GP 24-21 - Fire Hazard Analysis - 0900a86680322962

BP GROUPENGINEERING TECHNICAL PRACTICES

Fire Hazard Analysis


Group Practice

Page 2 of 53

BP GROUPENGINEERING TECHNICAL PRACTICES


Foreword

This revision of has been made consistent with GP 24-20 (Fire and Explosion Hazard Management) and reorganized to align with the flowchart and place technical methods in the annex. Because the reorganisation of this GP was so extensive, revisions have not been identified by bars in the margin as in normal practice.

Page 3 of 53

Copyright © 2009 BP International Ltd. All rights reserved.This document and any data or information generated from its use are classified, as a minimum, BP Internal. Distribution is intended for BP authorized recipients only. The information contained in this document is subject to the terms and conditions of the agreement or contract under which this document was supplied to the recipient's organization. None of the information contained in this document shall be disclosed outside the recipient's own organization, unless the terms of such agreement or contract expressly allow, or unless disclosure is required by law.

In the event of a conflict between this document and a relevant law or regulation, the relevant law or regulation shall be followed. If the document creates a higher obligation, it shall be followed as long as this also achieves full compliance with the law or regulation.


Table of Contents

Page

Foreword............................................................................................................................................2

1. Scope........................................................................................................................................6

2. Normative references................................................................................................................6

3. Terms and definitions................................................................................................................6

4. Symbols and abbreviations.......................................................................................................8

5. Fire hazard types, sources, and characteristics........................................................................9

6. FHA process...........................................................................................................................10

7. Identify the fire hazard sources...............................................................................................13

7.1. General........................................................................................................................13

7.2. Define control measure performance...........................................................................14

7.3. Select and analyze representative scenarios..............................................................14

7.4. Assess the impact........................................................................................................17

7.5. Assess the need to reduce hazard severity and consequences..................................18

7.6. Identify options to optimise the process, layout, and control measure performance...18

7.7. Document FHA and hazard management...................................................................20

7.8. Further analysis...........................................................................................................21

Annex A (Normative) Development and use of hazard source........................................................22

A.1. Wells.......................................................................................................................................22

A.2. Pipelines and risers.................................................................................................................22

A.3. Process...................................................................................................................................23

A.3.1. Release rate modelling................................................................................................23

A.3.2. Control measures and limitations on fire size and severity..........................................25

A.4. Fuel and chemical storage and transfer..................................................................................27

Annex B (Normative) Basic methodology for FHA...........................................................................29

B.1. General...................................................................................................................................29

B.2. Gas fire sources and characteristics.......................................................................................29

B.2.1. General........................................................................................................................29

B.2.2. Jet fires........................................................................................................................29

B.2.3. Low momentum gas fires.............................................................................................30

B.2.4. Cloud fires....................................................................................................................31

B.2.5. Release rate calculations.............................................................................................31

B.2.6. Transient release rates following isolation...................................................................34

B.3. Liquid fire sources and characteristics....................................................................................35

B.3.1. General........................................................................................................................35

B.3.2. Pool fires......................................................................................................................35

B.3.3. Spray fires....................................................................................................................35

Page 4 of 53


B.3.4. Two phase fires............................................................................................................36

B.3.5. Release rate calculations.............................................................................................36

B.3.6. Spray fire dimensions..................................................................................................38

B.3.7. Pool fire dimensions.....................................................................................................39

B.3.8. Fireball.........................................................................................................................42

B.3.9. Flash fire......................................................................................................................43

B.3.10. BLEVE.....................................................................................................................43

B.3.11. Blowout...................................................................................................................44

B.3.12. Thermal radiation....................................................................................................45

B.4. Confined fires..........................................................................................................................47

B.4.1. General........................................................................................................................47

B.4.2. Internal effects.............................................................................................................47

B.4.3. External flaming...........................................................................................................48

B.4.4. Smoke generation........................................................................................................48

B.4.5. Calculation methods for confined fires and external flaming.......................................48

B.5. Smoke from fires.....................................................................................................................53

B.5.1. General........................................................................................................................53

B.5.2. Outdoor smoke concentration......................................................................................53

B.5.3. Smoke impairment.......................................................................................................56

B.5.4. Smoke ingress into accommodation, TR and control spaces......................................57

Annex C (Normative) Impact criteria................................................................................................59

C.1. General...................................................................................................................................59

C.2 Personnel exposure................................................................................................................59

C.2.1 Personnel exposure to fire...........................................................................................59

C.2.2. Personnel exposed to radiant heat..............................................................................59

C.2.3 Personnel exposed to smoke and heat.......................................................................59

C.3. Equipment and structural response........................................................................................60

C.3.1. General........................................................................................................................60

C.3.2. Process equipment response......................................................................................61

C.3.3. Structural response......................................................................................................61

Bibliography......................................................................................................................................62

List of Tables

Table B.1 - Gas jet and liquid spray fire hazard ranges...................................................................46

Table B.2 - Obstruction attenuation factors......................................................................................47

Table B.3 - Plume compositions - open or well ventilated fires........................................................54

Table B.4 - Plume compositions - poorly ventilated fires.................................................................54

Table B.5 - Typical burn rates for hydrocarbon fuels.......................................................................56

List of Figures

Figure 1 - FHA process....................................................................................................................11

Page 5 of 53


Figure 2 - Application of fire hazard analysis throughout the project lifecycle.................................12

Figure B.1 - Typical initial release rate for methane at 10°C (50°F).................................................32

Figure B.2 - Methane gas jet fire flame dimensions.........................................................................33

Figure B.3 - Transient release rate of gas as a function of initial inventory.....................................34

Figure B.4 - Crude oil release rate as a function of vessel pressure...............................................37

Figure B.5 - Example liquid releases................................................................................................38

Figure B.6 - Crude oil liquid spray fire flame dimensions.................................................................39

Figure B.7 - Pool fire flame height as a function of pool diameter...................................................41

Figure B.8 - Flame deflection (degrees from vertical) for gasoline pool fires as a function of wind speed......................................................................................................................................42

Figure B.9 - Downwind heat flux from stabilised crude oil fires of various diameters......................46

Figure B.10 - Smoke dilution factors for seven representative fires.................................................55

Page 6 of 53


1. Scope

a. This Group Practice (GP) gives requirements on an approach for fire hazard analysis (FHA) in capital projects and existing operations. This provides the level of understanding needed to make informed decisions regarding the most effective combination of fire hazard management strategies.

b. FHA is applicable to offshore and onshore facilities.

c. FHA is applicable to:

1. Design of a new facility during each of the CVP stages, and

2. Design changes to an existing operating facility in accordance with a management of change process.

d. For an existing facility where the BP entity determines that fire hazards need to be better understood, the general principles of FHA should be implemented.

e. This GP is intended to conform to the requirements for continuous risk reduction defined in GP 48-50.

f. Fires not within the scope of this GP are:

1. Fires resulting from cellulose, electrical, metal fires, etc.

2. Fires inside buildings.

2. Normative references

The following referenced documents may, to the extent specified in subsequent clauses and normative annexes, be required for full compliance with this GP:

• For dated references, only the edition cited applies.

• For undated references, the latest edition (including any amendments) applies.

BP

GP 24-22 Gas Explosion Hazard Analysis.

GP 48-50 Major Accident Risk (MAR) Process.

3. Terms and definitions

For the purposes of this GP, the following terms and definitions apply:

Capital Value Process (CVP)

The CVP contains a number of distinct stages between project appraisal and facility operation. The majority of these stages address project development while the last one, “operate” addresses the majority of the facility lifecycle. The following sections describe the FEHM in the CVP. For an existing operation whose development did not follow CVP, these same FEHM activities will be applicable, but not necessarily in stages and timeline documented in the following sections.

Cause

Event, situation, or condition that results, or could result, directly or indirectly in an accident or incident.

Page 7 of 53

http://sunapps.bpweb.bp.com/cirrus/


CIRRUS

BP suite of consequence modelling tools. Can be found at http://sunapps.bpweb.bp.com/cirrus/

Consequences

Direct, undesirable result of an accident sequence usually involving a fire, explosion, or release of toxic material. Consequence descriptions may be qualitative or quantitative estimates of the effects of the accident in terms of factors such as health impacts, economic loss, and environmental damage.

Escalation

This occurs when the fire from the original source burns long enough to cause hazardous fire conditions (at tanks, vessels, piping containing flammable substances), or weaken structures supporting such systems and equipment. When loss of containment occurs, the released inventory adds fuel to the original fire, enlarging the fire envelope and putting more adjacent equipment at risk of fire.

External flaming

External flaming is the effect that occurs outside a roofed module when the fire size or ventilation conditions are such that a significant proportion of the flames extend beyond the module boundary affecting the areas to either side and above.

Facility

General term for a complete onshore or offshore installation, vessel or rig.

Fire and explosion hazard management (FEHM)

FEHM is the active management of fire and explosion hazards during each of the CVP stages from knowledge of the causes, likelihood, severity, and consequences of major accidents that could arise. It delivers a strategy for each hazard such that a major accident can be preferentially prevented, controlled such that evacuation is not necessary, or in extreme cases, that safe evacuation can be achieved.

Hazard

Condition or practice with the potential to cause harm to people, the environment, property, or BP’s reputation.

Hazard identification (HAZID)

Brainstorming approach used to identify possible hazards. HAZID studies are very broad in their scope. The HAZID is sometimes called a preliminary hazard analysis.

Modification

Changes to existing facilities.

Muster location

Designated place where personnel can muster and survive the effect of a major hazard accident while awaiting evacuation.

Pool spray transition point

The pressure below which an ignited pressurised liquid release will burn as a pool and above which, a significant portion will burn as a spray.

Radiant heat flux

Amount of radiant heat from a heat source, such as a fire, which flows through a unit area of a piece of equipment or structure per unit time

Page 8 of 53


Risk

A measure of loss/harm to people, the environment, compliance status, Group reputation, assets, or business performance in terms of the product of the probability of an event occurring and the magnitude of its impact. Throughout this GP the term “risk” is used to describe health, safety, security, environmental, and operational (HSSE&O) undesired events.

Safety critical design measure (SCDM)

Those safety systems, devices, and controls which are designed to prevent, detect, control, or mitigate a major accident, or facilitate the escape and survival of people. These are also referred to as ‘safety critical elements’ in some regions and ‘protective systems’ in GDP 5.0-0001 and GRP STD 01. Three function categories are:

Prevention

Design measures to prevent a major accident through effective monitoring and process safety systems, with automatic executive action to remedy abnormal plant behaviour where appropriate.

Control

These are the active and passive measures to prevent escalation of a major accident and to bring the plant to a stable condition over a defined time period without operator intervention. Operator intervention should be possible, but not relied upon to carry out essential incident control.

Mitigation

These are measures that provide reliable protection for personnel and critical equipment or structure from the effects of a major accident. It is often through barriers, structural strength and active fire protection.

