This action might not be possible to undo. Are you sure you want to continue?
ISSUE 1 October 2003
Whilst every effort has been made to ensure the accuracy of the information contained in this publication, neither UKOOA, nor any of its members will assume liability for any use made thereof. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without prior written permission of the publishers. Crown copyright material is reproduced with the permission of the Controller of Her Majesty’s Stationery Office. Copyright © 2002 UK Offshore Operators Association Limited
UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions
PUBLISHED BY UK OFFSHORE OPERATORS ASSOCIATION London Office: 2nd Floor, 232-242 Vauxhall Bridge Road, London, SW1V 1AU. Tel: 020 7802 2400 Fax: 020 7802 2401 Aberdeen Office: 9, Albyn Terrace, Aberdeen, AB10 1YP Tel: 01224 626652 Fax: 01224 626503 Email: firstname.lastname@example.org Website: www.oilandgas.org.uk
Issue 1, October 2003
UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions
This document has been prepared under a joint industry project sponsored by UKOOA and the UK HSE. The project was managed by fireandblast.com limited, and the production of the initial text and of the back up documentation was undertaken by a consortium headed by MSL Engineering Ltd. The other members of the consortium were Aker Kværner, Century Dynamics, Genesis Oil and Gas, IC Consultants, Morgan Safety Solutions and WS Atkins inc. This document is part of a series being produced by UKOOA and HSE on fires and explosions, the full series being: Part 0 Part 1 Part 2 Part 3 Hazard management (formerly FEHM) Avoidance and mitigation of explosions Avoidance and mitigation of fires Detailed design and assessment guidance
This Part 1 document is taken from MSL Engineering Reports C26800R006 Rev 2 and C26800R007 Rev 2.
Part 0:- Fire and explosion hazard management
Describes Hazard Management principles and practices with particular emphasis on the management of fire and explosion hazards
Part 1:- Avoidance and mitigation of explosions
Describes design considerations for the prevention, control and mitigation of explosions
Part 2:- Avoidance and mitigation of fires
Describe design considerations for the prevention, control and mitigation of fires
Part 3:- Design practices for fire and explosion engineering
Contains advice on the engineering implementation of the measures outlined in principle in Parts 1 & 2
Basis Documents for Parts 1, 2 & 3
Contains base position papers as guidance was developed. Available on www.fireandblast.com for those wishing to understand the logic and data gathered for the positions taken in the guidance
The treatment described in this part of the guidance draws on the experience gained during the period since the Interim Guidance Notes  were prepared. This has allowed simplifications to be made and a more clearly defined approach to be adopted in some circumstances, without compromising safety.
Issue 1, October 2003
UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions
This Guidance does not have the force of a Standard and contains information on good practice which may or may not be on a firm scientific basis and may require clear justification. Where this is the case the uncertainties are highlighted and the limitations of the methods are identified. There is a recognized need to provide such guidance to avoid decisions being made out of context during the explosion assessment process. The term ‘assessment’ is taken to include the assessment of a design in progress and the assessment of an existing installation. This part of the Guidance identifies methodologies for explosion assessment, the circumstances in which the methods may be applied and their limitations. Part 3 of the Guidance will deal with the detailed implementation of these and other methodologies. Alternative methods to those presented may be used so long as they are justified by a risk assessment and provided their use leads to reducing risks to As Low As Reasonably Practicable (ALARP).
Issue 1, October 2003
.............................. 47 3................................................ 80 5............................ 38 3....................... control and mitigation ................ 28 3..........................10 Generating Exceedance curves ...............................................6 Explosion damage to passive fire protection (PFP) Systems ................ 98 5. 36 3...11 Loads on piping and equipment ..................... 76 4........................... 111 5........................................ October 2003 5 ............................................................ 84 5...............................................................................1 Introduction to explosion load determination ..... 75 4....................................................................................................................................................................................... 89 5...............................................................................................................5 Potential areas of interaction and conflict ..........4 Legislation..............6 Explosion overpressure determination ....... 7 1...........................................12 Existing installations .....................................................................................2 Safety Management Systems............................4 Special issues relating to installation type .......... 70 4....................5 Inherently safer design ...................................................................................................... 105 5.......... 61 3........................4 Overview of Guidance ..............................4 Risk screening ....................................................................................................................................................................................................................................................................... 41 3.................................1 History ................6 Detection.... 7 1.............7 Control systems and Safety Critical Equipment. 69 4............................................... 65 INTERACTION WITH FIRE HAZARD MANAGEMENT ..................................................8 Impulse and duration related to peak overpressure .....................................................................10 Implementation and monitoring .............................................5 Hazard identification and scenarios...................................................... 18 2.............................................2 Objectives ............................................................................................. 10 Explosion hazard philosophy ..................................... 76 4........................................................... 89 5........................................................................... 8 1..........2 Tasks for the determination of explosion loads .........................9 Acceptance criteria ....... 12 2.......................................................... 29 3.......................................11 Analysis methods ............................3 Determination of explosion frequency ................................................... 14 2.. 70 4.............................................................................................................................................................................................................................. 12 2.... 82 5...3 Goals................................................................ 7 1................................................................. 51 3................................................9 Design explosion loads................................................................................. 20 Explosion hazard management ................................................................. 58 3..............................................8 Mitigation and consequence minimization ......................................2 Common areas ....................................................................................................1 General ......................................12 Reporting Template for ALARP demonstration .............................................7 Development and application of Nominal Explosion Loads.............................................3 Risk reduction.................. 28 3................................................................1 General ...................... 79 Derivation of explosion loads...... 17 2........................ 101 5................................. 69 4....................... 73 4................1 Overview fire and explosion hazard management................ 85 5................................ aims and Principles ............................UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Contents 1 Introduction ......... 116 2 3 4 5 Issue 1........................ 33 3............... 102 5........... standards and guidance ..................4 Dispersion.........................................................3 Considerations by design phase........................................................................ 80 5......................... 60 3..........5 Ignition .........8 Operational issues...............................................................................................................................2 The explosion hazard .........................................................................................7 Use and effectiveness of deluge .....................3 Fire and Explosion Hazard Management .......
........ 180 7 8 APPENDIX A – ACRONYMS ....................................................10 Response of the Primary Structure..........9 Non-Linear Finite Element Analysis............................................................................ 117 RESPONSE TO EXPLOSIONS..............5 Structural performance standards ......................................................................................................................................................... 164 7...................................3 Response considerations .......................................................................... 127 6...... 178 REFERENCES ....13 6 The NORSOK Procedure for probabilistic explosion simulation.................................................................... 135 6.......................................................................4 Material properties for explosion response........................................GLOSSARY OF TERMS ............. 123 6.................................................................... 119 6.. 132 6................4 Industry and Regulatory Authority Initiatives .............CHECKLISTS...................... 142 6.. 164 7...................................... 164 7......8 Single Degree of Freedom Idealisations......................................................................................... 188 APPENDIX B ............................................................................................1 General ......................................................7 Response prediction methods ......................................................................................... 154 6......... 121 6... 138 6.............. pipework and vessels ........................................................................3 Best practice in explosion hazard design ......... 174 7..............1 Overview of explosion response................................... 163 DETAILED DESIGN GUIDANCE/DESIGNING FOR EXPLOSION RESISTANCE ...........................................2 The Design Sequence .......2 Information required for explosion response calculations.......... October 2003 .................................................................................................................. 119 6................. 215 6 Issue 1..............6 Structural assessment ................................................. 192 APPENDIX C ......................................................................................................................................11 Response of equipment................................................................... 159 6.... 151 6...12 Areas of uncertainty.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 5.................
Issue 1. At about this time the Fire and Blast Information Group (FABIG) was set up and has subsequently issued a number of Technical notes on specific aspects of fire and explosion engineering [2 to 8]. integrates loading and response development and provides a rational design approach to be used as a basis for design of new facilities and the assessment of existing installations. During the same period. 1. and in making operational modifications to. The concepts of Inherently Safer Design or ‘Inherent Safety’ are central to the approach described in this document both for modifications of existing structures and new designs.12] which are amongst the source documents for this Guidance. the HSE and others have funded approximately 300 individual projects and a number of Joint Industry Projects (JIPs) at a cost of £31million. October 2003 7 . mostly executed in Norway and following the probabilistic approach. The results of these and other major investigations are summarised in the new Engineering Handbook published by CorrOcean .1 Introduction History Following the Piper Alpha disaster a large Joint Industry Project called ‘Blast and Fire Engineering for Topsides Structures (Phase 1)’ was carried out between May 1990 and July 1991. Preventative measures are the most effective means of minimising the probability of an event and its associated risk. Appendix B. Other valuable work. the environment and the integrity of offshore facilities exposed to the explosion hazards. Risk is defined as the product of the probability of an event and its consequences.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 1 1. Three further phases of the Blast and fire engineering project JIP were conducted from 1994 to 2001. This document consolidates the R&D effort from 1988 to the present day. has resulted in the new NORSOK guidance documents [11. Phase 3a and Phase 3b  consisting mostly of experiments to define and determine explosion overpressure load characteristics and to provide a basis against which load simulation software may be validated. The relative maturity of the subject has enabled a better defined approach to be adopted in this Guidance document relying more on ‘good practice’ which has been developed and implemented over the intervening years. This Guidance is intended to assist designers and duty holders during the design of. Alternative definitions are given in the Glossary. which have contributed significantly to the understanding of the key issues. Phase 2 .2 Objectives The primary objective of this document is to offer guidance on practices and methodologies which can lead to a reduction in risk to life. The main deliverable from this project was the Interim Guidance Notes (IGNs)  and 26 background technical reports written by the participants and published by the Steel Construction Institute (SCI) in November 1991. The development of the Interim Guidance Notes was a major step forward which consolidated the then existing knowledge of fire and explosion hazards. offshore installations in order to optimise and prioritise expenditure where it has most safety benefit.
En se R A ba g. The philosophy of using past experience for the estimation of explosion loads has also been applied to the definition of explosion design load cases. the estimation of dynamic pressure loads. areas affected.3 Fire and Explosion Hazard Management A thorough understanding of all hazards and hazardous events. 8 Issue 1. Ri d ju d B Lifecycle implications Some risk tradeoffs / transfers Some uncertainty or deviation from Standard or best practice Significant economic implications Company Values Societal values C Very novel or challenging Strong stakeholder views and perceptions Significant risk trade offs or risk transfer Large uncertainties Perceived lowering of safety standards Figure 1 . the key figure of which is illustrated in Figure 1 below .).UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions An additional intent of this Guidance is to move the decision-making processes within the fire and explosion design field as much as possible towards a ‘Type A’ process from ‘Type B or C’ as defined in UKOOA’s document on decision-making. For these hazardous events the management process is given below: • • • identification of the hazardous events (coarse assessment). A substantial number of installations will lie in Areas A or B of the chart resulting in an approach which involves codes and Guidance based on experience and ‘best practice’ (as described in this document) and supplemented by risk based arguments where required. magnitude of the consequences. likelihood.The UKOOA decision making framework The framework in Figure 1 defines the weight given to various factors within the decision making process. ranging from those decisions that are dominated by purely engineering matters to those where company and societal values predominate. analysis and assessment of the hazardous events (type. This overall process is outlined in the OGP “Guidelines for the Development and Application of Health Safety and Environment Management Systems”. and to the assessment of explosion response. Q sk e. duration. Means of calibration Significance to decision Making process Codes & standards pr ac t ic e m ge t en Decision context type Nothing new or unusual Codes and standards Verification Peer review Bench marking Internal stakeholder consultation External stakeholder consultation G oo A Well understood risks Established practice No major stakeholder implications g rin is ee lys n na CBA gi da . etc. 1. October 2003 . reduction of the risks from fires and explosions through inherently safer design. These Guidelines add more detail to this process and applies it to fires and explosions. including fires and explosions. is at the heart of the Safety Management System (SMS) and it should be proactive to reduce risks.
control and mitigation measures needed for each hazardous event confirmation of the suitability and effectiveness of each of the measures selected. communication and implementation. duration and effects of each hazardous event. • The management of hazards to reduce the risks involves many interests which may often appear to conflict with each other. Thereafter. can operate them properly and that adequate maintenance schemes are in place. particularly design engineering disciplines and those who will have to operate and maintain the plant. involving all levels of personnel from senior management to junior staff from a number of different organisations. “Fire and explosion hazard management”. it is used to assess these arrangements to make sure that the high level performance standards have been achieved. The hazard management process should be employed in a timely manner and in accordance with the type. More information on specifically fire and explosion hazard management can also be found in Clauses 3 and 4 of this Part 1 of the Guidance. for an existing installation the process should be applied to current arrangements and modifications. scale. October 2003 9 . Issue 1. The SMS of each organisation should identify the relevant roles and responsibilities. verification. detection.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions • • • • • • • design to reduce the likelihood. i. identification and specification of the particular prevention. It is essential that all parties who can contribute to the reduction of hazards. The results should then be communicated to personnel operating the installation to ensure that they know the purpose and capability of all the systems. documentation. The fire and explosion assessment process is used in the lifecycle to provide information on which to base decisions and design systems. The FEHM process can be applied to new or existing installations: • for new installations it should start during feasibility studies and be fully developed during detail design. The lifecycle approach shows how to prepare and implement a strategy for the management of fire and explosion on an offshore installation throughout its life. control and mitigation measures. specification of the measures adopted. The process is a multi-disciplinary activity.e. It is important that the input and activities of these personnel are fully coordinated and managed. severity and likelihood of each hazardous event. from design through commissioning and operations to decommissioning. A more comprehensive and homogeneous view of the role of fire and explosion hazard management within the overall Hazard Management System (HMS) can be found in Part 0 of this guidance. This is developed firstly by inherently safer design. These should be assessed to determine if the high level performance standards are achieved and that risks are as low as is reasonably practicable. followed by prevention of identified fire and explosion hazardous events and then by the selection of detection. understand the hazards and are involved during the appropriate stages of the lifecycle. intensity.
A description of the dynamic nature of the response leads to a definition of the material effects of large strains and strain rates. brings together the approaches identified in the other clauses and incorporates additional design and operations experience to provide guidance on methods of detailed design. Clause 6.4 Overview of Guidance The main clauses of the Guidance summarise the content of the basis documents. discusses the choice and management of detection. The principles of ‘Inherent Safety’ are presented and appropriate methods of analysis dependent on the expected risk level are identified. ‘Response to explosions’ . This subject area will be re-visited in Part 3 of the Guidance. Workbook methods of calculation of the dispersion of a gas cloud are discussed. Guidance is then given on the information to be presented when preparing an ALARP justification.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 1. 10 Issue 1. This clause sets out the process by which it is decided what has to be done in any particular context and what factors to take into account. ‘Interactions with fire hazard management’ . Clause 5. describes the features of an effective Safety Management System. In this clause there is a review of the available methods for the simulation of explosions and the selection of the appropriate tools. A method of the development of nominal overpressures is described and the desirable characteristics of data on which such nominal overpressures could be based is discussed. areas of potential conflict between the management of these two types of hazardous event and the role of deluge in the mitigation of fires and explosions. Finally a brief review of the NORSOK protocol  is included. control and mitigation systems and identifies the main characteristics of the hazard discussing minimisation of the consequences of residual events which lead to an explosion event. October 2003 . The generation of exceedance curves is discussed along with their use in the definition of achievable design explosion load cases for response analysis. Clause 3. Clause 2. Clause 4. ‘Explosion hazard philosophy’ . ‘Explosion hazard management’ .19] describes how appropriate design explosion overpressures and dynamic pressures are derived. The tasks identified are linked with the relevant phase of a design project or the stage in the life of the installation. defines the goals which should be achieved in designing for and managing explosion hazards. Available methods for calculating the response of structures and equipment are presented with their relative merits and limitations. describes the determination of the response of structures and SCE’s to explosion loads. This is achieved by a screening process of the proposed or existing installation so that resources may be effectively targeted where they are needed. ‘Guidance on detailed design’ . ‘Derivation of explosion loads’[18. Performance standards are given for component and member response. Methods for the calculation of drag and dynamic pressure loads on piping and equipment are given. identifies situations where fires may precede or follow an explosion and deals with common areas in fire and explosion management. Clause 7. The guidance is presented in the context of best practice and identifies other industry initiatives which have generated detailed design and operating practice guidance.
2. 7.3. dynamic pressures. 2.5. ignition time and frequency Main References in this Document 3. release rate. strong shock and missiles Test results of response analysis against the appropriate performance standards. explosion loads duration and dynamic pressures on the structure and other SCE’s to be used in design Determination response Evaluation of Determine the response of the structure and other SCE’s to explosion loads including overpressure. Investigate the hazard with a view to prevention.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Task Installation screening Scenario definition Description Determine installation category low risk or higher risk Define leak position.3 Issue 1. 0. material type.12. mitigation detection. 6. control and mitigation to reduce the severity and frequency of the hazard Determination of Determine the explosion loads.4 0 2. 6.11. Clause 3 Clause 5 Prevention. overpressure. demonstrate that ALARP has been achieved Clause 6 2. hole size. October 2003 11 .3. control. 3. 6.5.
1 Explosion hazard philosophy General A release of hydrocarbon with immediate ignition will result in a fire.7. a risk screening method is described which classifies installations and compartments according to their risk level. In other situations a scenario based approach will generally be used where release scenarios are postulated and their consequences and probabilities of occurrence determined. (Terms in italics are defined in the Glossary given in Appendix B). Explosion events. mitigation and consequences are the subject of this Guidance. An essential part of the process of explosion assessment is the determination of the design explosion loads. The features of an effective Safety Management System (SMS) are identified and the choice and management of detection. These measures are combined in a risk matrix to give low. 12 Issue 1. detection. their elimination. reliable estimates of explosion loads may be available from previous assessments. If this is followed by later ignition the result may be an explosion. Design explosion loads are selected on the basis of their frequency of occurrence and should be accommodated by the Safety Critical Elements (SCEs) of the installation which will include parts of the structure. prevent and mitigate the explosion hazard for new designs. It is suggested that the number of SCEs which need to be considered in detail is reduced by classification into criticality categories with respect to the explosion process. These nominal overpressure values should be available for inclusion in Part 3 of this Guidance. control. prevention. For existing installations. The legislative basis is reviewed and some high level performance standards are given. The worst credible case of a stoichiometric cloud engulfing the whole installation or filling an entire compartment will frequently give rise to loads which are too severe to be reasonably practicably designed against. October 2003 . control and mitigation systems is discussed. The main characteristics of the explosion hazard are identified and the relevant issues relating to interaction with fire hazard management are discussed. A number of explosion scenarios on various installations have now been assessed using the techniques described in this Guidance. The measures for frequency and consequence severity are based on process complexity and the exposure potential for people on board. the goals which should be achieved in designing for and managing the explosion hazard are identified. Methods of deriving useable nominal explosion loads are discussed in 5. Explosion loads will include both overpressure loads and dynamic pressure loads on piping and equipment which are due to the movement of gases around them. medium and high risk categories. release of a flammable vapour or gaseous hydrocarbon.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 2 2. In this Guidance. The techniques of inherently safer design described in this Guidance are central to the approach to eliminate. In order to focus effort where it is most needed. The risk level is an indication of the level of sophistication to be used in the explosion assessment process. It is proposed that this data be used in the determination of nominal overpressures or for use in the early quantification of explosion hazards for new offshore facilities.
both through the justification of the choice of design load and from a determination of the impairment frequency of the SCEs under these loads. Typically this frequency of exceedance will be of the order of 10-4 to 10-5 per year depending on the risk to people on board. Further material has been included from Reference . Guidance on detailed design is included which brings together the approaches identified in the other sub-clauses and incorporates additional design and operations experience. other SCEs and their supports.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Reducing risks to ALARP must be demonstrated in all cases. to be robust enough to resist the loads without excessive deformation and ductile enough to undergo local deformations. There is at present a lack of consistency in the methods used for the generation of exceedance curves. A lesser explosion event. the impact on the SCEs and the overall individual risk from all sources. Further detailed information on all the topics addressed in this clause may be found in these Basis Documents. is also recommended which examined using elastic response techniques. The extreme design event explosion load cases are represented by the Ductility Level Blast (DLB). This additional load case may detect additional weaknesses in a design and serves as a robustness check of the structure. the results of which should be available for inclusion in Part 3 of this Guidance. The guiding principle here is that risks should be reduced to ALARP. Issue 1. which identifies the acceptable frequency of exceedance of the severity of these scenarios. October 2003 13 . The structure therefore is required. globally. This clause defines the approach to be taken in the assessment of explosion overpressures and related phenomena and describes how those factors are incorporated into the features of the installation to bring about a position where the risk from explosion events is reduced to as low as reasonably practicable (ALARP). The determination of the acceptability of the results of an assessment is also discussed. escape routes or means of escape. the Strength Level Blast (SLB). Other industry initiatives which have generated detailed design and operating practice guidelines are presented and discussed. This is determined using the associated frequencies of occurrence or exceedance and the space averaged overpressures derived for the relevant scenarios. preferably without rupture. A common terminology has been developed to facilitate the transfer of complete and unambiguous information between the disciplines. The aspects include comparison with high and low level or element specific performance standards and risk acceptability criteria. This clause is based on References  ‘Explosion hazard philosophy and  ‘Explosion hazard management’. The HSE has embarked on a study to clarify these. An acceptable level of risk should be identified within the ALARP framework. Guidance is given on the information required by the response disciplines from the explosion loading specialists. panels and barriers must be shown to resist the design explosion loads without collapse and with minimal risk of impairment of the Temporary Refuge (TR). The supporting structure.
This guidance incorporates the findings of recent experiments particularly the full scale tests at the Advantica Spadeadam Test site [9. having previously identified the potential explosion scenarios.2 The explosion hazard Gas explosions can be defined as the combustion of a premixed gas cloud containing fuel and an oxidiser that can result in a rapid rise in pressure. This clause sets out the process by which the engineer/manager decides what has to be done in the particular context. October 2003 . in the definition of nominal overpressures or for use in the early quantification of explosion hazards on new offshore facilities. A method of installation screening is described which enables the general risk level of the installation to be determined. equipment or pipework response provide results with a much reduced variability compared with those used in the determination of the explosion loading. where possible. 2. structural response and consequent risk. with negligible reduction in reliability. 14 Issue 1. Where such comparisons are not possible. 10]] . These tests have produced a large amount of data which has enabled modelling techniques to be greatly refined to give more accurate definition of explosion characteristics. and managing explosion hazards. normally manned or unmanned. Approaches for assessing existing installations are also provided. equipment or piping responses to explosion These factors may lead to the need to assess previously determined overpressures. alternative means of explosion mitigation must be sought in order to reduce the magnitude or risk of exceedance of those design load scenarios The structure and other SCEs should be able to accommodate these design loads. Where practicable designs cannot be achieved. These results can then be used in Quantitative Risk Assessment (QRA). In order that these goals can be achieved the philosophy lays out the requirements for a Safety Management System (SMS). methodologies are proposed for the quantification of explosion hazards using the latest modelling techniques. including explosion hazards. This clause defines the goals that should be achieved in designing for. assessed and controlled to achieve a design where risk is reduced to ALARP. It may then be necessary to consider further measures to mitigate risk and reduce this to ALARP. Gas explosions can occur in enclosed volumes such as industrial process equipment or pipes and in more open areas such as ventilated offshore modules or onshore process areas. The structure and other Safety Critical Elements (SCEs) should be designed for the identified explosion scenario design load cases. The techniques used in determining structural. The load exceedance data can therefore safely be applied to the response results. A number of explosion scenarios on various installations have now been assessed using these refined models and it is proposed that this data is used. A Safety Management System establishes a safety environment whereby hazards in general. These design loadings will have an associated probability of exceedance. which is the subject of the next clause. The philosophy is applicable to all offshore design concepts whether fixed or floating. A philosophy for existing installations is also proposed where the design explosion scenarios require to be assessed in the light of improved understanding and modelling of explosion loads and structural. which then determines the appropriate level of analysis sophistication which should be used.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions The approach described in this Guidance will enable explosion scenario design load cases to be rigorously defined. are identified.
All of the above points from 1 to 5 can affect the explosion overpressures in this type of environment. A detonation is propagated by a shock that compresses the flammable mixture to a state where it is beyond its auto-ignition temperature. location) 6) Non-homogeneous Cloud density 7) Ignition timing Confinement is defined as a measure the proportion of the boundary of the explosion region which prevents the fuel/air mixture from venting which is the escape of gas through openings (vents) in the confining enclosure. The overpressure caused by the explosion will depend. amongst other things. October 2003 15 . This further increases the velocity and turbulence in the flow field ahead of the flame leading to a strong positive feedback mechanism for flame acceleration and high explosion overpressures. The disturbance is subsonic relative to the un-burnt gas immediately ahead of the wave. Two types of explosion can be identified depending on the flame propagation rate: A deflagration is propagated by the conduction and diffusion of heat. on: 1) The gas or gas mixture present 2) The cloud volume and concentration 3) Ignition source type and location 4) The confinement or venting surrounding the gas cloud 5) The congestion or obstacles within the cloud (size. This turbulence increases the flame surface area and the combustion rate. The shock wave and combustion wave are coupled and in a gas-air cloud the detonation wave will typically propagate at 15002000m/s and result in overpressures of 15-20 bar. hot combustion products are created that expand to approximately the surrounding pressure. Most vapour cloud explosions offshore would fall into the category of deflagrations. It develops by feedback with the expansion flow. The overpressures are not limited to the 8 bar maximum typical of completely confined explosions. Gas explosions in more open environments can also lead to significant overpressures depending on the rate of combustion and the mode of flame propagation in the cloud. shape.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions For an explosion to occur a gas cloud with a concentration between the upper flammability limit (UFL) and lower flammability limit (LFL) must be ignited. Issue 1. As the surrounding mixture flows past the obstacles within the gas cloud turbulence is created. Typical flame speeds range from 1-1000 m/s and overpressures may reach values of several bars. number. A typical vapour cloud explosion on an offshore installation would start as a slow laminar flame ignited by a weak ignition source such as a spark. Congestion is a measure of the restriction of flow within the explosion region caused by the obstacles within the region. The combustion wave travels at supersonic velocity relative to the un-burnt gas immediately ahead of the flame. As the gas mixture burns.
For this type of object the dynamic pressure associated with the gas flow in the explosion will dominate. Secondary. external explosions may result as the unburnt fuel/air mixture comes into contact with the external (oxygen rich) atmosphere. 16 Issue 1. The peak energy for typical projectiles may be calculated from the dynamic pressure load time history and their mass. Typically within an explosion there will be a strong variation of pressure in space and time. The importance of operational aspects is also shown in proportion of leaks attributable to poor inspection and monitoring. The duration of the positive phase in an explosion can vary greatly with shorter durations often associated with higher overpressure explosions. the overpressures applied to the front and reverse side of such items will be of approximately the same magnitude at any moment in time and in this case the overpressure difference will not be the only load component on the object.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Large components of the structure such as solid decks or walls experience loads due to the pressure differences on opposite sides of the structure. The overpressure at a location within a gas explosion will typically rise to a peak value and then fall to a sub atmospheric value before returning to zero overpressure. There will typically be localised high regions of overpressure with lower values of average pressure acting on large components. Small objects may be picked up during the explosion. such as piping. Typical durations range from 50 to 200 milliseconds with longer durations common in large open areas such as the decks of Floating Production. October 2003 . A blast wave will be generated which will propagate away from the explosion region and may impinge on adjacent structures. Avoidance of potential leak sources in design therefore needs to consider these above issues in particular. For smaller objects. • • • • 11 % are due to incorrect installation 26 % from degradation of materials (excluding corrosion and erosion) 11 % of all releases are due to vibration/fatigue 19 % of all releases were due to corrosion and erosion It is considered that 40 % of equipment related releases are attributable to poor design and 38 % to inadequate inspection and condition monitoring. It is instructive to be aware of the nature of the hazard and focus effort on the most frequent sources of the hazard as given by the history of releases experienced to date. Storage and Offloading vessels (FPSOs). creating secondary projectiles. The HSE document OTO 2001 055  states that for the UK sector of the North Sea: • • • • 61 % of all releases are from pipework systems 11 % of all releases are from small bore piping 15 % of all releases are from flanges 14 % of all releases are from seals and packing Of the causes. These can affect the venting of the compartment and enhance the overpressure within.
The following high level goals therefore define what is necessary to achieve an ‘ALARP’ design with respect to explosion hazards: • • • • • • the identification of all areas of the installation where there is potential for explosion events to occur. control. 2. In this case it is: • the achievement of a condition where the explosion risk for the installation is reduced to As Low As Reasonably Practicable (ALARP) Explosion risk is defined as risk from the initiating event and subsequent escalation. the implementation of such mitigation measures that are necessary to reduce residual explosion risk to as low as reasonably practicable (ALARP). Issue 1. the minimisation of exposure of personnel to explosion and subsequent escalation events.27] and the HSE/UKOOA publication on the subject  . October 2003 17 .3 Goals. • • • • • • • A further set of more specific goals ensure that personnel will reach a safe location in the event of a major accident event: • that one escape route to the Temporary Refuge (TR) remains functional at all times. the minimisation of the frequency of explosion events and. the minimisation of the consequence of explosion events the implementation of a safety management system which ensures that the above goals are consistently achievable. or if this is not achievable. The Minerals Management Service (MMS) of the US also publishes data on incidents on the Gulf of Mexico Further similar considerations are included in sub-clause 0 ‘Hazard identification and scenario definition’. the design of structures and the specification of safety critical elements (SCE’s) to prevent. the minimisation of asset loss from explosion events. the implementation of operational management systems to minimise the potential for explosion events to occur through the life of the installation including decommissioning and abandonment. the minimisation of the environmental impact from explosion events. The means of achieving this can be broken down into a number of high level goals whose attainment will achieve the top level goal. the determination of key explosion hazard parameters.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Sources of release data include WOAD  . the performance of suitable and sufficient assessments of the consequences and risks associated with the defined explosion hazards. aims and Principles Goals define what the philosophy requires of the end product. mitigate and withstand the consequences of design explosion events. The top level goal defines what is required from the overall process. detect. OREDA  release statistics published annually by the HSE [26. the elimination of the potential for explosion events to occur.
The regulations also require mitigating measures to be specified and put in place and for performance standards to be set for safety critical measures to prevent. October 2003 . These include the ‘Management of Health and Safety at Work Regulations 1999’. The Safety Case Regulations SCR  require that all installations in UK waters have an acceptable Safety Case. These measures include identification of explosion hazards and the evaluation of their consequences and likelihood (reg. The Duty Holder must ensure that effective evacuation. that a means of evacuation to be available at all times. 18 Issue 1. Information regarding the following issues is required to be addressed in the Safety Case:• • • • identification of major hazards. Various regulations are enacted under the HSWA.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions • • • • the TR and its supports will maintain their integrity in all design explosion events. This imposes a responsibility on the employer to ensure the safety at work for all employees. evaluation of risks associated with the identified hazards. safety and welfare of their employees.4. Regulations 9 to 12 specify the types of measures which are required for prevention. More specifically related to fire and explosion risk are the Prevention of Fire and Explosion and Emergency Response on offshore installations (PFEER)  regulations which place on the Duty Holder the requirement to take appropriate measures to protect persons from major hazards including fires and explosions. Legislation may be added during the lifetime of this guidance. which place an obligation on the employer to actively carry out a risk assessment of the workplace and act accordingly. standards and guidance UK Legislation This sub-clause details the major legislation covering explosion risk.4 2. details of the Duty Holder’s (Safety) Management Systems. details of appropriate measures taken to reduce these risks to as low a level as is reasonably practicable. detection. Employers have to take reasonable steps to ensure the health. the ability of personnel to escape from. control and mitigate explosion hazards. The primary legislation governing safety in the workplace is the ‘Health and Safety at Work Etc. Act 1974 (HSWA)’. Risks assessed will include those from fire and explosion.1 Legislation. 5). communication and control of emergencies. It is not exhaustive as any legislation covering general safety or requiring a safety risk assessment to be performed. escape recovery and rescue will occur in the case of an explosion event (Regulations 14 to 17). will be relevant where the potential for an explosion exists. that all large off-site inventories are isolated in the event of a design explosion event. and to shelter safely from the effects of an explosion event and the ability to evacuate to a safe location where recovery can take place is not compromised 2.
2. relating specifically to the design of installations against explosion events.SCEs).4. • • Institute of Petroleum. describes how any quantified risk assessment (QRA) has been used and how uncertainties have been taken into account (paragraphs 75-82).2 APOSC The ‘Assessment Principles for Offshore Safety Cases’ document (APOSC)  modifies SCR with respect to the development of PFEER SI 1995/743. 2.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions The Design and Construction Regulations DCR  amend the SCR by placing a responsibility on duty holders to prepare a suitable verification scheme for their installations to ensure independent and competent evaluation of those elements of the installation which are critical to safety (known as Safety Critical Elements . which may be used in potentially explosive atmospheres. • • • Issue 1. Performance standards are used to define the functionality and integrity of these safety critical elements. Classification of hazardous areas’.4. standards and guidance Guidance documents are available from the Health and Safety Executive for the legislation mentioned above. BS EN 60079-10. . evaluates the risks from the identified major accident hazards (paragraphs 49-74). standards and guidance are available covering elements related to the explosion hazard. The ATEX (Atmospheric Explosion) Directives 94/9/EC and 1999/92/EC cover electrical and mechanical equipment and protective systems. The structured approach listed above is generic. escape and rescue arrangements (paragraphs 113-144). They will define how these SCEs are expected to function during and after explosion events. Explosion prevention and protection’  is available. so that there are no specific requirements for how an assessment of explosion hazards should be carried out. However there is little guidance. describes the evacuation. October 2003 19 . Various codes.3 Codes. describes how major accident risks are managed (paragraphs 90-112). identifies and describes the implementation of the risk reduction measures (paragraphs 83-89). MAR-1995 and the DCR 1996 The APOSC document requires the following be demonstrated for Major Accident Hazard assessments: Acceptable safety cases will demonstrate that a structured approach has been taken which: • • • • • • identifies all major accident hazards (paragraphs 38-48). namely: For ignition prevention and Hazardous Area Classification. apart from ISO 13702  and PFEER. ‘Electrical apparatus for explosive gas atmospheres Part 10. The most relevant are the Interim Guidance Notes (IGNs)  which this Guidance will replace. (IP15) . For equipment in hazardous areas ‘BS EN 1127-1 : 1998 Explosive atmospheres. August 2002. . BS 5958: 1991 ‘Code of practice for control of undesirable static electricity’. ‘Area classification code for installations handling flammable fluids’.
htm HSE Books have published a guide which sets out an overall framework for decision taking by the HSE (Reducing Risks. 39].uk/hid/spc/perm12. environmental or commercial risk by setting a Safety Integrity Level (SIL). Guidance Note PM84 HSE. Recent experimental evidence is available relating to dispersion of explosive clouds [10. Alongside UK legislation.hse. The advantage of an inherently safer design or the ‘Inherent Safety’ design approach is that it attempts to remove the potential for hazards to arise. • • Policy and Guidance on reducing http://www.1 Inherently safer design Introduction Having determined the installation concept it is necessary to manage fire and explosion risk within the constraints imposed by the subsequent offshore layout.gov. It does not rely on control measures. HSE guidance relating to Gas Turbine enclosures is relevant as the principles mentioned are applicable to explosions in general.2 Goals of inherently safer design The key goals of ‘Inherent Safety’ may be summarized as:• Fewer and smaller Hazards Issue 1.htm risks to ALARP in Design  . Guidance on the demonstration of ALARP is available throughout this Guidance and from the following sources.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions These documents only cover operational leaks rather than accidental releases. Critical loops are designed according to their criticality in mitigating personal.5.gov. essential elements and performance standards for each of the systems required to manage possible hazardous events on the installation. Inherent safety avoids this potential by aiming for elimination rather than protection and the preference for passive protection over active systems.hsr. 2.5.uk/dst/alarp1. In setting a SIL it is acknowledged that there is failure potential although this is designed to be inversely proportional to the importance of the loop in risk mitigation. • ‘Control of Risks at Gas Turbines Used for Power Generation’. EN ISO 13702  also addresses the need to develop a Fire and Explosion Strategy (FES) which describes the role.uk/dst/r2p2. October 2003 20 . 2. Protecting People.5 2. Principles and Guidelines to Assist HSE in its Judgement that Duty Holders Have Reduced Risk as Low as Reasonably Practicable  . which is available in hard copy form  and as a free download from http://www.hse. systems or human intervention to protect personnel. often referred to in shorthand as R2P2). They do not define the extent of hazardous areas from the point of view of explosion and fire risk. There is always the potential for the systems to be damaged in a hazardous event. All control systems have the potential for failure to operate as intended – generally expressed as the Probability of Failure on Demand (PFD). http://www.pdf..gov.
Table 2. October 2003 21 .1 summarizes the major ‘Inherent Safety’ and control features necessary to achieve the goals stated above: There may be some conflict between the various features of inherent safety.1 Inherent Safety Features to Achieve Goals Goal to minimise explosion Inherent Safety features to achieve goal risk Minimisation of potential leak sources and release potential • • • • • • • • minimise number of pipe joints maximise welded pipe joints minimise intrusive instrumentation eliminate/minimise small bore pipework minimise offshore processing and process complexity minimise vibration minimise corrosion/erosion ensure effective inspection Minimisation of ignition potential • no naked flames in live plant • audit and review safety management system with respect to hot work procedures • insulate hot surfaces* • effective earth bonding • hazardous area zoning (area classification) • effective maintenance regime • separation of quarters and non-operational personnel from process areas • minimisation of maintenance load • remote operation of processes • simplification of the offshore process • separate accommodation platforms • fully rated blast barriers and primary structures • simplification/minimisation of offshore processing • use of small isolatable inventories • effective isolation from large inventories upon gas/leak detection • effective blowdown of inventories Minimisation of Personnel On Board (POB) exposure to blast effects Minimisation of hazardous inventory Issue 1. For example. The balance between such features will be discussed further in this Guidance. increased compartmentalisation will reduce the size of a potential explosive gas cloud however this will also decrease the potential for natural ventilation and increase confinement.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions • • • • Fewer Causes Reduced Severity Fewer Consequences More effective management of residual risk The following Table 2.
October 2003 .UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Table 2.1 Inherent Safety Features to Achieve Goals Goal to minimise explosion Inherent Safety features to achieve goal risk Minimisation of potential explosive cloud size • • • • • • maximisation of ventilation potential minimisation of inventory pressure minimisation of potential leak rate minimisation of hazardous inventory minimisation of module size/compartmentalisation subsea completions Minimisation of congestion • simplification/minimisation of offshore processing • optimisation of module layout • segregation of congestion and explosion leak sources • • • • grated decks open sided modules blow-out panels/louvres optimisation of module layout Minimisation of confinement Maximisation of ventilation • minimisation of confinement • platform orientation to make maximum use of prevailing wind direction • equipment layout to avoid ‘dead spots’ • platform aspect ratio to maximise ventilation • • • • • • • • • • • • • • • limit module size minimise offshore processing add partitions to limit maximum dimensions review of aspect ratio of module dimensions effective inspection programme effective maintenance procedures resistance to overpressure effects protection from blast wind effects protection from severe vibration effects protection from structural displacement effects separation of process areas from critical nonhazardous areas Limitation of potential flame front length Monitoring and maintenance of SCE integrity and functionality Maintain effective hazard management safety leadership and focus effective safety management system prevention rather than protection preference for passive systems of control and mitigation over active systems Note *Corrosion under insulation is a major cause of line failure and high operational cost. 22 Issue 1. hence insulation should be avoided where practicable.
3 Effective Management Of Residual Risk The risk which cannot be eliminated by the application of inherent safety methods is referred to as residual risk. 2. that minimise the consequences and risk to personnel. Inherent safety methods can also be applied to the management of the residual risk by consideration of the general principles indicated below: CONTROL is better than MITIGATION is better than EMERGENCY RESPONSE. Further details regarding the inclusion of inherent safety and control features are given in other sub-clauses of this Guidance.1 details inherent safety and control features that minimise the potential for explosion to occur.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Table 2. rather than by including mitigation measures at a later date. These features should be built into the early design of the installation. As regards systems to reduce risk PASSIVE systems are more reliable than ACTIVE systems are more reliable than OPERATIONAL systems are more dependable than EXTERNAL systems This indicates:• • The use of passive rather than active control and mitigation systems No reliance on personnel to prevent.5. October 2003 23 . or if an explosion should occur. control or mitigate hazards The Table below defines and expands upon some of the terms used above  Issue 1. Inherent safety carries on through the life of the installation continuing through the operational phase by adherence to effective inspection and maintenance regimes and by ensuring that management systems and related procedures are followed as intended.
testing and maintenance. they are less reliable. requiring only inspection and maintenance. depressurization systems. Selection of lowest risk option unless cost is grossly disproportional to risk gain. They require inspection. physical damage or removal. 2. thereby reducing the need for people to be in hazardous locations. They are preferred because they are inherently the most reliable. There is clearly further room for error due to the longer communication lines and frequent changes of the people involved. Typical examples are. relief panels and natural ventilation. maintenance. HAZOP. inspection and condition monitoring. These are systems that may require mechanical or electrical facility. Typical examples are. The project phases at which these processes are most appropriate are also identified Table 2. Examples are. Concept selection Process review. HAZID Issue 1. with associated minimum competences and procedures. They are susceptible to failure and downtime of these systems. Typical examples are corrosion allowances. welded connections. They do not need to react to the hazard or need operator input at the time of occurrence in order to be effective. Their effectiveness is wholly dependent upon the future operator. October 2003 Conceptual FEED 24 . and its direct workforce. and are thus susceptible to human error or omission. Effectiveness is dependent upon effective contracts and audit. As such. They also cause increased numbers and activity on the facility. either to initiate the system. layout.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Table 2. External These are systems that depend on the correct reaction of people beyond the company itself. Separate accommodation platform fully rated barriers and primary structure subsea completions.4 Processes for Achievement of Inherently Safer Design Goals The Table below details some of the processes which should be used to achieve the goals of inherently safer design.3 Achievement of goals Goal Means of Achievement Fewer Hazards Minimise offshore Giving Rise to processing Explosions Simplify the process employed Phase Conceptual Key Processes Concept analysis. and isolation of a third party feeder pipelines.5. the dependence on the competence of a supply boat master to avoid riser impact. particularly where their failures may not be visible. Active Operational These are systems that depend primarily upon people. The only modes of failure are long-term deterioration. or control signals in order to work. risk ranking.2 Systems to minimise the consequences of accidental events Type Passive Description These are systems that act upon the hazard simply by their presence. who should agree to the dependence on these measures. or to carry out the whole function. fire and gas detection and active fire and blast suppression systems. As such they can be the least reliable and require sufficient trained people to be on the facility in order to ensure their operation.
No welding/gas cutting at live plant Layout – minimise piping and small diameter items. corrosion policy (piping specification). instrument philosophy Concept selection.no naked flames Insulation specification (hot surfaces) Replacement of light fittings with flood lights  Operation/Maintenance philosophies. Minimisation of process complexity Area classification (hazardous area) designation. choice of materials/piping specification Process design.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Goal Means of Achievement Less maintenance burden Phase Conceptual FEED FEED FEED Conceptual FEED Key Processes Concept selection Process review. minimisation of process HAZOP. HAZOP Piping philosophy to promote welded connections Piping philosophy. Seek to minimise process operating pressures Layout (may conflict with confinement) Fewer Causes Less piping joints of Explosions Less small bore piping Less maintenance Less intrusive instrumentation Less corrosion Less vibration FEED FEED / Detailed Design Detailed Design Fewer ignition sources Conceptual FEED/Detailed Design ALL Operation Reduced Severity Explosions Less congestion of Less confinement Lower process pressures Smaller potential explosion zones FEED / Detailed Design FEED FEED FEED Issue 1. fewer walls/partitions Design Basis/Process Philosophy. Elimination or minimisation of hot work at live plant. Layout of small bore items Layout – grated decks. resilient mountings for mechanical plant. maintenance philosophy (replacement v. October 2003 25 . Safety Philosophy . maintenance) Instrumentation philosophy stating preference for nonintrusive instrumentation Corrosion philosophy. piping supports.
October 2003 . minimise maintenance Concept selection .removal of lose items from module. remote operation. minimise Normally Unattended Installation (NUI) visit frequency. Reduce confinement. Aim to achieve greater than ‘adequate’ level (> 12 air changes per hour (ach)) Concept selection minimisation of offshore processing Concept selection minimisation of offshore processing. no storage of equipment Segregation of explosion risks and major fire escalation HAZID/HAZOP .use control systems to make decisions in structured manner according to ESD hierarchy.minimise manning levels. Fewer Unmanned Consequences installation Resulting from Explosions Lower manning levels Conceptual Conceptual Lower exposure of personnel to the explosion hazard and escalating events No missile generation Conceptual / FEED Detailed Design Operation Escalation More effective Prevention rather management of than protection residual explosion risk Passive systems rather than active No reliance on personnel to prevent. Ventilation Study. control or mitigate hazards Conceptual/FEED All All FEED/Detailed Design 26 Issue 1.identify hazards then use hazard management hierarchy – elimination of hazard being first goal Aim for inherent safety rather than use control systems which have failure modes Control/Process/ESD Philosophies . Housekeeping . maximise separation of control and quarters areas from process areas Layout & fixing details . minimise maintenance.fix small items to robust equipment away from high blast wind areas. installation orientation.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Goal Means of Achievement Shorter flame path length More ventilation Phase FEED FEED Key Processes Layout – avoid long narrow modules Orientation Study.
October 2003 27 . maximise separation of control and quarters areas from process areas No exposure of safety critical systems to hazards All Issue 1.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Goal Means of Achievement Minimise number of safety critical systems Phase All Key Processes Minimise process complexity and manning levels so that quantity of safety critical elements are reduced in number Minimise the process.
If this is the case an overall. The HS&E philosophy for the project should reflect the goals listed in sub-clause 2. When hazards are identified with potential to give rise to risk to personnel or the environment. • • • the probability of explosion occurring.3 focus on the project management system and how getting that right forms the basis of managing the explosion hazard. Thereafter the aims shall be (in order): • • • • remove personnel from the consequences of the explosion or escalating events. bridging plan should be available to cover the interfaces between the individual plans. their control and mitigation. This clause deals with two main topics. the overpressure experienced once the explosion has occurred. The means of achieving this are encapsulated in the Safety Management System (SMS) of the Duty Holder and designer and more specifically the HS&E Plan for the project. October 2003 . install mitigation measures to reduce escalation and otherwise protect the workforce. safety and environmental issues will be handled throughout the life of the project. then risks shall be reduced to As Low As Reasonably Practicable (ALARP). These sub-clauses contain a description of the elements of a safety management system (SMS) that should be in place to ensure that a framework exists within which a successful design can be produced. The HS&E Plan(s) will define how health. inherently minimise the potential size/severity of the explosion event. sub-clauses 3. Management of explosion hazards and Safety Managements Systems are dealt with in the next clause. the assessment of those hazards and their elimination or minimisation. The safety management system SMS should specify the need for an HS&E Plan for all projects where significant risk to personnel or the environment is possible. 28 Issue 1.3. the control and mitigation of consequences of explosion. More than one plan may be in place.1 Explosion hazard management General Effective safety management requires a structured approach to the identification of hazards. Much of the discussion concerning the SMS is applicable to all aspects of design not specifically explosions.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 3 3. environment and asset If risks cannot be eliminated and where risk to individuals falls below that which is unacceptable but above the level of broad acceptability. This clause provides the engineer with the means of managing the various factors which contribute towards. covering different parts or phases of the project. use detection and control systems to warn personnel and to minimise the explosion event. in the first instance the aim should be to eliminate or minimise the hazard and the risks associated with it.2 and 3.
October 2003 29 .2 Safety Management Systems Safety Management Systems will have common features for dealing with the fire hazard.2 Content Each management system will have the same basic elements. these will include systems relating to:• • • • • • • The policy and objectives Organization resources and procedures Risk identification and evaluation Planning of work activities (including emergency response) Implementation and monitoring (including means of measuring performance) Audit to assess the operation of the SMS in practice Review of the system – implementation of the audit process and review of the targets based on feedback from the monitoring system. The system described in this sub-clause is generic in this respect. The structure of the SMS should comply generally with accepted standards from the regulator and industry . The SMS will demonstrate the means by which the organization’s safety policy is put into effect.2. with effective leadership to ensure that the above elements are diligently carried out. The management needs to be aware of the safety policy and aims and provide the necessary resources to ensure that these aims are fulfilled. All of the above elements need to be underpinned by a commitment to safety from the organization’s management at the highest level. the explosion hazard or any other accidental loads which can be envisaged. 3.1 Purpose A Safety Management System (SMS) provides a framework whereby an organization can be assured that So Far As Is Reasonably Practicable (SFAIRP) its operations can be undertaken in a demonstrably safe manner.2. 3. Issue 1. 3. Further guidance on health and safety management can be found in the Basis Document covering management systems .UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions The remaining sub-clauses focus on management of the explosion hazard itself and how an installation may be designed to reduce risk from explosions to as low as reasonably practicable (ALARP).
This HSE plan defines how safety and environmental aspects of the project will be managed. usually at the Front End Engineering Design (FEED) stage. These should.2.Identification .Evaluation .Management Audit Figure 2 . • • • • • • be focused to the activities of the organization be consistent with other sectors of the organization be publicly stated and accessible be consistent with or improve upon current legislation have at least the same level of importance as other policies of the organization commit the organization to improving safety performance 3.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Policy & Objectives Review Organization Resources Procedure Leadership and commitment Implement Monitor Planning Risk .2.4.3 Policy & Objectives The organization must define and document its HSE policy and objectives. and outline of the project 30 Issue 1. Typical contents will include:• an introduction to.Industry Standard SMS 3.2.1 Organization. October 2003 . Resources & Procedures Project HSE Plans It is general practice within the Offshore industry (for all but minor projects) to have specific HSE Plans drafted at the earliest practicable stage.4 3.
as a minimum. Inputs from external organizations may include:• • • • • • CFD modelling of natural ventilation and gas dispersion simulation of the explosion to give overpressures and exceedance curves the design and manufacture of pre-assembled units (PAUs). the quality of work carried should match the standards expected from the company system. e. including accident history and details of any prosecutions or improvement notices served supply evidence that they have a quality management system complying with ISO 9001 or equivalent. process skids construction activities the installation of the facility hook-up and commissioning Poor standards of workmanship. prospective suppliers should be required to. from board level downwards details of the roles and responsibilities of those persons named in the safety organizational structure. checking or other incorrect procedures are potential contributing factors in receiving a sub-standard product.g. services and sub-contracts relevant legislation and procedures available to undertake the above activities an audit plan. The issues noted above are also important when dealing with bought-in items and services.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions • • • • • • • • • • a declaration of project’s specific HSE policy statement and of the organization’s HSE policies a summary of the safety criteria and targets set for HSE design the project’s organization and reporting structure specific to HSE issues.e. they should be excluded from the supply process or subjected to an audit to identify shortcomings and preferably be coached to improve. Issue 1. All reasonable efforts therefore need to be made to confirm that the supplied goods and services are of an appropriate standard. including all levels within the project details of how the technical competency of the operational and design personnel is guaranteed details of the HSE activities on the project needed to ensure the appropriate level of safety and environmental integrity a schedule of HSE deliverables (i. documentation to be produced) the control of quality regarding bought-in items. October 2003 31 . • • • • supply a copy of their safety and environmental policies supply evidence that they have a functioning safety management system supply details of their safety and environmental performance. To achieve this. The high standard of design and product may be compromised by the input of others. If the supplier cannot give satisfactory information on these topics.
• • • • at concept selection stage to have safety and environmental input to ensure that the option chosen is the one with the lowest risk. It is a requirement that procedures and design guides will be available for engineers to follow and consult at all stages of the project. should include the process by which prospective suppliers are assessed and the criteria that need to be met to qualify for the work.3 Procedures It is one function of the SMS. during the detailed design stage to ensure that control and mitigation systems are incorporated as necessary. Consideration of the explosion issues at different stages in the design of an installation is discussed in sub-clause 2. All personnel should be aware of the standards required with respect to safety and environmental issues and what their role is in achieving this. as far as is reasonably practicable. Resources in the context of competent personnel must be available during the following project stages.2 Resources In order to carry out the work in a timely manner and with the necessary degree of competence sufficient resources must be made available. to ensure that activities are carried out in a correct and consistent manner and at the correct stage. With respect to explosions the definition of the hazard must be identified at the correct time in the design process such that actions can be taken to accommodate it.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions The Plan should reinforce the premise that HSE is a line management responsibility.4. They should state the methods by which the activity should be carried out. at the FEED stage to ensure that layout issues are handled to minimize explosion hazard and risk. 3. Where hazards or potential benefits are identified it is the individual’s responsibility to identify the scenario and ensure that it is resolved. that each discipline and each individual is responsible for the safety and environmental impact aspects of his work. 32 Issue 1.4 of the Commentary. October 2003 . Personnel undertaking the work should be suitably trained or experienced in the methods that need to be used. 188.8.131.52. The Plan should be approved by the appropriate representative of management and circulated to all disciplines for them to cascade to individual worker level. Procedures for dealing with bought-in services. throughout the design period to quantify as necessary the overpressure that may be encountered and the resultant risk. With respect to explosions suitable training is particularly important as new research is continually developing fundamental information on the mechanism as well as innovative means of control and mitigation. The role of the Safety Group to police the work of others beyond the usual interdisciplinary review of outputs cannot be relied upon. So Far As Is Reasonably Practicable (SFAIRP).4. by whom the work should be done and any performance standards or criteria that should be met.
