REPORT DNV RESEARCH AND INNOVATION

DNV RESEARCH AND INNOVATION FORMAL SAFETY ASSESSMENT ON THE INSTALLATION OF INERT GAS SYSTEMS ON TANKERS <20,000DWT PROJECT NO. 582624 REVISION NO. 4

EXECUTIVE SUMMARY This document details the work carried out to complete a Formal Safety Assessment (FSA) to assess the effectiveness of inert gas systems (IGS) in reducing the risk associated with cargo tank fire and explosions on chemical and oil tankers <20,000dwt. Two different IGSs are considered for two deadweight categories of 4,000-8,000 and 8,000-20,000; N 2 for chemical tankers and conventional oil burning type for oil tankers. These IGS systems are evaluated with Cost Benefit Assessments (CBA) to ascertain their economic viability with regards to reducing potential loss of life (PLL), potential loss of cargo (PLC) and potential loss of property (PLP). In this respect, the Gross Cost of Averting a Fatality (GCAF) and Net Cost of Averting a Fatality (NCAF) are calculated using the standard FSA method recognized by the International Maritime Organisation (IMO) [2]. Further, the Cost of Averting one Tonne of oil Spilled (CATS) proposed by Skjong et al. [3] is applied to understand the cost effectiveness of IGSs in preventing environmental pollution. From a GCAF point of view the results suggest that IGS would be cost effective only for oil tankers of 8,000-20,000dwt; from an NCAF perspective IGS is considered cost effective for chemical tankers 8,000-20,000dwt, although the NCAF for 4,000-8,000dwt tankers is not considered grossly disproportionate. However, there are concerns with regards to suspected under reporting in the casualty databases used, particularly with regards to tankers in the 4,000-8,000dwt category, which could significantly alter the PLL and individual risk scenarios, hence affecting the overall results. In this respect the final recommendation is for IGS to be implemented on all new build 4,000-20,000dwt chemical and product tankers. Page iv

CONTENTS 1. DEFINITION OF THE PROBLEM...6 2. BACKGROUND INFORMATION...7 3. METHOD OF WORK...8 3.1 STEP 2 RISK ANALYSIS...8 3.2 STEP 4 COST BENEFIT ASSESSMENT...9 3.3 STEP 5 RECOMMENDATIONS FOR DECISION MAKING...10 4. DESCRIPTION OF THE RESULTS ACHIEVED IN EACH STEP...12 4.1 STEP 1 HAZARD IDENTIFICATION...12 4.2 STEP 2 RISK ANALYSIS...12 4.2.1 Shipyear and Accident Data...12 4.2.2 Cargoes with a <60 C Flashpoint...12 4.2.3 Risk Analysis for Chemical, Parcel, Chemical/Oil, Crude and Product Tankers...13 4.3 STEP 3 RISK CONTROL OPTIONS...13 4.4 STEP 4 COST BENEFIT ASSESSMENT CHEMICAL, PARCEL AND CHEMICAL/OIL TANKERS...13 4.4.1 Risk Reduction of N 2 IGS...14 4.4.2 Cost of Implementing N 2 IGS...15 4.4.3 Economic Benefit of Implementing N 2 IGS on Chemical Tankers...15 4.5 STEP 4 COST BENEFIT ASSESSMENT CRUDE AND PRODUCTS SHIPS...19 4.5.1 Risk Reduction of Conventional IGS...19 4.5.2 Cost of Implementing Conventional IGS...19 4.5.3 Economic Benefit of Implementing Conventional IGS on Crude and Products Tankers...20 4.6 INDIVIDUAL RISK...23 4.7 STEP 5 - RECOMMENDATIONS FOR DECISION MAKING...24 4.7.1 Under Reporting...25 4.7.2 Root Cause of Reduction in Fire and Explosion Accidents...28 4.7.3 Fatalities Due to Tank Entry...28 4.7.4 Turnaround Time in Port...29 5. FINAL RECOMMENDATIONS FOR DECISION MAKING...30 6. REFERENCES...31 7. ANNEX...33 7.1 ACRONYMS...33 7.2 EXPERTS CONSULTED...34 7.3 INTERTANKO STATEMENT ON INERTING IN PORT...35 Page v

1. DEFINITION OF THE PROBLEM This Formal Safety Assessment addresses the potential of Inert Gas Systems (IGSs) to reduce the cargo tank fire and explosion risk on chemical, products, chemical/oil, crude and parcel tankers <20,000dwt carrying cargo with a flashpoint of <60 C. Cargoes with a flashpoint >60 C are not considered to benefit from IGSs with regards to reducing incidences of fires and explosions and thus are disregarded in the study. Currently tankers <20,000dwt, including those carrying volatile cargoes, that is with a flashpoint of <60 C (as identified in Chapter 17 of the International Code for the Construction and Equipment of Ships Carrying Dangerous Chemicals in Bulk & Index of Dangerous Chemicals Carried in Bulk (IBC Code)) [1], are exempted from having to operate with an IGS under SOLAS Chapter II-2. Moreover, under the same legislation (Paragraph 5.5.2) chemical tankers >20,000dwt with cargo tanks <3,000cbm are also exempted from having to operate an IGS (except in those cases when they are carrying MARPOL Annex I products). However, numerous fire and explosion events suspected of originating in the cargo hold have been recorded for tankers <20,000dwt over the last three decades, as noted by the Inter-Industry Working Group (IIWG) at MSC81. In this respect this FSA aims to quantify base risk and fatality (Potential Loss of Lives (PLL)) frequencies using the latest casualty data from Lloyd s Register FairPlay, and thereafter perform a cost/benefit analysis to ascertain if installation of IGSs on the above named tankers <20,000dwt may be considered cost effective. Page 6

