ANNEX IV RISK CONTROL OPTIONS

Transcription

1 ANNEX IV RISK CONTROL OPTIONS IV.1 INTRODUCTION The Risk Control Options (RCOs) are selected based on the recommendations resulting from the HAZID (ANNEX I), identified high-risk areas resulting from the risk assessment (ANNEX II) and measures identified in a meeting at Høvik The following RCOs are selected for further evaluation with respect to cost benefit: Sheltered mustering and life boat area Remote control of the ship from the mustering area Level alarms to monitor water ingress in all holds and forepeak Individual immersion suits to all personnel Free-fall lifeboat Free-fall lifeboat with an additional free float mode Marine Evacuation Systems for throw overboard liferafts Enclosing open lifeboats for all existing ships with open lifeboats Redundant trained personnel Improved pick-up function (crane) In order to evaluate the effect of the different RCOs, the probabilities of fatality and the PLL for the different accident events are presented in Table 1. Table 1 Probabilities of Fatality Associated with Evacuation PLL 1 [per ship year] Collision Contact Fire/Explosion Foundered Hull/Machinery failure Wrecked/Stranded IV.2 FREE-FALL LIFEBOATS IV.2.1 Free-fall lifeboat without free float mode Estimation of risk reduction See ANNEX III covering the risk analysis for free-fall lifeboats. The resulting probabilities of fatality associated with the RCO of replacing the conventional lifeboats with a free-fall lifeboat (without free float mode) are presented in Table 2. 1 PLL = Evacuation frequency [23.7 ] 16/01/2001 Annex IV, Page 1

2 Table 2 Probabilities of Fatality: RCO - Free-Fall Lifeboat (without free float mode) Collision Contact Fire/Explosion Foundered Hull/Machinery failure Wrecked/Stranded As seen from Table 2, a free-fall lifeboat is associated with smaller probability of fatality values compared to conventional lifeboats for all types of events except fire/explosion. The main reason for the increase in fatality probability associated with fire/explosion is the fact that 1 free-fall lifeboat has a greater unavailability compared to two conventional lifeboats in a fire/explosion. The cost of 1 free-fall lifeboat with the ramp and crane is not generally higher than for two conventional enclosed lifeboats with two davit systems. A packaged solution from some suppliers also includes the rescue boat. When collecting information from the yards the prises were equal. When collecting data from suppliers, data varied. This variation may be due to slightly varying solutions. The free-fall lifeboat has an impact on the liferaft solution. For most bulk carriers throw over board liferafts are preferred, because throw over board liferafts are more reasonable than davit launched liferafts. With a free-fall lifeboat, at least the liferafts on one side must be davit launched. The cost difference is estimated in Table 3 to 5,200 between the two solutions. Table 3: Cost Estimation Liferafts, Capacity 25 persons Cost element Type Davit Launched Throw Overboard Procurement Cost 3,000 2,500 Installation Cost Davit 4, h check, 6 h service, service equipment each year, 5% 5,600 5,100 depreciation of future costs (replace release mechanism) Training (half crew) /raft, 1h every 4 months, 5% 5,100 5,100 depreciation of future costs (Regulation 20, Paragraph 8.1) Total Life Cycle Cost, one raft 18,000 12,800 Two rafts on each bulk carrier 36,000 25,600 Cost difference by one davit launched instead of throw overboard, as 5,200 currently required for free-fall lifeboats For illustration of the decision problem it is therefore assume that the free-fall lifeboat is 7.5 % cheaper or 7.5 % more expensive than the conventional solution, or a low and high cost estimate ranging from - 13,000 to + 13,000. The corresponding Gross Cost of Averting a Fatality (GCAF) is ranging from - 356,000 to + 356,000. In addition to the likely situation of 16/01/2001 Annex IV, Page 2