The cost effectiveness of safety effort expended during the selection and definition phases is significantly better than during the later phases. The impact on project schedule of safety effort expended during the selection definition phases is significantly less than during the later phases. There should also be a driving force from management to achieve a continual improvement in HSE performance. on the overall HSE performance of the project and the end product. 3. or may have.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Procedures should be available at key locations (for instance in the design office and with the QA engineer) for consultation at any time. They should be regularly reviewed and updated as necessary to include new systems or as a result of the findings from audits undertaken.3 Risk reduction The exploitation of offshore hydrocarbon reserves may be considered to consist of the following phases: • • • • • • • • Concept selection Concept definition Front end engineering design (FEED) Detailed design Fabrication/Construction Installation/Start-up Operation Decommissioning The concept selection and definition phases are traditionally the time in which the most significant project decisions are taken. Figure 3 below shows that: • • • Maximum safety leverage is available during the selection and definition phases. This should seek to promote each member of the design team to look at the work he/she is doing and understand the effect it has. Issue 1. The majority of these decisions have a significant influence on the hazards which must ultimately be addressed by the engineering work and subsequently during the operational phase of the project. with HSE issues featuring prominently at regular project progress meetings and at meetings to all levels within the management hierarchy.4.2. 3. October 2003 33 . Project management should maintain a high profile for HSE issues throughout the life of the project. Effective leadership should also encourage involvement and participation in the HSE management process at all levels.4 Leadership Senior management should provide strong. visible leadership and commitment to HSE issues and ensure that this commitment is translated into the provision of the necessary resources to ensure that all the elements of the SMS as described above can be carried out effectively. This should further encourage each member of the design team to air any concerns they might have with regards to HSE performance.
It should be noted that small bore pipework added late in the project may increase the severity of the explosion overpressure. the incremental costs of dealing with the hazard later in the project life increase significantly. In many cases a late design/construction change can be a magnitude greater in cost than had the feature been incorporated in the design basis of the project. whereas had the measure been introduced earlier in the project life this would not have been the case. develop an explosion simulation model (as described in Clause 5) and continually update this model in line with design development. 34 Issue 1.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions • Unsuccessful safety performance during the selection and definition phases may result in failure to attain the maximum safety potential. the demonstration of ALARP is issue specific.The cost of Safety performance by design phase By not robustly addressing the explosion hazard early in the phase of a project. it is recommended that the following general philosophy be applied for addressing explosion hazards on risk offshore facility: • • • Establish and apply nominal overpressures if available. or if it were to be attained it would be at the expense of significant cost and schedule penalties. As described above. October 2003 . Attainable Safety Performance Maximum Attainable Safety Performance Level Executed Successful Safety Input During Early Phases Define Select Define Select Concept Execute Unsuccessful Safety Input During Early Phases Cost (CAPEX & OPEX) Figure 3 . Use this Guidance (Clause 3) to manage the explosion hazard during the FEED and detailed engineering phases For high and medium risk installations. This may not be apparent until late in the project. This can result in a measure that becomes grossly disproportionate to the risk reduction achievable.
• • • • Maintenance Inspection Change control Hot work procedures Operation Decommissioning • Implementation of decommissioning procedures • Implementation of disposal procedures Issue 1. firewall blast wall location • Determination of constructability • Element specific or low level performance standards set • Safety Critical Element (SCE) categorisation – identification of high criticality items • • • • • • • Design for overpressures. piping.1 Summary of Life-cycle Factors Design Stage Concept selection and definition Design Factors Affecting Explosion Hazard/Risk • Offshore processing content (amount of equipment and potential leak and ignition sources) • Offshore structure type • Location of living quarters (may be carried over to FEED) • Equipment layout • Decks plated or grated • Operation and manning philosophy (exposure of personnel to explosion risk) • Philosophy for engineering. location and schedule Supports for safety critical elements determined Control systems designed Verification of constructability Assembly/writing of maintenance and inspection procedures FEED Detailed design Construction • Small bore piping runs located • Changes to ensure constructability • Competence of construction • Assembly of Decommissioning procedures and assurance of integrity during decommissioning. dynamic pressures Finer points of layout Firewater and vent piping. October 2003 35 . etc. • Deck sizing • Nominal explosion loading • Fire area sizing.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions The Table below summarizes the factors affecting explosion risk by project phase Table 3.
oxygen depletion and burns from flames and hot gases. and strong shock response occupancy of area immediately affected vulnerability of people in adjacent areas the relative location of the TR Issue 1. loss of TR integrity by failure of the boundary partitions. The major risk lies in effects such as being blown over by the blast wind. The human physiology can withstand relatively high overpressures. Likelihood is a more appropriate term in this context where a qualitative assessment is being performed. For the installation under consideration the direct and indirect effects of explosion overpressure should be identified. being struck by missiles picked up by the blast. such as:• • • • fires resulting from loss of inventory from damaged equipment.1 Risk screening General The higher the life safety risk (or risk to life) on an installation or within a compartment/module the greater should be the rigor that is employed to understand and reduce that risk.4 3. It is important therefore to have a means of early estimation of the risk level of an installation to determine the appropriate approach to be used in installation explosion assessment. Where the risk associated with an outcome is low. Where quantitative values are not available a qualitative measure of risk can be estimated to a degree of accuracy sufficient to make a decision on the assessment approach to be adopted.4.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 3. In human terms the direct consequences of an explosion are potentially considerably less than those which may occur as an indirect result of the explosion. The effort expended should be proportional to the risk. This risk can be calculated as a numerical value expressed as Individual Risk (IR) or in terms of a value for the installation such as Potential Loss of Life (PLL).2 Consequence severity The consequence side of the matrix comprises an assessment of the effects of credible explosion scenarios including escalation. any inaccuracies in determining that risk will also be low in absolute terms. It is more likely that the major consequences will involve escalation. structural failure. This should be achieved by assessing parameters such as. missiles. the terms probability and frequency imply that numerical values are available. inaccessibility of means of escape and/or evacuation.4. Risk is the product of consequence and frequency of occurrence. The approach to explosion assessment needs to be decided early in the design process when absolute values for explosion frequency and detailed consequence analysis are not available. • • • • the vulnerability of Safety Critical Elements to dynamic pressure. pipework and vessels. A simple approach which is frequently adopted for qualitative risk analysis uses a 3 x 3 matrix of potential consequence and likelihood of an explosion event  is described in this subclause. overpressure. 3. October 2003 36 .
typified by the amount of equipment installed. Alternatively. jack-up or FPSO) with quarters on the same structure as the process.g. e. Manning would be consistent with a normally unattended installation with a low attendance frequency. influencing manning levels and occupation frequency the operating and control philosophy influencing extent of operator intervention and the potential for human error and inventory loss. hub type flanges. A high consequence installation would encompass remaining installations and compartments where there is significant processing on board leading to significant congestion and potential confinement with populated areas within the consequence range of escalation scenarios. Congestion. Manning would be consistent with a normally unattended installation with a moderate attendance frequency. a medium consequence installation may be a processing platform necessitating permanent manning but with low escalation potential to quarters. valves or pipework. semi-sub. valves.e. The following parameters will influence the potential likelihood of an explosion: • • • • hazardous inventory complexity. will be greater than for the low case. i. less frequent than 6-weekly. being limited to wellheads and manifold with no vessels (i. October 2003 37 . Some generic types of flange tend to have lower leak frequencies associated with them.4. no associated process pipework) resulting in low congestion and inventory. it is recommended that the category with next higher consequence is used. A medium consequence installation would be typically a platform or compartment where the congestion and confinement exceeds that defined for the low consequence case but with still a low manning level consistent with a normally unattended installation. This may typically be characterised by a PDUQ/PUQ installation (jacket. Issue 1. utilities and control areas which are located on a separate structure. The equipment count would probably be low. with no more than 2 solid boundaries including solid decks. Where there is doubt regarding the category into which an installation should fall. Confinement should also be low. the ventilation regime.e. 3. more frequent than 6-weekly. The following gives some guidance on the assessment of consequences: Low consequence outcomes would be predicted where the overpressure level is predicted to be relatively low and immediate and delayed consequences are also low. compressors and other potential gas leak sources. A 6-weekly visit by a maintenance/intervention crew results in occupancy of a little over 1 %. the type of flanges. both isolatable and non-isolatable the operating and control philosophy.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions • • • • • levels of congestion and confinement the suitability of the layout the dimensions of potential explosive gas clouds hazardous inventories. the number of ignition sources within the potential gas cloud. the number of flanges.3 Likelihood The likelihood of an explosion will depend upon the likelihood of occurrence of a gas cloud and delayed ignition.
Medium event likelihood is suggested by an NUI with equipment count greater than for the ‘low’ case.4 The Risk Matrix The risk category for the installation is assigned using the risk matrix below: Consequence Low Frequency/ Likelihood High Medium Low Medium risk Low risk Low risk Medium High risk Medium risk Low risk High High risk High risk Medium risk 3. Low event likelihood installations and compartments will have a low equipment count.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions • the equipment reliability and the maintenance philosophy. it is necessary to identify the initiating events which may lead to an explosion. 3. control and mitigation will become necessary. where the planned frequency of maintenance/intervention is greater than a 6weekly basis then this suggests a higher or less reliable class of equipment with medium level of potential for explosion. The frequency intervention of 6 weeks or more is also recommended as a criterion as this will be a surrogate for equipment count and reliability as well as a measure of maintenance risk with respect to explosion. 38 Issue 1. Similarly. it is recommended the category with next higher likelihood is used. An example methodology for the definition of scenarios is shown in the flow diagram of Figure 4. Explosion scenarios must also be considered from the point of view of the fire hazard.1 Hazard identification and scenarios General In order to manage the explosion hazard. Large installations will have more potential leak and ignition sources and therefore a greater requirement for intervention and maintenance. Fire and explosion interaction is discussed in Clause 4.5. Where the complexity of the process in a compartment requires a permanently manned installation this suggests a high equipment level and therefore potentially a high likelihood of an explosion event. The likelihood considerations tend to align closely with the consequence factors in that the low consequence installations will tend to be small and therefore less complex.4. a large number of potential leak sources and high ignition potential.5 3. October 2003 . Where there is doubt regarding the category into which an installation should fall. These initiating events should be eliminated if possible otherwise prevention.
Hazards may also be identified by the individual in the course of his work. or sample/drain line. vibration. ship impact.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 3.2 gives an historical overview of the sources of releases for the UK sector of the North Sea from Reference . fatigue. tracked and eventually closed out. reports and project concerns so that they can be formally held in a single data-base. extreme environmental conditions. It should be the aim of all projects to identify the major safety and environmental hazards as early as possible in the project so that they can be assessed and acted upon as necessary in a timely manner. Pump Seal or Gasket Leak: Often represented as a 5 mm orifice or the annular space between the flanges without a gasket or between the pump shaft and the caseless seal. in order to obtain as broad an understanding as possible of potential hazards. it is necessary therefore to undertake the necessary exercises to ensure they can be identified. Additionally. or similar title. intervention. October 2003 39 . and how each is handled. Small Fitting or Line: Often represented as a 20 mm orifice or the typically installed diameter for an instrument connection.3 Causes Causes of a release may be dropped objects. including operational personnel. A log should be kept of all actions arising from hazard identification studies. and is generally managed by the Safety Group within the project. Sub-clause 2. A key element in the management of hazards is the establishment of an action tracking system. This is called the Safety & Environmental Action Monitoring System.5. imperfections.2 Identification It is not possible to manage hazards without understanding them. their cause.5. These causes are often referred to as initiating events. 2. exceedance of design conditions and human error. 3. 3. Generic release scenarios based on historical evidence in the Gulf of Mexico sector of the North Sea  are listed below starting with the most credible: and the UK 1. The means by which hazards are identified may be by formal processes such as: • • • • • hazard identification studies (HAZIDs) and environmental hazard identification studies (ENVID) hazard and operability reviews (HAZOPs) layout review hazardous area review safety studies Formal reviews should be attended by all relevant disciplines. it may aid the management of hazards to compile a hazard register listing all significant hazards. Issue 1. escalation from a fire. Medium Line or Partial Large Line: Often represented by a 50 mm orifice and evaluated as a possible credible release scenarios especially when considering dropped objects. thus ensuring that none is forgotten or ignored.
5. 5. A hierarchy should be used when considering suitable means of risk reduction. use control systems to warn personnel and to minimise the explosion event. Injury may then result from contact with hard objects or exceptionally by being thrown overboard. then its functionality is assumed to be impaired. Evaluated usually for off-facility or facility separation distance determination and emergency response planning purposes.5..5 Hazard & Risk Management Where significant risk is identified. In the first instance. or other explosion effect such as severe vibration. As far as explosion impact on humans is concerned there are seven primary hazards. • if the overpressure. These are direct overpressure effects. including explosion. With respect to consequences. the environment and the asset. • • • 3. the impact of objects on to personnel and the inhalation of hot products of combustion It is probable that personnel around the periphery of the module. removal of protective clothing. the impact of blast wind on personnel. oxygen depletion. October 2003 . Rarely considered as credible for design although occasionally appears in loss databases usually as a result of inappropriate vessel materials selection (e. mechanical impact as well as incomplete inspection. the aim should be to eliminate the hazard entirely. that is identified an evaluation should be made of the potential consequences and the risk to personnel.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 4. exceeds the quoted survivability criterion in the element specific performance standard for the SCE. minimise the intensity of the explosion event. Quantification of the event frequency should be undertaken as detailed elsewhere in this Guidance using recognized QRA techniques. where gases vent. It is suggested  that at 20m/s most people would be knocked over.g. If explosion events cannot be eliminated the aim should be to: • • • • remove personnel from the consequences of the explosion or escalating events. Large Transfer Line & Vessel Nozzle Failure: Full pipe diameter. systems shall be put in place where practicable for its reduction. hydrogen embrittlement. chloride/caustic stress corrosion cracking). for example. In QRA it is often conservatively assumed that anyone within the same module as the blast becomes a fatality. burns from the flame front. a number of rule sets can be compiled. Rarely considered as credible for design although frequently appears in loss databases. It is difficult to quantify the risk from blast wind and inhalation of hot gases. improper relief sizing or relief plugging. 3. 40 Issue 1. install mitigation measures to protect the workforce.4 Evaluation For each hazard. will be more susceptible to blast wind than static overpressure. Vessel Failure: Some suggestions by failure mode propagation or vessel de-inventory within 10 minutes.
6 3. October 2003 41 .1 Detection. Control implies the detection of the initiating events and actions initiated after the detection to reduce the hazard or exposure to it. For the explosion hazard. systems will have a role in: • • • • detection and initiation of control measures avoidance of the explosion event reduction of people on board (POB) exposure reduction of the severity of the explosion hazard Issue 1.6.2 Management Hierarchy Level of Hierarchy Eliminate risk Typical Management Options • • • • • No leak sources No ignition source No congestion No confinement Ensure inherent strength to withstand the event Remote operation No offshore processing Split facilities (separate quarters structure) Provide and maintain emergency escape and rescue (EER) facilities Reduce congestion Reduce confinement Increase ventilation Reduce hazardous inventory Reduce inventory pressure Gas/leak detection Alarm to warn personnel Deluge initiation ESD and blowdown Electrical isolation Forced ventilation control Remove personnel from the • consequences • • • Minimise inherent strength of • explosion event • • • • Use of control systems • • • • • • Mitigation for the workforce • Blast Walls.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Table 3. however there will usually be residual hazards and some risk that needs to be managed. Using the safety management hierarchy these residual risks should be approached firstly by controlling the risk and then by mitigation. TR. The employment of inherently safe features is an important part of any new design or assessment of existing installations. control and mitigation Introduction Aspects of inherent safety have been discussed in Clause 2. means of escape and evacuation and individual 3.
This will include any area where hazardous fluid is handled.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions • • minimization of the escalation potential protection of the SCEs Where choices exist. In locating the detection devices some knowledge of the dispersion of the accumulation is desirable.3.3 Summary of Control and Mitigation Options Role Control/Mitigation Measure Gas detection Acoustic leak detection Operator/Manual alarm call-point/Phone Emergency Shutdown (ESD) Detection and initiation of control • measures • • • Avoidance of the explosion event • Isolation of electrical equipment • Increase ventilation – start stand-by fans (in explosion zone) • Shutdown ventilation intakes (on detection in adjacent areas) • General alarm • Blast walls. October 2003 .) 42 Issue 1.5. (Deluge should not be initiated on gas detection in confined areas as the result may be to increase overpressure due to the added turbulence induced by the spray. TR.3 below. Measures that can be used to fulfil the control and mitigation roles are given in Table 3. Gas detection devices should be located where accumulations are likely to occur.2 Detection The aim should be to detect hazardous gas. so that personnel can escape from the local area and the potential direct consequences of the explosion. vapour or oil spray accumulations and to initiate control measures at the earliest opportunity. a fluid handled above its flash-point or within 20C of it. means of escape and evacuation Initiation of area deluge upon gas detection Increase ventilation – start stand-by fans Reduction of POB exposure Reduction of the severity of the • explosion hazard • Minimization of the escalation • potential • Protection of SCEs Isolation of hazardous inventories Blowdown/depressurisation • Blast walls • Blast walls/enclosures • Resilient mountings • Inherent robustness 3.6. Upon detection deluge should normally be initiated where congestion dominated explosions may occur when the danger from the creation of additional ignition sources is considered unimportant. Detection should result in an immediate alarm. Table 3. i. preference should be given to passive rather than active systems of control and mitigation as stated in sub-clause 2.e.
It is often used for alarm only. This will ensure the earliest possible detection of the leak by this means. control measures will include. Ultrasonic leak detection is another means of detecting a loss of containment. A critical feature to be taken into account during design is the choice of detection devices to provide rapid response and avoid spurious action.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Confirmed detection should give rise to executive actions to help limit the potential for an explosion to occur or limit the consequences. High and Low Levels may be combined to a single action level.g. but this depends upon the potential for spurious initiation. generally the executive action level. should be established at the lowest level that detector can reliably be set without spurious tripping. They also provide a rapid response which is important where deluge is used as a means of overpressure reduction.6. If the operator loses faith in the system installed there is a risk that either the outputs will be disabled or that it will not be correctly acted upon when a hazardous event actually occurs.3 Control Systems Control includes actions initiated after the detection to reduce the hazard or exposure to it. This will involve Emergency Shutdown ESD at an appropriate level. October 2003 43 . generally the alarm level. initiation of blowdown and isolation of non-essential electrical supplies. • • • • • • • alarm to personnel Emergency Shut Down (ESD) blowdown/depressurisation electrical isolation shutdown of fans for detection in external areas start of standby fans for detection in enclosed areas deluge initiation Issue 1. In hazardous gas detection. e. Increased confidence in these types of detection is likely to lead further to their use in automatic initiation of executive actions. The latest infra-red hazardous gas detection devices are reliable and if correctly installed should not give rise to spurious trips. 3. This can be of significant benefit in the early detection of leaks. The High Level Gas action point. If the potential for spurious/faulty action is low then it should be considered for executive action especially in the initiation of deluge when acoustic detection may be expected to give a more rapid response to a leak that conventional gas detectors. It is then recommended that this level is set as low as practicable consistent with the avoidance of spurious trips. should be set at the lowest level whereby it is considered reasonable to instigate executive actions. All electrical equipment remaining in operation within the explosion zone should be rated for operation in a Zone 1 hazardous area ]. Infra-red devices may also be beneficial as they too do not need to be in the cloud to detect a gas release. ‘low level gas’ and ‘high level gas’. two levels of action are usually specified. This type of detector picks up the high frequency noise emitted by a sonic leak. For the explosion hazard. leaking pneumatic systems. such as including process shutdown. this does not rely on the device being within the gas cloud to be tripped. The Low Level Gas action point.
It is expected that the installation will have a hierarchy of shutdown levels depending on the degree of risk posed by the hazard identified. The Probability of Failure on Demand (PFD) of the system must be compatible with the SIL to ensure the appropriate level of reliability of operation. The following figure indicates the actions that may be taken upon gas detection. process shutdown and isolation of hazardous inventories should occur.4 Shutdown Philosophy Upon confirmed gas or leak detection. Detection of hazardous gas at the executive action level would be likely to initiate a high level of shutdown. Low level Gas Detection High level Gas Detection •Initiate general alarm •Start firewater pumps •Open deluge valve •Re-Alarm •Initiate ESDV at appropriate level Ventilated Area (hazardous) •Start all available HVAC ventilation fans Ventilated Area (non hazardous) •Stop ventilation fans •Close HVAC dampers ESD •Isolate non-essential electrical equipment •Close ESDVs •Open EDPDs •Trip mechanical plant Figure 4 . Further actions at the appropriate level of shutdown may also be initiated.Actions that may be taken upon gas detection 3.6. 44 Issue 1.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions • other systems such as inert gas flooding Control systems will aim to minimise the magnitude of the explosion event and prevent or minimise the potential for escalation once it has occurred. October 2003 . the executive action level should initiate a level of emergency shutdown appropriate to the risk of explosion and fire. In order to minimise the size of an explosive gas cloud that may accumulate and to minimise potential escalation. The control loops from initiation to output should be subject to assessment of the Safety Integrity Level (SIL) of the system  .
Upon gas detection. The inventory will then tend to drop to the lower pressure through the downstream control valve. Artificial ventilation may be either induced or forced. API 521 is now known to recommend blowdown rates which are too low for offshore use. October 2003 45 . In general the API standard requirements  should be regarded as a minimal rather than an optimal and are not related to the explosion hazard. Rooms containing hazardous gas inventories are generally operated at negative pressure compared with adjacent non-hazardous areas to prevent leakage of potentially explosive gas clouds to those adjacent areas. all fans should be stopped and the intake and discharge isolated by (nominally) gas tight dampers. In order to remove the inventory from the explosion zone more quickly. Blowdown of relevant inventories should commence at once or as soon as it is safe to do so.6 Artificial Ventilation Artificial ventilation is defined as that ventilation which is not supplied from the action of the environmental wind alone.5 ESD and Blowdown At the executive (High Level) action point ESD should be initiated to minimise the potential inventory that can be lost to the environment. The aim for explosion mitigation should be to depressurize the process as rapidly as reasonably practicable. balancing the gains from the reduced period of flammable gas cloud duration and reduced escalation consequences against size and cost of the vent system and the impact of thermal radiation from the flare or ignited vent. The use of standard API blowdown rates  is intended for the mitigation of fire escalation only.6. Downstream isolation valves and actuators should be outside the explosion zone or protected from the potential effects of blast so that closure after an explosion event is still possible. it may be feasible to close the upstream isolation valve and leave the downstream valves open if there is a significant pressure drop in the process. Non-hazardous rooms adjacent or close to hazardous areas are normally operated at a positive pressure to prevent ingress of explosive gas clouds. Upon detection of flammable gas. the standby fan(s) should be started to give maximum possible ventilation in order to aid dilution of the leak to prevent or limit the generation of an explosive cloud. Closure of the downstream valve should occur after a delay built into the shutdown software or initiated by a pressure sensing device as long as this will remain functional after the explosion and any foreseeable escalation.6. only a relatively small amount of gas is required to develop a large explosive cloud. The level at which this executive action is initiated should be chosen based on the specification of the detection device and the avoidance of spurious shutdowns. the room is then under negative pressure compared with areas around it. Forced ventilation means that air is forced into the room by fans in the intake ducting resulting in the ventilated space being at positive pressure relative to adjacent areas. for a typical methane rich hydrocarbon release with a Lower Explosive Limit (LEL) of 5 % volume concentration with 9 % as the stoichiometric volume concentration level. Full isolation from downstream inventories should however be completed as the pressure equalises. which would normally occur at the fresh air intake. these rates are under review even for this use. Unless blowdown is rapid it will tend to have a small impact on the flammable cloud size generated since. 3. Induced ventilation means that the air is drawn into the space by fans located on the ‘extract’ side of the room. Issue 1.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 3.
limit ventilation and mix an inhomogeneous cloud. This acts on the droplet to break it up and give a greater overall surface area so that it more efficiently achieves the quenching effect on the combustion mechanism. cover a number of such fire zones. and immediately adjacent zones.6. The usual general area coverage rate of 10 l/min/m2 . Firewater pumps are sized to deluge the zone in which fire detection takes place. In enclosed modules deluge will be ineffective in lowering overpressures and may even result in an increase due to the turbulence caused by the water spray. falls outside the rate investigated in the JIP tests . however. Fears that the deluge would provide an ignition source through electrostatic effects have proved unfounded  although there is a residual doubt about the ignition potential due to charge accumulation on isolated conductors e. there is not sufficient kinetic energy in the flame to break up the deluge droplets in order for them to be effective. even with 15 % added for hydraulic imbalance. water curtains may be used to arrest the flame velocity and prevent runaway flame acceleration. Greater flows up to about 20 l/min/m2 will have a marginally increased benefit. If explosion mitigation is considered critical a deluge flow-rate of at least 13-15 l/min/m2 is recommended for general area coverage. These effects could give rise to higher overpressures. Equipment specific protection only applies deluge to small discrete areas and this is not sufficient to give any meaningful control of overpressure. For these applications and where equipment specific deluge only is installed. Congestion generated explosions are characterised by a fast moving flame front. such as FPSOs. October 2003 . This means that there is insufficient water supply to cover the whole flammable cloud. This will have a significant effect on lowering peak overpressures. Deluge would then be applied to the fire zone where initial gas detection takes place with water curtains supplied simultaneously from a separate deluge valve. Where the overpressure is generated by confinement. Deluge from standard MV or HV nozzles has been found to be suitable for reducing overpressure in congestion generated explosions. On large open decks firewalls are not generally practicable so the deck is divided into fire zones where fire events may be expected to be confined. the process area is generally divided into fire zones in order to limit the size of the firewater pumps. these may be discrete process blocks or pre-assembled units. 46 Issue 1.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 3. for example in completely enclosed modules. scaffolding. Large gas clouds can. It is important that the deluge is employed as area coverage rather than equipment specific protection. The mechanism by which this mitigation occurs is by inhibition of the combustion process and by cooling of the products of combustion as the deluge water is converted to vapour. A double curtain of MV nozzles supplying water at a rate of about 15 l/min/m2 for the area covered by the curtain and spaced at 10m intervals should be effective in arresting runaway flame acceleration and prevent excessive overpressures. For large open deck areas. The initiation of deluge and water curtains on a quiescent gas cloud may induce turbulence in the cloud.7 Deluge The Phase 3a Advantica Spadeadam tests  confirmed that deluge can significantly reduce high explosion overpressures. The benefits of deluge have now been incorporated into the major CFD and some phenomenological simulation software. Other devices such as fan sprays have also been tested and may give coverage more suited for curtain use.g.
It is essential however that control systems have a reliability that gives sufficient confidence that they will be effective. This is due to at least two factors. Issue 1.7.g.1 Control systems and Safety Critical Equipment Safety Integrity Levels Instrumented control systems rely the following generic loop to effect the control mechanism. e. • • • a means of detecting an upset condition. For any control system there will be a potential for failure to operate on demand due to the inherent unreliability of the elements involved. for the majority of low pressure events there is no benefit particularly if the coverage is incomplete These factors can be however estimated from a knowledge of the equipment and systems within the module. the reliability of the deluge system. from detector device through to deluge nozzle.2 Safety Critical Elements Safety Critical Elements (SCEs) are those items or systems that prevent or reduce the impact of major accident events. The procedure should be applied to each instrumented control loop used in the management of explosion hazards. The use and effectiveness of deluge as a fire and explosion mitigation measure is discussed in sub-clause 4. Guidance on the setting of SIL can be found in Reference . • • • the delay between leak initiation and deluge being applied to the module. a gas accumulation an electronic data processing system to process outputs from the detection system output signals to activate the controlling mechanisms. The level of reliability to function on demand should be commensurate with the level of risk being averted. 3.7.7 3. A major problem in determining risk levels for an installation with explosion mitigation by deluge is determining the level of risk reduction that can be assumed. The benefits of explosion mitigation by general area may deluge may only be considered only if a time dependent ignition probability is used and the deluge can be applied within the reliability constraints of the system.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Some CFD overpressure modelling programs can accommodate the effect of water curtains so an optimisation of frequency and coverage of curtains within the practicalities of firewater demand can be undertaken. 3. They are identified as part of the detailed design of an installation as required by the ‘Safety Case Regulations’ . October 2003 47 .7. This can be assured by determining the loop’s Safety Integrity Level and designing the elements of the loop to meet this level of reliability.
October 2003 48 . Reliability or Availability. A design accidental event is one for which SCEs on the installation should perform their function as designed.g. SCE 001 SCE 002 SCE 003 SCE 004 SCE 005a SCE 005b SCE 005c SCE 006 SCE 007 SCE 008 SCE 009 SCE 010 SCE 011 SCE 012 SCE 013 SCE 014 SCE 015 SCE 016 SCE 017 SCE 018 SCE 019 Safety Critical Element Hydrocarbon Containment -Pipelines.4: Table 3. e. element specific performance standards defining how the SCEs will operate under normal and extreme conditions. apart from the crane/lifting appliances. Extreme events against which design is not reasonably practicable may result in loss of these systems. The element specific performance standard defines the item’s Functionality. The performance of each SCE is defined in element specific performance standards which remain with the SCE throughout its life and against which its performance is assessed. should remain functional during and after an explosion event. Survivability includes its endurance under explosion loads. The design of the system must therefore match its stated survivability so that its functionality is maintained in an explosion event. catastrophically collapse. and the remaining items should not fail so as to give rise to a hazard. A typical listing of SCEs would include the items given in Table 3.g. firewater pipework Hydrocarbon Containment -Topsides Process Facilities Hydrocarbon Containment -Wells Fire & Gas Detection System Riser Shutdown System Topsides Shutdown System Wellhead Shutdown System Ignition Prevention Platform Sub-Structure Topsides Structure Uninterrupted Power Supply Emergency Lighting Evacuation & Escape Systems Rescue & Recovery Telecommunications Navigational Aids Personal Protective Equipment Helideck Escape Routes Temporary Refuge Platform Crane / Lifting Appliances Of the above SCEs all. No. Vent lines. explosions. Survivability and some measure of interaction with other safety systems.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions In order to demonstrate that these systems remain functional and fulfil their duties. • overpressure Issue 1. Risers. SCEs that mitigate the effects of a major accident will have to remain functional after initiation of design accidental events. The specific explosion effects which it may need to withstand are. e.4 Typical Safety Critical Elements Ref.
UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions
• • • • dynamic pressure displacement effects strong vibration exposure to missiles
Equipment specific performance standards
The term Performance Standard has become interpreted as having a very specific meaning when related to the DCR which amends the SCR. One of the aspects of DCR is the identification of Safety-Critical Elements and the preparation of a Written Scheme of Verification for assurance of continued integrity. A Performance Standard can be described as a series of statements which can be expressed in qualitative or quantitative terms the performance required of an SCE can be compliant with DCR. The equipment specific performance standards or low level performance standards define the requirements in terms of functionality, availability and survivability. The Performance Standard is a means of condensing the Design Specification and Operating Procedures into a requirement related to a particular system in a particular mode of operation in the face of an identified hazard. The justification for the requirement and the degree to which it is achieved is contained within the Written Scheme of Verification or Examination.
Levels of criticality of equipment items
The DCR states: “Any structure, plant, equipment, system (including computer software) or component part whose failure could cause or contribute substantially to a major accident is safety-critical, as is any which is intended to prevent or limit the effect of a major accident.” So for our purposes, the following equipment and systems at least need to be addressed when considering the response to explosions: • • • • • Those necessary for the safe shut down of the installation. Those necessary for personnel protection and escape. Those necessary for fire detection, suppression and control. Those necessary for communications. Those necessary for hydrocarbon processing, transport and storage.
Safety-Critical Equipment may require a blast protection structure to enable satisfactory operation within the full range of design explosion events. Protection against fire is a further consideration, both before and after the effects of an explosion; such considerations are covered in Clause 7. A practicable limiting explosion needs to be identified in the specification of these equipment specific performance standards, such that the risks to personnel, the environment and the commercial viability of the asset are as low as reasonably practicable (ALARP). This Ductility Level Blast (DLB) is an achievable design event with a realistic frequency of occurrence. It will be paramount in this Guidance Document to achieve consistent terminology to identify those events with: • • The maximum credible peak overpressure, but very low frequency of occurrence; The practicable limiting overpressure for design purposes, giving ALARP risk levels;
Issue 1, October 2003
UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions
• Overpressures giving elastic response levels in components, for early design and robustness checks.
With systems such as gas detection, the ability to withstand a major explosion may not be warranted, so long as the presence of the gas build up is identified. The element specific performance standards thus need to be realistic and fit-for-purpose, taken on a element by element basis. The performance standards need to take into account initial design and manufacturing cost, practicability of construction, maintenance, accessibility for inspection, reliability, robustness and redundancy. To this end OTO 1999 046 and 1999 047 similar to the one given below: Criticality 1 Criticality 2 Criticality 3
, detailed the use of a criticality rating system
Equipment and pipes designed to withstand the same explosion severity as the main structure. Equipment and pipes designed for a less severe but non-the-less substantial explosion event. Equipment not designed or assessed for explosion.
The difference between Criticality 1 and 2 items would be associated with the consequence of fire escalation and escape from the facility. It was recognised in the paper that the assessment of Criticality 1 items would be very much constrained by project schedules and costs under current design project philosophies. The key question though must be whether current design practice delivers equipment and their supports which, if designed primarily for in-service loading, provide safe and fit-for-purpose details when subjected to extreme explosion events. Based on the limited amount of strengthening or modification to existing equipment required for operators to delivery satisfactory Safety Cases upon their most recent submission, it must be presumed that past designs are reasonably robust and a significant increase in design effort would not therefore be warranted or necessary. Criticality 1 Items whose failure would lead direct impairment of the TR or emergency escape and rescue (EER) systems including the associated supporting structure. Performance standard – These items must not fail during the DLB or SLB, ductile response of the support structure is allowed during the DLB. Criticality 2 Items whose failure could lead to major hydrocarbon release and escalation affecting more than one module or compartment. Performance standard – These items must have no functional significance in an explosion event and these items and their supports must respond elastically under the strength level blast (SLB)
Items whose failure in an explosion may result in module wide escalation, with potential for inventories outside the module contributing to a fire due to blowdown and or pipework damage. Performance standard – These items have no functional significance in an explosion event and must not become or generate projectiles.
Issue 1, October 2003
UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions
The Strength Level Blast (SLB) and the Ductility Level Blast (DLB) are defined in the glossary, Appendix B and sub-clause 3.6.
Mitigation and consequence minimization
The severity of explosion consequences may be mitigated by the introduction of:• • • • • • • • Lower inventory pressures More ventilation Judicious ignition source location Smaller potential explosion zones Optimised layout resulting in less congestion and less confinement The use of blast relief panels Robust design of SCEs and minimisation of escalation potential Segregation of explosion release sources from congestion and major escalation targets.
Flammable cloud size is determined by the leak dimension and the pressure of the inventory. The mass rate at which gas is released from a hole is directly proportional to pressure. Reduced inventory pressure will reduce explosive cloud dimensions and the severity of the explosion event. A reduction in pressure will also result in a lower inventory mass within the system which will give the potential for more rapid blowdown and reduced escalation consequence
Assuming that some loss of inventory from a hazardous process will occur at some time, ventilation is critical in ensuring that a significant flammable gas cloud does not form or that the cloud is reduced in size, particularly for small leaks which are the most frequent. Industry guidance  defines ‘adequate ventilation’ as ventilation achieving at least 12 air changes per hour for at least 95 % of the time, with no dead spots. The term ‘adequate ventilation’ is not meant to imply an optimum level of leak dilution; it should be regarded as minimum target. Increased ventilation rates will ensure dilution of larger hazardous leaks and reduce the potential for explosions to occur or reduce its severity. In the case of natural ventilation adequate ventilation needs to be confirmed. For small wellhead platforms and NUIs this can generally be demonstrated qualitatively since confinement is low. For more complex installations the only way of demonstrating adequate ventilation will be either by CFD modelling of the ventilation regime, or wind tunnel testing. Where there is inadequate ventilation the options are to re-orientate the installation to make more use of prevailing winds, to remove confinement, to relocate equipment to provide a freer air flow through the module and to add some fan assistance at dead spots.
Issue 1, October 2003
UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions
A special case, but one which typifies that various considerations that arise in the ventilation of a hazardous area, is that of turbine enclosure ventilation. The means of managing this risk is detailed in the HSE guidance. This relies on fast dilution of the leak at source to prevent significant flammable clouds forming.
If a module is open on all four sides, orientation of the installation is relatively unimportant. If, as is often the case, one side forms a solid partition then orientation will affect the ventilation air change rate within the module. The installation is often aligned such that the TR is directly upwind of the process areas. This minimises potential for smoke logging or toxic gas ingress into the TR. Apart from smoke logging and natural ventilation, operational factors also enter into the orientation assessment as supply vessels and helicopter approaches tend to be towards the flow of current or wind respectively. The preliminary QRA is likely to give an indication as to the relative risks of fire, toxic gas or explosion. It is probable that explosion risk (including escalation from an explosion event) will be well below that from process fires and that operational constraints of supply vessels, at least on the larger installations, will tend to set the orientation, with natural ventilation being a secondary consideration. FPSOs are unique in that they orientate themselves according to the relative forces of the current, wave and wind, called weather-vaning. Frequently the vessel will be positioned with the wind blowing lengthways from bow to stern. This is not the preferred arrangement from an explosion point of view as a gas cloud will extend along the deck within congested areas rather than being blow out to the sides. With fan assisted ventilation, for instance in a closed module, it is relatively easy to show that ‘adequate ventilation’ has been provided. The provision of 12 air changes per hour as an absolute minimum  with a qualitative demonstration that there are no dead spots by the suitable location of supply grilles should suffice.
Ignition Source Location
Highest overpressures in congested modules tend to arise when the ignition point is at the furthest point from a main vent. There is potential for ignition to occur at practically any point within the module (e.g. from hot work activities), however removing other ignition sources away from such extremities will, to some extent, lower the potential for high explosion overpressures to occur.
The layout of a module should be optimised in order to; • • • • • Minimise confinement Minimise congestion Minimise flame front path length Reduce potential gas cloud size Reduce unfavourable ignition source locations
Issue 1, October 2003
A balance needs to be sought in reaching an optimum solution.8. its speed and ease of use. as opposed to a long narrow plan. Any blockage of these will result. For this explosion type. will reduce the final overpressure. pipework located near vent areas will be subjected to drag forces proportional to the square of the exit velocity. Since smaller items tend to give rise to the greater level of turbulence critical for overpressure generation a suitable measure of overpressure generation potential may be the mean diameter of equipment.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Generally. Additionally. the largest overpressures for a given configuration tend to arise when the ignition point is furthest from a main vent. rather its ability to identify trends. This will however tend to increase confinement and reduce the effect of natural ventilation. This also means that for a given deck area the minimum potential overpressure occurs for a square plan area. Phenomenological models may suffice for this purpose. October 2003 53 . 3. vessel may have little effect on overpressure but a number of small bore pipes of similar volume will have a significant effect since the turbulence generated will be greater. These measures are in need of further investigation. layouts that result in maximum potential for venting and minimise the potential for turbulent flow generation. The volume blockage ratio is defined as the ratio of the volume occupied by the obstacles to the total volume. The intention is to identify trends in overpressure. To judge the degree of congestion in a given arrangement the volume blockage divided by mean diameter may give a measure of congestion. it is not then the accuracy of the model in predicting overpressure values that is required. The area around vents should be kept as clear as possible. Careful layout at an early stage in a design project by orientating major vessels parallel to the expected gas flow during an explosion and avoiding the blockage of vents. 3. This does not however take into account the amount of equipment present. The Volume blockage ratio is a suitable measure of this. particularly in the early stages of flame travel. can reduce the peak overpressures by a considerable amount. This issue will be pursued further in Part 3 of the Guidance The effect on overpressure of the above recommendations can be variable and in some cases conflicting. Keeping the dimensions of an explosion zone small will limit the potential for overpressure generation. A large. To achieve this for a large complex process suggests that there will need to be a number of partitions to limit the size of each zone.8. It is necessary to optimise the overpressure management techniques by reference to a calculation model to confirm the downward trend.8 Congestion minimisation and layout Congestion is not necessarily related to blockage ratio. but also in the velocity of the gases being vented. Issue 1. It is therefore preferable to locate pipework away from vent areas so that they do not block the vent path but also because the forces exerted on them may be difficult to accommodate. not absolute values. correctly aligned.7 Reduction in Explosion Zone Size For a congestion dominated explosion event the strength of explosion is dependent on the flame path length from a main vent or from a target. not only in increase in overpressure. more the type of blockage. Installations tend to have escape routes around the periphery of modules so that minimisation of explosion consequences at vents will have less impact on escaping personnel and prevent damage that may affect escape routes.
October 2003 . Increase Transverse spacing Figure 5 . 54 Issue 1. Poor Effect Better Reduce volume Safe area Reduce Blockage ratio And number Of obstructions Move obstacles to Inner part of module Sideways venting Reduce Blockage ratio.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Figure 5 shows comparisons of good and bad layout options from the point of view of explosion overpressure. • • • Effective venting reduces the magnitude of the peak overpressure When gas is allowed to expand freely the resulting overpressure is much lower than when the gas is confined For the same blockage ratio several small obstacles create a higher overpressure than a smaller number of big obstacles.Relative merits of layouts The following observations may also be noted for the mitigation of the effects of explosions.
5 kg/m2 depending on the method of attachment.9 Blast Relief Panels Blast relief panels. 3. 3. Equipment on anti-vibration mounts may be displaced if the mount topples. These requirements will probably result in panels of mass/unit area less than 0. escape routes. Sharp cornered objects generate higher peak pressures than rounded obstacles. It is unlikely that loosening of cladding panels will have the desired effect. otherwise they can be free of restraint.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions • • • Blow-out panels take a finite time to move out of the way of the expanding gas and may not prevent further overpressure increases. • • • • static overpressure dynamic overpressure displacement effects strong shock (strong vibration) effects Heat effects tend to be transitory and the thermal inertia of most items should ensure that thermal effects of the explosion event are negligible. Normal louvres create sufficient obstruction to cause increased overpressures. which open quickly during an explosion in order to reduce peak overpressures. Issue 1. the TR and evacuation systems). The specific explosion effects which an item may need to withstand are.8. If this is considered possible restrained mountings should be specified or sea fixings retained during operation. A typical blast relief panel will have the following properties: • • • • It will be light – probably of aluminium construction It will start to open at about 50 millibar or 0.05 bar (wind loads are usually a factor of ten lower) It must open quickly (within about 50 milliseconds) and stay open It must be located to open a clear vent path to prevent further flame acceleration occurring due to venting through a congested area.10 Damage to Safety Critical Elements and Escalation Potential Most SCEs should be designed to operate after the explosion event (e. These issues are discussed further . The smaller safety critical elements can either have enhanced support to withstand blast wind effects or be located near a bulkhead or structure where the wind effect will be low. The intention should be that these items survive an explosion event and maintain their functionality to prevent escalation.  and . must be carefully designed. hinged) if there is significant risk posed by them becoming missiles. October 2003 55 .8.g.g. Panels may be retained (e. This may be appropriate for fire and gas detection devices.
10’s of centimetres. by avoiding arrangements where items are attached to floor and ceiling.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Flare/vent and firewater piping can be especially susceptible to blast forces since they are generally of low pressure design and therefore light-weight.11 Strong shock response In an explosion event the forces acting on the structure will result in movement of decks and members. pipes that run from floor to ceiling. These phenomena may occur in a wave as the overpressure moves through the module in a congestion dominated explosion. bookcases. sufficient resilience should be included in the design to accommodate the differential movement. slack in cables where displacement of control or electrical equipment is possible. October 2003 . this involves. The result of this displacement is that initially. 58]. that is. Means of protecting these lines will include and or all of the following: • • • • not running these lines near vent areas where blast wind forces will be high increasing the schedule of the lines to increase inherent robustness increasing pipework supports running these lines in the shelter of structural elements Detailed guidance on the methods of designing for static and dynamic overpressure is given in Clauses 6 and 7. • • • • the location of vibration sensitive safety critical elements distant from potential explosion zones. so that forces potentially acting upon items spanning structural elements can be large. 3. Good design practice however should be employed to mitigate the consequences of this severe vibration. If this is so. As with all potential hazards the aim should be to avoid the situation occurring.g. e. especially not attaching them to common boundaries. 56 Issue 1. These displacements will have significant impacts on any items fixed to structural elements that are moving differentially to each other. • • • roofs/ceilings will move upwards decks will move downwards walls will move outwards Subsequent to this will be the rebound effects as the structure then oscillates. the use of shock or resilient mountings to absorb vibration effects. This may be achieved using resilient mounts or by the building in of flexibility to the piping runs. cabinets. Damage to firewater lines would result in the loss of fire control measures within the explosion zone. This displacement absorbs energy plastic by deformation and if it did not occur the structure could not survive [57. Relative displacements can be considerable. Damage to flare/vent lines can result in large quantities of gas entering the module. This would lead to jet fires and possibly secondary explosions.8. This may not however be reasonably practicable. or run from a floor mounted vessel to the ceiling. especially as the installation will probably be being blown down at the time. additional restraints to items that may topple and cause damage or lose function. for example.
The blast winds can reach sonic speeds. items that are not fixed to the structure or substantial equipment items. Care should be taken in providing such systems as they can pose a significant asphyxiation risk to personnel and they should be discounted if the overall risk of using them is greater than the explosion/fire risk saved. This is most appropriate for enclosed volumes where the environment can be controlled. October 2003 57 . The explosive gas can then be taken below its lower explosive limit. The philosophy for minimising the potential for missile generation is to ensure general tidiness.13. • • • • inert gas barriers soft barriers micromist 3. and being knocked over.13 3. especially in the hydrocarbon processing areas and areas adjacent to the TR. Inergen (CO2 + N2 + Ar) is an alternative inerting material but it has a volume 30 times that of Halon systems (now banned) which it may replace. Their advice should be sought before any mountings are fitted since. The natural frequency of the mount must not coincide (within a factor of 2) with the natural oscillation frequency of the loaded floor. are available from companies marketing anti-vibration equipment.13.13.8. therefore there is potential to move objects of a significant size.12 Missile Prevention Secondary missiles are those picked up by the blast wind produced during an explosion.8.2 Inert Gas Inert gas can be used to dilute the flammable mixture by flooding the volume within which the gas has been detected. 3. Small items such as fire extinguishers and items of safety equipment should be kept in cabinets firmly fixed to the structure preferably against a partition or structural member where blast wind will have little effect. In most modules this is where escape routes are located so these are the areas where personnel are in most danger from impact from missiles. shock mounts. This might particularly be the case for generally manned areas. N2.8. 3.3 Barriers Barriers can be used for five purposes. There should be no loose items in the module. The structural design group should be able to estimate the approximate natural frequency of the structure at the mounting point so that the correct mount can be specified. Issue 1.8. i. 3. • protecting areas behind the barrier from the overpressure effects generated in the explosion zone. Maintenance equipment such as scaffold poles and tools should not be stored in process areas and should be removed when no longer required.1 Explosion Mitigation Systems General Other mitigation systems that have been proposed include.e. enhancement of the vibration effect can occur.8.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Specialist products. Maximum blast wind speeds are found around the edges of the module where explosion gases are venting. if the wrong product is installed. with for example CO2.
These walls are designed to absorb blast energy by displacement.4 Soft Barriers Progress is being made in the manufacture of the Micro-mist device which has been tested as part of a recent joint industry project and consists of a cylinder of superheated water which is released quickly as a fine mist in response to pressure. The method reviewed here uses logarithmic frequency and severity ratings which are added together to give a risk rating. by how much and what is the residual risk level). B15 partition) on the non-hazardous side which will move sympathetically with the blast wall.13. 58 Issue 1.g. One simple approach to defining risk  and hence identifying whether risk is tolerable is the use of risk matrices. shock sensitive equipment should not be located along this wall unless it can be demonstrated that the effects will be negligible. As long as confinement is controlled decreasing the size of a module by inserting barriers can result in a reduction in maximum flame path length and therefore a reduction in the overpressure generated.8. is likelihood or outcome reduced. There may be a partition wall (e. October 2003 . It can be related to specific quantitative targets as is the case in some legislative regimes. 3. Protection from the direct effects of blast can be achieved by the barrier method. limiting cloud spread to ignition sources limiting the number of people directly exposed Blast walls have long been used to protect adjacent areas from the effects of overpressure. flame sensors during an explosion.e. This device suppresses the explosion and has been shown to be successful in significantly reducing overpressures. Accidental activation of these devices is not a hazard to people. that is. These come in a wide variety of forms but provide a simple and effective means for design teams to assess the likelihood (probability of and event) and the severity (consequence). By limiting the size of an explosive gas cloud the overpressure can be reduced.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions • • • limiting the size of the explosion zone and thus the largest dimension over which the flame can propagate. That is. 3. It is important that these barriers maintain their integrity and do not give rise to their own hazards by displacement and impacting on safety critical plant or by becoming missiles. Mitigation measures aim to protect personnel or equipment from the explosion event. Generic definitions for likelihood and consequence can be easily established. If the wall forms the boundary to a control area housing safety critical equipment the effects of this movement needs to be taken into account. Tolerable can be defined in many ways.9 Acceptance criteria The objective of a fire and blast hazard management process is to reduce risks associated with potential hazards to a level that is ALARP. the intervention of a blast wall between the explosion zone and the area to be protected. These are added rather than multiplied because they are logarithmic representations of the actual frequency and severity. the air gap between the two acting as a spring. it can be related in part to cost (risks being reduced to a level that do not incur disproportionate costs) and an array of other criteria defined by legislation and/or corporate goals as part of internal safety management systems. This enables the risks to be semi-quantitatively defined (positioned) and offers a mechanism for mitigating measures to be evaluated (i.
may be ranked using a coarse system based on the frequency of the cause.000 yrs Less than once every 10. ignition times and ignition location. Table 3. October 2003 59 .5 Risk Acceptance Matrix – Frequency against Severity Frequency Frequent Severe Severity Critical Substantial Marginal Negligible A A A B B Occasional A A B B C Infrequent A B B C C Unlikely B B C C C Rare B C C C C FREQUENCY RATING (FR) or Likelihood: Each generalised hazard scenario. Issue 1. taking into account existing controls and including releases in all directions under all environmental conditions.000 yrs Once every 1.000 to 10. Risk is set at three levels by the matrix: A .6 Frequency Rating (FR) Criteria Category Frequent Occasional Infrequent Unlikely Rare Annual Probability of Occurrence (/yr) > 10 -1 Frequency ‘Score’ 0 -1 -2 -3 -4 Rating 5 4 3 2 1 More than once every 10 yrs Once every 10 to 100 yrs Once every 100 to 1. This scenario specific frequency will not be representative of the general frequency of a late ignited release giving rise to a significant overpressure and will grossly underestimate the frequency rating and hence the risk.000 yrs 10-1 – 10-2 10-2 – 10-3 10 – 10 < 10 -4 -3 -4 The frequency ‘score’ is effectively the logarithm of the annual probability of occurrence.risk is tolerable with controls – evaluate additional controls/design changes C .UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Risk rating = Frequency rating + severity rating A simple example of the use of risk acceptance matrices is presented here.risk is not normally tolerable – additional controls/design changes required B . It is necessary to consider a generalised frequency in this calculation as the frequency associated with a particular release direction under particular ventilation (wind) conditions is likely to be very small. The frequency rating is a normalized representation of the frequency ‘score’ for use in the semiquantitative hazard assessment.risk is tolerable. The frequency rating may be categorised using the following as a guide: Table 3.
October 2003 . revised to suit this. It is essential that a checking procedure is instigated so that the work can be verified by a competent person. the variability from ignition time and position under a range of environmental conditions may be ranked using a coarse system based on severity of the cause. This should occur for all work produced.7 Severity Rating (SR) Criteria Category Severity (safety. or the procedure. In many cases it will be the efforts of a single person that will result in the quantification of the explosion hazard and resulting risk. Table 3. If practices are not in line with the written procedures then a review should be undertaken to establish which indeed is the correct practice and either the method of working. asset) Large scale loss of life Large scale environmental impact Large scale loss of asset Loss of life of several persons Extensive environmental impact Major loss of asset Loss of single life or serious injury to several persons Significant environmental impact Significant loss of asset Severity 'Score' Rating +2 5 Severe Critical +1 4 Substantial 0 3 Marginal Single serious injury or minor injuries to several persons -1 Minor environmental impact Minor loss of asset Single minor injury Little environmental impact Little loss of asset -2 2 Negligible 1 The severity rating is a normalized representation of the severity score for use in the semiquantitative hazard assessment. taking into account existing controls.10 Implementation and monitoring Guidance on design activities is given in the procedures and design guides which should be readily available for consultation by the design team. Activities should be undertaken in line with these procedures. The assessment must consider the risk profile from all causes to satisfy the legislative requirements in particular relating to TR impairment. environment.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions SEVERITY RATING (SR) or Consequence: Each release scenario. Severity is assessed using the following guide. 60 Issue 1. 3.
