Page 586 - Marine Structural Design
P. 586

5 62                                                      Part V Risk Assessment


                Romer et a1 (1993) presented a risk assessment of marine transport of dangerous goods based
                on historical data that consist of 15 1 accidents in the period of 1986 to 1991. Their paper gave
                frequencies for various kinds of accidents, FN curves and frequencies and size of spills.
                Loading Errors
                Improperly loaded cargo may adversely affect the ship’s stability, as well as put undue strain
                on the hull and subsequently increase the failure probability at sea. In rare cases, a vessel can
                sink due to improper loading.
                Vessels such as ore carriers are susceptible to payload shifts during periods of rough seas.
                While  for  container  ships,  container  lashings  can  become  loosened  or  broken,  causing
                containers to shift or move freely. This will not only jeopardize ship stability, but also pose a
                threat to personnel, machinery, or the hull.  There is also an economic intention to fill the
                vessel to its maximum capacity, especially for a short time period. However, some overloaded
                fishing vessels have foundered due to this reason.
                Extreme Environmental Condition
                Many  ships  were  lost  at  sea  during  the  extreme  weather.  The  ocean  is  a  hazardous
                environment for both man and ships.
                The crew is likely to lose their lives overboard if they go on deck during a sea storm. Secured
                components may also be broken and cause damage to the equipment. This typically happened
                for BRAER tanker, where pipes stored on the deck broke free and damaged the vents serving
                the diesel fuel tank during a storm. This led the diesel fuel to be contaminated with seawater
                and finally, the power in the vessel was lost. The vessel then drifted ashore under a prevailing
                current, leading to a significant oil spill. The heavy weather may  also lead to tiredness and
                seasickness of the crew; this therefore increased the likelihood of operational errors. Rough
                seas may pose other hazards as well, e.g. loss of visibility, position, or communications, which
                potentially increase the risk of grounding and colliding.

                31.3.3  Frequency Analysis of Ship Accidents
                The risk associated with an event is a function of two quantities: the likelihood of the event
                and the consequence from that event. Therefore, the fiequency analysis forms an essential part
                in the risk estimation. CCPS (1995)  listed ocean-going vessel failure modes (e.g.  collision,
                grounding,  fire  and  explosion  and  materiaVequipment  failure),  discussed  parameters
                influencing accident rates  and  hazards release probabilities, and  suggested procedures for
                failure (release) frequency calculations.
                For ship accidents, sometime sufficient historic data are available for some critical events. One
                of the best sources of data is the U.S.  Coat Guard’s vessel casualty file for U.S. waterways, the
                Marine Casualty Information Reporting Systems (USCG, 1992). The fiequency can then be
                derived simply hm the past data. However in most cases, a frequency analysis may not be so
                straightforward and it must examine the contributing factors that lead to the actual accident. In
                this process, it is necessary to break the compound event into individual events and put them
                together in  a  logic sequence to  model  how  the  hazards are developing into accidents via
                different failure paths.
                After  a  synthesis  of  individual  events  according  to  certain  scenarios,  the  occurrence
                probability of the accident may be quantified by using the fault-tree technique and the event-
                tree technique. Normally, the fault tree is used to explore the causes of a critical event, while
   581   582   583   584   585   586   587   588   589   590   591