3 both saving money and lives, free-fall lifeboat solutions also in the assumption of being 7.5 % (or up to at least 20%) more expensive, would also fall within the criteria of being recommended. Including the extra cost of the davit-launched liferaft that must be installed the cost estimate ranges from -7,800 to 18,200. The resulting GCAF is ranging from -214,000 to 499,000. Free-fall lifeboats would therefore still be recommended. It is also worth noting that the free-fall lifeboat solution has a crane that can be used for picking up the lifeboat after exercises and for lowering lifeboats in the case of e.g. grounding. This crane is often also used for loading supplies, and it may be used in the rescue phase to pick up lifeboats from other ships. The recommendation is limited to new ships. For existing ships, where retrofitting is necessary this solution is judged as not cost effective. IV.2.2 Free-fall lifeboat with free float mode Estimation of risk reduction See ANNEX III covering the risk analysis for free-fall lifeboats, with the following adjustment: Fatality due to untimely decision to launch in collision, grounding and foundering scenarios = 0, Otherwise no change The resulting probabilities of fatality associated with the RCO of replacing the conventional lifeboats with a free-fall lifeboat with free float mode are presented in Table 4. Table 4 Probabilities of Fatality: RCO - Free-Fall Lifeboat with free float mode Collision Contact Fire/Explosion Foundered Hull/Machinery failure Wrecked/Stranded As seen from Table 4 fitting the BC with a free-fall lifeboat with free float mode results in reduced probabilities of fatality for all types of events except fire, as compared to conventional lifeboats. The improvement as compared to free-fall lifeboats is due to the reduction in probability of untimely decision to muster. (Note that this effect is an effect more pronounced for bulk carriers than other ship types.) To equip a free-fall lifeboat with a hydrostatic release mechanism is not expensive, but requires that the release mechanism is thoroughly inspected and regularly replaced. Lacking a detailer specification of a technical solution, the assessment is based on an assessment from a manufacturer. This estimate was 5,000, corresponding to an increase in cost by less than 16/01/2001 Annex IV, Page 3

4 5%. To allow for inspection costs, maintenance etc., 5,000 is used as the lower estimate, and 7,000 as an upper bound. It is clear from the discussion of the free-fall lifeboat that such marginal cost increases would be defendable, with the significant improvement in life saving at a minor cost. Formally the GCAF is ranging from 551,000 to 772,000, which is within the criteria for recommending the solution. This is the GCAF for the free float mode of a free-fall lifeboat. It should be noted that this is a marginal improvement of the free-fall solution, and does not apply to other solution. This solution can not be recommended for free-fall lifeboats without making improvements to conventional lifeboats or making the requirement of free-fall lifeboats mandatory, as this may lead to decisions by owners to switch to conventional lifeboats, whilst the free-fall lifeboats are safer without this improvement. If free-fall lifeboats are made mandatory, a free-fall mode is also considered cost effective. The recommendation is limited to new ships. For existing ships, where retrofitting is necessary this solution is judged as not cost effective. The cost effectiveness (GCAF) of the free-fall lifeboat with and additional free float mechanism may also be estimated directly as an alternative to conventional lifeboats. Based on the consideration above the low and high cost estimate are - 2,800 and 23,200 respectively. The corresponding GCAF is ranging from - 61,000 to 509,000. IV.3 WATER LEVEL ALARMS Estimation of risk reduction IACS UR S24 requires water level alarms no later than the first intermediate or the first special survey, to be held after 1 January 1999 whichever comes first. (Water level alarms in all holds - no continuous water level indicator.) This study assumes a continuous water level indicator, as this is not adding much to the costs (the cabling and protection of the alarm is the main cost). The advantage of an indicator based on measuring hydrostatic pressure is that the master is actually going to understand that the scenario is developing. With an alarm (light/sound) the master may still be having the problem of deciding if the scenario is developing, or the alarm is false. To deny the development of a crisis is a natural human reaction. It is believed that continuous display of water ingress is a much better indication than an alarm. Reference is made to the HAZID (ANNEX I), which focused on these decision problems. The following adjustments are made to the model in order to reflect the effects of the RCO: For foundering events only - Probability of faulty evaluation of situation: Probability reduced by 50% As a result of earlier understanding the seriousness of the situation, the distress signal will be sent at an earlier time, which results in the following: Other vessel available: Increase in availability of 50% Helicopter available: No change, i.e. changes in availability is negligible, as the vessel most likely will be located far from shore. Lifeboat/Liferafts - Fatality associated with being rescued: 16/01/2001 Annex IV, Page 4