11. The nomination of a typical installation to represent a family of platforms is acceptable. Alternatively. However the premise is that the variables listed would be more closely analysed than might be the case for a low explosion risk installations or compartments. substantial physical differences between this and previous cases. 3. Issue 1. This involves the checking of work by independent personnel against the procedures for that activity. Discrepancies may involve amendment of the design or the procedure to bring about compatibility. 3. it may involve the scrutiny of a proportion of the design output by persons independent of the project audit. It can involve additional review by an independent engineer or by presentation of the overall design (including explosion) to a group of independent engineers. and no extrapolation of data allowed unless published data exists.3 Medium Explosion Risk Installations Where valid nominal overpressures are available or past cases exist that are relevant. 3. Extrapolation of data for the relevant parameters is not generally recommended but may be valid if a sound basis exists. They may be used as the ductility level blast overpressures (DLB).UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Other means of monitoring and verifying the design include peer review. piping. Such comparisons should be supported by evidence that a structured assessment has been undertaken to identify areas of difference and that the original means of calculation were sound. This may be achieved in a number of ways. the resolution and precision of the grids used in the calculation.11. A comparative assessment method may be used drawing on experience from a demonstrably similar structure geometry and scenarios. the version of the explosion modelling software used. and other SCEs. that sufficient and competent resources have been allocated and all risks are being managed correctly.1 Analysis methods General The risk category of the installation does not preclude the use of more sophisticated methods of assessment which may result in reductions in conservatism and hence cost. October 2003 61 . A suitable low sophistication means of defining the explosion hazard is the use of valid nominal overpressures derived from previous assessments of similar structures. This applies not only to the definition of the explosion hazard but also to methodologies in handling the response of structures. these values may be employed.11.2 Low Explosion Risk Installations Where the explosion risk category for the installation is low. The comparison process would incorporate consideration of the following factors: • • • • the validity of the model used in the initial assessment.11 3. It will also aim to check that activities are carried out according to the project plan. the low risk methodology may be used. if they are considered more appropriate. Another acceptable means of overpressure derivation for low risk installations is comparison with a specific past cases.
The philosophy recommended in this Guidance is that for medium risk installations the choice of methodology for any particular task must be justified where it deviates from the higher risk methodology. Table 3. Dynamic possibly non-linear modelling of the installation and systems response.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Where no nominal overpressure has been defined for the installation and there are no suitable past cases for comparison. These checklists are taken from Reference  the Genesis ‘Explosion assessment guidelines’. Determination and assessment of the structure and SCEs against the SLB and DLB design explosion loads including blast wind dynamic pressures.6. This level of analysis would involve. zonal models. this will warrant a commensurately high level of analysis. Medium risk methods are described in this Guidance which substitute for some of the tasks defined in the Higher risk methodology.7 of this document. The treatment of uncertainties is discussed briefly in Subclause 2.3 of the Main Guidance and Subclause 5. 62 Issue 1. then the level of analysis appropriate for high risk installations should be used for explosion hazard assessment. Appendix C gives by way of checklists the issues which should be considered for low medium and high risk installations. • • • • • • • • A complete set of explosion scenario investigations CFD simulations of gas dispersion.4 High Explosion Risk Installations Where the potential risk level on an installation or within a compartment is high. 3. A combination of CFD and phenomenological explosion simulation with generation of exceedance curves representing frequency of overpressure exceedance.11. Determination of equivalent stoichiometric cloud size  . the Shell DICE model or the workbook approach  if it can be justified.8 indicates appropriate methods of analysis by risk category for the installation. October 2003 . Explicit consideration of escalation and interaction between fire and explosion scenarios including the collapse of tall structures and the external explosion. The ability of the installation and the safety critical systems on it to withstand explosion need to be accurately determined as any error could have a significant risk impact. Consideration of strong shock and missile generation by the explosion.
additional activities2 This task determines the risk level of the installation Medium Risk1 3. characteristics design credible event Consider a number of Consider external explosions Consider and justify ignition source use of previous or similar assessments positions by use of design basis checks. Installation screening 2.8 . values not available. October 2003 63 . control and mitigation primarily by reduction of the frequency of an ignited release then by limitation of immediate and escalation consequences Categorise SCE’s determine criticality 1 and 2 SCE’s Consider representative leak Consider design Consider design directions.assessment complete if satisfactory. credible event. Preventcontrol-mitigate 4. overpressure peak and durations. Use appropriate nominal appropriate nominal overpressures and overpressures and durations durations. Use credible event. dispersion handbook wind speeds and directions.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Table 3. and duration – DLB only Consider intermittent ignition Calculate dynamic Calculate dynamic sources and time history of overpressures on overpressures on cloud geometry giving worst Criticality 1 & 2 Criticality 1 SCE’s.Appropriate assessment methods Task 1. Explosion load determination (continued) Consider worst case Consider a number of Consider ingress of gas from scenarios apply release scenarios neighbouring modules and practical limiting with their associated dispersion/ventilation factors to arrive at frequencies. ignition time at maximum SCE’s. Determine Calculate ignition probability Determine overpressure peaks time history. Calculate release rates and durations All risk levels should consider the options of hazard elimination. Issue 1. overpressures for DLB Criticality 1 and SLB if values not available. Explosion load determination 4. Calculate cloud evolution Calculate equivalent Calculate equivalent using CFD dispersion stoichiometric cloud cloud sizes from simulation for representative size. approach. Use 1/3 of Use 1/3 local peak equivalent stoichiometric local peak overpressure if cloud size. Scenario definition Low risk High risk .
Check deflection and rotation limits for DLB. Dynamic analysis required. Use full load-time histories. Check SCE’s barriers and connections to DLB. Assess integrity of structure and other SCE’s (continued) Non-load bearing barriers and cladding may be checked for DLB using simplified methods (Biggs). (Simplified 1 point method acceptable). Non-load bearing barriers. Check primary and secondary structure response to DLB and SLB loads using modified code checks.additional activities2 Calculate exceedance diagrams for overpressure and dynamic pressures. Assess integrity of structure and other SCE’s Identify and assess escalation targets. October 2003 . 5. cladding and their connections may be checked for DLB using simplified methods (Biggs) if ductilities less than 5 are expected. High risk . Determine SLB and DLB overpressures and dynamic pressures from exceedance curves.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Task Low risk Medium Risk1 Calculate exceedance diagrams for overpressure and dynamic pressures. Idealise load time histories nontriangular idealisations possible. Check primary and secondary structure response to DLB and SLB loads using modified code checks. Equivalent static loads allowed if justified. Check integrity of penetrations. Non-load bearing barriers. Idealise load time histories nontriangular idealisations possible. 64 Issue 1. If required check primary and secondary structure to DLB using non-linear dynamic analysis. Check deflection and rotation limits for DLB. cladding and their connections and supporting structure may be checked for DLB using simplified methods (Biggs) if ductilities less than 5 are expected. If required check primary and secondary structure to DLB using nonlinear dynamic analysis. 5. Otherwise use NLFEA. Check primary and secondary structure response to DLB using modified code checks. Determine SLB and DLB overpressures and dynamic pressures. Check deflection and rotation limits for DLB. blast wave effects Identify and assess escalation targets. Equivalent static loads allowed if justified. Determine far field.
Determine if ALARP has been satisfied. High risk activities2 - additional 6. every three years when the Safety case is required to be updated (triennial submission). Determine if the SCE’s satisfy element specific performance standards. The assessment of existing structures differs from the assessment of a structure during design in three important respects.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Task Low risk Medium Risk1 Check SCE’s criticality 1 to DLB and SLB using dynamic pressure loads. all the tasks from low and medium risk methodologies should in principle be considered in addition to those listed. Verify that the installation satisfies high level performance standards.12. October 2003 65 . The philosophy recommended in this Guidance is that for medium risk installations the choice of methodology for any particular task must be justified where it deviates from the higher risk methodology. Determine if the SCE’s satisfy element specific performance standards. Issue 1. Evaluation Verify that life safety risk is acceptable. Verify that the installation satisfies high level performance standards. Determine if the SCE’s satisfy element specific performance standards. Medium risk methods which substitute for some of the tasks defined in the Higher risk methodology. Check SCE’s criticality 2 to SLB using dynamic pressure loads. For high risk installations. Determine if ALARP has been satisfied.12 3. Verify that life safety risk is acceptable. 3.1 Existing installations General In the UK sector of the North Sea. Even if an installation has not been modified or its use has not been changed an assessment may be required to take account of advances in methodology. it is a requirement (SCR)  that significant changes in an installation or its operation will require the Safety Case to be resubmitted which should address a review of hazards including a re-will require an assessment including those due to explosion hazards. Notes 1. 2. Existing mobile installations entering UK waters will also require assessment. Determine if ALARP has been satisfied. Verify that the installation satisfies high level performance standards. Verify that life safety risk is acceptable.
The HSE reference  states “It should be borne in mind that reducing the risks from an existing plant ALARP may still result in a level of residual risk which is higher than that which would be achieved by reducing risks to ALARP in a similar. For existing installations the ‘Explosion Hazard Review’ is and additional activity. Intervention may give rise to an additional hazard which must be assessed. then the work to be undertaken should not in itself pose such hazards and risk to personnel that this compromises the gains to be achieved by such modifications. what may be demanded by reducing risks (to) ALARP for a new plant (and what may have become good practice for every new plant). The checklists described previously and located in Appendix C. The tasks to be performed are listed below with references to the sub-clauses where a detailed treatment is given. the original CAD and structural computer models of the facility. are applicable to the assessment of existing structures. 3.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 1. Information should be available from the Safety Case. Information may be available relating to expected explosion loads. the extra cost of retrofitting measures compared to designing them on the new plant. irrespective of cost. 2. assessment and other controls as determined by the Safety Management system as well as method statements for their implementation. If risks are in this intolerable region then risk reduction measures must be implemented.” The overall individual risk and the TR impairment frequency (TRIF) from all hazards must still be less than 10-3 per year. 66 Issue 1. All temporary structures and equipment used during the modification work should be removed as soon as the work is complete. All modification work should be accompanied by hazard identification. the risks involved in installation of the retrofitted measure (which must be weighed against the benefits it provides after installation) and the projected lifetime of the existing plant. The computer data files and design reports should be checked to confirm that they are a faithful representation of the present state of the facility. Use may be made of experience gained from the operation of the un-modified installation and from similar installations. There is relatively little scope for the reduction of the frequency of a release and scope for mitigation of the severity of an explosion may be limited. for example. Factors which could lead to this difference include the practicality of retrofitting a measure on an existing plant. that it is not reasonably practicable to apply retrospectively to existing plant. October 2003 . An explosion hazard review may be necessary in the case of an existing installation. The Individual Risk (IR) per annum TRIF and PLL will have been used in the demonstration of ALARP in the existing Safety Case for the installation. All this may mean. new plant. Should modifications be necessary to improve the safety performance of the facility. structural and equipment response from the detailed design stage of the design and construction project.
The general philosophy should be to bring all critical SCEs up to a similar level of integrity for all accidental events. There may also be a problem accessing design data which should include modifications that have taken place since the facility was installed.12. e. the facility should be reviewed with regard to the level of inherent safety that exists and what additional control mechanisms there may be for overpressure reduction. Ideally. The implementation of effective change procedures is essential in these circumstances. These critical systems include but may not be limited to: • • • • the TR escape routes to the TR evacuation systems systems that could threaten the integrity of the above systems. there should be no disproportionate contribution from any one hazard to the risk associated with the installation. poses a number of difficulties. These problems generally relate to the difficulty in retrofitting structure and equipment in modules that are operational and congested. with the usual constraints in working in a cramped offshore environment.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 3.g. October 2003 67 . This is referred to as the ‘balanced risk contribution principle’. Before investigating the ability of the systems on the installation to withstand the blast effects. Issue 1. For an existing installation there may be potential for: • • • • • • • • • • • • increased venting additional and/or improved gas detection (acoustic gas detection) introduction of flame detection lowering the level at which gas detection initiates executive actions voting for executive action on a single detector initiation of deluge on gas detection the removal of redundant equipment the relocation of equipment blocking vent paths vibration reduction enhanced robustness of small bore connections use of past operational experience improved inspection/maintenance regimes Some of these actions may not necessarily result in a reduction of calculated Individual Risk or loss of life (as shown in the QRA) but this should not be a reason for failing to undertake modifications if actual reduction in explosion risk can be foreseen.2 Explosion hazard review Review of explosion risk for existing installations. The existing wells may be operating at a lower reservoir pressure than originally designed for. however subsequent tie-ins may have offset this possible benefit. ESDVs isolating offsite inventories. For example fire and explosion hazards should contribute to the total risk at levels comparable with those for a similar installation.
then recalculation and assessment will be appropriate. it is frequently the case that the integrity of systems is greater than originally assumed. If risks are in the intolerable region then risk reduction measures must be implemented. some safety critical elements may be in criticality levels 1. Simpler phenomenological models or comparison with similar installations would be adequate for the purposes of quantifying the residual risk. Where this is high for explosion events the premise should be that greater cost is justified in providing mitigation and protection than where the risk posed is relatively low. The calculation of overpressure should however be reasonably straightforward if the original CAD model is available and has been updated to include changes made since the design stage. The overall individual risk from all hazards must be less than this value. October 2003 . Where the original design took account of explosion overpressure. The experience of the JIP full scale tests [9. there would seem to be little to be gained in determining to the last degree of accuracy what the actual overpressure characteristics will be. 50] tended to confirm this in that damage incurred was significantly less than might have been expected for the high overpressures experienced.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions In effect the list may include all the identified safety critical elements but. irrespective of cost. Where it is clear from preliminary inspection that it will be impracticable to meet maximum overpressures that may arise. (Note: Schedules 1-3 of the SCR  imply that some form of overpressure calculation will be necessary in order to quantify the risks involved in mustering within the TR. 2 or 3. 3. The need to determine potential for TR damage would then seem to be necessary.3 Use of previous analyses A review will need to be undertaken of the ability of safety critical systems to withstand overpressure/drag effects.10.) The aim should be to bring the design up to the same level of integrity for all major SCEs with the presumption that more effort should be made where the level of risk from explosions is high. Whilst increased knowledge of the explosion phenomenon has resulted in a general increase in the overpressure loadings that are calculated for a given configuration. Where a significant ‘cliff edge’ occurs this may indicate that design beyond the base of the cliff edge is not reasonably practicable and that design to this level is ALARP. A means of determining the cut off point for additional mitigation would be to determine the costs involved with design enhancement to achieve integrity for successively increased overpressure values. 68 Issue 1. Existing installations will have QRA data from which explosion risk can be determined.12. with respect to the ability to muster. assess and evacuate. The accepted level above which the overall risk is considered intolerable relates to an individual risk of greater than 10-3 per year or a TR impairment frequency of greater than 10-3 per year. Hence the risk from other hazards may indirectly affect the acceptability of risk from explosions and these may need to be considered in setting the target risk levels for the explosion hazard. but latest knowledge indicates that this needs to be reviewed.
The early ignition of a release will normally result in a fire. This clause is a link between Part 1 of the Guidance and Part 2 which deals with the avoidance and mitigation of fires. In fact 80% of ignited releases have been observed to result in fires. The approach is to assess the fire risk.1. Where any conflict exists between explosion and fire management it is the latter which will tend to take priority. The severity of many of these events may be limited by immediate and rapid. This could give rise to the accumulation of a vapour or gas cloud leading to an explosion on ignition or a secondary fire near a source of oxygen. The release of inventory may be sudden as in vessel rupture giving rise to a BLEVE (Boiling Liquid Expanding Vapour Explosion) or a secondary vapour cloud explosion. assess the explosion risk and manage accordingly. Some jet fires may become unstable either by flowing too fast for the flame to remain within the envelope of the release or through ventilation limitation leading to self-extinguishing of the flame. Fires may also cause escalation when in contact with vessels and piping through further release of inventory that is subsequently ignited. A blast wall may deform or bow in direct contact with flame but this should not be a problem for a wall designed to deform in a ductile way.1. Further information may be found in the basis document for this Clause .UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 4 4. Combined blast and firewalls are available. Blast walls will generally have good fire performance in isolating heat sources from their targets. active reduction of inventory (blow-down) or by segmentation of the inventory into isolated sections following gas. The consequences of combined fire and explosion scenarios with either component occurring first should be considered in the structural design or assessment.2 Generic scenarios combined fires and explosions A fire that becomes ventilation controlled may give rise to a cloud of unburnt fuel. however the optimum solution will generally be a balance between the two. Issue 1. Their performance as fire walls is not usually adequate from the point of view of limiting the temperature of the cold face. October 2003 69 . and highlights areas where interaction and conflict exists between these and aspects of fire management. which on contact with the oxygen rich atmosphere outside a compartment may lead to a secondary explosion. The late ignition of a hydrocarbon release will allow a flammable gas or vapour cloud to develop and could give rise to an explosion.1 INTERACTION WITH FIRE HAZARD MANAGEMENT Overview fire and explosion hazard management General This clause reviews the explosion management tools highlighted in other clauses. the design can usually be adjusted to accommodate this requirement. near the vent. subsequent release of inventory and ignition. The most common scenario is when an explosion gives rise to secondary fires due to vessel or piping rupture. Columns of the primary structure may be significantly weakened by this effect. 4. heat or fire detection.1 4. The object of this clause is to aid assessment of this balance.
These are thus common requirements. 4. Corresponding to the ‘fire triangle’ is the ‘explosion hexagon’.2. This device suppresses the explosion and has been shown to be successful in significantly reducing overpressures. this tends to give preference to concepts where. oxygen and ignition but also includes pre-mixing.1 Common areas Leaks and ignition Explosions.1 Concept Selection The aim at this stage is to select a concept where risk to personnel and the environment is inherently low. fuel.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 4.2. The latter includes fuel. Water curtains and drop down blast walls have been used. jet and pool fires are various forms of the same fire phenomenon. Conflict with fire management is likely to occur in these latter areas. congestion and confinement.3 Considerations by design phase Management of fire and explosion continues through the lifetime of the installation from Concept Selection until Decommissioning.1. 4. Water curtains will reduce the radiation from a fire and can be used to protect escape-ways. flame sensors during an explosion. Far field effects can be experienced by propagation of the sonic blast wave away from the explosion zone which may exceed the far field impacts from fire. All management techniques and inherent safety features proposed for eliminating or minimising leaks or ignition potential are therefore valid for both fire and explosion management. 70 Issue 1. • • the number of potential leak sources and ignition points is low number of persons exposed to major hazards is low. 4. However the general goal to remove personnel from processing facilities will benefit risk from both fire and explosion. October 2003 . Being common factors all management methods employed to eliminate any or all of these features will be effective with respect to fire and explosion. depending on pressure and leak size. It is necessary to ensure that escalation potential is low. air and ignition are needed (the fire triangle).3.2 Removal of personnel from consequences Removal of personnel from proximity to process areas will similarly benefit risk reduction for both fire and explosion events. Accidental activation of these devices is not a hazard to people. This sub-clause gives an overview of the considerations of inter-dependencies in the management of the two phenomena at various stages in this process. 4. To initiate a fire of any sort. Management of both fire and explosion risk at this stage is similar.2 4. Progress is being made in the manufacture of the Micro-mist device which has been tested as part of a recent joint industry project and consists of a cylinder of superheated water which is released quickly as a fine mist in response to pressure.3 Temporary and soft barriers It may be possible for a fire or explosion barrier to be present only when required.
4. The main interaction comes in the location and type of boundaries between adjacent areas. is that relating to smoke. • • • • • key philosophies installation orientation deck sizing general layout fire area definition and deluge/firewater pump sizing 4.3.3. This is not generally the optimum orientation for natural ventilation. They should aim to promote inherent safety features and the minimization of leak potential by minimization of connections and intrusive instrumentation. Is it better to allow smoke to vent from the immediate fire area to allow escape. or should it be contained to prevent migration to the TR or evacuation facilities.3. Congestion generated overpressures generally increase with distance travelled by the flame front. piping and instrument philosophies.2.2. Layouts that aim to achieve a minimisation of potential flame path length by deck sizing and aspect ratio should not significantly impact fire risk. Orientation of the installation to minimise smoke logging (and gas ingress) potential to the TR will tend to result in the TR being upwind of the process areas with respect to prevailing conditions. Due to cost constraints.3. FPSOs and single deck installations have large continuous equipment areas. Since smoke is generally the major cause of fatality in fire events it is preferable to make its passage towards the TR as difficult as possible but otherwise allowing it to disperse.4 Deck Layout Deck sizing is determined largely by the amount of equipment to be installed.3. 4. These should include the setting of safety goals and the methods of identifying hazards and controlling risk. All these features are consistent with the management of fire and explosion.2. 4.2. the tendency is to minimize topsides area and therefore jacket size.1 Front End Engineering Design (FEED) General Parameters defined at this stage of design include.2 Key Philosophies Key philosophies will include HS&E. It is generally the philosophy to contain fires within the immediate area of initiation to Issue 1.2 4.2.5 Confinement This is probably the main area of difference in management between fire and explosion hazards. As stated elsewhere in the guidance natural ventilation considerations generally are subservient to the smoke and gas ingress hazard concerns and operational factors when determining the orientation of installations with a TR. October 2003 71 . especially in deeper water areas.3 Installation Orientation and Natural Ventilation One significant fire consideration with respect to ventilation control.3.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 4. which generates confinement issues.
3 4. October 2003 72 . External flaming may occur which can impact escalation targets and transfer smoke nearer to safe areas such as the TR.3.3. Issue 1.3. It is important that. these valves can close if they have not already done so. • • • • • pipe routing passive fire protection specification ESD and F&G design severe vibration impact design of firewater and vent/flare Lines Firewater piping systems. vents and flare lines are generally fabricated from low pressure. 4. e. spindle and actuator are not distorted in the explosion event. The valve vendor should be able to advise the overpressure which the valve actuator can withstand without loss of functionality.g. Significant damage may then leave the installation exposed to fires that result from process damage following an explosion. This is achieved by the inclusion of fire barriers.3. the platform can be isolated from these inventories to minimise fire escalation potential. Explosion proof enclosures or blast walls may be installed to protect the valve system. Confinement can cause ventilation controlled fires which produce more smoke and carbon monoxide due to incomplete combustion. With respect to fire and explosion management this will include. although this is not always so. It is necessary to confirm that it will withstand the dynamic as well as static overpressure. For medium and large installations it is not usually practicable to size a firewater pump to deluge all areas simultaneously. Grating would normally be used to disperse gas clouds but this may lead to liquid fuel migrating to decks below.2 Isolation of Offsite Inventories Inventories beyond the topsides process. Apart from trying to limit the escalating potential of fires. It is generally the case when managing the explosion hazard that areas are kept unconfined. following any significant release. Reference  is a good starting point for the assessment of equipment against fire attack. Their relative fragility however leaves these systems vulnerable to explosion damage if exposed to high drag or relative displacement induced forces. It is important that. Riser ESD valves are installed for this purpose. are often significantly larger than that of the process itself.3. lightweight piping. 4. export lines and flowlines from remote wells. These barriers include fire rated decks as well as walls. A compromise is generally sought between the size/cost of the pumps and the practicalities of adding fire divisions. In order to ensure this it is important that the valve. low strength. for example in the use of partitions to minimize potential gas cloud volumes. after the explosion.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions minimize escalation potential and to discourage the movement of smoke to muster and evacuation areas. a major aim of this compartmentalisation is often to limit the fire to an area where the consequences can be mitigated by firewater.1 Detailed design General In detailed design the considerations are focused on the finer points of the installation and the full specification of control systems. the pumps (including stand-bys) would tend to be unreasonably large and expensive.
Fire control/ESD systems should not be placed on or close to boundaries and blast walls that will experience significant movement in a blast. By contrast all blowdown valves should generally be fail open. 4. blast and missile barriers would only be used to protect items of large or particularly hazardous inventory. If not. This will include blast walls which are designed to deflect and absorb explosion energy. firewater pumps will not be started automatically and active fire protection may be delayed or lost.e.4 Severe Vibration and Displacement Explosion induced vibration will propagate throughout the structure to some degree  .3. Outputs that do not fail safe will commonly fail ‘as is’ meaning that ESD valves do not close. Suitable slack should be also incorporated in critical cabling systems to accommodate predicted displacement. 4.3. These devices may also be damaged by the vibration itself. Differential displacement between modules and displacement of module boundaries (walls or deck) may place high loads on to firewater and vent/flare pipework. Severe vibration or blast wind may result in dislocation of instrumentation such as fire detection/CCTV devices such that it makes monitoring of subsequent fires uncertain. protection should be applied to the actuation system so that functionality is assured following an explosion event. if not. but they should be located where drag forces will have little impact. Flexibility should be built into these pipework systems to accommodate these movements such that active fire protection is not lost and vent/flare lines maintain functionality.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Generally.3. Similarly fire protection control measures may be inhibited by loss of signal due to cable failure following displacement forces. Each safety critical element should be reviewed with respect to severe vibration. They should however remain operable during and after an SLB event.3 Gas Ventilation Systems Detection systems may be considered to fall into Criticality class 2 as they will have fulfilled their function prior to the DLB event.1 Special issues relating to installation type Integrated Platforms For FPSOs and other single deck installations. October 2003 73 . fire.4. Resilient or special shock mountings may be specified as recommended by the specialist vendors to mitigate these effects. 4. In the Cullen report. Robust fixing of these instruments should prevent movement. ESD valves should be ‘fail closed’ ensuring that damage to the initiation mechanism results in closure. protection should be applied to the actuation system so that functionality is assured following an explosion event. i. it was reported that ‘a number of witnesses spoke of the severe vibration associated with the initial explosion. In the worst cases this vibration may be capable of damaging control systems. blowdown valves cannot be opened. acceleration forces. For fire area segregation walls will usually be impractical as flames can extend over the barrier and set off deluge in the Issue 1. People were thrown off chairs and knocked over’. These control systems are usually configured to ‘fail safe’ so that.4 4.3. for instance ESDVs close and blowdown is initiated.
Should high barriers be considered their impact on area confinement and natural ventilation should be assessed.plated deck to minimise ingress of gas into the interdeck area Fuel Gas area .4. October 2003 . 4. partitioning may have explosion benefits even though confinement is increased. This would then reduce fire area size and may provide some benefit in firewater pump sizing. 4. It is usually necessary to include as much free area in the process deck as possible by incorporating grating. barriers to limit firewater pump size/cost. The disadvantages of grating are:• • • process fires can impact the tank tops and tank deck piping through the grating.normally grated to maximise ventilation Utilities .4.4. 74 Issue 1. The potential for liquid spillage incidents to escalate to the cargo tanks via the main deck is reduced whilst maximising interdeck ventilation by the following choices for the process deck . Partitioning may be used to limit the maximum size of the flammable gas cloud and actually reduce the maximum overpressure experienced.4 Barriers A compromise has to be sought between increased overpressure due to confinement and: • • barriers to protect specific items of equipment and people. giving an environmental impact. On some integrated platforms.bunded skids are used for each package to confine liquid spillage Gas Compression . 4. • • • • Separation .grated to maximize ventilation. liquid process leaks fall through the grating and run off the tank tops and into the sea. This will permit vertical venting along the potential flame path and also promote natural ventilation to limit explosive cloud size. Removal or breaching of solid boundaries may extend hazardous areas beyond the original limits . process gas leaks can be directed through the grating to produce an explosive cloud in this narrow space. Extended hazardous areas may then encompass unrated equipment or impinge on enclosed non-hazardous areas where additional provisions may be necessary such as the installation of air locks at access points or the relocation of HVAC intakes. The horizontal distance is long and the confinement in the vertical directions is high.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions next area. These disadvantages must be weighed against the consequences of an explosion confined between plated decks.2 FPSO Raised Decks The area between the storage tank tops and the process deck on FPSOs is a particular arrangement where high overpressures can be generated.3 Existing Installations Removal of boundary walls or the addition of vents to mitigate explosion overpressure in previously enclosed modules can lead to additional fire risk or ignition frequency.
decreased congestion generated overpressure due to flammable cloud size reduction. cloud spread to ignition sources. Issue 1. 4. following on from the possibility of the introduction of new ignition sources the rate of area coverage required for optimum explosion mitigation. October 2003 75 .UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions • • • barriers to prevent environmental spills.5 Congestion Congestion which generates high explosion overpressures is generally that from the large number of small elements within the module.4.1 Barriers for fires and for explosions Fire Explosion Fire carries the highest risk to personnel. Protecting these lines by locating against structural elements would mitigate this effect. the delay between firewater initiation and its arrival. • • • • • • explosion damage to PFP the suitability of deluge for explosion mitigation. extinguishments of a jet fire may result in an explosion. Poor explosion management can lead to fire particularly smoke scenarios via escalation Barriers limit escalation by preventing spread Barriers give confinement and may increase of fires and smoke and protecting specific overpressure items of equipment and people. Firewater pipework can be a contributor to additional congestion and therefore higher overpressures. barriers to limit the extent of hazardous areas. Barriers between fire areas limit firewater Barriers will limit natural ventilation pump sizing Barriers can limit gas cloud size and cloud spread Removal of barriers on existing plant may impact hazardous area designation Barriers to prevent environmental impact 4. Table 4. On an offshore installation equipment is always in close proximity so fire escalation should be catered for in any case by ESD and depressurisation which will to some extent limit explosion hazard.5 Potential areas of interaction and conflict Potential areas of interaction and conflict between fire and explosion management are. the damage that explosions can do to lightweight firewater piping thus compromising its ability to fight subsequent fires. This will have little impact on fire management except that high levels of congestion would suggest a high potential for escalation in a fire event. this is critical for explosion management.
g. Most companies marketing intumescent systems have tested their products in blast conditions with the materials generally maintaining integrity in a blast of about 1. PFP rated for jet fires will tend to have a higher degree of mechanical integrity since they must withstand the erosion forces of the ignited jet. to attempt to extinguish liquid process fires. Deluge is not however suitable for mitigation of completely confined explosions and.2 Suitability of Deluge Tests have shown that area deluge coverage is suitable for explosion mitigation where high flame speeds are generated. since the area of influence is so limited. In an explosion the forces exerted on the structure and equipment in the blast zone can generate large deflections/distortions in the structural substrate which can result in cracking of the PFP. There exists some experimental evidence that low expansion foam.e.64]. Such data should be sought from prospective PFP suppliers to aid assessment of explosion risk. Cladding materials should be rigidly fixed as blast winds at high velocity potentially have the power to displace them exposing insulation material beneath which may not then be able to withstand jet fire impact. delamination or even loss of the material entirely from the surface. However.1 Use and effectiveness of deluge Firewater deluge Firewater deluge has historically been used to prevent process or structural failures by its cooling action or.6 Explosion damage to passive fire protection (PFP) Systems Passive systems provide a reliable means of fire management since they do not depend on any control systems for initiation. 76 Issue 1. Deluge is not an effective management tool when applied as specific equipment protection. rather than as an extinguishant since in some rule sets it has only been credited with about a 50% success rate in extinguishing hydrocarbon pool fires and very little in extinguishing hydrocarbon jet fires. being available at all times. explosion forces will place more extreme loads on PFP and its substrate [63.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions • the amount of firewater that can be delivered relative to potential gas cloud dimensions. Credit for deluge has been taken mainly for its cooling effect. due to the additional turbulence generated. However care needs to be taken in using this data as it is the deflection of the substrate which is likely to be more important than the headline overpressure figure. may actually increase overpressure.7 4. with foam. 4.7. e. These are probably pessimistic assumptions since recent work  has indicated some success of deluge in extinguishing or significantly reducing the impact of jet fires. For maximum benefit it should be provided as general area coverage. 4.7. to a process vessel. A review of the response of PFP to explosions is given in Reference  and .5 bar  . delivered to an area deluge systems may mitigate congestion (flame acceleration) dominated explosions. The large droplets in the deluge spray are broken up by the momentum of the flame to give droplet size suitable for quenching the flame mechanism. Deluge curtains can be useful in slowing flame acceleration in large modules to prevent flame speeds accelerating to extreme levels. 4. i. opening of the joints. October 2003 . in high congestion areas. The size of droplet from standard MV (medium velocity) and HV (high velocity) deluge nozzles will be suitable for both fire and explosion control.
This would then result in gas cloud accumulation where previously the gas had been burning.4 Firewater Extinguishing of Jet Fires Initiation of deluge on fire detection is general practice where firewater systems are installed. Recent research however has shown that deluge can successfully extinguish some jet fires. • • • • delay in detecting the gas cloud delay in starting the firewater pump driver delay pressurising the system (if no jockey pump is used) delay in flowing water to the nozzles.3 Rate of Deluge Coverage The minimum deluge rate for general process area coverage as recommended in ISO 13702  is 10 l/min/m2. Any subsequent explosion should however be attenuated by the firewater deluge generally will continue to run. The total delay may be in the region of 60 to 90 seconds. If ignition of the flammable cloud occurs within this time there will be no explosion mitigation by deluge. Special circumstances may occur where an installation is particularly susceptible to blast (even at the deluge attenuated level) and where fire scenarios are short lived due to inventory depletion with low escalation risk. If lighting seals are routinely replaced at the same time as expired light bulbs then the impact of this will be reduced or eliminated. however the increase in benefit between these values is slight. Issue 1. For gas fires the aim is not to put out the fire but to cool equipment and the structure to prevent escalation. the fear of extinguishing a jet fire should not result in any decision not to activate deluge since the initial jet fire is likely to hold the highest risk potential. especially as it does not need to be in the hazardous cloud to detect the release.7. The usual general area coverage rate of 10 l/min/m2 (even with 15% added for hydraulic imbalance) falls outside the rate tested. 4. 4. Full-scale tests undertaken as part of the Phase 2 JIP  investigated coverage in the range 13 to 21 l/min/m2. Unless special circumstances exist.5 Deluge Delay There will inevitably be a delay in the supply of deluge to the module.7. this is caused by. Acoustic leak detection is likely to give quicker detection. especially as hot surfaces will have been created by the initial fire which can then act as ignition points.7. October 2003 77 . For fires this delay can still be significant.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions There is always the possibility that the application of deluge may generate extra ignition sources if acting on unsealed electrical equipment and in particular lighting units. If explosion mitigation is considered critical a deluge flow rate of at least 13-15 l/min/m2 is recommended for general area coverage. Until deluge coverage occurs the module is susceptible to full overpressures. but a delayed supply will still provide considerable benefit especially for cooling where direct jet impingement has not occurred. This showed an increasing benefit of deluge in reducing overpressures as area coverage rate increases. The potential for explosion is then high. 4. Time delay can be minimised by: • rapid detection of the gas cloud.
by comparison of risk with and without deluge/firewater provisions. sparking due to water ingress into electrical equipment.8 Firewater on Small Installations Addition of firewater to a small installation to mitigate fire and explosion will raise maintenance requirements and visit frequency thus increasing the exposure of operators to hazards from the installation. Deluge would then be applied to the fire zone where initial gas detection takes place with water curtains supplied simultaneously. there will be significant risk of ignition from deluge interaction with electrical plant. the use of more than one deluge valve to minimise distance between the valve and the furthest nozzle. for a properly designed and maintained installation.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions • • • • reduction in the gas detection level at which deluge is initiated. For existing plant this may be a problem unless seals are routinely replaced at the same time as expired light bulbs. The question to be answered is whether the increased exposure to hazards from more frequent visits is compensated by the protection and mitigation effects of deluge both for fire and explosion and subsequent escalation. 4. such as the decks of FPSOs.6 Flammable Gas Cloud Dimensions An approach in dealing with large areas. is to size the firewater pumps to deluge any single fire area and a set of water curtains throughout the process areas. There is no evidence that. Studies  have shown that the static potential caused by deluge flow is negligible compared with that necessary to ignite a gas cloud.7.7.7 Deluge Initiation as an Ignition Source In the past fears have been expressed that deluge initiation upon detection of a gas cloud may increase the probability of that cloud being ignited. October 2003 . 4. This situation is unlikely to change unless explosion risk is high. This may be due to: • • static charge generated by the deluge water flow at the nozzle. firewater systems are frequently not installed because overall risk is increased due to the increased hazard exposure from additional maintenance burden. The water curtains may arrest flame front acceleration as it propagates through the process area and thus limit maximum overpressure  but they may inhibit through ventilation and encourage re-circulation and mixing. employing a pressurised ring main with jockey pump. 4. For NUIs. 78 Issue 1. minimising the distance from the deluge valve to the furthest nozzle (consistent with suitable location of the deluge valve for manual initiation). Electrical plant should however be designed with suitable ‘IP ingress rating’ to cater for deluge.7. Suitable arguments will need to be put forward to justify the chosen arrangement.
and Permit To Work controls. vibration) and that proper procedures are followed in undertaking any interventions. Deluge may extinguish fire and give rise to explosion – but subsequent overpressure then attenuated by deluge Deluge should remain functional to fight fires after explosion event 4. The aims are to prevent degredation of the hazardous inventory containment system and potential ignition sources.g.Deluge Considerations For Fire For Explosion Increase General Area Coverage flow to 15 at least l/min/m2 (10 l/min/m2 would still be expected to provide benefit) Specific equipment protection has negligible attenuation effect on overpressure Explosive gas cloud may cover a number of fire areas which cannot be deluged simultaneously Delay in receiving deluge will negate any potential benefit if ignition occurs before spray is applied. erosion.2 . maintenance. corrosion.8 Operational issues Management for fire and explosion risk in operation is similar. Change control is also critical in managing the impact of modifications over the life of the installations.7. Cannot rely on deluge to be active when ignition occurs (delay and reliability issues) Deluge pipework needs to be protected from explosion drag effects Deluge pipework needs to be protected from differential movement/displacement of decks and structures General Area Coverage of not less than 10 l/min/m2 Specific equipment protection can prevent escalation Firewater capacity generally linked to largest fire area Deluge protection at normal rates is not usually effective for jet fire impingement. threats to the systems are scrutinised (e.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 4. Issue 1. October 2003 79 . centred on inspection.9 Summary of deluge considerations Table 4.
If the rate of volume generation within the explosive region is greater than the ability of the vents to release this volume then the overpressure will continue to rise. on: 1. For an explosion to occur a gas cloud with a concentration between the Upper Flammability Limit (UFL) and Lower Flammability Limit (LFL) must be ignited.1. amongst other things. The cloud volume and concentration 3. Gas explosions can be defined as the combustion of a premixed gas cloud containing fuel and an oxidiser that can result in a rapid rise in pressure. These basis documents contain more detailed information on the topics addressed in this clause. Typical flame speeds range from 1-1000m/s and overpressures may reach values of several bars.1 Derivation of explosion loads Introduction to explosion load determination General This clause is based on the basis documents References . All of the above points from 1 to 5 can affect the explosion overpressures in this type of environment. October 2003 . The overpressures are not limited to the 8 bar maximum typical of completely confined explosions. filling a closed volume initially at atmospheric pressure. Ignition source type and location 4.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 5 5. The disturbance is subsonic relative to the un-burnt gas immediately ahead of the wave. It develops by feedback with the expansion flow. location) For stoichiometric hydrocarbon gas clouds. Gas explosions can occur in enclosed volumes such as industrial process equipment or pipes and in more open areas such as ventilated offshore modules or onshore process areas  . Gas explosions in more open environments can also lead to significant overpressures depending on the rate of combustion and the mode of flame propagation in the cloud. The gas or gas mixture present 2. This type of explosion is referred to as confined. combustion without heat loss will result in overpressures of close to 8 bar. shape. The overpressure caused by the explosion will depend. This pressure rise is mainly due to the temperature rise caused by the combustion process and is generally not dependant on the congestion within the volume. Two types of explosion can be identified depending on the flame propagation rate: • A deflagration is propagated by the conduction and diffusion of heat. 80 Issue 1.1 5. number. The confinement or venting surrounding the gas cloud 5. The congestion or obstacles within the cloud (size. A stoichiometric air/fuel mixture is such that it contains exactly the required amount of oxygen to completely consume the fuel.  and  and describes the process of derivation of explosion loads.
The combustion wave travels at supersonic velocity relative to the un-burnt gas immediately ahead of the flame. As the surrounding mixture flows past the obstacles within the gas cloud turbulence is created. As the flame front progresses outside the compartment. which may distort the pressure time history and result in double peaked pressure traces inside and outside the compartment. 5. This further increases the velocity and turbulence in the flow field ahead of the flame potentially leading to a strong positive feedback mechanism for flame acceleration and high explosion overpressures.11. In an environment typical of gas explosions on offshore platforms the gas cloud engulfs many obstacles including equipment. blast walls. equipment rooms etc. October 2003 81 . Most vapour cloud explosions offshore would fall into the category of deflagrations. There will typically be localised high regions of overpressure with lower values of average pressure acting on large components.2 Dynamic pressures and overpressures Gas explosions can generate both high overpressure and high-speed gas flows as a result of the gas combustion process. structure etc. Large components of the structure such as solid decks or walls experience loads due to the pressure differences on opposite sides of the structure. often referred to as congestion. Issue 1. Typical durations range from 50 to 200ms. This turbulence increases the flame surface area and the combustion rate. piping. Typically within an explosion there will be a strong variation of the spatial and temporal pressure distribution. The overpressure at a location within a gas explosion will typically rise to a peak value and then fall to a sub atmospheric value before returning to zero overpressure. The explosion typically starts as a slow laminar flame ignited by a weak ignition source such as a spark. 5.1. and in particularly the TR. Methods for calculating loads acting on intermediate sized objects such as large vessels are described in sub-clause 5. Further descriptions of the external explosion phenomenon are available Reference . an external explosion may occur.3 External explosions Secondary or external explosions may result as vented unburnt fuel/air mixture comes into contact with the external (oxygen rich) atmosphere. As the gas mixture burns. hot combustion products are created that expand to approximately the surrounding pressure.1. For this type of object the dynamic pressure associated with the gas flow in the explosion will dominate the applied loads. This may have consequences for the escape. The behaviour of the explosion will be influenced by both the degree of confinement and the congestion within the combustion region. For smaller objects such as piping the overpressures applied to the front and reverse side of such items will be of approximately the same magnitude at any moment in time and in this case the overpressure will not apply any net load to the object. The duration of the positive phase in an explosion can vary greatly with shorter durations associated with higher overpressure explosions. The shock wave and combustion wave are coupled and in a gas-air cloud the detonation wave will typically propagate at 1500-2000m/s and result in overpressures of 15-20bar. It is also usually confined by a number of solid planes such as plated decks.ways and may give rise to a blast wave which may impinge on neighbouring structures.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions • A detonation is propagated by a shock that compresses the flammable mixture to a state where it is beyond its auto-ignition temperature.
The category of the installation does not preclude the use of more sophisticated methods of assessment which may result in reductions in conservatism and hence cost. a pressure pulse will propagate into the surrounding atmosphere. An ignition source within the explosive part of the gas cloud is then assumed to ignite the local fuel/air mix causing expansion resulting from combustion in the region surrounding the ignition point. The received pressure on a flat surface may be greater than that in the incident blast wave. At pressure levels typical of a blast wave generated by a hydrocarbon explosion.(see sub-clause 5. Immediate ignition is assumed not to occur. Simple methods exist for calculating the form of this blast wave and its effects on targets in its path [67. 5.5) 3. the references refer to other sub-clauses. 70]. A hole of a given size is assumed to be present in a vessel. 1.2 Tasks for the determination of explosion loads The tasks in the determination of explosion loads are discussed in Subclause 5. The time history of the release rate is calculated. A blast wave with near zero rise time will then develop. Where given. piping or riser. This has the effect of decreasing the duration of the positive part of the pulse and steepens the leading edge. (sub-clause 5. (subclause 3. October 2003 . The convection in the flow will tend to make pressure disturbances at the back of the pulse catch up with those at the leading edge of the pulse. the ‘High risk methodology’ and the ‘Low risk methodology’ applied as appropriate for low medium and high risk installations as defined by the Risk Matrix given in Subclause 3.  2. a process which is referred to as ‘pressure doubling’.1. some are taken from military and nuclear codes. The treatment in this clause concentrates on higher risk methodology techniques with simplifications for the low risk methodology suggested for each of the tasks described. leading to a gas or spray release. The blast wave will be affected by other confining objects such as decks. The peak overpressure in this blast wave will then decrease with distance while the blast wave duration will typically increase and as a result the impulse will decrease more slowly than the overpressure. The probability of the occurrence of the release may be estimated from published failure statistics or even from simulation. Part of this cloud will be within the explosive concentration limits of the gas/air mixture. An explosion assessment is performed in the following steps.9.4 Far field effects Following a gas explosion that generates significant overpressures.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 5. There are two suggested levels of sophistication of approach.5 and Reference ) 82 Issue 1. this received pressure may be up to twice the incident pressure. Often this wave has zero net impulse as it travels away from the explosion.4. blast walls and accommodation blocks that will result in reflection and diffraction of the blast wave.4) 4. This may affect the decay of the blast wave and in some cases can increase local overpressures where a blast wave is reflected from a surface or object. A dispersion analysis will predict how the gas or vapour cloud develops and disperses under wind and ventilation conditions.
Issue 1. increased flame area.(sub-clause 5. Example High level performance standards for the installation are given in the Part 0. which the structure and other SCE’s must be designed to resist must then be derived. In order to provide some measure of asset protection as well as protecting life and the environment two levels of explosion should be considered.(sub-clause 5. October 2003 83 . creating secondary projectiles.9.1. Explosion loading software is then used to calculate how the flame front accelerates through the surrounding environment.6) 6. A blast wave will propagate away from the explosion region and impinge on adjacent structures (sub-clause 5. 11. ‘external’ explosions may result as the unburnt fuel/air mixture comes into contact with the external (oxygen rich) atmosphere (sub-clause 5. 9. Small objects may be picked up during the explosion.5 on the classification of SCE’s for explosion response assessment. Design explosion loads with a known frequency. increasing overpressures and increasing gas velocities within and outside the gas cloud. A method is discussed which uses exceedance curves for representative peak overpressures for the compartment/installation.11) 8. a process referred to as ‘Consequence Analysis’. For extreme explosion events it is conventionally assumed that fatalities will occur in the vicinity of the explosion. The peak energy for typical projectiles may be calculated from the dynamic pressure load time history and their mass. The methods for the determining structural and other SCE response and capacity are described in the next clause. This is discussed in sub-clause 5.1. Fuel/air particle velocities are also calculated to determine dynamic pressures or drag loads on any structural members. It is unlikely that the SCEs will withstand a worst case explosion scenario defined as that scenario which considers a maximal stoichiometric cloud filling the compartment or engulfing the installation ignited at the worst position. The explosion loads on piping and equipment may result in further releases with consequent fires or explosions. (sub-clause 5.10 on the generation of exceedance curves and sub-clause 6. Interaction with obstacles gives rise to increased turbulence. Fire and explosion hazard management.6) 7.4). This clause deals with tasks 3 to 11 above. It is only possible to provide barriers to protect adjacent areas.3). piping or vessels in the vicinity. Secondary. A more detailed treatment of these issues will be given in Part 3 of the Guidance. The strength level blast (SLB) and the ductility level blast (DLB) explosions are described in Clause 3 and sub-clause 5.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 5. 10. flame folding. Overpressures may then be calculated for any barriers in the vicinity.
the WOAD database . dependent on release rate or the rate of generation of new flammable volume Determine proportion of ignitions for which explosions occur (delayed ignitions).5mm) • • • • Leak frequency data 10mm representing medium releases (6. The first step in the compilation of a frequency versus overpressure curve is the determination of overall explosion frequency.1 Release and ignition data sources Item Release Sizes Possible Data Source Selected to a give spread of potential consequences (from low release rate long duration to high release rate short duration). The above steps are familiar to most safety practitioners and use generic data collected from past hazard events in the UK petrochemical industry [23 to 28]. Leak Rate Calculated from release size. October 2003 84 . inventory pressure and gas properties  .3 Determination of explosion frequency The main reason for using a probabilistic assessment method is to derive relationships for the probability of exceedance of a given explosion parameter such as overpressure or dynamic pressure at a specified location or in a specified region.10 and the Century Dynamics basis document . companies undertaking QRA tend to have their own individual procedures. Some possible methods of deriving exceedance curves are described in sub-clause 5. The following table however gives an indication of where the data can be located. Table 5. composition and pressure Calculate ignition frequency. OREDA  and the UKOOA release statistics review  . Proportion of ignitions Classification of Hazardous Locations  .5-20mm) 30mm representing large releases (20-60mm) 100mm representing catastrophic releases (60+mm) (Optional) Full bore if significantly different consequence from 100mm Hydrocarbon Leak and Ignition Data Base . This enables the range of scenarios to be compared with their probability of occurrence leading to the definition of design explosion loads which can and must be designed against.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 5. A typical spread might be: • 3mm representing small releases (0-6. There is no specific guidance on the above process. giving explosion Issue 1. This can be achieved by standard means used in Fire Risk Analysis (FRA) using the following steps: • • • • • Select representative leak sizes Determine leak frequencies from equipment item counts and generic release frequency data per item Calculate representative release rates based on release size. Ignition frequency Hydrocarbon Leak and Ignition Data Base  or Classification of Hazardous Locations  .
the workbook approach developed in the JIP on ‘Gas build up from high pressure natural gas releases in naturally ventilated offshore modules’  . The objective of these experiments was to investigate the effects on the dispersion process of release location. orientation.4 5. Cleaver  describes an investigation of gas build up in naturally ventilated offshore modules following high-pressure releases. A correlation for the flammable volume of gas within the rig (between 5 and 15 % concentration) was derived based on simple experiments and dimensional considerations. The software used for the dispersion simulations may be the same as that used for the explosion simulation itself. Gas concentration during the gas release event was recorded at up to 200 locations within the rig.13 on the NORSOK procedure. A Monte Carlo approach may be used to explicitly represent the variability of these and other parameters. Three main methods are discussed in this document for the determination of the extent of the gas cloud after a release.4. Reference . and the simplified method is given in the ‘Explosion Handbook’. 5. The weak points in this approach are the variability of gas cloud size with ventilation and environmental conditions and the choice of ignition time and location. This approach is discussed further in sub-clause 5. The majority of the releases were at a constant rate of between 0. The increased accuracy in the representation of frequency and consequence variability will offset the reduced accuracy in load determination in the determination of explosion risk.5 to 10kg/s but the effects of declining release rates were also considered.4. A plot of an upper bound estimate of the fraction of the module filled with flammable gas against the non-dimensional parameter R is given below: Issue 1. The direct CFD method has the advantage that ignition sources and their position in the gas cloud may be modelled with the effect of wind speed and direction being represented. The disadvantage is that at present these simulations do take some time to perform in view of the long time scale of the dispersion process as compared with the explosion simulation. pressure and diameter as well as module perimeter confinement and the wind driven ventilation rate.2 The workbook approach for calculation of gas cloud size. vapour cloud or liquid spray release must be calculated. Usually a phenomenological method of calculation of overpressures is used in view of the large number of simulations necessary.1 Dispersion General Once the release rate and frequency has been determined then the dispersion characteristics of the gas cloud. They are the direct CFD method whereby the cloud evolution is tracked through time. Other approaches include the zonal and Shell DICE methods. This work was based on an experimental program of 66 experiments in the full size Spadeadam test rig.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Two possible options are described for this approach . This may include a time history of the cloud’s extent and the identification of the likely ignition sources in the flammable range. 5. October 2003 85 .
Two estimates of the flammable volume are made: • A upper but realistic estimate of the flammable volume that a given release would produce for use at the ‘screening’ stage 86 Issue 1.Plot of Non Dimensional Flammable Volume versus Release Parameter  The Release parameter R is defined as: R= (dm/dt) / ( ρs Uv L2 ) Where: dm/dt = mass flow rate of gas released into module (kg/s) ρs Uv L = The density of the released gas = The average ventilation velocity in the module before the gas is released (m/s) = cube root of the module volume It was generally found that greater confinement gave lower ventilation rates and larger flammable cloud volumes.001 Figure 6 .UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 1 + + + + + + + + × ¤ + + ¤ × # ## ¤# ## # * × # × Non-dimensional flammable volume 0.01 0. October 2003 . This requires the separate determination of the ventilation flow rate through a confined and congested region and the determination of the flammable gas cloud volume.01 0. Note that the larger rate releases themselves changed the ventilation flow within the test rig. In Reference  this approach is extended to a ‘workbook’ form.1 Release parameter.1 ~~ ~ ~ ~ + × 0. R 1 0.
only the volume expansion filter should be used. Note that it is also stated that for very confined modules the burning rate filter becomes less important and in the case of a fully closed vessel. This process calculates the total mass of gas in the equivalent stoichiometric gas cloud. 5.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions • A mid-range estimate of a typical flammable volume that a release will produce to give a ‘best estimate’ of the likely outcome. For more complex flow fields the model uncertainty increases.4 Equivalent Stoichiometric Clouds At the current time it is not recommended to directly use dispersed non-homogenous and turbulent gas clouds in CFD or phenomenological explosion simulations due to a lack of testing/validation and therefore there is a lack of confidence in the codes for this application . Instead an equivalent quiescent stoichiometric gas cloud intended to give overpressures similar to the non-homogenous and strongly turbulent clouds ignited in some full scale tests should be used. provided that the ventilation flow field is close to uniform. The filters used in this process are shown in Reference . The normalised volume of the flammable gas cloud is also defined by a non-dimensional relationship. Three methods for the conversion of dispersed clouds into equivalent stoichiometric clouds are summarised in this sub-clause. It is stated that the workbook could be combined with accidental release rate/frequency data held by regulatory authorities. Finally. Issue 1. The flow through the module is characterised by a representative ventilation velocity that is assumed to be related to the external wind velocity. To calculate the Q5 parameter the mass of gas at non-stoichiometric concentrations is multiplied by the burning velocity and the volume expansion ratio at the concentration. a relationship for flammable gas volume between two given concentrations is given. 5. both normalised to that for a stoichiometric mixture.4.4. October 2003 87 .3 Explosion Handbook approach  A further simple method of calculating gas accumulation in a module is described in the ‘Explosion Handbook’ by Czujko : Mg = LVent RLeak 3600 LModule RVentilation = Mass of gas in the process area (kg) = Distance to module vent of end of congested area (m) = Length of module (m) = Gas leak rate (kg/s) Where: Mg LVent LModule RLeak RVentilation = Ventilation rate in air changes per hour It is stated that this simple model will often give a good estimate of the amount of gas within the module. Four generic module layouts are referred to specific details are not described. Hansen  describes a parameter called ‘Q5’ that is calculated during FLACS dispersion simulations.
it has not been validated as a methodology for general use. It has not been validated as a methodology for general use.6 0. For non-stoichiometric concentrations a weighting proportional to the square of the product of the laminar burning velocity times the expansion ratio.8 1.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Pappas  describes how the complex shape of a dispersed cloud can be represented using a cubic cloud typically extending from floor to ceiling in the module of interest.8 0. In Reference  it is stated that this method will generally give pressures of similar strength for the equivalent quiescent clouds as for the non-homogenous and strongly turbulent clouds ignited in full-scale tests. The data and correlation for the idealised partial fill experiments are also shown in Reference . In the full-scale dispersion and explosion tests conducted in Phase 3b  Advantica proposed an alternative method for calculating a stoichiometric gas cloud volume equivalent to the dispersed cloud in an experiment. This method was proposed for data evaluation purposes. It is also stated that the explosion overpressure durations may be shorter than for the non-stoichiometric clouds that may in turn affect the structural response. For fill fractions above 80 % the explosion overpressures are predicted to be as high as those with a stoichiometric cloud completely filling the test rig. October 2003 . The experimental data and a possible curve fit to it are shown in Figure 7.4 0.6 0. This method was proposed for data evaluation purposes.0 Equivalent Stoichiometric Fraction Fill Figure 7 . 88 Issue 1. This was used to generate an equivalent volume based on the dispersed clouds with results from partial stoichiometric fills of the test rig.0 0.Comparison of normalised overpressure plotted against the equivalent stoichiometric fill fraction for each large-scale realistic release experiment.4 0.0 ¤¤ ¤ ¤ ¤ ¤× ¤ ¤ ×¤ # # # ## × # 0. 1.0 80%ile test overpressure 80%ile max fill overpressure ¤ 0. both normalised to a stoichiometric concentration was used.2 × Partial fill correlation ¤ All Lean or Near Stoich # Some Rich # × # # # × Ideal Part Fill (Stoich) 0. It is suggested in Reference  that a simple straight line could be added to this figure to form an upper bound for the likely overpressures at small fractional equivalent stoichiometric fill volumes.2 0. A curve relating the likely overpressure to the fraction of the module filled with a gas cloud normalised to the fully filled case was derived.