2. BACKGROUND INFORMATION Following a series of fires and explosions on chemical and product tankers, ICS, IAPH, IACS, CEFIC, OCIMF, INTERTANKO and IPTA (see list of acronyms in 7.1) formed a steering committee in January 2005, which appointed a working group drawn from members of the individual organizations together with the International Group of P&I Clubs. The findings and recommendations of the Group have been submitted to MSC 81 as MSC 81/8/1 [8]. MSC 81 considered the reports on incidents of explosions on chemical and product tankers (MSC 81/8/1 and MSC 81/INF.8 [6]), which recommended that the Committee give consideration to amending the SOLAS Convention to provide for the application of Inert Gas Systems (IGSs) to new chemical tankers and new product tankers of less than 20,000dwt (MSC 81/25, paragraph 8.22) [7]. In this respect, MSC 81 resolved that a Formal Safety Assessment (FSA) should be carried out before decisions are made (MSC 81/25, paragraph 8.23) and agreed to refer the issues related to the proposals on IGSs to FP 51 and DE 50, for consideration under the agenda item dealing with casualty analysis, and reporting to MSC 83 (MSC 81/25, paragraph 8.30). Subsequently, at the fire protection sub-committee FP 51 Japan (FP51/10/1 and FP51/10/2) [16] presented a preliminary FSA on extending the requirement for IGSs to chemical, product, chemical/oil, crude, bitumen and asphalt tankers <20,000dwt. The FSA was 100% based on historical casualty data from Lloyds Register FairPlay (LRFP). For evaluating the effectiveness of installation of IGSs, comparison was made on risks resulting from accidents involving fire and/or explosion in the cargo tank of tankers of 20,000dwt and above for the period of 1978-1983, in which most such tankers were not provided with IGS; and for the period 1990-2005, in which all such tankers were provided with IGSs under the requirement of SOLAS Chapter II-2. The comparison showed that such risk for the latter period was 18.2% of that for the former period, that is, the risk reduction ratio between the periods is 81.8%. The words risk reduction here refer to the reduction of a Potential Loss of Lives (PLL) value. The PLL is defined as the expected loss of life and is usually quantified on an annual basis with unit per ship-year. The risk reduction potential for tankers <20,000dwt was calculated by multiplying the risk of cargo tank fire/explosion by 0.818. Moreover, tankers <20,000dwt were split into two dwt categories, 4,000-8,000dwt and 8,000-20,000dwt, to provide a degree of sensitivity to the results. This FSA aims to undertake a comparable study to that of FP51/10/1 except that only those ships that carry cargoes with a <60 C flashpoint, that is chemical, parcel, products, chemical/oil and crude tankers of <20,000dwt, will be studied; bitumen and asphalt tankers will not be included as the cargoes they carry are not considered to have an autoignition temperature of less than 60 C. Page 7

3. METHOD OF WORK This study follows the standard reporting format for an FSA as detailed in the Guidelines for FSA in the IMO Rule Making Process. Figure 3.1 below shows the five main steps of the Formal Safety Assessment (FSA) approach, detailing what each step is comprised of and how the various steps are interrelated. The total risk, defined as the combination of frequency and severity summed up over all identified accident scenarios may be controlled by a number of well-known or newly identified risk control options. Finally, the objective of the cost benefit assessment step is to identify and rank the risk control options in order to determine the most cost efficient ones, i.e. those that provide most risk reduction in relation to cost. This study is concerned with steps 2, 4 and 5 described in Figure 3.1; a Hazard Identification (Step 1) and Risk Control Options (Step 3) are not required to be undertaken as these have already been identified as cargo hold fire/explosions and inert gas system respectively in MSC 81/INF.8 [6] and MSC 81/25 [7]. Figure 3.1: The five steps of a Formal Safety Assessment The following subsections describe the FSA Steps undertaken by this study in more detail and are based on the IMO FSA Guidelines [2]: 3.1 Step 2 Risk Analysis The purpose of the risk analysis in Step 2 is a detailed investigation of the causes and consequences of the more important scenarios identified in Step 1. This can be achieved by the use of suitable techniques that model the risk. This allows attention to be focused upon high risk areas and to identify and evaluate the factors which influence the level of risk. Page 8

Different types of risk (i.e. risks to people, the environment or property) should be addressed as appropriate to the problem under consideration. The construction and quantification of fault trees and event trees are standard risk assessment techniques that can be used to build a risk model. An example of a conceptual risk model is the Risk Contribution Tree. Whilst the example makes use of fault and event tree techniques, other established methods could be used if appropriate. Quantification makes use of accident and failure data and other sources of information as appropriate to the level of analysis. Where data is unavailable, calculation, simulation or the use of recognized techniques for expert judgement may be used. The output from Step 2 comprises the identification of the high risk areas which need to be addressed. 3.2 Step 4 Cost Benefit Assessment The purpose of Step 4 is to identify and compare benefits and costs associated with the implementation of each RCO identified and defined in Step 3. A cost benefit assessment may consist of the following stages: 1. Consider the risks assessed in Step 2, both in terms of frequency and consequence, in order to define the base case in terms of risk levels of the situation under consideration; 2. Arrange the RCOs, defined in Step 3, in a way to facilitate understanding of the costs and benefits resulting from the adoption of an RCO; 3. Estimate the pertinent costs and benefits for all RCOs; 4. Estimate and compare the cost effectiveness of each option, in terms of the cost per unit risk reduction by dividing the net cost by the risk reduction achieved as a result of implementing the option; and 5. Rank the RCOs from a cost-benefit perspective in order to facilitate the decision making recommendations in Step 5 (e.g. to screen those which are not cost effective or impractical). Costs should be expressed in terms of life cycle costs and may include initial, operating, training, inspection, certification, decommission etc. Benefits may include reductions in fatalities, injuries, casualties, environmental damage and clean-up, indemnity of third party liabilities, etc. and an increase in the average life of ships. There are several indices which express cost effectiveness in relation to safety of life such as Gross Cost of Averting a Fatality (GCAF) and Net Cost of Averting a Fatality (NCAF) as described below. Other indices based on damage to and affect on property and environment may be used for a cost benefit assessment relating to such matters. Comparisons of cost effectiveness for RCOs may be made by calculating such indices. For the purposes of this study the Gross Cost of Averting a Fatality (GCAF) (Equation 1), Gross Cost of Averting one Tonne of Oil Spilled (CATS) [9] (Equation 2) and Net Cost of Averting a Fatality (NCAF) (Equation 3) are used. The definitions of GCAF, CATS and NCAF are: Page 9

