there are few countermeasures immediately available in the event of b1ockage or operating difficulty.

The

position and depth of the inlet ports is generally fixed and the quality of water admitted cannot be controlled.

Nonethe1ess, when

properly designed, many such structures have a record of reliable operation with low maintenance cost; they can be, and have been, located at considerable distances from shore when the requirements of depth or water quality make such positioning necessary.

These

reasons have made them particularly suitable for lake intakes, although their use in rivers is not excluded provided that the velocity and bed-load conditions are satisfactory.

TYpical instal-

lations include the following: Location of Intake

Distance from Shore

Source of Su:eE1l

Cleveland, Ohio (*13) Regina, Sask. (*14) Cleveland, Ohio (*22) Kodak Park, N.Y.(*23 ) Kingston, Ontario(*24) Gary, Indiana Erie, Pennsylvania

2-1/2 miles 800 feet 3-1/2 miles 1-1/2 miles 1200 feet

Lake Erie Buffalo Pond lake Erie Lake Ontario Lake Ontario Lake Lake Erie

6000 feet

In take Pi:ee Diameter 96" 54" 12011 54" 3011 72" 72 11

17 SIMPLE PIPE INTAKES:

Under certain operating conditions, it may be

possible to modify the intake itself and eliminate the protective structure which surrounds it, thereby creating a very simple form of intake.

Examples of this type would include:

i) Extending the end of the intake pipe beyond a small supporting crib (stable bottom), the supporting piles (unstable bottom), or the shelf of very stable strata (rock, shale, hard pan, etc.) in, or on which the conduit is laid; attaching to the end of the pipe a special flared fitting or "bell-mouth" which will reduce inlet velocities.

The

Beamsville Water Intake from Lake Ontario (*12) is an example of this simple form of intake. ii) Admitting water to the intake conduit through perforations in the outer sections of the pipe itself.

In this instance,

the conduit should be supported either by low

pier~,

small

cribs, or by a piling framework in the shape of A-frames with the pipe cradled between the cross-ties; all these measures serve to raise the pipe well clear of the bottom to prevent the formation of sandbars beneath the pipe and to reduce the turbidity of the influent water.

Such an

installation is in use at Knoxville, in Tennessee. iii) Branching the end of the intake conduit with pipes of smaller diameter; fitting the ends of these branches with upturned elbows and tees.

This type of intake serves the Town of

Glencoe, Illinois, which has a population of six thousand.

18

Simple pipe intakes can seldom be resorted to unless the naturel conditions of the surface supply provide a measure of reliability in themselves.

For instance, stable bottoms,

low turbidities, freedom from bed-load, low velocity of flow are conditions which are almost mandatory; where they exist, full advantage should be taken of these assets; where they do not, the requirements of reliability will often preelude such simple solutions to the problem. SHORE INTAKES:

As the name implies, these structures are located

on or near shores where deep water is available, and they are very often designed as units housing bath the intake proper and the suction-well.

Such positioning of the intake obviates the need

for any great length of intake conduit and as a result, this form of intake system is probably the least expansive. In theory, the shore intake combines the reliability and accessibility of the intake tower with the economy of a shore installation; however, the requirements of efficiency often outweigh these advantages because: i) Water of sufficient quality is seldom available along the shores of rivers and lakes; on the contrary, these areas are generally the most polluted. ii) Water of sufficient depth is seldom available along the shore.

Almost inevitably, dredging must be resorted to

and an approach channel created and maintained especially where the supplying source undergoes substantial fluctuations in water level.

In many cases as well, the shore in-

take is more prone to the difficulties presented by ice

19 formation than are the other types of intakes, and on this consideration alone, more fully discussed in Chapter III, the shore site may have to be abandoned as impractical. iii) The availability and cost of riparian sites may be a factor; soil characteristics of the site may prove unsuitable for the size of the structure contemplatedJ protective works such as wing-walls and breakwaters, and the dredging of approach channel may present expenditures which will nullify the economy anticipated. For these reasons, the shore intake is most commonly encountered where: a) Water quality is not paramount in importance, as in the case of many industrial intakes (*16, *18). b) Water quality along the confining banks can be assured and the depth of water controlled, as in the case of many impounding reservoirs and sorne rivers. c) Natural topography provides deep water together with stable and steep banks along the "inside" curve or bend of rivers, as in the case of the intake for Danville, Kentucky, reported by Turneaure and Russell* and endorsed in principle by

Mr. Roberts (*6). d) The qualities of reliability and accessibility outweigh other considerations, as in the case of the Missouri River intake for the City of St. Louis, Missouri, reported by Turneaure and Russell*. SPECIAL INTAKES:

While the great majority of intake problems can be

solved by using modified versions of the common types mentioned before,

20 occasions do arise where the rather unique character of the surface supply, the locality, or the particular demand will suggest innovations in structural form.

In many cases, the term "innovation" is,

strictly speaking, a misnomer since the application may be new but the principle involved is not.

On this basis, the following intakes

may be considered as novel: Bellaire, Ohio.(*l5).

The intake, built in 1951, consists of

400 feet of reinforced and perforated concrete pipe laid in a trench 22

feet below the river-bed.

The Trench is 100 feet from

shore and parallel thereto; the maximum depth of water in the river is 12 feet.

After the pipe was positioned,carefully graded gravel

in increasing size was used to back-fill the excavation to within 9 feet of the river-bed elevation; at this point, an impervious layer of clay was placed over the gravel and the remainder of the trench back-filled with selected gravel removed during the original dredging.

The layer of clay acts to prevent percolation of river

water through the unconsolidated back-fill which mightclog the porous gravel filter above the pipe. The resulting "structure" is a submerged infiltration gallery, the principle of which has often been employed in the case of shallow rivers.

The intake in question has a capacity of 6 mgd and

effects a reduction in turbidity of nearly 100% - from 23 parts per million in the river water to only traces in the raw water delivered from the intake. Ware River, Boston, Massachusetts.

This example, reported by

Mr. Camp in the Handbook of Applied Hydraulics (cf. Bibliography),

21 involves the diversion of the Ware River at discharges in excess of 131 cfs; the excess is impounded b.y a low-diversion dam, and withdrawn therefrom by nine automatic siphons which act as intakes. combined capacity of the siphons is 3,100 cfs.

The

When the river dis-

charge exceeds the capacity of the whole works, water flows over the dam as in the case of a suppressed weir.

Here, the use of fully

automatic siphons is somewhat novel while the use of a diversion dam is rather commonplace. In summarizing the matter of special intakes, it may be said that they are economically practical wherever unusual factors are encountered which would make the use of more conventional intakes particularly difficult.

They are the exceptions and not the rule;

hence, they will not be of primary concern in the discussion to follow. Selection of Structural Form.

The selection of structural form for an

intake normally involves weighing the advantages and disadvantages of the alternatives referred to above.

In most instances, the following

considerations will rapidly narrow the range of selection. A.

ECONOMY.

The capital

cost of an intake is a function not only

of its capacity, but of its structural form as well.

For

example, a submerged crib is generally less expensive than a tower intake of comparable capacity.

The difference in cost

may often be so great as to preclude the consideration of the tower altogether.

In this connection, the following quotations

from Turneaure and Russell* are of interest: "The tower has the advantage over crib construction in permanence and reliability" ••• "For these reasons,

22

this form of construction is to be commended, but it is much more expensive than crib construction and is therefore suited only for the larger and more important works." "The greater number of lake intak:es are protected by submerged cribs, but a few of the largest, notably those at Chicago, Cleveland and Buffalo, have large exposed cribs. All these protect shafts at the ends of tunnels. Such cribs are much more expansive than submerged onés and require constant attendance after completion, but in the case of tunnel intakes an exposed crib is necessary in the construction of the end shaft, and to make it permanent is of great advantage in case of future extensions. It also enables water to be drawn at different levels." "The best type depends largely upon the depth of water at the intake." ••• "For economy alone the limiting depth for the exposed crib is estimated at about 40 feet." Within limits, equivalent arguments can be advanced for each of the alternative types mentioned previously.

The designer

must therefore establish at the outset the economie limitations which exist in any given case before proceeding to more detailed considerations. B.

RELIABILITY OF OPERATION.

It is probably in the decision on

structural form that the requirements of reliability are given greatest weight by the designer, because, in a general way, good positioning can only protect the intake from abnormal hazards protection from normal hazards must come from structural strength. In this regard, the following considerations are of basic importance: 1.

Foundations.

Where stable bottom conditions prevail and

bearing capacities permit, as in the case of rock or heavy clay, the structure may be supported directly on the bed of the river or lake.

In such cases, the base

23

area is often prepared by dredging a smooth eut in the bed strata to permit

11

keying11 of the structure.

Where bearing capacities are unsatisfactory, piling may be used to support the structure.

In this event,

care should be exercised to prevent underscouring of the structure by the current. In all cases, the design should incorporate features which will prevent this undermining of the foundation, the creation of overturning pressures qy deposits of silt to one side of the intake, and any radical change in the flow patterns of the prevailing currents. Bearing capacity of the soil which constitutes the supporting bottom material should be determined by actual soil analysis, and the foundations designed accordingly.

The following general values given byE. L.

Corthell in Allowable Pressures on Deep Foundations (Wiley and Sons) may be of value in the preliminary design: "Actual Pressures on Dee which showed no settlement Average Value TonsZsg. ft.

Range of the Values

Mate r i al

Number of Samples

Fine Sand

10

4.5

5.4 - 2.25

Coarse Sand and Gravel

33

5.1

7.75- 2.4

Sand and Clay

10

4.9

8.5 - 2.5

Alluvium and Silt

1

2.9

6.2 - 1.5

Hard Clay

16

s.o -

Hard Pan

5

5.1 8.7

2.0

12.0 - 3.0

24

While these figures have a general value, a close inspection thereof will only confirm the necessity for actual soil analysis at the proposed site. 2.

External Forces - Loading.

Apart from the normal hydraulic

pressures to be expected, the effects of ice, wind, waves and current must be foreseen in the case of tower intakes. Submerged cribs, while free from the direct loads of wind and waves, may still be subject to their indirect effects; even where the fetch of the wind is limited, as in the case of small reservoirs, wave action may cause turbulence to depths of 15 to 20 feet.

Submerged intakes are

generally designed to be free from the lateral pressure of ice (cf. Chapter III), but the effects of current and bed-load may be aggravated. In this connection, Professors Babbitt and Doland* state that: "Little is known about the pressures from waves and currents. Waves 10 ft. high may cause pressures of 1,800 to 3,000 psf, with the maximum pressure at mean water level. Wind pressures up to 20 psf should be allowed against the portion of the tower exposed above low-water level." With the foregoing statement as a general premise, the following information, gleaned from a number of sources which are given recognition in the text itself, may be considered applicable to intake structures.

i) WAVES Pressure:

Wave pressure has been formulated by

Mr. D. A. Molitor (Trans. ASCE, Vol. 100, p. 984, 1935) as:

25 P

=

125 Hw

2

where P = total wave pressure in lbs. Hw = height of wave in feet from crest to trough. Height:

Stevenson's Equation (modified by Molitor)

1 1 2 Hw= O.l7(V.F.) / + (2.5 - F / 4 ) where the fetch is less than 20 miles. 2 Hw= O.l7(V.F.) 1 / where the fetch is more than 20 miles.

= height of the wave in to trough. V = velocity of the wind,

where Hw

F

= fetch

feet, from crest

miles per hour. of the wind, miles

Zuider Zee Formula for the set-up of the waves above pool level, take~ into account the depth of water and the angle of approach of the waves. 2

vF s = --~~~ 1,400 D where S

cos A

= set-up

of the waves above pool level in feet. V,F = as previously defined above. D = average depth of water in feet. A = angle of approach of waves, degrees.

Height of Wave Action: v2

Ht = 0.75 Hw+ 2g (1.5 Hw) where Ht

= height

above pool level at which the force of the waves may be considered to act, feet. Hw,V =as previously defined. g = acceleration due to gravity, taken as 32.2 feet/sec./sec.

Generally, Ht is about three-eigbths of Hw. The pressure exerted b,y the wave action upon submerged intakes can only be arbitrarily estimated. All of the foregoing applies to waves on moderately deep or deep waters where the wind velocity exceeds 60 m.p.h.

26 ii) FLOWING CURRENT.

If an intake structure is con-

sidered to be somewhat similar to a bridge pier, at least in respect to the forces exerted by the flowing water, the following formula may be applicable:

where P k

= pressure

of flowing water lbs./sq. ft. which depends on the resistance offered b,y the shape of the end of the pier or "cut-water'' face. w = weight of water lbs./cu.ft. = abt. 62.5 g = acceleration due to gravity = 32.2 ft./ sec./sec. V = velocity (Mean) of current, ft./sec.

= constant

Values of "k": Cummings and Hart - Civil Engineering Handbook 4th Edition, 1959, Chap. 8. k

= 1.33

for square ends.

= 0.50 for ends with interna! angles

of 30 degrees or less. = 0.66 for circular piers. Jacoby and Davis - Foundations of Bridges and Buildings, 1941, Chaps. 9, 14. and Hool and Kinne k

= 1.5

= 0.75

- Foundations, Abullmetts and Footings, 1923, Chap. 7. for flat surfaces for round surfaces

Minimum Values: 1.

Piers subjected to Flood Flows P = 150 lbs/sq.ft. for flat surfaces P = 75 lbs/sq.ft. for round surfaces

2.

Piers in Tidal Streams P = 50 lbs/sq.ft. for flat surfaces P = 25 lbs/sq.ft. for round surfaces

27

Professor Ketchum - Structural Engineers Handbook 1924, Chap. 6. k

= 1.33 = 0.48

for prism with length to width

= 3:1

for piers where length to width is either 5:1 or 6:1, and the cutwater is plane-faced and interior angle is 30 degrees. = 1.28 for square piers. = 0.64 for circular piers. Center of Pressure:

The distribution of velocity

is recognized to be different throughout the depth of the flowing stream.

Jacoby and Davis comment

as follows: "Experiment shows that the velocity varies with the depth approximately as the ordinates of an ellipse, the maximum being somewhat below the surface. The center of pressure is commonly assumed at one-third the distance from the water surface to the river-bed. This assumption is on the safe side." In any event, the velocity gradient in depth for any given location can probably be established by the "location-survey" mentioned previously. iii) ~.

Exposed superstructures of tower intakes

must be designed to withstand direct wind-load. The following values are suggested by the authors concerned: Babbit and Doland* =

r/

M.

s.

Ketchum

= 20

~ Cummings and Har t=

20 lbs./sq.ft. on Vertical Projection -

30 lbs./sq.ft. on Vertical Projection 30 lbs.jsq.ft. on Vertical Projection

28

~ Hool and Kinne

=

30 lbs./sq.ft. on 1.5 times Vertical Projection.

~ Indicates values for Bridge Piers.

Submerged intakes are not directly affected by wind. iv) ICE.

The affects of ice formation on the operation

and design of intakes is a matter of very great importanceJ indeed, it may be considered as the great problem for intakes in northern latitudes.

For

this reason, and because it is not a problem common to all intakes, the discussion of ice as an operational hazard is not included here, but is set over to Chapter III.

However, for completeness, the pressures

produced by sheet ice as an ''externa.l force" are included under the present heading. The experimenta of Ernest Brown and George

c.

Clarke at McGill University, reported in their paper, "Ice Thrust in Connection with Hydroelectric Plant Design", (Engineering Journal, January 1932), showed that ice expands with a rise in temperature, and being plastic, it flows under sustained pressure. The rate of this rise in temperature is a very important factor in determining the pressure exerted by the confined ice sheet.

Where this temperature

change and the time required for the change can be estimated, a value for

the'~ate

of Temperature Rise

Per Hour (deg. Fah.)" can be assigned.

In their

29 report, Messrs. Brown and Clarke have derived a curvilinear relationship between this rate of. temperature rise and the "Pressure Increase in Pounds per square foot per hour" which may be expected. Pressures determined by this method apply to a confined ice sheet behind a dam or in a reservoir, and for this reason are probably subject to modification for the present purposes. Earlier experimenta by Ernest Brown, reported in Engineering for Dams, Vol. II, by Hinds, Creager and Justin (Wiley & Sons), suggests that pressures of 3,000 lbs./sq.ft. to 5,000 lbs./sq.ft. are generally adopted in this country, although pressures of 47 9 000 lbs. per linear foot have been used in the design of dams in New York State where ice thickness in excess of two feet is rare. Calvin V. Davis in his book, Handbook of Applied Hydraulics,* states that: "The magnitude of ice pressure has been variously estimated from 5,000 to 30,000 psf of contact with the vertical face of a dam. It is believed that 10,000 psf would be ample under any ordinary circumstance." v) HYDRAULIC UPLIFT.

The stability of intake structures

must be assured under all operating conditions and for this reason the effect of hydraulic uplift as one of the hydraulic forces acting on the structure is perhaps worthy of special mention, particularly

30

where the following conditions obtain: Tower Intakes - when the inner well is pumped dry to permit inspection and maintenance. Shore Intakes - when the flood stage of the supplying source reaches unprecedented levels(*l8). Submerged Intakes - when air is entrapped in the intake conduit by accident or as a result of excessive entrance velocities.

3.

Protective Works.

The positioning of the intake may be

such that the structure lies in an area adjacent to a high velocity current, in an area where seasonal variations produce such a current, or in a surface supply where occasional commercial use would increase the amount of floating debris; in all of these cases, the danger to the intake from floating or submerged objects may be important.

If such is the case, protective works in the

form of pile-clusters, should be considered.

Professor

Azerier (*5) recommends that such clusters, located upstream from the intake site, be placed at sixty degrees to the approaching current, and that each cluster comprise three piles bound together

b,y

steel collars.

These protective deviees should not be confused with the auxiliary works, which are structures required for efficient operation of the intake rather than by consideration of reliability proper. For shore intakes, protective works might include break-

31 waters and retaining walls for the approach channel, wing-walls and revetted shore lines to guide the water in a smooth curve to the intake ports as well as to arrest wave-action and erosion, and the use of log-booms or rafts to produce quiescent conditions around the inlet.

The advisability or necessity of such structures

must be determined, in each instance, by the local conditions.

A common experience for the designer is to

construct a protective deviee solving one problem, only to discover that the alteration has created severa! new and unexpected difficulties. C.

EFFICIENCY OF OPERATION.

Whereas economy and reliability of

operation involve considerations which affect structural form in the broadest meaning of the term and which produce intakes whose very appearance is indicative of the solution arrived at, efficiency of operation is more concerned with the functioning of the smaller elements of the structure, elements which are lesa apt to attract the attention of the casual observer but which are, nonetheless, of considerable importance.

Indeed, it is

questionable whether some of these elements should not be classed under the heading of things essential to reliability, rather than place them, as i s done here, among the requirements of e fficiency.

In any event, let the question be resolved by

acknowledging that where these elements fail to function properly, the result is poor efficiency; where they fail completel y in their function, the result may be equivalent to unreliability.

32

The structural elements which contribute to the efficiency of an intake are principally: 1.

The material used to weigh down the structure and maintain it in position.

2.

The racks or screens which cover the inlet ports and serve to exclude debris from the intake and intake conduit.

3.

The inlet ports which by their position, size and shape determine not only the quantity and entrance-velocity of the raw-water admitted, but also affect to some extent the quality of that water.

4.

The auxiliary structures which artifically create conditions of flow in the surface supply which are beneficia! to the operation of the intake.

The material used.

Any dense material is suitable for weighing

down an intake as long as it does not provide, at the same time, an interstitial mass in close proximity to the inlet ports.

The

widespread use of rock to weigh down intake towers and submerged cribs may not be sound practice. Azerier

(*5),

According to Professor

conclusive evidence exists to show that the

interstices of rock-weighted cribs are quickly filled with sediments and algae; when seas onal variat i ons in current and temperature occur, the sediments and algae are dislodged and enter the intake.

Concrete, reinforced or otherwise, is to be

preferred in all cases.

33 Racks - Screens.

For the purpose of this discussion, racks may

be considered to be very coarse, unmeshed, bar-screens, the bars being set one, two or more inches apart and placed in front of the intake ports.

The bars may be of cast iron, steel, wood or

reinforced concrete construction.

They are usually removable as

a unit and are set with the bars in a vertical position; their purpose is to exclude large objects from the intake. Screens, on the other hand, having a true mesh and smaller openings - two to eight meshes to the inch - require continua! inspection and cleaning.

This factor limits their location to

either the inner central wet-well of tower intakes, or, in the case of all other intakes, to the entrance of the shore-based suction-well.

Screens, however positioned, operated or main-

tained are not permitted at the intake.

On this point, all

authors agree; for this reason, they are not discussed further. Where racks or bar-screens are provided, Professor Azerier

(*5) recommends 2 inch clear openings as the minimum distance between bars; Babbitt and Doland consider one inch as the minimum, while Professor M. G. Malishevsky (as reported by Professor Azerier) would prefer that racks be omitted entirely because, in his view, either they are too effective and become clogged with debris and ice, in which case the intake as a whole is badly positioned, or their necessity and effectiveness has been eliminated by correct positioning in the first instance, in which case they have no function or purpose. The problems presented by frazil ice, discussed in Chapter III, lend support to Professor Malishevsky's views.

However,

34 Professor Azerier, commenting on these views, summarizes the approach most common today when he says: " ••• while such opinion may be considered as a general solution, it is not always true, and protective bars are often extremely valuable. It is advisable to provide same, even if they are later removed." North American practice would seemingly endorse this approach, and bar-screens are generally incorporated into the structure on the basis that their function is intermittent rather than continual, that their purpose is to guard against occasional hazards rather than to provide constant screening. The Inlet Ports.

The design of the inlet ports involves decisions

which can greatly affect the long-term efficiency of the intake. These decisions will not only predetermine the capacity of the installation, but they will serve as vell to establish the entrance velocity of the raw-water through the ports.

This

velocity is a critical factor in the design and can be used as a two-edged weapon in the attack on turbidity.

Firstly, by

designing for very low entrance velocities at the mouth of the inlet the suspended load of the surface supply can be substantially reduced before the water enters the ports.

Secondly,

because some of the suspended material will inevitably penetrate beyond the ports, it then becomes necessary to design for suffioient flow velocities to prevent settling of this material in either the inlet section of the intake, or in the intake conduit leading to the suction-well.

In other words, either the sediment

and debris should be settled out from the raw-water prior to

35 entry, or if such material has entered, it should be conveyed in suspension to the suction-well where it can be more readily removed. This general theory has been followed out in practice and the "bell-mouth" form of intake is the simple result.

Examples

of these "bell-mouth" or "horn-shaped" sections, which are more specifioally considered by McDonald (*3) and Professor Azerier (*5) than by most authors, are quite commonly reported, but the dimensions for the "flare" or "horn" are conspicuously absent from the accounts in many cases (*12, *14, *23, *24, *37). For the moment however, the question of the shape of the inlet will be put aside, and the matter of intake velocity allowed to take precedence. i)

Intake velocities.

All authors agree that low entrance

velocities are a major factor in efficiency.

The con-

sensus may be summarized as follows: (a)

Where no danger of ice formation exists, velocities in excess of 1.5 feet/sec. will nonetheless invite clogging of the bar-screens, formation of vortices at the inlet and entrapment of air.

These diffi-

culties are not often found where the velocity is 0.5 feet/sec. or less.

Professor Azerier (*5),

assuming as a premise that the approach velocity of the surface supply will never exceed 2.5 feet/sec. even for swiftly flowing rivera, recommends intake velocities between 1.0 and 1.3 feet/sec; where the approach velocities of the surface supply are below

36 1.6 feet/sec, he suggests intake velocities of 0.70 to 1.0 feet/sec.

The important contribution

by Professor Azerier in this instance is not only the helpful values given above but the recognition which he gives as well to the inter-relation between approach velocity and intake velocity, recommending in a general way that the latter be one-half of the former and suggesting, at least by inference, that excessive approach velocities can completely nullif,y the benefits anticipated by designing for low entrance velocities. (b)

Where the danger of ice formation exists, intake velocities above 0.5 feet/sec. invite the particular difficulties discussed in Chapter III.

(c)

Where large municipal installations are concerned, all authors agree that in the interest of keeping the intake within reasonable size, the intake velocities suggested above as absolute maximums must sometimes be exceeded.

In this event, provision

should be made in the design for the resulta that almost inevitably follow; i.e., increased maintenance to keep the ports free of debris and ice. A typical example of this situation is reported by Holton (*19) for a tower intake four miles out in Lake Erie, where the maintenance staff must be increased during the winter months to combat the ice problem and free the ports.

37

ii)

Shape of the Inlet Port.

Reference has already been made

to the "bell-mouth" or t'horn-shaped" sections used as inlets and the reasons for their use.

However simple the

theory involved, little detailed information has been uncovered by the vriter which would be of great practical help to the designer, or which would establish with certainty the approach used in the past to design the inlet mouth.

The following discussion may perhaps prove of

assistance in this respect. On Page 38, Figure 2 depicts a diverging tube or flared section attached to the end of an intake conduit which for simplification of analysis has been shown in a horizontal position.

The mouth of the inlet is at point

"M", where the cross-sectional area is A sq. ft., the 1 velocity is vl feet/sec., the pressure is pl lbs./sq.ft., and the center of the mouth is situated z1 feet from the wa ter sur face.

At the point "N", a dis tance L fee t from

the mouth, the area has been reduced to A sq. ft., in 2 order to provide a scouring velocity of v2 feet/sec., and the resulting pressure at "Nu is p 2 lbs/sq.ft.; the point "N" is located

z2

feet below the surface.

Under static conditions: Pl

= P2

= wzl

= wz2

where ~ = z 2 and w = weight of water in lbs/cu.ft. Where a quantity Q cu.ft./sec. is flowing in the inlet port and in the conduit, we may write Bernoulli's

... ·

,.·:.·..

':" --~---·-··-..----r---1-,.---~ . , ~-~ .. .

·.

.~

•·

.

·•.

.

,

.

.. ... ~-

Q

--

--- ·- --Q

. .·~·.. • ..., . 6

.. ..• , .. .. •

. ·:

~

.•

,

~

· r {

•· ..

Pig. 2 - Beraouilli's Bqu&tion applied to a Converging Tube.

39 Equation for th~ points "M" and "N" as follows: 2 v2 P2 vl P1 = z2 + --;- + 2g + hf + he zl + --;- + 2g where g hf

= Head

h

= Head

e Since z1

.!L -

= z2

_!g_

w

w

= acceleration due to = 32.2 ft./sec./sec.

gravity

loss in feet due to friction between points "M" and "N". loss in feet due to entrance loss at inlet mouth.

by assumption: v2 v2 1 2 = 2g 2g

(Eq. 1)

This Equation can be translated as: Difference in Pressure-Head = Difference in VelocityHead + hf + he or, equiva1ent1y: "Suction-Head"

= Difference

in Ve1ocity-Head + hf + he

For Continuity of Flow:

or,

v2

1 2g =

Q2

v2 and

2 2g

=

2 2gA 1 whence Equation 1 above becomes:

92 21!). 2 2

The above indicates that the reduction in area at point "N" produces a decrease in pressure at that point with a corresponding increase in ve1ocity.

In terms of energy,

the pressure energy at "Nu which was equal to that at "M" under static conditions, is converted into kinetic energy

40 to produce the increased velocity and is therefore lowered in value.

If the flow vere reversed through the tube,

the same Equations would be true if the values of hf and he are discounted therein.

In this case, with flow re-

versed, the flow would be subjected to gradua! enlargement in going from "N" to "M" and the high value of kinetic (velocity) energy at "N" would be gradually converted into pressure energy at "M".

The end result would be much the

same, and low pressure energy and high kinetic energy would be present at point "N".

This is exactly the prin-

ciple invoked in the design of the draft-tube connecting turbine-runner and tail-race where the pressure at exit from the runner is kept as low as possible, the velocity of the water leaving the runner is tremendous, and the draft tube by gradua! enlargement converts a part of this velocity-energy into pressure energy thereby maintaining low pressure at the runner exit and a maximum differentiai in pressure across the turbine.