6 5. 5. During the Phase 3b tests  it was observed that for realistic releases giving a non-uniform cloud density. Assuming a largely uniform distribution in space of ignition sources will result in a time dependent probability of ignition proportion to the rate of generation of flammable cloud volume. 5. This should then be multiplied by a value for the total ignition probability for which a figure of 0.1 is suggested.5. An overpressure will result from the ignition of an accumulation of mixed air and fuel.5 5.1 Explosion overpressure determination Explosion prediction methods and tools The commonly used methods for predicting gas explosion loads are have been reviewed : Issue 1. 50 mbar is the commonly accepted lower limit for the opening of pressure relief panels. Further information may be found in the OGP (formerly E&P Forum) hydrocarbon leak and ignition database. This has a bearing on the accuracy of the overall process and whether high levels of detail in other parts of the chain are needed or can be justified It is hoped that this process can be improved or that the CFD/phenomenological codes may be validated against experiment for the case of varying density within the gas cloud. A significant overpressure is defined in this Guidance as one above 50 mbar and may serve to provide a workable definition of an explosion.6. This gives an overall delayed ignition probability of 0. As there are significant unknowns in both values suggested here it is recommended that a factor of two is applied to both recommended numbers. Based on data presented by Cox Lees and Ang  a probability of 0.08 for a delayed ignition probability is suggested. see Reference .UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Note The process of the representation of a dispersed gas cloud of varying concentration as an equivalent stoichiometric cloud is considered by Advantica to introduce uncertainties in the chain of calculations leading to the explosion loads.2 Estimation of delayed ignition probability Prign  Delayed ignition of a gas cloud implies that the resultant combustion would produce overpressure effects and can be characterised as an explosion. October 2003 89 .5. This may occur during ignition of a jet fire and may result in an initial fire ball. the ignition of the cloud did not occur every time the ignition source was activated.1 Ignition General The probability of the occurrence of a detectable overpressure being developed ‘Prexpl‘ is the product of the probability of the release ‘Prrel‘ and the probability of delayed ignition ‘Prign‘. 5.032. This was due to the fact that the source may be located in a non-flammable region of the cloud or that the initial combustion ceased as the small initial flame front progressed to nonflammable regions before it could develop. If a full dispersion analysis is performed then intermittent and continuous time dependent ignition may be considered.
The increased accuracy in the representation of frequency and consequence variability will offset the reduced accuracy in load determination as compared with CFD simulations in the determination of explosion risk. momentum and energy including turbulence and combustion in a large number of relatively small control volumes covering the region of interest. October 2003 . Multi-energy method andequivalency . Generally they represent the actual scenario geometry using a simplified system. Short run times make this type of model suitable for running large numbers of explosion scenarios. 90 Issue 1. These tools can provide a wide range of information about the flow field and the explosion behaviour at the expense of significant effort required to set up a suitable geometry model and significant computational power requirements. Drag Resistance (PDR) models). Usually a phenomenological method of calculation of overpressures is used in view of the large number of simulations necessary. The limits of applicability and accuracy of these methods are generally determined by the extent of the experimental data that they are based on. Some codes can also predict the blast wave that will propagate away from the gas combustion region into the far field. Examples of phenomenological models include CLICHE/CHAOS. In practice accuracy is limited by: • • • Available computation power limiting the numerical resolution that can practically be used Accuracy of numerical models The underlying empirical sub models for Reaction zone Turbulence generation Turbulence length scale Turbulent combustion In addition the numerical grid is typically insufficient to resolve smaller items of equipment and most pipe work. Congestion Assessment Method. Can be simple and quick to use so they can be a practical design aid but in general the least accurate and cannot address general scenarios. These tools solve the conservation equations of mass. Phenomenological models typically generate a peak overpressure or a single pressure-time history taken as representative throughout the area under consideration. Phenomenological methods are simplified physical models that attempt to model the essential physics of explosions.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Empirical models based on correlations with experimental data and usually used to predict far field blast effects outside the gas cloud combustion region. for example a small number of interconnected chambers with turbulence generating source terms between them. Examples of empirical models include Baker-Strehlow. These items must be represented as they are responsible for a large proportion of the turbulence generated during an explosion so they are represented as drag and turbulence source terms within each cell (so called ‘sub grid’ modelling or Porosity. Computational Fluid Dynamics (CFD) are in principal the most fundamentally based of the methods discussed here and have the best potential for accurate prediction of gas explosion behaviour over a wide range of geometries and explosion scenarios and in both the near and far field. as they should strictly be used within the range of data on which they are based. SCOPE (note CLICHÉ and CHAOS are now included in the Advantica ARAMAS package) A Monte Carlo approach may be used to explicitly represent the variability of these and other parameters. This can be a reasonable representation of some geometries such as an offshore module but may not be adequate for more complex situations. to represent the fully 3 dimensional nature of the real geometry.
Further data are available in References [78. The codes do not directly model flame distortion and diffusion/hydrodynamic instabilities that occur between ignition and fully turbulent combustion. 5. • • Many of these deficiencies are partly due to the current limited understanding of the various flame propagation phenomena. 3a and 3b of the Fire and explosion JIP therefore provide important reference data for CFD explosion prediction tools. Importantly Phase 2 of this JIP gives a comparison of predictive capabilities for selected explosion simulation tools. Important inputs to the analysis are the laminar and turbulent flame velocities. and due to inadequate computer resources which prevent sufficient spatial and temporal resolution.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions These models are therefore calibrated against experimental data of which full-scale data is preferable to avoid scaling effects known to exist in gas explosions . Some of the relevant conclusions/discussion from a recent review completed by HSL summarised below: • • •  are The phenomenological code SCOPE and ‘simple’ CFD codes AutoReaGas. A simplified formulation CEBAM has been developed in the US. and due to lack of validation with experiments where quenching can be directly observed. 79]. The turbulent flame velocity is related to the laminar flame velocity in Appendix A of the Reference . The full-scale tests performed in the Spadeadam test rig during Phases 2. October 2003 91 . Examples of specialist CFD gas explosion simulation tools include: AutoReaGas. The codes should not be expected to accurately model flame quenching due to deficiencies in the models. EXSIM and FLACS are in widespread use Phenomenological and CFD methods generally give fairly good accuracy (within an factor of two*) so these models yield solutions that are approximately correct The limitations associated with empirical and phenomenological methods (simplified physics and relatively crude representation of geometry) can only be overcome through additional calibration Issue 1.6. EXSIM. The codes do not model flame propagation phenomena involving instabilities associated with acoustic waves and shocks. they all have the following properties:• • The codes are mostly designed to deal with fully turbulent flames. The information described here was instead based on a recent survey conducted by the Health and Safety Laboratories .3 Explosion code review/selection The relatively limited time available for the preparation of this Guidance precluded a detailed survey of the current status of explosion prediction tools.6. FLACS. Consequently the codes must rely on empirical flame folding models that are valid for a limited range of fuel-air concentrations and boundary conditions.2 Limitations of CFD codes Whilst CFD codes are considered to be the most accurate and stable codes for the generation of explosion loads and are in widespread use. 5.
’ (number 8 of main findings).UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions • It is recommended to develop ‘advanced’ CFD codes that will allow fully realistic combustion models and resolution of all obstacles but is stated that it is likely to be 10 or more years before such tools are available. and ‘simple’ CFD codes that are described in the report. One must also be aware of the uncertainties associated with whatever modelling approach is used’. This is primarily due to the large computational expense of this type of model. of main recommendations).e. given the uncertainties that remain – especially if the model is used outside its range of validation. it is not unreasonable to move the models outside there region of experimental validation as long as this is done with a knowledge and understanding of the limitations.g. the models yield solutions which are approximately correct. Further extracts from this report can be found in Reference  along with a brief description of the empirical. i. It is considered however that the present phenomenological and CFD models are adequate for the assessment process being carried out given the uncertainties in the other stages of the process.e. it may be unwise to rely on the predictions of one model only.30 % in peak overpressures [12.6. ‘Perhaps the safest that can be advised at this point is that it would be unwise to rely on the predictions of one model only. especially if a model is being used outside its range of validation. However. only for a scenario for which the model parameters have been tuned. but.’ (number 9. Note: The view in Norway is that a correct application of the NORSOK protocol for explosion simulation will result in an accuracy of +/. one must make the best use of the currently available models. be able to more adequately model turbulence and turbulence-combustion interaction as well as being able to accurately represent all important obstacles in real. 11]. better to use a judicious combination of models of different types. simplified physics and relatively crude representation of the geometry. It could be said that all could not therefore be used outside their experimental validation. complex geometries. This limits the accuracy of these models as truly predictive tools. e.4 Summary of main conclusions of HSL report : ‘The limitations associated with empirical and phenomenological models i. 92 Issue 1. Note It needs to be remembered that the all phenomenological and CFD codes represent the physics to some degree and all are ‘calibrated’ or ‘validated’ against experiments. 5. but it should be recognised that because there is an understanding of the physics involved. October 2003 .’ ‘In light of the fact that gas explosion predictions are needed now. phenomenological. but that it will probably be ten or more years before the CFD-based models will incorporate fully realistic combustion models. ‘The accuracy expected from. say phenomenological and ‘simple’ CFD models is generally fairly good (to within a factor of two). This limits the scope for improvements. can only be overcome through additional calibration. importantly.
For AutoReaGas the recommended constants provided for turbulent combustion modelling (factor slope and turbulent combustion modelling constant) should be used with close to 1m3 cells. secondary structure with cross section dimensions > 5”. The level of geometric detail usually increases as the project proceeds as smaller pieces of equipment and objects such as piping and cable trays are defined.1 Practical Use of CFD Explosion Prediction Tools Geometry requirements. Fewer cells are necessary with confinement. • Make allowance for later increases in explosion loads by multiplying the explosion overpressures predicted for a given level of geometry detail by applying a factor for equipment growth (and hence congestion) based on previous project experience Addition of anticipated ‘probable’ congestion into the explosion geometry model to allow for as yet undefined geometry • Detailed investigation of an integrated deck platform typical of the central/northern North Sea showed that reasonable prediction of the likely final overpressures required the definition of all major equipment. 72) . 5.5. Several possible measures of the ‘completeness’ of the current geometry model such as the total length of the defined obstacles or various measures of the blockage ratio have also been proposed. Primary structure. For explosion overpressures predictions at each project stage the likely effects of geometry detail to be added later in the project should be accounted for. boundaries (decks.6. methods for early project phases During the early stages of a typical offshore project. Two possible methods of addressing this have been postulated .5 5.42.2 Hints. (P4. TR/accommodation blocks). The anticipated additional congestion likely to be added as the design is progressed can be based on historical data for similar previous projects. October 2003 93 . It is recommend that uniform stoichiometric concentration gas clouds are used in explosion modelling due to lack of calibration/validation for non-uniform nonstoichiometric concentration clouds A sufficiently detailed geometry model should be used with explosion prediction models that rely on a detailed geometry model of the facility An explosion prediction code with a demonstrated predictive capability should be used.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 5. consult your code supplier. The numerical mesh should be extended sufficiently far from the region of interest to prevent boundary conditions from affecting the simulation results of interest to the project • • • • Issue 1. • • • Uniform cells should be used in the region of a CFD model where turbulent combustion will take place It is recommended that in FLACS simulations there should be at least 13 cells across an unconfined gas cloud. tips and recommendations for use Further details of these methods are summarised in Reference . equivalent equipment with all of its associated pipe work.5. all piping with diameters > 8”. only the general layout and location of the major pieces of equipment are known.6.6. If this is not done it is likely that the calculated explosion characteristics such as overpressure and impulse will increase significantly as detail is added throughout the project duration.
6. It has been found experimentally that gas explosion behaviour in confined and congested environments is dependent on the scale of the test [81. October 2003 . The key to this validation process is to demonstrate the “predictive” capability of the tool rather than simply a calibration to one set of experiments. 5. were conducted in Phase C of the Model Evaluation Exercise conducted during Phase 2 of the Fire and Explosion JIP  . made before the relevant test results were publicly available.9]. revealed that the overpressure time histories had a large number of short duration spikes which seemed to be superimposed on a generally smooth curve.7 Example overpressure traces Results from the explosion tests at the Advantica Spadeadam test site.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions • The computer code EXSIMPO is available for use with EXSIM which determines the appropriate cell size from the layout. It is worth noting that the only assessment of full scale blind explosion predictions. Comparison between results for complete stoichiometric methane fills of a 1:3. Figure 8 shows a typical trace from inside the test rig. 5.6 Validation/calibration of Gas Explosion Prediction Tools If an explosion prediction tool is to be used to calculate design loads for offshore facilities it should be ensured that the tool is appropriate for the purpose. Thus one of the main objectives of the Phase 2 BFETS JIP  was to provide specific information on the characteristics of explosions in a full-scale test rig intended to be representative of an offshore module. It is therefore recommended that the suitability of explosion assessment tools should be at least partly assessed by comparison of simulation results with a selection of test cases taken from full scale tests in the Spadeadam test rig. 94 Issue 1. This was a demonstration of certain predictive capabilities of the range of gas explosion simulation tools available at this time.5 times greater at full scale.2 scale and full size Spadeadam test rig showed overpressures between 5 and 2.6. The full-scale test results from the Phases 2 and 3a can now be obtained from the UK HSE and are an important resource for validating/calibrating explosion assessment tools.
Whatever the cause investigations have shown  that the influence of these short duration loads is insignificant for components with natural periods more than 0. Existing CFD explosion simulation codes do not generate these spikes as they cannot represent any of the processes identified above so overpressure traces from CFD simulations may be used directly as the input loading for some finite element packages (FEA) packages. These spikes which can take the observed overpressures beyond 12 bar may therefore be ignored and pressure traces may be smoothed by time averaging over a period of 1.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Overpressure (bar) 2. A typical trace for this case is shown in Figure 9.0 400.5 milliseconds.02 seconds (natural frequency less than 50 Hz) which includes most components.0 300.50 0.0 Time (msecs) Figure 8 .00 100.0 200.00 0.Example overpressure trace (inside a compartment) The origin of these spikes could be: • • • • local supersonic flow and the resulting generated shock waves localised centres of combustion and pressure generation acceleration of the pressure transducers during the explosion local response and subsequent generation of shock waves through the (bolted) structure.50 1.00 1. This form is not characteristic of far field blast waves outside a compartment.0 600.0 700. Other more refined smoothing techniques have been investigated .0 500. Issue 1. October 2003 95 .
8 Extrapolation of the results of a worst case explosion simulation For an initial screening analysis the following method from Reference  is a method for the extrapolation of the results from a simulation based on a worst case scenario with the whole volume is full of a stoichiometric mix of gas for use when partial fills are more realistic.0563) 2 − ( E1 −1. The method is based on the assumption that any gas within the module is perfectly mixed and that the probabilities of an explosion occurring at any concentration within the flammable range are equal. such as both ends and the centre of the module.6. The peak smoothed overpressure for this worst case scenario simulation will be Pult The reliability of the results of this method will depend on the reliability of the base simulation. 5.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Overpressure (bar) 400 300 200 100 0. again assuming that the module is fully filled with a stoichiometric concentration cloud.900 1.693•[( E 2 −1.200 Time (msecs) Figure 9 . It is recommended that the variation of explosion overpressure with ignition point location is investigated by considering at least three ignition points at representative points.100 1. October 2003 . This may occur if the released inventory cannot be sufficient to form a cloud of this size.700 0.00 -100 -200 0. The variation of explosion overpressure with concentration for unmitigated explosions is defined by the following relationship: PE 2 = PE 1e17.000 1.800 0.0563) 2 ] Where: PE1 = overpressure at concentration equivalence E1 (Pult for a stoichiometric mix) PE2 = overpressure at concentration equivalence E2 E = Concentration of Interest / Stoichiometric Concentration 96 Issue 1.Typical pressure trace at some distance from a vent – blast wave.
2 represent median values.4 Normalised Cloud Volume 1 Figure 10 .0 8 A relationship for mitigated explosion peak overpressure is also given in the same form as the earlier relationship for unmitigated explosions: PE2 = PE1e18.5 0.unmitigated/mitigated Peak Overpressure Maximum unmitigated explosion overpressure (Bar) 0 to 0.215•[( E2 −1.0007 ) 2 ] The factors given in Table 5.0 Pr 1 2 3 3.1 0 0 0.0 1. The relationship between the fraction of the module filled with stoichiometric concentration and the fraction of the overpressure corresponding to a fully filled module. based on a small data set.1 0.6 0.2 0. The use of these factors may considerably under or over estimate the real local overpressure peak values. Issue 1.7 0.4 0.3 0.0007 ) 2 −( E1 −1.0 2. is quantified by the use of a ratio of the unmitigated overpressure to the mitigated pressure ‘Pr.5 0. The method was developed based on experiments for one explosion geometry.5 to 1. Normalised cloud volume The influence of water spray mitigation. This varies with the unmitigated explosion overpressure as given below: Table 5. with a large spread in the data for all overpressure ranges.8 0.2 to 0. October 2003 97 .0 to 4.2 0.9 Normalised Overpressure 0. produced by a given cloud is shown below in Figure 10: 1 0.2 Pr .’.4 6 > 4.Normalised overpressure Vs.0 to 2.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions This relationship for the variation of the maximum peak explosion overpressure with cloud size has been developed based on limited medium-scale experimental evidence and FLACS simulations of these events.
It may be more useful to develop a library of example simulations of typical well defined modules. At this stage of the project very little detailed information will be available and this will almost certainly be insufficient to allow the use of explosion prediction methods that require the input of geometry models with any level of detail. This could be based on data from previous explosion load simulations for the range of concepts. for the assessment of existing platforms or for high risk or novel installations. As the spatial variation of explosion loads will not be well represented. it is not recommended that nominal overpressures or dynamic pressures are used as a design basis at later project stages. module types and process characteristics.7. and for assessment of existing facilities it is recommended that a suitably validated phenomenological or a specialist CFD explosion simulation tool is used. If it is necessary to calculate blast wave effects in the far field.7. At later project stages.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 5. installation types. It was recommended that UKOOA/HSE should fund such a project in time for the compilation of Part 3 of this Guidance. In addition normal project timescales will not allow time for detailed explosion assessments to be carried out at this time.1 Development and application of Nominal Explosion Loads Intended use of nominal explosion loads During the concept selection/definition phase of a project it would be useful to have an indication of likely gas explosion loads for use in assessing alternative concepts. calculation methods and design explosion scenarios. Nominal overpressures are defined as peak representative overpressures by installation/module type determined on a non-statistical basis from acquired experience or simulation for a demonstrably similar situation. October 2003 98 . A Rule Set defining nominal.7 5. space averaged. Frequency information should also be developed alongside the corresponding explosion load. for example at an adjacent platform.2 Factors influencing the overpressure values The Rule set could also include information on the expected variation of the nominal overpressures with respect to but not limited to a subset of the following factors: • • • • • • • Production rate Gas compression pressure Gas composition Number of production trains Module area/footprint Confinement measures Congestion measures Issue 1. The external explosion which may occur as a result of the ignition of a vented unburnt fuel/air mixture is not explicitly included in the approach although in some of the data the external explosion may have contributed to blockage of the vents and increased the overpressures in the combustion region. module sizes. 5. peak explosion overpressures or nominal overpressures may be developed to assess alternative concepts at early project phases. this could be done using one of empirical blast wave propagation methods from a knowledge of the vent areas and the mass of fuel in the gas cloud.
October 2003 99 . A number of ignition points should be considered preferably with information on the probability of ignition associated with each one. frequencies and durations should then be a good indicator of the expected severity of the design explosion event and the level of risk arising from the explosion hazard. 5. An indication of the variability of the overpressure peaks throughout the compartment should also be part of the data set. acetylene and hydrogen require special consideration. There should be sufficient gauge points to represent the spatial variation of pressure within the explosion. These values are the minimum overpressures that will be acceptable for design according to the DnV Offshore Standard DNV-OS-A101. The scenarios simulated should represent the DLB design level explosion for the ductility level event occurring at the appropriate frequency level. has been located. and impulse) should be included. Contemporary analysis methods and up to date software should be used. It is assumed that in the case of measured overpressures they have been smoothed by time averaging over a window of 1. An indication of the form of the overpressure traces (duration. Section D600. If they are used.7. It should be noted that large quantities of ethylene.5ms. Peak dynamic overpressures at SCEs of Criticality 1 and 2 should be included.7. The desirable characteristics of a suitable data set are: • The values need to represent space-averaged overpressures over the affected area which are defined as the average of all simulated (or measured) overpressure peaks within the affected region. Note that the bounding overpressures represent space averaged peak overpressure values. P9 of these rules provides guidance for hydrocarbon explosions. Reference .4 Bounding (minimum) Overpressures and durations The following source of ‘bounding overpressures.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions • Aspect ratio It would be desirable if these factors could be as near as possible independent to aid processing of the base data. The detail on the model representing the geometry should represent obstacles down to piping of 3” in diameter. This data was found to be unsuitable for the generation of reliable nominal overpressures. The following extracts are taken from the guidance: In a ventilated compartment the explosion load given by overpressure and duration is mainly determined by the relative ventilation and the level of congestion. The overpressures are not defined precisely and may not be representative of the general level of severity of the explosion event simulated. they will need to be justified with respect to ALARP. A CFD or validated phenomenological simulation method should be used for all the data. Issue 1.3 Characteristics of a suitable data set Some data has been collected on a range of concept types and is given in Appendix A of Reference . • • • • • • • The resulting nominal overpressures. 5.
For process areas on open deck covering not more than 20m x 20m with an un-congested arrangement a design overpressure of 0. Conversely with a very open layout strong flows and rapid turbulent combustion can result in high localised pressures but the typical space averaged overpressures within the region are likely to be lower than the peaks because hot combustion products will be able to dissipate. October 2003 . The effective diameter can be estimated as D = √A where A is the smallest cross-section area and L is the largest compartment dimension. Puttock  describes a negative feedback mechanism between confinement and congestion that tends to limit the typical pressures within an explosion to an overpressure of 8 bar. Larger volumes also tend to increase overpressures. If panels or walls are intended to give explosion relief by failing a peak overpressure of 2-3 times their failure pressure can still be expected in the compartment. High congestion levels may increase overpressures by a factors of 2 to 3.2s for fairly open compartments to 1s for quite closed compartments. If an explosion is completely confined typical pressures will reach 8 bar but there will be only limited flow within the explosion due to the confinement. Thus the congestion must be considered to account for the phenomena likely to be seen in an offshore gas explosion. The vent area Av may be taken as sum of free opening and blow out panel areas where static opening overpressure is less than 0.2s unless a more detailed assessment is carried out.05 may be considered as not congested.2s (200 milliseconds) unless a more detailed assessment is carried out.1bar combined with a pulse duration of 0.7. bulkheads that must survive an explosion will be designed for 4bar. This is only the case if ventilation dominates.2bar with a pulse duration of 0.2s may be used.05 may be considered not to be congested. For large and congested compartments local overpressures may be greater. Design overpressure on an open drill floor area may be 0.5 Other sources of bounding or generic overpressures.05bar. For completely enclosed compartments generally.5. Durations for explosions is expected to vary from 0. Long compartments with length/diameter ratio greater than 3 will tend to give higher overpressures due to long flame acceleration lengths. 100 Issue 1. Volume Blockage Ratios (VBR’s) of 0. Design overpressure in a ventilated shale shaker room with less than 1000m3 volume and moderate congestion may be taken as 2bar. combined with a pulse duration of 0.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions For compartment volumes of up to 1000m3.2s may be used. with relative ventilation area of about 0. and therefore little turbulence. 5. For larger or congested process areas a design overpressure peak of 0.5bar with a pulse duration of 0. stoichiometric gas cloud ignition expected to give approximately 1bar with a medium level of congestion. VBR is the ratio of the blocked volume to the total volume considered. See Table D1 in paragraph 617 of the reference for a summary of these results. A volume blockage ratio of 0.
084 + 13. The explosion overpressure impulse.000 pascals) is the impulse in Pascal seconds is the positive phase of overpressure duration in the combustion region in seconds.8 Impulse and duration related to peak overpressure It has been observed that faster combustion results in higher peak overpressures but the gas is consumed in a shorter period.Generic variation of impulse with overpressure  Issue 1.5 1 1. the peak overpressure loads can be translated into useable loads for preliminary response calculations.000/ P (Figure 12) where P I t+ is the peak overpressure in Pascals (1 bar = 100.042 P + 6. is given by:I = 0. Without undertaking detailed modelling it is not practicable to establish the overpressure time history for the specific situation. I. However.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 5. Impulse kPas Vs OP Bar 25 20 Impulse kPas 15 Impulse kPas 10 5 0 0 0. These relationships are based on CFD simulations of a small number of geometries and should only be considered as approximate example values. October 2003 101 . This fact may be used to estimate durations and impulses associated with given peak overpressures. by establishing the blast duration.5 2 2. in seconds is then: t+ = 0.500 (Figure 11) The positive phase duration.5 3 3. It is recommended that in the absence of project specific data the relationship between impulse and peak overpressure described in Hoiset  is used to derive a positive duration for the overpressure based on the assumption of a triangular pressure-time history with equal rise and fall times. This load time history may be translated into an equivalent static load which is more readily usable by the project in the concept definition and FEED stages (see Clause 6). t+.5 4 4.5 Overpressure (Bar) Figure 11 .
4 0.5 1 1.5 Overpressure (Bar) Figure 12 .5 4 4.4 1. They are the ductility level blast (DLB) and the strength level blast (SLB).9.5 2 2.2 0 0 0.9 5. This load case is recommended for the following reasons:Issue 1.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions The curves given in Figure 11 and Figure 12 may not be applicable for large open areas such as occur on FPSOs and Spars. as it was derived for enclosed compartments. If an overpressure simulation is available then a point may be positioned on the impulse/overpressure chart in Figure 11.2 1 Duration (s) 0. The strength level blast represents a more frequent design event where it is required that the structure does not deform plastically and that the SCEs remain operational. An external explosion may result in a double peak in the overpressure. 5. The blast wave from a distant explosion (>20m from the vent) will develop into a sharp fronted wave with a negative phase often represented with duration twice that of the positive phase. Low risk installations may be assessed using only the DLB.Overpressure duration relationship Hoiset  Note that the assumption of a symmetrical triangular positive phase will not apply if there are two clear paths for the pressure disturbance to reach the observation point such as would be observed on the deck below an explosion. October 2003 102 .1 Design explosion loads Load cases for explosion response Two levels of explosion loading are recommended for explosion assessment by analogy with earthquake assessment. These curves do not apply to blast waves from a distant explosion as in this case the impulse may be near zero and the positive phase duration may be much shorter. This will serve to calibrate the model for the specific situation.6 0. Duration Vs Overpressure (Hoiset) 1. A line may then be drawn through this point parallel to the line given. The ductility level blast is the design level overpressure used to represent the extreme design event.8 Duration s 0.5 3 3.6 1. This new line may then be used for extrapolation of nominal impulse and duration.
The NPD code  also gives a lower limit on TR impairment frequency of 10-3 per year. The assessment principles for offshore safety cases document APOSC  states that ‘the frequency with which accidental events result in loss of integrity of the temporary refuge within the minimum stated endurance time. • • 5.000 per year is a reasonable estimate for explosion impairment. Usually the ultimate peak overpressure ‘Pult’ derived in this way is far too large to be accommodated by the structure. Impulse exceedance curves may also be generated which take into account the duration of the load and its peak value are a better measure of the expected response of the target which will be dynamic in nature. ALARP arguments are appropriate and can be used to demonstrate that risk levels have been reduced to satisfactory levels which itself relies on frequency and risk arguments. The space averaged peak overpressure for the compartment is used for determination of the design explosion loads as it is more generally representative of the severity of the event.9. The prediction of equipment and piping response in the elastic regime is much better understood than the conditions which give rise to rupture. An explosion event in the process area will be separated from the TR by a barrier or blast wall which should withstand the load and have an impairment frequency of much less than 10 % giving a target frequency for such an event of the order of 10-4 exceedance per year. The SLB enables these checks to be made at a lower load level often resulting in good performance at the higher level (strength in depth). does not exceed the order of 1 in 1000 per year’. Hence a target of 10-5 exceedance per year for an explosion event which directly impinges on the TR is reasonable. Pult will often correspond to an event with a return period out of proportion with the design life of the installation. It is reasonable and conservative to assume that the threat from fires exceeds that from explosions by a factor of 10 to 1 and an impairment frequency of 1 in 10. An SLB event could give rise to a DLB by escalation – this should ideally not occur as elastic response of SLB and supports should be maintained.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions • • • The SLB may detect additional weaknesses in the design not identified by the DLB (robustness check). Design explosion loads were in the past derived from the a worst credible event assuming a gas cloud of maximal extent with stoichiometric composition ignited at the worst time in the worst position. October 2003 103 . This load case offers a degree of asset protection. A local overpressure peak may be used to generate exceedance curves for the determination of load cases for local design of blast wall for instance. The SLB is a low consequence event important for the establishment of operability. A frequency of between 10-4 and 10-5 exceedance per year is considered a reasonable frequency for the ductility level design event by analogy with the treatment of environmental and ship impact loads which are often considered at the 10-5 level. Issue 1.2 Determination of explosion design loads Overpressure acts directly on loaded surfaces and is available directly from computational fluid dynamics (CFD) and some phenomenological explosion codes.
This parameter gives a good general measure for the choice of design scenarios.0 0.89]. These topics are discussed in detail in Clause 6. The reason for the reduction factor of one third is related to the expected reserves of strength in the structure and the observation that the primary structure will often only experience received loads of this magnitude [88. The horizontal axis is a linear scale usually with the peak space averaged overpressure for the combustion region plotted in bar. ‘Pduct’ whichever is the greater. Each of these scenarios may have a large range of local peak overpressures and associated durations within it.5 1. October 2003 . P rexp Frequency (per year) 10 -3 10 -4 10 -5 and frequency of exceedance 10 -6 P str 0.Example overpressure exceedance curve – location of DLB and SLB design load cases (Pstr and Pduct) The SLB overpressure. as that overpressure corresponding to a frequency one order of magnitude more frequent or with a magnitude of one third of the DLB overpressure whichever is the greater.5 2. The reason for the reduction factor of one third is related to the expected reserves of strength in the structure and the observation that the primary structure will often only experience received loads of this reduced magnitude. This curve is conventionally plotted with a logarithmic scale for the vertical frequency axis which gives the frequency of per year which the given overpressure will be exceeded.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions The SLB may then be identified from a space averaged peak overpressure exceedance curve.0 1.0 3.5 Overpressure (bar) Figure 13 . this gives the impression that the curves differ in shape which depends on the axis choice. Figure 13 represents an example (simplified) overpressure exceedance diagram. Other forms of curve may be plots of local peak overpressures for a particular location 104 Issue 1. These topics are discussed in Clause 6.5 3. Many other forms of this curve have been produced with various combinations of log and linear axes.0 P duct 2. ‘Pstr’ may then be identified as that overpressure corresponding to a frequency one order of magnitude more frequent or with a magnitude of one third of the DLB overpressure.
4 The PRESTO screening model Advantica have recently undertaken some work to produce a simple screening model for overpressure prediction. The model includes parameters that characterise the degree of congestion and confinement explicitly.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions There is some evidence that the curve has underlying exponential distribution characteristics [89. They make no claims for the validity or otherwise of the present version which is still very much at an early stage of development. 10-4 per year. This is based on the Phase 2 and 3 data and proprietary test data. The same convention may be used for plotting other measures such as dynamic pressure exceedance curves for a particular part of the combustion region. The tool includes an explosion assessment method based on FLACS simulations of more than 15 modules/geometries conducted using the NORSOK procedure to generate an explosion pressure exceedance curve. The model has been developed based on data from ordinary offshore modules and is therefore not recommended for extreme cases. These curves can relate to overpressure at a point. The program is used by building an approximate model of the various field development concepts from a list of elements so that each concept can be ranked and potential ‘concept stoppers’ can be identified at an early stage. or averaged over a wall. Issue 1. The generation of exceedance curves is discussed in detail in sub-clause 5.9. An exceedance curve will always be a monotonically decreasing (discrete) function. For a specified solid barrier the model can be used to calculate the explosion pressure for a defined annual frequency of occurrence e.9. 5. The dynamic pressures corresponding to the SLB and DLB load cases may then be obtained in the same way by selection of the appropriate frequencies of exceedance.3 The COSAC Risk Assessment Tool COSAC  is a commercially available risk analysis software tool intended for concept evaluation and screening at earl project phases.g. 5. Exceedance curves are typically plotted on a graph with overpressure plotted on a linear scale on the horizontal axis and annual exceedance frequency plotted on a log scale on the vertical axis. or other explosion properties such as dynamic pressure or impulse.1 Generating Exceedance curves General There is currently much industry interest in the generation of curves of the probability of exceeding a specified explosion load at a given location. 5.10 184.108.40.206] which gives a straight line when plotted with log-linear axes as in Figure 13. October 2003 105 .
It is important to identify the explosion scenarios with the higher overpressures as these will determine the required exceedance probabilities in the required range (10-5 to 10-4 exceedance). Decisions must therefore be made about how to represent these influences on the explosion. October 2003 .UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions ‘Several methods of generating exceedance curves are in use. in for example. The total probability of an explosion gives a method of determining if enough scenarios have been considered and convergence is being reached.4) 106 Issue 1. However. 5. some of these are described in Reference .7 describe methods of extrapolating from a small number of explosion simulations representing differing ignition points to represent various (equivalent stoichiometric) cloud sizes and the expected overpressures.10. An exceedance curve may be constructed using these extrapolated results. The most rigorous and repeatable method of exceedance curve generation is that following the NORSOK procedure  described briefly in sub-clause 5.5 presents a simplified method based on the assumption of exceedance curve shape which in theory requires only requires a small number of simulations to be performed.5) and determine release probabilities (sub-clauses 2.13.2 and 5. The other point uses the probability of a release forming a cloud combined with the probability of a late ignition Prexp. select leak scenarios (sub-clause 3. The range of methods represents the fact that at an early stage in a project the information available may only be sufficient to apply a relatively simple methodology but that more complex methodologies can be applied as the project develops. Figure 14 illustrates the sequence of tasks for exceedance curve generation. The release. release or ventilation rate. ignition and explosion must all be considered in generating exceedance curves.’ Reference  and sub-clause 5. Sub-clause 5.5). and uncertainty in both the data and models used should be considered. calculate release rate time history 3.3) 2. cloud formation.10. In addition the choice of models will also affect the methodology applied due to the resources they require. the range is also due to the number of factors that can contribute to an explosion occurring and that different methodologies account for these factors differently. Natural variability.2 Generation of exceedance curves for design explosion load case determination The process below is a method of medium complexity for the generation of exceedance curves for the purpose of identification of the design explosion events corresponding to the SLB and DLB.6. It is advisable to consider space averaged peak overpressures for this purpose as they are more representative of the general severity of the load case. This method may be used only at an early project phase when sufficient information for more sophisticated methods is not available. calculate cloud size time history (sub-clause 5. Explosions with a frequency of exceedance of greater than 10-3 will also dominate the accuracy of the exceedance curve around the 10-4 exceedance level. The chosen scenarios will themselves give rise to simulations which have large local variations of peak overpressure. 1. corresponding to the zero overpressure exceedance probability (subclause 5.
calculate equivalent stoichiometric cloud size (sub-clause 5.4. At a lower level the expected higher explosion overpressures would occur in situations of poor ventilation with the wind from a direction blocked by a barrier or equipment. The probability distributions of the input variables are considered in giving these bounds. In addition it will be difficult to determine its probability of occurrence.10. October 2003 107 . release and ignition.6) A high level of sophistication method could use CFD for the cases where the highest overpressures are expected with calibrated phenomenological simulations for interpolation between scenarios if required. Issue 1.6. A simplified method which may be used for the purpose of design explosion load determination. assemble the space averaged peak overpressure exceedance curve from probabilities of occurrence of the overpressures for the scenarios simulated. At a lower level the ignition probability could be assumed constant irrespective of cloud size but with say 3 widely spread locations for the ignition sources. will be estimated as a time history and will be proportional to the rate of generation of inflammable cloud volume which should be available from a CFD dispersion calculation. is presented at the end of this sub-clause in 5.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions At the highest level of sophistication this would involve consideration of ventilation rates by wind speed and direction by CFD simulation. 6. calculate the space averaged peak overpressure (sub-clauses 5. 4. calculate ignition probability At the highest level the probability of ignition from a continuous source. The wider range of parameter variations offsets some of the uncertainties in the extrapolation techniques which could be used if a CFD approach with fewer simulations is used. A simplified method might assume a near constant release rate to generate a series of cloud sizes with equal probability. a Monte Carlo simulation method may be used based on a phenomenological model used such as ARAMAS which generates space averaged overpressures implicitly. 7.4) note the reservations quoted in this sub-clause on the accuracy of this step in the analysis process. The worst credible scenario assuming a stoichiometric cloud filling the module and ignited at the worst position may not be representative of the general population of explosion events.6 and 5. A representative scenario at around the 10-4 per annum exceedance level or the 10-5 frequency of occurrence level should be used if it can be identified. Reference  forms the basis of this clause and contains a number of reviews of the literature relating to this topic. If local overpressure exceedance curves are needed for a reliability analysis of a blast wall for example these may be generated based on the probabilities for the averaged pressures and transfer functions relating these two values for the chosen location  .5. A low sophistication method may use one of the larger cloud size cases with interpolation for smaller clouds using approaches described in sub-clause 5. The workbook approach as described in Reference  and sub-clause 5. Alternatively.6.4 could be used. Bruce  describes a typical methodology for estimating overpressure exceedance distributions with associated confidence bounds.8 and Reference . Several thousand simulations may be performed for variations of all input parameters governing environmental. 5.
1 0.04 0.08 1 0. DENSITY 1.001 0. PROBABILITY 0.04 0.55 0.9 0 1.0 0.6 0.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions WIND ROSE N NW W SW S NE E SE VENTILATION AREAS FOR EACH FACE OF MODULE PROB.5 0.1 1.8 OVERPRESSURE (BAR) Figure 14 .000001 0.02 0 200 100 150 250 300 0 50 CUMM.0001 0.8 0.5 0.00001 0.06 0.02 0 AIR CHANGES/HOUR FORCED VENTILATION RATE HOLE SIZE DISTRIBUTION FREQUENCY FUEL CONCENTRATION VERSUS TIME % FUEL UEL LEL TIME HOLE DIA.4 0.5 1.2 0.06 0.5 RANDOM CHOICE OF IGNITION LOCATION WITHIN MODULE OVERPRESSURE EXCEEDANCE PROBABILITY 0.2 1. INVENTORY VOLUME AND PRESSURE FUEL CONCENTRATION PROBABILITY DENSITY IGNITION/EXPLOSION PROBABILITY VERSUS TIME PROB.Overview of Probabilistic Blast Modelling Approach  1 1 6 108 Issue 1.7 0.08 0. 6 1 1 1 FUEL/AIR MIXTURE OVERPRESSURE DISTRIBUTION PROB.4 1.01 0.00 E+00 0.3 0.3 1. DIST.3 0. October 2003 .05 0. DENSITY DISTRIBUTION OF VENTILATION RATES 0.
08 −1− P ) Where P is the overpressure in bars. The SCOPE model that only takes a few seconds to run is used for this purpose. for example over a complete congested region. As the worst-case overpressure will vary from module to module the generic exceedance curve is plotted as the ratio of overpressure to the worst-case overpressure. If P is replaced by S in an expression fitted to some experimental data then at low overpressures the predicted overpressures will be unchanged. Transform (or transfer) functions that relate the median overpressure produced by SCOPE to features such as localised high overpressures or geometry effects such as shielding or pressure wave reflection are then defined.3 Generic exceedance curves Puttock  describes the possible use of generic pressure exceedance curves designed to give a conservative estimate of the required probability for all modules of one generic configuration. E is the expansion ratio at atmospheric pressure.10.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 5. An expression for S is given as: S = P. S tends to infinity at 8 bars that would correspond to adiabatic combustion of a stoichiometric hydrocarbon without any expansion. The method takes account of variations in leak rate. Example exceedance frequency and exceedance probability curves are given for pressure and impulse in the reference. Severity index. If the correlation is used for S and S inverted to get P using the above equation it will automatically be limited in a realistic way to 8 bar. Where further detail is required CFD runs are performed using EXSIM. It is stated that the only feasible way to take into account the complexity of the physical processes in a gas explosion is to use a Monte-Carlo method that requires performing many thousands of model runs to obtain valid statistics. October 2003 109 . Flammable gas cloud sizes can be calculated in two stages using EXSIM (ventilation flow followed by dispersion simulation). One problem with the use of this type of generic curve is that they do not allow for the fact that the level of overpressure will affect the shape of the plot because if the overpressure is already close to 8 bars it is unlikely to significantly increase. wind speed and direction. is defined in Reference  based on runs of SCOPE 3. S. or using a zone model to predict ventilation flow followed by a random-walk dispersion method. The idealisation of the module used in the SCOPE input is adjusted if necessary so that with the module fully filled with gas and ignition at the centre of the module would give the same median overpressure with both models.4 P E1. Ratios of a parameter called severity index rather than the overpressure can be used to allow for this. ignition location and stoichiometry. Transform functions at each ‘receptor’ are calculated by running EXSIM for range of initial conditions. fuel type. It is however difficult to prove the general applicability of generic curves as their detail is likely to vary between different modules/plants.e (0. Note that this approach assumes that P is a typical overpressure. Issue 1.
110 Issue 1. A high overpressure. A suitable representative overpressure may be defined as the average of the highest 20 % of the peaks in any particular scenario.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 5. The difficulty is the determination of this frequency and it is argued that this case is not a representative point to take. This definition is used to avoid distortion of the resulting curve resulting from the variability of local overpressures. This method is based on the a priori fact that the statistics of random events such as explosions indicate that the intervals between such events will often have an Exponential distribution.5 Simplified Methods for Pressure Exceedence Curve Generation This method requires that two overpressure/probability of exceedance points are defined. The two plotted points are then joined by a straight line to give a simple relationship between overpressure and exceedance frequency.10. Four separate approaches are described in ascending order of complexity. October 2003 ..4 Mobil North Sea Methodology for Early Design Blast Analysis Reference  describes four methods to allowing designers to select appropriate gas explosion overpressures for use in early stages of the design process when only a small number of simulations have been performed (one for each ignition point considered). The guidance given is conservative in certain aspects in an effort to ensure that the estimated blast loads remain firm throughout the design process. A recent (unpublished) paper by Yasseri  has processed the historical data on explosions in the North Sea given by Vinnem  and has shown that there is a good fit for the data to the exponential distribution. These are then plotted on a graph with overpressure plotted on a linear scale on the horizontal axis and exceedance frequency plotted on a log scale on the vertical axis.10. The relationship between overpressures for clouds of different sizes and simplifying assumptions about the probability of ignition enable the construction of an exceedance curve for use in identifying design explosion loads in the most sophisticated method described. low frequency event based on the ignition of a full stoichiometric concentration fill of the module Pult and its associated probability. 5. Figure 13 shows an exceedance curve generated in this way from a sample representative overpressure and frequency of exceedance obtained by simulation. If the sample simulation corresponds to the worst credible case with a full compartment of a stoichiometric gas cloud ignited in the worst position Pult. A representative explosion overpressure at an exceedance frequency of 10-4 obtained using the COSAC software. An exponential distribution gives a straight line for the exceedance diagram with the axes chosen. A curve based on local peaks will be specific to the locations chosen. The first two approaches are hazard consequences based while the last two are risk based. then the frequency of exceedance is the absolute frequency. Possible methods of calculating the two required overpressure and frequency points include the following: • • • A point at zero overpressure corresponding to the probability of an ignited gas leak Prexpl .
5. Exceedance curves for dynamic pressures may be developed from simulations and used in the same way as for overpressures in deriving design dynamic overpressures. The performance of the structure and SCE’s for these scenarios must then be tested against the appropriate high level and equipment specific performance standards.3m or on cylindrical obstacles less than 0.7. Issue 1. In open areas. SCEs of criticality 1 will be assessed against the DLB as derived in the previous sub-clause. October 2003 111 . If the general level of dynamic pressure loads is not known then it is acceptable to take a load equal to 1/3 of the overpressure at the location for the DLB load case.3m in diameter in particular in regions of high gas velocity near vents.11. its density and the area presented to the flow by the obstacle. Pressure difference loads.53.97] must be determined and are referred to as dynamic pressure loads. Inertia loads proportional to the gas acceleration and the volume of the obstacle.11. however there is then the difficulty of estimating the frequency of exceedance. these loads should also be applied in the vertical plane. These may also be obtained from CFD simulation results and consist of: • Drag loads (similar to the Morison drag loads experienced in fluid flow) proportional to the square of the gas velocity.3 identifies three categories of criticality of safety critical elements (SCE’s). 5. In this example all walls except the West wall are solid. The dynamic pressure loads should be evaluated on SCE’s of criticality levels 1 and 2.1 Loads on piping and equipment Load cases for piping and equipment response Sub-clause 3. as there will be some cases where the overpressure exceeds this value each with their own frequency of occurrence. • • Drag loads dominate for obstacles with dimensions less than 0.96.11 5. The duration should be chosen so that the impulse matches that of the overpressure trace. SCEs of criticality level 1 and 2 should be assessed against the SLB.2 Dynamic pressure loads The explosion loads on equipment items and piping which are classified as SCEs [45. This load must also be applied in the reverse direction.3m and 2m in the flow direction. such as the decks of FPSOs. Both drag and pressure difference loads are significant on objects between 0. The blast overpressure causes a pressure front to move from left to right from the point of ignition at about the speed of sound in the unburned mixture.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions It is preferred to take a representative sample overpressure nearer to the DLB at a frequency of exceedance of between 10-4 and 10-5 . The situation in Figure 15 represents an explosion in a module compartment.
3 meters) at any time is given by the drag term in the Morison equation. The air ahead of this front is pushed out through the vent in the West wall over a vessel and pipe work spanning the vent giving the possibility of vessel or piping failure. The gas velocities in this case will occur predominantly near the vent. The force should be calculated for engulfment of the obstacle in the burnt and unburnt gas mixture as the burnt gas/air mixture may be travelling at a speed of ten times the speed of the unburnt mixture even though the unburnt mixture is ten times denser. Further details will be given in Part 3 of this Guidance. A typical time history for the drag load is shown in Figure 16 from Reference . Drag coefficients are given in Czujko and Reference . The magnitude of drag forces on the pipe work (with diameter D typically less than 0. with further release of inventory and consequent escalation.Explosion within a compartment [98.99] The unburned gas is pushed out of the module through the vented area with a velocity. October 2003 .UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions S E N W West wall open Explosion Vessel 70kN/m Overpressure load Dynamic Pressure Load Figure 15 . 112 Issue 1. This velocity will be available directly from a CFD simulation or by the approximate method given in the next sub-clause.
56 1. Catlin’s paper describes some experimental validations of formulae for both vented gas velocity from a compartment open on one face and the loads on vertical cylinders in the vent area  . For larger diameter obstacles or vessels.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 5000 4000 Drag (Pa) 3000 2000 1000 0 1.50 1.Example dynamic overpressure trace Two peaks are shown corresponding to the peak gas velocities ahead of and behind the flame front. Drag loads are particularly important in open areas such as on the deck structures of an FPSO.62 1.60 1.44 1. October 2003 113 . the pressure difference across the vessel will also need to be calculated and added to the drag force above.68 1.58 1. in particular in the immediate vicinity of the perimeter of the enclosure where the explosion vents.42 1.66 1.” These validated methods are based on TNT equivalent explosions with non-turbulent fluid flow at a distance allowing the use of Morison’s Equation to determine the load on a body.80 Time (s) Figure 16 .48 1.3m. Issue 1.74 1.76 1.78 1. The IGNs  state that the validated methods for determination of loading due to blast wind are strictly only applicable to distant explosions: “there is no equally established methodology for structures either within or close to an explosion typical of offshore geometry.72 1.46 1. In a vented compartment flow reversal of gases into the compartment could also occur at a later stage in the explosion. Secondary projectiles may be a problem for FPSOs in view of the higher gas velocities.64 1. The gas clouds associated with explosions on FPSOs may be very large and gas velocities up to 500 m/s could be experienced. This would give a trace with a negative phase.54 1.52 1.70 1. The drag loads may be used to represent the total force on obstacles with in flow dimensions less than 0. The direction of gas flow may also be very variable for example in the case of the pipe rack of an FPSO acted on by an explosion ignited at low level.
8.050 1.00 ∆P 1.8. Overpressure (bar) 3. October 2003 .3 Loads on Vessels It is possible to estimate an upper bound to the pressure difference force from an explosion pressure trace by assuming or from knowledge of the speed of propagation of the pressure disturbance. Under these assumptions.00 0.900 0.11.00 ∆ t=D/U -2.00 0.Estimation of the pressure difference across a vessel from a pressure trace This process is conventionally applied for the phase of the explosion where the vessel is in the unburnt air/fuel mix using the appropriate speed of sound.00 -3. This same model may be used to estimate the tension effects in one-way spanning blast panels discussed in sub-clause 6.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 5. 114 Issue 1.00 2. this same load may be applied after the peak acting in the opposite direction.100 Time (secs) Figure 17 . The part of the curve relevant to this calculation is that before the peak.950 1.00 -1. If the pressure pulse is considered to propagate with a velocity U then a time interval ∆t may be associated with the diameter of the vessel D through: ∆t = D/U The pressure difference across the vessel ∆P may be read off the trace as indicated in Figure 17. For piping the response to the drag load may be estimated using a single degree of freedom method such as the Bigg’s method described in sub-clause 6. A time history of drag load may then be estimated.000 1.
5. The DLM can then be applied.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 5. In the DLM method. 5.98] For the situation of a vented compartment open at one end. October 2003 115 . Tests  have shown that the force perpendicular to the grating may be calculated from a Morison drag formulation using the cross sectional. the pressure differential across equipment in each direction is taken from gauge (or measurement) points located on the up and down wind sides of the object and multiplied by the obstacle windage in each direction. This should not be a problem for objects large in relation to the size of the control volumes use in the simulation (3 times the cell dimensions). the velocity of the unburned gas may be estimated using the Rankine-Hugoniot equations for the change in pressure and density across a shock wave .6 Estimation of Vented Gas Velocities [100.3m in diameter.54]. For the simulations considered during the referenced work. This may add short duration loads (duration a few milliseconds) particularly a high Mach number flows. If the object is large and the gauge points are located at the centres of the object cross section then the computed pressure differential may be multiplied by 2/π to allow for the sinusoidal variation of load direction over the surface of the object.11. area of the grating presented to the gas flow. It is unlikely that drag loads would in any case be large at these locations so the pressure difference approach is recommended in this Guidance. with the ambient internal pressure and instantaneous overpressure is . Methods for dealing with multiple objects that contribute drag forces to the total load on an item of equipment are briefly outlined [53.4 Loads on grating Grating is often used for decking in an effort to allow venting of explosion gases. it was stated that the peak drag pressures were about 30 % of the maximum local field pressures and 60 % of the average overpressures experienced within the module. it is stated that near module vents forces calculated by the DLM and drag methods are similar. Alternatively. including a method for dealing with groups of objects that are partially shielded from the blast wind.5 Considerations in the use of CFD In References  and  dynamic pressure loads were calculated using the conventional drag formulation given above and directly from CFD simulations. For intermediate obstacle sizes it is suggested that the CFD numerical grid be locally refined until the object is sufficiently resolved in the flow field. It is recommended that this is used for objects greater than 0.11. the Direct Load Measurement (DLM) method.11. There are only two published circumstances where explosion gas velocities have been measured: Issue 1. Near the centre of a module the loads on obstacles it is recommended in the references that both approaches should be used and the larger of the two loads used for design. It should also be noted that the gauge points would not capture local pressure increases where the flow stagnates on the object if they are far from the obstacle in relation to the object diameter. One problem with this method is that overpressures are generally recorded at the cell centres.