Equation 1: Gross Cost of Averting a Fatality ΔC GCAF= Δ R S Equation 2: Cost of Averting a Tonne of oil Spilled ΔC CATS= Δ R E Equation 3: Net Cost of Averting a Fatality NCAF= ΔC - ΔB Δ R S Where: ΔC ΔB ΔR is the cost per ship of the risk control option during the lifetime of the vessel. is the economic benefit per ship resulting from the implementation of the risk control option during the lifetime of the vessel (includes environmental and property benefits). is the risk reduction per ship, either in terms of the number of fatalities averted ( R S ) or tonnes of oil spilled prevented ( R E ), implied by the risk control option during the lifetime of the vessel. The output from step 4 comprises: 1. Costs and benefits for each RCO identified in step 3 from an overview perspective; 2. Costs and benefits for those interested entities which are the most influenced by the problem in question; and 3. Cost effectiveness expressed in terms of suitable indices. For the purposes of this study only points 1 and 3 described above are addressed. 3.3 Step 5 Recommendations for Decision Making The purpose of Step 5 is to develop recommendations that can be presented to the relevant decision makers in an auditable and traceable manner. Those recommendations are based upon the comparison and ranking of all hazards and their underlying causes; the comparison and ranking of risk control options as a function of associated costs and benefits; and the identification of those risk control options which keep risks as low as reasonably practicable. There are several standards for risk acceptance criteria [3][4], however none are presently universally accepted. IMO has published criteria to be used in rule making for GCAF and NCAF [5] and to comply Page 10

with IMO s requirements these values have been used to assist judgements about the effectiveness of RCO s in this work. The output from Step 5 comprises: 1. An objective comparison of alternative options, based on the potential reduction of risks and cost effectiveness, in areas where legislation or rules should be reviewed or developed; and 2. Feedback information to review the results generated in the previous steps. Page 11

4. DESCRIPTION OF THE RESULTS ACHIEVED IN EACH STEP 4.1 Step 1 Hazard Identification As described in MSC 81/INF.8 [6] the majority of cargo tank fires and explosions on chemical and product tankers were caused by tank cleaning, venting and gas freeing, with static electricity and sparks/friction the most common ignition sources. It was found that the most significant contributory factor to the incident causes was a failure to follow or understand cargo operation guidelines and procedures. None of the incidents occurred during the use or operation of inert gas systems. 4.2 Step 2 Risk Analysis With regards to calculating risk of fire and explosion due to a lack of statistical data on specific ship types those vessels included within the study have been grouped together. In terms of calculating GCAF, NCAF and CATS the ships have been split into two categories based on the type of inert gas system required; thus, chemical, parcel and chemical/oil tankers, which require N 2 IGS, are included in a first analysis described in 4.4; while product and crude tankers, which require a conventional oil burning IGS, are included in a second analysis in 4.5. 4.2.1 Shipyear and Accident Data Shipyears are calculated using Lloyd s Register FairPlay database. The total number of ships scrapped per year was subtracted from those delivered per year for all ship types included in the study, giving total number of ships in operation for the periods 1978-1983 and 1990-2007. Ships operating per year were then multiplied by the number of years of operation in each time period, giving total number of shipyears. All accident data is sourced from the latest version of Lloyd s Register Fairplay (LRFP) casualty database (2008). The total number of fire and explosion accidents to be included within the study was selected from LRFP by studying the free text description of the accidents, and checking with other sources whenever possible. The results were than sent to two chemical/product tanker experts for validation 1 ; the rationale for counting accidents was those which could have conceivably been prevented by IGS technology. 4.2.2 Cargoes with a <60 C Flashpoint This study considers that cargoes with a flashpoint <60 C will principally benefit from IGS technology and thus only that portion of tankers that carry volatile products is considered. In this respect the base risk frequency and risk of fatality is multiplied by 0.53 to ensure that cargo tank fire/explosion accidents are only applied to those ships carrying cargoes with <60 C flashpoint; the figure of 0.53 was determined by studying recent annual voyage data, which included cargo type and associated flashpoint, provided by two global chemical carriers, Odfjell [11] and Stolt Nielsen [12], and then taking the mean between the two. This was considered representative of the world fleet. 1 Svend Foyn-Bruun (Odfjell) and Otto Nyquist (DNV) Page 12

4.2.3 Risk Analysis for Chemical, Parcel, Chemical/Oil, Crude and Product Tankers Table 4.1 shows the results of the risk analysis for chemical, parcel, chemical/oil, crude and products tankers in three size categories, 4,000-8,000dwt; 8,000-20,000dwt; and >20,000dwt, for two separate periods, 1978-1983 and 1990-2007. The >20,000dwt category was included to calculate the ratio of fire/explosions that could have been prevented by an IGS; to do this initially the PLL from fire/explosions for 1978-1983, when the majority of tankers did not have IGS technology, and 1990-2007 when all >20,000dwt tankers carrying MARPOL Annex I products had IGS installed as per SOLAS Chapter II-2, was calculated. Thereafter, the risk reduction potential of installing an IGS can be calculated by dividing the PLL for 1990-2007 by the PLL for 1978-1983. The percentage difference between the PLL figures for the two periods is considered the potential risk reduction possible by using IGSs on tankers <20,000dwt. The decision on which accidents were counted in 1978-1983 and 1990-2007 was based on reviewing each accident by experts from Odfjell and DNV. It should be noted that reporting and recording of accidents has improved since the first period of 1978-1983, thus the accident frequencies, and hence PLL outcomes, used in this FSA are considered conservative. It is considered that if all accidents could be included from 1978-1983, the difference between >20,000dwt PLL for 1978-1983 and 1990-2007, used for calculating risk reduction, would be even greater, leading to a higher potential risk reduction. Table 4.1: PLL Calculations for Chemical, Parcel, Chemical/Oil, Crude and Product Tankers % of cargoes Period dwt Shipyears No. of accidents <60 C flashpoint Acc frequency Lives lost PLL 1978-1983 1990-2007 >20,000 12,001 46 0.53 7.2E-03 209 3.3E-02 >8,000 1,569 4 0.53 4.8E-03 3 3.6E-03 >4,000 1,962 4 0.53 3.8E-03 9 8.7E-03 >20,000 53,357 41 0.53 1.4E-03 68 2.4E-03 >8,000 12,799 17 0.53 2.5E-03 34 5.0E-03 >4,000 16,108 10 0.53 1.2E-03 10 1.2E-03 4.3 Step 3 Risk Control Options The installation of N 2 IGSs on chemical, parcel and chemical/oil ships, and conventional oil burning IGSs for crude and products ships, are the only risk control measures to be considered in this study. 4.4 Step 4 Cost Benefit Assessment Chemical, Parcel and Chemical/Oil Tankers The risk control measures of N 2 and conventional oil burning IGSs are analysed in this chapter using the methods and criteria set out by the IMO [2][1] [5]. Page 13

