REVISION OF THE DAMAGE STABILITY REGULATIONS FOR RO-RO PASSENGER SHIPS

Transcription

1 E SUB-COMMITTEE ON STABILITY AND LOAD LINES AND ON FISHING VESSELS SAFETY 55th session Agenda item 7 SLF 55/INF.9 14 December 2012 ENGLISH ONLY REVISION OF THE DAMAGE STABILITY REGULATIONS FOR RO-RO PASSENGER SHIPS The GOAL based Damage Stability project (GOALDS) Development of a new risk-based damage stability requirement for passenger ships based on Cost-Benefit Assessment Submitted by Denmark and the United Kingdom SUMMARY Executive summary: This document contains a report on the development of a new risk-based damage stability requirement for passenger ships based on cost-benefit assessment carried out within the GOAL based Damage Stability project (GOALDS) Strategic direction: 5.1 High-level action: Planned output: Action to be taken: Paragraph 19 Related documents: SLF 45/3/3; SLF 46/INF.6; SLF 52/11/1; MSC 84/22/12 SLF 52/WP.3 and MSC 91/7/2 Introduction 1 This document provides information on the results of the "GOAL based Damage Stability project" (GOALDS). This document should be seen in conjunction with documents SLF 55/INF.7 and SLF 55/INF.8 concerning the GOALDS project. 2 The study has been partially funded by the European Commission under the 7th Research Framework Programme Theme "Sustainable Surface Transport". 3 In the framework of GOALDS project, risk models covering collision and grounding of passenger ships (RoPax and cruise ships) have been developed and used for cost-benefit assessments of design changes on representative sample ships. The design changes (implemented Risk Control Options) were aiming at improving the ship's survivability in case of collision or grounding. 4 In addition, multi-objective optimization studies based on parametric ship models have been carried out to derive cost-effective designs with high degree of survivability. I:\SLF\55\INF-9.doc

2 SLF 55/INF.9 Page 2 Background 5 MSC 84 agreed to include a new item on "Damage stability regulations for ro-ro passenger ships" in the Sub-Committee's work programme. 6 The Sub-Committee, at its fifty-second session, discussed in detail the impact of the SOLAS 2009 amendments on the damage stability requirements for ro-ro passenger ships after the presentation of three European Union funded research projects (GOALDS, FLOODSTAND and EMSA2), and whether in this regard any amendment to SOLAS should be considered. 7 The Sub-Committee noted the general view of the SDS Working Group that more research and the evaluation of further studies were important and necessary before considering any possible additional measures. Following a request by the Sub-Committee, MSC 89 agreed to extend the target completion year for this item to The Sub-Committee, at its fifty-third session, instructed the SDS Correspondence Group to further consider the impact of the SOLAS 2009 amendments on ro-ro passenger ships, as compared to the SOLAS 1990 regulations in association with the Stockholm Agreement, taking into account document SLF 52/WP.3, and any research results in the matter as they become available. Main contents of the GOALDS project 9 The GOALDS research project contained the following main sub-projects:.1 extending the formulation introduced by resolution MSC.216(82) for the assessment of the probability of survival of passenger ships in damaged condition, based on extensive use of numerical simulations;.2 performing comprehensive model testing to investigate the process of ship stability deterioration in damaged condition and to provide a basis for the validation of the numerical simulation results;.3 compiling damage statistics and probability functions for the damage location, length, breadth and penetration in case of a collision/grounding accident, based on a thorough review of available information regarding these accidents over the past 60 years worldwide;.4 formulating a new probabilistic damage stability concept for passenger ships, incorporating collision and grounding damages, along with an alternative method for the calculation of ship survival probability;.5 establishing new risk-based damage stability requirements of passenger vessels based on a cost-benefit analysis to establish the highest level for the required subdivision index;.6 to demonstrate that a potential commercial viable passenger vessel could be built to a significantly higher Attained Index than set forth by current requirements; and.7 investigating the impact of the new formulation for the probabilistic damage stability evaluation of passenger ships on the design and operational characteristics of a typical set of ROPAX and cruise vessel designs (case studies). I:\SLF\55\INF-9.doc

3 SLF 55/INF.9 Page 3 Main findings and results included in this document 10 Risk models in the form of event trees covering collision and grounding have been established. 11 The event trees have been populated with frequencies and consequencies derived from IHS Fairplay, in addition to GISIS and the GOALDS databases. 12 Cost-benefit assessments have been carried out on sample ships provided by the project shipyards; STX-France, STX-Finland, Meyer Werft and Fincantieri. 13 Financial data; costs and benefits for specific design changes have been provided by the shipyards listed in item 12 and ship operators; RCCL, Carnival Cruise and Color Line. 14 The study has been carried out following the IMO FSA Guidelines. In line with these guidelines an update of the Cost of Averting a Fatality (CAF) criterion has been proposed. 15 The attained index A for collision as applied in the calculations is calculated based on the new formulation of the factor s for survivability presented in document SLF 55/INF.8. Additionally and for reference the attained index A calculated according to current SOLAS has been included for comparison as well. 16 Optimization studies have been carried out for two small ropax ships, a medium ropax, a large ropax, a medium cruise ship and the large cruise ship providing additional data for the set up of the new risk-based damage stability requirement. 17 It is concluded from these studies that potentially commercially viable passenger vessels (of RoPax and cruise type) could be built to a significantly higher Attained Index than set forth by current requirements. 18 The highest attained indices A obtained for ships and for design changes (RCOs) found to be cost-effective are suggested to be used as the basis for the proposed new level of R for passenger ships. Action requested of the Sub-Committee 19 The Sub-Committee is invited to note the research carried out by the GOALDS project and included as an annex to this document, and to consider its findings within its work on the item on "Revision of damage stability regulations for passenger ships". *** I:\SLF\55\INF-9.doc