5 Early distress signal results in availability of other vessels to perform rescue to increase with 20%. Fatality as a result of not successfully rescued from sea: Early distress signal results in availability of other vessels to perform rescue to increase with 20%. The resulting probabilities of fatality associated with the RCO of installing water level alarms are presented in Table 5. Table 5 Probabilities of Fatality: RCO - Water Level Alarms Collision Contact Fire/Explosion Foundered Hull/Machinery failure Wrecked/Stranded As seen from Table 5, installing water level alarms result in a reduction in the probability of fatality from 55.4% to 40.6% for foundering events. The primary reason for the great reduction in fatality probability is the effect that the water level alarms have on the probability of faulty evaluation of the situation. have been estimated by yards that have been involved in installing the alarm according to UR S24. Prices were given for each alarm multiplied by number of alarms (holds) or as total costs. The costs are obviously a function of the size of the ship. The costs ranged from 14,000 to 21,000. These are costs for existing ships, and to assume similar costs for new ships is quite conservative. The GCAF for water level alarms, based on the assumptions described above is ranging from 146,000 to 220,000 and is recommended for both new and existing bulk carriers. IV.4 MARINE EVACUATION SYSTEM (IN CONNECTION WITH THROW OVERBOARD RAFTS ONLY) Estimation of Risk Reduction The following adjustments are made to the model in order to reflect the effects of the RCO: For all events except foundering: 16/01/2001 Annex IV, Page 5

6 Fatality due to jumping over board: Fatality probability reduced by 50% Probability of unsuccessful boarding: Probability reduced by 90% For foundering: Fatality due to jumping over board: Fatality probability reduced by 10% (Foundering is associated with heavy weather in which the effect of the marine evacuation system is reduced.) Probability of unsuccessful boarding liferaft: Probability reduced by 90% The resulting probabilities of fatality associated with the RCO of installing a marine evacuation system are presented in Table 6. Table 6 Probabilities of Fatality: RCO - Marine Evacuation System Collision Contact Fire/Explosion Foundered Hull/Machinery failure Wrecked/Stranded The probability of fatalities due to jumping over board and the probability of unsuccessful boarding both are reduced significantly by marine evacuation systems. As seen from Table 6, the system results in overall reductions in probability of fatality for all types of events. The reduction is small, as this is a marginal improvement of a life saving appliance with limited effect. No detailed cost estimation has been carried out, as the manufacturers contacted were not willing provide cost estimates. The likely reason for this is that some manufacturers are preparing marine evacuation systems that could be used for e.g. bulk carriers. Current systems are much too large, and have been designed for passenger ships. Applying the decision criteria of 1 million as suggested in MSC 72/16 the assessment indicates that for marine evacuation system would be recommended as cost effective if the life cycle cost of the marine evacuation system less the life cycle costs of throw overboard liferafts do not exceed 4, /01/2001 Annex IV, Page 6