Temporary refuge (TR)

A location (typically in an enclosure or building) that will enable occupants to survive defined major accidents, including fire or blast events. Other commonly used names are “protected muster location” or “safe muster location”, or less commonly used term “safe refuge”. A temporary refuge could be the living quarters and is generally where the majority of personnel reside.

4. Symbols and abbreviations

For the purpose of this GP, the following symbols and abbreviations apply:

BLEVE Boiling liquid expanding vapour explosion.

BOP Blowout preventer.

CFD Computational fluid dynamics.

CVP Capital value process.

D Diameter.

DHSV Downhole safety valve.

ESD Emergency shutdown.

FCCU Fluid catalytic cracking unit.

FEHM Fire and explosion hazard management.

Page 9 of 53


FHA Fire hazard analysis.

HAZID Hazard identification.

HCU Hydro cracking unit.

HDS Hydrodesulfurization.

HVAC Heating, ventilation, and air conditioning.

IM Integrity management.

LPG Liquefied petroleum gas.

P&ID Piping and instrumentation diagram.

PHAST Process hazard analysis software tool.

RHU Resid hydrotreater unit.

SCBA Self contained breathing apparatus.

SCDM Safety critical design measure.

TR Temporary refuge.

5. Fire hazard types, sources, and characteristics

a. Types of fire that may result from accidental releases of process, pipeline, or well hydrocarbon inventories are defined below.

b. Detailed characteristics and analysis methods are given in the Annexes.

c. This GP does not address the analysis of nonhydrocarbon fires, however the principles are similar.

d. Fires within the scope of this GP are classified as follows:

1. Gas fires - jet fires and low momentum releases.

2. Liquid fires - pool fires and pressurised spray fires.

3. Two phase fires - flashing liquids or gas with entrained liquid droplets.

4. Blowout - drilling, workover, well service or subsea.

5. Flash fire - combustion of flammable gas cloud without developing any explosion overpressures.

6. Fireball - fire from a catastrophic release of a large quantity of pressurised gas or volatile liquid.

7. BLEVE - catastrophic failure of a vessel containing volatile liquids following sustained fire exposure.

8. Boilover - sudden steam generation in atmospheric storage tanks throwing tank contents high in the air.

6. FHA process

a. The objectives of conducting FHA are to:

Page 10 of 53


1. Provide opportunities to reduce the risks from fire hazards.

2. Provide an understanding of the scale, intensity and duration of credible fires.

3. Provide a dynamic picture of the fire hazards as a capital project design evolves.

4. Identify the impact of fires on operating personnel, plant equipment, and facilities.

5. Provide a process and tools to optimise the design to minimise fire severity.

6. Define the performance of SCDMs which control or mitigate the effects of fires.

b. FHA process is described in Figure 1 and each step is described in further detail in the following clauses.

c. For capital projects, fire hazard management shall begin during the early stages in the CVP to be the most effective. Figure 2 describes FHA process through the CVP stages.

d. If the FHA identifies a fire hazard that is larger than that considered in the MAR analysis for the site, the MAR shall be updated to include the FHA fire hazard.

Page 11 of 53


Figure 1 - FHA process

Page 12 of 53

Identify the fire hazards

Quantify the source term inventories and characteristics

Define control measures performance

Select and analyse representative scenarios

Assess the impact on people and critical equipment

Assess the need to reduce the hazard

severity and consequences

Identify options to optimise the process, layout and control measure performance

Document assessment and hazard management

No

Yes


Figure 2 - Application of fire hazard analysis throughout the project lifecycle

Page 13 of 53

APPRAISE

SELECT

DEFINE

EXECUTE

Identify primary risk drivers.Initial assessment of their effects on each conceptInitial assessment of the practicality of controlling them or evacuating safely

For major fire hazards on each option identified using FHA:

Develop an understanding of fire hazard sources on each option using quantification of primary inventories and vulnerabilities

Perform preliminary fire assessments to give initial categorisation input to FHA

Identify key causes and estimate likelihood

Identify key parameters and SCDMs needed to make hazards manageable (e.g., use of FHA to challenge flammable inventories/location and number of isolation devices)

For further major fire hazards identified using FHA:

Further develop hazard sources for process, wells and pipelines for a range of representative scenarios

Analyse hazard effects for worst case controllable and evacuation events

Optimise, process design, control measures and layout to minimise severity and impact

Rerun hazard calculations and confirm of hazard categories.

Assess of causes and likelihood to determine if evacuation category frequency is acceptable

Identify damaging exposures to people and plant

Confirm performance needed from SCDMs

In line with improved detail for major fire hazards identified using FHA

Further develop hazard sources for process, wells and pipelines for a range of representative scenarios

Update hazard analysis to incorporate any design changes

Document categorisation, hazard effects and SDCMs in Hazard Register

Confirm hazard categorisation and perform final risk assessments

Hand over documentation, data to operations

OPERATE Update hazard source, hazard analysis and categorisation to account for process changes and modifications


7. Identify the fire hazard sources

7.1. General

Fire hazard sources shall be modelled as defined below.

a. Dynamic modelling of hazard sources with the input data being progressively updated as greater detail is defined or understood.

b. Fire hazard sources are hydrocarbons with the potential for release and fire being handled in operations including production, delivery, storage, processing, packaging, and shipping.

c. Properties (e.g., volume, mass, pressure, temperature, and location) of each of the identified hazards are accurately quantified so that the type, location, and rate of release over time can be modelled for a range of scenarios.

1. Models should be created which have the capability of simulating the performance of measures to control the severity.

2. Such models should be developed for the wells, process inventories, and pipelines. The development and requirements of these hazard sources are described in Annex A.

d. The potential for harm is determined by the fuel source and the way in which it is released.

1. An accurate model should be developed for each of the primary sources; isolated process inventories, pipelines, and the wells.

2. The hazard source is a definition of the possible hydrocarbon release rates over time associated with each process system, pipeline, or well, taking into account the isolatable sections, the riser capacity (e.g., export pipeline), or the reservoir. This is described in more detail in Annex A.

e. The following time dependent characteristics of fire are quantified to the extent necessary to manage them:

1. Release rate vs. time.

2. Fire type, size, shape, and location.

3. Heat flux.

4. For offshore facilities:

a) External flaming beyond the source module or process area.

b) Smoke location and density throughout the platform.

f. Annex B presents some simple calculations that may be used to analyse fire hazards.

7.2. Define control measure performance

a. Hazard sources include control measures which limit the rate of release and severity of fires.

1. In early stages of the FHA, assumptions about their provision and performance may be made based on default design standards and good engineering practice.

2. Thereafter, these may be optimised based on greater understanding and definition to give an enhanced reduction in severity.

b. The following systems should be included. These are discussed in Annex A together with their performance standards:

1. Inherent breaks in the process plant.

2. ESD valves.

3. Depressurisation.

Page 14 of 53


4. Containment and disposal of liquid spills.

5. Limitation of hole size.

7.3. Select and analyze representative scenarios

7.3.1. General

a. Fire hazards should be analysed by initially analyzing a range of release rate calculations and selecting the bounding cases with release rates which may cause the greatest impact.

b. In analysis of offshore facilities, these cases should be subject to scoping fire hazard calculations to show the effects on the facility at different times during each scenario. This data is then used to consider escalation and exposure to TR and evacuation systems.

c. Fire analysis calculations may be repeated several times in order to optimise inherently safer design and control system performance. In cases where the hazard severity warrants, more sophisticated modelling should be applied.

7.3.1. Release rate analysis

a. Fire hazard source information should be used to predict time dependent hydrocarbon release rates for a series of representative scenarios, as defined below.

b. Release rates at the varying times should be recorded, i.e., 0, 2, 5, 10, 20, 30, 60, and 120 minutes. In the case of low pressure oil releases with a lower ignition probability, the case of an unignited pre-release building up in the bunded (diked) area should be considered. It should be assumed to build up to a depth of 50 mm (2 in) or to continue for 30 minutes before ignition whichever is smaller.

c. Release rate modelling should cover the following range of events:

1. A range of credible hole sizes; typically this would be based on pipe sizes, equipment connections or possible leak routes through equipment seals, etc. The range should be sufficiently wide to model a release from small to full bore in order to obtain knowledge of potential fire effects right up to the worst case.

2. Scenarios taking credit for control measures such as ESD and depressurisation operating to their intended performance standard.

3. Scenarios taking no credit for the control measures such as ESD and depressurisation operating to their intended performance standard.

4. Full bore pipe rupture scenarios shall be modelled for the following units:

a) Crude and vacuum distillation units.

b) Fluid catalytic cracker units.

c) Delayed coker.

d) Hydrotreatment units (including HCU/RHU/HDS).

d. Pipeline events with the following conditions.

1. Full bore release.

2. Small bore failures of cases representative of the relevant types of causes.

e. Well or blowout events.

1. Drilling blowout.

2. Well service blowout.

3. Gas lift annulus backflow through the injection port diameter.

4. Wellhead blowout at the Christmas tree through the swab valve joint.

Page 15 of 53


7.3.2. Selection of release cases for fire analysis

a. The following cases should be identified from the release rate analyses:

1. Liquid release and gas fire cases with the greatest potential impact on health, safety, and environment.

2. For offshore facilities, liquid release and gas fire cases with the greatest potential for escalation.

b. It may be necessary to model escalating scenarios. These can arise from initial explosions or fire.

7.3.3. Fire hazards effects analysis

The selected release cases should be analysed using the methods described in Annex B to give the following data:

a. Location, shape, and size of the fire.

b. Preparation of schematic drawings with the dominant case fire effects overlaid upon them.

c. Heat fluxes within the fire.

d. Radiation contours from fires with effects beyond the source and into other areas of the facility.

e. Smoke effects from large sustained oil fires.

7.4. Assess the impact

7.4.1. General

Detailed methods and criteria for impact assessment for people and plant are given in Annex B and Annex C.

7.4.2. Personnel

Personnel present in or adjacent to the source, those evacuating, and those in adjacent buildings should be considered.

7.4.3. Structures and equipment

a. Analysis of the effects of flame impingement on structures and equipment should be carried out to determine whether any critical structure could fail or deform and lead to escalation through further loss of hydrocarbon containment or impairment of emergency response provisions.

b. FHA should define the extent and duration of the flame envelope for leaks from significant hydrocarbon inventories.

c. In offshore analysis, fire exposure, the size and strength of the plant, and the consequences should be considered. Those items with a significant risk of failure and a consequence which could require evacuation or impair the TR and evacuation systems should be subjected to more accurate response analysis if effective protection has not been specified.

d. A list of equipment and structures that, on failure, could lead to escalation of fire hazard should be developed by FHA. Typical items to be considered include:

1. Significant process inventories.

2. Risers and pipelines.

3. Wells.

4. Hydrocarbon storage and associated piping.

Page 16 of 53


5. SCDMs including:

a) Valve actuators.

b) Flare lines and equipment.

c) Active fire protection systems, including piping.

d) Firepumps and emergency generation.