4

5 Grant Agreement No: Project Acronym: GOALDS Project Title: GOAL based Damage Stability Deliverable D Development of a new damage stability requirement based on Cost-Benefit Assessment Document Id. Due date of Deliverable: Actual Submission Date: Rainer Hamann, Odd Olufsen, George Zaraphonitis -document author/s- Christian Mains -document approved by- insert adress to reply to -revision type PU 1 -date of last update- -distribution level- 1 dissemination level PU Public PP Restricted to Programme Participants (including Commission Services) RE Restricted to a group specified by the Consortium (including Commission Services) CO Confidential, only for members of the consortium (including Commission Services) GOALDS-D-WP.7.3-GL-INF Paper rev Page 1 of 70

6 Disclaimer The information contained in this report is subject to change without notice and should not be construed as a commitment by any members of the GOALDS Consortium or the authors. In the event of any software or algorithms being described in this report, the GOALDS Consortium assumes no responsibility for the use or inability to use any of its software or algorithms. The information is provided without any warranty of any kind and the GOALDS Consortium expressly disclaims all implied warranties, including but not limited to the implied warranties of merchantability and fitness for a particular use. (c) COPYRIGHT 2009 The GOALDS Consortium This document may be not copied and reproduced without written permission from the GOALDS Consortium. Acknowledgement of the authors of the document shall be clearly referenced. All rights reserved. Document History Document ID. Date Description GOALDS-D-WP.7-GL IMO INF3 rev Final update Page 2 of 70

7 Document Control Sheet Title: Abstract This report summarises the work for developing a proposal for the required damage stability index R of passenger ships using the Coast-Benefit Assessment (CBA). For the CBA risk models for collision and grounding accidents were developed based on comprehensive investigation of casualty reports and under consideration of the expertise provided by the project partners. The developed risk models allow a calculation of risk in terms of fatalities and loss of ship related to the damage stability of the ship. For representative sample ships design variations focusing on increased damage stability were developed, so-called Risk Control Options. These design variations were developed by yards and application of numerical optimisation methods. For CBA the additional costs of these RCOs were determined and evaluated as specified in the FSA guidelines. Cost beneficial RCOs were used to develop a proposal for R-Index related to people onboard. Summary Report: This section must provide information under the following headlines: Introduction One main objective of GOALDS is the determination of R-Index values by means of a Cost-Benefit Assessment (CBA) for the probabilistic damage stability model used in SOLAS. In cost benefit evaluation the costs for risk control options are evaluated with respect to their risk reducing effect (refer also to FSA Guidelines, 2007). In context of GOALDS the risk reducing measures, also named Risk Control Options (RCOs), were focused on the reduction of the consequences. In order perform a cost benefit evaluation a quantitative risk model is required which is used to calculate the risk reduction of risk control option. For the present analysis the risk is expressed in terms of fatalities and, to consider also the monetary benefit, also in terms of property loss (US$). State of the Art In order to support the IMO discussion on introducing new regulations IMO has agreed the process of Formal Safety Assessment (FSA). This process is based on a risk analysis and a cost-benefit assessment of potential risk control options. Value added to GOALDS This report summarises the results of WP 5 in order to support the discussion on new damage stability requirements for passenger ships. Achievements Summary report to be submitted as INF paper. Not achieved. -- Input from other Deliverables This report is based on the results from WP 5 and WP 6. This executive summary may be published outside the GOALDS consortium. NO Work carried out by Approved by Rainer Hamann, Odd Olufsen, George Zaraphonitis Christian Mains (GL) - signature on file - - signature of internal reviewer and date of acceptance - Page 3 of 70

8 Apostolos Papanikolaou (NTUA) - signature on file - - signature of external reviewer and date of acceptance - Page 4 of 70

9 Table of contents 1 INTRODUCTION RISK MODELS DATA SOURCES AND STATISTICAL DATA COLLISION GROUNDING CBA METHODOLOGY AND ASSUMPTIONS SAMPLE SHIPS AVAILABLE FOR THE COST BENEFIT ASSESSMENTS SELECTION OF RISK CONTROL OPTION GENERAL FINANCIAL KEY FIGURES Present value and interest rate Lifetime of ship and investment Fuel cost Cost of added built steel Cost of increased installed power Increased harbour fees FINANCIAL KEY FIGURES FOR CRUISE Cost of increased outfitting area Cost of increased hotel load FINANCIAL KEY FIGURES FOR ROPAX Financial benefit of increased capacity COST OF AVERTING A FATALITY (CAF) CRITERION SUMMARY OF COSTS/REVENUE FOR RCOS APPLIED ON SAMPLE SHIP SUMMARY OF COST AND BENEFITS Results based on collision model A and grounding formulation Results based on collision model B and grounding formulation Results based on collision model A and grounding formulation Results based on collision model B and grounding formulation RESULTS FOR RCOS WITHIN CAF CRITERION AND THE IMPACT ON ATTAINED A-INDEX DISCUSSION OF THE RESULTS FROM THE CBA Fuel impact Financial benefits Cost level uncertainty Recommending the level of required index R Impact of using different risk models Simplifications and optimisations INNOVATIVE SHIP CONCEPT DESIGNS BASED ON THE NEW DAMAGE STABILITY REQUIREMENT INTRODUCTION OPTIMIZATION PROCEDURES OPTIMIZATION STUDIES First small RoPax Second small RoPax Large RoPax Medium RoPax Post Panamax Cruise Ship Panamax Cruise Ship CONCLUSIONS GOALDS-D-WP.7.3-GL-INF Paper rev Page 5 of 70

