《DERBYSHIRE-LossofaBulkCarrier.doc》由会员分享，可在线阅读，更多相关《DERBYSHIRE-LossofaBulkCarrier.doc（11页珍藏版）》请在三一文库上搜索。
1、Case Study IIDERBYSHIRE - Loss of a Bulk Carrier-Vessel Particulars Summary of Structural Failure Background Detailed Description of Structural Failure (Proposed) End Result Acknowledgements -Vessel Particulars: Length: 281.94 MetersBeam: 44.20 MetersDepth: 24.90 MetersConstruction date: 1976Locatio
2、n: United KingdomClassification: Lloyds RegisterVessel Type: Oil/Bulk/Ore (OBO) CarrierArrangement: Accommodations and machinery aft. Nine bulk cargo holds separated by double skinned bulkheads.Arrangement and Construction Details:Last of six sister ships designed in 1969 in the UK and built from 19
3、70-76. The hull was double skinned and double bottom equipped. Each cargo hold was covered by 2 hatch covers (port and starboard). At the time of the loss, she had been in service only 4 years, and was considered a well-maintained ship, and was piloted by an experienced captain. -Summary of Structur
4、al Failure This paper discusses two possible explanations for the loss of the MV DERBYSHIRE during a severe storm, both of which postulate foundering due to structural failure of critical components. One scenario assumes the breakup of the ship due to fatigue failure of longitudinal structural membe
5、rs in the aft portion of the ship, while the other assumes foundering resulting from collapse of hold covers under sea loading. The two loss scenarios were compared in the light of the results of recent technical investigations and wreckage surveys. In the past decade this topic has produced intense
6、 debate in the maritime community, so the authors consider the topic worthy of review. The reader should bear in mind the complexity of ship systems and the highly irregular environment in which they operate, and realize that the loss of a ship is usually the result of many factors, including enviro
7、nmental, structural, and operational. -Background MV DERBYSHIRE was a British Oil/Bulk/Ore, (OBO), carrier transporting ore from Canada to Japan when she was lost during the typhoon ORCHID on the 9th or 10th of September 1980. She went down with all 44 on board without any distress signal. DERBYSHIR
8、E is the largest British bulk carrier ever lost and has been the object of several investigations and discussions regarding bulk carrier safety. This loss is just one in a long list of bulk carrier losses from the 1970s to the mid-1980s. Every year 10 to 20 bulk carrier losses occur, where structura
9、l failure might be the cause, (Ref. 2). The shared opinion among marine engineers is that this is an unacceptable number of losses, and that it is necessary to improve the design criteria of the classification societies.Because of the poor history of bulk carrier safety, a great deal of energy has b
10、een put into the investigation of the DERBYSHIRE casualty. It was hoped that if it could be recovered, the complete picture of the catastrophe would give valuable insight into potential structural weaknesses and inadequate ship design procedures.History of Investigation Since there initially was no
11、evidence that structural failure caused the loss, the UK Government did not hold a formal investigation into the loss the DERBYSHIRE. In 1982, eighteen months after the loss, her sister ship, TYNE BRIDGE, experienced severe brittle fractures. These cracks initiated at frame 65 and propagated into th
12、e deck. After this fracture, the DERBYSHIRE Family Association (DFA) started investigating frame 65 cracks on the other sister ships. They hypothesized that this might be the cause of the loss the DERBYSHIRE. In 1986 another sister ship, KOWLOON BRIDGE, broke at frame 65 after grounding. These event
13、s caused a closer look into the loss of the DERBYSHIRE.In 1994 the DFA had raised enough funds to conduct a survey of the wreckage. This first survey had the goal of locating the DERBYSHIRE and investigating the cause of the loss. The survey was completed in May 1994. The results of the survey were
14、inconclusive, but they located an object assumed to be the stern, separated from the rest of the wreck by approximately 600 meters.The survey led to further investigation. The new goal was to examine the loss of the DERBYSHIRE to learn more about structural weaknesses and to increase the safety of s
15、hips for the future . Thirteen loss scenarios were listed after an investigation of the service experience for the class and general casualty data for ships, specifically for bulk carriers, (Ref. 1). A lot of effort was put into the analysis of each loss scenario with the results that some were excl
16、uded while others were confirmed as more probable. This remainder of this paper discusses the two failure modes/loss scenarios that have been offered the most attention: Fatigue Failure at Frame 65 and Hatch Cover Collapse. -Detailed Description of Structual Failure (Proposed) Fatigue Failure at Fra
17、me 65Before the possible structural defects are discussed, the general arrangement for the region around frame 65 must be described. Frame 65 is the transverse watertight bulkhead that separates Hold 9 from the pump room. The longitudinal structure makes a drastic change at Frame 65 from a main long