6. Critical structures including:

a) Primary structures including:- Those supporting process modules, drilling modules, or the TR.- Offshore jacket members.- Elevated process decks- Refining reactors and towers with large inventories.

b) Secondary structures supporting major inventories including:- Tall structures including the derrick and flare.- Vessel supports and skirts.

7. Vessels with the potential for catastrophic failure (BLEVE).

7.5. Assess the need to reduce hazard severity and consequences

a. The need to reduce hazard severity and consequences may be dependent on the type of or reason for FHA.

b. The need to reduce consequences as prompted by a MAR analysis.

c. The need to reduce the hazard severity may simply be based on the principal of continuous risk reduction and knowledge gained through FHA.

d. A risk matrix may be used to prioritize fire scenarios for further risk reduction.

7.6. Identify options to optimise the process, layout, and control measure performance

a. In a capital project, there should be progressive efforts to reduce the severity of the events through process design and optimisation of control measures. As project and thus hazard source details are defined, consequences should be analyzed, and control measures optimized.

b. In both existing facilities and those in capital project design, the knowledge gained during FHA should provide the focus for risk reduction through optimizing the process, layout, and control measures. Approaches may include the following:

1. Minimisation of the inventory that can be released. This includes minimising the exposure of large inventories to explosion overpressures.

2. Optimisation of the layout and ESD location such that the number of release points is limited and that their location is restricted to a single area where the effects are controllable.

c. Minimisation of release pressure.

d. Minimisation of release rate.

e. Minimisation of the burn rate of pool fires.

f. Fire containment to origin.

g. Minimisation of the impacts including those to occupied buildings and, on offshore facilities, exposure of the TR and evacuation systems to flames, heat, and smoke.

Page 17 of 53


1. TR should be safely accessible by personnel not in immediate vicinity of event, provide protection from event, and provide life support systems for designated period.

2. Direct access to primary evacuation facilities (e.g., lifeboats) should be available.

7.7. Document FHA and hazard management

a. The final version of FHA conducted in capital projects shall be provided to operations so that any future changes can be assessed.

b. FHA documentation shall include the following information:

1. References to applicable hazard identification studies.

2. Reference to input data used. For example:

a) P&IDs.

b) Plot plan and layout drawings showing location of equipment, escape routes, and embarkation areas.

c) Framing plans or other drawings showing critical structural members.

d) List of SCDMs and the assumptions made regarding their performance specification.

e) For each fire hazard:- Unique identifier.- Type of fire hazard.- Details of safety systems that are assumed to operate.- Calculations showing magnitude and duration of hazard zones for each leak size.- Review of impact of fire hazards on personnel and equipment.- Where escalation was considered, the list of paths to escalation and timing of that escalation.- List of assumptions used in analysis.

7.8. Further analysis

a. In some cases, more detailed analysis may be required to understand and manage fire hazards. However, the need for more detailed studies should be carefully managed.

b. Detailed methods should only be used if early scoping studies indicate cause for concern (i.e., hazards contribute significantly to overall risk, and deeper understanding is critical to managing them effectively and specifying systems).

c. Detailed studies may include more detailed analysis of the following:

1. Hazard source – if necessary using a computerised transient process model such as HYSYS or pipeline release tools.

2. CFD models to predict flame shape and smoke impact on the whole platform.

3. Thermal response modelling of vessels and structures.

Page 18 of 53


Annex A(Normative)

Development and use of hazard source

A.1. Wells

a. Potential release scenarios are identified in the wells HAZID. Release rates should be quantified initially by the maximum well production rate estimates or by the release rate through a restricted orifice as defined by the scenario.

b. Hazard source definition

1. The best hazard source definition is the reservoir and well flow model itself.

2. The hazard source definition should be used both to calculate the stable flow rate and the timing of the loss of control.

3. This is an important input to assessing the practicality of completing a precautionary evacuation before hydrocarbons reach the surface.

4. The hazard source definition should consider the diameters of the well bore, tubulars, and drill pipe when calculating the flowrates for each scenario.

A.2. Pipelines and risers

a. The pipeline release model requires accurate input data of length, diameter, elevation at either end, profile along the seabed, and contents.

b. Contents may range from dry clean gas and dry stable oil, to spiked crude, wet gas with gas liquids, and well fluids. They may be one, two, or three phase.

c. Calculations should be performed for a full bore release and realistic hole sizes in accordance with the potential failure modes. These should be modelled at the top of the riser, at the sea surface where ship impact may occur, and on the sea bed immediately under the platform.

d. Single phase are the most easily modelled using initially simple calculations, such as those in the SINTEF/Scandpower Handbook for Fire Calculations and Fire Risk Assessment in the Process Industry. Stable oil cases should take into account the compressibility of the oil and the stretch of the pipeline.

e. Two and three phase pipelines are more difficult to model. Simple computer models, such as CIRRUS, may be used for scoping calculations. The dynamics within the pipeline affect the following critical factors:

1. Evolution or vaporisation of gas from liquids and the maintenance of pressure.

2. Condensation of liquids in gas or miscible gas pipelines.

3. Pressure time profile along the length of the pipeline.

4. Stratification of gas and liquids and subsequent slug or mist flow.

f. The primary strategy for risers should be prevention of the release, but analysis should be performed for small hole release that is likely to be of long duration and lead to structural failure or escalation to other risers.

g. A full bore rupture would be an evacuation category, and the transient flowrates may require more accurate modelling using transient, two phase flow methods.

Page 19 of 53


A.3. Process

A.3.1. Release rate modelling

An accurate hazard source definition should be developed for the whole process plant, including the flare system as follows:

a. Divide the process into isolatable inventories separated by actuated isolation valves.

b. For each inventory:

1. Identify each element: pipe section, vessel, pump, compressor, heat exchanger, etc.

2. Identify for each element

a) Dimensions.

b) Locations.

c) Elevations.

d) Liquid levels.

3. Identify for each element:

a) Pressure.

b) Density.

c) Molecular weight.

d) Water contents.

e) For liquids, the flash gas fraction.

4. Identify and name ESD and depressurisation valves and determine their performance - see A.3.2.2.

5. Identify any inherent barriers that limit the maximum gas or oil release, weirs, or phase breaks at which gas separates two halves of a liquid inventory or visa versa.

6. Identify pool fire area in each area in which there are inventory elements.

c. Develop methods to calculate the maximum inventory that can be released from each element.

d. Develop methods to calculate jointly the release rate and pressure versus time with the following variables or cases:

1. Specific hole size within the chosen hole size range.

2. Release point by element (optional).

3. With and without selected ESD operation.

4. With and without depressurisation.

5. Different depressurisation rates.

6. Time delay on operation of ESD and depressurisation valves.

7. Leak rates through ESD valves (optional).

8. Different heat inputs from the fire.

9. Different relief valve settings.

10. Different values of the pool spray transition point.

11. Specific pool burn rate for the liquids in each of the isolated inventories.

Page 20 of 53


12. Release or burn rate at which significant external flaming occurs (from module calculations given under confined fires in Annex B).

13. Minimum damaging fire size in terms of release rate.

14. Minimum time to failure for the surrounding equipment.

e. Calculation methods should give the following information:

1. Matrix showing which inventories can be released in which areas.

2. Temperature versus time profile for each scenario.

3. Release rate versus time profile for each scenario.

4. Duration of the longest damaging fire for each scenario and for the inventory as a whole.

5. Size of the maximum damaging fire for each scenario and for the inventory as a whole.

6. Time at which a liquid inventory passes through the pool spray transition point.

7. Highlight where the liquid release rate exceeds the pool area burn rate once the pressure has dropped beneath the pool spray transition pressure.

Release rate calculations are given in Annex B. These can be used manually in the early stages of a project, but it is advisable to build a spreadsheet model to accommodate a greater level of detail.

A.3.2. Control measures and limitations on fire size and severity

A.3.2.1. Inherent breaks in the process plant

a. Certain features within the plant act as inherent barriers to segregate inventories and reduce the potential mass that can be released.

b. These include displacement pumps, reciprocating compressors and the weirs within separators.

A.3.2.2. ESD valves

a. ESD valves limit the inventory available for release.

b. The following functional performance standards should be defined:

1. Location: (i.e., inventories that are isolated from each other and location of the valve itself.

2. Logic: automatic, automatic with time delay, and abort or manual.

3. Speed of response, including both detection and closure time.

4. Tolerable leak rate.

c. The reaction time (e.g., for operators to activate unit ESD) should be assumed to be at least equal to:

1. 20 minutes for first detection if no specific gas or fire detection system is available at the scenario location;

2. 10 minutes if a comprehensive fixed detection (fire or gas depending on scenario) system is available at scenario location.

d. If there are additional delays to activate some mitigation measures (e.g., need to manually isolate a distant valve, slow reaction devices), these times should be added to the reaction time before credit can be taken for that measure.

Page 21 of 53


A.3.2.3 Depressurisation

a. Depressurisation can reduce:

1. Size and duration of a gas fire.

2. Size, duration, and intensity of a liquid fire by reducing the pressure and flashing off the volatile components.

3. Time at which a liquid release passes through the spray/pool transition point, ideally such that it is less than that for major escalation from primary oil inventories.

b. The following functional performance standard should be defined:

1. Inventories that should be depressurised.

2. Logic: automatic, automatic with time delay, and abort or manual.

3. Speed of response, including both the detection and closure time.

4. The depressurisation rate.

A.3.2.4. Containment and disposal of liquid spills

a. The burn rate of a pool fire is proportional to its surface area. This can by limited by the use of bunds, dykes, gullies, and drains, which can be used to reduce the overall impact of an oil fire on the platform such that external flaming is reduced, and the event comes into the controllable category rather than an evacuation event.

b. The following functional performance standard should be defined:

1. Pool area.

2. Location.

3. Drainage capacity and flowrate.

4. Location of any discharged flammable material.

A.3.2.5. Limitation of hole size

a. It may be necessary to identify the maximum hole size that corresponds to a controllable event and the hole size for evacuation events.

b. Holes greater than these sizes may result in fire hazard events that cannot be managed easily.

c. In these cases, special arrangements in design and future integrity management should be specified to reduce any failures greater than these sizes.

d. These should be identified via the hazard register with the following information:

1. Potential failure points or causes of larger bore failures.

2. Causes of the failures and prevention systems.

3. Requirements for enhanced integrity monitoring.

4. Operational limitations and precautions.

A.3.2.6. Deluge systems

a. Deluge can limit fire severity, but its assessment is excluded from the FHA. Deluge effects should be judged separately.

b. Deluge is described in more detail in GP 24-22, GP 24-23, and GIS 24-231.

c. Deluge can reduce:

Page 22 of 53


1. Burn rate of pool fires in which the liquid release pressure has passed through the pool spray transition point.

2. Intensity of a fire filling a roofed module by vaporisation, heat absorption, and steam generation. It may completely suppress external flaming, but this would be substituted with steam mixed with carbon particles.