10 5 RESULTS NEW REQUIRED DAMAGE STABILITY INDEX (R-INDEX) REFERENCES Page 6 of 70

11 1 Introduction One main objective of GOALDS is the determination of R-Index values by means of a Cost-Benefit Assessment (CBA) for the probabilistic damage stability model used in SOLAS. In cost benefit evaluation the costs for risk control options are evaluated with respect to their risk reducing effect (refer also to FSA Guidelines, 2007). In context of GOALDS the risk reducing measures, also named Risk Control Options (RCOs), were focused on the reduction of the consequences. In order perform a cost benefit evaluation a quantitative risk model is required which is used to calculate the risk reduction of risk control option. For the present analysis the risk is expressed in terms of fatalities and, to consider also the monetary benefit, also in terms of property loss (US$). Because GOALDS has focused on the probabilistic damage stability model in view of Loss of Watertight Integrity (LOWI) and ship s flooding, in this investigation only accident categories collision (CN) and grounding or grounding (GR) are considered. In this section of the report the risk models developed for these accident categories are explained in detail. The risk models for the two accident categories were developed in form of event trees. GOALDS risk models were developed based on a comprehensive analysis of casualty reports and an investigation of risk models developed in previous studies, e.g. the FSA on Cruise ships (MSC 85/INF.2) and the FSA on RoPax vessel (MSC 85/INF.3). The database analyses provide information not only for the development of the event sequences but also to estimate the initial accident frequencies as well as dependent probabilities in the risk models. This report contains results of the database investigations as well as a description of the risk models. The CBA was performed for a set of sample ships representative for Cruise and RoPax of current world fleet. In view of the above, for each of the selected sample ships the following steps have been followed: 1) Calculate the damage stability and establish the attained index A of the design according to the new GOALDS formulation and additionally according to the currently in force SOLAS ) For each design modification or combination of modifications (applied RCOs) re-calculate the attained index A for the two alternative formulations. 3) For each design modification assess the financial costs and benefits in terms of reduction of risk and economic impact.. 4) Use IMO FSA methodology/guidelines to conclude whether the proposed design modification is cost beneficial. The results, i.e. the obtained attained indices A for the designs modifications which were found cost-efficient are used as a basis for recommending the level of the required index R. The recommended level of R is supported by a series of results from systematic optimisation studies for sample ships, though which the feasibility of the recommended level of required index is proven. 2 Risk models 2.1 Data sources and statistical data IHS Fairplay casualty data for the period (17 years) and IHS Fairplay ship register were used to calculate initial CN and GR frequencies for the following ship categories: 1. Cruise; 2. Pax; 3. RoPax and; 4. RoPax-Rail. Page 7 of 70

12 In addition, CN and GR accidents of GISIS and GOALDS databases were also analysed for the same period of time to ensure that the maximum number of casualties were included in GOALDS dataset. In this respect, a further two CNs were added from GISIS, and from GOALDS an additional six CNs and seven GRs were added. 17 years of casualty data was deemed sufficient to develop robust and up-to date accident frequencies. In order to base the analyses on a representative sample of the fleet that represents the current status of ship building technology IHS Fairplay ship register was analysed and the following selection criteria were specified for the further development of the risk models: Ship types: Cruise and Ro-Pax; GT 1,000 most ships below 1,000 gross tonne operate on non-international voyages; 80 m length - most ships below 80m in length operate on non-international voyages; Built 1982 the sample shall represent the current ship design and its safety level; therefore older ships were excluded; IACS class at time of accident to reduce the potential effect of under-reporting; IACS class for determination of ship years; Froude No. 0.5 to eliminate HSC from this study because HSC are built following other requirements and have a different operation scheme; Serious CN/WS only as only interested in accidents where the damage stability of the vessel could be potentially compromised (i.e. due to water ingress). These selection criteria were applied as filters for the fleet at risk as well as for the casualty reports. However, all casualty reports were analysed in detail aiming on a correct classification of accident severity as well as consequences, e.g. collision between a collision between a large RoPax and a fishing vessel that was originally classified as serious which it was definitely not for the RoPax. The calculated number of ship years for the period for each ship type is: RoPax shipyears ; RoPax-Rail shipyears ; Cruise shipyears ; Pax shipyears The number of ship years was calculated for the time between due or delivered and either one of the following: the end of the time interval ( ); the scrap date; the date of loss. The distribution of the ship years over the period considered for this analysis is shown in Figure 1. As shown by this figure the fleets of Cruise and RoPax have been characterised by continuous growth over the past 17 years, whereas the fleets of RoPax-Rail and Pax are mostly constant with respect to number of ships. The current fleet of RoPax vessels is nearly twice as large as the Cruise ship fleet. Both, IACS classified world fleet of passenger ships (Pax) as well as RoPax-Rail are rather small. Page 8 of 70