This analogous function

of the draft-tube of a turbine, considered in reverse, would be an unwarranted digression from the topic were it not that Mr. McDonald (*3) suggests that the inlet ports be designed along the lines of a hydraucone, while other authors remain entirely silent on the matter.

If

Mr. McDonald 1 s premise can be admitted, even with modifications, the problem of the shape of the inlet can be tackled with greater assurance of success because detailed information does exist on the preferable relationship

41 between dimensions for a draft tube, relationships which have been found by experiment, and which minimize the turbulence of flow.

It need hardly be added that minimum

turbulence is very much the goal of the present investiga ti on. Messrs. Stevens and Davis, reporting in the Handbook of Applied Hydraulics* on the matter of draft tubes, state that: "The most efficient draft tube is a vertical tapered pipe expanding at the rate of about 8 degrees central angle and approximately 4 to 5 inlet diameters in length." Before accepting these dimensional relationships, even as a guide to the design of the inlet ports, it should be stated that Mr. McDonald's suggestion is probably open to modification on the following counts: a) Draft-tubes are very much larger than the average inlet port, the inlet diameters for draft-tubes being in the order of 15 - 25 feet. b) Draft-tubes are more subject to turbulence because of the fact that the process involves an expansion of the flow, whereas the inlet port produces acceleration through converging walls, a process which entails less turbulence under normal conditions. If these two considerations are related to the fact that turbulence is known to increase with an increase in the "flare" of a diverging tube, and if it can be assumed

42

that the elimination of turbulence in a draft-tube is of paramount importance; then, as a minimum deduction, it may be concluded that the values for the suggested by Messrs.

11

flare" as

Stevens and Davis are very conserv-

ative when applied to the shape of an intake port.

In

the only example of a converging tube supplied by Mr. McDonald

(*3),

the dimensions were as follows:

Inlet Diameter • • • • • • • = 8.67 ft. Length •

...

• • • • • • •

= 25.0

ft.

Diameter at maximum contraction = 4.0 ft. Consequen tly, -1

The expansion (central) angle = 2 x tan = abt. 11

The ratio, Inlet Diameter : Length

2.33 25.0

0

= abt.

1:3

The ratio, Diameter at Mx. Contraction : Length, (which for a draft-tube = Inlet Diameter : Length) is about 6:1. It may be concluded therefore, that this particular converging tube is designed with dimensional relationships that are not very far from those recommended for draft tubes. Since reduction of turbulence at entry to the port is almost synonymous in this instance with the maintenance of full-flow at all sections and minimum contraction of flow at entry, the values for the coefficient of discharge "C" which is normally applied in estimating the discharge

43 from short converging tubes, may also provide a clue as to the preferable angle of taper.

Professor King in his

Handbook of Hydraulics (3rd edition, McGraw-Hill) reports on the experimenta of Mr.

w. c.

Unwin, and he gives values

of "C" corresponding to different values of the angle of taper "9°".

Examination of these values, which are given

in Figure 3, Page 44, indicates that the effect of contraction becomes quite significant where the angle of taper is 11 0 or more, which would correspond to an 0

expansion angle of 22 • While the value of "C" can never be made unity because it compensates for velocity deficiency ("Cv"), as well as the effect of contraction ("C ") in the flow c through the orifice; stïll, the value of

nctt (

= CC ) v 9

can be made to approach unity by rounding the surface of the inlet to conform to the shape of the contracted stream, and in effect make the coefficient of contraction ("C") equal to unity. c

The dimensional relationships

suggested by Professors King and Russell in this regard are shawn in Figures 4 and 5 respectively on Page 44. To summarize the discussion on the shape of the inlet port, the following suggestions might be considered as guides in the design thereof: (a) The mouth of the inlet should be shaped as a converging tube.

The angle of taper which is

one-half the expansion or central angle should be less than 10° if possible.

Page 44 ..

- ··-·-

-....-

·-

- - ÛrnerJ

J/~-h/7 rél~,;t/(!,/

~aenr J/a/ns tj 8"

tl "-tl

/-{,/v~ s

t?-.97

oj" 'C ·

.

s·-

o/

Ü/.sch6~ge

·c ··

.

22

, #

tJ. ;If

./.1 "•

/5"

1.7·32

#-

30

,

tJ.b'S

( From "Handbook of Hydraulics" - KING - Jrd • .Bdition McGraw Hill ) 1

'

1

1

/ . (; z~ _J) -1 '

1

-----·---}) - . ----·-i-

j __ - - ! 1

-

.., tJ.

'zsiJ'----

Di.ensional relationships for the curvature of the aouth taten f roa "Handbook of Hydraulics" - DNGJrd. Edition - McGraw Hill

hdvr<: "5 Dimensional relationships for the curvature of the aouth taken from "Hydraulics" - Rusaell 5th Edition, Holt & Co.

45 (b) Where the taper must exceed 10°, the leading edges of the mouth should be rounded to conform to the contracted stream. (c) The length of the converging tube should be 2.5 4.0 times its least diameter. (d) The velocity of flow at the point of maximum contraction; i.e., at the entrance to the intake conduit, should be such as to prevent the settlement of silt.

Velocities of 2.0 - 2.5 feet/sec. are

generally satisfactory. (e) The effect of approach velocity to the intake structure has been disregarded in the foregoing discussion on the basis that it was negligible in value and quiescent conditions existed.

Where

such is not the case, approach velocity may well modify all of the foregoing to a very great extent. iii)

The Position of the Inlet Port.

This facet of the design

involves two questions; namely, the vertical position of the ports relative to water depth, and their horizontal position in relation to the direction of the approach current. Vertical Positioning should provide: (a) Sufficient depth of water over the ports to prevent air entering the conduit and to minimize accumulations of floating or partly submerged debris which might clog the bar-screens.

For

46

these reasons alone, a minimum submergence of three times the diameter of the inlet is recommended by Babbit & Doland* and endorsed by Steelf (b) Sufficient depth of water to minimize ice difficulties; a consideration which may well govern, and dealt with in Chapter III. (c) Sufficient depth over the ports at all stages of the surface supply.

This aspect of the problem,

discussed in greater detail in Chapter III, often creates the necessity for ports at different elevations, which in turn suggests the advisability of a tower form of intake. (d) Sufficient depth to ensure that the raw-water drawn into the ports will be at a desirable temperature during summer months (about 45° F.) without at the same time locating the ports in an isolated thermal layer where oxygen deficiency produces stagnation.

This problem is not uncommon

where the surface supply is a small lake or impounding reservoir. (e) Sufficient depth of water below the port to minimize the effect which bed-load, approach velocity and unstable bottom material might have on the turbidity of the raw-water admitted. Turneaure and Russell* advocate that this distance be 6 ft. to 8 ft., and this is endorsed by Steelf Professor Azerier considera 5 ft. to be the minimum

47 for deep rivers, suggesting an absolute minimum of 1.5 ft. to 3.0 ft. where bottom stability is excellent. Horizontal Positioning of the inlet mouth should provide the following: (a) A minimum accumulation of debris, ice and bedload against the bar-screens.

For this reason,

most authors favour placing the ports in the downstream face of the intake structure.

As a

second choice, a lateral location seems to be preferred. (b) A minim1rm obstruction to the back-flushing process of conduit maintenance.

This consider-

ation may be of primary importance where such back-flushing is the only means immediately available to clear obstructions from the intake mouth.

Again, on this count, the preferred

position would appear to be on the downstream face of the structure. (c) A maximum reduction in the suspended load or turbidity of the raw-water before it entera the ports.

All other considerations asi de, the

possibility of r educi ng turbidity by a judicious positioning of the inlet mouth in relation to the direction of the flowing current, would appear to be worthy of investigation; s ince both approach velocity and intake velocity are thereby involved,

48 it might be assumed that a preferable relationship between these two entities, in magnitude and direction, actually does exist. Although Professor Azerier infera that such is the case, the matter is not discussed by other authors, and the latter presumably approach the problem of horizontal positioning on the basis that the considerations outlined under (a) and (b) above will almost inevitably govern.

Partly

for this reason, and partly because the matter would appear worthy of further examination, the inter-relationship of intake and approach velocities in effecting turbidity reduction was made the subject of the experimental investigation reported in the present Thesis.

A detailed

account will be found in Chapter V and following chapters.

The Auxiliary Structures.

Under this heading, only those common

structures or structural modifications will be discussed which contribute to the efficiency of operation of the intake, thus leaving to Chapter III the consideration of special structures required to combat the specifie conditions of ice and large variations in stage of the surface supply. Thus limited, the discussion becomes one which is largely concerned with the maintenance of ideal flow conditions adjacent to the intake structure. might be significant:

In this connection, the following

49 i) The introduction of any structure into a surface supply tends to interfere with the pattern of currents or hydrological regime of the source.

The effect of this inter-

ference will depend on the comparative aize, shape and position of the structure in relation to equivalently important features of the supplying source.

Thus, for

example, the size and form of a submerged crib may be quite important in a small river, but of negligible effect when placed in a large lake.

Suffice to say that the

interference of these structures with the normal regime of the source should be minimal.

To this end, Professer

Azerier (*5) recommends that: (a) The shape of submerged or exposed cribs be somewhat streamlined with the length of the crib set parallel to the current and the ends tapered. The length to width should be about 3:1 and the interior angles of the tapered ends should not exceed 50° to 60° (cf. Fig. 6, Page 51). (b) The intake conduit should not be exposed along the bed of shallow rivers since the ridge so formed will !essen the live cross-section and seriously disturb the hydrological pattern.

In

general, the intake conduit should be buried in a shallow trench if at all possible. ii) Whereas the construction of auxiliary structures to combat a special problem, say turbidity, has sometimes proved successful, these structures should be used sparingly

50

since their operation is often unpredictable.

For in-

stance: (a) Where the suspended load of a river or stream flowing with more than moderate velocity produces a raw-water turbidity of undesirable proportion, the creation of a stilling-basin, either by building a diversion dam (small rivers) or intake reservoir (large rivers) may suggest itself.

The purpose of

such structures, sketched in plan in Fig. 7 and Fig. 8, Page 51, is to reduce approach velocities and encourage settling of the suspended load, and to varying degrees this condition is generally obtained.

However, since the deposited sediments

are often quickly accumulated and are rather easily shifted by currents within the basin, and since these deposits are not easily removed without causing turbulanœ and resuspension, there is a tendency for them to "creep" forward to the more quiescent reaches of the stilling-basin directly in front of the intake ports.

For

this reason some attention in the past has been given to the design of structures immediately adjacent to the intake mouth which would provide a sudden increase in velocity and prevent these accumulations of deposited material.

Thus,

(b) The Ejector-type of intake places the intake mouth in a narrow gap between two cribs (or equivalent

Page 51.

1

1

\

Fig. 6 -

Streamlined Shape for Cribs ( Professor Azerier ). Dimensions are in metres.

Fig. 7 -

Low Diversion Dam ( OVerflow Type ) with Gate-Controlled Sluiceway to regulate water leve! on a small river of variable stage. Gate can be raised to provide adequate submergence of intake ports and lowered to flush out deposits.

-

---

Fig.8 - Intake Basin adjacent to river of high velocity. ( See also Fig.lO, Page 80)

52 structures) situated at the outlet from the stilling-basin.

Here, in theory, the decreased

cross-section produces an increase in velocity which effectively scours the area surrounding the intake mouth.

Such an installation is

depicted in Fig. 9, Page 53.

Professor Azerier

in discounting the practicality of this form of intake, is reinforced in his opinion by the view of his colleague Professor Malishevsky who points out that to be at all effective the structures concerned must function, not as a suppressed weir, but as a constriction of the flow channel itself.

Whereas this in itself is seldom pos-

sible, and whereas the placing of two such cribs, closely gapped, on the bottom of the average river in order to obtain an increased velocity of flow between them, is almost inadmissibly futile; it would appear that the application of the ejector principle is very limited indeed. If the prerequisite conditions for an adequate defence against ice be added to the argument, the ejector-type intake is virtually eliminated from further consideration. iii) Under the heading of Auziliary Structures must be included certain brief remarks concerning the inter-operation of the intake proper with the other two elements of the intake system; namely, the Intake Conduit and the

Page 53.

f

I

J_

j

1

1

J.oJ ...

Pig. 9 - Bjector Type Intake discussed by Professor Azerier. The Ejector Intake illustrated is constructed of timber bolted to supporting wooden piles. The piles are of small diameter and short in length, and · are grouped in clustere at the corners of the cribs which,in each instance, are tightly compressed by peripheral steel strapping. n1e dimensions shown are in metres.

54

Suction-Well.

Without embarking on a digression which

would contradict the general preface to Chapter I, the following observations would appear pertinent: (a) Where possible the Intake Conduit should be laid along a gradually rising slope as it proceeds shoreward from the Intake proper.

This measure

will help greatly to eliminate entrapped air from the pipe, improve efficiency of back-flushing, and reduce the depth of excavation required for the suction-well and for the trench in which the conduit itself is laid. (b) However, none of the reasons given in (a) above would justify allowing the conduit to rise above the hydraulic grade line, or to have the shore end of the pipe above the Low Water Level of the supplying source.

More specifically, and perhaps

more conservatively as well, it is generally desirable to have enough head available, even when all pumps are operating, to ensure that self-priming conditions will exist in the SuctionWell; or, in other words, that the lowering of the water level due to friction loss does not lower the f ree-water surface in the sucti on-well below the level of the pump casings.

In calculating

the head losses throughout the intake system, it should be r emembered that the hea d l oss acr oss the screens in the suction-well may amount to

55

1.0 ft. or more, and that the bottom of the suction-well must be set some feet below the suction inlet to the pumps. III The Problem of Installation.

The manner of installing an intake once

position and structural form are known, is largely decided by weighing the advantages of alternative constructional procedures, selecting the most economical and expedient of these alternatives, and then modifying the design, where necessary, to insure that the installation can be carried out with efficiency and safety.

Except for a general summary,

discussion of the many procedures and methods which have been resorted to in the past would be beyond the scope of the present treatise, and for this _reason the following outline presents the larger considerations only: A.

The intake installation problem is intimately bound up with the problem of constructing the intake conduit.

The method

of connecting these two elements of the intake system often presents a major problem in the constructional procedure. Thus, for example: i)

Where large intakes are concerned, especially tower intakes, the conduit diameter may be of sufficient size to warrant tunneling.

Where diameters are six

feet or better, soil conditions suitable and gas pockets absent, a drift may be eut between a shore shaft and one directly beneath the site of the intake structure itself.

The latter is often left with an

unexcavated "plug" or "core" to seal the tunnel until

56 such time as the overlying intake structure has been properly seated and or

11

11

keyed 11 to the bottom, the "core"

plug11 being then removed by excavation from within

the dewatered tower or caisson. ii)

For smaller intakes, particularly submerged cribs, the intake conduit is commonly a pipe with flexible joints which might, for example, be assembled on barges and then laid in the trench or on the supporting piers or cribs.

Alternatively, the sections of pipe might

be assembled on shore, bulkheaded and floated into position above the trench where, by slowly filling with water, they could be lowered into position.

In bath

methods, the final connections and inspection would probably require divers who, in one way or another, are almost indispensable when installing an intake. The methods of assembly and construction of a submerged pipe conduit (*27, 28) are tao numerous and detailed to be pertinent here; suffice to say, that any method that will assure the conduit being laid economically and safely is worthy of consideration. B.

The procedure used to install the intake structure proper will be largely determined by the distance from shore, the size of the structure and the depth of water to be encountered. i)

Thus,

Large intakes, commonly equipped with floatation chambers and bulkheaded, are most often constructed in drydock or shoreward areas (*21).

From here, after

floating they are towed to the site and sunk into

57 position on a previously prepared base. Tower intakes are designed either as "wet" or structures.

11

dry 11

In the "wet" tower, water fills the in-

terior of the structure to the level of the source of supply.

Although equipped with floatation chambers,

"wet" towers are often designed without an enclosing base slab, the walls of the structure being supported directly on the bed strata.

At times, where this strata

is largely impermeable and of good bearing capacity; e.g., heavy clay, an "embedding shoe" or peripheral steel ring of tapered cross-section is incorporated into the footing of the tower wall.

As the floatation

chambers are filled and the weight of the intake comes to bear on the cutting edge of the "shoe", the entire structure can be evenly settled through the bed strata and

11

keyed 11 thereto by excavation from within.

In "dry" towers, all water is excluded from the interior other than that which is contained within the intake pipes proper. ''Dry" towers are commonly designed with an enclosing base slabJ where poor bearing capacities require such a slab, or where special foundation work such aspiling must be undertaken, the "dry" tower is often selected although it is generally more costly to construct and subject to floatation when completed. ii)

Smaller intakes, including submerged cribs, are often assembled on shore or in shallow water, towed to the

58 site and slowly sunk into position by filling with concrete or rock.

Final connections and

is accomplished by divers.

11

fitting-outn

More recent design would

seem to faveur the use of sectional prefabricated structures or open caissons which incorporate the intake as the bottom section.

The entire caisson can be

floated to the site, sunk into position, fitted-out and connected to the conduit.

When the work of installation

is completed, the upper sections are removed and may be stored for future use.

Somewhat analogous procedures

are reported by ~œ. E. I. Brown (*17) and Professer Azerier. iii)

Shore Intakes are built in a number of ways with two methods commonly predominating: a) Where sail conditions and current velocity permit, a cofferdam of earth, rock and dredged material is constructed adjacent to the shore site.

At times, sheet-piling is used to re-

inferee the cofferdam and reduce the thickness thereof.

Once constructed, the interior of the

cofferdam is dewatered to permit laying of the foundations. b) Where soil conditions or current velocity make a cofferdam unwise or uneconomical, the shore area immediately adjacent to the site can often be filled to create a pier-like extension of the shoreline.

On such artificially created or

59

"filled11 ground, the foundations for the intake can be laid in sections or ulifts 11 with a steel ring or "embedding shoe 11 at the base of the footing.

By uniform excavation within the

periphery of the foundation the entire structure can be made to settle gradually and evenly, first through the unconsolidated fill, and then finally into the original underlying strata of the shore bottom.

Excavation is discontinued when the

foundations reach the desired level, at which time the walls are carried up to grade, the remaining fill surrounding the structure removed and the installation completed. iv)

River Intakes - Auxiliary Structures.

Where dams must

be constructed to divert or impound water, or to increase the depth thereof, the construction of the intake proper is generally a subsidiary problem in the overall scheme.

One or more of the methods described above may

be applicable, but in almost all cases, the specifie conditions of site, flow and size of installation are of such importance as to make generalization impossible. IV The Problem of Maintenance.

Once installed, an intake should be char-

acterized by trouble-free operation.

This opt i mum condition is obtained

as a result of a realistic approach to the question of maintenance which, from the very outset, should be foreseen and provided for in the following ways: a)

Avoiding as many of the foreseeable difficulties as possible by

60

correct positioning and design of the structure. b)

Providing adequate countermeasures against those probable contingencies which cannot be economically or effectively controlled or eradicated in their entirety.

In deciding which of these two approaches should govern, the question of the accessibility of the intake will be of paramount importance - for this point is very clear, endorsed by all authors and almost a selfevident theorem - facility of maintenance is largely a matter of accessibility.

An equally important corollary suggests that maintenance

operations can seldom be expected to compensate for deficiencies in design or for poor positioning of the structure. Within the framework of such general limitations on the applicability and effectiveness of maintenance operations, it must be acknowledged that sorne form of maintenance is normally required from time to time, and consequently, provision must be made to allow such operations to be carried out with efficiency and thrift.

Briefly, the problem is that of

preventing accumulations of ice, weeds, silt and debris from interfering with the free flow of raw water through the intake and intake conduit. Accumulations of ice are part of a larger problem, discussed in Chapter III, and not considered here.

For the moment, two contingencies will be

dealt with; namely, clogged bar-screens and detritus deposits in the conduit. Accumulations of debris occur, for the most part, at the bar-screens covering the entrance ports of the intakes.

Waterlogged branches, water weeds,

rags, paper, garbage and trash of all kinàs, have in the pas t found their way to the fa ce of the bar-screens, there to become l odged and effectively reduce the cross-sectional area of the opening.

This reduction inevitably

61

increases the entrance velocity and suction head at the port, which in turn produces vortices attracting further submerged material and creates pressures which will cause the smaller trash components to become firmly lodged in the openings between the bars, or to Pass through these openings and into the conduit proper.

In most instances a partially blocked or

clogged port produces a sizable increase in operating head loss and warning deviees can be installed to advise of this condition.

In the

case of towers and shore intakes, the difficulty can be combatted in a number of ways, including manual raking and

11

back-flushing11 of the port

in question using a special pump provided for this purpose.

In a few

instances, a blast of compressed air in close proximity to the clogged rack has created a surge sufficient to dislodge the obstruction.

Here

again, the factor of accessibility is a major asset because, in all of the above methods, clearing operations can be carried out at close

ran~

using equipment handy to the scene of the difficulty. Where the intake is submerged, the most obvious solution, and often the only remedy short of diving operations, is to conduit.

11

back-flush 11 the entire

Such a procedure, although frequently mentioned, is not always

as easy or as effective a solution as might first appear.

Nonetheless,

it is probably true that most submerged intakes are designed in such a way as to permit back-flushing should the need arise.

Since these

smaller, less expensive intake units are the primary concern of this thesis, the subject of back-flushing becomes of more than passing interest and will be discussed in some detail towards the end of the present chapter. Apart from the difficulty presented by clogged bar-screens, maintenance operations must contend with the elimination and removal of

62

unwanted deposits of silt, mud and detritus from the intake and intake conduit.

Under normal operating conditions, such deposits are prevented

by the self-scouring flow velocity within the conduit.

For shore intakes,

this problem is generally minor since the conduit is short in length. For towers, particularly "dry" towers where mesh screens are installed and serviced as part of the purification process, the mechanical straining of suspended matter obviates the problem entirely.

"Wet" towers,

like all submerged cribs, are quite frequently devoid of all straining deviees, screening being delayed until the raw water arrives at the suction well, and in these cases, particularly where intermittent flow conditions in the conduit are encountered, deposits of detritus may accumulate.

Under such circumstances, back-flushing of the conduit is

commonly advocated, and published reports (*3, *5) indicate that this procedure is effective.

Few alternative remedies are suggested by the

authors concerned, although it probably can be assumed that hydraulic pipe cleaners and mechanical scrapers are used where practicable. Back-flushing.

It would appear from the foregoing that back-flushing

facilities are important assets of an intake system.

Especially in

the case of submerged intakes, these facilities are most frequently provided for in the design of the suction-well and treatment plant, and are therefore not properly pertinent to the discussion here. Nonetheless, for completeness, it may be stated that the major problem involved is to obtain a reversa! of flow of such velocity that the resulting force exerted at the intake port will be sufficient to dislodge and clear away the accumulations mentioned before.

If one assumes

that the back-flow velocity in the conduit should be two or more times the normal velocity of in-flow; that such a velocity mtist be obtained

despite correspondingly high frictional losses, then the use of a high capacity pump, or the discharge from an elevated reservoir suggests itself. In a fully serviced municipal water supply aystem, the pump with the largest capacity is generally the one used to back-wash the filters in the treatment plant, and therefore its usetulness is sometimes

ex~

tended to include occasional back-flushing of the intake conduit.

In

auch cases, the flushing water is drawn from the clear well of the treatment plant; hence, the assignment of auch a dual role to the back-wash pump frequently involves problems of reserve storage, pump efficiency and cross connections between treated and untreated water mains. Back-flushing the intake by discharging from an elevated storage tank or reservoir circumvents few of the problems mentioned before; in addition, where the purpose of the elevated storage tank is primarily to equalize pressures, its location will often be at a point furthest from the intake making it quite useless for the purpose in question. In some instances, such as that reported by Mr. McDonald (*3), the pumps in the suction-well have been used to back-flush the conduit and intake.

Although details are lacking, it might be inferred that

the combined capacity of the pumps in the suction-well would be sufficient to produce the high flushing velocity required.

While

auch a condition could conceivably exist, it is difficult to imagine an installation of this kind where the pumps would not be overdesigned for normal operation; furthermore, costly piping arrangements would still be required and the problem of cross connection between raw and filtered water lines would remain unaltered.

64

As far as the intake proper is concerned, the effectiveness of back-flushing may be enhanced by attention to the following points:(a) Whereas an inlet-port is correctly designed as a converging tube in order to reduce entrance-velocities, discharge under reversed flow produces a large energy loss at exit from the port.

Although a clogged bar-screen will somewhat mitigate

this head loss, some reduction in back-flushing velocity must be accepted as inevitable.

For this reason, it becomes

i~

portant to conserve the potential energy of the back-flushing flow in every way possible.

Since frictional head losses

increase rapidly with an increase in velocity, the intake design should feature a minimum number of sharp bends or transitions in flow. (b) Where the intake is designed with several ports, the arrangement of the various inlets should insure that the debris and detritus cleared by back-flushing from one port will not foul another further downstream. (c) Although intakes should be positioned away from high velocity channels, it is often difficult, particularly in the case of rivers, to place them in areas entirely free of current.

In

point of fact, it is often the carrying power of these smaller currents which produces the accumulations on the bar-screen. Consequently, designing inlet ports to face upstream appears not only to invite the difficulties mentioned before, but reduces as well the effectiveness of back-flushing operations because the reversed flow must run counter to the natural current. Accordingly, inlet ports should preferably face downstream or be

65 pointed outwards at right angles to the current.

On the whole,

little information is available to confirm or deny the validity of this argument. in this regard.

Professer Azerier (*5) is a notable exception In his view, back-flushing operations are less

frequently required and more effectively carried out where the port faces downstream, and in a number of his illustrated examples, this preference is quite evident. The problem of back-flushing as well as other problems relating to maintenance are not insoluble.

A cursory recital of the difficulties

to be encountered in no way implies that effective measures cannet overcome operational disabilities.

On the contrary, careful selection of

pumping equipment, provision of by-passes in the piping arrangements, and attention to the design of the intake ports, will often readily solve the maintenance problem.

The point stressed here is simply that

such a problem can be more effectively solved by foresight than by wishful hindsight. The Common Factors - Conclusion.

In the foregoing chapter, the four

problems of position, structural form, installation and maintenance have been investigated in some detail and with particular reference to the smaller, less expensive intakes.

Although these four factors

are commonly encountered in the design of any intake, their relative importance, one to the other, depends on the specifie conditions surrounding the installation concerned.

Seldom are all four equally im-

portant, and qui te frequently one factor predominates and governs the design. In a similar way, two special conditions will now be discussed that often exert a governing influence in all aspects of the problem considered so far.

66

CHAPTER III "The Special Conditions - lee and Variable Stage" Of all the natural phenomena affecting surface water supplies and the intakes therefrom, none are probably more important or more easily foreseen than the formation of ice and the occurence of flOods..

Where

either or both of these conditions exist, their effect on intake design is no less severe than the change which they produce in the hydrological character of the supplying source.

Indeed, it may be said that these two

effects are roughly proportional, and in a number of cases this is tantamount to admitting that these special conditions govern the design. In the following pages, these conditions and their inevitable consequences are briefly and separately investigated. SECTION I - ICE Its Importance. In a paper presented before the Canadian Section of the American

Water Works Association, Mr. Norman McDonald (*3) summarizes the opinion of many authors on the importance of the ice hazard, saying: "Unlike most parts of the water works system the intake is rarely constructed in duplicata and its failure generally means the failure of the supply of water. Its design and construction is therefore of great importance. Should the intake fail through the inlet's being blocked and the available means for removing the blockage not be effective, the operator is faced with drastic action to secure water. In two municipalities taking water from Lake Ontario,.dynamite was used to open a hole in the intake pipe near the shore when ice could not be dislodged from the inlet. Such action appears drastic but under the circumstances may have been warranted."

67

"The greatest hazard in the operation of intakes in Canada is that due to frazil ice. On bodies of water subject to wave action, in which the sheet ice is broken up during its formation, frazil ice is formed and may be drawn into the intake by eddies or other currents. With low temperatures and high winds this ice may be formed in such quantities as to shut off completely the supply of water." Classification of lee.