A.3.2.7. Water content

a. The presence of water in the fluids reduces the release rate for a given hole size. In large quantities, it may inhibit ignition and combustion.

b. Experiments have shown that it is difficult to burn stable oil with more than a 65% water cut. This figure may be higher if there is a high gas content.

c. It is reasonable to assume that combustion does not occur with water cuts greater than 80%.

d. A stagnant plant or one that has been shut down for some time may allow oil and water to separate. This may be the case with a sell bore and should be considered prior to intervention.

A.4. Fuel and chemical storage and transfer

a. These inventories are not normally considered to cause initial events unless their storage and handling conditions makes them prone to a large release and ignition.

b. They should be considered as an escalating event if they are exposed to damaging fires and explosion and cannot be easily protected.

c. Releases are normally at atmospheric pressure, and judgement would be needed to determine the hole sizes and the effectiveness of isolation form the primary tanks.

d. The fuel burns as a pool, unless there are particular circumstances to raise the pressures. The fire size is a function of the release rate or the bund (dike) area if present.

Page 23 of 53


Annex B(Normative)

Basic methodology for FHA

B.1. General

This annex provides more detailed information on fire hazards and performing calculations.

B.2. Gas fire sources and characteristics

B.2.1. General

a. The primary gas sources are gas production wells, gas lift systems and annuli, separation from oil processing, drying, and gas liquids separation. The inventories may be augmented by flash gas from liquids.

b. The mass of the isolated inventories is limited by the volume and density. They would normally by insufficient to overwhelm a platform and require evacuation, unless there is an isolation failure from a pipeline, well, or well annulus.

c. Release rates are an order of magnitude smaller than liquid releases for the same hole sizes at moderate pressures (up to 30 bar [435 psi]).

d. Larger gas process inventories may have the potential for localised damage and escalation.

e. The action of depressurisation can further reduce that potential for escalation and widespread harm.

B.2.2. Jet fires

a. Gas fires are sonic momentum driven gas releases and are highly directional.

b. Obstructions and convection have little effect upon the flame shape.

c. Heat fluxes range from 200 kW/m2 to 300 kW/m2 (63 000 Btu/hr-ft2 to 95 000 Btu/hr-ft2) with the majority being convective.

d. Gases with higher molecular weights, such as propane, have a higher radiant heat component.

e. Gas jet fires result from pressurised releases.

f. In open, noncongested areas, fire is characterised by well defined narrow jet as gas and air are mixed and burned.

g. If release is from irregular shaped hole or flange leak or it impinges on adjacent plant, fire plume is wider and more diffuse than straight jet, although flames still have considerable momentum.

h. Burning characteristics are similar to those of jets, except that there is no single dominant jet direction.

i. If jet is totally obstructed (e.g., impingement onto wall), release loses its momentum and burns as diffuse fireball rather than well defined jet.

j. Because air-fuel mixing in jet/diffuse fireball is more efficient than in pool fire, jet fires burn with cleaner, less sooty flames.

k. As with pool fires, amount of soot increases for heavier hydrocarbons.

l. Methane jets burn with very clear bluish flame, and most of the heat is transported away as convective heat.

Page 24 of 53


m. For propane and butane, flame is more luminous (yellowish) due to combustion of soot (i.e., soot is produced and consumed within flame).

n. Gas releases above 2 bara (29 psia) are at sonic velocity. They have high convective heat transfer if flames impinge onto equipment.

o. Industry guidance suggests that, for natural gas releases of less than 10 kg/s (1 300 lb/min), total incident heat flux to engulfed objects can be from 50 kW/m2 to 300 kW/m2 (from 16 000 Btu/hr/ft2 to 95 000 Btu/hr/ft2). For larger flames or gases involving higher molecular weight components, total heat flux to engulfed objects may be higher.

p. Lift off

1. Close to source, shear stresses created as fuel expands from orifice are high, great, and flame cannot stabilise.

2. Therefore, there is a region close to release point where no flame exists. This is called lift off. The jet may become unstable and lift off completely.

3. In such cases, jet self extinguishes and may result in severe explosion hazard.

4. Jet is likely to be stabilised if release is from rough, irregular shaped hole or if jet impinges onto other process plant or structures.

5. It is unlikely that jet becomes unstable except in very open areas (e.g., a well bay).

q. Accumulated flammable atmosphere

1. If gas release does not ignite immediately, there is accumulation of flammable atmosphere.

2. If ignited, this flashes back from point of ignition to point of release, giving a cloud fire.

3. Depending on release orientation and density of gas released, resulting cloud fire may be either flash fire above grounded gas cloud or vertically rising fireball.

4. If area is congested or enclosed, explosion may also be caused.

5. Both situations (3. and 4.) represent significant risk to lives of people in immediate vicinity.

6. Heavier gases (propane, butane, etc.) can be more hazardous, as they may spread across ground, covering wider area and engulfing more people and more potential ignition sources. They are also more reactive gases that give higher flame velocities and are more likely to cause explosions.

7. If release continues after ignition, cloud fire is followed by sustained jet fire.

B.2.3. Low momentum gas fires

a. Low momentum gas fires only occur if the pressure is below that which could gave a sonic release; typically lower than 2 bar (29 psi).

b. As a result, the release rates are low and may be below those that could realistically cause escalation. One low momentum case that may require consideration is a subsea release from a pipeline or well which would give a fire on the sea surface under a platform.

B.2.4. Cloud fires

a. Given rapid transient nature of cloud fires, there is no realistic means of protecting people engulfed, so it is critical that both exposure and chance of ignited release should be minimised.

b. Control of activities and restriction of unnecessary visits to areas of concern should be minimised.

Page 25 of 53


c. In open, noncongested areas, significant overpressures are not generated, and structural consequences of cloud fires should be minimal.

d. In confined and congested areas, significant explosion overpressures are generated (i.e., explosions).

e. This GP does not cover assessment of gas cloud size or explosions. Areas that are partially enclosed by walls or ceilings or areas with significant congestion caused by process plant may cause both accumulations of gas and overpressures. These should be subjected to separate assessment.

B.2.5. Release rate calculations

a. Gas only

1. The initial gas release rate, Q(0), in kg/s is given by:

RT

MPCAREAQ d

1

2)0(

1

1

Where:

Cd = discharge coefficient (= 0,85).

= ratio of specific heats.

R = gas constant = 8 314.

P = pressure (N/m2).

M = relative molecular mass (g/mol).

T = temperature (K).

2. Typical values for are:

Methane 1,308.

Ethane 1,193.

Propane 1,133.

Butane 1,094 to 1,097.

Pentane 1,074 to 1,076.

3. Values for M (Molecular weight) are:

Methane 16.

Ethane 30.

Propane 44.

Butane 58.

Pentane 72.

4. An Example is shown in Figure B.1.

Page 26 of 53


Figure B.1 - Typical initial release rate for methane at 10°C (50°F)

0

5

10

15

20

25

30

0 10 20 30 40 50 60 70 80 90 100Pressure (barg)

Rel

ease

rat

e (k

g/s)

10 mm hole20 mm hole50 mm hole

b. Gas supported by vapourising or flashing liquid

1. For cases in which there is gas cap over liquid containing flash gas (e.g., pressurised LPG), initial inventory of gas, W, is taken as sum of mass of gas cap and liquid that will flash off.

2. For cases in which there is gas cap over liquid containing dissolved gas (e.g., live crude oil), initial inventory of gas, W, is taken as sum of mass of gas cap and amount of gas dissolved in liquid.

c. Flame dimensions

1. For quantification purposes, jet fires are often approximated by a cone. Base of cone will be “lifted off” from release point, and cone can be deflected by ambient wind.

2. For unconfined jet fires, the following equations can be used to give approximate size of flame. Fireball size is pertinent to cases in which jet is deflected by local obstructions to extent that fire burns as fireball rather than well defined jet.

3. Flame length (m) is calculated by:

0,46( ) 10( )Flame Length m Q

Where:

Q = release rate (kg/s).

4. Flame volume (m3) is calculated by:

1,35) . ( )3Flame vol (m const Q

Constant values are as follows:

Methane 100.

Page 27 of 53


Propane 90.

5. Fireball diameter (m) is calculated by:

3 12volFireball diameter (m) =

6. Flame volume and fireball diameter are appropriate for case in which jet flame impacts onto object and is deflected into diffuse fireball.

7. Calculation of flame volume can also be used to assess shape and dimensions of fire that is partially confined by roof or walls.

8. An example of fireball diameter for a methane gas jet fire is shown in Figure B.2.

Figure B.2 - Methane gas jet fire flame dimensions

0

10

20

30

40

50

60

70

80

90

0 10 20 30 40 50 60 70 80 90 100

Release Rate (kg/s)

Len

gth

(m

)

Flame Length for Unobstructed Fire

Fireball Diameter for Obstructed Fire

B.2.6. Transient release rates following isolation

a. Pressure in gas volume with a leak starts to decline once isolation valves close and blowdown valves (if present) have opened.

b. Reduction in pressure due to mass loss from fixed mass of gas is critical in modelling benefit provided by these safety systems.

c. If plant can depressurise or blowdown to flare, outflow from isolated inventory would be combined mass flow rate both of accidental release and discharge to flare.

d. For simplicity, it is best to assume that emergency isolation blowdown occurs at the same time. Initial release rate will be as calculated above, i.e., Q (0). The initial blowdown rate would be Q (bd). This should be available from the plant process engineers.

e. For case in which blowdown is initiated after a time delay of t (bd) seconds, subsequent transient release rate through the leak hole is given by:

W

bdttbdQQmExpQtQ

))(()()0()0()(

Page 28 of 53


f. Depressurisation reduces severity of fire, both by accelerating reduction in fire size and duration.

g. An example is shown in Figure B.3.

Figure B.3 - Transient release rate of gas as a function of initial inventory

TRANSIENT RELEASE RATE OF GAS AS FUNCTION OF INITIAL INVENTORYINITIAL RELEASE RATE = 20 lb/s

0

2

4

6

8

10

12

14

16

18

20

22

0 200 400 600 800 1000 1200 1400 1600

TIME (Seconds)

RE

LE

AS

E R

AT

E (

lb/s

)

INITIAL INVENTORY OF 1000 lb

INITIAL INVENTORY OF 5,000 lb

IINITIAL INVENTORY OF 20,000 lb

B.3. Liquid fire sources and characteristics

B.3.1. General

a. Primary liquid sources are the liquids production stream on oil platform and condensate collection and export on gas platforms.

b. Density and volume of some liquid inventories (i.e.,pipelines, separators, pumps, slug catchers) may be sufficient to have a major prolonged impact on a platform, and this may be sufficient to require evacuation or have catastrophic effects if not effectively controlled.

c. High pressures lead to high release rates from moderate hole sizes and to spray fires that cannot easily be controlled by deluge systems.

d. Depressurisation plays a key role in managing the fires and defines if and when spray fires become pool fires.