13 No ship years Cruise Pax RoPax RoPax-Rail '94 '95 '96 '97 '98 '99 '00 '01 '02 '03 '04 '05 '06 '07 '08 '09 '10 Year Figure 1: Distribution of ship years broken down into the ship categories Cruise, Pax, RoPax and RoPax-Rail. (Ships built after 1981, GT > 999, IACS classed, LOA 80 m) In order to calculate the initial accident frequencies casualty reports concerning collision and grounding accidents were composed using above mentioned databases (sources) in order to determine the maximum number of available casualty reports. The result for collision accidents was summarised in Table 1. As shown only few additional casualty reports could be added to that of IHS Fairplay. In total 62 casualty reports were considered for further evaluation, 19 for Cruise and Pax, and 43 for RoPax and RoPax-Rail. For the quantitative risk analyses within GOALDS two ship types were defined Cruise covering Cruise vessels and Passenger vessels; RoPax covering RoPax and RoPax-Rail. Table 1: Number of casualty reports collision for the period 1994 to 2010 for IACS classified ships Ship type IHS GISIS LMIU GOALDS Cruise Pax RoPax RoPax-Rail The annual accident frequencies for Cruise and Ro-Pax+Ro-Pax-Rail vessels are shown in Figure 2. As shown the annual accident frequency varies significantly between 1992 and Additionally, the ten-year average accident frequencies were calculated for the period 1994 and 2010 (Figure 3). This plot clearly shows that the accident frequencies for both ship types under consideration increase towards the end of the investigation period. Page 9 of 70

14 Accident frequency Acc. frequency 2.5E E-02 Cruise+Pax RoPax+RoPax-Rail 1.5E E E E Figure 2: Annual accident frequencies of Cruise and RoPax vessels between 1994 and 2010 Year 1.2E E-02 Cruise+Pax RoPax+RoPax-Rail 8.0E E E E E year interval Figure 3: Average collision accident frequencies plotted versus 10-year intervals between 1994 and Initial annual accident frequencies were calculated by means of number of accidents and fleet at risk (Table 2). Page 10 of 70

15 No accidents GOALDS was not the first project developing risk models for collision and grounding of ROPAX and Cruise vessels and therefore the GOALDS results were compared with these projects in particular with the FSAs developed in the EU funded project SAFEDOR. This comparison showed some deviation between the initial accident frequencies. These variations may be explained by the different assumptions on the fleet at risk, consideration of accidents (severe non severe), the effect of utilizing the more recent data as well as in case of RoPax the usage of different databases (LMIU IHS Fairplay). However, when taking into account the uncertainty the accident frequencies determined in SAFEDOR are in the 90% confidence interval of the statistical figures of GOALDS. Table 2: Number of casualty reports collision, ship years and accident frequency for the period 1994 to 2010 for IACS classified ships Ship type No of Ship years Accident frequency SAFEDOR FSAs accidents Cruise+Pax E E-3 Ro-Pax+ Ro-Pax-rail E E-2 Similar to the determination of collision accident frequencies the data for grounding accidents was analysed. The distribution of grounding accidents over the period under consideration broken down into the ship categories Cruise and RoPax is shown in Figure 4. It is noted that this figure shows the annual number and therefore allows no direct conclusions on the risk or the annual accident frequency Cruise Pax RoPax RoPax-Rail WS Year Figure 4: Number of grounding accidents per year between 1994 and 2010 The annual accident frequencies for Cruise and RoPax vessel are shown in Figure 8. Similar to the accident category collision a significant variation between the annual values is observed. Page 11 of 70

16 Accident frequency Acc. frequency 4.0E E E-02 Cruise+Pax RoPax+RoPax-Rail 2.5E E E E E E Figure 5: Annual accident frequencies (grounding) of Cruise+Pax and Ro-Pax+Ro-Pax- Rail vessels between 1994 and In Figure 6 the average accident frequencies for ten year intervals starting between 1994 and 2001 are shown for the same groups. The average ten year accident frequency nearly doubles for RoPax/RoPax-Rail vessels towards the end of investigated period whereas for Cruise/Pax a decrease by approximately 20% is observed. 1.4E-02 Year 1.2E E E-03 Cruise+Pax RoPax+RoPax-Rail 6.0E E E E year interval Figure 6: Average grounding accident frequency plotted versus 10-year intervals between 1994 and 2010 Page 12 of 70

17 The initial frequencies of grounding accidents for both ship types under consideration were calculated. The number of accidents as well as the sources of these casualty reports were summarised in Table 3 and corresponding frequencies in Table 4. Table 4 contains also the respective values of the FSAs for Cruise and RoPax. For Cruise vessel the difference between GOALDS values and that of the FSA are negligible whereas for RoPax the difference is about 40%. Potential cause for this difference is the development as shown in Figure 6 as well as causes that were already discussed in context with CN above. Table 3: Number of casualty reports grounding for the period 1994 to 2010 for IACS classified ships Ship type IHS LMIU GISIS GOALDS Cruise Pax RoPax RoPax-Rail Table 4: Number of casualty reports grounding, ship years and accident frequency for the period 1994 to 2010 for IACS classified ships Ship type No of accidents Ship years Accident frequency SAFEDOR FSAs Cruise+Pax E E-03 Ro-Pax+ Ro-Pax-rail E E Collision Four collision risk models for both ship types under consideration (two for each ship type each having a different consequence model A and B) were developed based on the high level event sequence shown in Figure 7. The corresponding event trees for the two ship types are shown in Figure 8 and Figure 9 with all input data. Dependent probabilities were determined based on the information provided by the casualty reports. Level 1 Level 2 Level 3 Level 4 Level 5 Collision Struck / Striking Operational State Water Ingress Sinking Consequences Figure 7: High level event sequence for accident category collision Page 13 of 70