7 IV.5 REMOTE OPERATION OF SHIP FROM MUSTER AREA Estimation of Risk Reduction The following adjustment is made to the model in order to reflect the effects of the RCO: Probability of untimely decision to muster: Probability reduced by 20% The resulting probabilities of fatality associated with the RCO of installing remote operation of the ship from the muster area are presented in Table 7. Table 7 Probabilities of Fatality: RCO - Remote Operation of Ship from Muster Area Collision Contact Fire/Explosion Foundered Hull/Machinery failure Wrecked/Stranded As seen from Table 7, the possibility of operating the ship from the muster area results in reductions in probability of fatality for all types of events due to the reduced probability of untimely decision to muster. The effect is small. The cost estimate is an educated guess from a bridge system manufacturer. Equipment for controlling only steering and propulsion was estimated to be cost between 10,000 and 20,000. These costs would just include the equipment, protection of the equipment, and equipment installation costs. In addition there would be inspection/testing and maintenance costs. The cost estimates may therefore be somewhat optimistic. The resulting GCAF is ranging from 2.1 million to 4.2 million. Even with optimistic cost estimates this is above the criteria and can not be recommended. IV.6 SAFER ACCESS TO LIFEBOATS Estimation of Risk Reduction The RCO is assumed to represent protection against fires (i.e. heat radiation and smoke) and wind/water (i.e. heavy weather). The following adjustments are made to the model in order to reflect the effects of the RCO: 16/01/2001 Annex IV, Page 7

8 Fatality due to unsuccessful boarding: Resulting fatality rate is negligible in moderate weather and (i.e. reduced by 50%) in severe weather. The resulting probabilities of fatality associated with the RCO of designing the access to the lifeboats more protected against fire and the environment are presented in Table 8. Table 8 Probabilities of Fatality: RCO - Safer Access to Lifeboats Collision Contact Fire/Explosion Foundered Hull/Machinery failure Wrecked/Stranded As seen from Table 8, safer access to lifeboats results in a small reduction in probability of fatality for all types of events. The cost estimate is based on enclosing the mustering area with a wall and a deck above the lifeboat, extending the full length of the lifeboat, and with a wall in front of the muster area/lifeboat. A Panamax Bulk Carrier was used to arrive at correct scantling. The vertical extent was 278 mm, the length 800 mm and the breath 470 mm. The total weight of the plating, girders and stiffeners for the deck and walls enclosing the mustering area was estimated to 15,800 kg/14.22 metric ton steel. The table below reports the detail of the cost estimates. The basic unit cost was estimated as: 400 US $ per metric ton of steel 140 US $ of consumables per metric ton of steel 190 US $ of coating per metric ton of steel 19 US $ per man-hour. The man-hours time was estimated as 6-8 days. Specific Panamax Estimate Material Cost (US $) Steel plates, stiffeners, girders 5,688 (14.22 metric tons) Welding oxygen and electrodes 1,991 Man-cost (56 h) 1,064 Cleaning 600 Coating 2,702 Grand Total 12,045 The cost will not vary much as a function of ship size but may vary to a large extent as a function of detail location of lifeboats and general design of the Bulk Carrier. Converting to and allowing for a ± 20% variation, 6,700 and 9,600 are used as lower and upper bounds in the cost effectiveness assessment. 16/01/2001 Annex IV, Page 8