Net Present Value (NPV) The cost and benefit of the Risk Control Options (RCOs) is spread over the lifetime of the vessel, which is considered to be 25 years for the purposes of this study. Some RCOs may involve costs annually while others only involve costs at given intervals. In order to be able to compare the costs and benefits and calculate the NCAF and GCAF, Net Present Value (NPV) calculations have been performed on applicable RCOs using Equation 4 as given below: Equation 4: Net Present Value T Xt t 1 ) NPV= A + t= ( 1 + r Where: X t = Cost or benefit of RCO for any given year A= Amount spent initially for implementation of RCO r = Discount rate T = 25 (years) A uniform discount rate of 5% has been used. (This is assumed to be a long term real (above inflation) risk free rate of return i.e. 3% inflation and 8% depreciation. The figure is used in most FSAs submitted to IMO). 4.4.1 Risk Reduction of N 2 IGS As the quality of chemical cargoes is affected by non-nitrogen atmospheres, an N 2 IGS must be used. In this respect Table 4.2 below illustrates the potential risk reduction achievable (93%) with regards to cargo tank fire/explosion by installing an N 2 IGS on chemical, parcel and chemical/oil tankers <20,000dwt. The risk reduction potential is calculated by totalling all cargo tank fire/explosion events for tankers >20,000dwt in the period 1978-1983 that could have been prevented by having an IGS this period was chosen as most tankers did not have IGS technology installed at the time; and then performing the same calculation for the period 1990-2007 when all >20,000dwt tankers licensed to carry MARPOL Annex I cargoes had IGS installed as per SOLAS Chapter II-2. As explained in 4.2.3 this exercise was undertaken by experts from Odfjell and DNV where every incident was reviewed while considering whether or not the accident could have been prevented by IGS. Table 4.2: Potential Risk Reduction of Cargo Tank Fires/Explosion Incidents for Chemical, Parcel and Chemical/Oil Tankers of <20,000dwt PLL 1978-1983 (A) PLL 1990-2007 (B) PLL reduction ratio (B/A) PLL reduction potential (1-B/A) 3.3E-02 2.4E-03 7% 93% The PLL reduction potential of 93% is translated into PLL/shipyear and PLL/ship lifetime, which is assumed to be 25 years, in Table 4.3 below. Therefore, an N 2 IGS is Page 14

expected to save one life every 920 years for ships >4,000dwt and <8,000dwt and one life every 215 years for ships >8,000dwt and <20,000dwt. Table 4.3: Estimated Reduction in PLL for Chemical, Parcel and Chemical/Oil, Crude Tankers of <20,000dwt Dwt PLL/shipyear PLL/ship lifetime* 8,000-20,000 4.6E-03 1.2E-01 4,000-8,000 1.1E-03 2.7E-02 4.4.2 Cost of Implementing N 2 IGS The cost for installing an N 2 IGS at new build stage differs depending on the cbm/hr capacity required. Advice from Odfjell suggests that 1,000cbm/hr x 95% capacity is sufficient for chemical carriers <20,000dwt [10]; the required N 2 capacity for 4,000-8,000dwt carriers is likely to be less than for 8,000-20,000dwt ships, although no information could not be obtained for the cost of an N 2 IGS for the smaller sized ships. Thus, a single cost of $505,000 is assumed for all <20,000dwt chemical carriers, which represents the mean cost of N 2 IGS prices received from Daewoo Shipbuilding and Marine Engineering (DSME) in Korea [13] and tanker owner Kristen Navigation of Greece [14]. Lifetime maintenance costs, provided by chemical carrier Odfjell and based on operational experience, are estimated to be $102,000 or $4,080 a year [15]. The yearly cost of $4,080 is incurred over the lifetime of the ship, thus to determine the cost of maintenance in Present Value, NPV is calculated, giving a total maintenance cost of $57,503. Thus, the total cost associated with fitting N 2 IGS is $562,503. 4.4.3 Economic Benefit of Implementing N 2 IGS on Chemical Tankers The economic benefit of introducing an N 2 IGS is in the reduction of cargo tank fires and explosions leading to fatal accidents, oil/chemical spills and property damage (i.e. to the ship), expressed as the Gross Cost of Averting a Fatality (GCAF), the Cost of Averting one Tonne of oil Spilled (CATS) and Net Cost of Averting a Fatality (NCAF). The limit cost of averting a fatality is currently agreed at $3 million, which is applicable for both GCAF and NCAF calculations; this is the standard value prescribed in the amendment to the IMO circulars MSC 1023 and MEPC 392, dated 16 October 2006 [5]. CATS was also used to assess the cost effectiveness of installing IGSs; from a CATS perspective a risk control measure is cost-effective if it is less than $60,000. This is the value proposed in SAFEDOR D4.5.2 Risk Evaluation Criteria [3] and is the implied intrinsic value of the environment and the cost that society is willing to pay to prevent an oil spill from occurring. Page 15

Gross Cost of Averting a Fatality (GCAF) The GCAF for 4,000-8,000dwt and 8,000-20,000dwt tankers is detailed in Table 4.4 below. Table 4.4: The Gross Cost of Averting a Fatality (GCAF) on Chemical, Parcel and Chemical/Oil Tankers <20,000dwt Dwt PLL/shipyear PLL/ship lifetime* Cost of N 2 IGS ($) GCAF ($) 8,000-20,000 4.6E-03 1.2E-01 562,503 4,843,000 4,000-8,000 1.1E-03 2.7E-02 562,503 20,725,000 * 25 years In addition to GCAF, the Cost of Averting a Tonne of oil Spilled (CATS) and Net Cost of Averting a Fatality (NCAF) is shown in the following tables. Cost of Averting a Tonne of oil Spilled (CATS) CATS is used to calculate whether a risk control measure is cost effective from an environmental point of view i.e. society puts an intrinsic worth on the environment, which is assumed to be $60,000/tonne of oil not spilled in the CATS formula (Equation 2). Thus, if the costs of implementing the risk control measure are less than $60,000 per tonne of spill averted, then the measure can be recommended from an environmental viewpoint. The potential environmental damage that crude and oil products can cause is well documented. For example, oil spills may seriously affect the visual appeal and use of coastal amenity areas; and may result in the widespread death of marine life through physical contact, ingestion and destruction of food resources, generating long-lasting indirect impacts on the complex inter-relationship of aquatic ecosystems. Although CATS has been developed to measure the cost effectiveness of a risk control measure in terms of avoiding an oil spill, in the absence of any other information it is argued that the formula is also appropriate for chemical pollution. Chemicals, such as organochlorides, are commonly shipped around the world and are widely recognised as highly toxic to aquatic organisms and the marine environment in general (e.g. MARPOL Annex II). Moreover, due to the diverse and complex nature of chemicals, the capacity to cause negative environmental consequences following a spill could be considered as significant as oil. Therefore, CATS is applied to chemical tankers as well as those carrying oil in this FSA. Table 4.5: The Cost of Averting a Tonne of oil Spilled (CATS) for Chemical, Parcel and Chemical/Oil Tankers 4,000-20,000dwt Size (dwt) PLC /lifetime* Value of oil not spilled($)** Cost of IGS ($) CATS ($) 8,000-20,000 9.46E+00 567,868 562,503 59,000 4,000-8,000 5.99E+00 359,255 562,503 94,000 * 25 years ** $60,000/tonne (based on value used in CATS) Page 16