B.3.2. Pool fires

a. A pool fire is the combustion of a hydrocarbon liquid from a surface with vaporisation caused by the heat feedback from the fire.

b. It may be restricted to a specific area (i.e., limited in area by bunds, dykes, or gulleys or it may be free to spread, in which case it may be a running fire.

c. Heat flux from pool fires depends upon size, fuel, and confinement, if any.

d. Pool fires have the poorest combustion and, together with confinement, give the highest smoke concentrations at the source.

Page 29 of 53


e. Pool fires can be effectively suppressed by deluge to reduce the burn rate and smoke production and can bring pool fires into the controllable category with the right design.

B.3.3. Spray fires

a. A spray fire is a liquid release in which the pressure and any dissolved gas cause atomisation such that a significant proportion burns as a spray. The point at which a liquid release burns as a spray is called the spray transition point. It depends upon the following:

1. Release pressure.

2. Shape of the orifice and degree of resultant spray formation.

3. Proportion of dissolved gas.

4. Volatility or proportion of light ends.

b. 7 bar (101 psi) should be used as the transition point for stabilised crude oil.

c. 4 bar (58 psi) should be used as the transition point for stabilised condensate.

d. Gas liquids (i.e., liquefied ethane, propane, and butane) should always be considered as spray fires unless they are refrigerated and at atmospheric pressure.

e. Heat fluxes from spray fires can range from 200 kW/m2 to 350 kW/m2 (from 63 000 Btu/hr-ft2 to 111 000 Btu/hr-ft2), with high levels of radiative heat transfer for fluids with gas liquids, dissolved gas or light ends.

f. Smoke density at the source reduces as combustion improves, again a function of high pressures, gas content, and volatility.

B.3.4. Two phase fires

a. The sources of two phase fires are:

1. Inventories.

2. Highly gaseous wells operating below their bubble point.

3. Gas lifted wells.

4. Gas pipelines in which there may be some liquid dropout.

5. Gas liquids inventories in which some gas can vaporise in the piping before being released.

b. Their characteristics are in between those of gas fires and high pressure volatile spray fires.

B.3.5. Release rate calculations

a. Initial release rate

1. Release rate of liquid, m (kg/s), is given by:

2d

Pm C Area

Where:

Cd = Discharge coefficient (=0,62).

Area = hole area (m2).

= Density of liquid released (kg/m3).

P = Differential pressure inside vessel (N/m2).

Page 30 of 53


2. Typically, density of crude is ~800 kg/m3, and condensate can be taken as ~600 kg/m3.

3. P is gage pressure of vessel.

4. An example is shown in Figure B.4.

Figure B.4 - Crude oil release rate as a function of vessel pressure

0

50

100

150

200

250

300

350

400

0 200 400 600 800 1000 1200 1400

VESSEL PRESSURE (Psia)

RE

LE

ASE

RA

TE

(lb/

s)

1/2 inch Hole

1 inch Hole

2 inch Hole

b. Reduction in release rate

1. Once process isolation occurs, liquid inventory in vessel or piping starts to reduce. Release continues until inventory is exhausted.

2. Releases downstream of pump see rapid reduction in release rate if pump is shutdown. At this time, pressure of release falls to pump suction pressure, and in many cases, pump acts as additional restriction to flow.

3. If there is gas pad over liquid, release rate continues at initial release rate until inventory is exhausted. The time pressure profile may be provided by process calculations of by using the equations in B.2.6 using just the Q(bd) term.

4. If gas pad is blown down, release rate falls as gas pad pressure reduces.

5. Example shown in Figure B.5 shows release rates for crude oil inventory of approximately 18 t (20 T) at pressure of approximately 30 barg (440 psig) with isolation and blowdown on gas pad (10 min for blowdown to 6,8 barg (100 psig)).

Page 31 of 53


Figure B.5 - Example liquid releases

0.0

10.0

20.0

30.0

40.0

50.0

60.0

70.0

80.0

90.0

100.0

0 600 1200 1800 2400 3000

Time (seconds)

Rel

ease

Rat

e (k

g/s

)Inventory exhausted for 2 inch leak

Liquid spray fire transition to Pool fire

Inventory exhausted for 1 inch leak

ESD and Blowdown start here

B.3.6. Spray fire dimensions

a. For quantification purposes, spray fires are often approximated by a cone.

b. Base of cone is “lifted off” from release point, and cone can be deflected by ambient wind.

c. If there is limited spraying or no immediate vaporisation, cone is narrow but terminates in large rising plume. Plume can have shape and characteristics of pool fire of same liquid.

d. Unobstructed fires

1. The following equation should be used to give approximate flame length (m):

0,4617( )Flame Length (m) Q

Where:

Q = release rate (kg/s)

2. An example of flame length for crude oil liquid spray fires is shown in Figure A.6.

e. Flame volume

1. The flame volume is calculated using simple formula derived from the volume of pool fires.

2. The results may be conservative.

Flame vol (m3) = const × m1,35

Constant values are as follows:

Crude oil 170.

Condensate 110.

f. Obstructed fires

Page 32 of 53


1. If jet flame impacts onto object and is deflected into diffuse fireball, diameter of this fireball are given by:

3 12volFireball diameter (ft) =

Where:

vol = the flame volume

2. Flame volume and fireball diameter are appropriate for cases in which jet flame impacts onto object and is deflected into diffuse fireball.

3. Calculation of flame volume can also be used to assess shape and dimensions of fire that is partially confined by roof or walls.

4. If the flame volume exceeds a significant proportion of the module volume, the residual portion extends beyond to form external flaming as calculated in B.4.5.

5. An example of fireball diameter for obstructed crude oil liquid spray fires is shown in Figure B.6.

Figure B.6 - Crude oil liquid spray fire flame dimensions

0

20

40

60

80

100

120

140

160

0 10 20 30 40 50 60 70 80 90 100

Release Rate (kg/s)

Len

gth

(m

) Flame Length for Unobstructed Fire

Fireball Diameter for Obstructed Fire

B.3.7. Pool fire dimensions

B.3.7.1. Size of pool

a. Pool size is governed either by release and burn rates or by area of the liquids containment.

b. If the release rate is greater than the burn rate on a confined area, there is a progressive buildup of the oil level that continues to burn after the release has finished, unless there are means to allow the oil to drain to a safe place.

c. The burn rates of common offshore liquids are:

1. Oil - 0,050 kg/sec/m2

2. Condensate - 0,059 kg/sec/m2

Page 33 of 53


d. The area needed to completely burn a release would be:

Free pool Area A = Q x (burn rate)-1

e. The burn rate within a containment area is:

Q (restricted pool area) = A(containment area) x ( burn rate)

f. Diameter (D) to which pool spreads depends on whether spill is continuous or instantaneous.

g. For liquid spills, area of fire is no larger than area of bund (dyke).

h. Exceptions to a. and b. are if:

1. Volume of spill is larger than volume of bund (dike).

2. Spill has high momentum and washes over bund (dike). This is a concern for sudden large spill in bunds (dikes) with angled sides.

3. Leak jets spray over bund (dike) wall.

i. For continuous spill in unbunded (undiked) region, equilibrium burning pool diameter is given by:

2

1

2

bmequilibriu m

VD

Where:

V = Volumetric spill rate (m3/s)

= Liquid fuel density (kg/m3)

mb = specific burning rate (kg/m2/s)

j. For instantaneous burning spill, in unbunded (undiked) area, average pool diameter, Davg, is given by:

max683,0 DDavg

Where:

8

1

2

3

max

b

L

m

gVD

B.3.7.2. Flame volumes

a. The flame volume calculations in B.3.6 may be applied to pool fires.

b. The pool burn rate should be used as the basis of the calculation.

B.3.7.3. Flame height

a. Calculation of the flame height is only required for large open top deck fires or for sea fires.

b. In roofed modules, moderate fires are distorted and spread by the ceiling.

Page 34 of 53


c. Height of flame (H, in m) is given by:

0,61

42 bmH D

gD

Where:

D = pool diameter (m)

mb = specific burning rate (kg/m2/s)

ρ = atmospheric density = 1,29 kg/m3

g = gravitational acceleration = 9,81 m/s2

d. An example is shown in Figure B.7.

Figure B.7 - Pool fire flame height as a function of pool diameter

POOL FIRE FLAME HEIGHT (ft) AS FUNCTION OF POOL DIAMETER (ft)

0

20

40

60

80

100

120

140

160

0 20 40 60 80 100 120

POOL DIAMETER (ft)

FL

AM

E H

EIG

HT

(ft

)

CONDENSATE

CRUDE OIL

B.3.7.4. Flame tilt

a. Flame tilt is only of relevance offshore for top deck fires, sea fires, and those cases where there is external flaming.

b. The following flame tilt figures may be applied to the external flames, but the strength and direction of the wind is affected by the bluff area of the platform.

1. Wind causes flame to tilt.

2. Angle of deflection, θ (degrees), from vertical is given by:

cos 1 1

U*

Where:

Page 35 of 53


13b

vap

wind

gm D

UU*

ρvap = fuel vapour density at boiling point with:

b

Wvap T

M18,12

MW = Molecular weight of fuel (g/mol)

TB = Boiling temperature (Kelvin)

c. An example is shown in Figure B.8.

Figure B.8 - Flame deflection (degrees from vertical) for gasoline pool fires as a function of wind speed

0

10

20

30

40

50

60

70

80

90

0 10 20 30 40 50 60 70

WIND SPEED (ft/s)

FL

AM

E D

EF

EL

CT

ION

FR

OM

VE

RT

ICA

L (d

egre

es)

POOL DIAMETER = 15 ft



B.3.8. Fireball

a. Fireballs are caused by catastrophic failure of a large pressurised vessel or a pipeline. This may be an initial event, but it is more likely to result from escalation due to a smaller fire or explosion.

b. The instantaneous release of fuel is such that it cannot burn completely until it mixes with sufficient air.

c. The duration of the burn and size of the fireball depends on the size of the release.

d. Unburnt fuel lifts with the combustion products causing the whole fireball to lift and expand.

e. The primary hazard is the radiation effects on personnel, as they should not have sufficient duration to cause escalation. The exception would be a riser failure that would be followed by a sustained release.

f. If the release is in the top deck, it is free to rise above the platform with the effects limited to those on that top deck.

Page 36 of 53


g. If the release is within a roofed module, the unburnt fuel is forced out through of the open sides, spreading radially before lifting with the convection.

h. In the case of pipelines, particularly gas risers, the effect of catastrophic failure depends upon the riser size and pressure. It could have sufficient initial size to totally engulf the whole of the topsides.

i. Unlike a vessel rupture, there is a continued flow but at progressively reducing rates over the first minute or so. Both the module and pipeline releases could put all personnel in the open at risk.

