A Risk Based Approach for the Evaluation of Firefighting Equipment on Liquefied Gas Jetties June 1999 Written by AEA Technology Ltd. on behalf of Society of International Gas Tanker and Terminal Operators Ltd WSIL - 209.indd 12/3/2009 10:26:01 AM v Contents Page No. Executive Summary v 1. Introduction 1 1.1 Objectives 1 1.2 Structure of the Guide 2 2. QRA Approach 3 2.1 Hazards and Risks 3 2.2 Safety Risks 3 2.3 Business Risks 5 2.4 Outline of QRA 6 2.5 Application of QRA 7 3. Case Studies 13 3.1 Description of Town Jetty 13 3.2 Description of Remote Jetty 14 4. Risk Assessment 17 4.1 Key Stages 17 4.2 Step 1 Identify Hazards 17 4.3 Step 2 Assess Likelihood 17 4.4 Step 3 Evaluate Severity 23 4.5 Step 4 Assess Escalation 27 4.6 Step 5 Assess Risks 31 4.7 Step 6 Ranking Risks 37 5. Fire Mitigation 39 5.1 Effect on Escalation 40 5.2 Effect on Safety Risks 41 5.3 Effect on Business Risks 44 6. Applying the Results 47 6.1 Using the Generic Assessment 47 6.2 Interpretation of Generic Results 47 6.3 Comparing the Results to a Specific Site 49 6.4 Summary 52 7. Glossary of Terms 55 8. Useful References 56 Appendix 1 Release Frequency Data 57 Appendix 2 Firefighting Systems 61 Appendix 3 Environmental Risks 63 Appendix 4 A Review of NFAP Guidelines 65 SIGTTO WSIL - 209.indd 52/3/2009 10:26:02 AM 17 SIGTTO Risk Assessment 4. RISK ASSESSMENT 4.1 Key Stages A safety risk assessment was carried out for the two jetties detailed in Section 3. The principles used in the assessment are detailed in Section 2. A summary of the results is given below. This section focuses on the potential for release and ignition of that release creating either fire or explosion. It also considers the escalation of initial fires into much larger and more damaging fires. Step 1Step 2Step 3Step 4Step 5Step 6 Identify hazards Assess likelihood Evalute severity Escalation analysis Assess risks Rank risks 4.2 Step 1 Identity Hazards The purpose of this part of the study is to identify all significant hazards associated with the operations. In this assessment, we are considering hazards which could cause either loss of life or business costs arising from the jetty operations; ie events with the potential to cause loss of product inventory and lead to a fire or explosion if ignited. There are generally two sources for external leakage of product. generic failure hazards • - this may include corrosion, erosion, maintenance error etc. leading to a leak external hazards • - collision of vessel with product line, human error during loading, structural failure of jetty due to collision, damage etc. extreme weather conditions. External hazards (such as ship collision) have not been considered here. These aspects vary considerably between sites and consideration of these here is not consistent with the generic nature of this study. However, it is important to know that these aspects may be of importance in certain situations and should be addressed in specific site studies. In a risk assessment the site is considered, whether this be a jetty or a complete refinery it is divided into a number of hazard areas. For a jetty assessment, two hazard areas are appropriate: area 1 - Jetty head• area 2 - Jetty pipeline• 4.3 Step 2 Assess Likelihood The likelihood of releases occurring is estimated from: Likelihood = ( ?Number of components × component failure rate) WSIL - 209.indd 172/3/2009 10:26:03 AM 18 Risk Assessment SIGTTO 4.3.1 Number of Components The number of components is estimated from a “parts count” which uses Process and Instrumentation, and General Arrangement drawings. A summary of the parts count for the jetties is given in Table 4.1. This gives the total number of parts in each jetty (i.e. all vapour and liquid lines are combined). Details (split by propane/butane and liquid/vapour) are given in Appendix 1. ComponentNumber/Quantity Town JettyRemote Jetty ItemUnitHeadPipelineHeadPipeline Piping (1?-3?) Piping (4?-8?) Piping (8? +) metre1 - 5 - - 50 50 50 50 - 400 2000 Valves (<2?) Valves (2?-4?) Valves (4? +) item- - 4 - - 1 6 - 27 - - 4 Flanges (<2?) Flanges (2?-4?) Flanges (4? +) item- - 10 - - 2 - 17 85 - - 4 LPG Pumpsitem--1- LPG Compressors item--1- Hard armsitem1-4- Cargo Transfer Operations visits/year10-160- Instrument Connections item5-25- Table 4.1 Summary of parts count (total parts) “Cargo transfer operations” in Table 4.1 refer to failures arising from the transfers themselves, including mechanical failure of arms/hoses, ship overfilling, mal-coupling/decoupling, human error and breakaway. The failure rates used here are per visit and the values recorded in the parts count are transfers per year. The duty of the systems is an important factor in determining the risks involved. It is important to understand what fraction of the time the jetty pipework will be “live”. In the case of the Remote Jetty, due to the frequency of operations, only the arms themselves are liquid free when not unloading LPG from the ship (all other product pipework is liquid full). For the Town Jetty, the entire jetty system is liquid free following unloading and only vapour remains in the lines (at a SVP of 6 bar). Prior to transfer, the system is slowly primed with the LPG from the site. While no vapour lines specifically exist, the valves in the example represent the vapour duty of the liquid lines. The duty of the Town Jetty is calculated