Three kinds of ice are commonly mentioned in re-

ports on this subject: i)

SHEET !CE - the common crystalline type that forms on the surface of lakes and on the quiescent reaches of rivers whenever the temperature reaches 32°F.

ii)

FRAZIL ICE - a name coined from the French Canadian term for the splinter-like cinders of a blacksmith 1 s forge - is a surface ice that is not allowed to form into a sheet due to turbulence in the water.

Borne along by wave action or current, frazil ice

appears in the water in the form of needle-like crystals, or in a flake form often mistaken for snow, or as formless slush.

At

temperatures of 32°F. or slightly below, frazil ice is rapidly formed and a very viscuous mixture of water and ice resulta; this in turn produces a rise in water level and a reduction in surface turbulence whereupon surface ice, enclosing masses of frazil ice, is formed. iii)

ANCHOR ICE - anchor ice is produced by excessive radiation and in this respect its formation is akin to that producing frost forms on vegetation.

On clear cold nights, rocks and boulders

lying on the bottom of clear streams tend to radiate earth heat and in this process they become super-cooled and coated with ice.

In the morning, the slight temperature rise accompanying

dawn causes this anchor ice to loosen, float to the surface

68

and drift downstream in lumps or blacks that will often have grave! and bottom material adhering to the underside. Provided that the intake is positioned away from areas where ice jams are liable to fill the water to considerable depth, and that the intake, if exposed, has been designed to withstand the pressures involved, sheet ice seldom interferes with normal operating conditions; on the contrary, its formation is to be encouraged since its presence acts as a deterrent to frazil ice penetration and anchor ice cannat form beneath its insulating caver. In contrast to sheet ice, frazil ice is unquestionably dangerous

and must be considered as one of the greatest hazards to normal operation.

On this matter, the explicit view of Mr. McDonald, quoted earlier, finds conforming opinion on all sides and to such an extent that much of the discussion which follows is directed at eliminating this threat to intake performance. Many authors, including Fair and Geyer, Turneaure and Russell, Babbitt and Doland, would appear to hold anchor ice as well as frazil ice to be jointly responsible for many of the blockages reported.

While this is

very likely true in sorne cases and sufficiently pr obable in ethere to warrant the generalization inferred by the texts, the fact remains that, of the two, frazil ice is the more dangerous.

Recalling to mind that the

formation of anchor ice depends on a number of climatic and environmental conditions; that even when these conditions are met, its existence is of short duration, it hardly seems likely that anchor ice would be primarily responsible for the serious blockages contemplated here. The general view on this matter is perhaps best summarized by the findings in a number of instances (*8, *29) where severe ice problems have

69 been investigated.

In these cases, the weight of evidence clearly im-

plicates frazil ice as the main offender whereas anchor ice appears for the most part as a contributing and complicating factor. m~

Therefore, it

reasonably be assumed that if the menace of frazil ice can be

successfully combatted, the hazards presented by other ice forms will have been overcome as well.

This is the approach adopted in the follow-

ing discussion on the counter-measures available to prevent ice from crippling the intake. Frazil lee and the Scope of the Problem.

Before discussing the alterna-

tive measures that can be invoked against frazil ice, it would seem advisable to establish by quick review what is known of its action and character.

Frazil ice has been defined as a surface ice which is pre-

vented from forming a sheet because of turbulence; consequently, if the turbulence be reduced, frazil can be expected to form sheet ice.

This

is indeed true, and little frazil ice is encountered beneath sheet ice provided that undercurrents do not exist.

An absnece of current is very

important since ice is only slightly lighter than water and it requires but little current to cause it to flow with water, particularly when it is broken into small pieces.

Where frazil ice is borne beneath sheet

ice by currents which are subsequently dissipated in the quiescent conditions, frazil ice will rise and attach itself to the sheet greatly augmenting the thickness thereof. Frazil ice is almost classically described as needle-like in form although it is generally

recognized that other forms are possible

depending on the manner of its formation.

Thus Baylis and Gerstein (*8),

in their report on the clogging of pumps at a Chicago intake, remarkt

70 nDiffering from the needle-shaped formations reported by many observers, frazil ice on the impeller was composed of thin, perfectly round, flat crystals, shaped somewhat like fish scales. After initial adherence to submerged surfaces, crystals grew rapidly from 1/8 inch to as large as 4 inches in diameter and 1/32 inch thick. With irregular edges, the flat surface stood out perpendicularly from the metal to which a section of the edge was attached. Clusters of prystals tended to form into flower-like rosettes on submerged surfaces in the intake basin and accumulated on impellers in dense, opaque ice masses." It would therefore appear difficult to recognize frazil ice by shape alone and other criteria must be used.

Such a prominent character-

istic is not difficult to find for all observers agree on the tenacity with which this form of ice will adhere to metallic abjects, and many operators hang short lengths of so-called "tell-tale" chain in the waters adjacent to the intake ports thereby obtaining confirming evidence of the arrival of frazil ice.

This propensity of the frazil crystals to coat

and agglomerate on submerged metal surfaces, which can probably be termed its most dominant characteristic and

identifying mark, results

from a very small temperature imbalance between the ice and the metal abject at the moment of contact.

An accurate, yet simple, account of

the process involved has been described by Messrs. Stevens and Davis in the Handbook of Applied Hydraulics and included here: "Except under rare cases of supercooling, when it lies undisturbed, water will change to ice whenever its temperature is lowered to 32°F. A mixture of water and ice cannat fall below 32°F. until the water has been converted into ice. In such a mixture, the slightest tendency toward lowering the temperature below 32°F. increases the portion of ice, because heat is being transferred from the mixture to calder media; conversely, the slightest tendency toward a rising temperature will reduce the portion of ice, because heat is being absorbed from the warmer media." "In the condition of losing heat, ice will adhere to any body in contact with the ice-water mixture that is absorbing heat, e.g. racks and gates. In the condition of absorbing

71

heat, ice will not so adhere. If then the temperatures of the racks and gates can be raised ever so slightly ( a hundredth of a degree) above that of the mixture, ice \-lill not adhere to them." In order to properly correlate the quotation above to the specifie

question of frazil ice, it should be recalled that by definition frazil ice is the product of turbulent, not undisturbed, surface conditions and hence it is most frequently accompanied by sorne supercooling however minute.

Thus, for example, there is no contradiction between the out-

line above, which applied to ice in general, and the actualities of temperature reported by Baylis and Gerstein (*8) where, in their words: "Precise measurements taken in the intake basin during frazil ice occurences revealed temperatures below 32°F., indicating supercooled water. For example, a water temperature of 31.893 degrees F. was observed on February 15th, 1946. Il On the contrary, the remedia! measures instituted in this instance corroborate the theory enunciated earlier inasmuch as the report specifically states: "As studies showed that frazil ice could be prevented by raising water temperature about 0.10 degree F., a pipe system for injection of steam into pump suctions was installed in the fall of 1946." The foregoing remarks and the excerps quoted are little more than a synopsis of the more important points applicable in this instance.

A more comprehensive and authoritative treatise will be found

in a number of texts such as Barnes' lee Engineering (Renouf Publishing Co., 1928) where the subject is pursued at a length not permissible here. Countermeasures Available.

However limited, the preceding outline is

sufficiently complete to suggest certain stratagems for the defense of intakes against the hazard of frazil ice.

In order to maintain the

72

continuity of approach developed in Chapter II, these countermeasures are discussed in relation to the four common problems already considered. A.

Effect on Position.

The avoidance of currents, channels and areas

of high turbulence becomes more important than ever.

As turbulence

normally decreases with an increase in depth, deep water in itself provides a measure of protection.

On this basis alone, 25 feet is

mentioned as the minimum depth by Steel* although he prefers, in common with ether authors, depths of 30-35 feet as a general rule. Consequently, i)

Submerged cribs in deep water gain in favour, particularly where sheet ice will overlie the site.

Three

examples reported by Turneaure*, all with submergence of more than 40 feet, are ice-free. ii)

Exposed cribs - Intake Towers, estimated by Turneaure* to have a limiting economie depth of 40 feet, become expensive structures and ice difficulties remain atical.

proble~

Reh (*9) claims that the accessibility of the

exposed crib compensates for the uncertainty involved and cites the example of the exposed crib at Toledo, Ohio, where the ports are submerged less than 25 feet in Lake Erie.

Here, the sixteen ports, each 10 feet square

and with an intake velocity of 0.175 fps., have encountered trouble from frazil ice and drifting sheet ice.

However,

the situation has been remedied b,y the maintenance facilities available to an exposed crib, which in this instance, projects 10 feet above water level.

More recent

73

design not only appears to confirm Reh 1 s view, but also appears to take the argument one step further suggesting that since the economies of the situation make cost proportional to height, the protection of towers and exposed cribs from ice difficulties by mere submergence of the ports is both an expensive and unnecessary course of action.

The more logical alternative is to adopt a tower

height that is economically feasible, assess the probable ice hazard that results, and equip the intake accordingly. In this way, full use is made of the asset of accessibility and heavy initial expense is avoided.

The recently con-

structed intake for Wayne County, Michigan, reported by Hardin (*31) is apparently designed on this principle. Although located in 35 feet of water, the exposed crib has ports that are only five feet below the minimum water level and ice difficulties are fully expected to occur; consequently, trash racks have been equipped with heating deviees, compressed air has been made available for clearing operations and other alternative methods for maintaining ice-free operation have been incorporated into the structure. iii)

Shore intakes, are possibly the most seriously affected by the additional depth requirements.

Since construction

and maintenance of forebays are costly items at best, port submergence rarely exceeds the 25 feet deemed sufficient for normal conditions and even this depth of water is often only obtained at the cost of considerable dredging or by

74 dam construction.

As a result, the deepening of forebays

or the construction of dams, simply to avoid ice difficulties, is generally too expensive a remedy to be practical and, as in the case of the exposed crib, other measures are used to overcome the hazard involved.

These include

the use of log booms or rafts to create quiescent conditions around the inlet thereby permitting sheet ice to form and obviating the problem of anchor ice.

With sheet

ice covering the forebay, frazil ice penetration is minimized.

Some difficulty in this regard will probably

still exist, however, because the currents in the forebay are hard to control and often unpredictable countercurrents develop with shifting winds; under these conditions, one or more of the remedies discussed later under maintenance must be applied. In summary then, ice affects the problem of position

inasmuch as it makes all the prior requirements more stringently binding and increases the desirable depths of ports.

With the possible exception of the submerged

crib, increased submergence imposes an economie load that may often be out of proportion to the benefit gained; henc~

particularly where the intake is accessible, other

means of ove rcoming the probl em should be considered. B.

The Effect on Structural Form.

Apart from the pressures to be expected

on the superstructures of tower and shore inta kes, a sharp reduction in inta ke velocities is required to eliminate eddy currents adjacent to the inlet.

Whereas velocities of 1.0 - 2.0 feet/sec. may be

75 justified under normal conditions, most authors consider 0.5 - 0.25 feet/sec. to be the desirable range when danger of ice exists.

Again,

this reduction is affected by the depths involved and in the following table, which lists the intake velocity and port submergence for a number of installations, the relationship of the two is quite apparent: Port Submergence and Intalce Velocities Source

Dept) (_

Intake Velocit:t: (ft./sec.)

14

0.175

20

0.40

24

0.35

- Prov. of Ontario

38

0.70

- Prov. of Ontario

0.75

- Prov. of Ontario

53 60

1.50

- Prov. of Ontario

63

2.60

Lake Erie - Toledo, Ohio

:/

Lake

- Prov. of Ontario

Lake Erie - Sandusky Intake River Lake Lake La.ke

(:J

:J :J :J :J

Data reported by McDonald (*3) - exact source unnamed). It should not be concluded however, that depth is the only

factor involved or that the tabulation above establishes a simple proportion applicable in all cases.

On the contrary, surface

turbulence, size of intake and the thickness of the. sheet ice cover may well produce some apparent contradictions.

For example, the in-

take at Buffalo Pond Lake near Regina, Saskatchewan, has a port velocity of 0.50 feet/sec. despite its submergence of less than 11 feet; conversely, the Nottingham intake for Cleveland, Ohio, has a port velocity of only 0.25 feet/sec. although its submergence is nearly 45 feet. Where towers or shore intakes are concerned, the lower limits imposed on entrance velocity require an increased area of cross-

76

section for the ports, and were a circular shape maintained, a reduction in submergence would result somewhat vitiating the margin of safety gained.

The same argument would apply to a submerged crib

where the ports are set in a vertical plane and kept bell-mouth in form.

In order to maintain maximum submergence and still comply

with velocity restrictions, the ports of towers and shore intakes are often designed in oval or rectangular form with the long axis horizontal, while those of submerged cribs are kept bell-mouth in shape but made to face upwards, the rim of the port being set flush with the top of the crib, and the connection between intake and conduit being effected by a ninety-degree, long-radius elbow.

Whereas

these va riations in shape of the intake ports are most frequently encountered in connection with i ntakes subj ect to ice problems, it is obvious that their application is not limited to such cases alone; on the contrary, such variations in shape are commonly resorted t o whenever dep th of water creates a cri tical situation. In summary then, it would appear that the restrictions imposed on port desi gn and velocity constitute one of the important and common eff ects of i ce on the structural form of intakes.

While

this conclusion is stressed in texts and authoritative reports, mention should also be made of those protective deviees and auxiliary structures that are a dvocated by s orne author s a nd yet not discussed in gener a l texts because their application is l imited or occasional. For example, in the case of: i)

Submerged Cribs, where the port f a ce s upwards as previously described , a cover plat e is some t imes i nstall ed a s hort distance above the inlet (*3, *14, *24).

When this plate

77 of steel or concrete is so placed that its clearance above the port is one-half to one-third the diameter of the port, it alledgedly acts to prevent any vortice from forming and compels the water entering the intake to approach along a horizontal line thereby reducing the danger of eddy currents and improving water freshness.

For all of these reasons,

Mr. McDonald (*3) advocates that such a plate be considered as a protective deviee against frazil ice, although he readily adroits that much of its effectiveness is lost in rivers where the approach current would be parallel to the plate.

In a dissenting opinion, Professer Azerier (*5)

considers any such deviee to be potentially dangerous as a source of blockage, a retardant when back-flushing and a poor substitute for a trash rack. ii)

Tower Intakes and Exposed Cribs.

The external force ex-

erted by ice pressure has already been considered in Chapter II (Page 28).

The stability of these structures

would appear to be greatly improved where the exterior walls are inclined and a buttress form adopted. iii)

Shore Intakes.

The effect of ice thrust against the shore-

ward wall of such intakes should not be minimized.

If the

conditions here can be considered somewhat similar to those experienced at hydroelectric plants, the suggestions of Mr. Paul E. Gisiger (*7) may apply; namely, (a)

Horizontal Pressures should. not be exaggerated and 10 kips/lineal foot is maximum requirement ancountered.

78 (b)

Uplift on band near top, due to rise in water leve! may amount to 1000 lbs./lineal foot.

This

will affect gates, projecting racks and similar appurtenances placed at or near the water level. In particular, the following comment by Mr. Gisiger appears pertinent to shore intakes: "where the intake is shallow, it should be protected by a barrier between the river and the forebay. This barrier can be formed by a skimmer wall or by heavy log booms floating between anchor piers. The skimmer wall consists of a solid .concrete wall supported by open arches through which the water has to pass. The top of these arches is 12 feet below normal pool level." In rivers, where current and frazil ice difficulties cannot be reduced to satisfactory proportions by a forebay of normal size, an artificial intake bay may have to be dredged inland from the main stream and connected thereto by a narrow inlet.

These stilling basins, already mentioned

in Chapter II (Page 50) in connection with the control of high velocity currents and turbidity, are effective weapons against ice when properly designed; in particular, the proportion of length to width, the size of the inlet and its position relative to the direction of the river current, are important factors.

The recommandations of

Professer Azerier (*5) in this regard, summarized and illustrated on Page 79, represent a valuable contribution of specifie information.

c.

The Effect on Installation.

On the basis of published reports it

would appear that few large intakes on this continent have been in-

Page 79.

Fig.lO - Intake Basins with Inlets in Alternative Positions •

.2H/er-"fe _/

).:_ ./;r/~..<:~

Recom.endations of Professor Azerier (*5) For comparative purposes only, two intake basins with different inlet locations are illustrated aboYe. In the case of Basin "A", the inlet presents an &!most direct opening to the approaching river current and the water entering the basin is mostly wsurface" river water with little sedt.entJ however, floating ice and libera~ed bottoa ice, borne along br current and wind, gain eaay access to 'i)le basin. Ill the' case of Buin "B", the river current IIUSt undergo a reversa! in direction at entry; consequently, ·~tta." river water predominates in the basin and increased sediments result. In both cases water freshness i.s good, and the probability of frazil-ice difficultiea about equal although this subject is open to sOJae discussion. Whereas Basin "B" is more subject to frazil-ice penetration because of its "bottaa" water content, it ia also the less turbulent of the two basins and sheet-ice will form early to protect the intate. Dy dredging the isthmls "C", a.:buü·.·1d.th c1ua1 inlets coald be coastructed and by the addition of inlet gates sa.e measure of water quality conUol obtained. In designing the inlet and basin, Professor Azerier aucgeat4; . (a) Inlet velocities be .reduced to 0.3- o.s ft./sec. to discourage frazil-ice penetration. (b) ~ volu.e of basin should be the equivalent of 20 ainutes coJWURption at II&Ximum intake capaci ty. (c) 1he length of the basin froa inlet to intake should be sufficient to perait the settling of sedt.ents, tbua ~re; H • Effective water depth,(œetres). L = tength of Basin, (metres). v Flow velocity at inlet, metres/sec. and s • Settling velocity of sediments (metres/sec.) or the upward flow velocity of frazil-ice crystaiS ( General value • 0.003 metres/sec.)

=

'nlen,

v s

L= H . -

80

stalled under winter conditions; hence, the effect of ice on the problem is rarely discussed.

Presumably the vicissitudes of weather,

water level and ice movement are major deterrents; in addition, since intake and conduit are preferably constructed at the same time, the use of dredges and divers would impose hazardous if not impossible conditions on men and equipment.

It may reasonably be con-

cluded that the installation of large intakes far from shore is not commenced when winter and ice conditions prevail. On the other hand, in the case of small crib intakes and some shore intakes the problem of installation may be simplified by the existence of ice.

Professer Azerier (*5) suggests that this is so,

and gives two instances where ice may be helpful: (a)

Small submerged cribs, assembled on shore, can be skidded across sheet ice, weighted and lowered by winch into their final position on the bottom through an opening eut in the ice itself.

(b)

Small shore intakes.

Construction can be speeded and oost

reduced by replacing sheet-piling with an ice wall created by removal of the top layer of ice from a semi-circumferential strip of the sheet ice adjacent to the site.

The

decrease in thickness along the strip will cause the ice immediately below to thicken from underneath and by repeated stripping an ice wall can be formed or "frozen down" until it meets and joins with the bottom. ftr. McDonald (*3) in confirming the usefulness of sheet ice as an

asset to installation, comments as follows:

81

"There are many small lakes or rivers on which no floatine plant is available and the intake has to be constructed by a different method." ••• "Fortunately, however, most of these lakes freeze over solidly and the intake may be laid quite easily through the ice during the winter months. In this method the shore end and part of the under-\.,rater section in the trench are constructed during the summer or fall, a coffer dam being built out into the water. Then when the ice is sufficiently strong the balance of the pipe and the inlet are put into place, a diver making the under-water joints. The Sudbury intake in Ramsay Lake was constructed in this manner." D.

Effect on Maintenance.

Various means are available to prevent and

counteract the adherence of frazil ice to submerged metal surfaces. In areas where intake location combines with the severe winter climate to make frazil ice a persistent hazard through the season, modern practice faveurs the use of steam, compressed air or electrical deviees to heat the submerged racks, gates or pump impellers and thereby produce the small but sufficient temperature rise (O.lOoF.) mentioned before.

Where the problem is infrequent

or occasional, backflushing and surging are often sufficient to prevent blockage. (a)

Thus for example,

Bay lis and Gerstein (*8), reporting on t he use of steam at Chicago's South District intake, say: "Steam injection at 90 - 100 psi. through three cocks into the suctions of large pumps produced a rise in water temperature of O.l30°F. Steam discharged through one nozzle of small pump suctions caused a simila r rise of 0.105oF."

(b)

Gisiger (*7) sugges ts tha t ice problems can be overcome by using compressed air and an air-bubbler system.

In the

example given by the author, two orifices, each discharging four cubic feet of a ir per minute, were set just below the level of the intake port for a hydroelectric

82

plant.

The orifices were located about five feet upstream

from the port in question and were submerged to a depth of ten fe e t.

The dis charge of air through these openings

caused the warmer water from the bottom to rise with the bubbles and prevented ice blockage at the bar-screens even wh ere th e (c)

·

a~r

t emperat ure was well below 0°F.

Heating trash racks electrically is advocated by many authors and the specifie recommendations of Professer Azerier (*5) in this regard are illustra ted in Figure 11 on Page 83.

This method requires special insulation for

the rack bars and the use of a low voltage electrical current. (d)

Under certain circumstance s, particularly in the case of industrial intakes, it may be possible to discharge relatively warm water at a point upstream from the intake port thereby tempering the water adjacent to the inlet and preventing frazil ice adherence.

Obviously, the

practicality of such a method is dependent on the quantity and quality of the warming discharge a s well as on the temperature differentiai that can be created. (e)

Back-flushing is an effective countermeasure against ice blockage provided that the problem is of a temporary or occas ional nature.

Confirming this general opinion ,

McDonald (*3) cites the example of a submerged crib intake supplying water from a river where serious ice trouble occurs onl y at infrequent i ntervals and back-flushing has proved a sufficient remedy when required.

~

,·

~,......."*-=--=---


'i:mn~.:-,1,.~"1

:

1

1,

----, f----:-4'8

..,j,

- - --LJ,,,.., 4A.A?

':f;t~,,/ S/'"/'l'''n/ 1(

!

1

~/Ml-- -- _ ~30111 _,- -__.J.

Pig.l1 - Blectrica11r Heated Bar-Screen déS~ribed by Azerier Dimensions are in millimetres.

(*5)

BLBCTRICAL HEATING OF BAR-SCRBENS Professor Azerier suggests that electrical heating of bar-acreens ia only practical where the position of the intake does not require laying long lengths of properly insulated cable, hence, electrical heating often becomes prohibitively expensive for the tower intake and submerged crib, and is generally considered as being applicable to the shore-based intate alone. Wbile frazil-ice ia normally accompanied by super cooled water, experiments have shown that the water temperature aelda. dropa below -o.03°C and that a teaperature of +0.01°C is aufficient to prevent ice adherence; consequently, the increment of beat that must be supplied is generally assumed to be 0.04°C. The flow of beat (q),in calories per hour, is then; g • 1000.Q.T Wbere Q water conauaption, cubic metres/ hour. T = temperature increment, degrees C • 0.040C.

=

and the corresponding Mechanical Power (M) •

m- kil&watts •

Thua, a flow of 1 cu. metre/sec requires 167.4 kilowatts,Which corresponds to practical data and experiment.

84 (f)

The use of more rough and ready methods such as pike-poles and explosives is mentioned by both Turneaure and Russell, and Babbitt and Doland.

Baylis and Gerstein (*8), report-

ing on the use of dynamite to create surges adjacent to ice-clogged racks, suggest that small charges (1/2 to 1/4 stick) be taped to long slender poles and lowered to within a few feet of the blockage and then detonated electrically.

In view of the risk which may be involved,

this method must be considered as a temporary expedient rather than a regular procedure in maintenance operations. The Effect of Ice - Conclusion.

From the foregoing discussion it

would appear that an intake can be effectively protected from the hazards of ice qy taking advantage of the defenses supplied by natural agencies, or through the use of artificial deviees, or by a combination of both.

While it is probably true that the very possibility of ice

blockage may, in sorne instances, virtually eliminate one form of intake from consideration; still, in most cases, the modifications which ice imposes in matters of position, structural form, installation and maintenance, can be accomplished without any radical departure in general form.

In short, ice does not demand any novel form of intake

but merely requires that the usual forms be adapted for operation under this condition. SECTION II - HIVERS AND STREAMS OF VARIABLE STAGE Scope of the Problem.

Throughout most of the preceding discussion it was

assumed that the fluctuations in water level of the supplying source were so controlled by natural or artificial means that the variations were either small in themselves, or at least, of minor importance in the problem under consideration.

85 At first glanee, such a premise might appear to seriously limit the applicability of the principles developed in the foregoing pages where the question of water depth has frequently been cited as a crucial factor. It should be recalled, however, that a number of sources of surface water are, in fact, free from great variations in level and therefore qualify under the premise in question. (a)

For example,

Large Lakes, where the effect of variation in stage of the tributary streams is dissipated over a large surface area.

(b)

Small Lakes, where ground water constitutes a major source of supply and the lake itself acts somewhat as a reservoir and surge tank between inlet and outlet streams.

(c)

Large Rivers, where industrial use, hydroelectric development, or flood-pr e vention measures have required the construction of dams to maintain artificial control over water level.

(d)

Small Rivers, where the water level is already controlled for industria l purp oses or wher e s uch artificial control can be contemplated in connection with the municipal water supply and intake problem.

It s eems evident therefore that the assumptions made earlier in connection with water level fluctuation are quite valid for a considerable number of river intakes as well as for t he majority of lake installations. Nonetheless, it must be recognized tha t there are a number of rivers, l a r ge and small, wher e wide varia tions in s tage a re quite common and where artificial control cannat be exercised without serious interference to navigation or wi thout i nfringing on t he prior le gal r ights of t hird parties .

Under such condi t i ons , va r iable s t a ge mus t be a ccept e d as a

considera tion of paramount importance which may well govern the design of the intake.

86

Effect of Variable Stage - General. As might be surmised from the remarks above, the general effect of variable stage on intake design is so great that it becomes impossible to consider the changes involved as mere modifications of the principles developed previously.

In short, a new approach must be adopted on the

questions of position, structural form, installation and maintenance which will take into account the far-reaching effect of varying water level; and for this reason, it would appear useful to establish at the outset of the discussion, the bread limits of the problem envisaged.

To this end,

the following assumptions are posited: i)

The surface supply under consideration is a river or stream where artificial control cannat be exerted and where the duration and intensity of the variations in stage have been recorded for many decades.

ii)

Fluctuations in stage produce equivalent variations in velocity, discharge and live cross-se ction of flow.

iii)

Varying velocity and cross-section produce aggravation of bedload conditions and the possibility of changes in the position of channels and currents.

iv)

Variation in level between extreme high a nd low stages may be fifteen to fifty fe e t.

With these ass umptions as a basis for dis cussion, the e ffect of va riable s t a ge on the i ndivi dual pr obl ems of posit ion, structural f or m, installation and maintenance will now be briefly considered. Ef fect on Position. In rive rs of va ria ble s tage, t he following considerations are particularly important:

87 (a) Since reliability of operation is an absolute requirement, the location of the intake should ensure adequate port submergence at extreme low water level. is inadmissible.

The alternative, shortage of water,

In many instances, adequate port submergence

at low flows makes it necessary to place the intake in, or adjacent to, the hitherto prohibited channel areas.

Although

Burdick (*4) reports an intake situated in the actual navigation channel of the Ohio River, the consensus among authors would advise against such positioning. (b) Even where the intake is located in or near a non-navigational channel, increases in turbidity, bed-load, and in the hazards presented by floating and submerged abjects must normally be anticipated and counteracted. (c) Where the intake is to be a permanent structure, as opposed to a movable type of intake (cf. Effect on Structural Form, Page 91) the river bed material should be sufficiently stable to ensure that the position of the existing channel will remain unaltered. Instances have been reported where the channel has silted up, moved away or eut behind costly installati ons.

Again, govern-

mental agencies have been known to alter the position of channels by dredging, moving them away from the intakes and rendering the latter inoperative (St. Charles, Mo. and Council Bluffs, Iowa.). (d) The value of the "location-survey", advocated and described in. Chapter II, is very much depreciated under conditions of variable stage .