Net Cost of Averting a Fatality (NCAF) The NCAF for 4,000-8,000dwt and 8,000-20,000dwt chemical tankers is calculated below with explanation of the methods used. Table 4.6: The Net Cost of Averting a Fatality (NCAF) on Chemical, Parcel and Chemical/Oil Tankers 4,000-8,000dwt Av dwt Av spill (t) Shipyears No. of acc PLC* (t) Risk Red PLC/lifetime# Value of oil not spilled#* Total benefit (PLC + PLP) 6000 342 5312 4 2.58E-01 93% 5.99E+00 359,255 4.61E+05 NCAF ($) Av repair ($) Shipyears No. of acc PLP** ($) Risk Red PLP/lifetime# 5.83E+06 5312 4 4.39E+03 93% 1.02E+05 * Potential Loss of Cargo ** Potential Loss of Property # 25 years #* $60,000/tonne (based on value used in CATS) 3,725,000 Table 4.7: The Net Cost of Averting a Fatality (NCAF) on Chemical, Parcel and Chemical/Oil Tankers 8,000-20,000dwt Av dwt Av spill (t) Shipyears No. of acc PLC* (t) Risk Red PLC/lifetime# Value of oil not spilled#* Total benefit (PLC + PLP) 14000 798 5881 3 4.07E-01 93% 9.46E+00 567,868 7.34E+05 NCAF ($) Av repair ($) Shipyears No. of acc PLP** ($) Risk Red PLP/lifetime# 8.39+06 5881 5 7.13E+03 93% 1.66E+05 * Potential Loss of Cargo ** Potential Loss of Property # 25 years #* $60,000/tonne (based on value used in CATS) With reference to Table 4.6 and Table 4.7 in the absence of any data the average spill is calculated by taking 5.7% of the average deadweight of (6,000 for 4,000-8,000dwt and 14,000 for 8,000-20,000dwt); this is based on Spouge [17], who based the result on historic data with regards to oil tanker mean spill sizes due to fire/explosion. Likewise, as there is very little information on repair costs for ships the average repair cost is based on using LRFP casualty data in terms of whether an accident is classed as serious or non-serious in conjunction with the two cost of repair formulas devised by Spouge [17] and detailed in Equation 5 and Equation 6 below. -1,473,000 Page 17

Equation 5: Cost of Repair for Serious Accidents GT CRsc= $58,000Tsc 75,000 Equation 6: Cost of Repair for Non-Serious Accidents GT CRnsi = $27,000Tnsi 75,000 Where: CRsc = repair cost in serious casualties ($) Crnsi = repair cost in non-serious incidents ($) GT = ship gross tonnage Tsc = mean time off-hire in serious casualties (days) Tnsi = mean time off-hire in non-serious incidents (days) Where a ship is recorded as a total loss in LRFP the value is estimated using further formulas from Spouge [17], which combined with the repair costs make up the average repair costs detailed in Table 4.6 and Table 4.7. The new building cost formulas are below: Where: N = new-building price ($) D = ship deadweight Equation 7: Oil Tanker New Building Cost N = $20 x 10 6 + $350 D Equation 8: Chemical Tanker New Building Cost N = $11 x 10 6 + $1900 D The factors in Equation 7 and Equation 8 are based on average new building costs from 2004-2007 supplied by Clarkson s, which in the absence of accurate historical data are considered the best way of estimating costs. Once the new build cost of the ship has been calculated, the value is depreciated by 4% of the new build cost annually for every year of operation up until the date of terminal loss to give the approximate value when lost [17]. Using the method described above in combination with the risk reduction afforded by an N 2 IGS the PLC/lifetime and PLP/lifetime can be calculated, giving an average total economic benefit for 4,000-8,000dwt tankers of $461,000/lifetime. The total economic benefit is used to calculate the NCAF (Equation 3), which for 4,000-8,000dwt tankers is $3,725,000, 24% above the $3 million upper limit; this implies that the economic benefit of fitting an N 2 IGS does not outweigh the real cost of implementation due to lack of reduction in risk and associated reduction in PLC and PLP. Page 18