B.3.9. Flash fire

a. Flash fires arise if gas or vapour fails to disperse, either through confinement or still air conditions. Flame velocities are below those that would cause significant explosion overpressures, but the effects can serious burns, which may be fatal.

b. Detailed examination of flash fire effects would not normally be required, as they would not be of sufficient damage potential to cause escalation or evacuation.

c. Flash fire effects constitute a portion of the overall risk of fatality and be included in the overall platform risk assessment.

d. Scientific judgement should be sufficient to judge the level of risk, given the sources, ventilation, and occupancy. They may also be covered in the explosion assessment, as these represent that high proportion of cases for which overpressures fail to be generated due to smaller gas clouds or ignition conditions.

e. Flash fire occurs if cloud of flammable material burns.

f. Duration of flash fire is typically short, and equipment damage is not a concern.

g. Typically, burn zone is estimated by dispersion modelling and is taken as boundary of flammable limit of cloud.

h. Flammable cloud dimensions should be calculated using CIRRUS or equivalent method.

B.3.10. BLEVE

a. This is the catastrophic rupture of partially filled process vessels in a fire and is most common with relatively thin walled LPG storage vessels.

b. The effect is caused by the heating of the dry top of the vessel until the steel strength drops below the stress caused by the vapour pressure of the contents. This causes the steel to fail and tear, opening the shell and potentially allowing parts of the vessel to become missiles.

c. The superheated liquid contents vaporise instantaneously and are ejected, causing the fireball effects described above.

d. If the BLEVE occurs within a roofed module, the effects are difficult to predict but would cause severe damage and external combustion of the released fuel.

B.3.11. Blowout

a. Blowouts can arise from drilling, workover, or well intervention via the christmas tree with wireline, slick line, or coiled tubing. The causes should be assessed by the drilling and completions departments. They should also give advice on frequency in conjunction with historical data.

b. Blowouts have the potential to become evacuation events, but this is time dependent. In assessing the event, the timeline of the loss of control should be taken into account to determine whether the primary evacuation would take place before hydrocarbons are released. This is an important consideration in the acceptability of classifying the event as an evacuation case.

Page 37 of 53


c. Blowout characteristics are a function of the fluids, the well conditions of temperature and pressure, and the routes and sequence of events by which it can occur. These should be carefully examined to understand the series of failures, timing, location of the release, flowrate, and distribution and numbers of personnel when the hydrocarbons come to the surface in large quantities.

d. Some of the potential events and locations are presented in e. through j. One factor that should be considered is the potential effect of delayed ignition. In the case of a gas or gas condensate blowout, particularly one where the release may spread through the platform, the risk of explosions should be considered. See GP 24-22.

e. In the case of an oil blowout, it could spread over the platform and cascade down through several levels before igniting; with the subsequent flames and smoke engulfing a significant part of the facility, impairing many of the safety systems and TR.

f. Drilling blowout

1. This may occur before or during production.

2. The location of release is the bell nipple leading initially to the collapse of the derrick, which should be designed to fall away from TR and well control systems.

3. There may be progressive escalation with loss of the drill rig support beams and collapse onto other wells.

4. Flowrates are a function of well status, the drilling equipment, well dimensions, and reservoir conditions.

g. Workover blowout. The effects would generally be similar to a drilling blowout.

h. Well intervention blowout, wirelining, or coiled tubing

1. There are several potential locations.

2. Failure of the swab valve joint in the wellbay is potentially the most serious, as the release location is lower, allowing the fire effects to impact on other wells, the supporting structure for the drill rig, and its support beams.

3. It is also likely to develop rapidly, spread to a wide range of ignition sources, and affect systems needed to recover from the incident.

4. Releases from the lubricators will be at a higher elevation, such as a BOP deck, and have a lower flowrate, but could still affect the drill rig or its support structure.

i. Wellhead blowout

1. This is a blowout from a completed well with the christmas tree in place.

2. Other than the swab valve failure listed above, there are few circumstances in which this occurs, other than a flowline washout and multiple failures of DHSV, upper master, and wing valves.

3. Common mode failure of these should be examined(e.g., by damage to the well control panel or hydraulic lines by an explosion).

4. The effects would be the same as for the well intervention blowout following a swab valve failure.

j. Subsea blowout

1. A subsea blowout brings hydrocarbon to the surface close to or under the platform.

2. It may be caused by a release around the outside of the casings or from a failure at a seabed BOP.

Page 38 of 53


3. If there is a significant gas content, it may be ingested into nonhazardous parts of the platform, causing explosions. Gas releases may impair the stability and buoyancy of floating facilities.

4. Any fire may affect the structure, pipelines, and floatation of the facility.

B.3.12. Thermal radiation

B.3.12.1. Pool fires

a. Calculation of radiant heat levels outside pool fire flame envelope should be performed using CIRRUS or other equivalent computer code.

b. Figure B.9 shows radiant heat against downwind distance for stabilised crude oil pool fires on land of various dimensions. CIRRUS was used to develop this figure.

c. If applicable, Figure B.9 may be used in place of CIRRUS or equivalent model.

d. These figures may also be used to approximate the radiation form external flaming

Figure B.9 - Downwind heat flux from stabilised crude oil fires of various diameters

Pool Fire

0

10

20

30

40

50

60

70

80

0 10 20 30 40 50 60 70 80 90 100

Downwind from Edge (m)

Heat Flux (kW/m2)

50m

40m

30m

20m

10m

Downwind

5 (m/s)

B.3.12.2. Jet fire and liquid spray fires

a. Maximum extent of radiant heat fluxes outside flame envelope for gas jet or liquid spray fires can be calculated using simple multipliers on flame length.

b. Multipliers are shown in Table B.1.

c. Factors do not account for impact of objects blocking radiant heat.

d. Multiplying factor for each heat contour is applied to flame length.

Page 39 of 53


Table B.1 - Gas jet and liquid spray fire hazard ranges

Heat flux Multiplying factor (on flame length)

kW/m2 Btu/hr/ft2

37,5 12 000 1,2

12,5 4 000 1,45

4,7 1 500 1,75

B.3.12.3. Reduction in radiant heat due to obstructions

a. Solid objects, such as walls, should be considered impervious to thermal radiation.

b. If there is partial barrier, such as process equipment or decks on open sided structures, the reduction factors in Table B.2 may be used for heat flux 10 m (30 ft) or more beyond edge of obstructed zone.

c. Obstruction factors should be applied to unobstructed flame length even if obstructed flame diameter is used.

Table B.2 - Obstruction attenuation factors

Unobstructed heat flux Obstructed heat flux

(kW/m2) Btu/hr/ft2 (kW/m2) Btu/hr/ft2

37,5 12 000 12,5 4 000

12,5 4 000 4,7 1 500

4,7 1 500 2 600

B.4. Confined fires

B.4.1. General

a. Confined fires are those in which the shape of a module governs the flame shape and potentially the combustion by limiting or distorting flow of the air supply. Typically, it applies to roofed normally ventilated modules in offshore platforms.

b. A fire is only deemed to be confined if either its size or the limitation to its ventilation is such that it caused sustained external flaming. In general, process gas inventories do not have sufficient inventory to cause sustained external flaming and analysis of the phenomenon should be restricted to the major oil inventories.

B.4.2. Internal effects

a. A confined fire causes flames to spread across the ceiling and descend to engulf all high level plant. In large fire cases, it may fill 60% of the free volume in the module, affecting the ceiling structure and columns and high level piping, including the flare lines, deluge piping, and cable trays.

b. For liquid fires, pool and spray fires give similar characteristics, as it is the lower velocity convection driven flames that fill the module. The tops of large separators, slug catchers, and vertical gas liquids knock out vessels would be engulfed. These are normally dry internally and could heat up moderately quickly.

c. The average ceiling temperatures are of the order of 1 000°C to 1 100°C (1 832°F to 2 012°F) with heat fluxes of the order of 100 kW/m2 to 200 kW/m2 (32 000 Btu/hr-ft2 to 63 000 Btu/hr-ft2).

Page 40 of 53


d. The local effects from spray fires would give the heat fluxes above and the point temperatures and heat fluxes immediately above a pool or spray fire source could reach 1 250°C (2 282°F) and 300 kW/m2 (95 000 Btu/hr-ft2).

B.4.3. External flaming

a. External flaming appears in two forms:

1. The first is simply the excess flame volume that cannot be accommodated within the module.

2. The second is the combustion of unburnt vapours because insufficient air can be drawn into the module so the vapours burn when they mix with their air outside.

b. The impact of external flaming on a platform can be such that evacuation is clearly necessary if it continues for more than a few minutes. It could also cause escalation of plant in other modules, multiple impairments of walkways and escape routes, and the simultaneous operation of deluge systems in adjacent areas.

c. As oil is the predominant source, smoke volumes and concentrations have the potential to impair every part of the platform.

B.4.4. Smoke generation

a. The smoke densities and volumes are a function of the fuel, burn rate, and efficiency of the combustion. In the case of confined fires, all these factors are present such that the smoke effect will be maximised.

b. The presence of a roof and the flame shape diminishes the convective effect seen in open fires that would lift the smoke above the platform. As a result, it should be assessed with the assumption that if the base level around the rest of the platform is the same as the neutral plane in the module.

B.4.5. Calculation methods for confined fires and external flaming

B.4.5.1. General

a. In general, confining fire has the following three effects:

1. Distorting flame shape and concentrating its effects upon adjacent plant and structure.

2. Restricting ventilation supply of fresh air to fire.

3. Causing external flaming where the size of ventilation conditions cause it to do so.

b. How these influence characteristics of fire is dependent on a number of factors, such as:

1. Fuel.

2. Compartment size.

3. Fire size.

4. Ventilation.

5. Type of burning.

6. Fire location and, in case of jet fires, direction.

7. Insulation in compartment.

8. Mass of steel, process plant, etc., in compartment.

9. Obstructions and resulting turbulence.

10. Depth of hot gas layer.

Page 41 of 53


c. Knowledge of confined fires is incomplete. In general, expert advice should be sought on how to assess confined fires.

B.4.5.2. Development, distortion, and confinement of combustion products

The progression and development of the fire in a confined area is as follows:

a. For pool fire, vertical jet, or the low momentum tail of a spray fire, the rising plume impinges on ceiling and spreads radially.

b. Depending upon the size of the fire and available ventilation, flames cover the ceiling and begin to build up underneath it to form the “hot gas layer”. Hot, less dense, combustion products begin to accumulate at top of compartment. This is a transient process while air in the module is used up and fire becomes completely dependent upon external ventilation.

c. The time for stable combustion to be established is a function of fire size compared with air volume. It is considered to be stable if two thirds of the air has been used up. This may be calculated as follows:

tsc = 0.045 x Vmod x (Qbr) -1

tsc = time to stable combustion in seconds

Vmod = free volume in the module in m3; i.e., the height of the openings x the floor area.This excludes the dead space volume amongst the ceiling beams.