During a vented internal explosion there may be an out of balance lateral force on the compartment depending on the distribution of vent areas around the module. 5. The asymmetric expulsion of vented gases from a compartment may give rise to out of balance loads or strong shock response.7 Strong Shock and Global Reaction Loads The two major sources of out of balance loads from explosions are: • • Reaction loads from the expulsion of vented gases.12 Reporting Template for ALARP demonstration The following should be considered when preparing the final ALARP justification for the management of explosion hazards: Method of selection of explosion scenarios • • • • • Available inventories Hole sizes and leak rates Leak locations and directions Wind speeds. if the local speed of sound is exceeded then the flow becomes ‘choked’ and a second shock wave front is set up at the vent. The SCOPE experiments executed by Shell where gas velocities were measured directly 2. The backpressure for confined flow may be represented by a ‘loss factor’. that validation of pressure may be sufficient to assume that the gas velocities are being correctly calculated.11. directions and ventilation rates Cloud build up Summary of explosion scenarios considered 116 Issue 1. Obstructions at the vent such as louvers may accelerate the flow locally and induce these standing shock waves. However. The net force on the gas volume contained in the module may also be calculated from rate of change of momentum of the enclosed gas. Side loads due to the ignition of an external gas cloud which has drifted to one side of the platform – the external explosion. piping and vent obstructions. The Phase 3b tests where an instrumented cylinder was placed in the vent area of the Spadeadam test rig . 5. which represents the loss of momentum encountered by the flow .UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 1. October 2003 . The general view is that in CFD codes the pressure and gas velocity is so closely linked at each point in space and time. In fact the time for pressure disturbances to cross the module may be appreciable compared with the load duration and the inertia of the un-burnt products and external atmosphere will affect the net force. Initially it seems that the out of balance force on the module will be equal to ∆PA less any net forces on internal equipment. which will restrict the flow.
even if the analyses are performed in different circumstances and by different personnel’. The procedure can be used to calculate exceedance curves for the overpressure and frequencies can be established for unacceptable explosion consequences. installations and concepts. Conversion of ‘typical’ overpressure/dynamic pressure time histories into equivalent triangular form.13 The NORSOK simulation Procedure for probabilistic explosion The NORSOK Z-013 standard  describes a procedure for probabilistic explosion simulations. The procedure briefly also covers response calculation and consideration of overpressure mitigation methods such as deluge. Issue 1. decks Equipment Piping Far field blast 5. Statoil and Saga to accelerate and harmonise the development of probabilistic explosion assessment tools. shape and location used in simulations Ignition point locations considered Summary of explosion overpressure results • • • Peak/typical overpressures at a point and over structural panels Peak/typical impulses at a point and over structural panels Dynamic pressures Method of derivation of design basis • • • Selection of equivalent static pressures – if used. with justification that this form is appropriate Use of full ‘typical’ overpressure/dynamic pressure time histories Statement of explosion loads used in design • • • • Structure. One purpose of the procedure is to ‘standardise the analyses so that the risk of explosions can be compared between different areas. This procedure is based on an initiative started in 1998 by Norsk Hydro. It is intended to be used for detailed analysis of platforms in operations or the project phases where the necessary detailed information is available for all design elements that will influence the risk picture. size.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions • • • Gas type and concentration Cloud gas concentration. blast walls. October 2003 117 .
In cases where low loads are expected or where the structure has high strength so that larger conservatism can be accepted.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions The procedure is meant to be used for detailed analyses of platforms in operations or in the project phases where the necessary information on all design elements influencing the risk picture is available. In particular the methods described in the NORSOK standard Z-013 represent the most technically rigorous methods available at the time of writing and are the benchmark against which the methods used should be gauged. The amount of equipment shall be estimated based on equivalent areas in previous studies. but the general structure and principles must be maintained. This Guidance is consistent with the underlying philosophy of the procedure but also aims to identify ways by which the volume of computation effort may be safely reduced. Sensitivity studies shall be used to establish whether minor changes in amount of equipment. the procedure may be simplified provided the conservatism is under control. In early project phases the project must be simplified according to the design information available. pressure relief areas and ventilation will change the gas explosion loads significantly. 118 Issue 1. October 2003 .
which enable design explosion loads to be determined which should be accommodated by the structure and other SCE’s. Issue 1. The ‘robustness’ approach is still valuable and should be considered in addition to the more rigorous probabilistic methods described in this Guidance.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 6 6. The methodology for the assessment of explosion response is illustrated in Figure 18. This approach is to an extent scenario independent and may give added protection against unidentified scenarios and in particular combined fire and explosion scenarios. Over the last ten years. October 2003 119 . piping and safety critical elements (SCE’s) to explosion loads derived following the methods given in the previous clause. many structures have been designed to resist uncertain explosion loads by the calculation of the capacity of the structure and the SCE’s and the demonstration of robustness in the structure as reflected in an insensitivity of response to variations in load.1 RESPONSE TO EXPLOSIONS Overview of explosion response This clause is based on Reference  and deals with the methods of analysis of the response of structures.
UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions
Inputs Peak overpressures (DLB and SLB), durations/time histories, dynamic pressures Pressure exceedance diagrams, element specific performance standards Probability of late ignition of releases Structure geometry, Loads on SCEs and supports, SCE criticality levels From Explosion load determination
Structural Safety Critical elements
Safety Critical piping and equipments items
Check connections and reactions from decks barriers and SCE's to DLB loads
Use dynamic pressure loads From (4) above or estimate Ductility level and Strength Level dynamic pressure load = ½ of DLB or SLB overpressure
Check primary structure to DLB - non-linear or modified code check method Check primary structure to SLB - elastic accidental load case Check secondary structure to DLB and SLB
Check SCEs and Supports to DLB and SLB Dynamic pressure loads
To ‘Evaluation’ Outputs Response durations/time histories, peak response of structure and SCEs to design explosion loads, ductilities, displacements, stresses and associated probabilities. SCE criticality levels, rupture and associated probabilities, element specific performance standards, escalation targets
Figure 18 - Explosion response
Due to the extreme nature of the ductility level blast (DLB) loads, it is essential to prevent overall structural collapse but local failure may, however, be tolerated, as long as these do not lead to an escalation of the event, whether through progressive collapse, leakage of hydrocarbon inventory or failure of safety critical elements (SCE’s). This places considerable emphasis on the requirements for structural and equipment robustness, i.e. ensuring that systems can absorb significant amounts of energy and be able to redistribute internal forces via the provision of adequate alternative load paths.
Issue 1, October 2003
UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions
One way of achieving structural robustness, is to prevent premature failure by buckling and shear failures which may produce brittle, as opposed to ductile response. Adequate connections to adjoining members will ensure the ability to redistribute load and enable the structure to mobilise its ultimate capacity. The goal of robustness ensures that the structure has the ability to absorb some of the uncertainty in the loading to which it may be subjected. A review of the research in the area of response to explosion loads since the generation of the Interim Guidance Notes  , or IGNs, is contained in Reference .
Information required for explosion response calculations
Clause 5 describes the information which will be produced by the explosion loading specialists in order that an installation may be assessed for its ability to resist the explosion event.
Information from the explosion load simulations
For High risk installations, these will include:-
Pressure exceedance diagrams for each combustion zone. Identified strength level blast (SLB) and ductility level blast (DLB) load cases – these should include special distributions of peak overpressure and overpressure time histories at critical locations. SCE’s and their criticality levels Dynamic pressure exceedance curves for SCE’s of criticality 1 for the DLB with time histories of dynamic pressures at these locations. Dynamic pressure exceedance curves for SCE’s of criticality 1 or 2 for the SLB with time histories of dynamic pressures at these locations. Other overpressure and dynamic overpressure information for significant scenarios with similar frequency, where the loading pattern differs appreciably from the design explosion events. If this is not available then a range of likely durations should be supplied.
• • • •
If exceedance curves are not generated then the probabilities associated with the design explosion events should be available. If time histories are not supplied load durations and general shape of the load time histories should be supplied. Low and some medium risk installations will only require the information for the ductility level blast explosion loads. A comparative assessment method may be used drawing on past experience from a demonstrably similar structure geometry and scenario. The nomination of a typical installation to represent a fleet of platforms is acceptable.
Issue 1, October 2003
UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions
Other information from non-structural disciplines
The project should also supply
• • • • •
High level and system specific performance standards for evaluation of the results of the assessment. Layout drawings, including process, escape routes, muster areas, the TR, the SCE’s and their support structures Any structural drawings and the location of tall structures which may become a hazard during escalation. Details of the assumptions for detection, control and mitigation systems, location and required availability. For dynamic response analyses, the location and masses of all major items should be supplied.
Further exceedance curves will be required from the explosion specialists for dynamic pressure loads, to enable the dynamic pressure loading on pipework to be developed. In order to determine the loading on large items of equipment or vessels, the required items can be identified to the explosion specialist for direct load extraction from the CFD model for various explosion scenarios. Alternatively, a less accurate, but also less time consuming method for the explosion specialist, is to provide generalised pressure-time histories for the equipment locations, together with the speed of travel of the pressure wave. The structural engineer then determines the maximum pressure difference between locations of known separation, having determined the time taken for the wave to travel that distance. If the general level of dynamic pressure loads is not known then it is acceptable to take a load equal to 1/3 of the smoothed peak overpressure at the location for loads on the relevant SCE’s and piping. The duration should be chosen so that the impulse is matched to the positive phase of overpressure trace. This load must also be applied in the reverse direction. In open areas, such as the decks of FPSOs, these loads should also be applied in the vertical plane.
Overpressure load considerations
It should be borne in mind that the design explosion loads represent only one of many scenarios and that the reliance on a particular overpressure time history or spatial distribution of overpressure is not advised. The fact that the detail of a given load distribution is only one of many possible scenarios also implies that reliance should not be put on the apparent result that some parts of a blast wall for instance seem to receive less load than others, as a different ignition point may give rise to a different distribution of load. It is therefore not acceptable to design adjacent areas to different pressures without justification. The structural engineer should develop the idealised pressure-time histories for various pulse durations in order to capture the range of dynamic response for the various components of the system to be designed or assessed or preferably use a representative range of overpressure time histories directly if they are available.
Issue 1, October 2003
UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions
If a range of pressure time histories is not available then extrapolation of the range of durations supplied should be investigated. The peak pressure/duration relationships  given in subclause 5.8 may be used to estimate the peak overpressure for shorter duration loads. The peak averaged overpressure is plotted against impulse in Reference . It is accepted that the design overpressure for the design of a panel will be greater than that for a section of a blast wall which will in turn be less than that for a whole wall. This is a result of the fact that the pressure disturbance is of finite extent relative to the target and hence the averaged pressure will vary with target dimensions. The explosion load specialist may supply a lower level of design load for an extended structure (deck or wall) than for a blast wall. Reference  gives a simple method of deriving scale factors. The approach should be applied with care as the direction of travel of the overpressure pulse relative to the target affects the way the pulse is reflected. The approach is suitable for large deck areas or if sufficiently detailed information on the loading pattern is available. In modelling the response of a structural frame, it is conservative not to model the cladding or plate as the shear restraint from such surfaces may be overestimated. The overpressure load on a plate may be applied directly to the bounding members of the plate using the ‘area tributary method’. For example the loads on a square plate may be applied to the bounding framing members with one quarter on each, for a rectangular plate the members on the long edges would receive a load proportional to the triangular area joining the edges with the plate centre.
Elastic Dynamic response
Most of the loads resulting from an explosion are dynamic in nature hence dynamic response must be considered unless some form of equivalent static load can be justified. This sub-clause deals with elastic dynamic response where the primary structure remains elastic during the explosion event. This may be a system specific performance standard for the temporary refuge (TR) primary structure for example. Elasto-plastic response of components is considered in sub-clause 6.8 where Biggs method is discussed . The response of a structure to a dynamic load is commonly characterised by the ratio of the load duration ‘td’ to the natural period ‘T’ of the structure. Depending on the value of this ratio, three response regimes are defined, denoted as either impulsive, dynamic or quasi-static. Table 6.1 is a modified version of the one in the IGN to reflect more recent work published in NORSOK  and summarises the influence of the loading characteristics on response in the three regimes.
Issue 1, October 2003
Multi-Degree Of Freedom (MDOF) systems may have a large number of modal periods associated with them. The impulse duration ratio at the impulse to dynamic boundary has been reduced from the 0. Preserving rise time is not important. Accurate representation of the impulse is important.0 Preserve peak value – the response is sensitive to increases or decreases in peak load for a smooth pressure pulse Preserve load duration since in this range it is close to the natural period of the structure. Accurate representation of impulse is not important.. Duration Impulse Accurate representation of impulse is critical.3 Peak Load Preserving the exact peak value is not critical Preserving the exact load duration is not critical Dynamic 0.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Table 6.0 values given in the IGNs. 124 Issue 1.0 Quasi-static td/T >3. Even slight changes may affect response. October 2003 . the rise time may be short and this can cause significant dynamic amplification as may also be seen in Figure 19 taken from Reference  for the sharp fronted pulse shape. although it is often possible to identify a mode or shape of response with each modal period. but is critical when response is plastic. The Dynamic Amplification Factor ‘DAF’ (dynamic/static response) does not approach unity for a triangular load pulse until a value of between 5 and 6 is attained. Not important if response is elastic. This table assumes that there is some identifiable natural period for the structure. It is important to note that although the load duration may be long.4 and 2.1 Regimes of dynamic response Impulsive td/T < 0. which indicates that a Single Degree Of Freedom (SDOF) idealisation of the structure is possible.3< td/T < 3. ignoring it can significantly affect response. If the loading is impulsive much higher peak loads can be tolerated than the static capacity of the target. Rise Time Preserving rise time is important. The limits specified in the Table above have been changed to correspond to those in the Norsok standard  .
8 0.2 /T 1.t Figure 19 .6 0.4 0. Figure 20 shows the conventional idealization of a pressure trace.2 0.6 1.8 2.4 1.0 1 1.2 Idealisation calculations of overpressure time histories for response For some design applications a simplified form of pressure-time history is required.3.0 Impulse length ratio.4 Rectangular 2.8 1. October 2003 125 .Idealised pressure trace for a hydrocarbon explosion  Issue 1. O verpressure (mbar) 2000 1750 Maximum overpressure (> 1 ms duration) (PMAX ) 1500 1000 750 500 250 10 % of maximum overpressure 320 340 350 tr td 380 400 420 Time (ms) Figure 20 . The usual method is idealization of the pressure time history into a triangular form with a positive phase only.Shock spectra  (elastic response) 6. D 1.2 0.4 0 0 0.0 Half sine wave 1.6 Triangular D y namic magnific ation factor.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 2.
This can then be translated into transient dynamic response as shown in Figure 21.8. The use of Figure 19 is only applicable when elastic response is expected and the structure does not yield locally. For plastic deformation a time history simulation may be used to estimate the DAF including plastic response using Biggs’ method described in 6. γ. These topics are under investigation and will be dealt with in detail in Part 3 of the Guidance. This form is not characteristic of far field blast waves outside a compartment. the natural periods are likely to be longer with DAF’s less than one. should be calculated for all loaded members such that the equivalent static load. A DAF may then be calculated for this form of response. For example consider an explosion overpressure of 1. Lstatic. A typical trace for this case is shown in Figure 8. as a static load which causes yielding would deform the structure without limit. the use of equivalent static explosion loads is generally acceptable to allow progression of the primary structure design using a global static overpressure. Another technique is to perform a simplified dynamic analysis and find an equivalent static load distribution which gives similar displacements to the peak dynamic response. A Dynamic Amplification Factor (DAF). by: Lstatic = γ Lpeak For single structural components γ may be as much as 2. Lpeak. A triangular profile may be appropriate as shown in Figure 20.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions For determination of deck response to an explosion from below.1. An equivalent static load may be obtained from consideration of the modes of response of a MDOF structure and by choosing the mode (and modal period) corresponding closely to the expected shape of the response if this is available.5 bar with a duration of 150 ms as the design load condition.3.3 Equivalent static loads During early design stages. for typical large structures such as module and topside structures. Reference . The single peak idealization shown in Figure 20 will not apply when external explosions are involved or when there are two paths to the observation point or in some cases of explosions in the moonpools of FPSOs. is related to the peak dynamic load. 6. There are also concerns that the sharp changes in slope in the idealization introduce spurious frequencies into the exciting force and that it is preferable to use the original pressure traces from CFD or use some spectral approach. The DAF obtained for an extended structure and the appropriate design load may not be applicable for member or local checks. For plastic deformations the dynamic amplification factor has a restricted meaning. 126 Issue 1. discusses the accuracy of various methods of idealization of pressure pulses. October 2003 . This only applies for elastic response but may enable code checks to be performed using conventional software. It is important that consideration is given to the impact of the dynamics on these loads. it is important to represent the negative phase of the loading to enable the calculation of rebound.
This equates to static overpressure loads between 1.0020 Dynamic Response Displacement (m) 0.000 0.350 0.0005 0.400 Time (second) Figure 21 .0015 0. The equivalent static load is that would cause similar response to the blast impulse.68 bar and 2.150 0.250 0.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Transient Dynamic SDOF Response 0. Further detailed information may be found in References  to .050 0.300 0. For a scenario with load duration 50 ms or for a member with natural period of 150 ms the corresponding DAF would be about 1.0010 0.1 Material properties for explosion response General The expected large deformations resulting from and explosion and the dynamic nature of explosion loading means that strain rate and strain hardening of the material may occur locally. The method could be refined to take account of the variation of impulse and duration with peak overpressure given in sub-clause 5. only a brief summary is presented here.100 0. 6. October 2003 127 .200 0.125 for a scenario with a duration of 150 ms and for an element with a natural period of 50 ms. The treatment of material properties in Reference  is at a level more appropriate to Part 3 of the guidance.85 bar for the static design load conditions.9.0025 Static Response 0.4 6.8. Issue 1.0000 0.4.Sample SDOF dynamic response The example presented here indicates that a dynamic amplification factor of 1.
is followed by a phase of strain hardening at a continuously diminishing rate until reaching the ultimate stress-strain point. and partly to allow for variations not explicitly covered in the geometric and physical description of the structure. The reason for this is partly to allow for deficiencies in the adopted analytical modelling tools. if detailed and precise material data is available then potentially a more economical design can be achieved.1 Static material properties General In a design or assessment project.4. such as those described in BS EN 10002-1:2001 is typically characterised by four distinct phases: • from the onset of loading. The yield plateau. when designing the individual structural members it is safe to ignore the upper yield stress. The quasi-static stress-strain curve for a typical structural steel specimen tested in accordance with standard tensile test procedures. In contrast. It should be emphasised that when performing a structural analysis the materials are often assumed to have a limiting strain well below their fracture strain.000 N/mm2. However in most design situations the engineer will typically have to rely on generic data specified in the relevant standards. where it remains for a while as the deformations continues to develop. the stress-strain behaviour of the specimen is nearly linear elastic with a modulus of elasticity of about 205. It is interesting that according to the guidelines given in the EC3 Part 1.2.4. the tensile strength and the rupture elongation of the steels.2 draft code for structural fire design the ultimate strain of structural steel can irrespective of the temperature be set to 15 %. which typically terminates at strains at least 10 times larger than the initial yield strain. According to the IGNs . it follows a rather sudden stress drop down to a lower yield stress. hereafter. 128 Issue 1.2 6. the limiting strain can be set to 5 % for a member in tension as well as for a member in bending or compression with a plastic cross-section. until strain hardening to rupture. for the purpose of designing connections and supports against dynamic loading. As an example the influence of specified tolerable variations in the dimension and mass of the structural components as well their alignment are seldom incorporated into the analytical model. • • • With respect to the upper yield it is known that stress concentrations and welding both have a tendency to reduce if not remove it. except by the most sophisticated and detailed of the analytical tools. In the special case of modifications to existing platforms (brown-field projects) mill certificates for the employed steels should be available. As a minimum these certificates will contain useful experimental results obtained from tensile coupon tests on the yield strength. ignoring the upper yield stress could prove unsafe. October 2003 .UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 6. Also local buckling of compression elements and localised necking of tension elements are in general not automatically captured. until reaching the upper yield stress. However.
2 % proof strength Rp0. As is the case for carbon steels.2 % proof strength. the actual yield strength. For the experimental test series described in Reference  the measured proof strengths were in average observed to be 28 % larger than the specified minima. S420 and S460.2 Stainless steels Compared to carbon steel the stress-strain curve for stainless steels are characterised by a smooth rounded response with no definite yield point.2.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions The specific and current mechanical requirements to structural steels to be used in the fabrication of fixed offshore structures are given in material standards . October 2003 129 . Another major difference in the mechanical performance at ambient temperature between the two types of steel is the improved ductility of stainless steel. where S0 is the original cross-sectional area of the test piece.65√S0 . It should be emphasised that the actual yield strength of steel frequently is significantly larger than its guaranteed minimum value. S420 and S460 steel are listed in Table 6.4. These steels are all suitable for installations in the North Sea sector.2 Mechanical properties typically specified for structural offshore steels Name S355 S420 S460 Reh N/mm2 355 420 460 Rm N/mm2 470 – 630 500 – 660 530 – 720 Af % 22 19 17 It should be noticed that the minimum yield strength Reh refers to the upper yield strength.3 lists typical values for the mechanical properties of various stainless steels as specified in Reference . which has superseded the recently withdrawn BS 7191: 1989. is usually significantly larger than the minimum specified in the standard. Issue 1. 6. with an average of about 13 %. In situations where the yield phenomenon is absent the yield strength should replace this parameter with the 0. and the ductility in terms of a minimum rupture strain Af. Table 6. In the case of the high quality offshore steels tested in Reference  the upper yield strengths were measured to be up to 26 % larger than their guaranteed value. the yield strength in the work hardened zones need to be adjusted upwards.2. This standard. Table 6. Furthermore. the mechanical requirements given in the table refer to material in the annealed condition. here taken as the 0. specifies the requirements to variations of the steel grades S355. Such differences between the guaranteed and the actual yield strength need to be accounted when estimating the magnitude of the forces transferred through the joint details into the supporting structure. Typical limiting values for some of the mechanical properties specified in EN 10225 for grade S355. The rupture strain is measured over a gauge length of 5. The requirements to the tensile strength Rm is given in terms of an upper and a lower limit.2. Thus in situations where the structural elements will be subjected to some degree of cold forming without a following heat treatment.
4362 (2304) 1. 130 Issue 1. Strain rate sensitivity is traditionally expressed in terms of the Dynamic Increase Factor (DIF).3 Strain Rate Effects It is well known that the stress-strain characteristics of steel depend on the rate of straining. For a given strain rate and a given material property the DIF is defined as the ratio between the value of the material property measured under dynamic and quasi static loading conditions. One of the most used models to describe the rate effects on the yield strength is the Cowper Symond’s model: f yd fy 1/ q ⎛ ε⎞ = 1+ ⎜ ⎟ ⎝ D⎠ Table 6.0 N/mm2 260 260 Rm N/mm2 520 – 670 520 – 670 630 – 800 640 – 840 Af % 45 45 25 25 Strain hardening should always be accounted for when designing for the dynamic reaction forces.4401 (316) 1.5x10-3 s-1 . an increase in the strain rate increases the stresses corresponding to given strains.2 N/mm2 220 220 400 460 Rp1. Although an increase in the rate of straining reduces the fracture toughness of steel it is unlikely that the strain rates associated with hydrocarbon explosions will significantly reduce the fracture toughness of the high quality structural steels used for offshore installations.4462 (2205) Rp0. The more precise effect that high strain rates have on the stress-strain characteristics of a given steel quality is a complicated function of its temperature. 6.3 Mechanical properties typically specified for stainless steels Name 1. However in general it is the case that the more the steel can be strengthened by heat treatment or cold working the more sensitive its mechanical behaviour is to an increase in the strain rates.4 [111. whereas it safely can be ignored when designing section sizes. and that. at least until the onset of strain softening. It is also generally accepted that the strain rate effects are the same whether the material is loaded in tension or compression.1] lists the coefficients D and q recommended for calculating the dynamic yield strength of various structural steels.4.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Table 6.4404 (316L) 1. chemical composition and internal microstructure. strain rates of 2. and that the Dynamic Increase Factors are isotropic in their nature. and that the strain rate sensitivity decreases with an increase in the plastic strain.As a consequence the tested value will typically be about 8 % larger than the long term yield strength. In general the dynamic testing of steel specimens has shown that the elastic modulus is unaffected by strain rates. October 2003 .
4 100 240 3489 5958 q 5 10 4. Likewise it is assumed that all conclusions regarding the uniaxial stress state can be generalised to the multiaxial stress state. For typical steel structures.1 1 10 100 1000 strain rate s-1 Figure 22 . However the risk of employing the Cowper Symond’s formulation to fit experimental strength data is the loss of some accuracy for the strain rates which are most relevant to gas explosion scenarios.2 1.74 5.4 1.001 0.8 1E-006 1E-005 0. This compares with the strain rates of 103 s-1 frequently encountered in metal working processes and hard impact situations. October 2003 131 .36 Figure 22 compares various empirical models with experimental data. It should be mentioned that the strain rates vary both spatially and temporarily within a structure. The experimental data for stainless steels were taken from Reference .4 Coefficients D and q for dynamic yield strength Material Mild steel Stainless steel (304) Stainless steel (316L Stainless steel (2304) Stainless steel (2205) D s-1 40.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Table 6. the strain rates associated with hydrocarbon explosions will seldom exceed 1 s-1. Thus when performing a non-linear finite element analysis it is necessary to know the complete stress-strain behaviour at all encountered rates. For strain rates less than about 1x103 s-1.01 0. an essentially linear relationship between the stress and the logarithmic strain exists.0001 0.3 1. 1.9 0. and for high quality carbon steels suitable for fixed offshore installations from Reference .6 1.5 Test data for BS 7191 grade 355 and 450 steels Upper Yield Strength ratio Lower Yield Strength ratio Mild Steel 316L 2304 2205 Dynamic increase factor 1.Strain rate effects on typical offshore steels Issue 1.0 0.1 1.77 6.
For a manned platform a Temporary Refuge (TR) or Safe Mustering Area must be available to protect those not in the immediate vicinity of an explosion and to survive the event without injury. October 2003 . at least one escape route must be available after the event for all survivors. screening or performance criteria. For the ductility level blast (DLB). In an explosion event. • • Tension members Members in axial compression or in bending and axial compression whose slenderness ratio does not exceed 15 and the cross-sections comply with the criteria of compact section Members in bending whose cross-sections comply with the criteria of compact section and whose slenderness ratio (λ) does not exceed certain values. necessary to demonstrate that the strains are high enough to mobilize the benefits of the strain hardening. Fire and Explosion Hazard Management.5. This effect should not be considered when designing supports against reaction forces.31]. defined as those elements critical to safety [30. person or procedure and which is used as the basis for managing the hazard. allowing for the possibility of helping injured colleagues. • 6. The following groups of structural members may sustain high plastic strain without losing strength from local or overall buckling. It is. the primary structure should not collapse with escape possible from safe areas after the event.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 6. High level performance standards have been given in Part 0. should have fulfilled their function or remain operational. however. Plastic deformation of the structure is acceptable provided collapse does not occur and barriers remain in-place and are able to resist any subsequent fires. The frequency with which accidental events.4. The ability of the structure to satisfy these requirements will depend on its ability to respond in a ductile manner and the ability of the barrier connection details to respond without rupture. they are reproduced below.4 Strain hardening Strain hardening may also be taken into account for tension members and plastic sections by taking design strength to be the ultimate strength divided by 1.25  . 132 Issue 1. Safety critical elements. The required endurance time is the estimated time for people to travel from their work stations to the TR. Performance standards may also be referred to as acceptance.5 6.1 Structural performance standards Introduction A performance standard is a statement which can be expressed in qualitative or quantitative terms of the performance required of a system. then to the primary and secondary means of escape. item of equipment. will result in loss of TR integrity within the required endurance time will not exceed 10-3 per year . from all causes.
in order to prevent an escalation of the event. equipment or functions) requiring detailed assessment are classified into three levels of criticality with respect to the explosion hazard as below. with potential for inventories outside the module contributing to a fire due to blowdown and or pipework damage. availability and survivability. In addition to the high level Performance Standards above it will be necessary to define measurable performance standards for specific key items or systems relating to the systems’ functionality. plant. are: • • • • Strength Deformation limits Local and Global ductility Rupture Issue 1. October 2003 133 . Performance standard – These items have no functional significance in an explosion event and must not become or generate projectiles. Criticality 2 Items whose failure could lead to major hydrocarbon release and escalation affecting more than one module or compartment.g. It is suggested that the number of SCEs (systems. These are sometimes referred to as low level performance standards. as is any which is intended to prevent or limit the effect of a major accident. Criticality 1 Items whose failure would lead direct impairment of the TR or emergency escape and rescue (EER) systems including the associated supporting structure. but overall collapse is prevented. The limits specified for the structural elements should not exceed those required for the supported or adjacent equipment to function adequately. system (including computer software) or component part whose failure could cause or contribute substantially to a major accident is safety-critical.2 Criticality categories for SCE’s It is helpful to consider a hierarchical approach to the identification of SCEs. As is common in the design of structural systems under other forms of accidental loading. as determined by the location of pipes and equipment and to the performance of these items. e. performance criteria for the DLB are set such that permanent deformations can occur. earthquakes. 6. Performance levels for structural elements. In addition to this is the requirement that the explosion must be contained. any structure. Performance standard – These items must have no functional significance in an explosion event and these items and their supports must respond elastically under the strength level blast (SLB) Criticality 3 Items whose failure in an explosion may result in module wide escalation.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions A Safety-critical element (SCE) is defined as. equipment. Performance standard – These items must not fail during the DLB or SLB. that are normally considered. Consideration needs to be given to limits of deformation. ductile response of the support structure is allowed during the DLB.5. These are referred to in this document as ‘element specific performance standards’ used in the evaluation stage of the explosion assessment.
3 Deformation Limits Deformation limits for structural elements are normally specified in terms of deflections. It should be noted that such levels of ductility will have design implications for members containing plastic hinges to ensure the ductility levels can be achieved. The performance standard relating to strength of members is more suitable for the strength level analysis. strength level analysis for the SLB and an ultimate capacity analysis or ductility level analysis for the DLB. deformation limits are normally adopted. N-004  gives the range of allowable values of µ depending on the section classification assuming no axial restraint to the beam and the boundary conditions. This is normally achieved by providing lateral restraint at hinge locations to the unbraced compression flange (Part 3 of the Guidance Document will cover these details). At the ductility level.5. as this is predominantly a stress based check using static design codes. 134 Issue 1.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions • Global collapse This guidance document recommends two levels of structural assessment. This is a global deformation parameter. This is given in Table 6. October 2003 . which is commonly used in a Single Degree of Freedom (SDOF) model. Norsok standard Annex A. Table 6.5 Allowable ductility values µ  Boundary Conditions Cantilevered Load Concentrated Distributed Concentrated Distributed Concentrated Distributed Cross-section category Class 1 6 7 6 12 6 4 Class 2 4 5 4 8 4 3 Class 3 2 2 2 2 2 2 Pinned Fixed Consideration of attached fire protection may impose a lower limiting strain or deflection. The displacement of the component must also be limited to prevent impacting SCEs and equipment items. ultimate capacity check the yield stress limit will be exceeded and member stress is not appropriate for judging member response. A local ductility parameter may also be defined in terms of the support rotation. For a class 1 section. For the DLB. Reference  allows a limiting support rotation of 2 degrees and a ductility factor of 10 (whichever governs) as reasonable estimates for absolute values to ensure safety for personnel and equipment and consistent with maintaining structural integrity in the inelastic range.5. 6. often given as a ductility ratio µ defined as the ratio of the maximum displacement of the element to the deflection required to cause first yield at the extreme fibres.
NLFEA is likely to be required to verify the design. It should be noted that variations in strain are quite common depending on the modelling assumptions used and care in interpreting these values at critical points is required. the philosophy for the explosion design cases and corresponding performance standards need to be established. The Norsok Standard  highlights a number of factors which influence the strain values such as material toughness. The extreme loading applied to a structure during an earthquake is often compared with an explosion event on a topside due to the small probability of occurrence. Screening of various options for blast walls or potential behaviour of walls and decks under a number of explosion scenarios requires cost effective. the structure or equipment should remain undamaged. Once the design has matured and safety critical elements of the structure have been identified. presence of defects. Strains are normally adopted as a failure criterion when using finite element models to judge response. Current guidance on allowable strain values are to set the critical percentage strain limit at 5 % at weld locations and 15 % in the parent material away from welded details. For a frequent earthquake with low seismic forces. It is also important to bear in mind that simple models may not give a satisfactory design for large overpressures as many of these tend to rely on bending capacity from a static analysis with some form of load factor to account for the dynamic response. 6.6. For earthquake design it is common to consider two load cases for the design of structures and equipment: 1. it may be important to obtain the ultimate capacity at an early stage which may require a more refined analysis at the outset. reliable and accessible solutions. October 2003 135 .1 Structural assessment Introduction The degree of complexity adopted in establishing the response of the topside requires considerable engineering judgement at an early stage in order to avoid unnecessary delays in the design process. For retrofit systems.5. 6. This will assist in determining the degree of retrofit required although other factors such as shutdown period if hot work is required and space available will also influence the options available. Issue 1.6. Also.6 6. These values refer to average local strain values. However it is important to be aware that strains at this level are quite sensitive to a number of parameters and tests show a significant scatter of strains at fracture. which if exceeded would cause rupture. In many cases this may mean that complex Non-Linear Finite Element Analysis (NLFEA) is unlikely to be an option in the early stages of the design cycle unless planned well in advance.4 Rupture Structural integrity is often assessed using strain limits. many of the concepts used in design for earthquakes such as attention to detail to connections and ductility requirements apply equally to topside explosion response.2 Design Criteria In order to decide on the level of assessment that is required for various parts of the structure and at different stages of the design process. strain rate and presence of strain concentrations. For the structure this is essentially a serviceability check and must remain elastic.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 6.
Some local rupture of structural elements can be tolerated if it can be shown that progressive collapse will not occur. October 2003 .UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 2.34 bar. For a strong rare event. equipment and piping. 3. two levels of explosion loads are recommended for explosion assessment. The strength level blast (SLB) and the ductility level blast (DLB) obtained from consideration of exceedance curves or from the use of derived nominal overpressures as discussed in sub-clause 5.4 Ductility Level Blast (DLB) At this load level. 6.definition In blast mitigation applications the prime objective is often energy absorption and not static strength. –. Undesirable modes of failure are: • • Those resulting in an overall collapse of the structure Those involving sudden brittle failure such as local buckling By analogy with the approach used to assess structures for their ability to resist earthquakes.3 Strength Level Blast (SLB) One aim of this approach is the use of existing design codes which are predominantly based on static design. It is important to be aware of limitations of this approach: 1. Design codes are based on static design and not dynamic loads. The use of simple models will allow a rapid assessment of the suitability of the layout of the primary steelwork and the screening of a number of pressure time histories to establish the bounding case from the loading. care needs to be taken to ensure that they remain intact to prevent an escalation of the event. although limiting deformations need to be set in some parts of the structure which may be influenced by plant.5 – 0. Membrane forces may occur in some elements which may not have be accounted for in some code checks 2. This is particularly important when checking members in compression as buckling is a time dependent phenomenon which may invalidate some of the static code checks. It should not be less than the 0. which is given in the building code  as a minimum lateral pressure to be resisted by walls in order to achieve a minimum level of robustness in the connection details. 136 Issue 1.7.6 bar. This is likely to involve simple numerical models such as the Bigg’s SDOF idealization  or a static finite element analysis for the primary steelwork. In general. the structure can sustain damage but overall collapse should not occur. with enhanced factors to allow for the material properties due to strain rate effects and differences between guaranteed minimum yield strength and coupon tests.6. increasing the stiffness of a structure can result in local stiff points which attract high loads and lead to brittle failure modes which may compromise the integrity of the whole system. The SLB may be typically in the region of 0. It is essential to allow for the enhanced reaction forces at connection details due to the enhanced yield stresses which may occur.6. particularly for the extreme low probability scenarios. 6. Performance standards based on design strength are adopted for assessment with this loading level. permanent deformations due to yielding are acceptable. For parts of the structure which need to contain the blast.
The validity of this method will depend on the severity of other load cases which have been used in the original design of the structure.10.115. Components which were subjected directly to the blast. which is crucial to the survivability of the structure under the extreme load condition. 6. in order to assess the possibility of a global collapse mechanism occurring. Analysis of this level of deformation inevitably requires the use of more complex analytical tools. Uncertainties still exist at such large strain values. unlikely that the differing levels of response to dynamic loads at the same peak level as determined by the natural periods of the target structural elements will be adequately represented without undue conservatism. It is also important to capture the effects of interaction of the structural elements as a system. For the primary structure. The definition of the SLB in this guidance may also serve to perform this check for the DLB with the more stringent requirement that the stresses in the primary structure remain below yield.5.6. This approach has largely been superseded by the modified code check method described in sub-clause 6. there will come a point where the incidence of failures rapidly starts to increase and begins to take in the majority of the members. The primary system is defined as those elements which have blast loads transmitted to them from the secondary system and consists predominantly the primary structure typically consisting of plate girders.6 The simple demonstration of ALARP One method of the demonstration of ALARP using a strength level analysis is to apply a static pressure load to the structure and observe. the code check results in members dimensioned to resist the worse credible event or ductility level explosion. however. deck plating including stiffeners and cladding are defined as the secondary system. columns and trusses. and care needs to be taken when interpreting the results of non-linear analyses. Design to this equivalent static pressure could then be said to be ALARP.7 General remarks on structural response Yasseri  recently published an article in the FABIG newsletter where a useful hierarchy of structural systems was described.6. 6. 6.5 Dimensioning Explosion Dimensioning explosion loads[89. At this point it may be argued that it would be unreasonable to strengthen or change the member properties as it would impact on members designed by the other load cases. Joints. when member failures occur. Issue 1. It is. The variability of pressure in the explosion load cases is also not represented in this method. such as walls.116] are of such a magnitude that when they are applied to a simple elastic analysis model. panels. beams. This will be predominantly the blast walls and plating. The adequacy of components is checked against performance criteria based on maximum deformation and strain rather than strength. If the pressure is then ramped up in stages.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions In some cases a maximum deflection limit may be specified to avoid compromising supports to vital equipment or pipework. which may lead to an escalation of the event. which if allowed to deflect will reduce the loads transmitted to the primary framing. such as non-linear finite element techniques which are capable of modelling large deformation and strain. it is necessary to ensure that it remains predominantly elastic in the overall failure mode and parts of the structure that are allowed to yield are detailed for high ductility. through code checks. which are inevitable in many instances. October 2003 137 . barriers and connections still need to be checked against the DLB directly preferably using a non-linear elastic plastic method of analysis.6.
UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions
This places emphasis on the connection details, which need to be assessed for large rotations in order to achieve a robust structure. The overall aim of the philosophy in this clause is to ensure that the concept of robustness is built in at an early stage. At the strength level blast, the structure should remain elastic and the structure should have enough ductile capacity to survive the ductility level blast load case.
Response prediction methods
Several techniques of varying complexity are available for predicting the blast response of topsides. These range from simple hand calculations and graphical solutions to more complex 3-Dimensional computer models capable of modelling geometric and material non-linearity, tearing of welded connections and contact with other parts of the structure or plant and equipment. However, with all of the models available, a considerable degree of engineering judgment and experience is required in order to convert the real structure into the idealised model. The general philosophy is to start with the simplest methods (ensuring a conservative approach) and if failure is indicated, proceed to more sophisticated methods of analysis. The three main levels of analysis are: 1. 2. 3. Screening analysis Strength level analysis Ductility level analysis
These are described below in sub-clauses 6.7.1 to 6.7.3. The following may enable simplifications to be made in the analysis method. Non load-bearing elements, typically panels, fire walls, and blast panels may be checked in isolation. The stresses in panels and cladding usually dominate the stresses due to frame movement. Panels and some blast wall systems may be conservatively idealized as one way spans. Resistance displacement curves may often be determined statically, dynamic response may then be determined using a modified Biggs’  method. Dynamics of simple structures may be represented with knowledge of dominant natural period. Biggs’ method extensions are available for loaded components  . It is preferable to model the primary structure response to the SLB using an elastic frame model as single member assessment does not take into account the transfer of loads between connected loaded surfaces and any co-existent static loads. It is preferable not to include nonstructural cladding and plates in the model as their response is mainly in membrane action and their shear strength may be overestimated.
Issue 1, October 2003
UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions
Screening analysis for an existing installation consists of condition assessment which may involve a survey followed by design basis checks. The transfer of conclusions and load characteristics from the analysis of a similar platform is acceptable for this and for Strength level and ductility level analyses.
Design basis checks
Design basis checks consist of checking the basis of design for the installation and determining if the methods used for the design are acceptable in the context of the fire and explosion events considered. If the structure and appurtenances have been checked for a Safety level or Ductility level earthquake loads following API RP 2A  (9th Edition sub-clause 2.3.6e.2 or later) then the ‘strong vibration’ response to explosions need not be checked.
Component checks may be employed if the component is non-load bearing in the operational condition or if the component does not form part of the main framing. Methods of dynamic response assessment such as Biggs method  may be used as described in sub-clause 6.8.2. Where loads from connected structures are represented component check methods may be employed. There are however many limitations on the method which are discussed in subclause 6.8.3. A major consideration in explosion response is that deflections of the structure must be limited to allow escape. Typically the deflections into escape ways should be limited to about 150mm. The allowable deflection into equipment spaces will depend on the clearances to equipment. The deflections of the Primary structure will normally be satisfactory if it passes the normal strength or utilization checks. Blast and firewalls are however designed to deflect to exploit the ductility of these items and so deflection checks for these items will be necessary. Buckling checks must also be performed to ensure that the full plastic capacity of a member can develop.
Strength level analysis
The integrity of an offshore structure may be checked using a linear, elastic, beam model. It is conservative not to include the restraint from the cladding or barriers in the primary frame computer model. In some circumstances it is advisable not to include the cladding as plate elements as the shear restraint from these elements will be overestimated in a small deflection elastic model. This applies particularly if an external explosion load is considered which puts the side walls with respect to the pressure into shear. In a dynamic analysis the masses and their distribution should be included in the model. In checking the primary structure it is conservative to include the loads from barriers but to ignore their strength contribution.
Issue 1, October 2003
UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions
The response of cladding panels and plates that form part of the primary structure may be analyzed in detail using finite element analysis assuming the supporting beams are fixed at the main nodes of the structure. The justification for this is that the stresses in the panel are dominated by local response of the panel out of plane and that the stresses induced by the deflection of the main framing are comparatively small. If necessary this assumption can be checked by application of prescribed displacements to the edge of the panel corresponding to the frame response. Buckling checks must be performed to ensure that the full plastic capacity of a member can develop. These checks should be made particularly for deck beams loaded from an explosion below as flanges usually in tension may be in compression during the explosion. These checks should ideally be performed using the loads for the DLB and SLB load cases. Deck beams loaded by an explosion below should also be checked for re-bound effects.
The inclusion of static loads
An equivalent static load as described in sub-clause 6.3.3, will need to be defined for application of explosion loads as a static load case. The dead loads in the structure need to be combined with a realistic estimate of live or operating loads to perform a strength level analysis treated as a design load case. Environmental loads need not be included. API  recommends that 75 % of the live loads are included in the combination for the earthquake analysis. It is recommended that the same proportion of live loads are taken for the explosion load case. The loads from an existing fatigue analysis load case may also be suitable.
Code checks for Strength level analyses
Blast resistant structures and their supports are designed to respond in a ductile way to explosion loading mainly in bending. The shear behaviour of structures at their supports and in the joints of the primary frame will also affect the utilization factors derived in the code checks. The safety factors implicit in the code checks will be different from those associated with bending. Shear strains at supports should generally be limited to the elastic limit. Additionally, for a dynamic situation these shear forces will be quite different from the values obtained by static analysis as a result of the inertia forces acting on the structure. The benefits of this effect are exploited in blast wall design, as the support or reaction loads may be less than the applied overpressure load. Plastic deformation of a blast wall also has the effect of limiting the reaction loads on the supports. For a loading duration near the natural period of the member the shear forces could be higher. Checks should be made where this is likely using a dynamic amplification factor based on the ratio of load duration to natural period (see Figure 19). Code checks may be re-interpreted to take account of the following inherent reserves of strength, both material and plastic reserves as discussed in sub-clause 6.4.3. Material effects are discussed in sub-clause 6.4.
Issue 1, October 2003
UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions
Ductility level analysis
220.127.116.11 Interpretation of the Output from a Non-Linear Finite Element Frame Analysis
For most ductility level analyses code checking will either not be appropriate or the response simulation software will not contain a code check module within the software. Checking of members will be done explicitly with regard to performance standards, which may take the following forms.
Strength limit checks similar to code checks; failure is defined to occur when the design load or load effects exceed the design strength. This criterion may be applied in the plastic region. Deformation limit checks. Permanent deformation may be acceptable so long as safety critical equipment is not impinged upon and collapse is not caused even in the presence of a fire. Mechanisms may be formed momentarily during an explosion. Buckling checks, which identify where plastic response may be limited by local buckling. Fracture checks to identify weld and member failures.
Members may be classified as plastic, compact or non-compact (BS5950). Plastic and compact members will generally reach their full plastic capacity before buckling. If the structural response software is capable of representing finite displacement effects, plates may be included in the model to represent barriers and loaded surfaces. The inclusion of plates with equivalent thickness to represent mid point deflection will also help to represent the tension and shear effects from these items. The restraining effect of cladding can conservatively be omitted from the computer model so long as the loads applied to them are applied to the bounding members according to the area associated with each one. Some packages will not take account of the loss of shear restraint from the cladding as it is deformed; in this case it is preferable not to include the cladding as plates in the model.
Deformation limits are discussed in sub-clause 6.5.3. Additional Guidance is given below. Compact sections may be designed using code checks without supplementary local buckling checks up to first hinge formation. Sub-clause 7.2 of Reference  gives formulae for checking for local buckling of beams working beyond the elastic limit. Checking of columns under axial compression is dealt with in Reference . In most situations blast walls are arranged to span from floor to ceiling without direct support from the columns. Isolated columns only receive load from dynamic pressures and could be in tension during an explosion. Allowable ductility ratios based on earthquake design practice are available for I sections, box sections and circular sections from Reference .
Issue 1, October 2003
A more detailed appraisal is given in Reference . or non-linear spring. 142 Issue 1. October 2003 .UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Buckling checks must be performed to ensure that the full plastic capacity of a member can develop. 6.8.3 Joint Design The principal ultimate failure mechanism for joints is rupture or brittle fracture unless the joint is stronger than the members which are connected to it. The magnitude of the mass and the stiffness of the spring are determined such that the displacements in the SDOF model are the same as a characteristic displacement of the real system. These are systems where the overall response may be represented by a characteristic displacement and the deflected shape is similar to the first or lowest mode of vibration of the system. These checks should be made particularly for deck beams loaded from an explosion below as flanges usually in tension may be in compression during the explosion.1 Single Degree of Freedom Idealisations General The basic analytical model used in many blast design applications is the Single Degree of Freedom (SDOF) system.2. Much of the guidance developed in the past such as Reference  is based on the Biggs method  Other methods based on an energy balance to obtain isodamage curves developed by Baker  are also adopted and the basis and limitations of these techniques are described in this sub-clause. which can be easily simplified to a single mass and spring. A detailed treatment of the Biggs method will be given in Part 3 of the Guidance. Large deflection effects such as web crushing and flange curling are dealt with in Reference . Single Degree of Freedom (SDOF) methods represent a structural system as a single mass whose motion is resisted by a single linear. 6. usually taken to be a point at mid-span.8.2 6.7.8 6. The basic design principle is to design them so that loading or imposed rotation causes ductile deformation of a connecting member. The SDOF method is limited to structural systems.1 Component Response – Biggs Method General Components may be analysed in isolation as long as the interaction with the surrounding structure through fixity and the applied loads are negligible or are represented in the component model  .8. These checks should be performed using the dimensioning explosion load level.4. 6. Reference  gives criteria for the prevention of lateral torsional buckling based on slenderness ratios for beams. This follows general practice in earthquake design.
This procedure is illustrated in Figure 23. In practice. Typical member response is shown in Figure 21. Methods of calculation of mass KM and stiffness KL transformation factors for beams and panels are given in References  and . The Biggs method requires two basic inputs. Member damping is not usually modelled in this analysis. the resistance displacement curve and an idealized loading time history. which is inaccurate for fully or partially fixed members as well as for members where tension effects are significant. xmax. and the ratio of peak overpressure load to component ultimate plastic resistance. given the load duration to natural period ratio. to allow the modelling of changes in structural behaviour at certain displacements. The equations of motion can be modified during the time-stepping process.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions The SDOF method can easily be programmed onto a computer. October 2003 143 . where the resistance-displacement relationship is complex. This is particularly useful in the analysis of blast panels. td/T. Explicit solutions of the equations of motion do not have to be found. as well as the loads applied to it. If the deflected shape of a member changes as it is loaded (for example if plastic deformation occurs) the transformation factors must be adjusted. hence realistically modelling the increased displacements that occur after plasticity has been reached. For example. Further loading results in no increase in the resistance of the member – it is assumed that the member is deforming purely plastically. The charts are based on a simplified bi-linear resistance behaviour. the change in structural stiffness caused by plastic deformation can be introduced when the yield displacement of a beam has been exceeded. are modified using transformation factors which are calculated taking into account the deflected shape and the loading and mass distributions [1. The mass and stiffness of a structural member must be carefully represented in the SDOF model to ensure that the model displacements accurately represent the displacements of the actual member. which can calculate the displacements at each time increment.101]. and hence it is relatively simple to use non-linear force and resistance inputs without significantly increasing the complexity of the program. and hence the member oscillates freely after the blast load has been released. which is important. Under increasing uniform loading. A typical idealized resistance-displacement curve for a simply supported beam is shown in Figure 23. and in any case it is the maximum displacement rather than subsequent oscillations. If the displacement reduces after plastic deformation then it is assumed that the resistance of the member returns to the preplastic (or elastic) form on a line parallel to the line representing the initial elastic deflection. The values of the transformation factors are dependent on the deflected shape of the member under the blast load. Issue 1.101] for the calculation of the peak response. The natural period of the member is given by: T = 2π where Me ke Me = Actual member mass M times KM. the member deflects elastically up to its yield displacement xe. as it is generally small for structural members oscillating in air. the stiffness or load transformation factor. damping is not important. The mass and stiffness of the member. Design charts are available [1. Fm/Rm. Ke = Actual member stiffness k times KL. the mass transformation factor.
3 0.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Pm F e( t) 1000 FORCE 800 600 400 200 0 0 R RE S I S TANCE R m 1000 2 800 0.6 TIME(s) td LOA D DURA TION 600 1 3 400 S LOP E K e 200 Xe 0 Xm DIS PL ACEM ENT X M F(t) B LAST WALL Figure 23 . October 2003 .4 0.5 0.1 0.2 0.Biggs’ single degree of freedom model  144 Issue 1.
The Biggs charts use an approximation of member stiffness to give a bi-linear resistance curve for all fixity conditions. A range of values of Fm/Rm and td/T were used to generate the sets of curves shown in the Biggs’ design chart Figure 24 for a simply supported member under a triangular load with a rise time equal to half the load duration.8. October 2003 145 . The bi-linear resistance – displacement relationship used in the Biggs’ design charts is only accurate for simply-supported members. where the member reacts either purely elastically or purely plastically. and hence the energy or work done to a given deflection is the same for the two curves. Similar charts for differing rise times of triangular loading are given in References  and . The stiffness used in the Biggs method is determined such that the area under the ‘actual’ and ‘idealized’ curve are equal. For partially or fully fixed members.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 6.2. elastic/plastic behaviour occurs before the member reaches its ultimate plastic resistance.2 Representation of Fixity Conditions The degree of rotational and axial fixity at the ends of a structural member affects its elastic stiffness and hence its natural period. Issue 1.
5 0.5 F(t) F1 R 1.1 0.2 0.3 0.2 0.5 t d t d Triangular pulse load 0.2 t 0.4 0.8 1 y t y el tm Resistance Displacement function function 2 5 8 t d /T Figure 24 .60 2.Biggs' design chart – overpressure rise time equals half load duration This chart enables the peak response of the member to be calculated.0 0.8 10 8 y m /y el = µ 5 0.1 0. which is the number of multiples of the yield displacement the member will reach under the loading. The vertical axis is in terms of the ductility.8 0.20 1.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 100 80 50 Rm /F1=0.5 0. To use this chart: 146 Issue 1.9 2 1. October 2003 .00 Rm ym 0.00 1.7 20 0.6 0.