18 Ship type Yes No Collision Struck 0.699% Initiator Striking % % En Route 37.5% Operational State Limited Waters Terminal 62.5% % 4.0% 23.0% 73.0% Yes No Yes No Yes No Water ingress Water ingress Water ingress Yes % Sinking (A-Index) No % 0.006% Yes % Sinking (A-Index) No % % Yes % Sinking (A-Index) No % % Fast-Capsize 27.0% Velocity Slowly 73.0% % Fast-Capsize 27.0% Velocity Slowly 73.0% % Fast-Capsize 27.0% Velocity Slowly 73.0% % % 2.16E % 9.82E % % 82.0% % 18.0% % 82.0% 0.005% Figure 8: CN model for Cruise vessel Page 14 of 70

19 Ship type Yes No Collision Struck % Initiator Striking % % En Route % Operational State Limited Waters Terminal % % 4.0% 23.0% 73.0% Yes No Yes No Yes No Water ingress Water ingress Water ingress Yes % Sinking (A-Index) No % % Yes % Sinking (A-Index) No % % Yes % Sinking (A-Index) No % % Fast-Capsize 27.0% Velocity Slowly 73.0% % Fast-Capsize 27.0% Velocity Slowly 73.0% % Fast-Capsize 27.0% Velocity Slowly 73.0% % % % 50.0% % 50.0% 0.007% 50.0% 0.007% 50.0% % 50.0% % Figure 9: CN model for RoPax vessel Page 15 of 70

20 In Level 1 (Figure 7) struck/striking and in Level 2 the operational states En route, Limited Waters and Terminal were considered. The scenarios for Striking were not further developed in this risk model because generally these accidents will not cause a loss of stability. In Level 3 the focus is on the possibility of water ingress. If water ingress occurs the ship may sink or not, and in case of sinking this can take place slowly or fast. The latter case considers also the possibility of capsizing. The probability of sinking was related to the A-Index which means that 1-A is directly the probability of sinking. Two different assumptions (model A and B) were used in GOALDS to estimate the number of fatalities: A If the ship sinks, it most likely that the ship capsizes and no orderly (SOLAS compliant) evacuation is possible. For this assumption 100% fatalities were considered. B Even if the ship sinks, different scenarios stretching from slow sinking to fast capsize are possible. People will not always follow the defined evacuation process and therefore fatality rate is below 100% even for rapid capsize are also in capsize events.. For example, the ferry Estonia 1 lost rapidly stability and sank. However the fatality rate is 87%. Therefore the second assumption considers slow and fast sinking. For cruise ships it assumed that the ratio fast/slow sinking is 18/82 % and the corresponding rate of fatalities is 80 % for the fast sinking case and 5 % for the slow sinking. For RoPax ships the distribution slow/fast sinking is assumed to be 50/50 % with the same fatality ratio as cruise ships. 2.3 Grounding The collision risk model for both ship types under consideration was developed based on the high level event sequence shown in Figure 10. The corresponding risk model for Cruise ships with all data is shown in Figure 11, respectively for RoPax in Figure 12. Like in the CN risk model the dependent probability of sinking was calculated using the A-Index, in this case the A-Index for grounding. In Level 1 of the risk model drift and powered grounding were distinguished. In the subsequent Level 2 the effect of different sea bed conditions (soft/hard) were considered. In case of soft sea bed penetration of hull is regarded to be impossible. Therefore, water ingress was considered only in scenarios with grounding on hard sea bed. Sinking of ship is possible for scenarios with water ingress. In GR risk models the consequence model B was used to estimate the number of fatalities due to the fact that grounding accident typically occur in coastal areas with better access for SAR. It has been assumed that the A-index relevant for grounding(a GR ) is a function of the A-index for collision (A CN ), and the following two formulations have been used in the CBA presented in Section 3 and are referred to as grounding formulation 1 and 2 respectively. A A 0.1, but less than 1.0 GR CN Equation 1 Grounding formulation 1 and A GR A 0.11 A CN CN Equation 2 Grounding formulation 2 1 The Estonia accident was not a collision event in terms of its inception, but resulted in flooding of ship s car deck, which could have been the result of a side collision in high seas. Page 16 of 70

21 Level 1 Level 2 Level 3 Level 4 Level 5 Level 6 Grounding Navigation Sea Bed Water Ingress Staying Aground Afloat Consequences Figure 10: High level event sequence for accident category grounding Page 17 of 70