Qbr = burn rate in kg/sec. This will be the release rate for spray fires or the pool area burnrate for pool fires.

d. Downward spread of accumulating gases is counteracted by buoyancy forces of hot gases and well defined layer of hot gases (including flame) develops and grows. This is known as “hot gas layer”.

e. Hot gas layer

1. Below hot gas layer is comparatively cool layer of almost clear air.

2. If ceiling is flat with no obstructions and clean openings extending up to full height and there is good ventilation, hot gas layer may be relatively thin, and it spreads easily across ceiling and out through openings.

3. If ceiling is congested with intersecting beams, cable trays, and piping causing dead areas and turbulence, layer is much thicker.

4. Hot gas layer descends below top of openings until excess heat and smoke can escape.

5. With poor ventilation or with large fires, the hot gas layer descends even further to a level more than half way down from the top of the openings in the side of the module. This is the neutral plane and its elevation from the floor is calculated as:

Hnp = 0.33 x Ho

Hnp = height of the neutral plane from the floor in metres

Ho = height of the openings from the floor in metres (generally measured to thebottom of the primary ceiling beams

Page 42 of 53


6. If there are openings in compartment, as hot gas layer builds up, unburnt gases and flames begin to escape through upper part of opening(s). These are examined in A.14.1.3.

7. Fresh air entrained through lower part(s).

8. In steady state, mass flow of hot gases leaving compartment equals sum of mass flows of fuel entering compartment and fresh air being drawn in.

9. Once inside compartment, fresh air is entrained into fire plume and, under certain conditions, into burning interface of hot gas layer.

10. The shape and heat of this type of fire causes simultaneous exposure of ceiling beams, flare and deluge systems, high level piping, pipe supports, and the tops of vessels that may be nonwetted tops of liquids vessels. This should be examined holistically when performing the escalation assessment.

B.4.5.3. Restrictions to ventilation

a. Amount of air that can be drawn into compartment to support combustion depends on both size and configuration of ventilation openings in compartment.

b. Poorly ventilated fire creates more smoke and carbon monoxide.

c. Unburned fuel may also burn outside compartment, endangering adjacent areas.

d. Fire in compartment may also cause other fuels to burn (cables, paint, plastic equipment, etc.) These add to the smoke.

e. Global equivalence ratio

1. Global equivalence ratio (φ) is used to describe entrainment of air into compartment fire.

2. This is defined as air to fuel ratio of fire in compartment divided by equivalent air to fuel ratio required for ideal, stoichiometric burning;

rm

m

f

a

Where:

ma = mass flow (kg/s) of air drawn into module.

mf = mass flow (kg/s) of fuel entering fire.

r = air to fuel mass ratio required for ideal, stoichiometric burning.

3. In all but the simplest cases, it is not straightforward to determine ma and expert advice should be sought. Simple calculations for modules with relatively open sides and with a solid floor are as follows:

ma = 0,5 Ao(Ho)0,5.

Ao = total area of the openings into the module on all sides in m2.

Ho = height of the openings in the sides of the module in metres – see above.

This formula should not be used for complex of highly blocked openings or for modules with grated floors.

4. For φ greater than 1, fire is said to be fuel controlled (fuel lean); more than enough air to support complete combustion can be drawn into module, and burning is limited by

Page 43 of 53


amount of fuel. Although, in theory, there is enough air within module to fully burn fuel, external flaming outside module may occur if module is not large enough to contain flame.

5. For φ less than 1, fire is ventilation controlled (or fuel rich). Insufficient air can be drawn into module such that not all of fuel released can burn within module. In ventilation controlled fires, unburned fuel leaves compartment and may ignite once it encounters fresh air outside compartment. In these cases, ventilation controlled external flaming occurs.

6. For φ = 1, just sufficient air can be drawn in to module that would, if contents of module were perfectly mixed, allow complete, stoichiometric combustion.

7. It is emphasized that, while equivalence ratio is a useful parameter to describe burning conditions, it is a hypothetical parameter based on the assumption that contents of module (air and fuel vapour) are always perfectly mixed. In reality, some parts of module may be fuel rich, others fuel lean, and others close to stoichiometric.

B.4.5.4. External flaming

a. External burning can occur in both ventilation and fuel controlled fires.

b. Fuel controlled fires

1. In fuel controlled fires, external flaming only occurs as a result of roof and/or walls of module deflecting flame.

2. Although there is sufficient oxygen within module to support complete burning, external flaming occurs because flame cannot “fit” into module.

3. The following scoping calculations can be used to gage the severity of the effects.

Vext = Vf – 0,5 Vmod.

Vext = volume of the external flaming.

Vf = flame volume calculated using the equation in B.2.5.

Vmod = free module volume as calculated above in B.4.5.2.c.

4. The shape of the flames is a function of the conditions in the source module, the flame volume, the shape of the outside of the platform, and the wind conditions. A very simple assumption is that the external flaming assumes the shape of a half cylinder rising up the side of the platform with the height equal to twice the diameter.

5. The base of the cylinder is at the level of the neutral plane. This simply gives an indication of the degree of exposure to the wider platform.

6. The cylinder may also tilt to the angles given in B.3.7.4 and Figure B.8. This is critical if the flames could tilt towards elevated accommodation modules. The flame size may also be used to estimate the radiation effects.

7. Using these flame shapes, flame heights and diameters may be estimated as follows:

H ext = 2,17 (Vext)0,33.

D ext

= 1,08 (Vext)0,33.

H ext = height of external flaming above the neutral plane (m).

Page 44 of 53


D ext = diameter of external flaming with base equally spaced either side of opening.

c. Ventilation controlled fires

1. For ventilation controlled fires, there is insufficient oxygen within module to allow complete combustion within module.

2. Unburned or partially burnt gases therefore leave module.

3. If temperature and composition requirements are met, out flowing gases may ignite and result in external flame.

4. Assuming external flow is into open environment, characteristics of external flame are generally similar to those of open, unconfined pool fire.

5. Ventilation controlled fires may reduce the burn rate of pool fires and cause spray pool transition to occur at higher temperatures.

d. There may be more soot and carbon monoxide, particularly if fire is ventilation controlled.

e. Characteristics of external flame depend on fuel involved. In particular, amount of thermal radiation is highly sensitive to amount of soot produced and whether soot burns or provides effective radiation shield.

f. Higher hydrocarbons produce more soot, and flame may be shielded.

B.5. Smoke from fires

B.5.1. General

a. Modelling of smoke in this GP is for initial scoping calculations.

b. Calculations presented do not take account of rise of smoke or detailed interaction of smoke around large obstacles (i.e., formation of a lee due to eddies around buildings, walls, and other solid obstructions).

c. Hence, the benefits of many good layout practices are not fully realised if applying this methodology.

d. More detailed analysis requires application of CFD codes, such as Chameleon from SINTEF.

B.5.2. Outdoor smoke concentration

a. In this methodology, fuel is given initial concentration of contaminants based on type of fuel and degree of ventilation.

b. Initial contaminant density is diluted as it travels downwind.

c. Downwind dilution is modelled using generic dilution curves.

d. Initial concentration of contaminates

4. Three categories of fuel should be considered:

a) Light (L): Predominantly methane but with no liquids.

b) Medium (M): Predominantly propane or butane but with no liquids.

c) Heavy (H): Predominantly pentane or heavier or any gas with liquids.

5. Two categories of ventilation should be considered:

a) Open (O): If four or more sides (of six) have no obstruction (e.g., deck and deckhead but no walls or large obstructions). Open fires are well ventilated, and burn rate is limited by available fuel.

Page 45 of 53


b) Enclosed (E): If three or more sides are obstructed. Enclosed fires are poorly ventilated, and burn rate is limited by available air.

e. Initial contaminant concentrations for well ventilated fires are given in Table B.3 and for poorly ventilated fires are given in Table B.4.

Table B.3 - Plume compositions - open or well ventilated fires

Component concentration

Component Unit Light fuel Medium fuel Heavy fuel

Carbon monoxide ppm 15 400 800

Carbon dioxide % 10 10,9 11,8

Oxygen % 0 0 0

Gas temperature °C (°F) 1 000 (1 800) 1 000 (1 800) 1 000 (1 800)

Optical density dB/m 1,5 15 47

Table B.4 - Plume compositions - poorly ventilated fires

Component concentration

Component Unit Light fuel Medium fuel Heavy fuel

Carbon monoxide ppm 28 500 30 000 31 000

Carbon dioxide % 7,7 8,2 9,2

Oxygen % 0 0 0

Gas temperature °C (°F) 800 (1 500) 600 (1 100) 600 (1 100)

Optical Density dB/m 5 29 70

f. Downwind dilution

1. As smoke travels downwind, it is diluted by air entrainment and expansion of smoke plume.

2. Dilution is modelled using dilution factors for dispersion of smoke for representative fires.

3. Dilution curves as function of downwind distance are shown in Figure B.10.

Page 46 of 53


Figure B.10 - Smoke dilution factors for seven representative fires

g. Applicable curve from Figure B.2 is selected by using representative fire with nearest burn rate to fire under consideration.

h. Size of fire is determined by burn rate of fuel in kg/s (lb/s).

i. Concentration of smoke components external to target point (Cext) are calculated from dilution factor (Df) and near fire concentration (Cs) as follows:

20,9(1 )

(1 )

ext f s

ext f s

ext f s f

ext f s f (ambient)

ext f s

Carbon Monoxide (ppm) C D C

Carbon Dioxide (% ) C D C

Oxygen (%) C D C D

Temperature (C) C D C D Temp

Obsuration (dB/m) C D C

j. For gas jet fires and liquid spray fires, burn rate is simply release rate. Rate is conservative. As release pressure decreases, percentage of fuel burnt decreases. Unburned fuel drops out and burn as pool fire.

k. For pool fire, burn rate is function of material being burnt and surface area of pool fire.

Page 47 of 53


l. Typical values of burn rate per unit area for fuel types on platform are given in Table B.5.

Table B.5 - Typical burn rates for hydrocarbon fuels

Depth burned per hr Surface burn rate

Fuel (mm) (kg/m2s)

Crude oil 180 0,05

Gasoline 180 0,05

Condensate 180 0,05

Kerosene 220 0,06

Diesel 250 0,07

Fuel oil 140 0,04

Hexane 290 0,08

Butane 290 0,08

LNG 330 0,09

LPG 400 0,11

m. Burn rates should be used to match fire being considered to one of representative fires.

B.5.3. Smoke impairment

a. Impact of smoke is primarily a concern for FHA on offshore installations.

b. FHA should consider potential for smoke to block escape routes and should also consider impact of smoke in areas where personnel need to reside for period of time (e.g., muster and embarkation areas).

c. Narcosis

1. Impairment occurs if dose received from various fire gases cause blood carbon monoxide (carboxyhaemoglobin) level to exceed given value.