9 The project also received an independent estimation from a Korean yard. This estimate gave a grand total of $ 10,000. This is almost identical to the low cost estimate. The resulting GCAF is ranging from 733,000 to 1,050,000. This is within the acceptance criteria and could be recommended. The recommendation is sensitive to assumptions in the risk estimation. Further, if implementing other RCOs reduces risks, this requirement may fall outside the criteria in MSC 72/16. IV.7 IMMERSION SUITS (FITTED TO EACH INDIVIDUAL) Estimation of Risk Reduction The survival time for personnel in immersion suits is more than 12 hours in water temperatures above 5 o C, see Table 9. In normal clothing this survival time is less than 1 hour. This make an significant difference, as the likelihood that other ships come to assistant increases from unlikely to likely in this period. It may therefore be assumed that the chances of being rescued could increases from 0 to 0.8 by use of immersion suits. In the calculation 0.5 is assumed in order not to be too optimistic. A statistical basis for this is the total loss accidents of 184 Norwegian fishing vessels in the period Of the 293 persons involved in these accidents, in this harsh environment only 3.4% lost their lives. Time to rescue exceeded 6 hours in several events, even in extremely low temperature. The full report is in a Norwegian submission to DE 44 (DE 44/1 and DE 44/INF.7). Thermal Protective Means. (Clothing is generally included) Table 9 General thermal protection performance standards For personal life-saving appliances. Clothing Thermal Protective (TP) Lifejacket Thermal Protective Aid (TPA) (Wet Suit) Anti-Exposure Suit (Wet Suit) Immersion Suit Uninsulated (Dry Suit) Immersion Suit Insulated (Dry Suit) (Table from DE 44/8) Level of Survival time (hrs) when exposed to seawater of temperature 20 o C (Sector 1) 1.6h 3) 4.0h 2) 6.0h 2) 10h 2) >12h 2) >12h 2) 1) Survival time according to IMO minimum requirements. 2) Survival time calculated. 3) Survival time laboratory test data. 10 o C (Sector 2) 0.8h 3) 2.0h 1) 3.0h 2) 4.0h 3) 5.0h 2) >12h 2) 5 o C (Sector 3) 0.5h 3) 0.75h 2) 2.0h 3) 2.0h 1) 1.0h 1) 12h 2) 0 o C (Sector 4) 0.3h 3) 0.5h 2) 0.75h 2) 1.5h 2) 1.5h 3) 6.0h 1) A more detailed assessment may be carried out by observing that many bulk carrier accidents occur in the regions North Atlantic, Pacific Ocean (east of Japan), and Indian Ocean (east of South Africa), in bad weather. These correlates with low sea surface temperatures, see for the details. The following adjustments are made to the model in order to reflect the effects of the RCO: Fatality probability associated with awaiting rescue: Probability reduced by 50% (The reason why it does not reduce the probability even more is that it is likely that the time until rescue exceeds the time that a person survives even in an immersion suit, in particular as it may be difficult to find individuals in open sea.) 16/01/2001 Annex IV, Page 9

10 Reduction for all situations in which persons are awaiting rescue, except for cases in which the person has jumped to sea as a result of faulty evaluation of situation, untimely decision to muster and unable to reach mustering station: In these cases the persons are assumed not to be wearing immersion suits (due to lack of time before jumping to sea) Lifeboat - Fatality associated with being at sea: Reduced by 5% for closed lifeboats Reduced by 10% for open lifeboats Liferaft - Fatality associated with being at sea: Reduced by 30% The resulting probabilities of fatality associated with the RCO of requiring personnel to wear individually fitted immersion suits in evacuations are presented in Table 10. Table 10 Probabilities of Fatality: RCO Immersion Suits Collision Contact Fire/Explosion Foundered Hull/Machinery failure Wrecked/Stranded As seen from Table 10, individually fitted immersion suits result in reductions in probability of fatality for all types of events. Immersion suits are produced in a variety of materials, with varying donning time, varying insulation, different strapping, different sealing, etc. The variation in costs may to some extent reflect variation in the improved chances of survival. A cost estimation of the life cycle cost of a high quality immersion suit is given in Table 11. Table 11: Cost estimation Immersion suit Procurement Cost 400 Service and Change of Batteries (Every 5 years). Pressure Test/Visual 135 Inspection/Repair of Minor Damages (Life Cycle Cost 5% depreciation) Annual Inspection and training (1h/suit/year) 183 Sum 718 Assume that 15% of the suits need to be replaced 2 during the life time of the 60 bulk carrier Total life cycle cost one suit 778 Assume 23 crew 17,800 2 The estimation is difficult, since little experience exists from using immersion suits for a full life cycle of a ship. Information has been received from a survey of 1000 inspections carried out in 1998 of immersion suits on board Icelandic vessels. Regulation no 179/1987 requires that all Icelandic vessels 12 m (LOA) or longer should be equipped with immersion suits for the entire crew. The survey showed that about 15% had some minor leakage. To assume 15% replacement is therefore very conservative. (The Icelandic administration indicated that a full report of their experience may be submitted to IMO) 16/01/2001 Annex IV, Page 10