22 Vessel II Yes No Grounding Drift % Navigation Powered % % Soft 11.0% Sea Bed Hard Soft 89.0% Sea Bed Hard Yes 25.0% Water Ingress No Yes 75.0% Water ingress No Yes 25.0% Water Ingress No Yes 75.0% Water ingress No 0.0% 0.0% 0 Yes 100.0% Staying aground No Yes 92.0% Staying aground No 8.0% 0.007% 0.0% 0.0% Yes 100.0% Staying aground No Yes 92.0% Staying aground No 8.0% 0.057% 100.0% % 0.0% 0.0% % 0.05% No - Proceed to % Sinking Yes 100.0% % 0.0% 0.0% % % No - Proceed to % Sinking Yes % % Slow % Velocity Fast % % Slow % Velocity Fast 82.0% 9.36E % 2.05E % % 18.0% % Figure 11: Grounding risk model values for Cruise vessel Page 18 of 70

23 Vessel II Yes No Grounding Drift % Navigation Powered % % Soft 11.0% Sea Bed Hard Soft 89.0% Sea Bed Hard Yes 25.0% Water Ingress No Yes 75.0% Water ingress No Yes 25.0% Water Ingress No Yes 75.0% Water ingress No 0.0% 0.0% 0 Yes 100.0% Staying aground No Yes 92.0% Staying aground No 8.0% % 0.0% 0.0% Yes 100.0% Staying aground No Yes 92.0% Staying aground No 8.0% % 100.0% % 0.0% 0.0% % % No - Proceed to % Sinking Yes 100.0% % 0.0% 0.0% % % No - Proceed to % Sinking Yes % 0.014% Slow % Velocity Fast % % Slow % Velocity Fast 50.0% 1.05E % 1.05E % 8.51E % 8.51E-06 Figure 12: Grounding risk model values for RoPax vessel Page 19 of 70

24 3 CBA 3.1 Methodology and assumptions The cost benefit analysis follows the structure applied for Formal Safety Assessment as described by IMO (1). Figure 1 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. Figure 13 - FSA Methodology The following subsections are based on the IMO FSA Guidelines (1). This work is mainly related to the FSA steps 3 and 4 but it is a part of the process to assess the risk reduction effect of identified risk control options as well. 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 measures (RCMs) or options (RCOs) 2. Finally, the objective of the cost efficiency assessment step is to identify and rank the risk control options in order to determine the most cost beneficial and efficient ones, i.e. those that provide most risk reduction in relation to cost. In order to compare single risk control measures or combinations of measures (risk control options) in a systematic and structured way, the risk models presented in Sec. 2 are used for the evaluation. 3.2 Sample ships available for the cost benefit assessments The studies on the effects of Risk Control Options have been based on sample ships made available by the project partners. All sample ships are in compliance with current SOLAS regulations and should thus be representative for the current standard. The following ship types were made available for this study: o A large cruise vessel, ref. Table 5 o A medium sized cruise vessel, ref. Table 6 o A large RoPax, ref. Table 7 o A medium sized RoPax, ref. Table 8 2 Refer to FSA Guidelines for further definition of risk control measure and risk control option. Page 20 of 70

25 Table 5 Main parameters, large cruise vessel: Length over all ~ 327 m Length between perpendiculars m Subdivision length m Breadth m Subdivision draught 8.80 m Height of bulkhead deck m Number of passengers 4,200 Number of crew 1,400 Gross tonnage 125,000 Deadweight 10,500 t No of cabins 1,664 Table 6 Main parameters, medium sized cruise vessel Length over all ~ m Length between perpendiculars m Subdivision length m Breadth m Subdivision draught 7.40 m Height of bulkhead deck 9.80 m Number of passengers 1,750 Number of crew 650 Gross tonnage 63,000 Design draft 7.20 m Design deadweight 6,100 t Light service draft 6.89 m No of cabins 1,664 People in life boats 1,800 SOLAS Required index Table 7 Main parameters, large RoPax vessel: Length over all Approximate 229 m Length between perpendiculars m Subdivision length m Breadth 32 m Subdivision draught 6.70 m Height of bulkhead deck 9.70 m Number of passengers 3300 Number of crew 200 Gross tonnage Deadweight 6900 t No of cabins 1000 Lane meter 1500 No of cars 1000 Page 21 of 70

26 Table 8 Main parameters, medium sized RoPax vessel: Length over all ~ 183 m Length between perpendiculars m Subdivision length m Breadth 27.6 m Subdivision draught 7.1 m Height of bulkhead deck 9.80 m Number of passengers 2080 Number of crew 120 Gross tonnage Abt. 36,000 Deadweight 5,000 t Lanemeters 1,950m No of cars 670 No of cabins 180 Design Speed 27.5 Kn 3.3 Selection of Risk Control Option The Risk Control Options are for the purpose of this study limited to those affecting ship s capability to survive a damage and subsequent water ingress when subject to collision or grounding, i.e. the ship s attained index A for collision and/or grounding. Risk control options that could reduce the probability for collision or grounding to occur are not considered in this context. The risk control options relevant to each of the sample ships were proposed and assessed in the design teams for each ship. Each design team consisted of one shipyard (coordinator), one operator, one class society, one flag administration and one university/research institute. It was subject to discussion how the RCOs shall be selected. It is sometimes argued that e.g. a change of the main dimensions is not a risk control option for a given design, and should be considered as the assessment of a new ship. In this study it was agreed that the risk control options should include changes in the main ship parameters as well. This is motivated by the fact that in particular the increase in breadth is the most efficient/drastic way to improve ship s stability. The RCOs were finally selected based on the assumption that basic ship requirements (ship specifications) are kept constant. The parameters that were subject to variations are briefly described as: Increase of freeboard Increase of breadth New subdivision, including watertight subdivision of Ro-Ro spaces Additional watertight volumes above bulkhead deck Combination of measures The purpose of step 4 (Figure 1) as described in (1) is to identify and compare the achieved risk reduction with the costs and benefits associated with the implementation of each RCO identified and defined in step 3 of the FSA analysis process (Figure 1). A cost efficiency assessment following the IMO procedure may consist of the following stages: Page 22 of 70