2. Principal gases concerned are: carbon monoxide, carbon dioxide, and oxygen (deficiency).

3. Dose at each 1 min time increment should be calculated using the following:

2( )in ico co ioF F V F

Where:

Fico = (0,000 829 × Cco1,036) / 15.

Cco = Concentration of carbon monoxide at end of time increment.

Vco2 = (EXP (0,2476 × Cco2 + 1,908 6)) / 6,8.

Cco2 = Concentration of carbon dioxide at end of time increment.

Fio = 1/(EXP (8,13 - (0,54 × (20,9 - Co2)))).

CO2 = Concentration of oxygen at end of time increment.

4. Impairment occurs if level of carboxyhaemoglobin in blood exceeds 15%. Using the equation in 3., this occurs if accumulated dose exceeds 1.

d. Heat

1. Heat impairment occurs if bodys core temperature exceeds a set temperature.

2. There is heat balance relationship between air temperature and ability of a person to lose heat.

3. Levels of activity and clothing have an effect on this calculation.

Page 48 of 53


4. Realistic values are used in the equation under 5.

5. FHA should calculate accumulated heat dose Fih for 1 min time increments using the following:

1/ 5,1849 0,0273 )ihF EXP ( T

Where:

T = temperature (C) at the end of the time increment

6. Impairment occurs if accumulated dose, Fih exceeds 1.

e. Obscuration

1. Particulates in hydrocarbon fire plumes do not cause serious irritation problems, but they may lead to reduction in visibility sufficient to impair escape or induce “irrational behaviour”.

2. FHA should use impairment criterion of visibility of 10 m (33 ft) or path obscuration of 1 dB/m.

3. Time at which level of obscuration exceeds impairment limit (tio) is determined directly from time/visibility calculations.

B.5.4. Smoke ingress into accommodation, TR and control spaces

a. The calculations in c. predict how conditions within a building change with time.

b. This is done for major components of plume that affect visibility, heat, narcosis, and irritancy, namely: level of obscuration, air temperature, and concentrations of various gases, e.g., carbon monoxide, carbon dioxide, and oxygen.

c. Levels of each “contaminant” in building (Cint) are calculated for time (t) using the following:

1

2

3

( )

( )

( )

n int ext n

e int ext e

s int ext s

If t t then C C A EXP R t



Where:

A1 = Cext - C0.

A2 = A1 × EXP (tn × (Re - Rn)).

A3 = A2 × EXP (te × (Rs - Re)).

t = elapsed time since smoke arrived at building.

tn = elapsed time at which HVAC air movement is effectively stopped.

te = elapsed time until external doors are sealed.

ts = elapsed time until maximum required endurance.

Cint = concentration, temperature or level of obscuration at time t.

Cext = concentration (or temperature or obscuration) outside building.

C0 = Initial concentration (or temperature or obscuration) inside the building.

Page 49 of 53


Annex C(Normative)

Impact criteria

C.1. General

a. Impact criteria should be applied to both personnel exposure and exposure of equipment and structures.

b. For a MAR study, the impact criteria in GP 48-50 should be used.

C.2 Personnel exposure

C.2.1 Personnel exposure to fire

The following criteria should be used for personnel exposure to fires:

a. May be caught in initial flash fire or radiant heat flux before they can escape.

b. Survivors of initial event may become trapped and unable to reach place of safety.

c. Effect of heat and smoke may overwhelm temporary places of safety, such as a control room or offshore TR or muster location.

d. Fire may escalate and lead to collapse of structures supporting personnel (e.g., escape routes and TR).

e. Fire may escalate (e.g., BLEVE, cause process column collapse, fail process column supports, cause boilover) engulfing emergency response and fire teams.

f. Fire may escalate, affecting spectators and members of public in off site areas.

C.2.2. Personnel exposed to radiant heat

The following criteria should be used for personnel exposure to radiant heat:

a. Personnel exposed to radiant heat levels in excess of 37,5 kW/m2 (12 000 Btu/hr/ft2) or who are inside flame boundary for any period of time should be assumed to become fatalities.

b. Personnel inside flash fire flame envelope (typically boundary of lower flammable limit) or inside any flame envelope should be assumed to become fatalities.

c. Escape routes should be considered blocked at between 6,3 kW/m2 (2 000 Btu/hr/ft2) and 12,5 kW/m2 (4 000 Btu/hr/ft2).

d. Areas where personnel can be exposed for over one minute while performing essential duties are assumed impaired at 4,73 kW/m2 (1 500 Btu/hr/ft2). This is a maximum heat flux for localization of fire actuation panels (deluge, etc.), manual emergency isolation, or embarkation areas.

C.2.3 Personnel exposed to smoke and heat

The following criteria should be used for personnel exposure to smoke:

a. FHA should consider the effects of immediate and sustained smoke exposure for offshore personnel who may be mustering in TR, trapped, or have to remain in the vicinity of fire. Effect should include narcosis, heat, and obscuration.

b. FHA should use narcosis impairment criterion of 15% carboxyhaemoglobin in blood.

Page 50 of 53


c. FHA should use heat impairment criterion of body core temperature rising to 40°C (104°F).

d. FHA should use obscuration impairment criterion for blockage of escape routes of visibility less than 10 m (33 ft) or path obscuration of more than 1 dB/m.

C.3. Equipment and structural response

C.3.1. General

a. Equipment and structures fail if exposed to fire for long enough to reach their failure temperature. The following can occur during fire leading to escalation:

1. Weakening of flanges and instruments.

2. Weakening of pressure vessels leading to rupture. See B.3.10.

3. Weakening of steelwork supporting escape routes, process plant, and piping.

4. Weakening of primary structural steelwork leading to progressive collapse of facility.

5. Weakening of secondary steelwork leading to collapse of tall structures, such as vertical process vessels, drilling derricks, and flare towers.

6. Weakening of pressure vessels and pipework.

7. Damage to safety systems needed to contain fire: fire pumps, ring mains, ESD and isolation systems, flare systems, and deluge system.

8. Damage to critical systems needed for buoyancy and stability of a floating facility.

9. Damage to risers, supports, and tensioning systems.

b. Failure temperature for heated plant depends on:

1. Inherent strength of plant, vessel, or structure.

2. Loading or stress in equipment or piping.

c. Heat input and subsequent rate of temperature rise of a piece of plant depends on the following:

1. Degree of engulfment (total, partial, or distance from flame).

2. Heat load from flame.

3. Duration of exposure.

4. Thickness, geometry, and specific heat of components.

5. Rating of passive protection.

6. Presence of insulation.

7. Presence of insulation.

C.3.2. Process equipment response

a. In many cases, detailed heat transfer assessment is not needed to estimate damage to process equipment resulting from fire. Approximate assessment of potential for failure should be sufficient.

b. For first pass screening of failure times, the following guidance is appropriate:

1. Lightweight equipment could fail in 5 min if exposed to fires with high heat fluxes. High heat fluxes are in range 200 kW/m2 to 300 kW/m2 (63 000 Btu/hr/ft2 to 95,000 Btu/hr/ft2) (e.g., gas jet fires, liquid spray fires, large condensate, LPG, or compartment fires).

Page 51 of 53


2. Heavyweight large equipment and thick walled piping could fail in 10 min if exposed to fires with lower heat fluxes. Lower heat fluxes are less than 200 kW/m2 (less than 63,000 Btu/hr/ft2) (e.g., smaller fires, flame extension from compartment, or edges of oil fire.

C.3.3. Structural response

a. In many cases, detailed heat transfer assessment is not needed to determine time to failure. Approximate assessment of potential for failure should be sufficient.

b. For first pass screening of failure times, the following guidance is appropriate:

1. Lightweight structures could fail in 5 min if exposed to fires with high heat fluxes. High heat fluxes are in range 200 kW/m2 to 300 kW/m2 (63 000 Btu/hr/ft2 to 95 000 Btu/hr/ft2) (e.g., gas jet fires, liquid spray fires, large condensate, LPG, or compartment fires.

2. Heavyweight large structures could fail in 10 min if exposed to fires with lower heat fluxes. Lower heat fluxes are less than 200 kW/m2 (less than 63 000 Btu/hr/ft2) (e.g., smaller fires, flame extension from compartment, or edges of oil fire).

3. Failure of any equipment or structure if located within 100 kW/m2 (31 000 Btu/hr/ft2) envelope for 30 minutes.

4. Structure exposed to lower level of heat flux over a long period of time is prone to collapse. A heat flux of 25 kW/m2 (8 000 Btu/hr/ft2) for exposure of 30 minutes or more is liable to significant weakening could lead to collapse.

Page 52 of 53

http://northsea.bpweb.bp.com/ask/pi/pi0393.pdf

http://northsea.bpweb.bp.com/ask/pi/pi0393.pdf

http://etplib.bpweb.bp.com/login/IntegratedLogin.jsp?docNumber=GN%2044-330&docType=gn




Bibliography

BP

[1] GDP 3.1-0001, Assessment, prioritization, and management of risk.

[2] GIS 24-231, Water Deluge System Design.

[3] GN 24 - 200, Safety Critical Design Measures .

[4] GN 44 - 330, Siting of Occupied Buildings in Relation to Buried High Pressure Natural Gas Lines on Onshore Facilities.

[5] GP 24-10, Fire Protection - Onshore.

[6] GP 24-20, Fire and Explosion Hazard Management (FEHM) of Offshore Facilities.

[7] GP 24-22, Gas Explosion Hazard Analysis.

[8] GP 24-23, Active Fire Protection - Offshore.

[9] GP 24-24, Offshore Passive Fire Protection.

[10] GP 30-80, Safety Instrumented Systems (SIS) - Specification and Implementation.

[11] GP 30-85, Fire and Gas Detection.

[12] GP 44-15, Offshore Platform Layout.

[13] GP 48-04, Inherently Safer Design (ISD).

[14] GRP 3.1-0001, Selection of Hazard Evaluation & Risk Assessment Techniques.

[15] GRP STD 01, Integrity Management.

[16] SPR/G/97/006, “Method for Assessing Smoke and Gas Ingress to Accommodation and TR Volumes” Issue 01, April 1997. (available in the Design Safety website library)

American Petroleum Institute (API)

[17] API 510, Pressure Vessel Inspection Code: In-Service Inspection, Rating, Repair, and Alteration.

American Institute of Chemical Engineers (AIChE)

[18] Guidelines for Consequence Analysis of Chemical Releases, 1999, Center for Chemical Process Safety (CCPS).

[19] Guidelines for Chemical Process Quantitative Risk Assessment, 2nd Edition, 2000, Center for Chemical Process Safety (CCPS).

[20] Guidelines for Fire Protection in Chemical, Petrochemical and Hydrocarbon Processing Facilities, 2003, Center for Chemical Process Safety (CCPS).

Centre for Marine Technology (CMPT)

[21] Handbook of Fire Protection Engineering, 3rd Edition, 2002.

Page 53 of 54