This stiffness gives the greatest elastic resistance for the member. A one degree of freedom idealization may be adopted to check the response of deck sections if a Screening Analysis is being performed. The degree of fixity significantly alters the natural period of the member. The same Biggs’ chart may be used to estimate peak response so long as the natural period and resistance function are modified to represent attached masses.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions • • • • • • • Determine the peak resistance Rm. which must be accelerated during blast loading. Determine the effective peak force Fm from the peak force Fpeak and the load factor Kl. These formulae are not accurate for two way spanning panels but are reasonable for the assessment of one way spanning panels or beams. plastic hinges are initially formed at the ends of the member. The transitional stiffness occurs at Ks=6EI/L. Issue 1. Similar charts may be constructed for partially fixed members by applying a rotational spring stiffness.3 Out of Plane Equipment Loads Attached equipment alters the natural period of the member by increasing the mass. (Fm = Kl x Fpeak) Calculate the equivalent mass of the member Me Determine the natural period of the member from the effective mass and stiffness Choose the nearest curve corresponding to the ratio of Fm/Rm Read off the ductility for the peak response corresponding to the td/T for the member Calculate the peak response from the ductility indicated on the vertical axis The reaction loads at the ends of the member may also be estimated using this method although the accuracy may be doubtful as detailed strain information is not represented in the model. 6. The initial deflection due to the out of plane loading effectively changes the position of the origin on the resistance – displacement graph. At high values of fixity. to the ends of the member  . The ability of a member to resist axial loads is reduced if its fixity is reduced. Ks tends to infinity for a fully fixed member and to zero for a simply supported member. Ks. The resistance function may be modified to take into account equipment loads. Increased deflection causes a plastic hinge to be formed in the middle of the member. effective stiffness and yield displacement Xe for the member. a plastic hinge is initially formed in the middle of the member. At low values of fixity. The effect of static out of plane loads is to cause an initial deflection of the member. The buckling load for a pin-ended column is 4 times smaller than that of a perfectly fixed column.2. The natural period for a fully fixed beam is less than half that of a pin-ended member. The deflection will be positive in the case of a floor beam subjected to blast loading from above.8. October 2003 147 . Here I is the second moment of area. with additional deflection causing plastic hinges to occur at the ends until the ultimate plastic resistance is reached. The degree of end fixity for a member is determined from the flexural stiffness of the connecting structure. L is the length of the member and E is the Young’s modulus for the material. and will be negative in the case of a ceiling beam subjected to blast loading from below.
or yield line pattern. 6. axial) loads have 4 effects on the response of structural members: 1. Local and torsional buckling of stiffeners and wall plate may also occur before plastic hinge formation.8. 2. A great deal of effort is usually expended in designing the edge connections of the blast wall so that the reaction loads are transmitted to the supports without damage to the supporting steelwork. stiffener failure. In-plane (i. At the dead and live loads normally applied to structural members the reduction in plastic moment capacity is small. Other formulations for resistance estimation are given in the API Bulletin 2V. The failure mode corresponding to the lowest ultimate resistance is the critical failure mode of the wall.e. Because of the inertia of the wall it is possible to design these connections such that the transmitted shear forces and moments are much less than the peak overpressure force on the wall.4 The Effect of In-Plane Loads Whilst purpose designed blast walls will have no static loads applied to them. creating an initial central deflection. The capacity of a blast or fire wall with stiffened plate construction may be estimated for Screening Analysis purposes using yield line analysis for the plate sections. The dominant effect on response is the p-delta effect. Axial loads reduce the plastic moment capacity of a section. October 2003 . Axial loads change the stiffness and hence the natural period of a member.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 6. Each mode corresponds to a different plastic hinge. The reduction of blast resistance due to eccentric axial loading is modelled by reducing the resistance of the member at the initial deflection. there is often a need to assess the capacity of module walls and columns with in-plane loads present. 148 Issue 1. In particular the finite deflection capacity of a simply supported plate is larger than the classical estimates of the capacity of a clamped plate with the same dimensions and span. The possible failure modes of the wall may include panel failure. This ‘P-delta’ effect is modelled by converting the moment caused by the axial load into a lateral resistance. As an axially loaded member deflects under blast loading an additional moment will be generated by the axial load about the original centreline of the member. Purposebuilt blast walls are usually free standing. Work by Shell indicates that the ultimate capacity of panels may be much larger than formerly thought. 3.2. spanning from floor to ceiling and are not an integral part of the supporting structure.2. 4.8.5 Barriers and Cladding Response Blast walls are usually designed to deform plastically and act predominantly in bending to minimize the reactions on the primary structural members of the platform. and whole wall failure. As blast walls are usually arranged to span from floor to ceiling the direct explosion loads on columns are small these effects may be ignored in a Screening Analysis. which is then subtracted from the original beam resistance An axial load may be applied eccentrically to the member. The stiffness of a member will reduce to zero at the Euler buckling load.
UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions
The calculated resistance will also depend critically on the edge support conditions assumed which may in turn depend on the form of loading. High pressure loading on adjacent panels may give rise to clamped edge conditions between panels even though the panels would not otherwise merit this approach. Tension and membrane effects often indicate an increased resistance but the restraint from the surrounding structure through inertia or stiffness may not be sufficient for these effects to be fully mobilized. Penetrations such as piping and cable runs should be arranged to pass through the wall at the top or bottom of the wall to limit deflections and induced strains in the wall. The penetration should be fire and gas proof. The penetrated section of the wall should be designed to have the same stiffness and strength as those parts of the wall where no penetrations exist. The dynamics of the wall and penetrations may need to be checked if large mass items are attached to the wall. Examples of connection and penetration details will be given in Part 3 of the Guidance. Efforts have been made to include tension and membrane effects into Biggs’ method [98,121,8] with mixed success. A detailed assessment of these will be made in Part 3 of the Guidance.
Rebound must also be considered. This occurs when the load has subsided and the stored elastic energy in the wall is released. This may be taken into account by assuming a reverse load or suction at a level of 1/3 of the original overpressure with the same duration. Biggs method may also be used to determine the rebound response if a suction phase is included in the load time history. Other authors have developed a rebound dynamic amplification factor which may be used to calculate rebound response (116].
Limitations of Biggs Method
The advantage of Biggs’ method is that it is simple and easy to apply once the structural system is idealised into a single degree of freedom system. However, the basic Biggs model is clearly based on simple bi-linear structural behaviour assumptions which may limit its applicability for design of modern lightweight systems. It is important to bear in mind that in the past, guidance which was developed for blast resistance design was based on using structural strength, weight and standoff from the explosion source to control response. Some of these limitations of the technique which need to be considered before adopting the method include the following: 1. It is not an easy task to convert anything other than the simplest structures into equivalent spring mass models. A great deal of engineering judgement is required in determining meaningful models and interpreting the solutions. For example, modules are normally framed for lifting and to give redundant load paths, making the overall structure of the module difficult to describe as a one degree of freedom system. The deformation of structural components such as walls and decks is highly dependent on the degree of restraint provided by the primary steelwork. Although some membrane action can be incorporated into the model, it is difficult to accurately model the stiffness of the boundary. This places severe limitations in trying to characterize the resistance function of the SDOF model when modelling plated structural components.
Issue 1, October 2003
UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions
3. Ductility factors which are used are normally for idealised simple beam models. These may not be appropriate for the actual MDOF structure where different patterns of yielding could give the same overall inelastic peak deflection but the local ductilities may be significantly different. The Biggs method implicitly assumes that the deflected shape under blast loading is normally the same as the static deflected shape. Many multi-degree of freedom systems or systems with complex mass distributions do not respond in this way. This also applies to some stiffened plate blast walls. The loads obtained from a CFD model may indicate that they vary significantly in both space and time requiring some conservative idealisation to simplify them so they can be used with the SDOF model. The method relies on the fact that the mass can be lumped, but deck type structures are likely to have significant mass components from plant and equipment. This spatial variation in the mass may well invalidate the assumed shape function and give nonconservative results. Biggs  does address this problem but often the refinements of his method are not implemented. If the rebound is more important, which may well be the case if the ceiling of a module supporting plant is being considered, care needs to be taken  . The prediction of rupture is very unreliable as there is no detailed representation of the strain distribution within the structure. High ductility deflections are not well represented by the flat portion of the resistance displacement curve representing plastic response.
Although some of the limitations above seem severe, the technique has been successfully used in design situations, particularly where the resistance function has been accurately determined. Reference  shows an example of a stiffened plate which whose resistance function was determined via a static NLFEA and combined with the spring mass model. The deflection time histories compared well with an explicit NLFEA up to pressures of the order of 3 bar. Beyond this, the model was unable to give accurate deflection values. This was thought to be due to the spread of plasticity which results in a change of the shape function used in formulating the spring mass idealization.
Pressure Impulse Diagrams
Rapid assessments of maximum response are very useful in a preliminary structural design, particularly where there are a large number of load cases to screen. The previous Biggs model described how transformation factors can be developed using an energy balance to convert the structural element into a spring mass model. The same concept is adopted in developing pressure impulse diagrams which is described in detail by Baker . For an impulsive load, the kinetic energy imparted to a structure or structural element by a short duration pulse, or external work for a long duration pulse, is determined and the strain energy is then equated to the strain energy the structure develops while deforming.
Issue 1, October 2003
UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions
A suitable shape function is required to describe the deformation under the loading and both elastic and plastic behaviour can be accounted for. The resulting expressions for the two loading regimes are then rearranged such that a non-dimensional plot of pressure against impulse can be constructed. A number of curves are normally plotted for varying levels of damage which may be defined in terms of the ductility parameter µ. This allows rapid assessment of the damage or vulnerability to a number of different loading conditions. The method is capable of representing membrane action in elements such as plate components. The method can also be applied to certain systems which are also not easy to define using a SDOF model. The method can also be taken a step further by constructing pressure impulse diagrams for a number of different elements such as beams, plates and columns for a series of damage levels. The damage calculated for each component is then combined in a weighted manner, depending on the vulnerability of each component, to define overall module damage.
Figure 25 - Pressure impulse diagram
Non-Linear Finite Element Analysis
The non-linear finite element analysis method (NLFEA) is potentially a very accurate tool, and often the only tool available, for predicting response. In contrast with the more conventional methods the NLFEA simulates the complete stress-strain history within the volume of the structure. Furthermore the impediments to the common use of this powerful modelling tool are continuously being pushed back by advances in computer technology. However, though the necessary computer speed and data handling capabilities may prove no longer to be a problem, it is still very much a specialist’s tool because of its general-purpose characteristics, and because the method inherently overestimates the resistance of structures. The successful application of NLFEA requires a sequence of carefully considered modelling, and possibly submodelling, decisions as well as a number of control checks before total confidence in the predicted results con be achieved.
Issue 1, October 2003
UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions
Choice of NLFEA tools
A suitable NLFEA tool should as a minimum have the capability to account for non-linear and rate dependent material behaviour, geometric non-linearities, large displacements, local and global instability in the structural response simulation. Geometric non-linearities are here assumed to include opening and closing of gaps, material ruptures and contact between separate components.
Construction of the finite element model
One of the first choices the analyst is facing when defining a numerical model of a given physical problem is to set the outer boundaries of the model, and the type, or combination of types, of elements to be employed. In general beam elements are employed for the analysis of complete modules, shell elements for the analysis of decks, walls and structural components, and solid elements for the analysis of connection details. The use of solid or continuum elements potentially leads to the most accurate predictions, but are also much more costly in terms of computer resources. Furthermore an increase in the complexity of the model also increases the possibility of human errors. The numerical model should include sufficient parts of the supporting structure to ensure that the end-restraints are modelled realistically. This is especially the case when determining the membrane forces developing in structural components undergoing large displacements. The membrane forces have the effect of increasing the resistance of a structural member by delaying the onset of buckling, and can also significantly alter the reaction forces. In this context it should be noted that when membrane forces develops in one part of the structure they are usually counterbalanced by compressive forces in other parts. These compressive forces may have an adverse effect on the stability of the structure. Rather than opting for a high overall level of detailing in the modelling of structural assemblies it is often possible, without loss of accuracy, to employ the output of one model to drive a displacement controlled analysis of a sub-model. An example of sub-modelling is the assessment of the containment pressure of a blast wall. Typically the stiffened or corrugated plate is represented by shell elements. A sub-model of solid elements is employed in order to assess whether or not the ultimate resistance of the welded connections has been exceeded. The converse of displacement driven sub-modelling is to use the results from a few lower level analyses, such as those carried out to find the deformation capacity of connections and welds, to define member performance standards applicable to the higher level analysis. A further complication often encountered in the NLFEA is the presence of mathematical singularities, which occur where elements meet at sharp corners. The analyst need to be aware of this numerical phenomenon, and may have to investigate this in further detail by employing refined models. In order not to overestimate the ductile capability of members, it is very important that the finite element analysis takes due account for all the potential buckling modes. As a consequence the mesh density need not only to be fine enough for apparent convergence, but need also to be sufficiently fine to represent all the dominant local buckling modes. Sufficient fineness can be obtained by dividing each structural element susceptible to buckling into at least 6-8 elements across the width. Furthermore the analyst may have to introduce imperfections or destabilising loads in order to trigger numerical buckling.
Issue 1, October 2003
and will as such require much more memory and disk space. Regarding meshing it is also considered to be good practice to keep the aspect ratio of the elements as close to unity as possible. Implicit methods on the other hand are essentially static codes running in time domain. For steel materials Von Mises yield criterion used in conjunction with an isotropic hardening rule and an associated flow can be employed in situations where the critical response is reached during the pass of the positive phase of the blast loading. However when carrying out a global structural analysis the ability to describe non-uniform loading. where the transient response are to be measured in tens of milliseconds. The spatial and temporal distribution of strain rates. However in most cases the reward is a more cost efficient design. Hence the individual structural components need to be checked until the analyst is satisfied that all potential failure modes have been accounted for. which also varies in time. 6. In the special situations where plastic strain reversal is significant a kinematic hardening model is preferred. Post processing tools can play an important role when investigating whether the meshing is adequately fine. and their adverse influence on the reaction forces need always to be considered when performing a plastic analysis. October 2003 153 . Ignoring the service loads when calculating the blast resistance will not necessarily be conservative. which in general means that a significant amount of time may need to be invested in setting up the models and interpreting the results. The degree of non-linearity of the problem makes the larger time steps permitted in the implicit codes of little practical value. Furthermore the analyst should recognise and possibly need to include the fact the structure is in a state of static equilibrium at the time of blast loading. and large plastic deformations develop. Blast loading should at all times remain normal to the surface of components undergoing large displacements. In general the explicit method is best suited for the analysis of blast loaded structures. Other service loads are included using a smeared approach. which can be mutually constrained. The most severe limitation for non-linear finite element modelling is the dependence on low level performance standards. The primary reason being that the explicit algorithm requires relatively little computational effort in each time step since no formal matrix factorisation is necessary.9. and only to gradually change the sizes of the elements in the mesh. Issue 1. When analysing structural components it is usually assumed that the pressure time history is uniformly distributed over the surface. is often required. For most models it is necessary to specify appropriate multiaxial constitutive equations for the material.4 Solution techniques in NLFEA Two fundamental different solution techniques can be adopted based on implicit and explicit time integration algorithms. Permanent loads associated with heavy equipment are best modelled by transferring the load through discrete fixing points.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions For large assemblies it is not practical to make the mesh fine enough to pick up all the buckling modes. Another advantage associated with the explicit codes is that they have an inherent ability to initiate most of the local buckling modes.
neglecting convection with the gas flow) giving a length scale of the disturbance ld = td x C0 of about 17 m for a 50 ms pulse.10 6. is the scale factor or coherence factor. 6.25 of the peak value. The structure has withstood these pressures with minimal damage. 154 Issue 1. The simplest factor.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 6. Often the direction of the loading on the primary framing may be changed locally due to panel membrane effects.10. These membrane loads may be balanced globally if panels of similar dimensions are attached to either side of the columns. the limitation of reaction loads by secondary member capacity and the effects of dynamics and plasticity are discussed in this sub-clause.3 Global Received loads Global loads on primary members are also dependent on the capacities and dynamic properties of the connected panels. The panel peak response may well occur long after the load has subsided and if a suction phase is present then the panel response itself may be reduced. Other mitigating effects including scale effects.10. Considering the situation of a pressure disturbance travelling along a wall or for a ridge of pressure travelling across a deck. Detailing of barriers and connections may however be required to bring the structure up to full capacity.75 of the peak. 88] and enable it to resist unexpectedly high loads without modification. October 2003 . A typical bracing member or panel of length l equal to 8 m will hence see a peak averaged pressure of 0. This indicates a scale factor of 1/3 on the global load/member load for this situation.2 Global Loading scale effects It is quite clear that the Advantica Spadeadam explosion test rig used in the UK project on Fire and Explosion engineering has repeatedly been subjected to local overpressures over 10 bar which it was not designed for. Isolated columns and beams will be loaded by drag or dynamic pressure loads which are only important near vents and are often much less than the overpressures. The loading direction resulting from an explosion may also necessitate further stiffening to prevent buckling if this is not allowed. The pressure disturbance is considered to be a pulse travelling with a velocity of about 340 m/s (the speed of sound C0 in ambient conditions. A module wall of length L equal to 35 m will see a peak averaged pressure of 0.5 6. The following cases are given only as an illustration of the scale effect. Often the loads transmitted into the primary framing will be reduced by panel capacities and the delay in load transmission due to delayed panel response. The method of load calculation is discussed in the context of piping and vessels in sub-clause 5. Finally these effects are discussed in the context of modified code checks in sub-clause 6.1 Response of the Primary Structure Introduction There are a number of effects which tend to mitigate the effect of explosion loading on the Primary structure [18.104.22.168. In situations where a large deck is considered these conclusions may only apply locally as the pressure loading pattern may be circular or irregular in shape. The precise geometry of the target structure and the loading pattern will change these conclusions and a detailed assessment of these factors should be made depending on the scenario considered. which affects global primary structure response.
89]. If a ductility of 10 is allowed then the resistance to load ratio is about 0. cladding and plate [88. 6. barriers. whose adequacy needs to be considered. AISC) may be re-interpreted to reflect the reserves of strength in the structure. October 2003 155 . Allowing for strain rate effects will extend the elastic range and this needs to be considered in the buckling checks.10. Unity or utilization checks to the appropriate standard (e.67 times the member resistance. the forces at the connections will increase.g. Values above 4 may be reached for structures with long natural periods (large frame structures or compliant structures). The checks can be carried out with some relaxation on the design yield strength which are normally based on guaranteed minimum values and an allowance for the increase in the yield strength due to strain rate effects.6 and the member will resist (dynamically) a load of 1/0.67 times the member resistance.6 = 1. Hence the ratio of the capacity allowing a ductility of 10 to the elastic capacity is 6.4 Plasticity and dynamic effects The Biggs chart shown in Figure 24 may be used to estimate the benefit to member capacity resulting from plastic deformation and dynamic effects (within the constraints that the chart is applicable to the structure and loading considered). this approach can be used before further checks and analyses are carried out. The allowed ductility will depend on the applicable performance standards. When compared with a Ductility level analysis this approach often result in a less efficient and heavier structure.10.5 Modified Code Checks The design of the topside components of an offshore hydrocarbon production facility is traditionally carried out using codes of practice based on conventional static design.6. and a ductility of unity is required (elastic response) then the static resistance to load ratio is about 0. For elastic response the dynamic amplification factor for this case is 0.15 and the member will resist a load of 6.6 bar. As an initial screening process. A ductility level blast performance standard could well allow a ductility level of 10 for decks and beams. Checking primary structure against the SLB (in the case where the load is 1/3 of the DLB) will often size the members for the DLB.67/1.2 for example.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 6. Also. If a member has a td/T ratio of 0. Issue 1. less than 0. Experience suggests that the Primary structure will not experience the full overpressure loading as the loading acts first on secondary members.5 – 0.2. particularly if the overpressures are not high. where the elements are required to remain elastic under normal service load conditions. These methods will not be suitable for the analysis of a situation where a fire has preceded the explosion unless allowance is made for material weakening and geometric imperfections and permanent deformations resulting from the fire.67 or a ratio of 4 for structures with a load to natural period ratio of 0. This normally involves comparing maximum stress levels in the member with allowable values permitted in the code of practice.
The material strain rate effect will generally give an increase of yield stress of the order of 20 %. Buckling checks for members working beyond the elastic limit are discussed in References  and . 156 Issue 1.2 may be applied to the acceptable utilization or to the yield stress. The occurrence of plastic hinging may be taken into account by factoring the acceptable utilization factor by the ratio of the plastic ‘Zx’ to elastic section modulus ‘Sx’.12 = 2. The benefits of this effect are exploited in blast wall design.20 x 1. Shear strains at supports should generally be limited to the elastic limit.e. Strain hardening will occur around regions of local plasticity. The shear behaviour of structures at their supports and in the joints of the primary frame will also affect the utilization factors derived in the code checks. Usually the benefit of both strain hardening and strain rate yield strength enhancement is not taken into account.5 x 1. i. 2.0 for primary members acting in bending and/or compression under explosion loading so long as the member does not buckle where local buckling is not acceptable. 4.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Blast resistant structures and their supports are designed to respond in a ductile way to explosion loading mainly in bending. A factor of 1/0.5) is then appropriate on the allowable utilization.5) for a tension member.67 (1. For a loading duration near the natural period of the member the shear forces could be higher. Code checks may be re-interpreted to take account of the following inherent reserves of strength. Taking into account all the factors above gives a possible acceptable utilization factor of 1. The recommended acceptable utilization factor is hence 1.1 (a factor of 2. This value may be higher locally. as the support or reaction loads may be less than the applied overpressure load. Alternatively the yield stress may be enhanced by the same factor to allow for nonlinear relationships for some aspects of utilization factor calculation. This factor is generally greater than 1. Additionally. both material and plastic reserves. Plastic deformation of a blast wall also has the effect of limiting the reaction loads on the supports.1 to 1.5 x 1. The following factors may be applied to the final utilization: 1. The safety factors implicit in the code checks will be different from those associated with bending. A factor of 1. The explosion event is an accidental event and hence the stress may be allowed to approach yield.20 x 1. plastic sections in compression and/or bending) allowance may be made for this by taking the design yield strength as the ultimate tensile strength divided by 1. in tension or be a ‘plastic’ section.25 x 1. October 2003 .5.25 (IGNs  ). Checks should be made where this is likely using a dynamic amplification factor based on the ratio of load duration to natural period (see Figure 19). The member must be able to sustain the formation of a plastic hinge before buckling.12 and will be in the range 1. for a dynamic situation these shear forces will be quite different from the values obtained by static analysis as a result of the inertia forces acting on the structure. this value is used in the nuclear industry . As discussed above shear checks should also be made using the correct dynamic reaction loads with strains being limited to elastic limits. 3. Where it is appropriate (for tension members. It is stated in Reference  that semi-compact and compact sections may be designed using code checks without supplementary local buckling analysis up to the formation of the first hinge. in particular in high strain regions of a blast panel.
The capacity of a blast wall or deck. so the need to achieve greater resistance or capacity in decks. Assessment and design tended to be component oriented.10. Non-linear and dynamic analyses became commercially viable for even small design houses at component level using PC based software. As the understanding grew. of the actual overpressures realized in an explosion.6 The use of NLFEA in global response One of the key limitations to effective assessment of the global response to explosion loading. and the delay of peak response to a time where the load has dissipated or reversed. which may reduce the explosion resisting capacity of the wall. making a wall more flexible may increase its ultimate capacity to resist blast. with little understanding of the interaction of effects between components  . The computer hardware was simply not generally available to enable the whole topsides structure to be modelled and analysed for the response to localised dynamic loading with a rapidly changing load function. in terms of explosion loading. In the mitigation of explosion loads. The deformation of the wall. The limit state approach may also be used with the equivalent elastic load level for the DLB (the dimensioning explosion load) being derived from consideration of the differing partial safety factors suitable for the ultimate limit state and elastic limit state. walls and connections has grown. dynamic techniques to identify additional capacity without the need for reinforcement. In the assessment of explosion resistance of existing platforms. The inclusion of penetrations and local reinforcement for the support of piping or equipment. October 2003 157 . Removing or revising over-stiff details. may also violate the integrity of any attached pipework or equipment. has been the time and cost of NLFEA structural analysis. will introduce hot spots in terms of stiffness. interpreted through its stiffness. through consideration of support stiffnesses and axial membrane resistance. 6. Issue 1. This could be the situation where the connection is made by means of a weld with a 5 % strain limit. These topics are discussed in References  and . the key consideration is the ability to absorb energy through ductility. as compared to the 15 % strain limit of the parent material. is limited by its weakest component. when subjected to explosion loading. A common misunderstanding in the design of blast walls is the relationship between the design overpressure and the strength of the wall. Finite element packages enabled single degree of freedom solutions to be improved. through full scale testing. say by making a heavy column base pinned rather than fixed. may allow the development of plastic hinges throughout a wall and through greater overall deformation before key component failures.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions The above approach assumes a working stress approach is being used for primary framing design. In some situations. The influence of penetrations could be addressed and the detrimental effect of local stiffening on global ductility became evident. Single component assessment could be expanded to the consideration of panels or complete walls and decks. simple linear elastic designs are often reviewed using non-linear. Later response may exploit the benefit of the suction phase in the loading.
especially if key components are limited in capacity by the 5 % weld strain limit. which ties two decks together. particularly if it has been detailed in such a way as to provide an explosion resistance load path between decks via pipe supports.10.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 6. as should the sensitivity to residual fabrication deformations and stresses. or on complete modules such as the Living Quarters. Destabilisation of the brace could result and potential toppling of any supported tall structures. Elevated masses may be required for significant items of equipment or vessels where rotational effects may be important and local horizontal loads are present. say to 600 mbar design overpressures. requires another order of magnitude of effort. Elastic global models may be conservatively used to determine accelerations on primary elements remote from an explosion. Accurate modelling of both geometry and mass distribution is required. This can be common with linear elastic designs. Application of point masses at deck level at primary model nodes may also be sufficient to achieve an acceptable accuracy for vertical dynamic response. or in the open truss work behind a blast wall. especially in fillet welded details. October 2003 . but modelling a complete module or integrated deck. such as penetrations and stiffeners. The sensitivity of the explosion capacity to deadweight and operating load should be addressed. Many different blast simulations may be required in order to generate a realistic envelope of responses for specific details. Inclusion of small details. an assembly of modules or a complete topsides of a fixed or floating facility. 158 Issue 1. may not influence the global response and thus becomes irrelevant for a global model. or for lifeboat and muster areas. in combination with the dead loads from the deck above. Finite element analyses may well consider a full wall or deck. Rigid boundaries for panel assessments are not necessarily conservative if the operating condition deformations are not considered. the total dynamic loading on the brace needs to be considered. as a considerable portion of the total capacity of a connection may be absorbed by secondary loads generated by eccentric load paths. computing power and cost. The danger here is that the framing braces may not have the capacity to carry the transferred loads from the blast wall.7 Global response considerations Another key area for not fully appreciating the response of the structure to explosion loads is at the interface between walls and decks. Welding residual stresses should also be considered in connection assessment. A balance has to be reached between the extent of the model and the level of detail in order to be cost effective. deflection criteria may become critical. including the explosion loading on connected decks. The weld geometry needs to be addressed for ultimate capacity calculations. In the assessment of explosion capacities of existing installations. Major pipework may need to be modelled if it could influence the dynamic and explosion response of the structure. In these circumstances. Dynamic amplification for such areas can thus be determined for use in local analysis. especially in circumstances where blast panels come into contact with primary framing braces at loadings beyond the design overpressure. which are revised to 2 bar overpressures following updated CFD modelling. in order to achieve a realistic dynamic response.
components that break free of their supports as a consequence of explosion loading effects. the following load conditions or actions need to be addressed for pipework. in order to prevent operating loads being transmitted into the ties. The potential for loose items to become missiles needs to be controlled by good working practices. potentially triggering a further hydrocarbon release. or for vessels and pipework to break apart and become missiles. cabling and their supports. Loose items subjected to blast wind. The loading required for design can be extracted from a global dynamic analysis which incorporates the ties – the ties affecting the dynamics of the deck girders. which would otherwise suffer consequentially from large displacements. Similar effects could be seen with pipework connections to equipment and with pipe branches. Tying long span decks together to limit the deflection under heave is a means of limiting damage to equipment and pipework. imparting loads on the vessel nozzle. Issue 1. a vessel will be supported from the deck of a module. the potential for items to break free and become missiles. the deck and roof will deflect in opposite directions. Dynamic pressure loads are discussed in sub-clause 5. pipework and cable trays.11. For instance. equipment. vessels. vessels.1 Response of equipment. or where supports are positioned either side of stiff deck elements leading to support rotations in opposing directions. Differential movements may generate significant loadings where adjacent supports respond in a differing manner to an explosion. October 2003 159 . needs to be controlled by design. such as scaffolding and loose items: • • • • • Dynamic pressure loads (comprising drag and inertia components) Differential pressure Differential movement Strong vibration Missile impact and generation Areas remote from. however. which may otherwise not be apparent when focussing on individual components or panels. or physically separated from. plus other miscellaneous items.11 6. pipework and vessels General In areas of a production facility with an explosion risk. vessels. but pipework connected to the vessel may be supported from the roof of the module. pipework or critical electrical and control systems. which could be catastrophic in terms of escalation of the event.11 and are potentially the most severe design loadings on equipment. or fragments of equipment or vessels generated by the break-up of those items in explosions all have the potential to become missiles which can then impact other equipment.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions An advantage of global models is the ability to visualise modes of mitigation. In an explosion. 6. including personnel escape provision. Care must be taken in detailing. the areas with the direct explosion risk should also be assessed for these actions if there is a potential consequence to personnel safety or integrity of safety critical systems.
as the design explosion loading would have been on a par with the accelerations allowed for in the transportation or earthquake scenarios. but normal operating vibration considerations are likely to govern. the external overpressure could be a significant cause for concern if ignition was to follow blowdown. Long spans for smallbore pipework should be avoided in order to limit vibration problems caused by flow turbulence. The deck plating or under-deck bracing must be assessed for adequate resistance to the lateral explosion loads. Too much movement may also contribute to vibration modes of failure for connecting pipework. as well as differential pressure and blast wind. but the uncoupling of loads onto the deck steelwork needs to be considered in conjunction with the deck loading from explosion.g. e. fuel gas lines and instrument pipework. Attachments to equipment must be assessed for the possibility of differential movement between the equipment supports and the attachment supports. there should be consideration of the near-uniform pressure loading. This is particularly important with small bore pipework. whereas simple NLFEA would be. flexible supports will generate lower reactions. In most circumstances. as process vessels are not normally designed to withstand net external pressure loads in excess of 1 bar. In many cases the blast overpressure will have negligible effect if the internal operating pressures are of a greater magnitude.2 Response of equipment and vessels to Explosion Loading For large diameter vessels. Process vessels need to be assessed for the vessel stability when subjected to explosion overpressures. or continued integrity during or following an explosion event. so modifications should be carefully assessed. October 2003 . Very stiff supports will generate high reactions. which would otherwise be removed after the transport and installation phase. this may no longer be the case and modification of the vessel supports may be required. this may be well above the design loading magnitude. A similar approach to the large vessels discussed above can be taken for the support reactions and the response of the under-deck steelwork. For equipment supported on Anti-Vibration Mounts (AVMs). Under the higher explosion loadings now prevalent. A simple SDOF solution may not be appropriate.11. 160 Issue 1. with full use of material certificates. Often a linear elastic design may be demonstrated to have adequate ultimate capacity if assessed using non-linear dynamic techniques. Too flexible a support condition may not be tolerable for operational reasons.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions The dynamic response of the vessel or equipment on its support steelwork is fundamental to the reactions that are generated. As discussed above. However. Tall equipment (relative to the support footprint) must be assessed for the possibility of toppling under explosion loading. the critical explosion loading effect on a vessel will be a lateral loading or overturning effect on the supports. The vessel should be considered rigid in order to develop support reactions. this could mean retention of lateral stops or shear keys. For the assessment of existing installations. 6. support strengthening may lead to an increase in the support reaction.
Plastic hinging soon occurs at each end of the pipe and hence the pipe may be assumed simply supported. Reduced piperack support reactions may be achieved by the use of baffles or slipstreaming aids. in order to present a single large rounded face to an advancing shock wave. This failure criterion applies to all reasonable pipe sizes and is insensitive to pipe diameter. In some instances. in order to prevent excessive midspan deflections or displacement of the pipe from its supports.11. whether at a pipe branch or transition from deck to roof support. It may well be prudent to design the first support from a pipe branch as a structural fuse. must be accounted for in considering explosion loading. Issue 1. line pipe is inherently ductile and consideration should be given to the adoption of flange class ratings which are higher than the design pressure rating of the pipe. Explosion loading assessments should not be restricted solely to oil and gas lines.3 Response of Pipework to Explosion Loading Pipework supports are generally designed to withstand gravity loads and allow thermal expansion of the line. should have the same considerations. the lateral loading generated by blast wind loads must also be accounted for.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 6. An average plastic strain limit of 5 % and a peak local strain limit of 10 % may be used to determine failure of the pipe. in order to prevent a line from breaching under explosion loading . Pipe runs on FPSOs are designed to accommodate significant movement as a result of vessel flexing. In general terms. the lateral loading caused by slugging forces must also be allowed for. Given the problems associated with developing a time-domain loading for a complete pipe run. Differential movement between adjacent supports. in order to prevent flange failure under explosion loading  . an appropriate test for robustness would be to consider each pipe span as simply supported and use either a Biggs SDOF approach or a static load solution in a standard design pipe stress analysis. October 2003 161 . Explosion loading on pipework will be significantly increased by turbulence and congestion. Flowlines are designed to tolerate a specified rotation or lateral movement of a well conductor or pump caisson. or between modules. If the pipe supports are attached to beams of similar section properties then the differential movement of the deck under an explosion should not be a problem. in regions of significant potential movement when subjected to explosion loading. safety critical pipework on fixed structures. as being essential for mitigation of potential fires following explosion. which in turn would lead to rupture of the line or flange leakage and an escalation of the event. but must also include deluge pipework. Consideration of the equilibrium of a length of pipe between supports as a catenary will give a time history of the tension Tp in the pipe and the extreme deflection of the mid point enables the plastic strain at each end to be estimated using methods given in the previous sub-clause. The tension capacity of the piping dominates the response and so a pure plastic resistance may be assumed for the pipe span considered with the mid point deflection being the relevant response variable. However.
For elastic response of the jacket. These need to be allowed for in the design of sensitive equipment. A reduced equivalent static load may then be derived which can be applied to the fixed platform computer model. The displacements and accelerations at points on the topsides may then be calculated. It is likely that the explosion loads will have a duration in the range of 50 to 100 milliseconds whereas the substructure will have a sway natural period in the range of 2 to 4 seconds. or large bending forces on long unsupported spans for pipework or cable trays.A. electrical connections and instrument panels. damage to valve actuators. It should be recognised that vibrations can be either vertical or lateral. In the Cullen report. a dynamic amplification (or reduction) factor may be calculated from the ratio of load duration td to natural period T or extracted from the curve given in Figure 20.. People were thrown off chairs and knocked over’.3. piping. Once the force is calculated it may be applied to a dynamic global platform structural model .11. These accelerations are potentially damaging to essential safety systems such as: • • • • • • • Emergency escape and rescue systems The temporary refuge and its supporting structure Emergency lighting systems Emergency power supply and battery systems P. October 2003 .6e. If the accommodation module or Temporary Refuge (TR) is supported on ‘anti-vibration’ mounts the acceleration may be further amplified due to resonance by the low frequency components of the shock.2) or later then the ‘strong shock’ response to explosions need not be checked. Strong vibrations could lead to the toppling of tall items of equipment or instrument cabinets. 162 Issue 1.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 6. depending upon the support arrangement of the item relative to the deck or wall transmitting the shock wave If the structure and SCE’s appurtenances have been checked for a Safety level or Ductility level earthquake loads following API RP 2A 9th Edition (sub-clause 2. it was reported that ‘a number of witnesses spoke of the severe vibration associated with the initial explosion. This would be particularly important for the fire water main. telecommunication systems and navigation aids Fire water systems Emergency shut down systems One further consequence of shocks of this kind is that pipe runs between modules may be stressed due to relative displacements between modules.4 Strong Vibration Strong vibrations are caused by shock or displacement waves transmitted through the decks and walls as a consequence of the explosion. For a fixed platform the integrity of the substructure should be checked against out of balance explosion loads.
This load case may then be eliminated from consideration. deluge pipework. EER etc. rotating machinery such as pumps. Awareness of strong vibration is therefore essential in the design and protection of safety systems used for emergency shutdown. The difference between guaranteed minimum yield stress values and actual coupon tests can give rise to a significant increase in reaction force.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Typical dynamic reduction factors of the order of 0. stair towers. Walkdown is now a commonly used tool to perform a qualitative assessment of the risk of damage from strong vibration  . The ductility values given in Table 6. particularly in strain concentration regions which will amplify strain rate behaviour. October 2003 163 . Isolation or shut down valve actuators. There are limited performance criteria in order to assess the reduction in available ductility. However in three dimensional systems tri-axial effects. 3. effects of bending and twisting in the third dimension are significant and will reduce the available ductility. should be assessed for the risk of toppling by strong vibrations. communications. These will be addressed in Part 3 of the Guidance. fire protection. Floating. 6. Battery racks should have a retaining bar across the front of each shelf. which may fall and damage control panels or cause injury to key personnel.5 are based on 2 dimensional beam models. 2. Systems which some may consider not to be “Safety Critical” also need to be assessed for strong vibration loading: e. HVAC dampers and lifeboat mountings should be reviewed for the consequences of shock or strong vibration loading. Assessment of response at high ductility levels is still difficult due to uncertainties in predicting both the performance of welded details under dynamic loading and in sensitivities in finite element modelling. turbines and compressors. which are protected from the direct influence of explosion loading. module stools. Massive objects with sensitive connections to adjacent machinery should be particularly addressed. 1. HVAC.05 may apply in which case problems with the substructure are unlikely. emergency power. tethered and moored installations are unlikely to suffer serious response to this form of loading because the natural period in the horizontal direction is much longer than the load duration. Even book shelves in critical areas such as the Control Room should be assessed for the risk of displaced objects. fire and gas detection.12 Areas of uncertainty There are a number of areas which still need to be addressed that relate to structural response. Issue 1. Control and instrument cabinets in utilities areas. grating on escape routes. Such systems are fundamental to the safe evacuation of personnel from the facility and strong vibration or blast wind loading may be the governing design load case. on the basis that the equivalent static loads on the substructure are smaller than those induced by waves and currents.g.
concept selection is normally achieved by means of comparative. with a minimum facilities platform.2. 7. generally the selection process for sub-sea or topsides wellheads is related to satellite facilities where processing and main export is accommodated elsewhere. will have constraints in terms of viable location. as best suits the delivery of the discussion and its understanding. but the topsides wellheads makes for improved accessibility for maintenance. Detailed guidance will be given in Part 3 of the Guidance Document. risk and timescale. layout. With sub-sea wellheads. Explosion hazard is not normally a primary consideration. An existing facility modification.1 DETAILED DESIGN GUIDANCE/DESIGNING EXPLOSION RESISTANCE General FOR This clause of the Guidance is based on Reference  and incorporates additional design and operations experience to provide guidance on methods and particular areas of detailed design that are not covered elsewhere in this Guidance.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 7 7. as the satellite structure. normally unmanned installation. primary separation may be included. Current technology has made the subsea option more viable. if selected. materials or structural form. identifying those areas that have an impact on explosion hazards. spares. especially where the main facility is in close Issue 1. as will the conflicts between explosion and fire hazard management or mitigation. Conflicts between the requirements for different engineering disciplines will be identified. accessibility and impact upon the existing Safety Case and HAZOP assessments for the facility.2. 7. verification and compliance) through all phases of project development. would be a minimum facilities.1 The Design Sequence Introduction The approach in this section is to work through the design process in a generally sequential manner. all fluids would be transferred to the main facility for processing.2 Project Appraisal and Concept Selection Given the number of competing options available. weight. materials. qualitative and quantitative assessment based on cost. The readership is intended to be broad and the issues understood by all disciplines (engineering. to enable co-ordinated best design practice. such that significant changes can be avoided late in the design process. by contrast. Some areas will be covered on a phase by phase basis. The following subject areas have an influence on explosion hazard: New build or modification a new build project essentially starts with a clean sheet of paper and thus has little or no constraints in terms of layout. power.2 7. The detail design aims have been identified in this part of the Guidance Document. execution and operation. October 2003 Sub-sea or topsides wellheads 164 . interfaces with existing plant. others on a discipline by discipline basis. Industry initiatives that have generated detailed design and operating practice guidelines are briefly discussed. management.
To date. Regulations vary in different parts of the world and cognisance is therefore required of the relevant governing regulations for the installation location. with the pressure of the flow. the availability of a convenient nearby pipeline. with sand entrainment being the key contributor to erosion. The temperature of the fluids has an influence on the degree of absorption of gas within the crude oil.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions proximity. The process area has the highest risk on the facility. excessive flaring is no longer considered to be an option in many parts of the world. gas has either been flared. pressure containing vessels and machinery. The required pipeline entry pressure will contribute to the compression requirements for the platform. high pressure and subject to vibration due to tree movement. but a risk is still present. but is much less significant in terms of explosion hazard than pressure. The other factor in determining compression requirements is whether gas or electrical submersible pumps are used for artificial lift of the oil. with available capacity. The utilities area is much less vulnerable to serious explosion hazards. It is normal to distance the gas compression plant from the separation vessels. or reinjected into the formation for later export. due to environmental considerations. October 2003 165 . being the key parameters in determining the magnitude of release for a given leak size. The gas lift lines are a major potential hazard due to them being small bore. These have the next highest risk. A major hazard is also presented in the loss of containment of the gas in the well annulus. Inventory The gas content of the received fluids is the primary contribution to the explosion hazard. Switching from gas lift to electrical submersible pumps is generally not an option on existing installations. due to the compressor design requiring a minimum throughput to maintain efficiency – for new installations the electrical submersible pump option is much safer. where possible. coupled with volume in the case of vessels. due to the amount of flanged pipework required to connect the trees to the manifold and the tendency of this pipework to move in concert with the conductors. The degree of gas compression has a significant influence in the potential for leaks and the resulting gas cloud sizes in the process area. utilities and accommodation Wellbay and drilling facilities Export route Gas lift Regulatory regime Issue 1. drilling. exported by pipeline. as a consequence of the high pressures and the number of potential leak sources in the pipework. in terms of explosion hazard. The sweetness of the crude and the water content are the primary contributors to corrosion risk. to reduce the escalation risk for any single event. for example with fuel gas for power generation. Separation of process. may determine whether export of oil is by tanker or pipeline. irrespective of where the design is executed.
concept selection is normally achieved by means of comparative assessments both qualitative and quantitative. in order to be meaningful and aligned with the design process. normally unmanned installation. coupled with volume in the case of vessels. being the key parameters in determining the magnitude of release for a given leak size. With sub-sea wellheads. would be a minimum facilities. will have constraints in terms of viable location.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Budget. risk and timescale. whether for new build or modifications to existing installations. This review process must be open. An existing facility modification. spares. • • 166 Issue 1. The sweetness of the crude and the water content are the primary contributors to corrosion risk. Explosion hazard is not normally a primary consideration. with a minimum facilities platform. Current technology has made the sub-sea option more viable. by contrast. with the pressure of the flow. materials. as well as to the capabilities of new technology or design methods. interfaces with existing plant. especially where the main facility is in close proximity. power. however. The initial hazard review can make judgements based on past experience and advise on the consequences of initial engineering or management decisions. to new understanding on explosion loading and response. resulting in a less than optimum solution. The temperature of the fluids has an influence on the degree of absorption of gas within the crude oil. timescale and external factors Best practice may sometimes be compromised by the available budget and project timescale. weight. The degree of compromise should be documented for evidence in hazard review and project safety audits. but the topsides wellheads makes for improved accessibility for maintenance. Inventory – The gas content of the received fluids is the primary contribution to the explosion hazard. accessibility and impact upon the existing Safety Case and HAZOP assessments for the facility. as the satellite structure. primary separation may be included. all fluids would be transferred to the main facility for processing. materials or structural form. The initial QRA must be based on current generic data for leak and ignition frequency. Initial hazard review Given the number of competing options available. with sand entrainment being the key contributor to erosion. based on cost. if selected. or take account of proposed high integrity solutions which offer better hazard management. The following subject areas have an influence on explosion hazard: • New build or modification – a new build project essentially starts with a clean sheet of paper and thus has little or no constraints in terms of layout. Similarly external factors such as lift vessel availability may influence or determine the outcome of a particular decision. Sub-sea or topsides wellheads – generally the selection process for sub-sea or topsides wellheads is related to satellite facilities where processing and main export is accommodated elsewhere. layout. but is much less significant in terms of explosion hazard than pressure. October 2003 .
The required pipeline entry pressure will contribute to the compression requirements for the platform. timescale and external factors – Best practice may sometimes be compromised by the available budget and project timescale. due to the amount of flanged pipework required to connect the trees to the manifold and the tendency of this pipework to move in concert with the conductors. The gas lift lines are a major potential hazard due to them being small bore. Switching from gas lift to electrical submersible pumps is generally not an option on existing installations. as a consequence of the high pressures and the number of potential leak sources in the pipework. pressure containing vessels and machinery. The initial QRA must be based on current generic data for leak and ignition frequency. to reduce the escalation risk for any single event. or take account of proposed high integrity solutions which offer better hazard management. drilling. Initial hazard review – The initial hazard review can make judgements based on past experience and advise on the consequences of initial engineering or management decisions. The degree of compromise should be documented for evidence in hazard review and project safety audits. may determine whether export of oil is by tanker or pipeline. but a risk is still present. or re-injected into the formation for later export. Export route – the availability of a convenient nearby pipeline. The wellbay and drilling facilities have the next highest risk. whether for new build or modifications to existing installations. high pressure and subject to vibration due to tree movement. for example with fuel gas for power generation. resulting in a less than optimum solution.3 Sub-sea Layout As discussed above. due to environmental considerations. Additional elements to those already covered above. include: Issue 1.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions • Separation of process. due to the compressor design requiring a minimum throughput to maintain efficiency – for new installations the electrical submersible pump option is much safer. exported by pipeline. irrespective of where the design is executed. gas has either been flared. in terms of explosion hazard. October 2003 167 .2. with available capacity. however. • • • • • • 7. This review process must be open. excessive flaring is no longer considered to be an option in many parts of the world. The degree of gas compression has a significant influence in the potential for leaks and the resulting gas cloud sizes in the process area. as well as to the capabilities of new technology or design methods. Budget. where possible. Similarly external factors such as lift vessel availability may influence or determine the outcome of a particular decision. the sub-sea layout can significantly influence the topsides inventory and the topsides hazard. To date. Gas lift . to new understanding on explosion loading and response. A major hazard is also presented in the loss of containment of the gas in the well annulus. utilities and accommodation – The process area has the highest risk on the facility. It is normal to distance the gas compression plant from the separation vessels. The utilities area is much less vulnerable to serious explosion hazards. in order to be meaningful and aligned with the design process. Regulatory regime – Regulations vary in different parts of the world and cognisance is therefore required of the relevant governing regulations for the installation location.The other factor in determining compression requirements is whether gas or electrical submersible pumps are used for artificial lift of the oil.
in-field pipelines. Sand is primarily a source of vessel and pipework erosion.2. given its current condition and design specification. sub-sea storage. Areas for consideration include: • • • • • • • • separation. cooling/heating. the process design then has to evaluate and specify equipment which will convert the fluids to export specification. water injection. metering. Several of these elements affect topsides risks by their vulnerability to dropped object hazards causing sub-sea leaks. The pipe-in-pipe solution requires the use of vents to control leaks. sand entrainment. cost and impact on the global structural integrity of a retrofit custom designed riser. together with any allowances for future expansion.4 Process Engineering Having got the reservoir fluids to the topsides. operational risks. generally through mechanical connections. October 2003 .UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions • • • • • • • sub-sea isolation valves (SSIV). compression. pipe-in-pipe risers. 7. The use of new or pre-installed risers is a judgement as to the fitness-for-purpose of the existing riser. in which the construction arrangement inhibits in-service inspection and repair. emergency shutdown valves (ESDVs). in such a manner that the changing circumstances through the life of the field can be met. The influences on explosion hazards here are leak sources and the magnitude of the release. as balanced against the construction risks. vessel size and configuration. storage prior to export. 168 Issue 1. but also associated with corrosion and human intervention. Water injection is included here as being associated with aquifer water and the pipework contributing to process area congestion. water is the primary source of corrosion. new or pre-installed risers. high integrity pipeline protection systems (HIPPS). This arrangement may be preferable to single skin risers for reasons of insulation to maintain oil temperature.
generally through mechanical connections but also associated with corrosion. Vessel size is also crucial in terms of reducing pressures during blowdown. flange locations to facilitate installation and maintenance. in terms of explosion hazards. Smaller vessels may thus be more controllable in emergency situations. The following also need to be considered in terms of explosion hazard management: • • • • • • large deflections overloading flanges. fatigue. use of pipe supports as structural fuses to protect pipework at branches. joint damage due to mechanical impact. material selection for robustness (wear control and corrosion resistance). October 2003 169 . selection of laydown and storage areas and mechanical handling routes between the laydown areas and the relevant equipment.5 Pipework The initial pipework sizing and routing follows the layout process. 7. Operational considerations included here are: provision of fixed and spring supports.g. control panels and junction boxes.2. There should be a challenge to lagging/insulation for personal protection or heat conservation due to issue of corrosion risks hidden under lagging and also to the materials selection in the context of late field life costs and integrity. The key areas for explosion hazard potential are leak sources and the magnitude of release.2. inspection and maintainability particularly in late field life. erosion and human error. switchgear and distribution boards. The main considerations here. are associated with hot surfaces and electrical sparks. area congestion. operability. bundling and shrouding of pipes to reduce explosion turbulence. shielding of safety critical pipework to reduce drag wind loading. Design should be influenced by the maintenance of piping joint integrity. The security of supply to critical services (e. The key design considerations are inherent safety and robustness. maintainability and robustness. Issue 1. transformers. 7.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Vessel size is a consideration in terms of the quantity of the topsides inventory and the potential for fuelling a fire in case of rupture.6 Power Requirements and Electrical Systems The power requirements can be determined after the compression and main pump designs are at an advanced stage. inspection. valve locations. but may be less efficient in operating terms. The layout stage is important in the provision of adequate inspection and maintenance access. Elements for consideration include: motors. The design should be influenced by reliability. TR pressurisation) and protection of generator and TR HVAC intakes from smoke/gas ingress.
Intrusive instrument tappings are a major potential source of leaks and non-intrusive options are recommended if possible.2.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 7. hence reducing congestion. Reference  is a good starting point on this topic. A review of potential standardisation of components is likely to lead to reduced human error. The Control Room layout is critical in terms of the potential for human error in crisis management – information overload is now recognised as a cause of delay in response times.2.8 ESDV. switches etc) need to be accessible to minimise or eliminate the need for scaffolding or other temporary access. These codes do not take account of the stress distribution in the vessels and the strength of the vessel steel is reduced faster than is applied in the codes. although work is continuing at the time of writing. Pressure Relief Devices and Isolation Systems The Emergency Shut-Down Valve (ESDV) system provides the means of isolating the installation from import and export pipelines. Blowdown. Pressure relief devices are provided on process systems to prevent rupture of pressure vessels and leakage of pipework joints under applied pressure arising from faults in the process control system or as a result of fire. along with the potential for upgrade or expansion to accommodate future platform modifications. 7. as a result of the higher heat load that would be typical. but both the blowdown and isolation systems require a very high integrity. The mixing of system types or overloading of components can lead to increased explosion risk. electricity or fuel gas.7 Instruments and Controls A key consideration in terms of explosion risk is whether instruments are powered by air. reliability and survivability in order to function safely. The isolation systems enable safe and secure isolation of key inventories and components to enable draining and purging of fluids prior to maintenance or inspection. in order to control the topsides inventory in an emergency or quickly terminate export in the case of a pipeline or riser leak. An area of potentially poor design that has become apparent recently is that onshore refinery based design codes for blowdown systems  are based on fire loads which are too low for the offshore industry  . flare or reservoir in a controlled manner. The blowdown system rapidly transfers the gas or oil inventory to the vent. Reviews of blowdown system efficiency are also now considering sequenced blowdown as a means of improving overall response time. Such reviews need to assess manual versus automatic or semi-automatic shutdown of systems on high-level gas detection. vent paths and potential missile material. 170 Issue 1. in order to reduce the potential for further escalation in the case of a fire or leak. October 2003 . the later obviously having an associated leak potential and the middle option an ignition potential. The durability of components and software should be an important design consideration. These systems can involve a large number of flanged connections and isolation systems are associated with components that are regularly subjected to manual intervention. Cabinets in hazardous areas need to be robust to withstand explosion loading and prevent missile penetration. Instruments and controls (valve wheels. The vessel integrity may thus be compromised before the blowdown is achieved. handles. The same influences and design requirements apply to general pipework. Local or remote control is significant in terms of manning levels within hazardous areas.