27 1. Consider the risks assessed in step 2 (Figure 1), 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 (Figure 1), 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 by reassessing the risk assuming the option under consideration is in place and comparing this risk level to the established base case; 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-efficiency perspective in order to facilitate the decision-making recommendations in step 5 (Figure 1) (e.g. to screen those that are not cost effective or impractical). Costs should be expressed in terms of life cycle costs and may include initial setup, operating, training, inspection, certification, decommission etc. Benefits may include reductions in fatalities, injuries, casualties, environmental damage and clean-up, etc. and an increase in the expected operating life of ships. There are two indices used by IMO that 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). The definitions of GCAF and NCAF are: C GCAF R NCAF C B R Where: ΔC is the cost per ship of the risk control option during the lifetime of the vessel. ΔB is the economic benefit per ship resulting from the implementation of the risk control option during the lifetime of the vessel. ΔR is the risk reduction per ship, in terms of ΔPLL The output from this step comprises: 1. Costs and benefits for each RCO identified in step 3 (Figure 1) 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. 3.4 General financial key figures Present value and interest rate All monetary values have been calculated to present values. All delta operational costs due to a RCO have been calculated as present values of annual costs over the lifetime of the vessel with a depreciation rate of 5%. This depreciation rate has also been used in previous FSAs submitted to IMO.. Page 23 of 70

28 3.4.2 Lifetime of ship and investment The lifetime of the ship and investment are both set to be 30 years). Most owners will use a shorter investment period for a new ship; however the costs are to be seen from the society s point of view. Therefore the investment time will be equal to the ship s expected lifetime Fuel cost The fuel consumption for each of the selected ships may be affected by the applied RCOs. In order to determine the added lifetime fuel cost of each RCO a fuel cost prediction will be applied to each type of fuel consumed. An important factor affecting the fuel cost for each of the case vessels is the fuel mix applied during its lifetime. New regulations on sulphur content may impact the fuel mix applied. Each of the case vessels will therefore apply a specific fuel mix which varies over its lifetime. The fuel mix evolution will be ship and trade specific. The following fuel mixes have been applied to the various case vessels, and is resulting from the qualified opinion of ship yards and operators on the possible future development and alternatives. The fuel mix for the large cruise ship is significantly different compared with the other ships, as it is expected that for large ships measures like scrubbers will be applied and HFO can be used for a longer period of time: Figure 14 - Fuel mix Large RoPax Page 24 of 70

29 Figure 15 - Fuel mix Medium RoPax Figure 16 - Fuel mix Large Cruise Page 25 of 70

30 Figure 17 - Fuel mix Medium Cruise In addition to the change in fuel mix used it is expected that the fuel cost will increase over the coming years, see Figure 18. In this study it is assumed that the marine fuel price evolution correlates to the evolution in global crude oil prices. Based on this the price path for heavy fuel oil (HFO), low sulphur heavy fuel oil (LSHFO) and marine gas oil (MGO) has been made based on EIA s reference case prediction for oil price evolution (2). EIA s predictions only go to 2035, in order to predict the remaining years of the lifetime of the vessels the fuel cost from 2036 to 2040 has been linearly forecasted based on the previous cost in the 3 years. Figure 18 - Fuel price paths A consideration that may have an effect on the future fuel cost for RoPax and Cruise vessels is the introduction of LNG as fuel. The use of LNG is likely to be driven both by environmental regulations and cost, and thus may impact on the added fuel cost that a RCO may impose on a ship design. The future LNG prices will not be used to determine the CAF figures in this study. However it is evident, based on the estimated future LNG price, when comparing to conventional ship fuels that LNG may reduce the NCAF/GCAF values considerably. For RCOs that are marginally above the required limit this could impact on the results. Page 26 of 70

