Scap-presentation At Csche 2003

Safety Evaluation and Control Measure Design In Offshore Process Facilities

Paul R. Amyotte Faisal I. Khan Faculty of Engineering & Applied Department of Chemical Engineering Science Dalhousie University Memorial University of Newfoundland Halifax, NS St. John's, NF

Outline of presentation

  

Risk Assessment Methodologies SCAP a New Methodology Application of SCAP to Offshore process facility

Methodologies available for safety evaluation and hazard assessment  

    

Hazard index – Dow index, Mond index Hazard and operability (HAZOP) study Failure mode effect analysis What-if analysis Fault tree analysis Event tree analysis Consequence analysis

fkhan:

Do we need a new methodology? 

No single methodology is able to answer:     

What may go wrong? How it may go wrong? How likely its occurrence? What would be the impacts? What control measures would reduce its impact and likelihood of occurrence?

A new methodology SCAP  

*

S- Safety, CA- Credible Accident, and PProbabilistic hazard assessment SCAP’s objectives:   



to identify hazards in an unit/industry, to quantify its probability of occurrence, to forecast its impacts in and around the industry, to suggest safety measures and then reassess the risk incorporating suggested control methods.

* Khan, F.I., Husain, T., & Abbasi, S.A., J of Loss Prevention in Process Industries, 15, 129-146, 2001

SCAP is developed by integrating: 

Safety Weighted Hazard Index (SWeHI),



Maximum Credible Accident Analysis, and



Probabilistic Hazard Assessment.

What would be the impacts?

What may go wrong?

Start

How it may go wrong? How likely its occurrence?

Hazard identification SWeHI Probabilistic hazard assessment-ASM

Quantitative hazard assessment- MCAA Accident scenario development •MCAS

Fault tree for the envisaged scenario

Consequences analysis •MAXCRED

What control measures would reduce its impact and likelihood of occurrence?

Fault tree development

Fault tree analysis •PROFAT

Apply safety measures and reevaluate risk

Risk estimation

Whether risk is in acceptance? Yes End

No

Suggest safety measures to control risk

Start

Hazard identification SWeHI Probabilistic hazard assessment-ASM

Quantitative hazard assessment- MCAA Accident scenario development •MCAS

Fault tree for the envisaged scenario

Consequences analysis •MAXCRED

Fault tree development

Fault tree analysis •PROFAT Apply safety measures and reevaluate risk

Risk estimation

Whether risk is in acceptance? Yes End

No

Suggest safety measures to control risk

Safety weighted hazard index (SWeHI) *

 It

relates hazards posed by a unit and safety measures effective on it  It represents the radius of the area of 50% probability of fatality/damage

* Khan F.I., Husain, T., and Abbasi, S.A. Transaction of IChemE UK, B79, 1-16, 2001

SWeHI continued… SWeHI = B/A B

is the quantitative measure of the damage potential  A represents the credits due to control measures and safety arrangements

Start Manageable units & take one unit Identify all hazardous chemicals Fire and explosion hazards

Match the unit with the predefined units

Type of hazards presents?

Calculate G factor Calculate penalties

Calculate Fs factor and different penalties

Estimate damage potential

Estimate damage potential using Fs & penalties

Estimate B2 factor

Estimate B1 factor B1

Toxic and Corrosive hazards

B2

Maximum of B1 and B2 as B factor

Credits for the safety measures Quantification of A Quantification of SWeHI All chemicals & units checked? Yes Stop

No

Quantification of B1 (fire & explosion hazards)  Energy

factors, Fs

Chemical Energy  F1 = 0.1*M * (Hc)/K  Physical Energy  F2 = 1.304 * 10-3*PP*V  F3 = 1.0*10-3*1/(T+273)*(PP-VP)2*V 

B1 quantification continues  Penalties        

for various parameters

Temperature, pn1 Pressure, pn2 Location with respect to others, pn3 Capacity of the unit, pn4 Chemicals characteristics, pn5 Degree of congestion, pn6 External factor such as earthquake, pn7 Vulnerability of the site, pn8

Quantification of B2 (toxic hazard)  B2

is quantified using one core ‘G’ factor and seven penalties G= S*m S is dependent on release condition, and  m is release rate or mass released 