HVAC. An impaired pressure relief valve may fail to prevent an explosion or lead to a sudden large toxic gas release in the event of rapid escalation of fire. ventilation. one third to one half of the DLB has been quoted elsewhere for initial estimates of SLB loading. The main layout considerations are: process economics. pipe racks. 7. trip hazards. incorrect installation. this could apply 50 % of the nominal load for elastic design. installation or environmental) but blast walls will be sized for these blast loads.11 Structural Arrangement – Topsides In the sequence of design phases. 7. congestion.9 Utilities Requirements Utilities systems in themselves are not contributors to explosion risk. storage.10 Layout The equipment layout is usually based on good past experience and an awareness of design details that can lead to problems with poor ventilation. a suitable nominal overpressure for a facility type and location may be available for initial structural sizing. they can contribute to congestion and their failure under explosion loading may reduce the ability to mitigate a following fire or contribute to secondary incidents such as chemical spills. Utility systems include: firewater deluge. Some operators have experienced a number of failures of bellows in balanced bellows type relief valves  . high integrity seals. chemical storage. mechanical handling provision. A load factor would be required to enable sizing by elastic methods. communications. Bellows failure may be caused by valve chattering.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Considerations for reliability and robustness of systems include the use of double isolation valves. maintenance access.2. The primary structural steelwork will often have been sized for other load considerations (operating loads. growth provision.2. as non-linear analysis would be too extravagant for initial sizing. seawater cooling. The design and routing of deluge mains is of primary importance in ensuring a robust system capable of withstanding the Strength Level Blast and providing continued deluge coverage outside the affected area in a Ductility Level Blast.0 and characteristic material properties. The finished layout may then be validated by CFD modelling. escape routes. Issue 1. with safety factors of 1. 7. the attachment of small bore bleed pipework in double block and bleed isolations and the selection of appropriate pressure relief valve devices. accessibility. pipework protection.2. lighting. HVAC. however. October 2003 171 . cable trays. blast wind turbulence and high explosion overpressures. incorrect material specification or pressure rating. incorrectly specified fatigue life or incorrect dimensional tolerances on the valve seat and bellows manufacture. congested explosion venting. diesel and potable water tankage.
UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Topsides structural and architectural design includes the following considerations: modules and module stools. however. Some fixed steel platforms have diesel tanks in the legs close to the topsides. growth provision. Jack-up production and drilling platforms. particularly with gas and chemical injection pipework. Conductor and riser support is important in the control of flowline vibration topsides. Control Room and Workshops. it is important to consider the negative pressure impulse immediately following the initial high positive pressure impulse. decommissioning. 172 Issue 1. integrated decks. lifeboat and liferaft supports.2. A robust sub-structure design will significantly reduce the explosion risk associated with strong lateral vibration resulting from major boat impact hazards. October 2003 . whose failure could lead to TR impairment. such tanks will require vent pipework to control gas build-up and allow tank purging. ensuring that gross deformation of the topsides does not result. tall structures – drilling derrick. bridges and bridge supports. stairways and escape routes. leading to secondary releases and explosion risk. The main structural forms for manned installations are: • • • • • • • Fixed steel platforms. In the design of architectural features such as doors and windows in the LQ. cranes and masts. jack-up production facilities will have diesel tanks in the side double skin or crane pedestal. blast walls. living quarters (LQ) and temporary refuge (TR). equipment supports and pipe supports. fabric maintenance and inspection. the substructure arrangement can have a bearing on the behaviour of the facility under extreme events. Subsea oil storage is not considered to be an explosion risk.12 Structural Arrangement – Substructure Whilst less susceptible to direct impairment from explosion loading. Floating structures – semi-submersible. The ability of doors and windows to resist outward loading is generally much lower than that of inward loading due to the construction details and assembly method. wind walls and fire walls. Helideck. passive fire protection (PFP). Structural redundancy will ensure adequate support to the topsides in the case of a surface pool fire. floors. FPSOs have storage tanks beneath the process modules. Spars TLPs Gravity base concrete platforms. Gravity based steel platforms. Such vibration can have a significant impact on attached small bore pipework fatigue. lift points. 7. doors and windows. FPSO. The provision of pre-installed risers and J-tubes for future satellite development can significantly reduce the explosion hazard associated with construction activities on the topsides.
Issue 1. requiring pressurised welding habitats the manning levels for which often require more space than is available. but welded connections are difficult to achieve on a live installation. Craneage requirements. With all piping systems the location of flanged joints must be considered in conjunction with the selection of support locations. as a consequence of joint relaxation upon cooling or accidental mechanical impact.2. leading to gas leaks. are that piping joints are cold-stressed and on-site fit-up rectification is required for pipework or structures. even where no intervention had occurred. mechanical handling routes. where the flange class ratings are higher than required for the design pressure rating of the pipe. which integrates data from all disciplines in a 3-D model. The use of PDMS (Plant Design Management System). 7. in terms of explosion hazard. If high integrity flanges are used. October 2003 173 . An additional requirement for works on existing installations is a well-developed installation scheme. The need for detailed dimensional surveys for pipework tie-ins at the hook-up stage is demonstrated time after time. equating cost. the degree of non-compliance needs to be established and various options described elsewhere in this Guidance will need to be considered to further determine the optimum route for reduction of the explosion hazard: A comparative exercise is required to document the pros and cons of one or more of the options.13 Explosion & Fire Hazards Review If the ALARP (as low as reasonably practicable) criterion has not been met. Flanged connections present an ongoing operational leak risk. cabling etc) all need to be clearly defined and prepared for. Hook-up and Commissioning The balance between construction risk and operational risk in pipework modifications needs to be carefully assessed. following assessment of an existing structure. clearances and temporary removals (pipework. timescale. It is sometimes impracticable to handle long lengths of new pipework offshore. laydown availability. risk savings. but as-built models are still rare. The consequences of unexpected problems can again be associated with human error. Leaks can often be detected in existing pipework following a shutdown. Reliance on “as-built” drawings has caused many project delays.2. that may involve hot work. whereas a single visit offshore to confirm the accuracy of the recorded data would have ensured a better fit-up and eliminated clashes with items not included on the referenced drawing.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 7. For modifications to existing installations factors such as lack of bed space and production downtime for modifications are significant factors in the calculation. has vastly improved clash control. especially where complex geometries are involved. ignition sources and hence explosion risk. The consequences of being ill-prepared at hook-up. The balance is thus drawn between the cost of lost production and the increased hazard presented by more pipe flanges. for example. Nitrogen leak testing is recommended for all hydrocarbon pipework following new construction and maintenance. to avoid joints being placed at locations where high ductility is required. then the system ductility will be tolerant of explosion loading and brittle modes of failure and flange leakage will not result.14 Modifications to existing installations: Survey. practicability etc.
2 Management of the Explosion Hazards High risk activities. Other high risk activities like construction and maintenance. the needs of each design discipline must be weighed against each other. as well as good working practices. such as start-up. Hot surfaces. The need for a calm weather policy should be considered. by design and operating procedures. however. lights. to best suit the particular geometry of the structure and identified ignition sources. October 2003 . 7. In each design process. helicopters and lifeboat engines. These considerations are constantly changing as new technologies and new explosion behaviour or response information becomes available.3. Measures to be considered under this category include: the minimisation of potential missiles. as well as those of the operating requirements. Similarly. the regulatory requirements. ALARP (as low as reasonably practicable) is one term used to measure the acceptability of the resulting risk to personnel safety.3 Derivation of Explosion Loadings. as the absolute minimum risk may result in a design that is not commercially viable. flushing and venting. but will be used here to describe the outcome of a design process that would result in the minimal explosion risk for a viable concept. well operations. such as ship exhausts. mechanisms of escalation and the stability of tall structures.3 7.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 7. or containment of damage. can have the associated risks reduced by good design. the strengthening or protection of critical safety systems and critical equipment. the consequential damage caused by displaced wind walls or blow-out panels must be considered.1 Best practice in explosion hazard design Introduction “Best practice” is a subjective term. the protection or behaviour of the primary and secondary structures subjected to explosion loads. relief valve testing and simultaneous operations will also have reduced associated leak risks through good design and planning. blow-down.3. 7. draining. External ignition sources. commercial considerations and safety. This sub-clause outlines briefly the considerations used to arrive at a best practice design. Design is a balance between conflicting objectives. such as hot work or poorly earthed scaffolding or equipment. also need to be considered in the design process.3. 174 Issue 1. Temporary ignition sources. behaviour and ultimate capacity of blast walls.3. An increase in venting area can be achieved in modifications to existing installations as a mitigation measure. the protection from loss of containment damage to process systems components. 7. and equipment earths are all fixed ignition sources that need to be designed or selected with recognition of the potential explosion hazard. Rarely will two designs have the same set of parameters and therefore the same outcome. electrical equipment. The term viable is necessary. commercial viability and the environment. Selection of a higher specification. need to be eliminated or controlled.4 Response to Explosions The key factor in the response of the structure or equipment to explosions is the limitation of escalation damage. Venting panel locations can also be optimised. testing. Design has a big influence on the prevention of ignition. sampling. different location or gas tight enclosure will reduce the ignition risk. Detailed guidance will be contained in Part 3 of the Guidance Document. The effects of venting on escape routes and adjacent areas needs to be considered. the deflection of decks.
The consequence of component failures within the collapse sequence must be appreciated. Increased capacity in existing systems can often be demonstrated through the use of non-linear analysis.5. where simple elastic methods were used in design – the strain limitations of the welded connections may be the limiting factor. with understanding of the real collapse and failure mechanisms. especially in the area of missile limitation. In other instances. say within a double skin wall. there may be no significant consequence from first failure. It is important to be aware of the interdependence of some of the controlling parameters such as the angle of corrugation which will control the degree of restraint to the flange. Also the web/flange ratio plays a crucial role in controlling buckling of the section. The walls typically span from floor to ceiling and are normally provided by specialist blast wall suppliers. as the failed component may be redundant.3. firewalls were used in situations where explosions were a possibility. the initial failure in the system may not represent the ultimate capacity. The design based issues.1 Specific Considerations for Blast Walls Corrugated Profiles Corrugated profiles are commonly constructed from stainless steel due to its high energy absorption characteristics and good corrosion resistance. Connection capacities can be more critical than span capacity. A weld failure along the edge of a blast wall may enable release of gas or flame into a previous safe area. Blast wall performance can be compromised by the need for pipe transits.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions The subject matter in this category is very broad and needs to be addressed by simple. However these were not initially designed for the blast loads and many were constructed from mild steel.5 7. corrugation angle and the end connection details. Capacity is not necessarily related to stiffness but to ductility. The behaviour of the wall is very much controlled by the web and flange slenderness limits. Several of the measures are procedural. Equally important. October 2003 175 . Complex loading situations need to be considered on space frame systems. cable conduits and access manways. with a potential tunnelling failure resulting. not just structural. where computer modelling may only represent idealised post-yield and joint behaviour. PrePiper Alpha.5. the column slenderness of the wall. 7. with detailed assessment being performed in accordance with Part 3 of the Guidance. Previous research  on profiled flooring systems suggests a maximum value of 1. Technical Note 5  provides guidance for web and flange slenderness limits based on stub girder tests in order to classify the section. the flange to web ratio. although limited profiles were given which may not be appropriate for blast walls. generic guidance in this Part 1 document. Typical failure strains for ASTM 316 stainless steel are in excess of 50 % with ultimate capacities in the region of 600 N/mm2. In traditional static design of typical universal beam members in bending. Issue 1. The height of wall that can be provided varies between 3 and 8 metres and thicknesses are typically from 6 mm to 12 mm. whether for new or existing installations. codes give slenderness limits for web and flange elements independently despite the fact that one provides restraint to the other. can involve multi-discipline aspects. Mechanical strengthening may have to be achieved through cold work solutions.3. if platform shutdowns are to be avoided. A guide issued by FABIG.
they are likely to be closed hat type sections. occur in the components of the connection rather than the weld itself. Walls are normally attached at their top edge to the flanges of primary beams. They should be designed for a capacity that is greater than its supported member. In some cases the wall is welded directly onto the deck which has no rotational flexibility. then little of the blast energy will be absorbed via bending deflections. which may not always be avoided. Double skin profiled systems have also been adopted recently where the skins were separated by an air gap with the same profile used for both layers to allow for a blast occurring from either side. relying totally on the top connection for this.2 Stiffened Panels Stiffened panel blast walls consist of a flat plate stiffened by a series of vertical and horizontal stiffeners and is supported by the main steel structure. 176 Issue 1. The bottom of the wall is commonly welded to a channel or angle section attached to the top of the deck in order to allow for rotation to occur. With profiled sections under blast loads there is little published data to provide the same extent of validation for using these results. For integrated decks the stiffeners are continuous and are required to carry large service loading.5. Where there is a need to assess the capacity of module walls and columns with in-plane loads the effects of static loads need to be considered on the response of structural members: 7. This also has the effect of reducing the forces applied at the supports. Also the web in this case is loaded by the pressure providing a destabilising load. This will attract high shear forces which will impose large forces at the connections and could induce buckling of the cross section. 7. The concept being that one of the skins becomes sacrificial while the gap increases the standoff from the explosion and impacts the second skin. normally allowing rotations to occur such that large membrane forces are not developed.3 Support Details for Blast Walls Support details for any blast barrier are crucial to ensure that the integrity of the wall is not breached during an explosion and must be detailed such that they are not the weak link. As a result. However.3. It is important to ensure that any stress concentrations. If a deeper wall is adopted which is very stiff. Buckling issues need to be carefully checked for deep troughed walls. this will be transferred to the main structure as an axial force and requires careful consideration.5 for plastic sections. Although plates can resist high loads through membrane action. particularly if the deck is carrying high service loading. The connection must be sized to transfer the computed reactions and to assure that any plastic hinges can be maintained in the assumed locations.5. Many stiffened panel walls have connection details which minimise the transfer of high axial forces to the main structure for this reason. October 2003 . For relatively shallow profiles the corrugations can open up under a blast pressure load allowing some of the energy of the explosion to be absorbed by the deflection of the wall. It should also be noted that Reference  provides guidance on the selection and design of profiled barriers.3. The wall is designed to be flexible to allow much of the energy to be dissipated in bending. this deals with design criteria for predominantly elastic wall profiles and recommends the use of NLFEA to provide performance standards for ductility factors greater than 1. It is important to be aware of the change in behaviour of the wall with increasing depth.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Extensive validation work has been carried out to ensure designs are conservative based on this specification with an effective width concept adopted to account for the actual stress state.
and providing adequate extension is available. Yasseri  has discussed a number of these.5 Transits penetrations in Blast Walls Ideally no penetrations should be allowed in the wall as these are likely to reduce the available ductility under blast loading. it is also necessary to define whether the service moves with the wall or not. can be anchored to the wall. have shown that flexible angle connections arranged in a dog-leg configuration performed well in terms of energy absorption capability .5. the greater strength in the door mountings being present to resist the positive pressure loading. which restrict the differential movements between the floor and the ceiling. if they are required they should be located at the top or bottom of the wall. 7.3. These need special attention particularly in relation to any likely stiffening which can introduce undesirable stiff areas which attract high loads. In order reduce local loading effects the ties should preferably link the stiffer members of the deck structures. How that system is operated and maintained is outside the control of the designer. Stiffer walls also tend to exhibit more brittle failure modes. as there are likely to be many penetrations including doors in some instances. which investigated tests on firewalls carried out at the Spadeadam test site. together with numerical studies. Inspection and Maintenance Issues Good design is intended to deliver a reliable. October 2003 177 . 7.3. maintenance and operating procedures. For existing installations the problem is likely to be more complex. 7. including adding a facing plate (flat or corrugated) to the wall or the addition of an X-bracing or stiffening in the form of a light grillage welded to the wall. may be incorporated into the design to enhance its blast resistance. although recommendations can be made regarding inspection. In addition to the above considerations. robust system to the workplace that meets the required specification.5.7 Operations. However. 7.4 Increasing Blast Capacity for Existing Structures In situations where walls have limited clearance strengthening schemes which considerably stiffen the wall have been adopted.3. The doors and seals need to retain integrity through both the positive and negative parts of the explosion impulse.3. Issue 1. However. The ties must be designed so that do not carry any compressive loading occurring as a result of operational loading or the explosion rebound of the deck structures.6 Integration with Fire Hazard Management The integrity of door strengths and seals is of primary importance on air lock systems. 7. However limiting the deflection of the wall means less energy of the blast will be absorbed in bending which will attract high shear forces at the connections. a large pipe can be regarded as a fixed object and the wall must move relative to it.5. or other locations where deflections will be a minimum.6 Combined Behaviour of Decks and Walls It has been suggested that internal ties. a cable has a degree of flexibility.3.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Results from a recent JIP. For instance.
In the case of bolted joints. “Protective Systems” are defined as items that prevent an explosion that has been initiated from spreading or causing damage. In the case of bolted flanges this is achieved through the use of design codes such as PD5500 and ASME VIII as well as correctly executed working practices with regard to joint make up. Rigorous testing procedures should prevail to ensure that any intervention activities with pressurised hydrocarbon systems are executed only after depressurisation and purging is confirmed. 178 Issue 1. which may be used in potentially explosive atmospheres (flammable gases. including the critical aspects of system design.4.4. pressure relief panels and fast-acting shut-off valves. nitrogen leak testing is recommended to confirm joint integrity. Operating procedures should recognise the explosion hazards associated with the misuse or poor maintenance of equipment. the designer can assist with the continuation of existing safe working practices and safety records by using the same equipment and material specifications as are prevalent. Inspection and maintenance procedures should be regularly reviewed to assure that best current practices are considered. method of tightening (torqueing or tensioning) and leak testing.2 Bolted Pipe Joints An essential aspect of maintaining integrity in mechanical joints is correct joint design and analysis. It will set out the principles of joint integrity management and provide examples of best practice to assist operators in develop their own management systems. Also included in the term “equipment” are safety or control devices installed outside the hazardous area but having an explosion protection function. failure will result. 7.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions In designing modifications to existing installations. maintenance and operation. vapours or dusts).4 7. Inspection methods should recognise the function and performance of insulation or fire protection systems and ensure that the performance is not compromised by removal of such systems in order to gain inspection access. this includes sequence of bolt tightening. Following reinstatement of systems. Training and competency are significant contributors to successful safe operating practices and demonstrate that good design alone is not sufficient where manual intervention is required with hydrocarbon systems. in addition to consistent sparing philosophies and tool sizes as are in use for that installation. The term “equipment” is defined as ‘any item which contains or constitutes a potential ignition source and which requires special measures to be incorporated in its design and/or its installation in order to prevent the ignition source from initiating an explosion in the surrounding atmosphere’. These directives have been adopted by the European Union (EU) to facilitate free trade in the EU by aligning the technical and legal requirements in the Member States for products intended for use in potentially explosive atmospheres. These factors all influence the integrity and pressure retention capacity of the joint and if any or all of these are not rigorously followed. October 2003 . A UKOOA guideline  is in preparation that covers the lifecycle activities of bolted pipe joints. construction. 7. quenching systems.1 Industry and Regulatory Authority Initiatives ATEX Directives The ATEX (Atmospheric Explosion) Directives 94/9/EC and 1999/92/EC cover electrical and mechanical equipment and protective systems. They include flame arresters. whilst retaining the intended operating philosophy and design parameters.
All Shell installations have been reviewed for compliance with this strategy. Tubing and Flexible Hoses These components are commonly used in process systems both for instrument connections and other process duties. As such. Design. which could lead to uncontrolled releases. such as vibration. the adoption of these guidelines go some way to assuring the integrity of tubing systems under operational loading. environmental impact and production or equipment losses.4. such as mechanical impact. October 2003 179 . which are a further cause of releases. They require special treatment due to this and the fact that it is often difficult to provide assurance of their capacity when exposed to accidental loads such as blast loads. This system uses a risk matrix to categorise the consequence of hose failure in terms of human injury.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 7. In actuality the loads derived from large differential movements between supports on these systems can be critical. as the received explosion loads are of small magnitude. Shell’s flexible hose management system has been offered as a starting point. and should be adopted as a basis for good practice for resistance to accidental loads. Most concerns with these systems are associated with vibration and fatigue. Installation and Maintenance of Small Bore Tubing Systems” ) records that "small bore tubing systems are the single largest contributor to the incidence of process containment rupture in potentially hazardous plants".3 Small Bore Pipework. Small bore piping and tubing is often site run without any particular assessment of the strength and integrity of the configuration. but then the additional released inventory may not be significant in terms of the unfolding event. A common definition of small bore is 50mm and under. A joint UKOOA/HSE workgroup is to be set up to review general industry best practice on flexible hose management. Issue 1. An inspection and hose replacement frequency is determined by the hose classification and performance standards for hose condition and competency standards for operators have been established. The recently published “Guidelines for the Management.
SCI 2001 FABIG Technical note 7. ‘Technical note on Explosion mitigation systems’. ‘Updated Guidance for fire and explosion hazards – Part 1 Avoidance and mitigation of explosions. Z-013. Genesis Oil and Gas Consultants. Engineering Handbook. October 2003 . FABIG Technical note 3. Ed. ‘Use of ultimate strength techniques for fire resistant design of offshore structures’. 1998 Johnson D. Pappas J. ‘Updated Guidance for fire and explosion hazards – Part 1 Avoidance and mitigation of explosions. ‘Design Guide for Steel at Elevated Temperatures and High Strain Rates’. Czujko J. FABIG Technical note 4.A. Cleaver R.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 8  REFERENCES Bowerman H. Design of Offshore Facilities to Resist Gas Explosion Hazard. ERA Report 2001-0575.1. Owens G W. Tolloczko J. SCI 1995. June 1999. ISBN 1 85942 078 8.. SCI 1996 FABIG Technical note 5. Steel Construction Institute Document SCI-P-112/487. CTR 103.. January 1992. Bleach R. Advantica report R4853. August 2002. SCI 1993. Burgan B A. Interactions with fire hazard management’. Selby C A. FABIG Technical note 1. Rumley J H.M.. Explosion hazards philosophy’. Steel Construction Institute. ‘Blast and Fire Engineering for Topside Structures – Phase 2. FABIG Technical note 6... ERA Conference. Genesis Oil and Gas Consultants. ‘Explosion resistant design of offshore structures’.P. The NORSOK Procedure on Probabilistic Explosion Simulation. CTR 104.. Bleach R..5.. August 2002. ‘Design guide for stainless steel blast walls’. ‘Gas explosions in offshore modules following realistic releases (Phase 3B) – Final summary report’. February 2002. CTR 107.J. ‘Industry Guidelines on a framework for risk related decision support’. Final Summary Report’. UKOOA. August 2002 Bleach R.. Genesis Oil and Gas Consultants. February 1999. Guideline for Risk and Emergency Preparedness Analysis. Paper 5. ‘Major Hazards Offshore’.                 180 Issue 1. FABIG Technical note 2. ‘Interim Guidance Notes for the Design and Protection of Topside Structures Against Explosion and Fire’. NORSOK. Management of explosion hazards’. ‘Simplified methods for analysis of response to dynamic loading’. CorrOcean 2001. ‘Technical note on Fire resistant design of offshore structures'. Issue 3. ‘Updated Guidance for fire and explosion hazards – Part 1 Avoidance and mitigation of explosions. 2002. SCI 1994. London 27-28 November 2001.
CTR 105. ‘Worldwide offshore accident database’. ‘Updated Guidance for fire and explosion hazards – Part 1 Avoidance and mitigation of explosions.. ‘Updated Guidance for fire and explosion hazards – Part 1 Avoidance and mitigation of explosions. HSG181. Guidance on detailed design’. Response to explosions’. DnV Publications. Derivation of explosion loadings – method support dossier’. HMSO. HSE books.. ‘Hydrocarbon release reduction campaign’.. ISBN 0 7176 1238 4. ‘Offshore Installations (Prevention of Fire and Explosion and Emergency Response) Regulations’ (PFEER). ‘Assessment principles for offshore safety cases’ (APOSC). HMSO. Penwell Books. HSE OSD report OTO 2001 055. HMSO 2001. ‘Updated Guidance for fire and explosion hazards – Part 1 Avoidance and mitigation of explosions. ‘Explosion assessment guidelines’. UKOOA. Atkins Inc. IC Consultants. 1998. Gall D. report J6838 for the HSE June 2001.) Regulations’ (DCR). HMSO 1998.A. ‘OREDA – Offshore reliability data’.A.. HMSO 1996. and Carney S. ‘Updated Guidance for fire and explosion hazards – Part 1 Avoidance and mitigation of explosions. IC Consultants and Aker Kværner. Aker Kværner. (SI 1995/738). January 1998. Friis J. (MAR) ‘Petroleum and natural gas industries – Control and mitigation of fires and explosions on offshore production installations – Requirements and guidelines’. HSE. Friis J. ‘Offshore Accident and Incident Statistics Report 1998’.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions  Fairlie G.. Preparation of updated guidance for fire and explosion hazards – Derivation of explosion loadings’. August 2002 Louca L. Genesis oil and gas consultants. August 2002.. W. OTO 2002 012. HMSO 1992. (IP 15)                  Issue 1. CTR 108. HMSO 2002.. ‘Offshore Installations and Wells (Design and Construction. ISO 13702. HSE Books. ‘Accident statistics for fixed offshore units on the UK continental shelf 1991–1999’.S. ‘Area Classification Code for Petroleum Installations’. DnV for the HSE.. Institute of Petroleum.. August 2002 Warwick A.J. HMSO. 1999.J. October 2003 181 . Louca L. 1st Edition. Century Dynamics Ltd. CTR 105. (WOAD). OTO 98: 952. etc. ‘The offshore installations and pipeline works (Management and Administration) Regulations. ‘UKOOA Hydrocarbon release statistics review’.. 1995 ‘Offshore Installations (Safety Case) Regulations’ (SCR). August 2002 Carney S. March 1990. CTR 106.
Policy and Guidance on reducing risks to ALARP in Design. HSL Report. and Lowesmith B.pdf . ‘Electrostatic hazards associated with water deluge and explosion suppression systems offshore’. (Phase 3A) HSE books OTO 1999 043.. ‘Guide for Pressure-Relieving and Depressuring Systems’. ‘Code of practice for control of undesirable static electricity’. ‘Recommended Practice for the planning.. http://www. OTO 95 026.. 1991. June 2002.              182 Issue 1. ‘Explosions in full scale offshore module geometries – Main report’.uk/dst/alarp1. Classification of hazardous areas’. ‘JIP – Gas build up from high pressure natural gas releases in naturally ventilated offshore modules – Workbook on gas accumulation in a confined and congested area’. Evans J. BS EN 60079-10. 4th Edition. ‘White paper – working group 2 – Safe design practice’. Alderman J. International workshop ‘Fire and blast considerations in the future design of offshore facilities’. ‘Successful health and safety management’.htm Principles and Guidelines to Assist HSE in its Judgement that Duty Holders Have Reduced Risk as Low as Reasonably Practicable.uk/dst/r2p2.A. Carter D. ISBN 0 7176 2151 0.fireandblast2002. 21st Edition.M.uk/hid/spc/perm12. http://www. Explosion prevention and protection’ BS EN 1127-1 : 1998.hse.A. International workshop ‘Fire and blast considerations in the future design of offshore facilities’. HSE books. Guidance Note PM84 HSE. May 2000. ‘Guidelines for instrument-based protective systems’. March 1997. et.gov.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions     ‘Electrical apparatus for explosive gas atmospheres Part 10.. BP Amoco. HSE web site http://www. ‘Control of Risks at Gas Turbines Used for Power Generation’. API RP 2A-WSD.htm ‘Reducing risks protecting people. HSE’s decision making process’ (R2P2). http://www.gov.hse. Designing and Construction Fixed Offshore Platforms – Working Stress Design’. 1999. ‘White paper – working group 1 – Philosophy and management strategy’. HSE books HS(G) 65.fireandblast2002. API (American Petroleum Institute).. Al. June 2002. FS 99/09. Houston Texas.hse. October 2003 . BS 5958 : 1991. UKOOA November 1999.com ‘A Review of Potential Ignition Sources on Offshore Platforms’. Wishart J. BG Technology.com API Recommended Practice 521. 1995. http://www. Johnson D.gov. Issue 2. CERC. HMSO 2001. ‘Explosive atmospheres. Houston Texas.
J. Response. Analysis and Design’. ‘Structural Analysis and Design of Nuclear Plant Facilities’. 'The Escalation Consequences of Accidental Shock Loads'. ‘Early design blast analysis’. 1990. Australia. Safety on Offshore Installations’. O’Doherty T. D. Gregory J. March 2000 ‘Control of risks at gas turbines used for power generation’. RISX Ltd. 99-0808.W. May 1997.J. Response. ‘Modelling the Internal/External Interaction from Vented Explosions. FABIG Technical meeting review. Report TN006.Y. R. Health and Safety Laboratory Report CM/00/04. 15 – 16 November 1994. Ledin H. Conference and Exhibition – Offshore Structures Design – Hazards and Safety Engineering.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions  Natabelle Technology Ltd. ‘Blast and fire engineering project for topside structures’. Bowen P. Tam. Nicholson. 13th Australasian Fluid Mechanics Conference. 3rd Intl. SCI. Analysis and Design’. ‘Explosion Loading on Topsides Equipment: Part 2 – Treatment of Explosion Loads. ‘The Design and Testing of a Louvered Blast Relief System’. London. 15 – 16 November 1994. November 1994. ‘The effectiveness of area water deluge in mitigating offshore fires’. ‘Fire and Explosion Aspects of FPSOs’.H. London. November 1998. International Coatings Ltd. S. October 2003 183 . N. and Tandberg. Tahan. Lea C. Health and Safety Executive Report OTO 1999 046. London. Walker S. OTO 200 051.                 Issue 1.. ERA Report. FABIG Technical Meeting Notes.. ‘Barrier Methods for Gas Explosion Control. Walker S and Klair M.. Report FR6 ‘Fire performance of explosion damaged structural and containment steelwork’. Safety and Engineering... December 1999 ‘Review of the response of pressurised process vessels and equipment to fire attack’. 1980. ‘The Public Enquiry into the Piper Alpha Disaster’. Morrison G. Melbourne. ‘Explosion Loading on Topsides Equipment: Part 1 – Treatment of Explosion Loads. ‘Strong Vibration and Design for Robustness’ HSE – ERA Conference Offshore Structural Design – Hazards. Report RISX/12-01-01/01. HSE. 3rd Intl. ASCE 58. ‘A Review of the State-of-the-art in Gas Explosion Modelling’. 1998. V. Health and Safety Executive Report OTO 1999 047. 1991. 19 February 2002 Marsano S. HMSO. March 2000 Natabelle Technology Ltd. Cullen The Honourable Lord. 13-18 December. ‘Explosion Resistance of Passive Fire Protection’. Conference and Exhibition – Offshore Structures Design – Hazards and Safety Engineering. HSE Guidance note PM84...
. pp315-335. Cleaver R P. Major Hazards Offshore Conference.H.J.C. ‘Validation of scaling techniques for experimental vapour cloud explosion investigations’. A 1990 431. 1998. G. Renoult J . ‘Offshore Standard DNV-OS-A101. ‘Studies of the turbulent burning velocity’. Halford A R. Process Safety Progress. 167-173. Robinson C.. ‘Methods for increasing the Accuracy of Explosion pressure predictions in early phase of new topsides projects’. ‘The use of a combination of explosion models in explicit overpressure exceedence calculations for an offshore platform’. Harris R. November 2000. 2. September 2001. May 1992 Mercx W P M.. ’ Analysis of results from large scale hydrocarbon gas explosions’. 1995 Shearer M J. Loss Prev. Mercx W P M. Loss Prev.. J. Lees and Ang . January 2001 Puttock J S. 1983. November 1990 Walker S. Appendix E Phase 3B – Lessons Learnt from CFD’. 2000. Tam V Y H. 1994 Hansen O R. Proc. Process Ind.. D. 14. London. Puttock J. Johnson D M.N. Det norske Veritas. J. Humphreys C.. 2000 CREA Consultants Limited. C – TM5-1300. Advantica Report Numbers R4853 and R5107. 2. 14. ISBN 0 419 13220 1. No..UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions   Structures to Resist the Effects of Accidental Explosions – Department of Army. BP Corporation.. Johnson D M. ‘Gas Explosions in Offshore Modules Following Realistic Releases (Phase 3B): Final Summary Report. 120-130. London. Document Reference CH152R002 API Draft 0. Puttock J. pp 273-280.. ‘Analysis of Structural Response Measurements – Phase 3B Spadeadam’. October 2003 .Wingerden K. Tam V. Roy. B. Process Safety Progress. Washington. ‘The investigation and control of gas explosions in building and heating plant’. Soc. Safety Principals and Arrangements’. 7. ‘Classification of Hazardous Locations’ E&P Forum. Process Ind.. ‘Accidental generation of gas clouds on offshore process installations’. E. ‘Validation of scaling techniques for experimental vapour cloud explosion investigations’. Gexcon Report Number Gexcon-01-F36018-1. 2001 Cox A W . HSE Report OTH 541. ‘Guidance for the protection of Offshore Structures against Fires and Explosions’. ‘Hydrocarbon Leak and Ignition Database’. Advantica Report Number R4853. 1995 Bray K. Bergonnier S.Y. Appendix H An Analysis of the Phase3B Large-Scale Gas Explosion Experiments’. 13. 120-130. Corr R. ‘Gas Explosions in Offshore Modules Following Realistic Releases (Phase 3B): Final Summary Report. Corr R. Brewerton R W. No. OTO 055/2000. 2001 Cleaver R P. E&FN Spon Ltd.B.               184 Issue 1.
B. Puttock J S. NPD. ‘Development of the limit state approach for the design of offshore platforms’ ERA Conference. Paper 3.A.. Clough. Conf. 3-4 November 1993. ‘Dynamics of Structures’. HSE L39. HMSO 1992. ERA Conference. 1996. J. Journal of Loss Prevention in the process industries.B. Structural design. Corr R B.1.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions     Hoiset S. AIChE Process Safety Progress 17. Walker S. ‘Major Hazards Offshore’.. 1999. D.. 1975. ERA Report 2001-0575. ‘The blast loading imparted to a cylinder by venting of a confined explosion’. Rev 3. ERA Major Hazards Offshore Conference. Hazards. 1999 ‘Acts. Simplified Approaches to Fire and Explosion Engineering. Hjertager B H. Regulations and Provisions for the Petroleum Activity’. ‘Properties of simulated gas explosions of interest to the structural design process’. August 2000. ‘Gas explosion drag loads in offshore installations’.H. San Francisco...A. Tam V Y H. ‘Blast Overpressure Prediction – Modelling the Uncertainties’. November 1998.E. and Tam V. Graybill. London.. ‘Target explosion loads for offshore installations – strength level and ductility level events’. Snell R O.F. Solberg T. Offshore Structural Design – ERA conference. Paper R429..Y. NORSOK Standard N-001. ‘Improvements in guidelines for prediction of vapour cloud explosions’. FABIG newsletter. September 1999 Bruce R L. October 2002. ‘A Guide to the offshore installations (Safety Case) regulations 1992’.. and O’Connor P. Boes.H. 3rd Edition. SCI publication. R.               Issue 1. ‘Blast load frequency distribution assessment of historical frequencies in the North Sea’. M. McGraw-Hill.(ACOP) Catlin C.Y. ‘Introduction to Structural Dynamics’. Walker S and Haworth M. Offshore structural design against extreme loads. Volume 2. November 2000.3. 1964. 11 (1998) 43-48. ‘COSAC.. Corr R. ERA Conference. ISBN 0 7008 0587 7. December 1992. Yasseri S. Frieze P A. FABIG Newsletter January 2002.. Yasseri S. Biggs J.. Penzien. The Tool for Efficient Hazard Identification and Risk Assessment in the early phase of a field development’. London. Bucknell J. et al.. October 2003 185 . Tam V. Lund J K. R. HSE OTO 97 043. McGraw-Hill.M.W. Preventor report 19816-04.. McGraw Hill.’ Sensitivity of response of topside structures to fires and explosions’. Corr R. An Approximate Method for Blast Resistant Design.. Mood. Int. London.W. Malo K A. ‘Introduction to the theory of statistics’.C. Safety and Engineering London November 1994. 1974. Vinnem J. and Workshop on Modelling the Consequences of Accidental Releases of Hazardous Materials.F. London 27-28 November 2001.A.
‘Blast and fire engineering project for topside structures’.G. Second impression’. Corr BR.S. J. 17th conference on Offshore Mechanics and Arctic Engineering. 1997. N. and Birch. Harwood R. Taylor P. R. BS EN 10225: 2001 ‘Weldable structural steel for fixed offshore structures – Technical delivery conditions’. January. W. ‘New guidance on fire and explosion engineering’. The Steel Construction Institute. ‘Elevated Temperature and High Strain Rate Properties of Offshore Steels’. Jan 2002.H. October 2003 . Cambridge University Press. ‘Structural Impact’. BRI. ‘The effects of simplifications of the explosion pressure-time history. Oslo. API Bulletin 2V (BUL 2V) 1st Edition. 1987. Wiley. Japan 1989. Kato B. ‘Stainless steel – Technical delivery conditions’. ‘Bulletin on design of flat plate structures’. Bucknell J. Stewart G. 1988. ‘Rotation capacity of steel members subject to local buckling’. Tam VHY. 1991 ‘Blast and fire engineering project for topside structures’. OTI:605 ‘Methodologies and available tools for the design/analysis of steel components at elevated temperatures’. – 2nd Edition. 1991. D.    . 1991 ‘Blast and fire engineering project for topside structures’. E. O’Connor PA. et al – Elsevier Science Publishers. Izzuddin. B. Report FR2. ‘Simplified Methods for Dynamic Response Analysis’. OTI:602 ‘The effects of high strain rates on material properties’. May 1 1987. 2002.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions    ‘Blast and fire engineering project for topside structures’. Earthquake Resistant Design for Engineers and Architects – Dowrick.. OTO Report 2001/020 EN 10088-2:1995..            186 Issue 1. UKOOA.. Proc. 1991. Jones. Norway.. OMAE 2002... June 23-28. Report BR4. paper OMAE02-28623. Explosion Hazards and Evaluation – Baker. N. ‘Blast wall strength estimated using finite displacement theory’. 1998. Walker S. Jones. 9th World Conference on Earthquake Engineering. OTI:604 ‘Experimental data relating to the performance of steel components at elevated temperatures’. Norsok standard Annex A. N-004 BS 6399 – Part 1. – FABIG Technical Meeting. ‘Loading for Buildings – Code of Practice for Dead and Imposed Loads’. Report FR1. Guidelines for The Management of Integrity of Bolted Pipe Joints. OMAE98-1423.
.. No. October 2003 187 .2. June 2000. K. Retrofitting of Blast walls for Offshore Structures’. H. Constructional Res. and Evans. and Berge Dr. HSE Offshore Division Safety Notice 2/2002 May 2002 “Balanced bellows pressure relief valves – problems arising from modification of the bonnet vent”. M. Issue No. 2001. 1967 Yasseri.. A. Guidelines for the Management. pp. ‘Fire and Blast Information Group Newsletter. Thin Walled Structures’.. and Harding. ERA Major Hazards Offshore Conference Nov. 1996.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions  Louca. 37. ‘The Behaviour of Corrugated Flooring Systems.      Issue 1.. Vol. S. Institute of Petroleum.F. 93-113. Rockey. ‘Why API Recommended Practice may Lead to Unsafe Design of Pressurised Systems’. L. Medonos S.C. ‘Non-Linear Analysis of Blast walls and Stiffened Panels Subjected to Hydrocarbon Explosions’ – J.R. Crosby Lockwood. J. 1996. G. Punjani. E. Installation and Maintenance of Small Bore Tubing Systems.15. Design.
Procurement. October 2003 188 . Evacuation and Rescue Euro Norm – International Standards Organisation Emergency Overnight Accommodation Emergency Preparedness Analysis Engineering. Construction and Installation Emergency De-Pressurisation Valve Issue 1.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions APPENDIX A – ACRONYMS ach AE AFFF AIR ALARP ALS API APOSC BDV BOP BSI CAD CAM CBA CCTV CFD CRA DAE DAF DAL DCR DIF DLB DNV DP DSHA E&P EDPV EER EN ISO EOA EPA EPCI EPDV air changes per hour Accidental Event Aqueous Film Forming Foam Average Individual Risk As Low As Reasonably Practicable Accidental Collapse Limit State American Petroleum Institute Assessment Principles for Offshore Safety Cases Blowdown Valve Blowout Preventer British Standards Institution Computer Aided Design Congestion Assessment Method Cost Benefit Analysis Closed Circuit Television Computational Fluid Dynamics Concept Risk Analysis Dimensioning Accidental Event Dynamic Amplification Factor Dimensioning Accidental Load Design and Construction Regulations Dynamic Increase Factor Ductility Level Blast Det Norske Veritas Dynamic Positioning Defined Situations of Hazard and Accident Exploration And Production Emergency Depressurisation Valve Escape.
UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions ERA ESD ESD ESDV F&G FAR FEA FEED FES FMEA FPSO FRA GBS HAZID HAZOP HC HCLIP HES HIPPS HMSO HRA HS&E HSE HSMES HVAC ICAF IMO IR IRPA ISO JIP LCC LEL LIER LQ MDOF Early Risk Analysis Emergency Shutdown Emergency Shutdown Emergency Shutdown Valve Fire And Gas Fatal Accident Rate Fire and Explosion Analysis Front End Engineering Design Fire Explosion Strategy Failure Mode and Effect Analysis Floating Production Storage and Offloading Fire Risk Analysis Gravity Base Structure Hazard Identification study Hazard and Operability study Hydrocarbons Hydrocarbon Leak and Ignition Project Health. Environment and Safety High Integrity Pressure Protection System Her Majesty's Stationary Office Health Risk Assessment Health Safety & Environmental The UK Health & Safety Executive Health. Safety and Environmental Management Heating Ventilation & Air Conditioning Implied Cost of Averting a Fatality International Maritime Organisation Individual Risk Individual Risk Per Annum International Standards Organisation Joint Industry Project Life Cycle Cost Lower Explosive Limit Local Instrument and Electrical Room Living Quarters Multi-Degree Of Freedom Issue 1. October 2003 189 .
Utilities. October 2003 190 . Drilling. Quarters (Similarly PUQ) Process Flow Diagram Prevention of Fire and Explosion and Emergency Response Potential Loss of Life Persons On Board Quantitative Risk Analysis Research And Development Risk Acceptance Criteria Risk Based Inspection Reliability Centred Maintenance Risk Reduction Measure Relief Valve Search And Rescue Scenario Based System Design Stand-By Vessel Safety Critical Element Safety Case Regulations Single Degree Of Freedom Safety Integrity Level The Foundation For Scientific And Industrial Research At The Norwegian Institute Of Technology Safe Job Analysis Strength Level Blast Safety Management System Issue 1.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions MODU NCS NFPA NLFEM NMD NNS NPD NPV NTS NUI OLF PA / GA PDO PDUQ PFD PFEER PLL POB QRA R&D RAC RBI RCM RRM RV SAR SBSD SBV SCE SCR SDOF SIL SINTEF SJA SLB SMS Mobile Offshore Drilling Unit Norwegian Continental Shelf National Fire Protection Association (Usa) Non-Linear Finite Element Method Norwegian Maritime Directorate Northern North Sea Norwegian Petroleum Directorate Net Present Value Norwegian Technology Centre Normally Unattended Installation Oljeindustriens Landsforening (The Norwegian Oil Industry Association ) Public Address / General Alarm Plan for Development and Operation Process.
October 2003 191 .UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions SNS TBL TEMPSC TR TRA TREPA UK UKCS UKOOA UPS VCE WSV Southern North Sea Federation Of Norwegian Manufacturing Industries Totally Enclosed Motor Power Survival Craft Temporary Refuge Total Risk Analysis Total Risk and Emergency Preparedness Analysis United Kingdom United Kingdom Continental Shelf UK Offshore Operators' Association Un-Interruptable Power Supply Vapour Cloud Explosion Written Scheme of Verification Issue 1.
Mitigation technique in which the release of suppressant is triggered as a result of flame or gas overpressure detection. injection or expulsion – in order to combat smoke. Active Fire Protection A fire protection method that requires activation – switching on. equipment. or damage to health. For instance. As Low As Reasonably Practicable. Active Mitigation ALARP ALARP Artificial ventilation As low as reasonably practicable (ALARP) Auto-Ignition Temperature Availability 192 Issue 1. To reduce a risk to a level which is 'as low as reasonably practicable' involves balancing reduction in risk against the time. trouble. or system is performing in the desired manner. planned operational exposure that may be hazardous to health or to the environment. a concept that forms the primary basis for determining the tolerability of risks and the adequacy of arrangements for managing risks to health and safety. trouble. difficulty and cost of achieving it. objectively assessed. The requirement to establish a cost effective solution implies that risk reduction is implemented until the cost of further risk reduction is grossly disproportional to the risk reducing effect. That ventilation which is not supplied from the action of the environmental wind alone ALARP expresses that the risk level is reduced (through a documented and systematic process) so far that no further cost effective measure is identified. Duty holders (The owner or operator of an existing installation or the designer of a new one) have a responsibility to reduce risk to ALARP . flames or thermal loadings. October 2003 . difficulty and cost of further reduction measures become unreasonably disproportionate to the additional risk reduction obtained. unwanted and unplanned.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions APPENDIX B . This level represents the point. the environment or assets NOTE 1 . directing.GLOSSARY OF TERMS Accident Accidental event (AE) See 'incident' Event or chain of events that may cause loss of life.The events that are considered in a risk analysis are acute. are usually not included in a risk analysis. The minimum temperature of a vapour and air mixture at which it is marginally self igniting. at which the time. The proportion of the total time that a component.
The average chance of any individual in a defined population sustaining a given level of harm from incidents which are considered to be limited to that population. Verbal form used for statements of possibility and capability. the result of a catastrophic rupture of a pressurized vessel containing a liquid at a temperature above its normal boiling point. The fraction of an area which does not provide open venting or free passage for a flame. October 2003 193 . A structural division which has been designed expressly for the purpose of resisting blast loads. (Sometimes called Individual Risk – IR) Parts of a module wall.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Availability The probability of the system being available to perform the function when required. possibly with a substantial fallout of heavy fractions which may also ignite on or around the platform as pool fires. Boiling Liquid Expanding Vapour Explosion. ceiling or roof which are designed to increase the area of venting in an explosion by being opened or removed by the force of the explosion. The sudden rupture due to fire impingement of a vessel system containing liquefied flammable gas under pressure. Also the velocity of the flame relative to the unburnt gas/vapour. running from the module deck to a sub sea level.a blast wave and potential missile damage. intense fireball. A form of jet resulting from ignition of a high pressure flow of oil and/or gas issuing from an uncontrolled well. The rapid controlled or accidental depressurization of a vessel or network. forming a service line for seawater suction or sewage/drilling disposal. whether material. A vertically aligned pipe. physical or causal Average Individual Risk (AIR) Blast Relief Panel Blast Wall Blast Wave BLEVE Blockage Factor Blowdown Blow-out (ignited) Boiling liquid expanding vapour explosion (BLEVE) Burning Velocity Caisson Can Issue 1. Simultaneous ignition of the vaporizing fluid gives a short duration. A pressure pulse formed by an explosion. The pressure burst and the flashing of the liquid to vapour creates . and immediate ignition of the expanding fuel-air mixture leads to intense combustion creating a fireball The rate at which the flame consumes the unburnt gas/vapour.
emergency power supplies) safety hardware (e. warmer. in the exploration for and production of oil and/or gas. pressure relief valves. or the presence of manually or automatically initiated ESD procedures which are intended to contain a developing situation so that escalation and a major accident may be avoided. The failure of components from the same cause An organisation engaged. October 2003 . A quantitative technique to assess the overall value of a proposal taking into account its likely benefits and detriments. although its intensity may ultimately reach or exceed that of a hydrocarbon fire. dump tanks.g. A mathematical model in which the region of the flow is subdivided by a grid into a large number of control volumes. as principal or contractor.g. a single site may be defined as a company.g. cotton). less dense fluid rises and is replaced by cooler. Heat transfer associated with fluid movement around a heated body. Chemical Composition Common Mode Failure Company Computational Fluid Dynamics (CFD) Confinement Confined Explosion Congestion Contractor Control Convection (Heat) Convection Cost Benefit Analysis (CBA) 194 Issue 1. A fire involving these materials is relatively slow growing. paper. An explosion of a fuel-oxidant mixture inside a closed system (e. directly or indirectly. vessel or module). coolant sprays). Transfer of material (fluid or gas) from one place to another. Confinement is defined as a measure of the proportion of the boundary of the explosion region which prevents the fuel/air mixture from venting or being released to the external atmosphere. For bodies or establishments with more than one site. company or other association that performs work for an Operator or in other ways is a supplier of products or services to the Operator.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Cellulosic Fire A fire with a fuel source predominantly of cellulose (e. timber. Means of intervention permitted by the design (e. The proportion by weight of constituent elements in a particular batch of steel as determined by ladle analysis.g. Congestion is a measure of the restriction of flow within the explosion region caused by the obstacles within that region. more dense fluid. Person.
in the context of this report. It provides the practical tool to aid decision making. It is a relatively slow process and the velocity is always less than the speed of sound in the mixture (see Detonation). Mechanism for propagation of an explosion reaction through a flammable gas mixture which is thermal in nature. either because they directly affect option performance or because they determine the context in which the decision needs to be made. Criticality (of SCEs) Decision Basis/Bases Decision Context Decision Factors Decision Framework Decision Option(s) Deflagration Deflagration Design accidental event Design basis checks Issue 1. An accidental event for which SCEs on the installation should perform their function as designed. The different options or solutions which are to be considered as means to resolve the problem. the term explosive deflagration may be used. at a fixed temperature. October 2003 195 .UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Creep The time dependent straining of a material. The background or environment in which the decision needs to be made. EER systems and associated supporting structure. Aspects of performance or influencing factors that can have a bearing on the decision. The decision context affects the importance of the various decision bases. The decision framework consists of the model and accompanying guidance on the decision context and means of calibration. The basis/bases for assessing the performance of decision options. Design basis checks consist of checking the basis of the existing design for the installation and determining if the methods used for the design are acceptable in the context of the explosion hazard. The chemical reaction of a substance in which the reaction front advances into the unreacted substance at less than sonic velocity. Where a blast wave is produced which has the potential to cause damage. These range from technical codes and standards based ways of assessing options to values based assessments based on company or wider societal values and stakeholder expectations and perceptions. Defined by the criticality classes 1 (highest) to 3 (lowest) with respect to the explosion hazard and the consequences of failure. For the purposes of the model the context is represented as a gradual transition from a technical basis to a values basis. In particular the impact on the TR. It is measured under a constantly applied load and.
frequency level) – Pduct. Detonation Detonation Detonation Limits Diffracted Wave Diffusivity Dilution Dimensioning accidental event (DAE) Dimensioning accidental load (DAL) Dimensioning explosion loads Drag Coefficient Drag force Ductility level blast (DLB) Ductility level design event 196 Issue 1. Representative peak overpressure used in design (10-4 to 10-5 p.a. Load (action) that is sufficient in order to meet the risk acceptance criteria. The drag load on a small obstacle due to the movement of gas past a small obstacle less than 0. SCEs must be designed to resist these load levels within the constraints of the associated element specific performance standards.3 m in the direction of flow. the code check results in members dimensioned to resist the ductility level explosion. Mechanism for propagation of an explosion reaction through a flammable gas mixture which is more mechanical in nature and acts through shock pressure forces. See DLB. form drag – Presented Area x Pdrag. Explosion loads which are of such a magnitude that when applied to a simple elastic analysis model. An expression relating to the range of mass transmitted through a given medium to that absorbed by the medium. It has units m2 s-1.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Design explosion loads Explosion loads used for design. in order Ito meet the defined RAC. That component of the blast wave which propagates into the sheltered region behind the structure. It is a very rapid process and velocity is always greater than the speed of sound in the mixture (see Deflagration). The range of fuel-air ratios through which detonations can propagate. October 2003 . AEs that serve as the basis for layout. dimensioning and use of installations and the activity at large. An empirical multiplying constant used to relate the drag load on a structure to the stagnation overpressure. Addition of inert gas to flammable mixture. An explosion caused by the extremely rapid chemical reaction of a substance in which the reaction front advances into the unreacted substance at greater than sonic velocity.
includes drag. establishment of performance standards for emergency preparedness and their fulfilment and identification of emergency preparedness measures . The ratio of the maximum displacement of the element to the deflection required to cause first yield at the extreme fibres. The process by which surrounding fluid is drawn into a jet to cause its dilution. Emergency preparedness Technical. Measurable performance standard for specific key items or systems relating to systems’ functionality. which can. The organisation. pressure difference and gas acceleration effects. valves and actuators. Used for calculation of loads on equipment and piping – Pdyn.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Ductility Ratio The ratio of the peak deflection to the deflection at first effective yield. pipe or marine riser specifically for the purpose of cutting of the supply of its normal contents in an emergency. enable the safe and effective shutdown of plant and machinery in a controlled manner. signalling and logical control. Emissivity of a perfect black body is 1. operational and organisational measures that are planned to be implemented under the management of the emergency organisation in case hazardous or accidental situations occur. A valve mounted in a pipeline. Analysis which documents the fulfilment of performance standards for safety and emergency preparedness. October 2003 197 . Ductility ratio Duty Holder Dynamic pressure Effectiveness analysis Emergency preparedness analysis (EPA) Emergency Shutdown System Emergency Shutdown Valve (ESDV) Emissivity Entrainment Issue 1. in order to protect human and environmental resources and assets. A safety shutdown system comprising detection. A constant used to quantify the radiation emission characteristics of a flame. in tandem with alarm and direct control mechanisms. Representative peak out of balance loads over target area. availability and survivability. (Sometimes referred to as low level performance standard). Analysis which includes establishment of OSHA. sometimes referred to as dynamic overpressure. including major AEs. or company responsible for the safety of the installation or its design under UK legislation.