With respect to 8,000-20,000dwt chemical tankers the total economic benefit is $734,000/lifetime, giving an NCAF of -$1,473,000; as the NCAF is negative the economic benefit of installing an N 2 IGS on 8,000-20,000dwt tankers implies that the reduction in risk with regards to PLC and PLP outweighs the cost of the equipment. 4.5 Step 4 Cost Benefit Assessment Crude and Products Ships 4.5.1 Risk Reduction of Conventional IGS Tankers transporting crude and hydrocarbon products do not need a high degree of nitrogen to maintain the quality of their cargoes, and as such can employ conventional oil burning IGSs. As with the N 2 IGS (4.4.1) the potential risk reduction of 93% in terms of cargo tank fire/explosion is used. As in Table 4.2, the risk reduction is calculated by totalling all cargo tank fire/explosion events for >20,000dwt crude and products tankers in the period 1978-1983 that could have been prevented by having an IGS; and comparing the PPL frequency with that of the latter period of 1990-2007. Table 4.8: Potential Risk Reduction of Cargo Tank Fires/Explosion Incidents for Crude and Products Tankers of <20,000dwt PLL 1978-1983 (A) PLL 1990-2007 (B) PLL reduction ratio (B/A) PLL reduction potential (1-B/A) 3.3E-02 2.4E-03 7% 93% As with the chemical carriers, the PLL reduction potential of 93% is translated into PLL/shipyear and PLL/ship lifetime, which is assumed to be 25 years, in Table 4.9 below. This translates into one life saved every 909 years for <4,000dwt tankers and one life very 217 years for <8,000dwt ships. Table 4.9: Estimated Reduction in PLL for Crude and Products Tankers of <20,000dwt Dwt PLL/shipyear PLL/ship lifetime* 8,000-20,000 4.6E-03 1.2E-01 4,000-8,000 1.1E-03 2.7E-02 4.5.2 Cost of Implementing Conventional IGS Guidance from Odfjell suggests that for <20,000dwt tankers a conventional oil burning IGS with a capacity of 4,000cbm/hr will be sufficient to inert the cargo hold [10]. The tankers in size category 4,000-8,000dwt may be able to operate safely with an IGS capacity of >4,000cbm/hr, although no accurate installation prices could be obtained; in this respect a single cost of $215,000 for providing and fitting a conventional IGS is estimated, which represents the mean between prices received from DSME of Korea [13] and Kristen Navigation of Greece [14], which has several new builds on order with various yards. In the absence of accurate information, the maintenance cost is taken from Japanese study FP 52/INF.2 [16] and escalated using OECD CPI yearly percentage point increases, giving a total of $29,800 for the Page 19

lifetime of the ship with the NPV cost being $16,800. Thus, the total cost of fitting and maintaining a conventional IGS over the lifetime of a tanker is $231,800. 4.5.3 Economic Benefit of Implementing Conventional IGS on Crude and Products Tankers The economic benefit of introducing a conventional IGS is in the reduction of cargo tank fires and explosions leading to fatal accidents, oil/chemical spills and property damage (i.e. to the ship), expressed as the Gross Cost of Averting a Fatality (GCAF), the Cost of Averting one Tonne of oil Spilled (CATS) and Net Cost of Averting a Fatality (NCAF). The upper limit cost of averting a fatality is currently agreed at $3 million, which is applicable for both GCAF and NCAF calculations; this is the standard value prescribed in the amendment to the IMO circulars MSC 1023 and MEPC 392, dated 16 October 2006 [5]. CATS was also used to assess the cost effectiveness of installing IGSs; from a CATS perspective an risk control measure is cost-effective if it is less than $60,000. This is the value proposed in SAFEDOR D4.5.2. Risk Evaluation Criteria [3] and is the implied intrinsic value of the environment and the cost that society is willing to pay to prevent an oil spill from occurring. Gross Cost of Averting a Fatality (GCAF) The GCAF for fitting a conventional IGS to 4,000-8,000dwt and 8,000-20,000dwt tankers is detailed in Table 4.10 below. Table 4.10: The Gross Cost of Averting a Fatality (GCAF) on Crude and Product Tankers <20,000dwt Dwt PLL/shipyear PLL/ship lifetime* Cost of IGS ($) GCAF ($) 8,000-20,000 4.6E-03 1.2E-01 231,800 1,996,000 4,000-8,000 1.1E-03 2.7E-02 231,800 8,541,000 * 25 years Further analyses, including the Cost of Averting a Tonne of oil Spilled (CATS) and Net Cost of Averting a Fatality (NCAF), are shown in the following tables. Cost of Averting a Tonne of oil Spilled (CATS) CATS is used to calculate whether a risk control measure is economically beneficial from a environmental point of view i.e. society puts an intrinsic worth on the environment, which is assumed to be $60,000/tonne of oil not spilled in the CATS formula (Equation 2). Thus, if the costs of implementing the risk control measure are less than $60,000, then the measure can be recommended from a CATS viewpoint. Page 20

Table 4.11: The Cost of Averting a Tonne of oil Spilled (CATS) for Crude and Product Tankers 4,000-20,000dwt Size (dwt) PLC /lifetime* Value of oil not spilled($)** Cost of IGS ($) CATS ($) 8,000-20,000 8.05E+00 482,745 231,800 29,000 4,000-8,000 2.21E+00 132,574 231,800 105,000 * 25 years ** $60,000/tonne (based on value used in CATS) Net Cost of Averting a Fatality (NCAF) The NCAF for oil tankers is calculated below; the formulas used to calculate the average repair costs are those described above in Equation 5 Equation 8 above. As with chemical tanker NCAF calculations, in the absence of any data the average spill is calculated by taking 5.7% of the average deadweight of (6,000 for 4,000-8,000dwt and 14,000 for 8,000-20,000dwt); this is based on research of historical data undertaken by Spouge [17] with regards to oil tanker mean spill sizes due to fire/explosion. Av dwt Table 4.12: The Net Cost of Averting a Fatality (NCAF) on Crude and Product Tankers 4,000-8,000dwt Av spill (t) Shipyears No. of acc PLC* (t) Risk Red PLC /lifetime# Value of oil not spilled#* Total benefit (PLC + PLP) 6000 342 10,796 3 9.50E-02 93% 2.21E+00 132,574 1.43E+05 NCAF ($) Av repair ($) Shipyears No. of acc PLP** ($) Risk Red PLP /lifetime# 1.60E+06 10,796 3 4.43E+02 93% 1.03E+04 * Potential Loss of Cargo ** Potential Loss of Property # 25 years #* $60,000/tonne (based on value used in CATS) 3,276,000 Av dwt Table 4.13: The Net Cost of Averting a Fatality (NCAF) on Crude and Product Tankers 8,000-20,000dwt Av spill (t) Shipyears No. of acc PLC* (t) Risk Red PLC /lifetime# Value of oil not spilled#* Total benefit (PLC + PLP) 14000 798 6,918 3 3.46E-01 93% 8.05E+00 482,745 6.15E+05 NCAF ($) Av repair ($) Shipyears No. of acc PLP** ($) Risk Red PLP /lifetime# 7.85E+06 6,918 5 5.68E+03 93% 1.32E+05 * Potential Loss of Cargo ** Potential Loss of Property # 25 years #* $60,000/tonne (based on value used in CATS) Using the method described in 4.4.3 in combination with the risk reduction afforded by a conventional IGS the PLC/lifetime and PLP/lifetime can be calculated, giving an average total economic benefit for 4,000-8,000dwt tankers of $143,000/lifetime -3,297,000 Page 21