11 All personnel should verify that they have suits that fit, and should also test a suit. Information about the suits is available by searching the web. Some alternative cost estimates are also available, see e.g. As the cost estimation is based on the best available suits and somewhat conservative, the costs are used as an upper bound. Allowing for the likely availability of more reasonable suits a lower bound of 15,000 is assumed. The assumptions above suggest a GCAF ranging from 481,000 to 571,000. Immersion suits to all personnel are therefore clearly within the acceptance criteria, and also more cost effective than the conventional lifeboats. The lifetime of the suit does not have to be related to the lifetime of the ship, and the recommendation therefore also applies to existing ships. In this respect the suit should belong to a person rather than the ship 3. IV.8 ENCLOSED LIFEBOATS FOR ALL EXISTING SHIPS Estimation of Risk Reduction After the requirement of enclosed lifeboats was enforced , some manufacturers enclosed the open boats by prefabricated elements. This may influence the capacity of the lifeboats. Manning has been reduced since 1986 and it would not normally be expected that enclosing a lifeboat had an effect that reduced the capacity below the required capacity. The following adjustment is made to the model in order to reflect the effects of the RCO: Lifeboat - Fatality associated with being at sea: The probabilities of fatality associated with being at sea are assumed to be 4% and 6% for enclosed and open lifeboats, respectively. In the model, it is assumed that 50% of the vessels are fitted with enclosed lifeboats and 50% of the vessels are fitted with open lifeboats, which results in an average probability of fatality of 5%. In order to evaluate the difference between fitting a ship with enclosed and open lifeboats, the different results obtained by using a value of 4% and a value of 6% is evaluated. Table 12 Probabilities of Fatality: RCO Closed Lifeboats for all Existing Ships Collision -0.7 Contact -0.7 Fire/Explosion -0.7 Foundered -0.6 Hull/Machinery failure -1.1 Wrecked/Stranded For example, in the Collective Bargaining Agreement (CBA) between The Norwegian Seamen s Unions and Norwegian Shipowners Association for Shuttle Tankers ( 07) it has been agreed that a seamen on ships in North-European waters should receive a personal immersion suit upon employment. The seaman is responsible for maintenance and cleaning, but documented costs of maintenance are refunded. 16/01/2001 Annex IV, Page 11

12 As seen from Table 12, enclosed lifeboats result in reductions in probability of fatality for all types of events. The effect is significant. Cost estimates were received from one manufacturer with previous experience with closing open lifeboats by prefabricated elements. The manufacturers estimate was 4,500. Allowing for uncertainty the cost is assessed to range from 4,000 to 6,000. The effect of this requirement could benefit the fleet of bulk carriers built before 1986 for about 5 years. The risks of fatalities associated with foundering accidents is increasing a factor of three (Actual: 3.15) above the average during these last years in the bulk carrier s lifetime, see Figure 1. By assuming that other accident scenarios are associated with time independent fatality rates, the result is a factor of two (Actual: 2.3). These assumptions result in a GCAF ranging from 906,000 to 1.36 million. This RCO is therefore in an area where implementation may be recommended or not dependent on if the upper or lower bound is used in the assessment. The recommendation may be sensitive to the accuracy of the probability estimates, as the RCO is associated with small costs and small effects. The estimation also make it clear that replacing existing open lifeboats with new enclosed lifeboats is not cost effective. Average PLL 0,050 0,040 PLL 0,030 0,020 Average PLL 0,010 0, Ship Age Figure 1: Potential Loss of Life as a function of ship ages (historical data) IV.9 REDUNDANT TRAINED PERSONNEL Estimation of Risk Reduction The following adjustment is made to the model in order to reflect the effects of the RCO: Lifeboat/Liferaft - Probability of unsuccessful preparation of equipment: The human error is in the model multiplied by a factor of 5 due to extremely high stress level and step-by-step task, as described in Handbook of Human Reliability Analysis with Emphasis on Nuclear Power Plant Applications. 16/01/2001 Annex IV, Page 12