18、itudinal girder to a longitudinal bulkhead separating the pump room from the slop tanks (see sketch of general arrangement below).The possibility of failure at frame 65 was realized after the KOWLOON BRIDGE ran aground and broke into three pieces. One break occurred at the bow where it was resting o
19、n the reef and the second near frame 65. Other sister ships began reporting visible cracking in the deck above frame 65. These incidents caused the industry to examine the structural design for weaknesses. The examination revealed that the method used to transition from the longitudinal girder to th
20、e longitudinal bulkhead could have resulted in the catastrophic failure of the MV DERBYSHIRE.FURNESS BRIDGE, the first ship of the class, used an accepted design method for the longitudinal transition at frame 65 by continuing the girder through Bulkhead 65 and then tapering and finally welding it t
21、o the longitudinal bulkhead in the pump room (the drawing below shows the general structural design).In the later ships, however, the design was changed to accommodate of coffer dams added to isolate the slop tanks from Hold 9 and the pump room. The new design ended the girder at Bulkhead 65 and sta
22、rted the longitudinal bulkhead on the aft side of the bulkhead (see drawing below for details).Ideally, the longitudinal girder and bulkhead would be in line at the transverse bulkhead, but if they are misaligned, the fatigue life of the structure could be reduced. As the ship cycles from hogging an
23、d sagging moments the misalignment causes the transverse bulkhead to be distorted fore and aft. The fluctuating distortions would result in high local stresses that could lead to cracking due to the shortening of fatigue life. The existing sister ships were found to have misalignments between the gi
24、rder and bulkhead. Of course, the alignment for the DERBYSHIRE is unknown. Although the magnitude of the misalignment is under debate, it could have been as much as 45 mm (see drawing below for a basic representation).Because of the high profile nature of the accident, eventually enough money was ga
25、thered to conduct a survey of the wreck. The survey revealed that the ship was in two main sections, but the stern section was only 600 meters from the forward section. Since the ship descended 4000 meters to reach the bottom, it seems unlikely that the break-up occurred at the surface. Therefore, f
26、ailure in the region of Bulkhead 65 was not the likely cause of the foundering of the DERBYSHIRE.Hatch Cover CollapseThe hatch cover design is another frequently discussed failure argument. In 1966 the International Convention of Load Lines, (ICLL), categorized a new freeboard class B-60, where the
27、freeboard requirements of the ordinary B class may be reduced by 60 cm. The only requirement is that the ship survive flooding of one compartment without loss of sea keeping. The new B-60 class resulted in a decrease of freeboard in the majority of newly built bulk carriers. The smaller freeboard ca
28、used an increase of wetted deck occurrences and pressure head experienced by deck plating and hatch covers. The design pressure was set to 1.75 tonne/m2 at the same conference, but it can be questioned if this pressure is high enough.Description of No 1 Hatch CoversThere are two covers on each hatch
29、 with length 14.68 m and beam 11.0 m (Cover No. 1)The stiffeners are T beams with 635 mm web an flanges of 280 mm x 25 mm. They are tapered at the end with a depth of 483 mm.The transverse center girder is 920 mm deep with a small 75 mm x 25 mm flange and the side girders have webs of 560 mm and fla
30、nges of 100 mm x 25 mm.All webs and plate thicknesses are 10.5 mm.The stresses in the hatch cover stiffeners met the requirements set by ICLL in 1966 according to Ref. 1. The requirements, taken from Ref. 3, page 3, are listed below:Hatch covers in the forward quarter-length (0.25L) to be designed f
31、or a uniform pressure of 1.75 tonne/m2.Level of stress not to exceed Ultimate Tensile Strength (UTS)/4.25. (which equals 0.40sigmay for mild steel)Limiting plate thickness for mild steel is b/100 or 6 mm, where b is spacing between stiffeners.To determine the maximum load a rule-designed hatch cover
32、 can carry, we set plastic collapse as the ultimate state. The T stiffener has a plastic shape factor, s, of 1.25. (The shape factor refers to how much the load can be exceeded from first yield to full yield of the cross section). Then the ratio of the full plastic collapse load and the design load
33、is:EQUATION: Ultimate stress / Design stress = 1.25sigmay/0.40sigmay = 1.25/0.40 = 3.125This means that we can increase the design pressure to a pressure of 3.125 (1.75 tonne/m2) = 5.47 tonne/m2 uniform pressure. This equals a water height of = 5.32 m over the hatch cover (Head = pressure/sea water
34、density = 5.47 tonne/m2 / 1.028 tonne/m3). The collapse head shows the true pressure a rule-designed cover can carry before collapse. But did the cover truly fulfill the requirements? Finite element analysis of the hatch cover design shows that they would collapse under a static pressure of about 4.