(Table 4.12). The total economic benefit is used to calculate the NCAF (Equation 3), which for 4,000-8,000dwt tankers is $3,276,000, nine percent above the $3 million upper limit; this implies that the economic benefit of fitting a conventional IGS does not outweigh the real cost of implementation due to lack of reduction in risk and associated reduction in PLC and PLP. With respect to 8,000-20,000dwt oil tankers the total economic benefit is $615,000/lifetime, giving an NCAF of -$3,297,000, which is considerably below the $3 million upper limit; thus, the NCAF for 8,000-20,000dwt tankers implies that fitting a conventional IGS is beneficial from an economic viewpoint. Page 22

4.6 Individual Risk With regards to ascertaining if risk to personnel is within the ALARP region individual risk is calculated. The tables below are compared in Figure 4.2 below. It should be noted that the table only include the individual risk from ship accidents from LRFP data. To calculate the total individual risk the risk from personal accidents would need to be added. Table 4.14: Individual Risk Resulting from All Fire/Explosions (Excluding Cargo Tank Fire/Explosions) Period dwt Shipyears No. of acc Acc. Frequency Lives lost other fires PLL Individual risk Crew no. 1990- >20,000 53,357 260 4.9E-03 253 4.7E-03 2.2E-04 22 2007 >8,000 12,799 89 7.0E-03 59 4.6E-03 2.6E-04 18 >4,000 16,108 77 4.8E-03 45 2.8E-03 1.6E-04 18 Total 82,264 426 357 4.3E-03 2.1E-04 20.6* * Weighted average Period dwt Shipyears No. of acc <60C cargoes Acc. Frequency Lives lost Lives lost all other acc PLL cargo FX >20,000 53,357 41 0.53 1.4E-03 68 186 2.4E-03 1990- >8,000 12,799 17 0.53 2.5E-03 34 58 5.0E-03 2007 >4,000 16,108 10 0.53 1.2E-03 10 27 1.2E-03 Total 82,264 96 0.53 112 270 2.6E-03 Table 4.15: Individual Risk Resulting from All Accidents (Excluding Fire/Explosions) Period dwt Shipyears No. of acc Acc. Frequency Lives lost not fire PLL Individual risk Crew no. 1990- >20,000 53,357 5050 9.5E-02 97 1.8E-03 8.3E-05 22 2007 >8,000 12,799 1157 9.0E-02 51 4.0E-03 2.2E-04 18 >4,000 16,108 1043 6.5E-02 5 3.1E-04 1.7E-05 18 Total 82,264 7,250 153 1.9E-03 9.0E-05 20.6* * Weighted average Table 4.16: Individual Risk Resulting from All Accidents inc. Fire/Explosions on Ship Carrying <60 C Cargoes PLL total Individual risk Cargo FX Individual risk total Crew no. 9.0E-03 1.1E-04 4.1E-04 22 1.4E-02 2.8E-04 7.6E-04 18 4.3E-03 6.5E-05 2.4E-04 18 8.8E-03 1.2E-04 4.3E-04 20.6 * Weighted average Page 23

4.7 Step 5 - Recommendations for Decision Making Risk Reduction R S Cargo Spill Reduction R E Table 4.17: Results Cost Benefit C B ΔC GCAF= Δ R S ΔC CATS= Δ R E ΔC - ΔB NCAF= Δ Tonnes 1) Risk Control # of saved $ 1) 2) $ 1) 2) 3) $ $ $ Measure lives 1) N 2 IGS 4,000-8,000dwt 2.7E-02 5.99 562,503 461,000 20,725,000 94,000 3,725,000 chemical tankers N 2 IGS 8,000-20,000dwt 1.2E-01 9.46 562,503 734,000 4,843,000 59,000-1,473,000 chemical tankers Conventional IGS 4,000-8,000dwt oil 2.7E-02 2.21 231,800 143,000 8,541,000 105,000 3,276,000 tankers Conventional IGS 8,000-20,000dwt oil tankers 1.2E-01 8.05 231,800 615,000 1,996,000 29,000-3,297,000 1) Per ship lifetime, assumed to be 25 years 2) Includes NPV at 5% per year where relevant 3) Reduced PLC and PLP R S Figure 4.1: GCAF and NCAF for <20,000dwt Tankers Carrying <60 C Cargoes 25 20 USD (millions) 15 10 5 $3 million upper limit 0-5 Chemical tankers - >4,000dwt Chemical tankers - >8,000dwt Oil tankers - >4,000dwt Oil tankers - >8,000dwt GCAF 20,725,577 4,843,532 8,540,734 1,995,955 NCAF 3,725,019-1,473,403 3,276,174-3,296,969 The results displayed in Table 4.17 and Figure 4.1 above show that the conventional IGS on 8,000-20,000dwt oil tankers is cost effective both from a safety and Page 24

environmental perspective alone. From an NCAF standpoint fitting N 2 IGS to 8,000-20,000dwt chemical tankers is considered cost effective. However, NCAF for both 4,000-8,000dwt chemical and oil tankers is not grossly disproportionate to the value of life threshold of $3 million, being only 24% and 9% respectively above the upper limit. In this respect it should be considered that the individual risk may be close to or above the intolerable limit when personal accidents are included. The GCAF calculations show that the outcomes for the 4,000-8,000dwt category tankers are far higher than for the 8,000-20,000dwt tankers, particularly with regards to the >4,000dwt chemical tankers. This situation is suspected to be influenced by under-reporting of accidents in the smallest tanker category; this issue is discussed in more detail below. 4.7.1 Under Reporting There is the suggestion that under reporting is inherent in accident databases, such as LRFP, with the result that the results of this FSA may be underestimated. This viewpoint is based on an exercise undertaken where the fire and explosion incidents of Norwegian flagged vessels (NOR/NIS) recorded in LRFP, of which there were 26, were compared with fire and explosion events documented in the Marine Casualty Database (DAMA) maintained by the Norwegian Maritime Directorate, of which there were 50. Norwegian flagged vessels account for 107 ships of the type included within this FSA, which is the seventh largest fleet in the world (3% of world fleet). Therefore, the comparison between LRFP and DAMA fire and explosion accident reporting is significant. Of the total number of fire and explosion accidents in both databases (76) there was matching reporting in only 15 cases (30%). The following formula is used to estimate under reporting taking into account the probability of accident events not appearing in either database. The true number of accidents is represented by N. The probability of an accident being reported in LRFP is 26/N. The probability of an accident being reported in DAMA is 50/N. The probability of reporting in both databases is 15/N. Thus: 26/N*50/N = 15/N or N = 26*50/15 = 87 So the estimated number of accidents is 87. Therefore, overall LRFP contains only 26/87 or 30% of all fire and explosion accidents, suggesting an under reporting of 70%. Thus, the accident frequencies used in this study could potentially be increased by a factor of 3.3 to take account of underreporting if the NOR/NIS fleet was considered representative, although the fatalities associated with these accidents is an unknown. Page 25