In this connection, the comments of Lischer and Hartung

(*1) are significant:

88 "In the design of an intake, stream stage records are the only reliable source of data for determining extreme low-level conditions. Site observations may be helpful but it is necessary to be extremely conservative in establishing the lowest operating leve! of the intake to al1ow for deficiencies of record information and for the variability of nature." (e)

Intake design requires accurate estimates of both extreme high and 1ow water levels.

Lischer and Hartung suggest that:

"The same care should be given to the selection of the safe flood height of the intake structure as is given to the determination of the spillway capacity of the dam. The extent to which it is possible for floods to exceed previously considered safe levels was demonstrated in the floods of 1937 at Cincinnati, 1951 at Kansas City and 1952 in Omaha. At Cincinnati, the previous maximum was exceeded by 8.9 ft., at Kansas City by 5 to 8 ft., and at Omaha by 5.5 ft." ••• " The designer has no choice but to provide a liberal margin of safety in establishing the flood protection level of the intake. 11 "Low water conditions have an important bearing on intake design. The intake must be capable of extracting the necessary quantity at the lowest stage which might occur. The depth of the lowest intake opening must be materially lower if the stream is subject to floating ice and extremely cold climatic conditions. If a minimum intake submergence of 3 ft. is adequate in warm climates, 8 to 10 feet and preferably more submergence may be required in cold climates. Hence, the season of the year in which low water occurs is an important consideration." In summary then, it would appear that the principal effect of variable stage on intake position is largely one which subordinates water quality to the more vital question of supply in quantity at all times. This is in agreement with the general theorem of Chapter I; namely, that reliability of operation far outweighs efficiency as a prime requisite in intake design.

In a more specifie vein, the effects of variable stage

outlined above auggest a number of conclusions of which two are perhaps pertinent here: (a) Where ice problems develop concurrently with low water level, the available water depth may prove dangerously inadequate.

In

89 such circumstances, infiltration galleries along the banks of the river or below the river bed (cf. Chapter II, Page 19, Special Intakes) are often to be preferred as a source of supply. (b) Since variable stage presents a problem where changes in turbidity, bed-load, current velocity and even channel position must often be expected, occasions do arise where no one position can be found for the intake that will meet the changing requirements imposed by all of these variables.

In

this event, the use of a movable type of intake suggests itself. Effect on Structural Form. In contrast to the mere modifications which ice imposes, variable sta~

demands major changes in structural form of intakes and in the

concepts which have hitherto governed their design.

According to Lischer

and Hartung (*1): "The means of extracting water from variable streams are as varied as the imagination of man." While this statement will be substantiated by illustrated examples in the following pages, it would appear advisable to begin the discussion by considering the effect of variable stage on the int akes previously described.

The following points are perhaps of particular importance

in this regard: i)

Submerged crib intakes can rarely be used because of bed-load and siltation problems.

ii)

Shore intakes.

Although increased in size and cost, shore in-

takes are commonly used for the larger intakes on rivers of variable stage.

Major requirements are increased height for

the structure, provision of ports at various levele and a

90

suitable site,

An ideal location is one which provides a steep

bank of unquestioned stability adjacent to deep water.

Lischer

and Hartung (*1) suggest that: "In general the intake site should be chosen on the outside bank of a well established river bend, a point at which the river is usually swiftest and deepest. The inside of a bend is generally shallow and has broad flat sand bars. If a site is available at which the river runs against a rock bank, a reliable intake can generally be built." In a partly dissenting opinion, Roberts (*6) recommends that the inside bend be selected as the site since there is less danger of the foundations being undermined by the current, trash accumulations and port maintenance are reduced, and adequate depth can generally be obtained by locating the intake well forward from the bank and partly or entirely in the river.

In sorne

instances, the gap which may then exist between the structure and the shore can either be filled with rock and stable material to artificially extend the shoreline, or simply bridged to provide access to the tower-like intake.

The former method,

although expensive, is often designed to compress the flow between stable banks and the fill so placed is frequently discontinued above average low water mark.

In this way, the major

portion of the live cross-section of the river is available to cope with flood flows. iii)

Tower intakes.

Although these structures must be strengthened

against the increased pressure of flood water, protected by piling from the impact of floating abjects (including inadvertent colli sion by river craft) and generally eauipped to combat the greatly augmented problems of bed-load, siltation and port

91 blockage; tower intakes are possibly the least affected by the operating conditions which variable stage imposes on the intakes so far considered.

This is largely because auch towers are

equipped with ports at different levels and in different positions. Provided that asuitable site can be f'ound which w:iill ensure stable foundations, adequate port submergence and a solution to the problem of inaccessibility during periods of heavy floating ice, the tower form of intake remains an effective, if expansive, means of obtaining water under conditions of variable stage. In addition to the foregoing types, a number of novel intakes, commonly termed Movable Intakes, have been designed to cope with the problem of fluctuating water level.

In the view of Lischer and Hartung (*1):

"These designs usually have the advantage of low first coat in comparison to the fixed installations required under similar conditions of moving river bottom and changing water elevation. The low first cost, however, must usually be balanced againàt the operating costa entailed each time a pump or structure is moved." Examples of these movable intakes are described and illustrated on the following pages.

In each instance, the drawing and oommentary has

been taken from a paper (entitled "Intakes on Variable Streams") presented by Lischer and Hartung at the annual conference of the American \iater Works Association in Kansas City, Missouri, on May 8th, 1952.

Pig. 12 - Intakes on Movab1e Carriages

( St. Louis County, Mo.)

INI'AK.ES ON MOVABL.E CARRIAGES

SOI!tetimes a river bank can be shapect, paved, and made stable to form an incline extending from the river bottom to a point above high water. Intake pumps on wheels or movable carriages which can be moved up or down the incline often constitute a low-cost, feasible intake. This type of intake is especially adaptable if inclines can be formed on river bank levees that have been made stable by riprapping or revetment work and that are near the water's edge,

but have soil-bearing conditions that preclude construction of heavier fixed intakes. Tyrical of this type of intake are the movable pumphouses and pumps at the st.Louis County \tater Co. on the Missouri River. Pumphouses, each containing a 16-in, or larger pump, are mounted on wheels which roll upon a railroad track. A pump discharge header parallels the rails on the pump incline. Planged openings into this header are provided at severa! locations to receive the pump discharge and each intake pump is designed to operate with a 15 ft suction lift.

( ,,;

., _,.

/

Pig.13

- Intate Pump on Ploating Barge

Mo.

PLOATING PUMPS

If both the river bank and the river bottom shift frequently, the floating type intake is sometimes applied. Intake pumps are installed upon a barge which can be moved with the ehanging river channel. Both horizontal and vertical pumps have been used on auch barges.

Diacharge pump piping is connected through flexible con-

nections to a pipeline header which is usua11y carried froa the shore to the barge on high piling or upon a rock-filled dike. If river bottom conditions permit, the pipeline can be extended to the floating barge by laying the pipe in a dredged trench deep in the river bottom and terminated in a cluster of piling at the barge location. Operating power can be obtained through an electric cable or by diesel engines.

l'a:;e 94 •

•

c;.,, R.ul ftr

/

~.~.,,. ~~(ly,j

Pic. 14 - Intate Pump in Series with RemoYable Motors SBRIBS BLBVATION PUMPS

Series paaps are installed in saall dry wells at eleTations at least 15 to 20 ft apart. When rising river eleYations threaten to flood any of the pUllp and we 11 units be low the top unit, tbe .otor of that unit ia re.oved &Dd the pump ia teaporarily abandone4 to the rising river. Puaping is continued through the remaining

paapa _in the series. There is no appreciable losa of capacity into the 8Jatea upon the abandonaent of a puap or puaps inas11111ch as the increaaed river stage and consequent _lower pWlping head compensate for the losa of pumping stages.

rage 95.

z

Pig.lS - Low River stage Pu!p,

LOtf RIVER STAGB PUMPS

When the total intake pumping head is relatively low, changes in river stages frequently cause important total intake capacity changes - for example, at the St,Louis County Water Co., a rise of 20 ft in the river elevation will cause an alm.ost equal drop in total pumping head ( under some conditions of intate station operation ), and an increase in capacity of one of the pumps from 12 to 18 llgd. Under such

.

conditions of changing river elevations, intake design must tate into consideration the loss of pump capacity at low river stages, One aethod of compensating for this loss of pumping capacity at low river stages - that is, high heads - is by the installation of a low river stage pump which becomes a standby pump or even inoperative at high river stages, In the example aboYe, the pump provides addititional capacity between river stages of 14 aad 36 ft. Wben the river stage approaches 36 ft, the vertical pump motor is removed and the pump is temporarily abandoned to the rising river.

l 'age 9 ~· .

-

Fig. 16 - SUbmersible Intate Puaps

Lïwe1is.

SUBMBRSIBLB INI'AX.B PUMPS Tbe submersible pump and intake has been success-

fully applied on streama and rivers which flood. Simple shore type intake structures are constructed and submersible pumps installed during low water. Inasmuch as pump and motor can be completely submerged, construction need not be carried to elevations above high water. Small submersible pumps can be installed in almost any position - vertical, horizontal, or upon an incline.

~age

97.

Pic. 17 - Intate Puap in Siphon ~11 in aotterdaa, Holland. SIPHOO WEU

'l11e siphon well intake is a f ixed shore type of

wt pit intate .tlich recei'Yes its wa.ter fr011 the river through a siphoo JJ.ne. Tbe siphon intake line requirea air e'Yacuation to maint&in the 'Y&euaa

aeceasary for proper operation. 'l'be chief advantage of the siphon

w11 la one of lower first cost aince well drilling methods can be uaed to conatruct the wet pit. Little construction is required UDder water. This type of intake is in use at St. Louis,

Mo.,

at .Bdc1ingtoa,

Pa.,

Uld at R.otterdul in Holl&D4. 'nlese intakes take water froa the Missouri, Delaware,aa4 Maas li.era, respectively. Tbe siphon well intate is unaffected by flood or low

water coaditions on the river. In addition, a very important consideration is the eaae by which it can be adapted to reaote-control operation.

..:. :J~:c ~~------~/cr ' - - CJ_p/,.?/?of/

h f .- .. / .:;n

>: ·-~.:;~

Pig.18 - SUspended Pump Intake SUSPBNDED PUMP INrAJœS The outstanding developments which have been made

in vertical well type, lli.xed- and axial-flow pumps malte possible the

construction of intates in which the pumping units can be suspended into the stream. The suspending structure may be a bridge or a special high level platfol'Jil founded on piling. Pigure 18 shows an intake at St.Louis on the Mississippi RiYer wbere three-stage well type pumps are suspended from a pier. The motors are well above high water. The pump columns are protected by a steel casing comparable to a well casing. The advantage of this type of intake lies in its

low cost and the elimination of costly underwater construction. The pump must be protected from drift and navigation and for this purpose it may be necessary to construct an outer protecting fender of piling.

99 Variable Stage - Effect on Installation. The conditions which affect the installation of any given intake, particularly movable intakes, are so unique that generalizations become difficult and unwise.

Nevertheless, the writer is tempted to include two

suggestions both of which are rather obviously inferred in the preceding pages; namely, (a) Whereas variable stage, or more specifically, low water conditions may greatly assist installation procedures auch as the construction of caissons, cofferdams and foundations, undue reliance should not be placed on the co-operation of nature when planning the operations involved.

Excessive rainfall during

normally dry periods may prevent the water level from falling to the anticipated stage and sudden storms may produce flash floods of dangerous proportions. (b) In the case of a number of movable intakes, pumps and other equipment must be moved to higher levels in times of flood.

It

seems only prudent that the operations involved should be carefully planned, the requisite equipment tested and the adequate personnel trained in advance of the actual need.

Although

periodic inspection and use of this equipment may be justly considered a part of maintenance, the fact remains that the installation cannot be considered complete until this matter has been attended to satisfactorily. Effect on Maintenance. Sufficient reference to the problems of bed-load, turbidity and trash accumulation has already been made to warrant the generalization that variable stage almost inevitably increases the scope and cost of

wo maintenance operations.

These three problems alone have far-reaching

affects that are difficult and expansive to counteract; for example: 1.

High bed-load and turbidity may produce severe abrasion of pumps and mechanical equipment in the suction-well thus lowering their operating efficiency.

The installation of grit chambers at the

treatment plant may also become necessary. 2.

Accumulations of trash in large quantities may require enlargement of the openings in the bar-screens and pumps capable of passing solids of larger size.

Mechanical or travelling screens

may be required in the suction-well to keep abreast of the cleaning problem.

3.

Back-flushing of the conduit, and the removal and disposa! of screenings may become part of the regular maintenance schedule thereby requiring the addition of more efficient, permanent and expansive equipment.

To all of these considerations must be added the effect of variable head on the problem of pump selection.

According to Lischer and Hartung

(*1): "Variations in level between extrema high and low stages of 30 to 50 feet are not uncommon. If the average static head is of the order of 30 to 40 feet, static pumping heads may vary from 10 to 60 feet. The selection of pumping units to operate over this range at good efficiency and without cavitation is difficult. "In selecting the pumping units, data on the duration or length of time that various river stages are experienced are necessary if maximum average efficiency is to be achieved."

101

CHAP'l'ER IV

lllustrations of Some :t'ypical Intakes

Forevord. On the folloving pages will be found illustrations of some of the

more common forma of vater intakes to which reference has been made throughout the previous chapters.

The drawings and the accompanying

summarizing commentary have been reproduced from sources which are given recognition in the text itself. Inasmuch as the illustrations are of general interest, they have been grouped here for ease of reference.

102

Pig. 19 - Tower Intake (Wet Type). TYpical Chicago Intake Crib

1

" At the present time the Chicago Water Worta obtains its supply froa six surface cribs varying in distance froa two to three ailes from shore. These are circular structures, wbich vary fr011 90 to 112 ft, in diaaeter and are located in water from 32 to 37 ft. deep at the various cribs. Water is taten in through the ports in the sides of the structure, which generally extend fr011 S to 12 ft. above l&te bed, and enters a circular well in the center of the crib from which a shaft extends downward to a connection with a tunnel extending inland to one or aore of the city pumping stations. The water is controlled as it enters this downtake shaft. The ports at the outer circumference of th~ crib are proportioned to a aaxiaua delivery of about 1 fps. and the drait per crib often exceeds 400 agd. A structure on the above dimensions, in a locality subject to navigation and in the relatively shallow depths of water prevailing, necessarily extends above the surface of the late. The parapet walls of the Chicago cribs are about 27 ft. above the ordinary low water leve! in order to provide for the fluctuations in late leve! and the action of the waves." Illustration and coiiiJilentary from "Water Works Intakes" by Burdict (*2)

·~

h.cc

10~.

_ L_ · - · ··· ··~,

··-r

_:~f

Fig. 20 {.\) -

:~

rnta':e Crib icsiyner1 f nr use in a bay. ( .\.W.'i' •.\. Journl.l, Vol.32- .\:,ril 104ù.) 'r.lis subner,;e·'l crih intal;c is clescribeJ by McDonald (*3) in the f ollowing way, " In la!~es the cr~ · ) usual 1Y con tain~ an inlet consi~tin:; of an elbow and bellfllottth f :tcinc up\'Jar:!s an··! over t 11i:::. i.:-; ;_)lace:! a steel or con crete covcr plate b,1 f o.rn :1. square or cyl i nrlric:tl in let L'f l:l.r ~ ~ arca and consequcntly low vclocity."

Hg. 20 (3) - An Intakc Crib dcsiçned for use in a river. ( ,\.~'l. ·.;r.A. Journal, Vol. 32 - ,\pril 1940.) ~cDonald (*3) report ~ t~:~.t etis intake is loratcct in a laree river in al>out 40 ft. of wüer. -rl~e center of t:Je inlet Ï" 7 ft. abovc the river bottor~ :Ul~l f:1ccs towards nirlstrc--t~ so t~ ':l.t tl1c fla\~· into the int:ü::e is at right an3lc!" to tl:c river fl0\i • .Jntrancc velor:ity ::t ma..·dnun flou is 7. 9 ins • .:-'er sec. Ice blocl;:age has occured n i nc ti-;c s in thirteen years but the inta'~c has bccn r:lc:1.rc ·1 by revcrsinc the flow.

104

-

--

~r ~

~--· -

T

=t:-

Pig. 21 - Submerged Crib in Lake Michigan at Gary, Ind. " The intake crib at Gary,Ind., is a sucessful example of the submerged crib for large water delivery. The Gary crib is located at the southern end of Lake Michigan and one and one-half miles from shore in water about 44 ft. deep. The crib is octagonal in shape, 60 feet in diameter and 10 ft. in height. It has a solid bottom consisting of two layers of 12 x 12 timbers on which is superimposed a cellular crib work of timbers of the same size, built up log-house fashion and solidly joined with through bolts and drift bolts, stone and concrete filled and decked over with 12 x 12 timbers. The crib rests upon the clay lake bottom with no more support than a thin layer of crushed stone to level up the bottom, and is surrounded by an embankment of heavy riprap stones, 26 ft. wide on the bottom, and sloping upward to the top of the crib. The water enters at twelve 5 x 5 ft. ports lying in the horizontal plane of the top at the outer edges of the crib, from tmence it passes to a central well extending downward to a tunnel in the clay located about 45 ft. below lake bottom. This crib was designed to take 60 mgd. at an intake port velocity of 0.33 fps. The crib has been in sucessful use since 1909 and there have been no ice or sand troubles." Illustration and Commentary from "Water Works Intakes" by Burdick.

105

· ..

-•

- ·--

Fig.22 - A Shore Inta.ke with Dry-\'lell

In the illustration above, compiled from drawings by Babbitt and Doland, Steel, and Professor Azerier, water is admitted to a wet-well tllat is divided into t\1o compartments by a partition of heavy timbers set between, and supported by, structural steel channels. These channels serve as guide-frames for the mesh screens \~lich cover openings in the partition, and extend upwards through the floor slab to within a few feet of the ceiling on the top floor. A pair of screens is provided for each opening to permit alternate cleaning. Having passed tllrough the screens to the inner wet-well, the wn.ter is withdrawn by pumps located in the adjacent dry-well. A sluice-gate, hatch-ways, steel ladders, I - beams, chain-hoists, and an exterior landing over the bar-screens,all provide facilities for the cleaning of the various compartments and appurtenances. Many shore intakes make use of fully submersible pumps and the dry-well is ommitted in the interests of economy. Where the accumulation of debris and screenings is large, travelling screens become a virtual necessity.

106

CHAPTER V

"The Experimental Investigation - Scope, Theory, Object and Limits 11 General Scope. In previous Chapters the effect of intake position on the turbidity of raw-water entering the ports has been discussed at some length, and throughout most of the discussion the term "position" has designated the general location of the intake as a whole in relation to the shore, currents, channels and bed of the river or lake concerned.

So defined, the ideal

position for an intake has been established as that which, among other considerations, provides quiescent conditions conducive to the settling-out of the heavier portions of suspended matter thereby, in effect, reducing the turbidi ty of the water immediately adjacent to the intake.

While this

reduction is often significant, it is seldom sufficient to reduce the turbidity to the desirable minimum; consequently, the raw-water approaching the ports is still turbid and the question arises whether or not a further reduction can be effected before entry of the water into the conduit.

Un-

doubtedly, the provision of low entrance velocities and a flared or bellmouth port, already amply discussed, is a partial solution - but what about the position of the mouth i tself in relation to the direction of the approaching current?

What difference in turbidity, if any, can be effected by facing

the intake port upstream in this current - or downstream - or straight out? Can intake position, in this specificsense, reduce turbidity?

The answer to

this question constitutes the object of the experimental investigation of this thesis as well as the subject of the following report and discussion.

107

Terms of Reference. Before proceeding to the more detailed facets of the experiment and the report thereon, a brief discussion of the terms of reference and of the theory involved would appear to be indicated. Turbidity is defined by Steel* as follows:"A water is turbid when i t contains visible material in suspension. While turbidity may result from living or dead algae or other organisms, it is generally caused b y silt or clay. The amount and character of the turbidity will depend upon the soil over which the water has run and the velocity of the water. When the water becomes quiet, the heavier and larger suspended particles settle quickly, while the lighter and more finely divided ones settle very slowly. Very finely divided clay may require months of complete quiescence for settlement. Ground waters are normally clear because the turbidity has been filtered out by slow movement through the soil. Lake waters are clearer than streams in flood because of the smaller velocity and because dry-weather flow is mainly ground-water seepage. Low turbidity caused by silt may result in a relatively high organic turbidity. The explanation for this is that low inorganic turbidity permits sunlight to penetrate freely into the water and stimulates a heavier growth of algae. Turbidities are expressed in parts per million by weight and may range in natural waters from 1 p.p.m. or less to as muchas 40,000." With respect to the turbidities encountered in natural waters, Turneaure and Russell* report the following values for the average amount of sediment carried by various river waters used as public water supplies:CITY

RIVER

SUSPENDED MATTER Parts per Million

lawrence

Merrimac

10

Pittsburgh

Alleghany

50

Washington

Potomac

80

Cincinnati

Ohio

230

Louisville

Ohio

350

New Orleans

Mississippi

650

St. Louis

Mississippi

1000

Burdick* reports that the inshore turbidity of the Great Lakes under turbulent surface conditions is about 100 p.p.m.

108

Of course, all of the above merely serves to indicate the general range of values to be anticipated since it is quite impossible to state any significant "average" value for turbidity in natural waters.

However, what-

ever the variations, the treated water delivered to the consumer must not have a residual turbidity in excess of 10 p.p.m. (silica scale) if it is to conform to the accepted standard of the United States Public Health Service. The term "silica scale" means that the standards of comparison for turbidity are mixtures of silica and clear water of known concentration.

Thus, the

question of importance here is not so much the amount of natural turbidity which exista, but rather the amount that can be removed before entry into the supply system when full advantage is taken of the action of subsidence. In other words, what is the nature of the material which can be removed by settlement prior to entry? Commenting on the action of subsidence, Turneaure and Russell* state:"The particles of sand and clay have a specifie gravity of about 2.6; they are, therefore, held in suspension only by virtue of the currents maintained in the water. When these currents become retarded the suspended matter is gradually deposited, the rate of settling varying with the size and form of the particles. The weights of similar particles are proportional to the cubes of their diameters, while the surface areas are proportional to the squares; consequently, the relative resistance to sedimentation is much greater with fine particles than with coarse. According to Hazen the rate of settlement of particles of sand and silt ~ still water is approximately as follows:- 11

Mate rial Coarse Sand Fine Sand Silt Fine Clay

Diameter m.m. 1.0 0.25 0.05 0.001

-

0.5 0.10 0.005 0.0001

Rate of Settlement Ft. per Hr. 1200 636 320 96 0.46 35 0.00018 0.018 -

Unfortunately, raw water surrounding an intake can seldom be classed as still water; hence, Hazen's settling velocities can only be considered as a guide in estimating the vertical component of particle movement.

In

109 addition, the use of such settling velocities as a basis for calculation presupposes that the flow in question is rigidly controlled as in the case of a sedimentation basin where baffles, weirs and other deviees can be installed.

However, even under such controlled conditions, particle

settling velocity is subject to wide variations and is not generally accepted as an absolute basis for the design of such basins.

In this con-

nection, Babbitt and Doland* comment as follows:"Attempts have been made to formulate mathematical expressions applicable to the determination of the dimensions of sedimentation basins. Notable among these attempts are those of Hazen, Slade, Camp, Carpenter and Speiden. In view of the number of conditions and constants involved it becomes the custom to base the design of sedimentation basins on experience with successful basins. "Conditions of importance frequently neglected in theoretical studies of sedimentation basins are velocity energy and density currents ••• Entering velocities are difficult to dissipate, particularly in large basins, and may even build up to create tornadolike "storms" among settling partiales. Density currents may create 11splash" near inlets that will prevent sedimentation in the region, and inadequate outlet deviees may concentrate or create currents that will draw settling partiales or previously settled sludge, from or near the bottom of the basin, at the same time that partiales suspended nearer the water surface are undisturbed." Quite obviously, the uncontrolled flow around an intake will produce currents many times more unpredictable than those encountered in sedimentation basins; hence, settling velocities alone are a weak tool in predicting the size of partiales that can be removed, and other approaches that will take into consideration the turbulent and uncontrolled conditions of flow must be considered.

In this connection, the known values for the currents

necessary to move solids in surface drainage sewerage systems, while neither completely satisfactory nor analogous, may be of sorne significance. The following tabulations, submitted by Greeley and Stanley in the Handbook of Applied Hydraulics*, represent the findings of the New York Metropolitan Sewerage Commission:-

110

Currents Necessary to Move Solids Kind of Material

·velocity Required to Move Material Along Bottom - fps.

Pebbles - 1 11 diameter

2.0

Pebbles - 1/2" diameter

1.0

Fine Sand

0.50

Fine clay and silt

0.25

General Assumptions. With all of the foregoing in mind as a basis for argument, including the desirable values for approach and port velocity discussed in earlier chapters, it would now seem possible to enunciate certain generally valid assumptions with respect to the reduction of turbidity of raw-water at an intake; namely, a.

Approach velocities in excess of 3 fps invite high turbidities and bed-load conditions.

b.

Approach velocities of 1.0 fps or less should normally permit settlement of the coarser sand particles and almost eliminate bed-load.

c.

Intake ports with velocities ranging from 0.50 to 0.25 fps will probably not entrain coarse sand, and at even lower values may largely prevent fine sand penetration.

This a s sumption pre-

supposes that the approach velocity is proportionately low and that the ratio between approach velocity and port velocity is approximately two to one as recommended by Professor Azerier. d.

The removal of all fine sand by simply reducing approach and port velocities is quite improbable; the removal of silt and clay particles quite impossible.

Optimum reduction of turbidity at

the intake woul d a ppear to exist when all particles with a diameter in excess of 0.10 m.m. (or 100 microns) are excluded.

111 If these assumptiona can be accepted as basis for further consideration, the discussion may now revert to the question of whether or not there exists a preferable position for the intake mouth in relation to the direction of the approaching current.

Stated in more general terms,

the problem is that of determining whether a reduction in turbidity can be effected by varying the directional relationship of approach velocity and port velocity. Figure 23 on Page 112 depicts, in plan view, three positions in which a simple intake vith flared mouth could be placed in a flowing stream. In each instance the velocity of the stream, or approach velocity, is designated by "V", the port velocity by "v", and the velocity through the conduit by "v0 ". Specifie Assumptions. A summary examination of Figure 23 would seem to indicate that certain specifie hypotheses are possible; namely, A.

Intake Port Facing Upstream. Some of the particles borne along by the main stream would appear to be able to gain entry to the intake port w1 thout change in direction.

B.

Intake Port Facing Straight Out. In contrast to the conditions prevailing with the intake port facing upstream, the straight out position is at right angles to the main current; consequently, particles entering the port must undergo a change in direction.

This consideration alone is

sufficient to suggest that the turbidity of the influent will be less vith the port in this position than with the port facing upstream.

112.

fi.:;. 2J - T1_:rc,; positions for . ,_ \ntcr i'1ta.:,c. V ::: ·,'(; L .> C L t:· uf f 1 o;·;

in t '•c ::;ui'i' lyin.ss nnrcc.

-~--v~

~ · 0 rt

v=

v0

=

._~~~~~~~~

t

Yo (a.) r:a.c i n <; Cp;: t re ·tr..

V _---~

f

v.; (b) Pn.cinç

D o \:n s tr c ~m .

r

'Yo ~

(c) racin: Strai;;ht uut.

vc1ot:.ity .ü t'1e int;ùce. inta.!:c veloci ty •

113

c.

Intake Port Facing Downstream. Whereas the straight out position for the intake port compelled partiales to undergo a change in direction of ninety degrees before entry, the downstream position for the port mouth requires a reversa! in direction of 180°, or more specifically, requires the partiole to deoelerate in the original direction of the main current until it reaches a velocity of zero, and then undergo acceleration in the reverse direction.