B2 quantification continues  Seven

penalties are:

Operating temperature, pnr1  Operating pressure, pnr2  Vapor density, pnr3  Chemical characteristics, pnr4  Population density of the area, pnr5  Site characteristics, pnr6 and pnr7 

Quantification of A A

incorporates the quantification of the various control measures  A is classified in two groups Measures to control the damage potential  Measures to reduce the frequency of occurrence 

Ranking of Hazard SWeHI = Maximum (B1 or B2)/A SWeHI 0

Not hazardous

1

Less hazardous

5

Moderately hazardous

10

Hazardous

20

Highly hazardous

Start

Hazard identification SWeHI Probabilistic hazard assessment-ASM

Quantitative hazard assessment- MCAA Accident scenario development •MCAS

Fault tree for the envisaged scenario

Consequences analysis •MAXCRED

Fault tree development

Fault tree analysis •PROFAT Apply safety measures and reevaluate risk

Risk estimation

Whether risk is in acceptance? Yes End

No

Suggest safety measures to control risk

Maximum credible accident analysis (MCAA)  Accident 

scenario forecasting

Maximum credible accident scenario (MCAS)

 Damage

estimation for envisaged accident scenario 

MAXCRED software

Maximum credible accident scenario *

 The credible accident is defined as ‘the

accident that is within the realm of possibility (i.e., probability higher than 1*e-06 /yr) and has a propensity to cause significant damage (at least one fatality)’.

* Khan F.I., Chemical Engineering Progress (AIChE, USA), November, 55-67, 2001

Take one unit Develop all plausible accident scenarios Consider one accident scenario Flammable

Is the chemical flammable &/or toxic?

Calculate factor A

Toxic and/or corrosive

Calculate factor BB

Calculate factor B Calculate factor CC

Calculate factor C Calculate credibility factor L1

Calculate credibility factor L2 Calculate total credibility factor L Classify credibility of the scenario Is it credible? Yes List the scenario No

Are all units over? Yes

Short list the most credible accident scenarios

No

Delineation of maximum credible accident scenarios  Credibility

of accident scenario is delineated using: L1 (fire and explosion)  L2 (toxic release)  L = (L12 + L22)1/2 for both type of events 

0.0 Uncertainty zone 0.2 Credibility zone 0.5

Maximum credibility zone

1.0

Damage estimationMAXCRED *

 MAXCRED

enables simulation of accidents and estimation of their damage potential

* Khan, F.I., and Abbasi, S.A., Environment Modelling and Software, 14, 11-25, 1999

Models in MAXCRED 

Fire    



Pool fire Flash fire Fire ball Jet fire

Toxic release  





Heavy gases Light gases

Domino effect model

Explosion   

Confined vapor cloud explosion Boiling liquid expanding vapor cloud explosion Vapor cloud explosion

Start

Hazard identification SWeHI Probabilistic hazard assessment-ASM

Quantitative hazard assessment- MCAA Accident scenario development •MCAS

Fault tree for the envisaged scenario

Consequences analysis •MAXCRED

Fault tree development

Fault tree analysis •PROFAT Apply safety measures and reevaluate risk

Risk estimation

Whether risk is in acceptance? Yes End

No

Suggest safety measures to control risk

Analytical simulation method (ASM) *

 Main

steps:

Fault tree development  Boolean matrix creation  Finding of minimum cutsets and optimization  Probability analysis  Improvement index estimation 

* Khan, F.I., and Abbasi, S.A., J of Hazardous Materials, 75(1), 1-27, 2000

Start Represent an undesired event in terms of fault tree Transform fault tree into boolean matrix Solve boolean matrix for minimum cutsets Optimization of cutsets

No

Optimization criteria

Is optimization over? Yes

Probabilistic analysis

Transformation of static probability to fuzzy probability set

Probabilities

Improvement index calculation Stop

ASM Procedure

PROFAT

*

 PROFAT

is the software developed based on ASM  It is coded in C++

* Khan, F.I., and Abbasi, S.A., Process Safety Progress (AIChE, USA), 18(1), 1999