in the worked example below. WSIL - 209.indd 182/3/2009 10:26:03 AM 19 SIGTTO Risk Assessment WORKED EXAMPLE: CALCULATING DUTY IN LINE An example of how liquid duty is calculated: The example given is for the Town Jetty, as used in the study. Number of transfer per year = 10 Average duration/transfer = 17 hours Priming duration/transfer = 24 hours Hours in a year = 8760 hours Lineduty=´+=(())%101724 87605 The line is therefore, liquid full for approximately 5% of the time. The duty is included in the parts count as follows. Appendix 1 shows how duty has been included for the Town Jetty. The worked example below indicates the calculation performed for valves. WORKED EXAMPLE: INCLUDING LINE DUTY IN PARTS COUNT An example of how duty is included in a parts count assessment is given. The example used is for valves (greater than 4? diameter) on the Town Jetty head. Number of valves (diameter 4? and greater) = 4 Liquid duty = 5% Therefore, “number” of valves in liquid duty = 0.05 x 4 = 0.2 (“average” number per year) “Number” of valves containing vapour = 4 - 0.2 = 3.8 4.3.2 Failure Rate Data The likelihood of a component failing varies considerably and some parts (eg. Pumps) are much more likely to leak than others (e.g. Welded pipework). A summary of the failure rate data used in the assessment is given in Appendix 1. There are many sources of failure data that could be used. Depending on the type of facility that is being considered and the level of detail which is required. The data used in this study is based mainly on E and P Forum [6] and Cox, Lees and Ang [7] data. 4.3.3 Release Frequencies The release frequency is the likelihood of an LPG release whether it is ignited or not. The results show that: the Remote Jetty is relatively more likely to experience a leak of LPG. An estimate of once every • 8 years (0.125/yr) is given for significant release of LPG from the Remote Jetty. An estimate of once every 100 years (0.01/yr) is given for the Town Jetty. liquid releases dominate for the Remote Jetty (99% of all releases). This reflects the much higher • proportion of liquid rather than vapour equipment present at this jetty (see the parts count). vapour releases account for 50% of all releases from the Town Jetty. This is due to the duty of the • system which means that the jetty is liquid full only 5% of the time. In the Remote Jetty, the liquid lines are always full of LPG (apart from the arms themselves). [Note: Routine liquid freeing of the Remote Jetty line is not performed due to the high number of unloading operations]. WSIL - 209.indd 192/3/2009 10:26:03 AM 20 Risk Assessment SIGTTO most of the releases are small, only a very small proportion is expected to be large. This distribution is • because releases are generally associated with deterioration of the equipment and pipework resulting from corrosion in obscured or poorly maintained areas. These initially tend to occur as smaller leaks that are either ignited or become repaired at that time. WORKED EXAMPLE: EXPRESSING FREQUENCY Example of how frequency can be converted from releases per year to “one release every ( years” format. The example used is for total release frequency from the Town Jetty. Total release frequency = 0.01 per year Converting = 1 001. = One release every 100 years Figures 4.1 and 4.2 show where most of the releases occur. The “pipeline” pipework (jetty approach) is a major contributor to the overall release frequency for both jetties. This is due to the very long lengths of pipework involved. The Remote Jetty has designated LPG compressors and pumps. These items are prone to leakage compared to other process components which may be encountered on jetties. This is mostly due to the complexity of these units and the operations they perform (when compared to a static piece of pipework or an infrequently used valve). Contributions from transfer operations are based on assumed transfer levels (10 per year for the Town Jetty and 160 per year for the Remote Jetty). Naturally, an increase or reduction in these levels would have a direct impact on their contributions. On the Town Jetty, release from instrument connections (gauges, etc.) dominate. These tend to be more easily damaged (e.g. by mechanical damage) causing small external leaks than more robust items of equipment. It is therefore, important to locate these items where possible in areas where damage is least likely and to reduce their numbers to the minimum. The LPG jetties considered assume a relatively simplistic manifold arrangement. A more complex arrangement of valves and associated flanges and fittings at the jetty head is often seen as a route to flexibility for transfers. However, such arrangements would also have a significant influence in increasing the likelihood of release at the jetty head. 4.3.4 Fire Frequency The likelihood of a fire occurring is estimate from: Fire Frequency = Likelihood of release × Probability of ignition Only a small fraction of releases (see Table 4.2) ever ignite to become a fire/explosion. The others disperse and evaporate before they come into contact with an ignition source (a flame, spark etc.) The probability that a release is ignited will vary with: the type of site considered• the number of ignition sources available• the characteristics of the flammable material released• WSIL - 209.indd 202/3/2009 10:26:03 AM