Such a

reversa! in direction suggests that a reduction in turbidity will result, and that of the three positions under consideration, the downstream position will produce the least turbid influent water. The Hypothesis. At very beat, the theories so far advanced would appear to suggest that a hypothesis can be formulated with respect to intake position and turbidity ,reduction; namely, "Where an intake is introduced into a flowing stream of known turbidity, the positioning of the mouth of the intake in relation to the direction of the current in the stream can effect a reduction in turbidity for the water entering the intake." As stated earlier, the proof of this general theorem constitutes the object of the experimental investigation undertaken in connection with this thesis.

Moreover, the theory advanced previously suggests a

corollary to the general theorem; namely, "Assuming that the particles producing turbidity act somewhat as discrete particles then, - when the intake port is faced upstream, the reduction in turbidity effected is lesa than that obtained when the port is faced straight

114

out at right angles to the ourrent& when faced straight out, reduction is again less than that which resulta from facing the port downstream. It should be noted that the adjunctive statement immediately above bas · been made oonditional to the nature of particle movement whereas the relationship postulated by the general theorem is not so oonditioned; hence, the adjunctive statement could be experimentally disproved without invalidating the hypothesis of the general theorem.

Such an observation,

unintentionally didactic, is merely meant to acknowledge the possibility that a relationship between intake position and turbidity could exist for reasons little concerned vith the theoretical character of particle movement.

Such a relationship, predicated by a different theory and produced

by other causes, might conceivably produce experimentally proven resulta

which are quite contrary to those forecast by the discrete particle theory. Discrete Particle Theoty. Mention has already been made of the simplitying assumptions implied in discrete particle movement.

a)

Notable among these are the following:-

The direction of flow in the main stream is assumed to be horizontal at all times and the velocity of flow equal at all points.

b)

The concentration of particles - or turbidity - is assumed to be equal at all cross-sections of the flowing stream.

c)

Particles are assumed to settle vith uniform velocity, without change in aize or shape and without interference or coalescence.

d)

Particles, once settled, are presumed to be removed or in event, not to be resuspended.

aQJ

115

How valid are these assumptions when compared with the aetual conditions tbat may normally be expected around an intake?

The previous dis-

cussion, the quoted opinion of authors given earlier, and common sense, would indicate that auch assumptions are far from identical vith aetual field conditions.

In brief, all of the assumptions are subjeet to con-

tradiction in varying degrees by the factual conditions of turbulent flow, non-uniform velocity, varying bed condition, varying turbidity and resuspension of particlesf consequently, in the present case as in the case of sedimentation basins, the discrete particle theory is best considered as an avenue of approach for experimentation rather than as a firm basis for design and application.

Furthermore, where the discrepancies

between theory and fact are large, it must be aoknowledged and foreseen that the probable application of the theory will be correspondingly limited.

Nonetheless, the theory cannot be dismissed in its entirety

without experimental evidence to show that it is inapplicable to the problem of intake turbidity. Experimental Objectives - Summary. In view of the immediately foregoing considerations and of the hypotheses postulated earlier in this chapter, the experimental investigation was undertaken in the hope of determining Primarily, That irrespective of particle theory, a reduction in turbidity can be obtained by preferential positioning of the mouth of an intake in relation to the direction of the approaohing current, and Secondarily, The extent to which the theory of discrete particle movement is applicable to auch an investigation.

116

Limits of Investigation. Practical considerations, founded either in the theory relating to intake operation or in the wise usage of laboratory equipment, made it necessary to adopt the following limits for the investigation:Intake Position - Three positions for the mouth of the intake were investigated - upstream, downstream and straight out. Approach Veloeities -"V"- were limited to the range from 1.0 - 5.0 ft./sec. which may be considered representative of conditions existing at the majority of intakes. Port Velocities - "v" - were selected in a normal range from 0.20 ft./sec. to

o.ao

ft./sec.

Turbidity - "T" - of the main stream was artificially created to correspond to a value of 1000 ppm.

While admittedly high when com-

pared to the turbidity of natural streams, this value is not preposterous and might be considered to be the equivalent of the turbidity encountered in some rivers at flood stage.

From a purely

experimental point of view, such a value was desirable in order that the amount of sediment collected in each sample of water would be of significant weight. Particle Size - Earlier discussions have established reasonable grounds for utilizing partiales of about 0.10 mm (100 microns) in diameter.

Larger particles two to five times this aize coulâ be

used in all probability without contradicting natural occurenees; however, the pump available in this instance was not designed for the circulation of water containing abrasive material, and the delivery lines connected to the pump bad a number of gate-valves, pressure gauges and piezometer connections interposed throughout

117 their length vhich made the use of large particles unwise and hazardous.

In view of these considerations, the particles added

to the water to create the "artificial" turbidity vere from 50 microns to 147 microns in diameter and vere composed of Magnesium Silicate (talc) - the least abrasive of the silicates commercially available at that time.

118.

Plate I - Discharge End of Flume at Velocity

= 3.0

ft/sec.

119

CHAPTER VI

"The Experiment - Apparatus, Procedures and Observations" Apparatus. The experimental apparatus, shawn on the following page in

diagr~

matie form, combined available laboratory equipment with certain additional elements which were designed and fabricated for this particular experiment. 1.

The available equipment consisted of the following:

A horizontally-driven centrifugai pump with a maximum rated capacity of 10 cfs.

2.

A supply reservoir of large capacity measuring 11 1 -2" in length, 7 1 -8 11 in width, by a depth of about 6 1 -6".

3.

Weighing assembly consisting of a scale balance and weighing tank with a capacity of about 5000 lbs.

The contents of the

weighing tank could be returned to the main reservoir through a connecting line. 4.

A steel-framed f.lume of rectangular cross-section which could be bulk-headed to form a receiving or mixing basin of any desirable capacity and into which the effluent from the overhead delivery line could be diverted.

The steel flume, being

of very sturdy construction, could also be used to support the test flume and receive any spillage therefrom during the course of the experiment. The apparatus which had to be adapted to the available equipment mentioned

Fig. 24 - Schematic diagram of Apparatus

-- -

...

r-

1

Basin

1

1

~ \...../

..

~ ~

Control Gate Valve

.·.~·

lntake pipe

1 I Mixing

1 1

__.,._"""'

Overflow

va:irb~ Head

~ei

Assemblv

weighing Assemblv -L_

- -T

Weighing Tank

Gate Valve

Reservoir

-....._d 1-' [\.)

0

121 previously, and designed and fabricated for this particular experiment, consisted of an experimental flume, intake pipes, bulkheads for the steel flume and an overflow weir.

A summary description of all of these

items is given below devoid of any explanation for the form and dimensions selected.

In the interests of brevity and continuity, these important

but secondary considerations have been put over to a special appendix (Appendix B) where the matter is fully discussed and a complete report on the hydraulic performance of the elements is presented. Experimental Flume. selected.

A flume of trapezoidal cross-section was

It had a bottom width of 1 1 -0", a top width of 3'-0" and

a depth of 1 1 -0".

It was prefabricated in two seètions of eight

feet by the firm of Daniela and Mannard in Montreal and assembled in the laboratory by the writer (cf. Plate I - Page 118, and Plate IV - Page 126). Intake Pipes.

Two intake pipes with an inner diameter of 0.684"

were fabricated by the writer.

Each was fitted with a bell-mouth

attachment 1-1/2" long to produce a port diameter of 1.813".

The

two pipes were identical in every respect except that one provided a ninety-degree bend for observations upstream and downstream whereas the other was entirely straight (cf. Plate II - Page 122). Bulkheads for Steel Flume.

Bulkheads of 1" marine plywood were

built four feet apart in a section of the steel flume to provide the necessary receiving or mixing basin at the outlet from the delivery line.

The plywood was fastened on 2" by 4" studs, the

two bulkheads wedged against the aides of the steel flume and the joints caulked with plasticine and oakum.

The bulkheads were braced

122.

Plate II -

Intake Pipes

123 by 2" by 4" struts placed between them and drawn up tight against these by six 1/4" steel reinforeing rods with threaded ends whieh passed through each bulkhead and were bolted on the outside of the supporting studs.

This work was done by the writer with the

assistance of the laboratory staff (cf. Plate III - Page 124). Overflow Weir.

In order to obtain samples of the water flowing

through the intake at a point where a continuous and free

dischar~

would oecur, it was necessary to construct a deviee somewhat like a drinking fountain - an overflow weir - in and out of which the intake water could flow eontinuously; furthermore, sinee the

dischar~

through the intake was to be varied, it was necessary that a differentia! head between this deviee and the flume should exist and the height of the weir varied at wi ll.

To aecomplish this a circular

tube of 3" diameter was buil t wi thin an outer circular shell of 6" diameter in such a way that water entering the intake would flow along a flexible plastic bose, pass through a funnel and into the inner tube. ed~

The water would then rise up the tube, spill over the

and fall into the lower encircling container from which it was

withdrawn by another hose connection and returned either to the reservoir directly or dischar~d into the weighing tank (as in the diagrammatic sketch).

The amount of water collected in the weigh-

ing tank over a known period of time served to establish the discharge through the intake.

This entire deviee - the overflow weir -

was attached to a sliding bloek which eould be made to move vertically between gui de rails; bence, the head on the weir was variable and the discha rge thereof adjustabl e.

The overflow weir

124.

Plate III -

Rear Bulkhead in position.

125 was constructed from galvanized iron sheeting by the firm of Daniels and Mannard of Montreal and supplied as a unit (cf. Plate V - Page 128). Additional Eguipment.

In addition to the apparatus listed previously,

certain additional items of equipment were also used; namely, 1.

A Gurley Current Meter (Priee Pattern) of the conventional bucket-wheel type.

When placed in a flowing stream the

buckets rotate at a speed which is proportional to the velocity of flow and every rotation (or multiple thereof) transmits a signal to a headset worn by the observer.

By noting the

frequency of the signal over a known period of time, the velocity of flow can be determined from a calibration curve supplied by the manufacturer for every instrument (cf. Plate IV - Page 126). 2.

A standard laboratory analytical balance to weigh objects in the range of lOO grams or less with a sensitivity of 0.001 grams (cf. Plate VI - Page 130).

3.

An evaporating oven of standard type providing temperature warm enough for the evaporation of water from samples of soil.

It was found that another additional source of heat

had to be supplied in order to increase the rate of evaporation of the all-liquid samples encountered in the present instance.

A single element electrical unit placed in the

bottom of the oven (cf. Plate VII - Page 119) proved sufficient for the purpose.

126.

Plate IV -

Current-Meter in position.

127 Procedure. With the apparatus assembled in the manner indicated by the diagrammatie sketch on Page 120, the slope of the flume was varied to produce the different velocities of flow investigated; namely, 1.0, 2.0, 3.0, 4.0 and 5.0 ft./sec.

Meanwhile, the discharge of the pump was also adjusted to

produce a desirable depth of flow in the flume which would minimize spillage and yet maintain sufficient submergence of the intake pipe. At each of these flume veloci ties (denoted by "V") the intake pipe was positioned so that it faced upstream, then downstream and finally straight out in relation to the approach current.

In each one of these

positions, for any given value of "V", the intake velocity (denoted by "v0 ") was adjusted through a range from 1.0 - 5.0 ft./sec. by raising or lowering the overflow weir.

In summary then, there were five flume

velocities at each of which there were three intake positions, and for each of these positions there were five intake velocities investigated making in all some 75 different sets of conditions. Sampling.

For any given set of conditions, three samples of turbid

water were taken, two from the intake weir and one from the flume.

The

flume sample was interposed between the two weir samples in order to be as representative as possible.

There were therefore 75 "sets" or 225

samples to be taken and processed. Flume Velocity . was determined by use of the Gurley current meter already described.

The meter was introduced into the flume at a point

just upstream from the intake position and observations carried out over a period of twenty minutes or until the conditions of flow became stabilized.

The meter was then removed and the intake pipe positioned in the

flowing current.

128.

Pla te V -

The ove r-f low

~ve ir.

~~

Intake Velocity.

As intimated earlier, the velocity through the in-

take was determined by measuring the discharge over a known increment of time.

Preliminary measurements were made by directing the effluent from

the overflow weir into a graduated container with a five gallon capacity and observing the time required to fill this container.

When the fill-

ing time closely approximated the precalculated interval required for any given intake velocity, the discharge from the weir was rerouted into the weighing tank and allowed to accumulate for about 5 minutes measured by stop watch.

The difference in weight at start and finish of

this interval could be translated into discharge in terms of pounds per minute.

Prior to commencing the experiment, the flow velocity through

the intake equivalent to discharge in lbs./min. had been established using the known dimensions of the intake pipe, and a calibration curve prepared relating these two quantities (cf. Appendix B).

It was there-

fore simply a matter of raising or lowering the overflow weir in order to obtain the desired discharge. Artificial Turbidity was created by introducing Magnesium Silicate (talc) in predetermined amounts into the water as it passed through the mixing reservoir.

Two particle sizes, 47 and 149 microns, were combined

in equal proportions by weight to provide a turbidity between 500 - 1000 ppm. Determination of Turbidity.

As each set of samples was collected, it

was removed from the test area by Mr. Wellington Chen, the writer's assistant, who carried out most of the sample weighings.

Each sample bottle

had been weighed empty and carefully marked in advance with an identifying number.

Each sample in the set was then weighed before and after evapor-

ation of the water content, and the amount of water as well as sediment

130.

Plate VI -

Analytical Balance and samples.

determined in each case.

Sample calculations of these simple arithmetical

procedures are included towards the end of the present chapter.

The re-

sults of these weighings were immediately translated into values of turbidity in parts per million and recorded on specially prepared sheets similar to those shown in the observations.

Observations. In

pr~senting

the experimental observations the following symbole have

been used: V

z

Q •

Flume velocity, feet/sec. Discharge of intake, lbs./min.

=

Intake velocity, feet/sec.

T

=

Turbidity of flume water, ppm.

t

=

Turbidity of intake water, ppm.

v0

Variation in turbidity between flume water and the water passing into the intake has been expressed mathematically by the relationship t - T • 100

T

which therefore expresses the difference in the turbidities of intake and flume water as a percentage of the turbidity of the flume water. The observations have been classified according to the value of flume velocity concerned and only the reduced observations, rather than the complete weighing reports, are tabulated.

Samples of the actual observation

forms are included towards the end of the present chapter together with sample calculations indicating the arithmetical procedures whereby the weighings of samples produced values of turbidity in parts per million (ppm). Supplementary observations were frequently made during the course of the experiment in order to test the consistency of the resulta. tions are indicated by an asterisk immediately.

(*)

These observa-

in the tabulations which follow

13J August 6th, 1957

Observations Flume Ve1ocity "V" Nominal Vo ft./sec.

Intake Q

lbs/min.

Actual vo ft./sec.

Flume T ppm.

= 1.0

tt./sec.

Intake t 2 samples ppm.

Mean t ppm.

t - T • lOO T

%

"Upstream Position" 1.0

9.4

0.99

411

564-547

555

+ 38

2.0

19.3

2.00

354

396-425

410

+ 16

3.0

28.3

2.95

265

525-600

562

+lOO

3.0*

30.0

3.12

518

552-472

512

-

1

3.0*

30.3

3.15

358

370-232

301

-

9

4.0

36.0

3-75

681

478-533

505

- 26

5.0

48.7

5.1

397

346-300

323

- 18

5.0*

48.2

5.0

521

594-598

596

+ 15

"Downstream Position" 1.0

10.9

1.13

563

623-530

576

+ 2

1.0*

10.0

1.02

572

430-502

466

- 19

1.0*

10.2

1.25

342

426-409

417

+ 22

2.0

20.8

2.15

487

403-520

466

- 19

2.0*

20.2

2.10

372

407-358

382

+ 10

3.0

29.0

3.00

471

500-428

464

-

1

3.0*

29.4

3.05

488

529-416

477 /

-

2

3.0*

30.3

3.16

394

288-374

331

- 15

4.0

38.0

3.95

424

390-514

452

+ 7

4.0*

39.5

4.12

458

522-518

520

+ 13

5.0

50.0

5.2

460

460-436

448

-

(co nt inued ••• )

2

133 Observations (continued)

August 6th, 1957

Flume Velocity "V" = 1.0 :ft./sec. Nominal v :ft./sec.

Intake Q

lbs/min.

Actual Vo ft.7sec.

Fltu1e T

ppm.

Intake t 2 samples ppm.

Mean

t- T

t ppm.

• 100

T

~

"Straight Out Position" 1.0

9.4

0.99

428

474-495

484

+

2.0

19.0

1.97

395

548- •••

548

+ 39

3·0

30.0

3.12

440

590-545

567

+

29

4.0

38.0

3.97

478

522-550

536

+

6

5.0

48.0

5.00

437

414-•••

414

5.0*

48.7

5.10

471

430-358

394

Commentary.

13

5

-

15

The resulta tabulated above and on the previous page, as

well as those vhich are recorded on the folloving pages for the other flume velocities, revealed certain conditions which required investigation and explanation; namely, a)

Although enough artificial turbidity had been adàed to the clear water of the reservoir to produce a turbidity of nearly 1000 ppm, the recorded turbidities of the flume samples varied ia the range from about 400 - 600 ppm.

It

was immediately suspected that a portion of the particles vere becoming lodged throughout the delivery system and in the reservoir itself.

All valves vere flushed out and the

pump alternately started and stopped in an effort to produce sufficient surge in the reservoir to resuspend the deposits. Supplementary observations (indicated *) vere then taken in the usual manner but again the samples showed a diminished

tvlttc11Q'

OYV

(ct. Appediz

tbat aa\1o1pate4.

:a - :hce

InRMil u ,_.,111« oapaoiQ' ,.. U.ld•••

ni), U w. deo14ed tbat \hta oODIU\ion wnld

baw to be aooepMd aa an experiMn\al liaUat:l.oll.

'b) n•tuUou 1n tlae turlt141\J aD4 1Dtaa 'llrb141tJ wn larll• lloreowr, . ftluea ob\aizae4 for tlle quan\1'J ' ; 'l' • 100 ~) l~J~Ül" abtoet ideDUaal

OOD41\1ona vere ao' oonaiaunt and ..a, ftlwa for \hia

,..utw

.. qt~U.U.t,

va"

in aign ind1oa\1Di thereb,v liat U• iD\ake ,.. U.o11arlla« (at

the . .Ill of MIIP11Dc) wat•r vith a h1Pel' turb141lJ \bu tha' vbiah W8 tlov~ in \be fluae (a\ \he --Dt Of AÇliJic} •

lt

W..

80Da1Wie4

tbat au.ah uama\ural oeadi\ieu nnlw« fi'• nwn "aton. of twU41t,. tba\ vere oonUmall7 puda, 40VD the

n-

and f!'ea the 1M~nsty of

\be . . .pli.Dg prooeëre \o cope vUh aaoh a oon41Uon

'bJ euuriq \bat

the - - "a\ona" W11 ~led bo\b at the 1'1- aDd &\ \ba Oftl'-tlOV ......

1\artherMre, ...pue the preoauUoaa \&ken 1D. \he &taip of lbl cwertlov wir (ot. Appendiz B - hp xi), U vu q111\e probaltle \bat he&Yier

partiolu ven oireulaU.JIB 1.11 \he qate• due \o ooaleaoeue aD4 ac«lc.eraUoa 4urinc perioù ot MUlaent iD the reaerroir u4 elaevben. Such beavr par\iolea, aoou.alaU.ac in \he owr-tlew

&117

ODe.

autple.

weir,

Jl:f..pt ...117

Por all of tbeae r.qona :1. t _ , lM aokaovlecSced

tbl auplillc prooedu.n aplOJed ill tbia

~H.noe

the oon4Uiona of v14eapna4 Q4 contilluoua

wu

1M.4eqJ~&M

~t

\o _.,

nuo,uaUou in tun14UJ.

priaillc aDJ ODe ae\ vi\hiD a aini ... \ta. illtei'Y&l.

~5

In view of the foregoing discoveries, it was decided to enlarge the

ecope of the observations in an attempt to evaluate the effect which the fluctuating turbidity values might have on the experimental objectives. It was becoming quite clear that unless the variations in turbidity due exclusively to intake position were large they might well be hidden by the pseudo-variations produced by a fluctuating turbidity in itself. Consequently, it was deoided to conduct an investigation, at every proposed flume velooity, of the extent to which the fluctuating flume turbidity would produce variations between flume and intake turbidities. This would require maintaining the intake velocity and the intake position as well as the flume velocity constant, or in short, making all the intended variables constant and leaving only the fluctuating turbidity as a variable.

It was arbitrarily decided to adopt an intake velocity of

2.0 ft./sec. and a "straight out" position for the · intake during the proposed tests and these constant conditions were adhered to at each ot the flume velocities investigated.

These additional tests, or "calibra-

tion runa" as they were termed for convenience, were conducted over a period of one hour during which two flume samples, one upstream and one downstream, as well as one intake sample were taken every ten minutes. The resulte recorded for these additional tests are included among the observations pertaining to the individual flume velocity concerned and may be considered representative of the accidenta! error of sampling involved in each instance.

136 Observations. Variations in Turbidity at Constant Velocities Flume Velocity •

1.0 ft./sec.

- - - Intake Velocity •

2.10 ft./sec.

Flume T 2 samples ppm.

Mean T = Tm ppm.

ppm.

Variation (%) t - T • lOO T

481 - 634

557

571

+ 12

490 - 754

622

551

- 11

513 - 427

470

446

-

452 - 545

498

658

+ 36

505 - 711

608

521

- 14

580 - 597

588

446

-23

lntake t 1 sample

Total Tm 3343 Average Tm = Ta • 557

Total t

a

Average t

a

5

3193 532 • t a

In the above it may be assumed that the average T&lues of flume and intake turbidities are closest to being correct; hence the percentage variation would be t

a

- T

Ta

a • 100

=

532 - 557 • 100 557

=

_

5~

In the test itself, single observations showed a variation from + 36% to - 23%.

These are the result of accidenta! errors in sampling and

arise from the chance that the samples at the overflow weir might or might not be representative of the water sample taken from the flume. While it is impossible to establish a correction for any one observation which is subject to accidenta! error, a series of such observations taken collectively tend to obey the law of probability.

Admittedly, in the

present case, the small number of observations makes the applicability of

137 the law somewhat doubtful; on the other hand, no equivalently rational method would appear to improve this condition.

The theory of probability

states that the probable error of a single observation is given by E =

r;-:2

0.6745J~

where E

=

probable error of a single observation (+ or -).

r

=

residual or difference between any one individual measurement and the mean value of all measurements.

n

2

number of measurements.

Therefore, in connection with the values of the relationship

t-T · T

100 listed previously, each value maw be subtracted algebraically from the mean value of - 5% and the value of r2 determinedf bence,

=

For Flume Velocity

1.0 ft./sec.

t - T • lOO

...L.

2 -L-

+ 12

- 17

289

-11

+

T

6

36

0

0

+ 36

- 41

1681

- 14

+

9

81

- 23

+ 18

324

-

5

and

Mean Variation = - 5%

Whence, E

+ = - 0.6745

E

= :!: 15%

~ 2~93

Total = tr2 = 2493 It may be conc1uded from the above that the precision of measurement of the variation in turbidity between intake and f1ume water for the present flume velocity (V= 1.0 ft./sec.) and with a constant intake velocity and intake position is

!

15%.

Moreover, if it can be assumed

138 that this variation is due mainly to a fluctuating turbidity throughout the delivery system, it can be foreseen that the precision of measurement might be some function of the discharge through the flume and ultimately of the flume velocity.

As a minimum deduction it may be concluded that

flume velocity is the governing factor in turbidity fluctuations and therefore in the precision of measurement of the variations; again, it would seem reasonable to assume that all other observations of turbidity variation carried out at this flume velocity, irrespective of intake position or of intake velocity, would have about the same range of probable error. As mentioned earlier, equivalent determinations of probable error

have been made at all flume velocities and the results have been presented in a summarized form in the observations pertaining to each of these velocities.

The foregoing commentary, sample calculations and

formulae, being applicable at all flume velocities, are not repeated.

139.

Plate VII -

Evaporating Oven and booster-element. The small quantity of residue can be seen in the sample bottles.

140 August 4th, 1957

Observations. F1ume Ve1ocity "V" Nominal v

ft./sec.

Intake Q

lbs/min.

Actual Vo ft.lsec.

F1ume T

ppm.

=

2.0 ft./sec. Intake t 2 samp1es ppm.

Mean

t - T • 100 .. T

t

ppm.

%

"Upstream Position" 1.0

11.0

1.12

645

620 - 770

695

+ 8

1.0*

10.0

1.03

804

705 - 807

756

6

2.0

22.0

2.28

720

610 - 790

700

2.0*

21.0

2.18

930

815 - 872

843

-

3.0

29.5

3.07

825

930 - 841

885

+ 7

4.0

42.0

4.40

826

902 - 900

901

+ 9

5.0

47.0

4.90

650

732 - 857

794

+ 22

5.0*

50.3

5.2

928

713 - 944

825

-11

5.0*

47.0

4.90

849

785 - 725

755

-11

3 9

"Downstream Position" 1.0

10.0

1.04

750

680 - 527

603

- 20

1.0*

10.5

1.10

798

795 - 818

806

+ 1

2.0

20.0

2.08

560

800 - 455

627

+11

2.0*

19.5

2.04

881

885 - 906

895

+

3.0

30.0

3.12

650

527 - 568

548

- 16

3.0*

30.0

3.12

745

891 - 980

935

+ 25

4.0

39.0

4.05

540

530 - 673

601

+11

5.0 3.0*

48.0

5.00 3.14

720

484 - 421 730·-- 760

452

- 37 3

30.3

769

745 (continued)

-

2

141 Observations (continued)

August 4th, 1957

F1ume Ve1ocity "V" = 2.0 ft./sec. Intake

Nominal

Actua1 v ft.fsec.

Q

y

lbs/min.

rt.,sec.

Fiume T pp m.

t - T • 100

Mean

Intake t 2 samp1es ppm.

T

t

ppm.

"Straight Out Position" 1.0

10.0

1.02

610

350 - 590

470

- 23

1.0*

10.2

1.05

789

716 - 837

776

-

2.0

19.5

2.05

520

325 - 550

438

- 16

2.0*

20.0

2.06

834

778 - 902

840

+ 1

3.0

31.0

3.24

470

635 - 567

601

+ 28

3.0*

30.2

3.15

860

765 - 923

844

-

2

4.0

39.8

4.15

620

520 - 575

597

-

4

5.0

48.3

5.04

668

820 - 720

770

+ 15

2

Variations in Turbidity at Constant Velocities Fiume Velocity

=

2.0 ft./sec. - - - Intake Velocity

=

2.17 ft./sec.

"Straight Out Position"

= Tm

ppm.

Intake t 1 sample ppm.

Variation ~~l t - T • lOO

759 - 603

681

600

- 12

512 - 551

531

618

+ 15

628 - 611

619

504

- 19

451 - 540

495

374

- 24

452 - 402

427

442

+ 3

624 - 430

527

610

+ 16

F1ume T 2 samples ppm.

Mean T

Total T Average

T:

:Il

=

Ta

=

3280 546

Total t Average t

= =

T

3148 525

=

ta

142

Using the average values ta and Ta t

a

-

T

a • lOO

Ta

=

Applying the lav of probability to the resulta of t -T T • lOO above, the probable error in determining turbidity variations at this flume velocity (2.0 ft./sec.) isE

= ! 1~

for a single observation.

143.

Plate VIII -

Flume discharging at a velocity of 3.0 ft/sec.

144 Observations.

August

lst, 1957

Flume Velocity "V" = 3 .o ft./sec. Nominal v ft.,sec.

Actual Q v lbs/min. ft.Îsec. Intake

Flume T

ppm.

Intake t 2 samples pp m.