13 The value of 5 is for skilled persons, while for a novice (less than 6 months experience) a value of 10 should be utilised as described in (Swain and Guttmann, 1987). The effect of requiring redundant trained personnel is evaluated by obtaining the overall difference between using a multiplier value of 5 and a value of 10. The resulting probabilities of fatality associated with the RCO of requiring redundant trained personnel are presented in Table 13. Table 13 Probabilities of Fatality: RCO Redundant Trained Personnel Collision -0.2 Contact -0.2 Fire/Explosion -0.2 Foundered -0.4 Hull/Machinery failure -0.4 Wrecked/Stranded -0.3 As seen from Table 13, redundant trained personnel results in reductions in probability of fatality for all types of events. The reduction is small, due to the small probability of unavailability of trained personnel under the current requirement in STCW. The various training centres provide training packages for various personnel as required or as adequate interpretations of STCW, Section A-VI/2 (Tables 1 & 2). The costs are varying to a large extent. A typical cost estimate in Europe could be as in Table 14. Table 14: Assumptions in cost estimation Cost Element Cost Cost of travel and 40 subsistence per day Cost of travel per person 250 per course Cost per work hour 40 Assuming that two persons have to go to one combined lifeboat and fast rescue boat course, this would take 5 days (some courses are longer). The repetition course, every 5 years, would take 3 days. The costs would add to: Course: ( 40 x x 40 ) x 2 = 4,100 Repetition Course: ( 40 x x 24 ) x 2 = 2,660 The total cost, depreciating with a 5% real risk free rate of return would result in a total cost of: 2,660 2,660 2,660 Total cos t : 4, = 9, /01/2001 Annex IV, Page 13

14 The assumption of a low and high cost estimate is therefore 8,000 and 10,000, receptively. It should be noted that considerably lower costs are possible for crew from the Far East. The cost effectiveness expressed as GCAF thus ranges from 2.05 to 2.56 million. The cost effectiveness may be much more favourable if cost estimates were based on crew cost from the Far East. IV.10 IMPROVED PICK-UP FUNCTION Estimation of Risk Reduction The following adjustment is made to the model in order to reflect the effects of the RCO: Lifeboat/Liferaft - Fatality associated with being rescued: Probability reduced by 50% The resulting probabilities of fatality associated with the RCO of installing/modifying a crane for easier pick up lifeboats and rafts are presented in Table 15. Table 15 Probabilities of Fatality: RCO - Improved Pick-Up Function Collision Contact Fire/Explosion Foundered Hull/Machinery failure Wrecked/Stranded As seen from Table 15, the changes in the probabilities of fatality are minor compared to not having such a system installed for all types of events except foundering. Note that the assumption is that the ship involved in the rescue is equipped with a pick up facility. In reality this would correspond to implementation on all ships, and not only Bulk Carriers. If only implemented on Bulk Carriers there would be some probability < 1 that the assisting ship was a Bulk Carrier. As indicated under free-fall lifeboats, the ships with free-fall lifeboats already have a possibility for picking up a fully loaded lifeboat from the sea. For a Bulk Carrier with conventional lifeboats there is no crane available (at least not as a result of safety regulations) that have the capacity to pick up a fully loaded lifeboat from the sea. As an indication of the costs, the cost of the crane on bulk carrier with free-fall lifeboats is used in the calculation. The costs from two suppliers of such cranes indicated a cost between 40,000 and 60,000. There would also be some maintenance and installation costs. 16/01/2001 Annex IV, Page 14

15 The cost estimate will result in a GCAF ranging from 12 million and 24 million. As this is based on optimistic assumptions on improved safety and optimistic assumptions of costs it is clear that this solution can not be recommended. However, in a safety equivalency consideration it should be noted that this function exists on ships with free-fall lifeboats including other ship types than Bulk Carriers. 16/01/2001 Annex IV, Page 15