Start

Hazard identification SWeHI Probabilistic hazard assessment-ASM

Quantitative hazard assessment- MCAA Accident scenario development •MCAS

Fault tree for the envisaged scenario

Consequences analysis •MAXCRED

Fault tree development

Fault tree analysis •PROFAT Apply safety measures and reevaluate risk

Risk estimation

Whether risk is in acceptance? Yes End

No

Suggest safety measures to control risk

Risk estimation  Risk

= damage potential * probability of occurrence F

 Risk

representation

F-N Curve  Iso-risk contours 

N

Risk contours over site layout

Start

Hazard identification SWeHI Probabilistic hazard assessment-ASM

Quantitative hazard assessment- MCAA Accident scenario development •MCAS

Fault tree for the envisaged scenario

Consequences analysis •MAXCRED

Fault tree development

Fault tree analysis •PROFAT Apply safety measures and reevaluate risk

Risk estimation

Whether risk is in acceptance? Yes End

No

Suggest safety measures to control risk

Safety measures design  Design

measures to control the damage Fire resistance barrier,  Blast resistance barrier, etc. 

 Design

measures to reduce probability of occurrence Automatic shut down system,  Safety relief valve, etc. 

Re-evaluation of Risk  Modify

the fault tree  Redo the fault tree analysis  Re-estimate the risk  Compare risk against acceptable criteria  Units for which risk could not be brought to acceptable level, develop Disaster management plan  Emergency resource plan 

Application of SCAP Process facility on a fixed Offshore platform

Problem Statement* 

To design the safety measure for process units of a fixed offshore platform



The platform is located in east coast region of Canada (Atlantic Canada), Newfoundland shelf, Canada

* Khan, F.I. et. al., J. Of Hazardous Materials , A94, 2002,1-36

Process facilities on offshore platform

Separator 1

Compressor 1

Flash Drum

Process area

30m Separator 1

Compressor 2

Offshore platform

50m

Drier

Compressor

Drier

Flash drum

Pump

Separator 2

Separator 1

Gas pipeline

150

Oil pipeline

Fire and explosion hazard index

Hazard identification Results 300

250

200

Screening limit

100

50

0

Maximum credible accident scenario 

Condensate separator 



Formation of vapor cloud due to release of flammable gas (wet natural gas) from the unit which on ignition causes “vapor cloud explosion”, unreleased chemical in unit burn as “Pool Fire”

Compressor unit 

Continuous release of flammable gas (wet natural gas) from compressor on ignition cause a “jet fire”

Damage estimation: MAXCRED results for condensate separator Parameters

Values

___________________________________________________________________

Unit: Separator Scenario: VCE followed by pool fire Explosion: VCE Total energy released by explosion Peak overpressure Variation of overpressure in air Shock velocity of air Duration of shock wave

(kJ) (kPa) (kPa/s) (m/s) (ms)

: : : : :

1.23e+07 320.00 345.00 353.00 8.0

Damage Radii (DR) for various degrees of damage due to overpressure DR for 100% complete damage (m) DR for 100% fatality or 50% complete damage (m) DR for 50% fatality or 25% complete damage (m)

: : :

53 74 86

Fire: Pool fire Burning area Burning rate Heat flux

(sq.m) : 265.00 (kg/s) : 10.00 (kJ/sq.m) : 2654.00

Damage Radii (DR) due to thermal load DR DR DR DR

for for for for

100% fatality/damage 50% fatality/damage 100% third degree of burn 50% third degree of burn

(m) (m) (m) (m)

: : : :

34 55 69 78

Damage estimation: MAXCRED results for compressor _____________________________________________________________ Parameters Values _____________________________________________________________ Fire : Jet Fire Flame length Burning rate Radiation heat flux

(m) : 5.45 (kg/s) : 10.0 (kJ/m2) : 1493

Damage Radii (DR) due to thermal load DR for 100% fatality/damage DR for 50% fatality/damage DR for 100% third degree of burn DR for 50% third degree of burn

(m) (m) (m) (m)