31 It is evident that the fuel price path comes with a great deal of uncertainty. For RCO s where the fuel cost is a dominating factor in the total cost of the measure, this uncertainty may have large impact on the conclusion Cost of added built steel As many RCOs involve increasing a certain ship dimension there will be an increased construction cost related to this. In order to determine at this early stage the added cost related to this RCO a simple assessment for the added steel weight is applied and combined with the cost of constructing 1 tonne of steel. This cost includes raw material, steelwork and painting. This cost is representative for the construction of RoPax and Cruise vessels built at a European yard. Lower costs outside of Europe would reduce the costs of some of the RCOs and therefore higher safety levels could be established by ALARP process. The specific cost of added steel weight is assumed to be 6,000 USD/tonne Cost of increased installed power All RCOs are to be developed in such a manner that the vessel will maintain or increase its operational capabilities; this implies that the vessel s speed remains unchanged after implementation of an RCO. In case of increased weight and/or breadth there will be need for increased propulsion power. The specific investment cost of increased power is set to be constant within all ship sizes. The specific cost of added installed power is assumed to be 350 USD/kW Increased harbour fees As harbour fees are in most ports based on the Gross Tonnage (GT) of the vessel, an increase of the ships dimensions will also lead to increased GT and harbour fees. 3.5 Financial key figures for cruise Cost of increased outfitting area RCOs that increase the ships GT will lead to increased areas that need to be outfitted. These areas are not assumed to be used as cabins, but may be used for common passenger areas, while not expected to be able to generate any added revenue. If added value could be generated, some RCOs would be more cost effective. It is expected that the added areas that will need outfitting will be common passenger areas and the cost reflects this, areas not in need of outfitting (void spaces) should not have an added cost of outfitting. Cost per added m² of outfitted area assumed to be: 3450 USD/m² Cost of increased hotel load Along with the increased outfitting area the hotel/energy load will also increase. The hotel load is assumed to be linearly dependent on the accommodated volume. This increase in hotel load only applies to areas that will be illuminated and air conditioned. Increased hotel load per added GT is assumed to be: 0,056 kw/gt. Page 27 of 70

32 3.6 Financial key figures for RoPax Financial benefit of increased capacity When increasing the breadth of the ship this may, in some cases, lead to increased capacity that can be used for increased income. For RoPax vessels this is assumed to only be possible if an RCO can lead to additional lane meters. The income per lane meter is dependent on the selected route and will vary as a function of the frequency of departures and ship size. The following figures are used to determine the increased income per lane meter: Large RoPax 17,000 USD/lanemeter per year, based on the expected operations of the case vessel Medium RoPax 45,000 USD/lanemeter per year, based on the expected operations of the case vessel It should be noted that the above increased income has not been considered in all explored design alternatives/rcos and during the systematic optimisation studies, as the increased capacity will lead to increased income only if the market conditions allow for this. Where the increased income has been considered, this is clearly indicated. 3.7 Cost of Averting a Fatality (CAF) criterion The criterion used for recommendations based on NCAF and GCAF can be found in the consolidated version of the FSA Guidelines (MSC83/INF.2, page 54). The criterion that has been used for all FSAs submitted to IMO so far has been at $3million/fatality, see Table 2, page 54 of MSC83/INF.2. However, it is stated in the FSA Guidelines that the proposed values for NCAF and GCAF have been derived by considering societal indicators, (1)UNDP 1990, Lind 1996). They are provided for illustrative purposes only. The specific values selected as appropriate and used in an FSA study should be explicitly defined. This criterion is not static, but should be updated every year according to the average risk free rate of return (approximately 5%) or by use of the formula based on the Life Quality Index (LQI). It is noted that the $3million is in reality derived from 1998 statistics. If adjusted for US inflation rates until 2010, this figure should be updated to $4.14 million (2010). If adjusted for a 5% risk free rate of return the figure should be $5,39million (2010), and if a full update based on LQI is carried out the result is $7,45million. The LQI formula used is g e/4 (1-w)/w) g is gross domestic product per capita (Statistics from world bank used) e is life expectancy at birth (statistics from CIA fact-book used) w is the portion of life spent in economic production (statistics from OECD used) As for the $3 million (1998) figures, the derived criterion is an OECD average. The main changes are due to the following: The number of OECD countries has increased, Gross Domestic Product per Capita has increased, Life expectancy at birth has increased and we spend less time in economic activity (now about 1/10 rather than the 1/7 in the nineties). In addition the US$ has decreased its value against most other currencies. In this report the $7,45million criterion is used to indicate when an RCO is considered cost effective. Page 28 of 70

33 New building price in USD All references above can be found in MSC83/INF Summary of costs/revenue for RCOs applied on sample ship In the following, the results for each sample ship have been presented in tables. The attained GOALDS A-index is calculated based on the s-factor as described in the INF 2 paper. For reference the corresponding attained index obtained when calculated according to current SOLAS has been presented as well. The costs have been presented as capital costs, operational costs and fuel costs. The added revenue is in general the benefit resulting from the reduced probability from loss of the ship, and has been based on an net present value adjustment of the new building price of the ship taken from open source, namely the IHS database and presented in Figure 19 New building price (2010 USD) for RoPax and RoPax-Rail vessels plotted versus ship length (LOA) for IACS classified ships with a length equal to or greater than 80 m. Trend line for potential regression amendedand Figure 20. Figure 20 New building price (2010 USD) for RoPax and RoPax-Rail vessels plotted versus gross tonnage for IACS classified ships with a length equal to or greater than 80 m. Trend line for potential regression amended It is therefore also a function of the value of A, and as shown previously there are two alternatives for the A GR which would have influence on the revenue. The revenue obtained by increasing breadth to such an extent that additional lane meters can be obtained has been accounted for in the relevant cases of RoPax. For cruise ships no revenue based on additional income have been accounted for. It is concluded from investigations using the same source that the scrapping value (value after 30 years) of the cruise ship can be assumed as 15 % of the new building price and 12 % for the RoPax vessels. 1.2E E E+08 y = 76976x R 2 = E E E E E E E E E E+05 Gross tonnage Figure 19 New building price (2010 USD) for RoPax and RoPax-Rail vessels plotted versus ship length (LOA) for IACS classified ships with a length equal to or greater than 80 m. Trend line for potential regression amended Page 29 of 70