Events for which design is not reasonably practicable may result in loss of functionality of SCE’s The fraction of the net combustion energy of a flame transmitted as radiation. including living systems (human and other) therein. A documented evaluation of the environmental significance of the effects of the company's activities.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Environment The surroundings and conditions in which a company operates or which it may affect. (see gas explosion) risk from the initiating event and possible subsequent escalation. The ratio of burnt to unburnt gas volumes of a given mass of gas. see above A plot of the value of a variable against the plot of the probability or frequency of exceedance of that variable. whether adverse or beneficial. A direct or indirect impingement of the activities. External or secondary explosions which may result as an unburnt fuel/air mixture comes into contact with the external (oxygen rich) atmosphere as it is vented from a compartment. Escalation Essential safety system Establishment of emergency preparedness Exceedance curve Expansion Ratio Explosion Explosion risk External explosion Extreme (residual) events F Factor 198 Issue 1. products and services of the company upon the environment. October 2003 . products and services (existing and planned). Environmental effect Environmental effects evaluation EPA Equivalent static overpressure An overpressure derived from an overpressure trace with respect to a target which gives the same peak deflection when applied as a static load as the original trace. Consequences subsequent to an explosion or other major hazard. A release of energy which causes a pressure discontinuity or blast wave. System which has a major role in the control and mitigation of accidents and in any subsequent EER activities Systematic process which involves planning and implementation of suitable emergency preparedness measures on the basis of risk and EPA.
An integrated detection. Conventionally. burning sufficiently rapidly for the burning mass to rise into the air as a cloud or ball. October 2003 199 . The temperature at which a liquid fuel flashes and then marginally sustains a flame. such tests are carried out in purpose-built furnaces operating to a defined temperature/time curve. The length of time that an element can resist fire either up to the point of collapse or alternatively to the point when the deflection reaches a limiting value. This phenomenon. a fire. Fire Fire Ball Fire Curve Fire Endurance Fire Point Fire Prevention Fire Protection Fire Protection System Fire Test Fire Wall Fire Water Pump Flammability Limits Issue 1. The term can apply to a floor or roof panel. mounted in or capable of being lowered into a fire water pump caisson. A temperature/time curve devised for a fire tests that is intended to represent (but not necessarily reproduce) the temperature development of a ‘real’ fire. A designated partition. A submersible electric or diesel pump. which feeds into the firewater ring main. Alternatively. signalling and automated fire control system. which by nature of its construction and certification status. Measures taken to minimize effects of damage from fire should it occur. Measures taken to prevent the outbreak of fire at a given location.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Fail-Safe Valve An actuated valve which automatically assumes a safe position (open or closed depending on the needs of the system) in the event of loss of power to the actuator. either cellulosic or hydrocarbon. A process of combustion characterized by heat or smoke or flame or a combination of these. The range of fuel-air ratios which can support non-detonative combustion such as laminar and turbulent flames. is warranted to resist a standard or hydrocarbon fire test for a particular time period. A test designed to quantify the resistance to fire of elements of construction and/or materials applied thereto. which may occur as the result of a deflagration of a vapour cloud that does not result in a blast wave.
Ventilation by mechanical means.g. damage to the environment or some combination of these. The rapid evaporation of a volatile liquid when suddenly released from pressurized storage. plant. A fire in which the rate of fuel consumption is controlled by the rate of supply of fuel to the fire. a fan. The potential to cause harm. products or the environment. October 2003 . rather than the availability of oxygen for combustion. The flash point of a liquid is the temperature at which the vapour and air mixture lying just above its vaporizing surface is capable of just supporting a momentary flashing propagation of a flame prompted by a quick sweep of a small gas pilot flame near the surface. including ill health or injury. production losses or increased liabilities. A physical situation with a potential for human injury. damage to property.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Flash Fire The combustion of a flammable vapour and air mixture in which the flame passes through the mixture at less than sonic velocity. temperature and pressure of the unburnt gas. The burning velocity of a laminar (non-turbulent) flame under stated conditions of composition. The capacity of a system to perform the function required of it during and after a major accidental event. damage to plant/equipment. Flash Fire Flash Point Flashing Forced Ventilation Frequency Fuel Controlled Fire Functionality Fundamental Burning Velocity Gas explosion Gauge points Hazard Hazard Hazard 200 Issue 1. The combustion of a flammable vapour and air mixture in which flame passes through that mixture and negligible damaging overpressure is generated. e. The number of occurrences anticipated during a unit of time. such that negligible damaging overpressure is generated. damage to the environment or some combination of these. damage to property. A physical situation with a potential for human injury. Gas explosions can be defined as the combustion of a premixed gas cloud containing fuel and an oxidiser that can result in a rapid rise in pressure Points defined in a CFD analysis where information on overpressure/dynamic pressure time histories are derived.
safety and environmental objectives. safety and environmental (HSE) strategic objectives Health. A hydrocarbon is a molecule comprised of carbon and hydrocarbon atoms. measured in degrees C per minute. procedures. Hazardous Substance Health. safety and environmental management system and its implementation. October 2003 201 . giving rise to its strategic and detailed objectives. A description of the means of achieving health. processes and management system (HSEMS) resources for implementing health.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Hazard Analysis The identification of undesired events that lead to the realisation of a hazard. conduction and convection. It is a total of heat transmitted by radiation. magnitude and likelihood of any harmful effects. safety and environmental (HSE) management plan Health. safety and environmental effects. and which should be quantified wherever practicable. responsibilities. safety and environmental (HSE) management review The formal review by senior management of the status and adequacy of the health. See element specific performance standard. safety and environmental issues. that a company sets itself to achieve. A convenient unit is kW/ /m2/s (1 kW m-2 /s = 317 Btu ft-2 h-1). in relation to health. that a sample is heated. The rate. regulations and new objectives resulting from changing circumstances. arising from the HSE policy. A substance which by virtue of its chemical properties constitutes a hazard. The rate of heat transfer per unit area normal to the direction of heat flow. The company structure. A public statement of the intentions and principles of action of the company regarding its health. The broad goals. policy. In two-phase flow the vapour and liquid phases are assumed to be travelling at the same speeds and to be mixed uniformly throughout. practices. Hydrocarbon Issue 1. safety and environmental management. safety and environmental (HSE) policy Health. Health. the analysis of the mechanisms by which these undesired events could occur and usually the estimation of the extent. safety and environmental management system (HSMES) Heat Flux (heat density) Heating Rate High level performance standard High (Higher) risk methodology Homogeneous Methodology of assessment appropriate for High risk installations or compartments as defined in this Guidance.
but the term also has connotations of 'bad luck' in common speech. illness and/or damage. The quantity and type of fuel stored. only the term 'incident' has been used-in the above sense which embraces the concept of 'accident'. The length of a jet flow over which the effects of its initial momentum are dominant. Hydrocarbon Fire Test Impulse Incident Individual Risk Individual Risk (IR) Inhibitor Initiating events Intumescent Inventory Jet Fire or Jet Flame Jetting Length 202 Issue 1. The frequency at which an individual may be expected to sustain a given level of harm from the realization of specified hazards. and is therefore avoided by others. burn fiercely and produce a high heat flux. having a high flame temperature achieved almost instantaneously after ignition. Substance which reduces explosion overpressures by chemically interfering with combustion. In this guidance. the environment or third parties. illness and/or damage (loss) to assets. Note: The word 'accident' is used by some writers and organisations to denote an incident which has caused injury. An epoxy coating.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Hydrocarbon Fire A fire fuelled by hydrocarbon compounds. The integral of a force or load over an interval of time. The events which initiates an explosion event. A platform inventory lists the fuel types and total volumes stored on board an installation. The term inventory is also used to describe the quantity and type of fuel stored in vessels and pipe assemblies. The frequency at which an individual may be expected to sustain a given level of harm from the realization of specified hazards. An event or chain of events which has caused or could have caused injury. The combustion of material emerging with significant momentum from an orifice. A hydrocarbon fire will spread rapidly. Incident Wave The blast wave as it approaches a structure just before it impacts on its surface. October 2003 . sealing compound or paint that swells up under thermal loading to produce a fire resisting covering. A furnace fire test using a time/ temperature curve which supposedly represents a hydrocarbon fire and which results in an ‘H’ rating for successful specimens.
per unit area of the pool. property or the environment or purely financial loss due to plant outage. such as damage to people. undisplaced shape. Mass Burning Rate may Means of Calibration Issue 1. The activities may be associated with financial loss or safety/environmental issues. especially one that may be released through an acute event. Units are typically kg/m2/sec. stakeholder consultation can be used to assess the values of the company or society at large. See element specific performance standard. For example what is thought to be 'good practice' or sound ‘engineering judgment’ can be checked by peer review and benchmarking of practices in other parts of the industry. Similarly. An imprecise term for a large-scale chemical hazard. Life Safety Risk Liquid spray release Loss Prevention Loss Prevention Lower flammability limit (LFL) Low level performance standard Low risk installation/compartment Major Hazard An installation or compartment identified as being low risk by the risk matrix and screening method described in this Guidance. The mass burning rate of a pool fire is the mass of fuel supplied to the flame per unit time. October 2003 203 . A systematic approach to preventing accidents or minimizing their effects. High pressure release of a liquid hydrocarbon which gives rise to a suspension of the fuel in the air similar to a gas cloud which is capable of deflagration on ignition. The lower level of gas concentration which will result in combustion of the gas. Verbal form used to indicate a course of action permissible within the limits of a standard Means of checking or benchmarking the performance of a decision option with respect to a given decision basis.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Laminar Flame Large Deflection Analysis A smooth surfaced flame with low burning velocity. Risk to life. A type of non-linear structural analysis based on the final deflected shape of the structure rather than the initial. Or. This is the same as Lower explosive limit (LEL). a popular term for an installation that has on its premises a more than a prescribed quantity of a dangerous substance. The general term used to describe a range of activities carried out to minimize any form of accidental loss.
Nominal overpressures are defined as peak representative overpressures by installation/module type determined on a nonstatistical basis from acquired experience or simulation for a demonstrably similar situation. ranges. Loading imparted to a body as a result of overpressure acting normal to the surface of the structure. This is used as the top point in the simplified generation of an exceedance curve Prexp Representative nominal overpressures and nominal dynamic pressures on equipment. The frequency of occurrence of a release which is ignited and results in detectable overpressure (>50 mb). All inspection. atmospheric pressure at any stage or location is called the overpressure. October 2003 . The time required for a freely vibrating structure to complete one cycle of motion. but the term overpressurisation is preferred. Value assigned to a basic variable determined on a nonstatistical basis. either through the action of an external wind or gravitational forces. test and monitoring work related to health.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Medium risk installation/compartment Mitigation An installation or compartment identified as being low risk by the risk matrix and screening method described in this Guidance. safety and environmental management. Monitoring activities Natural Period Natural Ventilation Nominal explosion frequency Nominal explosion loads Nominal overpressures Nominal Value Non-Linear Analysis Overpressure Overpressure Overpressure Loading 204 Issue 1. The ventilation of an enclosure by natural means. the pressure developed above . Overpressure is also sometimes used to describe exposure of equipment to internal pressure in excess of its design pressure. In a pressure pulse (blast wave). The excess pressure above ambient conditions. Means taken to minimise the consequences of a major accident to personnel and the installation after the accident has occurred. A type of structural analysis that allows the geometry and/or the material properties to be non-linear. outliers and sensitivity measures. typically from acquired experience or physical conditions.
Requirements to the performance of safety and emergency preparedness measures which ensure that safety . personnel competence. That deformation which occurs following yield. RAG. or seals formed in situ to ensure that penetrations to firewalls do not impair fire resistance. Strain beyond the elastic limit. Passive Mitigation Peak Overpressure Penetration Seal Performance criteria Performance standards for safety and emergency preparedness Plastic Deformation Plastic Hinge Plastic Regime Plastic Strain Issue 1.objectives. October 2003 205 . Phenomenological models Simplified physical models. A plastic hinge is a zone of yielding due to flexure in a structural member. replenishment or sustainment.) Performance standard A statement which can be expressed in qualitative or quantitative terms of the performance required of a system. Purpose-made seals. person or procedure and which is used as the basis for managing the hazard. Mitigation technique that is ‘non-triggered’ and generally uses the kinetic energy of an explosion to disperse an extinguishing agent or suppressant. reliability. mobilisation time. vulnerability. expressed as far as possible in a verifiable manner. capacity. mechanical or other means of initiation. It is associated with plastic strain and large plastic deformations.The term perfomance' is to be interpreted in a wide sense and Include availability. which seek to represent only the essential physics of explosions. The maximum overpressure generated at a given location. That region of structural behaviour dominated by plastic response rather than elastic response. authority minimum requirements and established norms are satisfied during design and operation NOTE . item of equipment. cladding or free-standing system that provides thermal protection in the event of fire and which requires no manual.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Passive Fire Protection (PFP) A coating. Performance criteria describe the measurable standards set by company management to which an activity or system element is to perform. functionality. (Some companies may refer to performance criteria as 'goals' or 'targets'.
Established and documented step by step activity which allows a specific task to be completed. Such means include management systems applied to the design. A number in a scale from 0 to 1 which expresses the likelihood that one event will succeed another. engineering and construction standards. resulting in the formation of missiles which may have the potential to cause damage. . major accident). October 2003 . All systematic actions that are necessary to ensure that quality is planned. The means of escape which was assumed during design. A documented series of steps to be carried out in a logical order for a defined operation or in a given situation. its inspection and maintenance. The (fractional) no of individuals predicted to become fatalities over a specified period of time – usually one year. Accepted methods or means of accomplishing stated tasks. Quantitative Risk Assessment: the quantitative evaluation of the likelihood of undesired events and the likelihood of harm or damage being caused. The structural components. together with the value judgements made concerning the significance of the results.3m in the direction of flow. obtained and maintained. The rupture of a system under internal pressure. and perhaps a blast wave. Proof stress. is a valuable concept in materials which do not exhibit a sharp yield point.be quantitative. calculated from the pressure time histories at the front and back of the obstacle.though not necessarily . Potential loss of life (PLL) Practice Preparedness analyses Pressure Burst Pressure difference Prevention Primary means of escape Primary Structure Probability Procedure Procedure Proof Stress QRA Quality Assurance 206 Issue 1. The pressure difference across an obstacle greater than 0.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Pool Fire The combustion of material evaporating from a layer of liquid at the base of the fire.2 %. the operation of the installation. whose failure would seriously endanger the safety of a significant part of the installation.e.Pdiff Means intended to prevent the initiation of a sequence of events which could lead to a hazardous outcome of significance (i. The analysis has to be traceable and will normally . for example 0. The stress at which the plastic strain of a material reaches a particular value.
A product. Time averaged over 1.5ms and space averaged over explosion affected area. then to the primary or secondary means of escape. Thermal radiation involves the transfer of heat by electromagnetic waves confined to a relatively narrow region of the electromagnetic spectrum. – Prep The estimated time for people to travel from their work stations to the TR. which are usually independent and may be identical. The time taken for the overpressure to increase from zero to the peak overpressure. through measurements. The probability that an item is able to perform a required function under stated conditions for a stated period of time or for a stated demand. allowing for the possibility of helping injured colleagues. October 2003 207 . The likelihood of a specified undesired event occurring within a specified period or in specified circumstances. Initiating events which cannot be eliminated by design or operating procedures. The product of the chance that a specified undesired event will occur and the severity of the consequences of the event. The performance of the same function by a number of equivalent means. tests or inspections.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Quality Control That part of the Quality Assurance which. Combination of the probability of occurrence of harm and the severity of that harm Quality Radiation Redundancy Release Reliability Reliability Representative peak overpressure Required endurance time Residual events Rise Time Risk Risk Risk Issue 1. service or activity is in accordance with specified requirements. ascertain whether a product. depending on the circumstances. It may be either a frequency (the number of specified events occurring in unit time) or a probability (the probability of a specified event following a prior event). The probability that an item is able to perform a required function under stated conditions for a stated period of time. service or activity's ability to fulfill specified requirements. The type and quantity of fuel which can possibly be ignited.
where risk acceptance is concerned) should not be included in the quantitative expression of risk. Risk acceptance criteria (RAC) Criteria that are used to express a risk level that is considered tolerable for the activity in question NOTE . Examples of the simpler analyses are SJA.e. NOTE 2 .The risk analysis term covers several types of analyses that will all assess causes for and consequences of AEs.Risk may be expressed qualitatively as well as quantitatively. It may be relevant to consider on a qualitative basis certain aspects of risk aversion in relation to assessment of risk and its tolerability. Use of available information to identify hazards and to estimate the risk NOTE 1 . together with the value judgements made concerning the significance of the results and their tolerability. Risk Analysis The quantified calculation of probabilities and risks without making any judgements about their relevance. subjectively evaluated risk performed by individuals) should not be included in the expression of risk. preliminary hazard analysis. Risk analysis Risk assessment Risk Evaluation 208 Issue 1. including issues of risk perception where appropriate.e. Overall process of risk analysis and risk evaluation The evaluation of the likelihood of undesired events and the likelihood of harm or damage being caused. involving a quantification of the probability and the consequences of AEs. The implication of the definition is further that perceived risk (i. October 2003 . The definition implies that risk aversion (i. etc. Risk acceptance criteria may be qualitative or quantitative.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions NOTE . and is the starting point for further risk reduction according to the ALARP-principle. with respect to risk to personnel. dimensionless) or as a frequency.Quantitative analysis may be the most relevant in many cases.RAC are used in relation to risk analysis and express the level of risk which will be tolerable for the 9 activity. in a manner which allows comparison with quantitative RAC. with the inverse of time as dimension. Probability may be expressed as a probability value (0-1. FMEA. environment and assets. Risk Assessment The quantitative evaluation of the likelihood of undesired events and the likelihood of harm or damage being caused together with the value judgments made concerning the significance of the results. HAZOP. an evaluation of risk which places more importance on certain accidental consequences than on others.
executed and maintained to achieve safety and protect the environment in accordance with the acts or regulations As for SCE. environment and assets towards which the management of the activity will be aimed All documented policies. for example. equipment. A pressure pulse formed by an explosion in which a sharp discontinuity in pressure travels as a wave through a gas at a supersonic velocity. Verbal form used to indicate requirements strictly to be followed in order to conform to the standard and from which no deviation is permitted. Means of escape which should be available if the primary means of escape is not available. or provided by the regulators. unless accepted by all involved parties The load imparted to a structure by a passing shock wave. A design limit beyond which the structure may become unserviceable. The ratio of heated perimeter (Hp) to cross-sectional area (A). those elements of the installation which are critical to safety. plant. standards and practices which the Operator shall initiate to ensure the activity is planned. a specified maximum deflection.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Risk Issues Issues that impact on the nature and perception of safety related risks and the role of risk based analysis techniques Insensitivity of response to variations of load. Any structure. as is any which is intended to prevent or limit the effect of a major accident. They should be based on sound scientific and technical information and may be developed by the company and industry bodies. Objective for the safety of personnel. The values or standards against which the significance of the identified hazard or effect can be judged. October 2003 209 . Robustness Safety critical element (SCE) Safety objective Safety Plan Safety-critical elements (SCEs) Screening criteria Secondary means of escape Section Factor Serviceability Limit Shall Shock Load Shock Wave Issue 1. system (including computer software) or component part whose failure could cause or contribute substantially to a major accident is safety-critical. organized.
measured in Joules. Units are Joules/kg/ºC. The amount of heat. J without mentioning or excluding others.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Should Verbal form used to indicate that among several possibilities one is recommended as particularly suitable. required to raise one kilogram of a substance by one degree C. or that a certain course of action is preferred but not necessarily required An overpressure > 50m bar. Overpressure trace which has been smoothed by using a moving average over a period of 1. A flow consisting of either a liquid or gas. An explosion response assessment where the SCEs including structure and supports are required to remain elastic. It is assumed in the case for measured overpressures that they have been smoothed.5 milliseconds (1 millisecond or ms equals one thousandth of a second) A peak overpressure for a compartment/area or installation which is the average of all measured or simulated peak overpressures within a compartment. Air/fuel mixture is such that it contains exactly the required amount of oxygen to completely consume the fuel. Significant Overpressure Single-Phase Flow Skin Friction Societal Risk Smoothed overpressure Space averaged overpressure Specific Heat Stagnation Overpressure Standard Fire Test Steady State Stoichiometric mix Strain Hardening (work hardening) Strength level analysis 210 Issue 1. The relationship between frequency and the number of people suffering from a specified level of ham1 in a given population from the realization of specified hazards. This is the pressure at which pressure relief panels should operate. A furnace fire test using a time temperature curve which simulates a standard fire and which results in an ‘A’ or ‘H’ rating for successful specimens. The excess pressure above that in the approach flow which occurs on the front face of a surface where the gas velocity is brought to rest. The frictional force caused by the passing fluid flow which acts tangentially to the surface of the body. October 2003 . Conditions which do not change with time. The tendency of an elastic-plastic material to exhibit increased resistance to high strains. An elastic method is used for structural response with code/utilization checks as the performance standard.
A calculation in which the results are temperature distributions for a given heat input. such as velocity or concentration. A flame burning in a turbulent flammable mixture. Defined as for VCE and is an imprecise term (see below). that arise due to the presence of eddies within the flow. – Pult. Elastic structural analysis is appropriate. Suppressant Survivability System audit Thermal Analysis Thermal Capacity Thermal Conduction Threshold overpressure Time Domain Turbulence Turbulent Burning Velocity Turbulent Flame Two-Phase Flow Ultimate overpressure Unconfined Vapour Cloud Explosion Issue 1. The burning velocity of the flame when turbulence is present in the flammable mixture.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Strength level blast (SLB) Representative peak overpressure used to test robustness of equipment and structure. Planned and systematic examination of systems to ensure that these have been established. Peak representative overpressure resulting from a CFD simulation where it is assumed that the area is engulfed in a stoichiometric gas cloud and ignited at the worst position and time. The heating may be described in terms of either temperatures or radiation levels. A flow in which both the liquid and vapour phases co-exist within close proximity. chemically interfering with combustion or diluting the flammable mixture. followed and maintained as specified. Rapid irregular local fluctuations in physical variables. Substance which reduces explosion overpressures by either removing heat from the flame. The ability of a material to store heat. October 2003 211 . A solution technique for structural dynamics in which results are obtained at regular time intervals when a time variable force or other load is applied. The ability of a system to function in the conditions of an accidental event for the time required. Peak overpressure below which explosion assessment need not be performed. Heat transfer through a medium via random molecular motion.
product or service is in accordance with specified requirements. Ventilation rates are often given for an enclosure in units of air volume changes per hour. The region on the downstream side of the structure. The overall rate of increase in volume caused by the combustion process. October 2003 . A fire in which the combustion rate is controlled by the availability of oxygen rather than the supply of fuel. The area of the topsides structure which accommodates the wellheads. The process by which an enclosure is supplied with fresh air. Examination to confirm that an activity. The fuel concentration above which combustion will not occur. A network of small bore pipe work nozzles connected to the firewater main which is capable of delivering the design water spray to the protected area. An opening through which gas escapes from a confining enclosure as a result of the expansion caused by combustion. The escape of gas through openings (vents) in the confining enclosure. The proportion of the field of view of a receiving surface that is filled by a flame. Upper flammability limit (UFL) Vapour Cloud Explosion (VCE) The preferred term for an explosion of a cloud made up of a mixture of a flammable vapour or gas with air in open or semiconfined conditions. smoke and billowing flames.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Unity Checks A summation of non-dimensional terms in an equation in structural design. A summation greater than unity indicates failure of the member. Vent Ventilation Ventilation Controlled Fire Venting Verification: View Factor Volume blockage ratio Volume Production Wake Region Water Deluge System Water Screen Wellbay 212 Issue 1. The ratio of the volume occupied by the obstacles to the total volume. each of which relate to a component of loading or structural response. A series of wide angle and/or mist spray nozzles connected to the firewater main that protect a specific area in the module from the passage of heat.
October 2003 213 .UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Worst case (explosion) scenario CFD or phenomenological simulation where it is assumed that the explosion area or compartment is filled with a stoichiometric gas cloud ignited at the worst position and time.2 % (see also Proof Stress). often 0. the stress corresponding to a particular plastic strain. The stress at which a steel sample departs from linear elastic behaviour to plastic deformation in a standard tensile test. Yield Point Yield. Effective Issue 1. The apparent yield stress exhibited by metals such as steel when they are strained at a rate which is significantly faster than the normal testing rate. Dynamic Yield. For materials which exhibit no clear yield point – such as mild steels at elevated temperatures.
October 2003 .UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 214 Issue 1.
Scenario identification • Simple Identification appropriate for installation • Use CFD to analyse emission dispersion. D. Assessment of explosion consequences • Direct impact of Personnel • Direct impact on SCEs • Has assessment of escalation potential been made? • Is fatality estimation method realistic for the module type? E. Structural and Equipment Optimization • Use dynamic analysis to assess • Use simple quasi-static load vibration magnitudes. • Is the structural response coupled to the blast analysis. G.. assess against redundancy. • Define explosion exceedence using ignition point array or other empirical method e. historical data. wind directions to generate an overpressure exceedence curve. • Use robust QRA to fully justify that no further measures can be taken to reduce explosion risks further. equipment selection and so on. • Use semi-quantitative arguments to demonstrate ALARP for structural & equipment protection. A summary of the assessment checklist contents is given in the Figure below. • Model all local turbulence and flame propagation effects using mesh model • Define a distribution of leak sources. Blast analysis • Simple estimate of venting and congestion is required to judge whether explosion can occur. ALARP assessment • Demonstrate best practice in layout. based on an analysis and review of practicable design limits of structure • Iterate by assessing all local and global explosion effects and potential risk reduction strategies • Has a detailed escalation analysis been carried out that considers the effect of blastinduced SCE failure (structure + designated systems) on the emergency response functions • Consider external explosions. influence of wind. • Consider external explosions and bang-box ignitions. B. Issue 1. F. vessels and pipework down to a level appropriate for phenomenological modelling tool. blastfunnelling. • Piping stress analysis using selected BOP pulse to demonstrate process integrity. ESD / Blowdown • Effect of wind on mixture build up • Special situations – external explosions etc. Layout optimization • Best practice e. Key techniques Low complexity method Medium complexity method • Emission modelling to estimate likelihood of mixture build-up within enclosure using CSTR-style calculations. October 2003 215 . • Use of case-by-case CBA to assess the worth of upgrading blast protection. EN ISO13702 etc. • Assess the effect of changing wall location. • Show that local BOPs and drag forces are not excessive. • Use model to assess the effect of equipment location changes. based on estimate of practicable limit of protection. for the HSE June 2001. Blast impulse selection • Assess / estimate worst case blast impulse • Assess “basecase” overpressure for iteration.g. • Discourage use of venting model • Use of historical comparison • Use CAD to represent module internals. local concentrations and ingress from neighbouring volumes to arrive at scenario definition. • Model should be able to assess mitigation • Assess “basecase” overpressure for iteration. C. Figure C1 . impulse estimate to demonstrate • Check for local plastic deformation & structural integrity. • Non-linear structural analysis to demonstrate residual strength of structure. analysis with worst-case blast • Assess overall drag forces on SCEs..Summary of assessment checklist contents High sophistication methods may be used where more sophisticated methods of assessment which may result in reductions in conservatism and hence cost are considered more appropriate. supported by CBA.CHECKLISTS This appendix contains the checklists published in Reference  ‘Explosion assessment guidelines’.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions APPENDIX C . release hole sizes. • Consider uncertainty affects • Iterate by assessing effect of layout & structural risk reduction measures High Complexity method A.g. far-field overpressures • Use model to assess the effect of changes in layout and venting areas. removal of blockages and congestion reduction. • Represent equipment.
UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions 216 Issue 1. October 2003 .
there may be no. Typical Application Heading Scenario Identification Limited access wellhead platforms Platforms with few SCEs. pipework. in order to show that an explosion scenario is credible. If the installation closely resembles other designs. Issue 1. for anything other than an empty box). Some sort of comparative assessment is necessary in the Safety Case.g. Continued use of these models for either Design or Operational Safety Cases is not desirable. Venting guidelines have been used in the past for wellhead platforms. it is probably impossible to dismiss an explosion scenario wherever there is the potential for the human factor to result in a release. It may be physically not credible for a flammable mixture to develop in an area. There is insufficient inventory to form a cloud of sufficient size to result in damaging overpressures (it is probably impossible to dismiss an explosion scenario on this basis. because the route taken by the flammable gas is so extraordinary. Blast Analysis Is simple assessment of congestion and confinement sufficient? Is use of simple vent model sufficient? Is use of historical comparison sufficient? There must be evidence of a structured approach to congestion and confinement. There are no ignition sources present (again. sometimes when their applicability has been questionable (in fact. that has considered whether the installation in question is significantly more congested or confined compared to historical precedent. This need not be particularly complex. the Shell CAM method appears to cover these issues well. as the persons present represent an ignition source). if there is no weather deck. so that industry experience is shared. then there seems to be little point in conducting an analysis if one has been done elsewhere for a similar installation.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Low risk methodology The low risk methodology may be applied to low risk installations and medium risk installations where valid nominal overpressures less than 1 Bar are available. Consideration may be given to databasing analysis results for installations. For a simple installation. or where the SCEs vulnerability to blast is low Platforms with little or no congestion or confinement Checklist Guidance Is the scenario The Safety Case should include an account of how scenarios credible? were identified. or no process vessels or other venting obstructions. cable trays and so on. or few credible scenarios for the following reasons: Little or no confinement e. The Safety Case must consider if the possibility of a flammable accumulation exists.g. and that anything up to a full bore pipe / valve / flange rupture is possible). and can be applied to local areas of congestion and confinement. given that leaks are generally at high pressure. Little or no congestion e. October 2003 217 . either by formal HAZID or by judgement.
It is simultaneously necessary to consider of the period of the pulse in comparison with the natural period of the structure and pipework. or drag forces. there are typically few options practice in venting available for layout. Impact on SCEs Unless the SCE is rated for the worst-case overpressure. the Case should seek to show that SCEs have to blast effects sufficient residual strength to withstand worst-case blast effects. indirect vibration. simple QRA methodologies assume that POB in the personnel fire area containing an explosion will perish immediately. Best practice is exhibited by demonstration e. will demonstrate the integrity of the structure. This would point toward the need for a medium sophistication analysis. this approach is more problematic given that it is more difficult to retrofit blast wall detailing and supports to withstand high overpressures. then the analysis should conclude functional impairment results. Where there is explosion analysis potential. It is suggested that this should be used as a basecase for new-build designs where the provision of protection is concerned. or that they are located out of the vent path of the explosion. Adoption of best For a wellhead platform or NUI. and so on. blast analysis these are as follows: Overpressure pulse magnitude Pulse duration Associated drag force Assessment of explosion consequences Layout optimisation Structural and Equipment Optimisation Simple quasistatic load analysis ALARP Assessment Best practice in layout.g. some information is necessary. equipment selection. to show that there is no possibility of damaging resonance effects. in the associated PS it is necessary to identify the principal explosion parameters.g.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Typical Application Heading Blast Impulse Selection Limited access wellhead platforms Platforms with few SCEs. it appears that most open wellhead platforms and NUIs can only experience overpressures up to approximately 1 bar in any case. For existing installations. Sufficient input In order to check whether the structure and pipework is information from vulnerable to blast. that vent paths are clear. but nonetheless there are impulse or lower SCEs protect against. and so on. The aim should be to show that there is value for structure sufficient inherent strength in the structure and equipment to and equipment withstand worst-case conditions. keeping vent paths open. It is envisaged that it would in fact be difficult to justify any risk reducing measures at all on cost-benefit grounds for small installations that are only periodically visited. The Case should show how layout best and congestion practice has been used to reduce overpressures e. Typically. No SCEs exposed As above. by reducing management congestion. or where the SCEs vulnerability to blast is low Platforms with little or no congestion or confinement Checklist Guidance Identification of It is intended that a low sophistication approach should be used Worst-case blast wherever personnel exposure is low. that grating has been used to ease overpressures where possible. Impact on Typically. either as a result of exposure to the pressure impulse. locating SCEs distant from the explosion source. based on the maximum pulse magnitude. A simple quasi-static load analysis. 218 Issue 1. October 2003 . These are discussed below. that pipe routings consider blast impact / drag loadings. venting and congestion management In order to take credit for SCE survival to blast. Note that this could be done on the basis of previous analyses based on similar installations. and so the optimal level of protection could be less.
This scenario is therefore not credible. How are Usually. Typically ignitions are explosions characterised as immediate or delayed. FPSO cargo tanks and so on.g. direction? Are confined Does the HAZID address the possibility of explosions in vents explosion and drains e. Medium risk methods are described in below which substitute for some of the tasks defined in the high risk methodology. compared identified? to single-component explosions. The philosophy recommended in this Guidance is that for medium risk installations the choice of methodology for any particular task must be justified where it deviates from the high risk methodology. Checklist Guidance Have scenarios For the low sophistication case. These are in turn variously described as explosions and flashfires. pipe supports. due to the back flow of air into a vent drum and scenarios the subsequent formation of a flammable mixture.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Medium risk methodology The medium risk methodology may be applied to medium risk installations. gas treatment. The analysis should consider "partial fill" situations. A flashfire or weak explosion is usually not considered in terms of overpressure damage. but with considerable segregation between TR and hydrocarbon hazard e. probably associated with the cause of the initial leak e. This could be the case for successful ESD / inventory? Blowdown. NUIs with relatively high manning e. Criteria for assessing a delayed ignition as either a flashfire or a strong explosion should be identified at the outset. the Fire and Explosion Risk Analysis will consider the unconfined potential for gas build-up and ignition. bridge-linked. The analysis must consider the full range of such as wind wind speeds and directions. However. Does the scenario Possibly there is insufficient inventory in the leaking system to account for build-up an explosive mixture throughout the domain being available considered. due to high equipment density. The inspector must ensure that the scenario identification accounts for this possibility and that this is incorporated into the uncertainty analysis as appropriate. Heading Scenario Identification Issue 1.g. Delayed ignitions are caused by "far-field" ignition sources e. which should consider all the influencing factors.g. if a blast scenario is dismissed as not credible.g. It is likely that some scenarios will be omitted from the analysis. the scenario identification is been considered primarily concerned with the geometry of the area where an and excluded? On explosion could potentially take place.g. Typical Application NUIs with confinement or congestion e. and account for uncertainty in order speed and to omit scenarios on this basis. vent pipes and headers. Are severe It is well known that enriched gases and mists result in changes explosions to known flammability limits and blast overpressures. October 2003 219 . These should identified? be addressed in the design of vent drums. static or intermittent sources such as faulty electrical equipment. Does the scenario Some scenarios may be omitted from the analysis on the basis account for that the ambient wind conditions prevent the formation of an external effects explosive mixture. This must in turn entail a treatment of the likelihood of a flammable mixture being present in an enclosure. only as a cause of fatality for persons directly exposed to the event. a medium what basis? complexity analysis will probably entail a probabilistic analysis of the likelihood of an explosion occurring.g. > four visits per year Manned platforms with separation process. or "strong" and "weak" explosions. Immediate ignitions arise addressed? from "near-field" ignition sources. hot-work. and so the inspector must be satisfied that the grounds for omission are firm.
UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Typical Application NUIs with confinement or congestion e. These . Has the effect of For new-build designs.g. but with considerable segregation between TR and hydrocarbon hazard e. Checklist Guidance General For a small. Small open structures should not experience severe overpressures and so it is simple and practicable to protect against worst-case or near-worst-case scenarios. or by random selection. the worst-case scenario will depend on ignition source strength and location. in order to demonstrate the survivability of vulnerable SCEs such as hydrocarbon pipework. For any mitigating effects? level of sophisticated analysis. Assessment of The analysis should use a variety of ignition point locations to ignition point array show that worst-case conditions have been determined. Heading Blast Analysis 220 Issue 1. simple installation it is probably sufficient to identify worst-case overpressure scenarios and determine if the structure and facilities are able to withstand it. walling Can the analysis Generally. October 2003 . For a given geometry. Larger geometries will inevitably lead to rather more severe explosions given the increase in inventory size or an increase in the amount of equipment representing both congestion and vent blockage.g. It is likely that the use of an ignition point array is used to derive a probabilistic exceedence curve (see below). blast models are most useful in assessing the effect of successfully model layout changes such as equipment location and venting. The analysis should attempt to show the degree of increase in blast overpressure brought about by design growth or alternatively should calculate the overpressure at the end of the design to demonstrate that there has been no more than a small effect. the wellbay. due to high equipment density. cable trays and so on. Other factors The calculation of blast overpressure should consider the effect affecting Blast of temporary blockages such as wind-walling / tarpaulins / overpressure scaffolding or a workover MODU placed against the platform. if MODU or windthis interacts with the module being analysed e. > four visits per year Manned platforms with separation process. vent pipework. Is there any effect The analysis must account for the possibility that a severe of flashing liquid / explosion can result from a mist or concentration of enriched condensate on gas. and also on the local gas concentration. more complex geometries it is likely that worst-case overpressures will be much more severe than the structure and equipment can cope with. cable trays etc. For larger. Calculated overpressures are the analysis been likely to be exacerbated as more detail is introduced to the accounted for? analysis e.g. NUIs with relatively high manning e. the initial analysis may only consider the design "growth" on geometry down to crude limits.. selection of the most confined or congested areas. It is difficult to say how many ignition locations and overpressure calculations have to be carried out to determine an adequate curve. gas treatment. deluge to mitigate explosions. bridge-linked. small diameter pipework.g.g.g.g.the blastwall has already been specified and ordered. blast overpressures? Have drag forces The analysis should be capable of being used to calculate drag been quantified? forces. however once the design overpressure is selected it is difficult to iterate the design . the model should be able to demonstrate the worth of mitigation measures e.random? could either be through expert judgement e. however the case should be convincing that a sufficient number has been carried out.
In fact.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Typical Application NUIs with confinement or congestion e. Input variables. Model limitations. A crude method is to simply adopt a conservative approach and factor up the predictions according to expert judgement. alternatively it may be possible to show that that modelling uncertainties are not significant given the inherent strength of the SCE being considered. and producing an exceedence curve. The literature describing most models generally claims that they are just as likely to underpredict as overpredict the overpressure for a given blast scenario. How has Analytical uncertainty arises as a result of inevitable modelling uncertainty been imperfections compared with actual events.g. riser ESDV. model limitations or uncertainties regarding the value of the input variables. this is combined with the exceedence curve probability to arrive at physical impairment frequencies. Therefore. October 2003 221 .g. due to high equipment density. that shows the probability that an explosion will be greater than a given overpressure. Issue 1. If an exceedence curve technique is adopted. and the concentration at the ignition location. If statistics are used to compile or compare exceedence probabilities. Checklist Guidance What is the As above .calculated worst-case scenarios. This approach readily lends itself to cost-benefit analysis in support of ALARP. and suggests that applying a log-normal distribution to the laminar burning velocity within the model.g. Uncertainty can arise if the model is unable to calculate overpressures at a single location e. it is impossible to know the ignition location for a gas cloud with any degree of certainty. the methodology needs to be reviewed carefully. Work has been done to compile data relating to estimates of blast magnitude and frequency for North Sea installations of varying sizes. If it is constructed from the results of explosion analysis.it is likely that for larger more congested modules it is methodology for not practicable to stiffen structure and supports to withstand selection of lower. > four visits per year Manned platforms with separation process. Heading Blast Impulse Magnitude Selection What are the practicable protection limits? The inspector should establish that the safety case has explored the practicable limits for overpressure protection on the basis of the worst-case overpressure and any probabilistic analysis. bridge-linked. This is quite sophisticated and may not be possible within the confines of "black-box" models bought off-the-shelf. Thus. Is it possible to design for the worst-case. Uncertainty is therefore typically reduced by carrying out many calculations for an array of ignition points and gas concentrations. if within the model domain. producing a "cliff-edge" risk effect on the QRA results. if an estimate of explosion frequency is available. This blast loadings for is likely to be probabilistically based. a methodology is needed than-worst-case to arrive at a design value that is lower than the worst-case. the inspector must ensure that a sufficient number of calculations has been carried out to reduce uncertainty to a tolerable level. Puttock  explains that model uncertainties mainly affect calculated overpressure by their influence on the flammable cloud burning rate.g. even though this is a remote event? This goes a long way to removing uncertainty. The inspector must be satisfied that cliff-edges have been identified. gas treatment. and result in the production design? of an "exceedence curve". Scenario definition. The average overpressure may be close to the SCE survivability. Relatively crude phenomenological models are usually unable to meaningfully calculate anything other than a peak throughout the domain. but with considerable segregation between TR and hydrocarbon hazard e. NUIs with relatively high manning e. Uncertainty in addressed? overpressure estimation could arise due to uncertainties regarding scenario definition. then the inspector must ensure that they are applicable to the installation type under scrutiny.
g. due to escalation effects may be low. high overpressures can shape been result if the available vent path is on the long axis of the module. Has published There is a lot of published guidance regarding the most guidance been advantageous layout for areas where there is an explosion used to lay out the hazard e. due to high equipment density. the QRA rule set may assume universal fatality amongst POB on the other side of the blast wall. panels. Have the effect of Location of equipment can affect the degree of blockage. vent been carried out? header pipework.g. it is generally accepted that a "long & thin" layout should serve to reduce explosion likelihood and have the benefits magnitude. removal changes. It is common to assume universal fatality in the area containing Is fatality estimation realistic the blast. TR functions may be under-reported. gas treatment. Heading Assessment of explosion consequences Layout optimisation Structural and Equipment Optimisation 222 Issue 1. On the other hand. and has included a process for assessing and of blockages and controlling the key factors affecting explosion risk in the design or congestion safety management process. and so the benefit of risk reduction methods intended to mitigate and control escalation e. assessed? The inspector should be satisfied that the advantages of this layout have been addressed within the design of the facility. the axis of vessels is along changes been the vent path. Performance vibration / accelerations been standards should be set for SCEs to withstand blast-induced determined? accelerations where appropriate e. As a result. then the likelihood of immediate fatality may be for the module inappropriately high.g. it is necessary to of structural analyse the deflection of primary structure arising from the deflections using design event.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Typical Application NUIs with confinement or congestion e. given that there could be a significant effect on risk attempted? levels.g. For an open module. UPS. Typically. Has quantification In order to assess the survivability of pipework. reduction been assessed? For new-builds. in the event of failure of a blast wall. in order to be able dynamic analysis to show that deflection does not result in escalation to e. but with considerable segregation between TR and hydrocarbon hazard e. the numbers of delayed type? fatalities e. The equipment location inspector should be satisfied that e. TEMPSC davits. October 2003 . > four visits per year Manned platforms with separation process.g. Have the effects of Leading on from the above. The module inspector should be satisfied that the explosion analysis has equipment? considered the effect of following the guidance.g.g. If it is assumed that POB have no chance to escape before the blast. the inspector should be satisfied that changing wall the dutyholder is aware of the impact of the effect of layout locations. the inspector should look to addressed? see that some sensitivity work has been done. to check that the effect of relocating equipment is not significant. Checklist Guidance Has assessment It is probably valid to consider escalation effects resulting from of escalation been SCE failure. van Bakke. The Safety Case should include an analysis of local structural Have blast accelerations arising from severe explosions. bridge-linked. whilst the vent changing module paths are short. EN ISO13702  and Bjerkervedt. If the layouts are fixed. NUIs with relatively high manning e. This may be inappropriately conservative. given that "failure" of the wall is merely exceeding the buckling failure limits of the wall and supports rather than rapid and complete demolition. The layout encourages dispersion. This must consider the failure modes and the likely escalation that results. This must be realistically done. say.g.g.
Issue 1. As exposure increases. Conservatism in cost-benefit analysis relates to ensuring: Costs associated with implementing remedial measures are not over-stated (some remedial measures can be implemented as part of an ongoing scope of work and so offshore mobilisation costs etc. the dutyholder must show that the preferred way forward follows best practice. and that the criteria used are valid.production and so on are accounted for. other losses . Checklist Guidance Do Survivability Following from the above. more complex installations.e. it is envisaged that these require greater risk exposure on the part of operations POB to inspect and maintain the plant items.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Typical Application NUIs with confinement or congestion e.g. so the use of cost-benefit analysis (CBA) becomes more meaningful.g. Standards Loading arising from flow effects i. gas treatment. POB exposure is not underestimated. drag forces. Where benefits are marginal. NUIs with relatively high manning e. the inspector must ensure that SCE Performance Survivability criteria are based on the ability to withstand: Overpressure effects (direct loading). Check for local Points to consider in the assessment are: Has the structural design accounted for heavy explosion loads giving plastic deformation rise to plastic deformation. but with considerable segregation between TR and hydrocarbon hazard e. due to high equipment density. These could point to a conclusive overall benefit. so that the benefit is accurately measured. bridge-linked. Care is needed to ensure that the analysis is conservative. and redundancy Has sufficient redundancy been build into the structure.g. Has the effect of plastic deformation on SCEs been accounted for in the Safety Case? Heading ALARP Assessment Is cost-benefit analysis used to assess risk reduction measures? Is it conservative? For larger. Gross disproportion is accounted for. October 2003 223 . > four visits per year Manned platforms with separation process. are therefore not applicable). consider all blast Acceleration resulting from blast effect on primary structure effects? Note that SCEs located in non-hazardous or safe areas are vulnerable to blast-induced acceleration. The costs of risk reduction measures require careful "reality" checking. Where the results of CBA are counter to historical experience.
October 2003 . response analysis and so on. The inspector should be satisfied that the Safety Case accounts for the possibility of small concentrations giving rise to a local explosion hazard.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions High risk methodology The High risk methodology may be used for medium and low risk installations where more sophisticated methods of assessment which may result in reductions in conservatism and hence cost. it is expected that they should respond to challenges present at the leading-edge of explosion and structural response research. It is envisaged that the explosion analysis for an clearly stated? integrated facility. and the expectation that blast protection can be optimised. this could include accounted for? gas migration and accumulation over the tank top. Typical Application Heading Scenario Identification PDUQ platforms. FPSOs. it may be emission appropriate to check for local high gas concentrations resulting concentrations from small leaks. The TR and EER facilities may be directly vulnerable to blast effects (see below). Blast Analysis 224 Issue 1. blast analysis. At this high level of sophistication. due to the inability of phenomenological models to tackle non-standard module configurations and the inability to calculate local overpressures with any degree of confidence. Nonetheless it is clear that such expertise is available in the market-place and so at this level there should evidence that the dutyholder has commissioned an appropriate level of consultancy support for scenario identification. Have local high For areas of high congestion around equipment. Have external The hazard from external explosions is most severe for large explosions been integrated facilities. CFD (design) or real-time tests (existing been identified? installations) can be used to see if there are occasions when natural ventilation is insufficient to disperse leaks and to check for mixture build-up. if the process decks are all plated. it is clear that the Dutyholder cannot always be expected to have specialist explosion expertise in-house to facilitate and advise. Use of CFD follows from the high explosion risk potential. are considered more appropriate. increased computer power. in the HAZID.g. For an FPSO. and which aim to deal with the numerous uncertainties associated with detailed calculations. as it is likely that SCEs are located in identified? modules adjacent to the explosion location. multi-module installations. will be carried out using a CFDbased code. platforms where it is clear that the potential for escalation exists Checklist Guidance General All the points from the low and medium sophistication checklists apply here also. Given the resources available to the operators of large integrated facilities. complex layouts. FPSO etc. Is the philosophy All the points from the Medium Sophistication Checklist apply for blast analysis here also. the potential for gas neighbouring migration and gas build-up from remote areas must be modules been considered e. Has ingress from For large. Operators must be prepared to resource detailed studies and tests which seek to demonstrate the clear interaction between the calculation of blast overpressure coupled with structural response.
Overpressures The degree of analysis should be sufficient to indicate areas of high local overpressure to allow suitable mitigating measures to be specified and implemented. This can have a magnifying effect on the overpressure in the partially confined volume containing the initial explosion as it prevents venting. It is clear that effort is being directed to using CFD codes to input to a probabilistic analysis. full-scale experiments have been based on quiescent clouds. CFD codes are of most use in determining local overpressure effects. Elsewhere. (See Scenario Identification / General for the duty holder's requirement to ensure they obtain competent consulting support where necessary). Blast Impulse Magnitude Selection What is the methodology for selection of lowerthan-Worst-Case blast impulse magnitude for design? CFD codes require a lot of computing power to operate and run times are lengthy. will be carried out using a CFD-based code. In the event of a strong explosion in an area of partial confinement. platforms where it is clear that the potential for escalation exists Checklist Guidance Has the effect of Explosion overpressures are exacerbated by high cloud cloud turbulence turbulence as is likely in the event of a large or catastrophic leak. Blast-funnelling. been accounted It is not clear if complex. explosion effects are not limited to domains and there are two important phenomena to account for in explosion analysis: External explosions arise as a result of turbulent unburnt gas being pushed out of the partially confined volume and subsequently ignited. The analysis should consider the effect of explosions on exposed SCEs that located outside of the explosion analysis domain. Puttock  has described a technique for combining CFD results at local locations to exceedence calculation results from a phenomenological model to arrive at a "transfer function" that can be used to provide an exceedence curve for a given location . the inspector must be aware of these and assess whether the installation Safety Case under scrutiny has used such tools in the ALARP demonstration. sophisticated CFD-based explosion for? modelling techniques can tackle this sort of turbulence. This in turn can result in potentially severe overpressure and drag force effects. Assessment of Blast analysis is typically carried out within a set "domain" that other blast reflects the extent of the partial confinement and effectively phenomena marks the limit of the calculation. Issue 1. External explosions also result in areas not exposed to the initial explosions receiving unexpectedly high incident overpressures and drag forces. FPSOs. these can have an effect on items some distance from the limit of confinement as the overpressure pulse decays. Full-scale trials are intended to investigate the effects of turbulence: hitherto. FPSO etc. producing high local gas velocities. Far-field overpressures. Talberg et al seems to describe a similar approach. Vented gases from a module can encounter confinement in neighbouring modules which will "funnel" the gases.or SCE. Identification of It is envisaged that the explosion analysis for an integrated local high Blast facility. The analysis should account for the possibility of external explosions. In fact.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Typical Application Heading PDUQ platforms. October 2003 225 . as the number of runs required to build up a convincing dataset would be prohibitively time-consuming. Huser et al has described a how a database of dispersion and overpressure results can be used as a basis of a probabilistic analysis. At present this probably precludes a full probabilistic analysis arriving at an exceedence curve. The analysis should address this phenomenon and its potential for increasing the likelihood of SCE failure.
The assessment must show that been determined? sufficient structural and piping stress analysis has been carried out to determine ultimate strength for the applied load. It is only by fully considering the effect on SCEs that full cost-benefit analysis of explosion risk reduction measures can be carried out. Evidence should be drag forces are not presented that the layout has been optimised to reduce these excessive potentially damaging local overpressures to a tolerable level in order to demonstrate ALARP. platforms where it is clear that the potential for escalation exists Checklist Guidance How have It is likely that overpressures in complex multi-module practicable installations can rise to high values.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Typical Application Heading PDUQ platforms. Has all vulnerable Following from the above discussion on "far-field" effects. Non-linear analysis can be used to demonstrate the point at which pipework loses containment. This could be conservative. Has non-linear It is possible that the calculated structural deflections have been analysis been used in a relatively simple fashion in assessing the survivability used to check the of process pipework. the blast wall and primary connections. This is additional to the simple calculation of direct and indirect fatality referred to in the Medium Complexity checklist. Uncoupled analysis is the norm. which is assumed unaffected by any resulting coupled to the deformation. Non-linear analysis of the structure the residual can determine the degree of plasticity and whether actual failure strength? of e. FPSOs. Robertson et al describe how "coupling" the blast analysis? structural response model to the blast overpressure model resulted in a more realistic representation of the effect of blast. Has non-linear It is likely that structure and supports have a great deal of analysis been residual strength after undergoing plastic deformation brought used to determine about by blast overpressure.g. accounted for? Assessment of explosion consequences Layout optimisation Structural and Equipment Optimisation 226 Issue 1. if the effect of structural survivability is linked to a code stress level rather than the deflection on ultimate strength of the pipe then effectively the pipe will never pipework? survive any significant deflection. an attempt should be made to properly demonstrate escalation analysis the effect of SCE failure on the capability of the emergency been carried out? response functions of the installation and ultimately the individual and group risk levels. the structural model allows the loading to interact with the structural deformation so that both loading and deformation were considerably reduced. and so it is Robertson suggests that coupling the analyses effectively removes unnecessary conservatism from the response analysis. Is the structural Typically. but perhaps expert judgement can be used to identify particularly vulnerable sections that can be subject to close study. This will naturally lend itself to determining whether additional strength can be built into the structure and supports in line with costbenefit analysis. Effectively. It will probably be prohibitively time-consuming to subject every pipe length to such a degree of analysis. structure is assumed rigid throughout the applied blast response analysis load profile. the structure and analysis must ensure that items outside the blast model domain equipment been are also optimised for blast resistance. Show that local The blast analysis should be of a suitable degree to identify overpressures and areas of locally high overpressures. SCE protection may not be protection limits practicable for the worst-case. October 2003 . Has a detailed At this level.
because sometimes it is difficult to directly relate SCE demonstrate failure to the QRA. a fully integrated QRA is quantitative risk necessary. platforms where it is clear that the potential for escalation exists Checklist Guidance Use of robust Following on from the above. where uncertainty arises in the analysis of causes of management fatality. October 2003 227 . This will reduce uncertainty in the final risk explosion results. FPSOs. reduces risks to ALARP Issue 1.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions Typical Application Heading ALARP Demonstration PDUQ platforms. This should address risk to SCEs as well as directly analysis to to POB.
uk Website: www.UKOOA FIRE AND EXPLOSION GUIDANCE Part 1: Avoidance and mitigation of explosions PUBLISHED BY UK OFFSHORE OPERATORS ASSOCIATION London Office: 2nd Floor. Tel: 020 7802 2400 Fax: 020 7802 2401 Aberdeen Office: 9. SW1V 1AU. AB10 1YP Tel: 01224 626652 Fax: 01224 626503 Email: email@example.com. 232-242 Vauxhall Bridge Road.org. London.uk 228 Issue 1.co. Albyn Terrace. Aberdeen. October 2003 .
This action might not be possible to undo. Are you sure you want to continue?