: 24 : 35 : 44 : 57

_____________________________________________________________

Vapor cloud explosion followed by fire

OR gate

Release of chemical from other units

Vapor cloud explosion

AND gate Basic event Ignition source

Ignition source

15

16

Chemical release

Formation of vapor cloud

18

17

19

20

21

Leak from pipeline Leak from separator 2

Leak from crude oil line

Leak from valves

Excessive pressure in vessel release of chemical

Leak from vapor line

5

6

7

8

9

10 12

1

2

3

4

13

14

Fault tree for a VCE followed by fire in condensate

Jet fire

Ignition

Jet release

Jet causing other units to fail

OR gate AND gate Basic event 13

14 15

Release from upstream pipeline

Release from downstream pipeline

17

16

Release from compressor

Release from pump section

3 1

2

12

8

4

9

5

11

7 6

10

Fault tree for a jet fire in compressor unit

Results of ASM 

Condensate separator unit 





The occurrence probability of the envisaged accident is 9.474E-04 per year Events 18 and 20 (release from connecting pipe and ignition due to external energy source) has maximum (about 17% each) contribution to the probability of the eventual accident High pressure in upstream pipeline, ignition due to electric spark, release from connective vessel, and ignition due to external fire are other important events that are making significant contribution to this accident

Results of ASM 

Compressor unit 





The occurrence probability of the envisaged accident is 1.364E-02 per year Events, external fire causing unit to fail and release of chemical and ignition due to of external energy sources, have maximum contribution (about 47%) to the probability of the eventual accident Ignition due to electric spark, release from pipeline, and leak from casing and seal of compressor are other important events that are making significant contribution to this accident

FN curve for condensate separator 1.00E+00

Frequency of occurrence (F)

1.00E-01

1.00E-02

1.00E-03

1.00E-04

1.00E-05 1.00

10.00 Fatalities (N)

100.00

Design of safety measures  Separator

and compressor unit

Flame arrestor  Cooling system  Flammable gas detector  Inert gas purging system  Preventive maintenance of pumps, pipelines and compressor  Installation of blast barriers 

Vapor cloud explosion followed by fire

Ignition

Occurrence probability reduced from 9.474 E-04 to 1.555E-08 /yr, individual risk from 1.4E-02 to 2.3E-06

Release of chemical from other units

Formation of vapor cloud

Ignition source Jet causing other units to fail

Ignition

27 28 21

22

23 Ignition source

24

15

16

25

Jet causing other units to fail

26

17 18

Leak from pipeline Leak from separator 2

Leak from crude oil line

Leak from vapor line

5

1

2

Leak from valves

3

6

7

8

9

10

19

Bursting of separator 2 causing release of chemical

Excessive pressure in vessel

4 11

12

13

14

Fault tree for condensate separator after implementing control measures

Jet fire followed with pool fire

Ignition

Jet release

Jet fire causing other units to fail

Ignition source

20 21 15

16

Jet fire causing other units to fail

22

17 19

18 Release from downstream pipeline

Release from upstream pipeline

3 1

2

4

Release from compressor

Release from compressor

13

Release from pumps

Release from pump section

9 11

7 6

14

12

8 5

Occurrence probability reduced from 1.364E-02 to 1.311E-06 /yr, individual risk from 1.24E-01 to 1.21E-05

10

Fault tree for compressor after implementing control measures

FN curve for condensate separator 1.00E+00

1.00E-01

before safety measures

Frequency of occurrence (F)

acceptance criteria 1.00E-02

after safety measures

1.00E-03

1.00E-04

1.00E-05

1.00E-06 1.00

10.00 Fatalities (N)

100.00

Comparison of individual risk with ALARP criteria Units After remedial meaures

1.0E+01 Unacceptable region 1.2E-01

1.0E-01

Before remedial measures

1.2E-01

1.0E-03

5.7E-03 ALARP

3.7E-04

4.2E-05 1.2E-05

1.0E-05 2.3E-06 1.0E-07

1.2E-05 Negligible risk

5.2E-07

5.5E-07

5.2E-07

Broadly acceptable region

Drier

Flash

Compr. 2

Compr. 1

Sepa. 2

1.0E-09 Sepa. 1

Individual Risk (IR)

1.4E-02

Thanks