Mean t ppm.

t-T T

•lOO

%

"Upstream Position" 1.0

9.80

0.90

764

684 - 715

700

-

9

2.0

21.0

2.18

680

720 - 697

709

+ 4

3.0

32.8

3.40

720

667 - 718

692

-

4.0

47.0

4.90

737

755 - 713

TH

0

5.0

37.6

3.90

690

690 - 685

687

0

4

"Downstream Position" 1.0

9.6

1.00

742

640 - 565

602

- 19

1.0*

10.0

1.05

428

377 - 366

371

- 13

2.0

20.0

2.08

547

612 - 664

638

+ 16

3.0

27.4

2.85

555

604 - 636

620

+ 11

4.0

40.5

4.20

549

454 - 530

492

- 10

5.0

48.0

5.00

524

522 - 595

558

+ 6

5.0*

49.4

5.15

456

451 - 487

469

+ 3

"Straight Out Position" 1.0

10.0

1.03

530

695 - 550

622

+ 17

1.0*

9.8

1.00

390

319 - 468

393

0

2.0

19.1

1.97

510

548 - 450

500

-

2

3.0

28.0

2.90

548

588 - 490

540

-

1

3.0*

28.2

2.95

422

475 - 429

452

+

7

4.0

39.0

4.05

536

455 - 477

466

- 13

5.0

51.7

5.40

435

462 - 415

438

0

145 Observations. Variations in Turbidity at Constant Velocities Flume Velocity

=

3.0 ft./sec. - - - Intake Velocity

=

2.10 ft./seo.

••straight Out Position"

= Tm

ppm.

Intalœ t 1 sample ppm.

Variation (~l t - T • 100 T

839 - 701

770

778

+ 1

648 - 432

540

571

+

604 - 559

581

669

+ 15

640 - 655

647

553

- 14

556 - 620

588

706

+ 20

678 - 627

652

592

-

Flume T 2 samples ppm.

Total Average

}

•

Mean T

Total t = 3869 Average t = 645

3778 630 =

= =

T a

6

9

= ta

Using the average values Ta and ta' t

a

- T

Ta

a • 100

=

+

2%

Applying the law of probability to the resulta of

t -T T • 100

above,

the probable error in determining turbidity variations at this flume velocity (3.0 ft./sec.) isE

= : 9%

(single observation).

146.

Plate IX

Discharge from Mixing Basin ( Velocity

= 3.0

ft/sec.)

147 Observations.

August 7th, 1957 Flume Velocity "V"

Nominal v ft. /sec.

Intake Q

lbs/min.

Actual Vo ft.7sec.

=

Flume T

ppm.

4.0 ft./sec. Intake t 2 samples ppm.

Mean t ppm.

t - T • 100

912

- 19

T

%

"Upstream Position" 1.0

10.8

1.12

1130

2.0

20.0

2.07

1150

1062 - 1016 1039

'3.0

28.6

;.oo

1242

1265 -

955

1110

- 11

3.0*

32.0

;.32

672

658 -

700

679

+ 1

4.0

37.6

;.92

955

1100 - 1041 1070

+ 12

5.0

45.4

4.75

1050

1091 - 1121 1106

+ 5

862 -

963

-

1

"Downstream Position" 1.0

9.75

1.00

1498

958 -

877

917

- 39

1003 -

9'30

966

+

950 - 1000

975

-11

1050 - 1030 1040

+ 10

2.0

20.5

2.12

904

'3·0

32.0

3.32

1100

4.0

39.1

4.06

940

5.0

46.6

4.88

1060

760 -

930

845

- 20

5.0*

49.3

5.13

627

540 -

664

602

-

870

- 17

990 1035

+ 10

7

4

"Straight Out Position" 1.0

9.1

0.92

1050

750 -

990

2.0

18.1

1.90

940

1080 -

3·0

30.1

3.15

960

820 -

940

880

;.0*

;0.5

;.17

696

657-

657

657

4.0

39.2

;.10

960

1110- 880

995

5.0

47.5

4.95

1020

1040 -

5.0*

53.0

5.50

588

549 -

960 1000 595

572

-

8 6

+ 4

-

2

-

3

148 Observations. Variations in Turbidity at Constant Ve1ocities Flume Velocity

=

4.0 ft./sec. - - - Intake Velocity

=

2.0 ft./sec.

"Straight Out Position" Flume T 2 samples ppm.

= Tm

Mean T

Variation (%) t - T • 100

ppm.

Intake t 1 samp1e ppm.

647 - 758

702

682

-

697 - 750

723

592

- 18

768 - 628

698

664

-

5

638 - 809

723

680

-

6

582 - 754

668

668

0

712 - 680

696

548

- 21

Total T = 4210 Average ~ = m Ta= 702

T

4

Total t • 3834 Average t = 639 • ta

Using the average values of Ta and ta' t

- T

a

Ta

a • 100 = -

'JI,

Applying the law of probability to the resulta of

t -T T • 100

above,

the probable error in determining turbidity variations at this flume velocity (4.0 ft./sec.) isE

= ! 1%

for a single observation.

149 August 9th, 1957

Observations. Flume Velocity "V" Nominal

Intake

v

Q

ft.fsec.

lbs/min.

Actua1 v

ft.~sec.

Flume T ppm.

=

5.0 ft./sec. Intake t 2 samp1es ppm.

Mean t

t - T • lOO

T

%

ppm.

"Upstream Position" 1.0

10.3

1.06

950

848 - 778

813

-

4

2.0

20.2

2.10

932

963 - 948

906

-

3

3.0

29.1

3.02

853

1030 - 987

1008

4.0

39.2

4.10

871

964 - 807

885

-

4.0*

40.5

4.2

874

864 - 961

912

+ 4

5.0

48.0

5.00

938

913 - 907

912

-

+ 18

2

.,

"Downstream Position" 1.0

10.0

· 1.04

856

777 - 707

742

- 13

2.0

20.8

2.17

794

826 - 832

829

+

4

3.0

30.2

3.15

870

952 - 763

857

-

1

3.0*

32.0

3.35

632

616 - 653

634

0

4.0

38.0

3.95

798

787 - 845

816

+ .2

4.0*

38.0

3.95

734

704 - 693

696

-

5.0

50.0

5.2

772

841 - 724

782

+ 1

5.0*

47.3

4.92

750

723 - 812

767

+

5

2

"Straight Out Position" 1.0

9.90

1.02

865

715 - 757

736

- 15

2.0

20.1

2.10

788

898 - 780

839

+ 6

3.0

30.4

3.13

712

802 - 703

752

+ 6

3.0*

29.7

3.10

740

750- 674

712

-

4.0

39.0

4.08

785

792 - 835

813

+ 2

5.0

48.2

5.02

822

834 - 784

809

-

4

1

150 Observations. Variations in Turbidity at Constant Velocities Flume Velocity

=

5.0 ft./sec. - - - Intake Velocity

=

2.15 ft./sec.

"Straight Out Position"

ppm.

Intak:e t 1 sam.ple ppm.

Variation (%) t - T • 100

905 - 947

926

867

- 7

877 - 838

857

792

- 8

794 - 899

846

894

+ 6

747 - 862

804

799

- 1

851 - 880

865

859

- 1

Flume T 2 samples ppm.

Mean T = Tm

Total T m= Ta Averaae Tm

= 429S = 860

T

Total t = 4211 Average t = 842 = t a

Using the average values Ta and ta' t a - Ta T

• 100

= - ~

a

Applying the law of probability to the results of t -T T • 100 above, the probable error in determining turbidity variations at this flume velocity (5.0 ft./sec.) isE

=!

5~ for a single observation.

151 Observations. SamEle E!eerimental Observations Flume Veloci ty "V"

=

Monday, August 6th, 1956

1.0 ft./sec.

Current Meter Readings - 26 revolutions/min. at 7.25 a.m. 26 revolutions/min. at 8.25 a.m. 27 revolutions/min. at 9.25 a.m. Intake Velocity (Nominal-Actual) v ft.,fsec.

1.0

2.0

4.0

3.0

5.0

-

0.99

-

2.00

-

3.75

-

2.95

-

5.10

Time of Weighinga hr.-min.

lbs.

7 - 51

15

=

lntake Position

62

8 - 16

159

8 - 23

275

8 - 45

526

8 - 50

706

9 - 05

815

9 - 11

984

9 - 15

984

9 - 19

1184

Sample No.

Taken from

1

Flume

2

3

lntake lntake

Upstream

4 5 6

lntake Intake

Upstream

7 8 9

F1ume lntake Intake

Upstream

10 11 12

Upstream

13 14 15

Upstream

7 - 56

Sample Calculation for Actual Nominal v0

Weight in Tank

Flume

c

Flume Intake lntake F1ume Intake lntake

Vo

5.0 ft./sec.

Difference in Tank Weight = 200 lbs. Difference in Time 4 mins. = Therefore, Discharge through Intake = 50 lbs./min. From calibration curve (Appendix B) for the intake pipe, a discharge of 50 1bs./min. = a velocity v0 = 5.10 ft./sec.

152 Observations. SamE1e Ca1cu1ations to Obtain Turbiditl Monday, August 6th, 1956

F1ume Ve1ocity "V" = 1.0 ft./sec.

grms.

grms.

Sediment grms.

Weigbt of Sediment grms.

1

32.665

69.2

32.680

0.015

36.5

411

2

36.063

82.2

36.089

0.026

46.1

564

3

31.551

66.3

31.570

0.019

34.7

547

4

32.312

63.4

32.323

0.011

37.1

354

5

34.850

67.7

34.863

0.013

32.8

396

6

33.551

64.2

33.564

0.013

30.6

425

7

32.80

69.5

32.825

0.025

36.7

681

8

33.682

67.2

33.698

0.016

33.5

478

9

30.560

68.1

30.580

0.020

37.5

533

42.259

80.1

42.469

0.010

37.8

265

11

32.231

74.2

32.253

0.022

41.9

525

12

35.538

85.6

35.568

o.o;o

50.0

600

13

56.448

93.4

56.458

0.010

36.9

271

14

32.306

76.1

32.335

0.029

43.8

662

15

40.901

8;.6

40.923

0.022

42.7

515

(1)

(2)

(3)

(4)

(5)

(6)

(7)

Samp1e No.

10

c

Container Empty (Zero Wt.)

Zero Wt.

Zero Wt.

+ water +

+

Sediment

Weight of Water grms.

Tpty of Samp1e ppm.

Note. Co1umns (1) and (2) are se1f-exp1anatory. Co1umns (3) and (4) are the 1st and 2nd weighings respective1y. Columns (5) and (6) are computed and self-evident. Co1umn

(7) is co1umn (5) divided by column (6) and multip1ied by 106 •

153

CHAPTER VII "The Experimental Conclusions" Before attempting a critical evaluation of the resulta of the experiment, it would seem wise to recall to mind certain outstanding conditions which characterized the investigation and which have been amply discussed either in the preceding pages or in the Appendix B to this thesis; namely, a)

The wide fluctuations in the turbidity of the "natural" source. The erratic and unpredictable fluctuations in turbidity of the flume water, while accepted perforee as an experimental limitation, cannot be considered as representative of natural conditions except perhaps in the case of streams at flood stage. This unusual condition produced a sizable probable error in the determination of turbidity variations which was proportional to the flume velocity, or more exactly, one which was proportional to the quantity of water discharged through the flume.

This is

not surprising since the greater the discharge through the reservoir and delivery line, the lesa chance there was for particle settlement and resuspension.

This relationship, be-

tween Flume Velocity and the value of Probable Error due to Fluctuations in Flume Turbidity 9 has been shown graphically on the following page and an examination thereof indicates that an almost linear relationship exista between these two quantities.

--" --1 f i f r·c·r- ft-_,LFfl_,_LJ _1 __ ~ - 1· - ~ +'-i-' : :,-r-lr: +- ~- ·-~ -~ :·i 'r~- !• 11 i i r1r~ -0- -,_-- '=~ -~~ ~ !; ! ,- r= '-Fr-j ~ ~ q ,-r!-,.:r.- T+~: ~ ~-: c+ ~ ;.; !- r;· : ~ r ;)+·t ~~ - ii --~1 1 i 1 -t,· .t ~-~~=r 1-_f_:l r r: r ;_: :_~:; . LJ-f ~; i- j + ~-T. · :___ - · r ., +· ' · i ~t ·l'- r -· -fi-

-~-

I-

r-

1

1

r-j

r:_

1

-

J-."i -:-L

-

1 1

1 - -- 1 1 -f- -·

1-HH-f"H-.Jo't-1-~-·

1

j 1

;

j'

1 •·

1 • 1 ' • 1 . 1

r- ·T I-

:

j1

11

1

1''

1

T

1

.

·-:-r---·

1

i

1 :

11

·-·i

1

i :

1

-

t· ' :

!_

,.

1.!

1 •

1

11 1

-,

t'

.'

1

1

j

,-

1 · 1 ,-1 i.' - -·!

1

-

1

: ..

t- · ·· 1 : ; - , i ·· 1.t - .1 ..;....~ 1 • ï ' •--~rL-,r-r

·r_u_ ij

l'Il

,

1',

.1l • i.,

11 1 1 '~;:t ~ i1·-[=j 11.r--9 't'· l:' ·i. ·1 i, i 1' : i1 ~.:-~11 1[-I~LIIj ' · 1! ~ 1 --!i+1-~riF:+f~ i ft i_~j! 1- . . -i . jl 1 h· ~-li 1·-t· . . -!- 4-Li .,. 1 1 . - . . t . ~ .!,,. ·r--t r 1 - : ,:Ir ~ : -~ --Il- - -T!l·L!Jt--~- ~---'-w= --,,1 -~ !,, H-H. .~u . -~ f' -lt · j-t_, 1-, -~ 1 ! ! :! l. 1 ~ '~ -t . _· ~~ 1- -~~r- _-t-l-: 1=;r:~tt_i_ Tl ït ;- ·Lf+_ 1 1- -t r-- 1_~ !· l:o~,.. - - 1 t f- -r'-, ~--H-i:-l-f~l1f :i. j--· l-+-t -,•. ,.:. ~ , i i :-!-. --- - --1 --r- -~- -: :· Î hi r 1 ·

j

1

1 J

1 •

j-r

·t1-l: 1l ::n~-t::-ft=.~ ~: J ; j ~

- --~ -~-l

1

1

,

1

:

;

' :

·

t ' 1 '! i

;

; 1 j -~- -l -;- r

f -

1

Ï 1 1 1 1 1

'j

1 l,

1

1

-

1 1 ·-

1

1

1

1

l-+-t--Hf-'~1'ftt.I,U'I~0 r

~

l f · '

1

i

• ,

- 1- . •

j \- 1 :

· 1::

-- -"l'ii;

n:

1 ~-

1

1-

---- -

1 Ht

-~H-

~

1

L; ~ [--~~~;=_! r~! -tt~_-:_ r· • ~ 1-:--:-1-l+- 'Il_i : -~- 1 t'

•

'

-t

T~

--

i

nTl ~-r

- .1.-

L

·r

; 1 ,-, ï f

f-t 1r ·+r+ --H ·-+·-Ill - .- ;..~~ j--r- tt---ttri-1.Î

-~-

;_~-111!1..,

- -t 1 i 1 1 - _., , j-I : i i 1 r- - L"''~ -:-r·-~-~~-:--+- - r-I r . '. i rn · -;.~ -rt-ï: ; . ::i-i;·: ~F;-:--~ · -:-I~f~!' r-l=-(-=Fr~+± l-~+-p-,l~---+·-rr r-fi-r- -t · : 1 1 t -t-~--r-- - r- T 1· t-r---1 EF . ! 1·i-,-'-H , --··-: . -+-r~ -r- ; 1 - f- --r-r +r1 +l' : -. , 1~~ ~ 1-rt+t-1--n r ! - i - ·- :- ·- •~ ·j-r-- --!-rr -ï r + "' - t-· 1 r : . r 1

1

1 ,

•

1 , ; .

i_

1 1

1

1---·

; -;- - .

1- .--

-

-w

jf·-L 1"'~ · 1•:;! : ~- - ! -f+-f- 1 -l-l-1--'~ 1 1 ' -~-'-- ' ' . ' ' . r-~·-·-t·i·j HT r-:- -~ ~~ tl~-=~1'1 .: ! ; 1 : : : i ~ - ~: ~: -R~~I -T ~ c :-~-- [f -~~t-,_h·--~-"-._r-tt::-'+ 1 b- -i i ! f· 1 1- - . · · j --tt i -•j-l~- - : ·~-'-L·- _[ !-;! : i;! i- t 1; +f~t 11--: _i-·-Ef-_ -~~ , !lH·i-~ -1+1·-

t-

.. n

1..:.

14

1

1

1

.- ,..-t~ rj' ii-~ ·j' - . t --i -~ --1--1-

-

1

i

.: 1 T~ -'

rr-

Î 1· t

l .

- -

l'

~ ~~

-

-l -

!

1

~--r- -~-r-i+ ~ 1j· li·jl . 1t !-+-- - + 1 1

'l i~i 1 - i-L ! ,-r ! 1 1

t

1

-

-- 1

1--

1

1 .

-:.

1-

-

1

1

-

. -

~+r- -

1-

-1-

-

·-

-

~· '1--r- ~

1

1

1 --:'

1

-·-

1:; 1·r i 1- h r 1 ~--! 1 -t

r -1- -l,_ f -- - - · 1-' +~- -i !

-- -- ---

·--

1-

-

'F

1-

+-- -

--

:=

-

-

~

---

.±.-

f--

- t--

'·

r- --

-

r-

- ~-

---

-

- -

r -'--'r-·H-.----1-t -rrr f--t-1--+l_j-~ i_j_j__; -11

-

. 1- .. -

: -

__ -~--

_

- - -·-- -1 -

1J-t _,__ -ff 'f- .-

--

t-

1

1

.

t--

1

· ~- -~~4--F~j- r-

~-

1

.- -

;

1-'

-

1

r

l-i r--t-r--rr

1 ri

_j

1

-

H --f-H:-t:-1-H-+-t-H·-t-t++++

I-

1- .

--

-

f-

1'1. l'

--

H-+-+--H-t-++-t,~H-1.

•-_ r

-

1-

-

r-

---

~-

.

~-.l·t·,l -~~- ~~f . -.~ -~ --~~~;: _+ 1-+·+- +t r-t t t+~~+ -+-1 -+++ ~-1-1 1· r

· j-~

__ Hr-

·1-+--+H-;----+ +

11

,.

--,-r- i-

1-

t-r-H-+-+-t--1-

•

155 The importance of the relationship between these two quantities lies in the fact that the value of the probable error at any given flume velocity corresponds approximately to the range of the majority of turbidity variations encountered at that velocity; for example, at a flume velocity of 2.0 ft./sec. the values recorded for variations in turbidity ( t -T T • lOO ) lie in a range from - 15% to + 15% for the most part and the range of probable error at this same flume velocity is

± 1~.

With

these resulta it is difficult not to come immediately to the conclusion that the so-called variations between intake and flume water are not variations at all but merely the result of the wide fluctuations in flume turbidity itself.

While auch a

conclusion is admittedly possible, there would appear to be grounds for further investigation before rendering such an absolute verdict. b)

Turbulent flow and variations in depth adjacent to the intake. As mentioned in Appendix B, flows at or near critical depth occurred at times in the flume, and turbulence and variation in head over the intake resulted.

What effect, if any, did this

condition have upon the variation in turbidities of intake and flume water?

Pending further experimental investigation, the

final answer to this question must remain in the realm of conjecture; nonetheless, it might be anticipated that this condition hindered, to some degree, particle action akin to discrete particle movement and therefore qualified the answers that can be given to the original questions under investigation.

156 AP.lzt1a ot !!aulta. 'the obJeoUwa ot the preaen' •ZJHJr1Mat vere tvoto14• ttraU;r, to

4e\el'lline vhetber tntake poa1Uon a.t-.. coul4 .-Hec' a aip1ftoaà n4uot10Jl in the turb1d1\;r ot the •••r en\ertaa \ha poriJ ae-ooDdl.f, to deteraine to vhat ex\eet, 1t ..,., ·U... \''!lll"J of diaorete puoUole ...,._

•et Jd.cht 'be applicable

w

auch an Q'ftaU,caU.fJl•

So •tated, ,.,. ezpert.ental objeot1yea. (ail to take in\o aocaaat the eQeriMn\.al

OODdUiona-.eDOoun\~6, ~nd

therefore preMDt .-attn.

auh are \oo pneral to be tull;r anawreq· bJ a

atucl•

~but

aDI5

oer\ainlT not by the preunt - · ..bt . .\i8ra ottered - . _ , be pntaoe4 b7 the oond1 t.io~ al&U8e - MVt;id•r. HIIJ)l1J!IB oon41 tiou" - in

Ol'der

'x1-i1M~Ul aD4

t_p 1Je )oth ftltd aa4 plau1ltle.

!he tiret objeatiw wa to detendall 'lUther poeitin aloae Jd.ch' atteot the tvb1di 'J' ot wa\er entef'i.Dc •

leut•.

t'be ftl'iûle ia ••~

tore iatùe poaitie ut, tor the .-nt, the etteot of intake wlooit;r vUl 1M ctiaoounte4.

If the obeena~ ~ual7 ù.1Nlate4 an re-

arraDCM vith 'hia 14ea 1n 111Dd, i.Mn, 'tW" aD1 li'ftD.

n-

ftlue• ahow UDder 'hl rnpect1w bu41Na •t "Up•tnu", aD4

"Straiaht Ou.t"

oa11

be adèled. tos•.'her M.c.ebraiaall;r

e:rtrao\N tor eaoh ot the

thrM

poe1Uona..

Bach

ftlooit,,

"hwi'r.a"•

uc

a • • ftl'M

8PD 10 • •,.....

tbe reault ot aoabini.Dg not lep thU .fift, uèl •r• ot'teD.IP i.Jid1Yi4.t ftl.uea.

ta

OJ'

vi1l be

•SCb'

In extracUnc *be . . . . . . tbrH lùl-~ Y&l.MI

( +1~, + ~ ed ..,~) baYe boen .4ie,I"'O&&'Md on tbe Maia \ba' lMJ po~al.J OI'I'OI'

exoeede4. the JU4D1 of the ether

1Moln4.

Vhere

'he••

are olearl;r indioated (•).

ftl.Ma

ftl•• 1I)' unral UM•

•coeur oa 'b• tollovins

~·

pqe•,

prebable

tbof

157 Below, the values previous1y tabulated on Page 147 for a f1ume ve1ooity of 4.0 ft./sec. have been rearranged on the basis described above. Mean Values of

t ; T • 100 (%) At Flume Ve1ooity

=

4.0 ft./sec.

(Reference - Original tabulation on Page 147)

"Upstream"

"Downstream11

- 19

- 39*

- 17

-

1

+ 7

+ 10

- 11

- 11

+ 1

+ 10

+ 12

- 20

+ 4

+

-

••••

-

-la%

-2-do

-

-

5

.... Total: - 13% Means: - 2%

"Straight Out"

-

4

4% (* Questionable value - disregarded)

8

6

2 3

3%

EquiVàlent determinations of mean turbidity variations in relation to intake position have been carried out for the other flume ve1ocities. The resulta are conso1idated in the table below. Mean Turbidity Variations Mean Values of

t ; T • 100

(%)

(%)

for Different Positions

(Neg1ecting Effect of Intake Ve1ocity) F1ume Velooitl 1.0

Upstream Position

-

2.0

Downstream Position

Straight Out Position

2*

0

+ 11 (x)

0

3

0

3.0

-

2

1

+ 1

4.0

-

2

4*

-

3

5.0

-

2*

1

-

1

158 In the immediately foregoing tabulation, the value of + 11% (x) ~ be

s~spect

since it rèsults from observations which are deficient in

number of samples taken.

Such a condition, particularly at the low flow

of 1.0 ft./sec., would greatly augment probable error.

Means that have

been computed by excluding the questionable individual values referred to earlier are indicated

(*);

however, this is probablr a refinement of

little significance here since only three values out of nearly one hundred are rejected. Conclusions. A summary examination of the preceding table indicates quite clearly that no significant reduction occurred continuously at any of the three positions investigated.

If the resulta shown in this table were again

averaged over the whole range of flume velocities, the mean values thus extracted would vary between + 2% and - 3%.

Under these circumstances,

graphical interpretation of the resulta would appear redundant and unwarranted and it may be concluded that; "Where turbulent conditions of flow exist adjacent to an intake, the position of the mouth of the intake in relation to the approaching current will not, in itself alone, effect a significant reduction in

the turbidity of water withdrawn from the source."

159 The Second Objective. The second objective of the experiment was to determine to what extent, if any, the theory of discrete particle movement is applicable to the investigation.

The earlier discussions of Chapter V would seem to

have established that if this theory is applicable to investigations of the present type, then auch application would be on the basis of the following hypothetical truths;

a)

Particles entering an intake in the "straight out" or "downstreamtt position must undergo a change in direction.

b)

Particles entering an intake in any position must undergo a chanBe in velocity if the port velocity is greater or lesa than the vel-

ocity of the approach current. c)

Particles subjected to changes in direction or to a reduction in velocity tend to settle out.

If these factors are translated in terms of the present investigation, it might be assumed that a reduction in turbidity of water entering an intake is a function of the position of the intake and a function of the ratio between intake and flume velocities, or, using the symbols already employed, t -l'...;;. T • 100 ___

..

f Eosition,

v: J t -

T

Consequently, if for any given intake position the values of T • vo 100 be obtained for varying values of V , it should be possible to test the validity of the hypothesis. In the following example, the observations tabulated earlier for the "straight out" position have been rearranged in accordance with these vo requirements to show, in ascending values of lf , the corresponding values of

t - T

T

• 100.

In preparing this new tabulation, the following

160

points vere observed: a)

Since it is a reduction in velocity at entry that may be presumed to affect a reduction in turbidity, only those values for vo ïf that are less than unity are of interest here.

b)

As might be expected, different individual values of V and v0 , vhen combined, give nearly identical values for the ratio of vo therefore, it is possible to group tvo or more single

vr ;

determinations of

t - T • 100 and to consider the mean of the T

group as corresponding to a "nominal" value for the ratio

Vo

V •

In following such a procedure, tvo points are perhaps worthy of particular notice§ firstly, actual values of v0 are used to vo determine the individual ratios of " in order to maintain accuraey; secondly, while it is true that mean value for t - T

T

• lOO is not obtained by averaging the same number of

observations in each instance, the resulting mean values are nonetheless equal in probable error involved.

This is because

more individual values are available from observations made at lover flume velocities (vhere probable error is highest) than from the observations made at the higher flume velocities (where the probable error is least). For example, the mean value for t - T Vo T • 100 corresponding to a value for " of 1.0 is the result of averaging seven observations§ on the other hand, the Vo value for t-T T • 100 recorded for a value of V of 0.20 is obtained by averaging only two observations.

Hovever, six of

the seven observations in the first case have a probable error of about ~ lo% vhile the two observations in the second group have a probable error of only ~

5%

in each case.

Now, according

161

to the law of Least Squares already mentioned in Chapter VI, probable error varies as the square of the number of observations and where E is the probable error of one observation, the probable error resulting from averaging the resulta of "n" observations would

be

E = m

E

Therefore it is evident that if seven observations are averaged, each of which is assumed to

have

a probable error of

! lo%, the resulting probable error in the mean value is about +

4%.

In the case where only two observations are averaged, each

of which has a probable error of

! 5%,

the resulting probable

error in the mean value will be again about !

4%.

This is the

condition in the present instance, and the values tabulated for the quantity

t ; T • 100, although not obtained by averag-

ing the same number of observations in each case, are nonetheless of about equal reliability and it can be assumed that the probable error in each case is about

! 5%.

With all of the above in mind, the values for the "Straight Out" position are retabulated on the following page to show the desired vo relationship between t -T T • 100 (d) ~ and the ratio ~ •

Such a re-

arrangement is only preliminary to the final tabulation wherein the vo . values of " are presented in ascending order; hence, it is included here only as a sample for the sake of completeness and will not be repeated for the other positions.

Final tabulations for these other

positions a r e, of course, presented.