Moreover, when individual risk is presented (Figure 4.2 and Figure 4.3) it is clear that there are far fewer accidents and associated lives lost in all accident categories for 4,000-8,000dwt tankers compared to the other tanker size categories, suggesting a certain degree of under reporting. The confidence intervals in Figure 4.2 indicate the statistical uncertainty (a confidence level of 0.9 was used), although this neither supports nor disproves the hypothesis of under reporting. However, the upper confidence level of the average total individual risk for 8,000-20,000dwt tankers is 8.6E-04, which is close to the intolerable threshold of 1.0E-03; it is presumed that when occupational risk is added it can not be concluded with confidence that the total individual risk is in the ALARP region, indicating that actions must be taken to reduce the risk irrespective of costs. Figure 4.2: Confidence Intervals and Individual Risk Comparison for All Accidents for Tankers Carrying <60 C Cargoes 1.0E-03 9.0E-04 8.0E-04 Individual Risk 7.0E-04 6.0E-04 5.0E-04 4.0E-04 3.0E-04 2.0E-04 1.0E-04 0.0E+00 >4,000dwt >8,000dwt >20,000dwt Average All cargo FX 6.5E-05 2.8E-04 1.1E-04 1.2E-04 All non cargo FX 1.6E-04 2.6E-04 2.2E-04 2.1E-04 All other Accidents 1.7E-05 2.2E-04 8.3E-05 9.0E-05 Total individual risk 2.4E-04 7.6E-04 4.1E-04 4.3E-04 Page 26

Figure 4.3: Confidence Intervals and Individual Risk Comparison for Cargo Tank Fire and Explosions for All Tankers Carrying <60 C Cargoes 4.0E-04 3.5E-04 3.0E-04 Individual Risk 2.5E-04 2.0E-04 1.5E-04 1.0E-04 5.0E-05 0.0E+00 >4,000dwt >8,000dwt >20,000dwt Average All cargo FX 6.5E-05 2.8E-04 1.1E-04 1.2E-04 Figure 4.3 above illustrates calculated individual risk and associated confidence intervals for cargo tank fire and explosions, all of which lie in the ALARP region; however, occupational accidents are not taken into account as this information is not included in the LRFP database. On the other hand, the GCAF and NCAF calculations in 4.4.3 and 4.5.3 illustrate that in the cases of 8,000-20,000dwt oil tankers and 8,000-20,000dwt chemical tankers the individual risk for these ship types displayed in Figure 4.3 may not be in the ALARP area. In support of the above argument on poor reporting the issue is also recognised by IIWG in MSC81/8/1, paragraph 3 [8]: A letter was sent on behalf of the IIWG to the owners and/or operators of the ships involved in the incidents identified, requesting that they provide any information they might have to the Group. Data provided on the various incidents was of variable quantity and quality, although the Group felt that some value was derived from all the data received. Despite the invitation made by MSC 80 to the relevant Administrations to provide further information on the findings of the investigations into recent casualties, little new information was provided to the IIWG. In addition, IACS suspects large under reporting for general cargo ships <20,000 gross tonnes and in MSC85/19/1 [18], paragraph 8, stated the following with regards to under reporting in general: The investigation of historical data on general cargo ships found in databases indicates that: 1. the issue of the completeness of incident reporting is unsolved and thus an unknown percentage of underreporting exists; Page 27

2. when broken down to sub-categories (e.g., flag, classification society etc.), the results are dominated by uncertainties; and 3. in order to complete an FSA of high quality, it is essential to complement the available data sources with additional data, which needs to be provided by flag States and any other organizations which have relevant data to contribute. 4.7.2 Root Cause of Reduction in Fire and Explosion Accidents The rationale for judging the risk reduction potential of IGS is based on the assumption that reduced PLL on <20,000dwt tankers between the two periods of 1978-1983 and 1990-2007 is due to the introduction of IGS post-1990. However, studying the accident statistics in Table 4.1, it is clear that there is also a marked reduction in PLL between 1978-1983 and 1990-2007 for 4,000-8,000dwt tankers of 86%; the significant reduction in risk to life between 1978-1983 and 1990-2007 cannot be explained by the introduction of IGS as it is not required for these size tankers for either period. Therefore, either other aspects besides IGS have reduced the risk of cargo tank fire and explosion, such as improved safety management systems, or there is a considerable degree of under reporting in the smallest tanker category; the second supposition is supported by the apparent under reporting identified and explained in 4.7.1. 4.7.3 Fatalities Due to Tank Entry With the introduction of IGS on <20,000dwt tankers the associated risk of asphyxiation due to entry into inerted tanks must be considered. As LRFP does not record death due to tank entry several Intertanko members provided fatality information which suggests that the PLL due to inerted tank entry is 4.4E-03. However, this calculation was based on a small dataset containing relatively few shipyears; in this regard it is suspected that additional shipyears should be included in the evaluation to encompass the entire period of time when crew were at risk from death due to entry into inerted tanks, resulting in a more realistic PLL. Moreover, it should be recognised that the risk associated with tank entry can be significantly lowered with adequate training of suitably qualified and experienced personnel and ensuring the enforcement of strict safety procedures. For example, Odfjell s Safety and Quality Management System requires that personal oxygen meters or O 2 /LEL multi-meters are used when in enclosed spaces and cargo tanks [19]. On the other hand risk of fire and explosion in the cargo hold cannot be reduced to a commensurate degree through training alone; this can only be achieved through IGS. There is also reason to believe that the risk of fatality associated with tank entry will be reduced if the inert requirement was based on flashpoint alone, irrespective of cargo carried, ship deadweight and tank size (i.e. the same rule for low flash chemicals as is already the case for low flash petroleum products) [20]. In this respect the IG rule would be absolute, eliminating uncertainty in terms of whether the tanks are inerted or not. Page 28