162 Preliminary Rearrangement of the Observations (Straight Out Position) Flume Velooit) (Aotual

Ratio Vo

v

Intake Ve1ooit) (Aotua1 vo

1.0

0.99

0.99

+ 13

1.0

1.97

1.97

+ 39

2.0

1.02

0.51

- 23

2.0

1.05

0.52

-

2.0

2.05

1.02

- 16

2.0

2.06

1.03

+ 1

3.0

1.03

0.34

+ 17

3.0

1.00

0.33

0

3.0

1.97

0.66

-

2

3.0

2.90

0.96

-

1

3.0

2.95

0.99

+ 7

4.0

0.92

0.23

- 17

4.0

1.90

0.47

+ 10

4.0

3.15

0.79

-

8

4.0

3.17

0.79

-

6

4.0

4.10

1.01

+

4

5.0

1.02

0.20

- 15

5.0

2.10

0.42

+ 6

5.0

3·13

0.62

+ 6

5.0

3.10

0.62

-

5.0

4.08

0.81

+ 2

5.0

5.02

1.00

-

v

Turbidity Variation t - T • lOO (%) T

2

4

1

163 Turbiditl Variations CorresEonding to Values of the Ratio of (Straight Out Position) Actual Ratio vo

t - T • 100 T

v

(%)

0.20 0.23

Nominal Ratio vo

Mean t - T • 100 T

v

(%)

- 15 - 17

0.20

- 16

0.33 0.34

0 + 17

0.33

+ 8

0.42

+

6

0.42

+

6

0.47 0.51 0.52

+ 10 - 23 2

0.50

-

5

0.66 0.62 0.62

-+

-

2 6

8 6 + 2

-

0.96 0.99 0.99 1.00 1.01 1.02 1.03

-

0

0.62

4

-

0.79 0.79 0.81

Vo V

o.8o

-

1.00

+ 1

4

1

+ 7 + 13

-

1 + 4 - 16

+

1

Equivalent tabulations have been prepared for the "Upstream" and "Downstream" positions, and these are reproduced on the fo11owing pages.

164 vo

V

Turbidity Variations Corresponding to Values of the Ratio ("Upstream" Position) t - T • lOO T

Actuel Ratio vo

(%)

v

-

0.20

Nominal Ratio vo

Mean t -T T • 100 (%)

v

4

0.20

-

4

0.30 0.28

-

9 - 19

0.30

- 14

0.43

-

0.43

-

0.51 0.56 0.52

+ 8

--

0.66

+ 18*

0.66

+ 18*

0.73 0.75

+ 4 -11

0.75

-

0.83 0.81

+ 1

-

3 6 1

2

3 0

0.53

3

0.82

0

1.00

+ 8

+ 12 + 38

0.98 0.99 1.09 0.91 1.03

- 9 + 4 - 3 *

Value is questionable and disregarded (cf. Page 156).

165 Vo

V

Turbidity Variations Corresponding to Values of the Ratio ("Downstream" Position)

Mean t-T • 100 T

Actual Ratio

t - T • 100 T

~

v

(%)

0.21 0.25

- 13 - 39*

0.2

- 13

0.33 0.35

- 19 - 13

0.34

- 16

0.43

+ 4

0.43

+ 4

0.53 0.52 0.55

+ 7 - 20 + 1

0.53

-

4

0.63

-

1

0.63

-

1

0.67 0.69

0 + 16

0.68

+ 8

0.79 0.79 0.83

+ 2 5 - 11

0.80

-

1.00

+ 2

-

Nominal Ratio vo

(%)

v

5

+ 1 + 2 + 10

1.00 0.98 1.01 1.04 1.02 1.13 1.02 0.95

+n

+ 2 + 2 - 19 +11

*

Value is questionab1e and disregarded (cf. Page 156).

166 With the foregoing information at band, it is now possible to graphically represent the relationship between Turbidity Variation and the ratio

Vo

if

for each position of the intake.

To do this properly

requires that the curve be fitted by using some precise and analytical approach such as the method of least squares.

For reasons which will

become apparent a little later on, the relationship between the two quantities

t - T • lOO T

(%)

and

Vo

ïf will

be assumed as linear in the

present instance. Curve Fitting - Method of Least Squares. This me thod is based upon the law of chance or random sampling and is designed to make the sum of the squares of the differences, or residuals, between observed and calculated values a minimum. Thus, if Rn

= the

residual of the nth pair of observations (xn' yn)

which are to be fitted by the equation y

=a

+ bx, then the residual

a + bx - y • n n

The term (a + bxn) is the value of y calculated to be paired with the value xn' whereas Yn is the observed value of y. and

~

The parameters a

are to be chosen in such a way that the sums of the squares of the

residuals are to be a minimum, i.e. 2 R

=

1: [
2

=

a minimum. 2

To meet this requirement, the first derivatives of ER with respect to

~

and

~

are set equal to zero, or, d(ER2) da

d(tR 2 ) db

=

2t [R ~ J = [a*] = 21:

21: (a+ bx- y)

=

21: (a+ bx- y) • x

0

=

0

167 For n pairs of observed values, the following simultaneous equations exist: I

-

na + bEx - Ey

II

-

aEx + bEx - Exy

=

2

0

=

0

Dividing both equations by n, we have the two normal equations: bi:x

I-a+-n

II

-

fz. • n

~ + bi:x2 - f!l. n

n

n

0

• 0

Thus we have two equations and two unknowns; whence, the values of a and b are determined. To 11lustrate the application of the method more completely, the values hereinbefore tabu1ated for the "Straight Out" position will be used as an example, and the linear equation relating t -l' T • 100 ( "y" ) Vo and V ("x") obtained. Sample Calculation - - - Application of the Method of Least Squares Observed "x" vo

v

Observed "y" t - T • lOO (%)

Calculated "y"

T

0.20

- 16

0.04

- 3.20

- 3.49

0.33

+

8

0.11

+ 2.64

- 2.72

0.42

+

6

0.18

+ 2.52

- 2.19

0.50

...

5

0.25

- 2.50

- 1.72

0

0.38

.o.oo

- 1.01

0.62 0.80

-

4

0.64

- 3.20

+ 0.09

1.00

+

1

1.00

+ 1.00

+ 1.24

- 10 1.42

2.60 0.37

- 2.74 - 0.39

- 9.80 - 1.40

Totals: 3.87 Means: 0.55

-

168

Substituting these values in the Equations I and II,

I

a + 0.55 b + 1.42

II

•

0.55 a + 0.37 b + 0.39

0 =

0

whence, the values of a and b are determined as a b

= =

- 4.67 + 5.91

and the linear equation for the relationship is given by t T

T • 100 = - 4.67

+

5.91

vo

v

Using this equation, values of y can be calo.ulated for the observed values of x and entered in the preceding table. values of calculated y culated y = - 1.40. observed

y = -

=-

A summation of these

9.80; therefore, the average value of cal-

This compares favourably with the average value of

1.42 and indicates that the square of the residuals -·

is undoubtedly a minimum. The straight line

y

=-

4.67 + 5.91 x may now be drawn as represent-

ing the line of best fit through the plotted observations for the "Straight Out" position.

The result of this procedure may be seen on

the following page. Equivalent calculations have been carried out for the

11

Upstream"

and "Downstream" positions and the corresponding equations have been derived.

These calculations as well as the graphical resulta appear

on following pages.

169.

L ! .. ,

1 1 1

1-r,--. !

. ! ,- .-1

!

: 1 ; 1

!

~:

QtJ1; 1 ,.,

1

i: : ;. ; 4~6fl ~

+

-

- -~-

170 Method of Least Sguares - - - "U:es tream Observed "!" t - T • 100 (~) T

Observed "x" vo

v

x2

11

Position ~

Calcu1ated

"z"

0.20

-

4

0.04

- 0.80

- 9.67

0.30

- 14

0.09

- 4.20

- 7.73

0.18

- 1.29

- 5.20

0.28

o.oo

- 3.26

0.56

- 2.25

+ 1.02

-' 0 -3

0.43 0.53 0.75 0.82

0

0.67

o.oo

+ 2.38

1.00

+ 8

1.00

+ 8.00

+ 5.89

- 16 - 2.28

2.82

- 0.54

-16.57

0.40

- o.oa

- 2.36

Totals: 4.03 Means: 0.58

Whence the two equations I and II are: a + 0.58b + 2.28

=

0.58a + 0.40b + 0.08

0

=

and 0

So1ving for !. and l' we have: a

=

- 13.56

b ...

+ 19.45

and

Therefore, the Equation of the 11ne of best fit is: y

= -

13.56 + 19.45 x

171 • .

+ +r+:-:

t- -1---1---J l- -~ -1- H - - r-: · ~· ~ · · ' ·! 1t :~Î- ~ ~~ j · -! l 1: t- ----'-f-1- 1-r----t-t---·- . .. t--~--~-~-- .....;_.,...... _· f-, ·- · ~

+ rtt

1

-·, :-. _·-_ ,: rl--i--H-rr· · ·P.=H= -=r=--,-r·- ' " 1-:-+--rti···-Ï! -t·

!

,

~

~

•

•

1

t

t

,

·

ÏÏ··'

1

- ,_ 1- - - - _,.. r- -i-+-'--,-,-~ -· - . . --- ~ 1

1 t-- r-h...LLL -• _:.....:....l.._L~ 1 ~ t\ t --,r-;

f---- ,+1

-rl,-:

t:-i-

- !-- r

1 1-

,.

..,

• T_,

-1 r·-1- 1 · - ~ ~-r·: i-H-r--tft-

1--ft '- f- ~-; t~ ~-~

• , .

-

r ,_

t-·

1

•

r--

r+

.1 •• , 1

r

r1

',

't

1 '

1 r

!1

1

1-1

'

1

=f=C±ti.-:-1::1~-~ ;

1

i

1

1

i

- . ,

1 1

;

i~ 1

,

1 • '

1

'

1

'1

, •

1 :

1

'

1

1

1

1 '

J:

l' .

1

[

i

1

i

i

· · : 1 ~~- • ; 1"' i i

~-

• 1

1

1

•

·

1

i·· t

!~'~ 1 ; 1 ! 1 -~--

1 1-t.

1

:1 ~ 1' : i

' 1

~-

i

••

1

•

1

j 1l ; -

.•

•

;

1 : ;

'

1

•

,

1

1

1

+±-!

1,- ·-ï

l1

· 1 1

': -· •

1

1·· -1

1 L - , --+~- · 1;- · ,ïr 1--+-++-t-+-H

i"' r-

1

1

li F--,wr-

Il

' T;

-- 1 1

~

· -

'

--H

. ,

- 1--t-if---H--+-1

j-

'

i t-+ --+--+-+-+++++

··1 · ! ~ '

-

ï - 1- · . · 1 1 ! ,-- ·- - +t 1

J

j

Î

1-r-

-~~}li ~~- 1

1 .

:-

1

u - :

1

~~ - - f - 1 t . i'o -1- '-l' 1 J

-+ i !-t-

-HI-IH~

' l l-+-t--r- H-+---+-+---+-+-1

'

1

i

! ~ -~ • ~ !1' --- - -' _

, , ';;! r ·-

1

1

ï

i; :e

;

g i·r:l -1-· ,i r,-J-,· 1 i 'jf ·-~ !- 1'1'--lL 1 ~r1-J··· 1 + . ·-r l -r_i ~ i · --~

\ ! li'-

1

1

i

l' 1

!•

1 · 1 1 1 r 1 1 j 1 1 , 1 '1 J i 1 1 1 j 1

l ! H-·f j r-i ! ii

1;

t

•

1

1: 1' i- - ~ · · ·

1 ;

, 1 ,,

1

:

1

, ' , .· ,~ .~ ' î . 1: 1l;1 ,~

1- "-r-t--~-- j- . . 1 1 1 I l -l~-rï-;- - -1--1 i -i i i. '

l'

·L!_:·. ·: i.

'i

'

i '

_j

rl

•

J"l

l

--

1---+-++ ++-+-+-t-

--,-f~·-HH'",~+++t---1 r1

-fr -t-

:-it ~--

~0

-

:--1-f---H l-~ r -rr' ·,'' f- . 1, - 1- ·r 1 1 , r--- '- 1 -l~ H--l --1-+-+-++-+-+-t- t--t 1- - ~ I~' Tj-: -!±~ï-+~t ri li 1r 1 T 1 l 1 .. ~-~--- ' : : J l-- 1 +-H-++-t-+-H :t-= -,.T -~-·1 +-:-1--J--~t ~-:·rt~ i-i tT· --j-: 'r FI:·-~;', 11 , 1 _ , - ~ - , r~1 ···~ ·_:_ïJ T l_r_ -' ·t11-+ ï- -! t 1 .- 1 · ; i 1 1 t i 1 · f 11 · 1 1l - ~~- t- H -. 1 t- î T , f- t·1+ - -1- -- ! !-- l · r i : 1. 1111! 1 1 i- i ·h -" - -+--+-+-~---+--+-+-+++ --r·t---j- 1.5L 1-r ; 1 ' 1 1 ! ' : - .. i . : 1 1 1 i : 1 \ 1 i . 1-·t f-

~-,- ~--

·· : tlL~ fTt r' r 1 ., .1 '· ~-~ -1 -· ' -1

•

·

-

H-.!--i-•1-ff+ +t+l11Rr[ -~-~:--t·-•-,' l :,. ',.. ~ ,i'. . ·n- · i-1-j-+- +t-r-n-· :· · i --· • .

'

1 1

1 :

· · · · · · · · ; 1 ~• ; ·t ~.

J

~-~-l--, + ~r.

1

'

'

1

1

'1 '-

1

.:

rt-

1

1

1

•

;

1

1 1 1

_--Lf --r -_cv~ -

----~ii 1,:-:: - ~": ,,:,:'!: ,~'- 1!J't--~~. --~=--r--~ ~ :- --t-t-1 ~ -t-t-t-H-+---t--t .± + -1 r· 1 ,- 1- r- 1 - •. '" , .. _E ~1 1 1 - - : ,- --: ; ,, i ; i 1 1 ' 1 1 -- -+-+-+-+-++-! -~-- r i F'j-1; ~H · -i i 1::: ! i '· -~ 1 ~ --~ 1 - 1 R~t-=-t:-: :t_-1-1-+-t-+-

1

1

~-.

~ -~T ~ -

r- ·1-

-

--

~

f-

_ ,_

-

11

~

1

- t---

- -

_ _ 1

-

_

·-

r-

· 1 r irH-; 1 1

,-

t-

1

1--i-

t---1- 1---

1

:·

~

Î

'

1

: 1· 11

'

· '

' ii· i

f ! rr fJi f[ . ! -~-1~- ;-,-FI=f:- ~~ - ~: r H-r-f ~-1 r·:--: ~-ï=-r~ --n-rr ---8 r-- -ï ï ·1

!

++--++-t--tl

---1- ---

--- ~

1- ·! '1"' 1 -,

:-i rt·-'-[~--~

1

1

. 1

:

i l' '

1'

1

1

1

:

1 1

-

-+ r

-- 1- - -lï

1'':

i-~-ï ï

1

1

~ 1 -- i- t· -t-t--HH·--++-H ' -·-1 rt - 1- j-·f---+-f---H---1 r...,---- f----1-'t-·H~

~~ --

lj

!

l r-·1·- il t· t-

1 1

:j. -:--

t- -

1

l

r

1

1 -_

--f-

1

1

1

1

-

--1----

rtf rl---

--

1

j -j

1

-r---t- t--++-++-1-

~ !.:~1.?j~~-~-~--t-+r-t-t~~-=: -_+f---++-1 1r

1 L~- -FF-t-r-t--t-1H-~-t:=t~-+-+ -- -+-1 1 -F·

IT

172 Method of Least Sguares - - - "Downstream" Position x2

Observed "!" t - T • 100 (~) T

Observed "x"

~ v

EL

calculated

0.20

- 13

0.04

- 2.60

0.34

- 16

0.11

- 5.44

0.43

+

4

0.18

+ 1.72

0.53

-

4

0.28

- 2.12

-

4.18

0.63

-

1

0.40

- 0.63

-

2.07

0.68

+ 8

0.46

+ 5.44

-

1.00

o.so

-

5

0.64

- 4.00

+ 1.54

1.00

+

2

1.00

+ 2.00

+ 5.78

- 5.63 - 0.70

- 25.62

Totals: 4.61

- 25

Means: 0.58

-

3.12

Whence the two equations I and II are: a + 0.58b + 3.12 •

0

0.58a + 0.39b + 0.70 So1ving for

~

=

and 0

and ~' we have:

a

=

-15.42

b

=

+21.2

and

Therefore, the Equation of the 1ine of best fit is y

=

- 15.42 + 21.2 x

- 11.18

-

-

8.21 6.30

J.20

";l''

' 1

1

1

L -, 1

'

1

1 1 1

175 Conclusions. A summary examination of the individual graphs for all three positions would appear to vindicate the earlier decision of assuming a straight line as the curve of best fit.

To have done otherwise would have required a

more definite pattern in the plot and many more points therein; moreover, it must be recalled that even the present curves are subject to a probable error which is equal to a Variation in Turbidity

(%)

of

! 5%.

Therefore,

whereas a curve of the exponential type might be equally suited, the present graphs are nonetheless suft.icient for the present purposes and may be accepted as truly representing the results of the experiment. An examination of the Combined Graphe on the preceding page indicates the following facts: a)

The slope of all three lines is identical in direction. It should therefore be concluded that reduction in turbidity is Vo a function of the ratio if and varies inversely with this ratio.

b)

The slope of the lines is guantitatively different. It should therefore be concluded that reduction in turbidity

is a function of intake position provided that the value of vo the ratio lr is suitable. vo c ) All three linas converge at a value of if egual to about 0.70. It should therefore be concluded that reduction in turbidity

is negligible at values of

vo ;r

which are greater than 0.70

irrespective of the position of the intake. d)

The slopes of the "Upstream" and "Downstream" lines are about equal but both exceed the slope of the "Straight Out" line by a considerable amount. It should therefore be concl uded that, under conditions similar to those of the experiment, the "Upstream" and "Downstream"

176

positions will

be

about equally effective in reducing turbidity

and that either of these positions will be considerably more effective than the "Straight Out" position by an amount which might be determined as follows: The "Downstream" line intersecta the "Straight Out" line at the point "n" on the eombined Graphs; therefore "O - n" is the range of values of

over which a comparison èan be made.

age value of

over this range is the value at the point "N" =

0.35 or one-half the value at the point "n".

The aver-

Drawing the line

"N - Q - S" to intersect both the "Downstream" and "Straight Out" lines at the points "S" and

"Q"

respeoti vely, the values for

corresponding Average Reduction in Turbidity can be read from the graph as: Average Reduction in Turbidity (Downstream)

•

a.Q%

Average Reduction in Turbidity (Straight Out)

=

2.5%

Therefore the ratio of the Average Reductions •

hl

8.0 ,

or, the "Downstream" position is more than three times as effective as the "Straight Out" position in reducing turbidity on the average. A similar analysis for the "Upstream" and Straight Out" positions, using the points "m", "M", "P" and "R", shows the Average Reductions in Turbidity to

be

the following:

Average Reduction in Turbidity {Upstream) Average Reduction in Turbidity (Straight Out)=

2.75%

Therefore the ratio of the Average Reductions== i:b~

,

or, the "Upstream" position is more than two and onehalf times as effective as the "Straight Out" position in reducing turbidity on the average.

177 e)

If the ordinates for any one of the lines are averaged for values ( 2! vo greater and less than 1.00 the resultant mean ordinateor mean Variation in Turbidity) would be nearly zero.

;r

It should be concluded therefore that the Mean Variation in Turbidity, obtained by combining the resulta observed for values of Vo

;r

greater than unity with those that are lees than unity, will

be negligible.

This confirma the earlier findings mentioned in

connection with the first objective and permits the two sets of conclusions to be correlated and summarized. Summary of Conclusions. The resulta of the experiment admit of the following conclusions: "Where turbulent conditions of flow exist adjacent to an intake, the position of the mouth of the intake in relation to the

direc~--

tion of the approaching current can effect a reduction in the turbidity of the water withdrawn from the source provided that the ratio between the intake velocity and the velocity of the current be substantially lesa than unity." "The reduction in turbidity effected is a function of both intake position and the ratio of these velocities, not of intake position aloneJ therefore, if suitably modified, the theory of discrete particle movement is applicable to the present investigation." "The reduction in turbidity effected is inversely proportional to the ratio between intake velocity and approaching current velocity irrespective of the position of the intake." "The reduction in turbidity effected by an upstream or downstream positioning of the intake is about three times more than that pro-

178 duced where the intake is placed straight out and at ninety degrees to the approaching current." Importance of Conclusions. . ,_ . _. ,.,

e~r

From a practical standpoint, the resulta of the experiment pr%+1de important information on the behaviour of intakes under conditions of turbulent flow and such information finds ready application in practice; for example, 1.

Rivera subject to flood stage and seasonal variations in discharge are frequently subject as well to large increases in turbidity due to the additional carrying power of the augmented flow velocity.

At such times of high turbidity full advantage

should be taken of the increase in river current by keeping the intake velocity at a minimum value and, if feasible and applicable to the case in question, by fitting out the intake mouth with a bell-shaped port facing upstream or downstream. 2.

Intakes placed in rivera of constant high turbidity because of the lack of suitable alternatives, could be designed in such a way as to improve the quality of the raw water admitted by merely selecting a proper velocity and position for the ports.

The general concept however of overriding importance is the conclusive evidence that turbidity reduction can be effected by a preferential positioning of the intake mouth and by a judicious selection of intake velocity. In effect, the intake becomes part of the water treatment plant and, b,y proper design, offers the prospect of improved water quality at less cost to the consumer.

179 Future Investigations. The resulta of the present experiment and the conditions which characterized it are sufficient in themselves to suggest the nature of future investigations.

The following questions in particular would seem

to remain unanswered in full: "What reduction in turbidity can be effected under conditions of tranquil flow adjacent to the intake?

Under auch conditions of

flow, does the order of preferential positions remain the same?" Rather than attempt to forecast the answers themselves - a procedure which has shown itself to be unwise - the writer intends to confine his concluding remarks to certain suggestions regarding experimental conditions that may now be anticipated in advanceJ namely, 1.

It would seem unnecessary to investigate flume flows in excess of 3.0 ft./sec. particularly where the size of the flume and other experimental limitations must be adapted for tranquil flow.

Professor Azerier's advice that approach velocity should

not be allowed to exceed 2.50 ft./sec. lends weight to this suggestion. 2.

Discharge through the flume or equivalent deviee must be sufficient to prevent fluctuating turbidity throughout the system; otherwise, the masking effect of the probable error produced is a decided disadvantage.

In the present case, a pump discharge of

about 7.5 cfs. would have been the minimum requirement to obtain ideal conditions in this respect.

As alternatives to such a

value for pump discharge, it might be suggested that a smaller pump reservoir be considered as well as a less circuitous delivery lina; ultimately, however, it becomes necessary to

180

actually test the assembled apparatus before the outcome can be known.

3.

Future investigators would do well to consider the use of turpidimeters now available for measuring turbidity.

The delaJs

and inherent unoertainties that accompany sampling are not favourable factors and can only be overcome by repetitions of observations. 4.

If a flume is again used for the investigations at tranquil flow, a different cross-section should be considered in order to avoid flows at critical depth and obtain greater submergence of the intake port.

5.

The present experiment has clearly indicated the importance of the ratio of intake velocity to approach velocity.

Indeed, thià ·:

ratio would seem to adhere more closely to the theory of discrete particle movement than does the directional factor involved in position.

Every effort should therefore be made to carry out

future experimenta at values for this ratio which are exactly similar for all three positions.

Appendix A

BIBLIOGRAPHY Note:

The Bibliography is numbered to correspond with the figures shown in brackets throughout the text of the Thesis, e.g. (*16). In addition, the reference material has been divided into three parts; PART l •••• Material which has a direct bearing on the experimental investigation reported in the Thesis. PART 2 •••• Material of particular interest in the positioning, structural design or operation of Water Intakes. PART 3 •••• Material of general interest in the investigation under consideration.

PART 1 1. LISCHER and HARTUNG, AWWA Journal, Oct. 1952, P.873. "In talees on Variable Streams". 2. BURDICK, AWWA Journal, Vol. 38, March 1946, P.315. "Waterworks lntakes". 3. McDONALD, AWWA Journal, Vol. 32, April 1940. "Waterworks Intakes~ 4. BURDICK, Engineering News Record, May 22nd, 1930. "Waterworks Intakes of the Middle West". 5. AZERIER, Transzeldorizdat, Moskva, 1940. "Vodosnabzenie na Zeleznodoroznom Transporte" (part 1).

PART 2 6. ROBERTS, Engineering News Record, ~roh 1949, P. 52. "How to Design River-Intake Pumphouses''. 7. GISIGER, Civil Engineering, January 1947, P. 24. "Safeguarding Hydro Plants against Ice Menace". 8. BAYLIS and GERSTEIN, Engineering News Record, April 1948, P. 80. "Fighting Frazil lee at a Waterworks".

Appendix A Bibliograpby PART 2 (cont'd) 9. REH, American Society of Civil Engineers, Proceedings 83, No. 1465 "Intakes". 10. AMERICAN SOCIETY OF MECHANICAL ENGINEERS, Transactions 75, P. 643. "Hydraulic Problems Encountered in Intake Structures". 11. ENGINEERING NEWS RECORD, Vol. 157, No. 25, December 20th 1956. "Near Detroit Clean Water is Where You Find It". 12. ENGINEERING & CONTRACT RECORD, Vol. 68, No. 7, July 1955. "Beamsville Waterworks Project". 13. ENGINEERING NEWS RECORD, Vol. 155, No. 22, December 1955. "Cleveland, Ohio, Taps Ie.ke Erie Again". 14. MUNICIPAL UTILITIES MAGAZINE, Vol. 92, No. 10, October 1954. "City of Regina Acquires New Water Supply from Buffalo Pond". 15. WATER & SEWAGE WORKS, Vol. 101, No. 3, March 1951. "New Type River Intake ••• Perforated Concrete Pipe". 16. ENGINEERING NEWS RECORD, Vol. 151, No. 45, November 1953. "River Intake for Industrial Water Supply to Tap Flood Flows Only." 17. ENGINEERING & MINING JOURNAL, Vol. 154, P. 84. "White Pine' s Water Intake". 18. GRANACHER, Civil Engineering, Vol. 23, August 1953. "Water lntake for Steel Company Built to Withstand Monogohela River Floods". 19. HOLTON, Water & Sewage Works, Vol. 101, June 1954. "Water lntake Cribe is Lonely Island". 20. RODEMEYER, Power Engineering, Vol. 57, February 1953. "New Boom Intake •• Will it Cut Screening Costs?". 21. ENGINEER, Vol. 193, March 14th 1952. "Water lntake ••• Coryton Refinery~ 22. MOSELEY, AWWA Journal, Vol. 42, June 6th 1950. "Design Features of Cleveland's Nottingham lntake".

23. BAILEY, AWWA Journal, Vol. 40, No. 7, July 1948. "New lntak:e Installation for Kodak Park".

Appendix A Bibliography PART 2 (cont'd) 24. WATER & SEWAGE, Vol. 83, No. 9, September 1945. "Kingston Filter Plant and Intake". 25. GRASSY, AWWA Journal, Vol. 35, April 1943. "Use of Turbidity Determinations in Estimating the Suspended Load of Natural Streams." 26. GLAZBROOK, Engineering News Record, Vol. 131, No. 23, December 1943. "Cantilever Bridge Buil t to Support Pumps and Intake of Water Supply System". 27. STEWART, AWWA Journal, Vol. 31, No. 5, May 1939. "Installation of 54" Steel Intake Line ". 28. ENGINEERING NEWS RECORD, Vol. 119, No. 12, September 1937. "Floating Water Intake into Place". 29. DRAKE, AWWA Journal, Vol. 27, No. 1, January 1935. "Anchor and Frazil Ice at Buffalo Intake". 30. ENGINEERING NEWS RECORD, Vol. 115, No. 25, December 1935. "New Intakes for Montreal". 31. HARDIN, American Society of Civil Engineers, Proceedings 1592, April 1958. "Water Intakes in the Detroit River". 32. MURRAY, American Society of Civil Engineers, Proceedings 1607, April 1958. "Water Intakes in the Niagara River and Lake Ontario".

PART 3 33. SHERMAN, Engineering News Record, September 1930. "River and Lake Supplies". 34. HUDGINS, Geographical Review, Vol. 27, No. 3, July 1937. "Waterworks Intakes of the Great Lakes".

Appendix A Bibliography PART 3 (cont'd) 35. HOWARD, Engineering & Contract Record, Vol. 61, No. 10, October 1948. "liater Supply in Canada During Past 60 Years~ 36. SCOTT, Municipal Utilities Magazine, Vol. 94, No. 3, March 1956. "Water Supply in Canada Taxes Engineer's Ingenuity". 37. ENGINEERING NEWS RECORD, Vol. 156, No. 50, April 12, 1956. "Extended Intake Improves City's liater". 38. MUNICIPAL UTILITIES MAGAZINE, Vol. 95, No. 1., January 1957. "Statistics of Canadian Waterworks Systems by Provinces".

REFERENCE TEXTBOOKS WATERMAN, "Elements of lia ter Supply Engineering". 2nd Edition, Wiley & Sons, Chap. XII, P. 152. INSTITUTION OF 'WATER ENGINEERS, "Manual of British Water Supply Praotice", 2nd Edition, w. Heffer & Sons Ltd. TURNEAURE &: RUSSEL, 'tpublic Water Supplies'', 4th Edition, Wiley & Sons, Chap. XIII. STEEL, "Water Supply and Sewerage", 3rd Edition, Chap. VI. McGraw-Hill Book Co. FAIR & GEYER, "Water Supply and Waste-Water Disposal", Chap. Wiley &: Sons.

x.

BABBIT &: DOLAND, "Water Supply Engineering", McGraw-Hill Book Co. DAVIS, "Handbook of Applied Hydraulics", 2nd Edition, McGraw-Hill Book Co.

APPENDIX B

Foreword. Stated in general terms, the experimental investigation reported in this thesis is primarily concerned vith the possibility of reducing the turbidity content of naturally flowing waters, not vith the laws governing the flow itself.

The interest here is in turbidity and not in

hydraulics; nevertheless, it is quite evident that the two considerations cannat be entirely divorced particularly when contemplating the design and assembly of experimental apparatus. In order to keep the primary objective of the investigation in sharp focus and still obtain a measure of brevity and completeness for the text, the apparatus listed in Chapter VI is very summarily described, little explanation is given for the dimensions of the elements used and only passing reference is made to their hydraulic performance.

Discussion

on all of these matters, deliberately held over until now, constitutes the purpose

and scope of the present appendix.

In the interest of completeness and continuity, the factors discusaed are presented in the approximate chronological order of their appearance; namely, Part I - Experimental conditions established by preliminaty stugy and investigation. Part II - Calculations preliminaty to the design of some elements of the Experimental Apparatus. - 1-

Part III - Report on the Performance of some elements of the Experimental Apparatus.

Part I - Experimental conditions established by preliminar.y investigation and study, The Known and Probable Factors, Preliminary investigation and study of the contemplated experiment had served to establish the following as either probable or known factors. In order to simulate the flow that might characterize a section of a flowing stream, it would be necessary to use a pump of large capacity.

The largest available in the Test Laboratory

was a horizontally driven centrifuga! pump with a rated capacity of 10 cfs. and a variable speed motor. The high discharge contemplated would make direct discharge into the test flume impossible; bence, some form of receiving reservoir or mixing basin would be required to receive the water at the upstream end of the flume. Vater delivered by the pump to the mixing basin passed through an overhead main that bad several bende, valves and crossconnections in it including a venturi meter.

It was anticipated

that these obstructions might harbour deposits of the material used to create the artificial turbidity of the water; however, any modification in the delivery line would have caused a major displacement of laboratory apparatus and the only feasible solution was to provide for regular flushing out of the delivery line and to precede all tests by a period of pumping to stabilize conditions.

- ii -

Test Flume.

Proposed flume velocities were from 1.0 - 5.0 ft./sec.

The high flume discharge contemplated made it desirable to locate the flume at a point where the effluent could be discharged directly above the pump reservoir.

This would also ensure that, to a

certain degree, a minimum of subsidence would occur in the reservoir. Some leakage and spillage from the flume had to be expected and provided for in advance.

All of these considerations made it advis-

able to place the test flume on and partially within an existing laboratory flume which terminated directly above the pump reservoir and being made of steel plate was aufficiently strong to carry the superim.posed load.

At a point some fourteen feet distant from the

end of the steel flume, a connection could be made to the overhead supply main by merely substituting one fitting for another and thereby delivering water to the mixing basin.

By allowing the

test fiume to projeot beyond the existing flume, a total fiume length of some 16 feet oould be obtained.

Since the exact discharge from the flume was not of primary interest, quantitative measurement vas not considered.

Instead,

the slope of the flume vould be determined as vell as the approximate depth of flow and these values together vith the known velocity would provide sufficient data !or computing the approximate discharge.

A piezometer was introduced into the flume opposite to the

intake in order to obtain the depth-of-flow measureaents and to provide a check in the event of fluctuations in water level. Velocity of flow in the fiume vas to be determined by the use of a Gurley Current Meter (Priee Pattern) located just upstream of

- iii -

the intake position and with the rotating buckets of the meter placed at the same elevation as the intake port.

In order not to

interfere with the flow pattern adjacent to the intake, the current meter was not to be used concurrently with any measurements at the intake. Intake Velocities - Port Velocities. were from 1.0 to 5.0 ft./sec.

Intake velocities contemplated

Port velocities were to be in the

normal range from 0,20 to 1.0 ft./sec. - the exact relationship being determined by the shape of the bell-aouth port attached to the intake conduit. Since a large capacity weighing tank and scala were convenient to the site of the experiment, it was decided to determine intake velocity by veighing the discharge therefrom over a known period of time.

With the diameter of the intake pipe known, the relation-

ship between velocity and discharge could be established with ease. Turbidity.

Turbidity would be artificially created by adding

Magnesium Silicate (talc) to the flume water.

This substance was

selected because it was the !east abrasive of the silicates commercially available at the time.

It had a specifie gravity of 2.6

and could be procured in 50 lb. bags.

Two particle sizes were

available - 49 and 149 microns (J.&) - and it was decided to add a mixture of equal parts by weight to the water in sufficient quantity to produce a turbidity between 500 and 1000 ppa. Determination of turbidity in parts per million (ppm) would be made by weighing each sample before and after evaporation of the water content (cf Sample Calculations - Chapter VI). - iv -

Evaporation

of the water content vould require a standard evaporating oven

vith the temperature augmented to a considerable degree by another source of heat because the saœples of water taken would have to be large in order to contain any appreciable amount of sediment. It was decided to make use of 50 c.e. sampling containers to ensure that the samples vould contain from 35-40 c.e. of vater and yet be handled without excessive danger of spilling.

Actual veighings

would be carried out on a standard laboratory ecale reading directly to 0.001 grama. Samplipg.

Sampling of the flume water vould be done directly from

the flume at a point dovnstream from the intake.

Sampling of the

intake water would be done at the rim of a small circular veir over vhich the influent water would be made to fall thereby vithdraving the samples from a freely and uninterruptedly discharging flow.

Part II - Calculations preliminary to the design of some elements of the Experimental Apparatus. Elements to be designed. With the aforementioned factors in mind, preliminary calculations vere necessary for the design and fabrication of the following. (a)

The model intake conduit.

(b)

The small circular wetr over which the intake water vould be discharged prior to sampling and termed the "over-fiow"

weir for convenience. (c)

The flume.

The Model Intake.

A section of brasa rod vith an inner diameter of

about 1/2 inch vas choaen for the intake pipe. -v-

lt vas fitted vith

V;?

a flared end-section made from copper sheeting l/32 in. thick and soldered to the rod.

Two auch intakes vere prepared in advance

and the dimensions of each are shown on the following page (Drawing D). Let

The se dimensions were arrived at as follows:Model lntake Pipe Diameter

=

o. 684 inches

=D

Model lntake Pipe Area Max. Desirable Port Velocity

=

• A

=

0.002562 sq.ft. abt. 1.0 ft./sec.

Max. Intake Velocity

=

~.o ft./aec.

Area of Port

= A1 sq. ft.

=

Diame ter of Port

D1 inches

Then for continuity of flow through intake,

whence, A1

=

0.0130 sq.ft.

These are minimum values.

and D1

= 1.53

inches.

The actual values used in the experiment

were as follows:Port Area = 0.0180 sq.ft. and Diameter of Port

= 1.813

inches.

Consequently, with a length of flared section selected as 1.75 inches, the ratio of length of section to minimum diameter vas 2.5 and the value of the angle of flare (8°) was equal to about 18°. At these values some contraction of flow at the port is to be expected, and a discharge coefficient (c) of 0.90 vas applied to the port area (cf. Handbook of Hydraulics - King), hence, the Ratio of Port Area to Area of lntake conduit is 0.0180 x 0.90 0.00256

=

6 •3

and the ratio between port velocity and intake velocity is the - vi -

Drawing "D"

~1

.. '"

~

~

'

~ \

~

~

~

'!(

!

'1

'\ll v'

,,'' '

.1

~

-vii _

'•,

,, i;

~

""

~~-tt~).

~l". '~) s

~~~~;

~

~"; ;

reciprocal or 0.159. With these relationships established, a calibration curve vas

"Q"

at the intake in

lbs./min. to the Intake velocity "vo " in ft./sec.

A similar rela-

plotted for the intake relating the Discharge

tionship was established between

"Q"

and the port velocity "v".

Both curves are reproduced on the following pages of this Appendix.

- viii -

Appendix

ï

l

.

1

ï

r-

ix

Ca ibr.ation

- x -

The Overflow Weir.

Water entering the intake was to be conve;red to

a circular weir over which it would fall in a free and uninterrupted discharge thereb;y providing an ideal sampling point.

It therefore

became important for the up-flow velocities through the weir to be sufficient to prevent any particle settlement at the lowest discharge from the intake, i.e. at an intake velocit;y of 1.0 ft./sec. The specifie gravi t;r of the magnesium silicate used in the experiment was 2.6 and the largest particle had a diameter of 149 microns (~) or 0.149 m.m.

According to Steel* the hydraulic sèttling

value of auch a particle would be about 15 mm./sec.

through still

water at 10°C. (cf. Water Supply and Sewerage - 2nd Edition)

and

ma;y be computed direotly using Stokes' law as modified by Hazen; namely, Where

v

= =

s s' •

velocity of subsidence in mm./sec. specifie gravity of the particle = 2.6 specifie gravity of the liquid (abt. = 1.0)

=

d = diameter of the particle in mm. 0.149 t = temperature of the 1iquid in degrees F. = then v=

and

v/305

418 (s - s ')d2 (t + 10) 60

=

=

15 mm./seo.

0.049 ft./sec.

Thus it is evident that the up-flow velocity in the overflow weir must exceed 0.049 ft./seo.

In the experiment a velocity of

0.0525 ft./sec. or 16 mm./sec. was selected as being sufficient to prevent any settlement and since continuity of flow requires that

-xi-

then if A

=

Area of the intake conduit

=

0.002562 sq.ft.

Minimum velocity in conduit = 1.0 ft./sec. v1 = Minimum velocity in up-flow • 0.0525 ft./sec. A1 = Area of circular weir in sq. ft. V =

0.002562 x 1.0 0.0525

•

0.049 sq. ft.

and the diameter of the circular weir

=

0.25 ft.

=

3 ina.

The overflov weir fabricated for the experiment had a diameter of 3 ins. and an overall height of 18 ina.

Details of the weir

are shown on drawing "D" (Page vii of this Appendix). The Flume.

Due to the length of the experiment, the pump could be

expected to operate eight hours a day and therefore, it was decided to pump at 2/3 maximum rated capacity or 6.5 cts. rather than attempt prolonged pumping at full capacity. In order to facilitate construction, only the rectangular and trapezoidal shapes were considered for the flume.

Since a

trapezoidal shape provides greater increase in oarrying oapacity per increment of depth than the rectangular shape, the fiume vas built with a trapezoidal cross-section and vith aide slopea of 1:1. Several bottom widths and heights vere investigated within the limitations imposed by laboratory space and vith the following points in mind, (a)

Non-uniform flow in the open flume had to be anticipated It seemed reasonable to presume that at some point in the

variation of flume velocity, flow at or near critical depth might occur and unstable flow conditions prevail. Adequate free-board had to be provided against this possibility. - xii -

(b)

Since the water arriving at the flume vas to flow upwards through a mixing basin, vater influent to the fllDDe would have little or no component of velocity in the direction of the flume and most of the energy at this point vould be potential energy.

Upon entering the flume, JIUch of this

potential energy would become kinetic energy; therefore, an entrance drop should be anticipated and the mixing basin carried to the full height of the f1 ume. (c)

Provided that adequate intake submergence vas obtained as vell as the desired velocity of flow, the exact quantity of water discharged vas of no interest in itself.

Further-

more, since the velocity in the flume could be controlled by varying the slope thereof, the discharge adjusted by throttling vith a gate valve as vell as by varying the speed of the motor, there appeared to be sufficient flexibility of control and little need for lengthy preliminary calculations. The following computations constituted the basis for the fiume design and for the dimensions shown on the drawing "B" on the following page. If and

Maximum Discharge = 6.5 cfs. = Q Maximum Velocity = 5.0 ft./sec. = V

Then, Area of Cross-Section Required

=

QjV

=

1.3 sq.ft.

Assuming trapezoidal flume, side-slopes 1:1 and bottom width If Depth Area

= =

(1)

= 1.00

0.75 ft. 1.312 sq.ft.

=

Area required in (1) above.

Therefore, assuming a flume constructed vith a bottom width of -xiii -

ft.

Drawing "B"

'\

- xiv _

Plate X -

Baffle placed in t-·!ixing Basin •

-xv-

1.0 ft., 8ide-slopes 1:1 and height of about 1.0 (to provide 80me free-board), the Friction Lo8s in such a flume flowing at 5.0 ft./8ec. and discharging 6.5 èf'a. can be computed from the Manning Formula, 1.486

v • where

n

r

2/., 1/2

(2)

8

V n

= a

velocity in ft./sec. coefficient of roughness vith planed face

r

=

hydraulic radius in ft.

=

area of cross-section of flow wetted perillléter loss in head in ft. per foot of length.

s

=

0.010 for wood flumes

=

Sub8tituting in the formula for, V = 5.0 ft./sec. n = 0.010 and assuming a depth of flow

=

0.75 ft.

then, Area of cross-section = 1.31 sq.ft. Wetted Perimeter = 3.12 ft. and

r

=

0.42 ft.

whence

8

=

0.0036 ft. per ft. of length

=

Assuming length of fhune

16 ft.

then, head lost in friction at maximum velocity

= =

hf

= abt.

0.06 ft.

1 inch (max.)

Specifie energy head He at that point where

=

=

5.0 ft./sec. Depth of Flow = D = 0.75 ft. Velocity

V

is given by the relationship, v2/2g + D

=

He, and by substitution H8 -xvi-

= 1.14

ft.

(3)

Critical Depth.

For a

section where

trape~oidal

B = width of flow at water surface (ft.) b = bottom width (ft.) He = energy heat (ft.) as given in (3) on previous page. Then,

!!

b x He = Critical Depth = D0 (4) 5 and for the assumed D = 0.75 ft., V = 5.0 ft./sec. and the proposed flume cross-section where b D c

= 1.0

=

ft. and B = 2.50 ft.,

0.84 ft.

Critical Velocity, or the flume velocity when flowing at the critical depth and with the discharge unchanged, would be .],_ A

where

At D0

= vc

c

(5)

A c

= =

=

0.84 ft., the value of A 0

Q

=

discharge cfs.

6.5 cfs. as before.

Cross-sectional Area of Flow at Critical Depth (sq.ft.)

= 1.55

sq.ft.; whence by Equation

(5)' V c

=

4.25 ft./sec.

The Critioal Slope, or the slope equivalent to the critical ve1ocity, may now be computed from the Manning Formula (2) as follows,

With V0

=

4.25 ft./sec.

and D0

=

0.84 ft.;

=

hydraulic radius at critical depth

=

critical slope =

Then r and s

0

c

A0

=

- xvii -

1.55 sq.ft.

=

0.454 ft.

0.0024 ft./ft. of length.

Resulting Effect.

From the aforegoing it is clear that the depth

of flow selected is less than the critical depth and the velocity of flow selected greater than the critical velocity.

Rapid flow

rather than tranquil flow will result; aoreover, since the flume is small in overall dimensions, the difference between the selected depth D (0.75 ft.) and the critieal depth D0 (0.84 ft.) is small and any irregularities in cross-section may produce unstable as vell as non-uniform flow. Channel Entrance.

An "entrance drop" between mixing basin and flume

bas already been anticipated in the earlier discussion of this Appendix.

The approximate height of this drop may be computed by

the following formula (Handbook of Hydraulics • King ), h = difference in head between energy gradient and the water surface just downstream from the entrance drop (rt.) H8

=

specifie ènergy head (tt.) available at the mixing basin measured from the flume bottom as datum.

D

=

depth of water (ft.) flowing in flume.

a

H

Then, h

(6)

e - D

Furthermore, when the depth of flow is less than ctitical depth, then D beeomes equal to D0 or the critical depth.

Such is the con-

dition in the example presently being considered; bence, In the present example,

h

=

He - Dc Since H8 = v2/2g + D (Equation 3) Then ha v2 j2g + D - D0 or h is approximately proportional to the square of the velocity in the flume; consequently, the maximum - xviii -

drop should be at maximum flume velocity as in the case under discussion where V= 5.0 ft./sec. and He= 1.14 ft. (Equation Since D0

= 0.84

3).

in the present case, then, h

=

0.30 ft.,

or about four inches.

Flume Flow at Different Velocitiea.

With a view to obtaining some

idea of the flow conditions that would prevail at the other velocities contemplated in the experiment (from V= 1.0 ft./sec. to 4.0 ft./sec.), equivalent theoretical computations vere carried out for these values. It vas readily foreseen that modifications large and small might be required as the experiment progressed, and the intention here vas

simply to obtain some guide and not exact information.

In the

folloving table are shown the computed values for Q, He and D0 where the value of Dis kept constant at 0.75 ft. and the velocity varied from 1.0 ft./see. to 4.0 ft./sec., the slope of the flume being adjusted accordingly. Constant Depth of Flow Constant Crosa-Sectional Area Constant Hydraulic Radius Velocitz Flume ft./sec.

v

Dis charge Q Flume ch.

1.0

1.31

2.0

= =

0.75 ft. 1.31 ft.

=

D

=

A

=

0.42 ft.

=

r

Energ[ Head He Flume ft.

Critical De2th D 0 Flume ft.

2.62

0.76 0.81

0.56 0.60

3.0

3.93

0.89

0.66

4.0

5.25

1.00

0.74

Examination of this table indicates that at the lover flume velocities (1.0- 3.0 ft./sec.) the flow might be tranquil since -~-

the critical depth is less than the proposed depth of flow.

At the

higher flume velocities of 4.0 ft./seo. and 5.0 ft./sec. this situation is reversed and rapid flow should occur; turthermore, at 4.0 ft./seo., the proposed depth and critioal depth are very nearly equal and fluctuations in water surface should be antioipated. Summary.

On the basis of these theoretical caloulations, the pro-

posed flume was oonsidered adequate for the experiment in hand as well as within the praotical limitations imposed by site and availability of laboratory apparatus.

The assembly of the flume

and other elements of the apparatus was therefore commenced along the lines already indioated and to the dimensions shown on the following drawings "A" and "C".

- xx -

Drawing "A"

i

~

~w

-xxi -

.

'

Drawing "C"

1

...

'1

1

1

~

·. ,.. 1

1

1

4:..

~·:.

1~

1

;-: ... .~......

1

~ .,,·:-

1

'

xxii

:

Part III - Report on the Performance of Some Elements of the Experimental Apparatus. The Model Intake. Performance of the designed intake tubes was satisfactory throughout the experiment.

The provision of a union immediately inside the flume

proved to be a great convenience and permitted changes in the position of the intake to be effected rapidly. The Overflow Weir. Performance of the weir itself was satisfactory throughout the experiment, and there was no visual evidence of particle suspension or settlement at minimum flows.

The method utilized for raising or lowering

the weir - a sliding block within a guiding channel - was neither convenient nor satisfactory.

With room humidity varying considerably, the

fit of the wooden blook in the wooden guide channel was either too tight or too loose and only by continua! waxing of the mating surfaces was it possible to obtain reasonably fine adjustment. The Flume. Prior to embarking on the actual experiment, the flume was set exactly horizontal and subjected to trial flow, at a velocity of 3.0 ft./sec.

During this period of trial, certain desirable and minor

modifications became apparentJ namely, (a)

Provision of a baffle in the mixing basin to reduce the high turbulence of the upflowing water.

(b)

Installation of a piezometer in a forward corner of the mixing basin in order to facilitate the observation of head at this point. - xxiii -

In general the hydraulic performance of the flume was very similar to the forecasted behavior outlined in Part II of this Appendix largely because the assumed depth of flow of 0.75 ft. did, in fact, prove to be a convenient depth. As mentioned earlier, the actual value of the flume discharge was of no interest in itself; nonetheless, during the course of the experiment certain additional quantitative measurements were made with a view to establishing approximate values for the discharge flow (D) and the energy head (He). s

= slope

(Q),

the depth of

These observations were as follows:-

of the flume (ft./ft.), using an Engineer's Level

and rod to an accuracy of 0.01 ft. He= energy head (ft.) • elevation of the water surface at entry to the flume in relation to the flume bottom as datum.

This was measured by the piezometer mentioned

earlier and could be ascertained within 0.02 ft. at the lower flume velooities (1.0- 3.0 ft./sec.) and within 0.04 ft. at the higher flume velocities.

Using these observations, caloulations were carried out in each instance to obtain the theoretical values for disoharge, depth of flow and critical depth shown on the following page in diagrammatic form. It must be stressed that these values are the result of theoretical calculations and not quantitative measurement; henoe, particularly in the case of the depth of flow at the higher flume velocities, they do not represent necessarily the exact conditions that occurred in the flume.

Their purpose at the time of the experiment, and here, is to

serve as a guide in understanding the phenomena reported. - xxiv -

CHARACTBRISTICS OP TIIB PL~ IN THB PLUMB AT TIIB VARIOUS VBLOCITIBS

11'1

+·-· 1

Htxmp' ....• ..

.

l ...-. ~ · ·- ., ,

Sketch is not to scale and diagrammatic only. tesend

WV • Velocity in Plume, ft/sec. *He • Total Bnergy Head above flume bottom he • Bntrance drop at entry, ft. he • Head loas due to friction, ft.

at entry, ft.

*D • Depth of flow near intake, ft. *hs• JJead 1oas ( or gain ) due to alope of flume, ft. 2

.!_ • Velocity head near intake, ft. 2g

-Note:

(*)

indicates observed valuea, the reaaiDder are obtaiaed by computation. The fo11owing ~lues were &lao c0111puted;

Q • Discharge of Plume, cfs. De• Critical Depth of flow a.t actual depth D, ft. OBSBRVBD AND Cœ1PU1'BD VALUBS V•

He*

1.0

1.02

o.1s -o.2s

0.01

o.o1

o.o1

1.31

o.s6

2.0

0.92

0.75

-0.10

0.06

0.01

o.o6

2.62

0.60

3.0

0.95

0.75

o.oo

0.14

0.06

0.14

3.93

0.66

4.0

1.00

0.60

+0.08

0.2S

o.os

o.2s

3.84

0.62

s.o

1.10

o. 75

+0.11

0.39

o.o1

0.39

6.55

0.84

D*

Q

h6 •

Note - ( a ) Velocity of approach at flume entry asaumed • 0; heoce Jte • y2 (b) Algebraic signa of ha indicate direction of alope, tbua, ~ i) + indicates forward alope as in sketch, ii) - indicates reverse alope. (c) Value hf c011puted as followa, hf • He - hs - D - he ; therefore, hf abaorbs &11 the : errors of measurement as well as the loaaes due to frlctiœ and turbulence. -xxv-

Flume Flow V= } ft./sec. In brief, the flume set in a horizontal position was found to have a velocity close to 3.0 ft./sec. and this exact velocity was obtained by adjusting the speed of the pump motor.

The flow in the flume at this

velocity was not tranquil and undulations in the water surface of about 0.15 ft. vere indicated by the piezometer adjacent to the intake which also showed a mean depth of nearly 0.75 ft. V = 2 ft./sec. The discharge end of the flume was then raised by 0.16 ft. and the

pump adjusted to give a velocity of 2.0 ft./sec. vith a depth of approximately 0.75 ft.

Flow at this velocity was more tranquil than before,

the undulations near the intake being reduced to 0.08 ft. on the average. V = 1 ft./sec. ·The discharge end of the flume was agàin raised to mak:e a total increment of 0.46 ft. and the pump discharge adjusted to produce a velocity of 1.0 ft./sec.

Flow at this velocity was definitely tranquil

and the variations in water surface recorded by the piezometer vere less than 0.04 ft.

The "entrance drop" at entry of the water from the mixing

basin was negligible and the depth of flow about 0.75 ft. V = 4 ft./sec. The inlet end of the flume was then raised to provide a total fall of 0.13 ft. and the discharge of the pump increased in an attempt to obtain 0.75 ft. for the depth of flow.

This proved impossible since the

undulations in water surface became too great for the free-board available - xxvi -

and spillage occurred.

The piezometer indicated fluctuations of 0.25

or more; indeed, it became almost useless as a gauge to water level since the undulations at this moment, where depth of flow approached critical depth, were extreme.

This possibility had been foreseen (cf.

Part II) and it had been hoped that the free-board allowed would have been sufficient to cope with the situation.

Since prior investigation

bad also shown that to increase the depth of flow within the limita available would not substantially separate the values of critical and actual depth at this velocity, the only remaining alternative was to reduce the flow depth and increase the available free-board by so doing. The resulting flow was rapid being at a depth equal or just less than the critical depth and although not stable, it could be contained within the flume without spillage.

The actual depth of flow under these con-

ditions of great variation in surface could only be estimated and a mean value of 0.60 ft. is recorded in the diary of the experiment. V = 5 ft./sec. The inlet end of the flume was again raised providing a total fall of 0.17 ft. and a velocity of 5.0 ft./sec.

As in the immediately pre-

ceding case, flow was rapid and turbulent; however, the fluctuations in depth were less pronounced and it proved possible to obtain a flow depth of about 0.75 ft. without spillage.

Fluctuations in water level recorded

at this velocity amounted to 0.20 ft. or more and a pronounced entrance drop occurred at entry to the flume. Flume Performance - Summarz. Among severa! conclusions that might be made with respect to the hydraulic performance of the f.lume, at !east two are probably of - xxvii -

consequence to the experimental investigation; namely, (a)

Equal depth of flow vas not obtained at all flume velooities, and a variation in port submergenoe of the intake resulted.

(b)

Flow conditions vere not the same at all flume velocities. Velocities of 2.0 ft./sec. and less produced relatively tranquil flow and nearly constant head conditions over the intake pipe.

At flume velocities of 3.0, 4.0 and 5.0 ft./sec.

fluctuations in depth were large and the variation in head over the intake vas considerable.

In terms of the submergence

of the intake, these fluctuations were as follows:Flume Velocity "V" - ft./sec.

Port Submersence (Max.)- ft.

Variation in DeEth Submergence

%of Port

1.0

0.50

~

2.0

0.50

16%

3.0 4.0

0.50

3~

0.38

65%

5.0

0.50

4~

The effect produced by these two conditions on the experimental observations can only be conjectured here; but in terms of the similarity to actual flow conditions over the average intake, such variations in head are difficult to imagine.

On the other hand, it must be recalled

that the head over the intake was only a small part of the total head differentia! being produced by lowering the overflow weir.

In other

words, the variations in flow depth within the flume had only minor effect upon the constancy of the intake velocity however greatly they may have affected the flow patterns at the intake port.

Unfortunately, much of the

interest of the present experiment is centered around the latter considera ti on.

- xxviii -