35、1 m (Ref. 1, pg 9) (hyperlink). This is above the ICLL requirement to resist a uniform pressure of 1.75 tonne/m2, but is still under the predicted collapse head of 5.3 m corresponding to the design pressure.Basic Calculation Procedure for a Stiffened Plate FieldThe stiffened plate of the hatch cover
36、 is modeled as a rigidly supported beam under uniformly distributed load. The beam length is taken as the stiffener span between the girders. The end moments, M, for a rigidly supported beam are given by:MOMENT EQUATION: M = pbL2/12where:p = pressure = 1.75 tonne/m2 = 17.17 kN/m2b = stiffener spacin
37、g = 0.994 mL = span of the member = 14.72 m/4 = 3.68 m.The longitudinals are assumed to be rigidly supported, because of symmetric loading on each side of the girder, and because the girder is usually much stiffer than the longitudinals.The design moment for the stiffeners with length 3.68 m is then
38、 M = 1.963 tonne-m = 19.56 kNm. This is the maximum load the member must carry according to the design pressure. This gives a section modulus, Z, for the stiffeners of: SECTION MODULUS EQUATION: Z = M/sigmawhere sigma is the acceptable stress level in the member. The stress was not to exceed UTS/4.2
39、5, which equals 0.40sigmay for mild steel, where sigmay is yield stress (This is commented as an irrationally low stress limit, but it still exists, 1998).Required section modulus is thus 19.56 kNm / 0.40(235 MPa) = 204 cm3. The actual section modulus of the T beam stiffener can be calculated or tak
40、en from a table (Z = Iz /ymax where Iz is the 2nd moment of inertia about the z-axis and ymax is the distance from the neutral axis to the furthest flange). We do not use the dimensions of the tapered end of the stiffener because the ends are approximated as simply supported because the loading is n
41、ot symmetric, and because the side plate at the end of the hatch cover is flexible. The most critical part of the stiffener is the end condition at the girder, so stiffener height at this location must be used in the calculations. The stiffeners used fulfill the requirements, according to Ref. 1. Se
42、e also calculations in the Appendix (MS Excel Spreadsheet). These calculations show that the required Z is about 204 cm3 (without corrosion allowance) and the actual Z is 1169 cm3, so the stiffeners seem to be over designed, if calculated in this manner.It can be questioned whether these calculation
43、s describe the structures capacity correctly, since the transverse girders are not as effective as assumed. One reason is that the depth of the side girders are smaller than the depth of the stiffeners, so in fact, the actual moment capacity of the girders is about 0.2MP, where MP is the plastic mom
44、ent capacity of the stiffeners. The narrow flange of the girders provides much less bending strength than necessary to restrain bending of the cover. This leads to significant bending of both the girders and the longitudinals, and therefore, compression at the top of the members and the thin welded
45、cover plating. The compression in this welded zone is substantial as bending increases toward collapse.This is not mentioned in the references, but as far as we can calculate, the side girders do not fulfill the section modulus requirements. According to Det Norske Veritas (DnV) rules (Part 3, Ch.1
46、Sec.8, D201) the section modulus of the girders should be Z = 8133 cm3 (without corrosion allowance), but our calculations show Z = 2330 cm3 for the side girders, and Z = 4251 cm3 for the center girder. This is dramatically lower than the required scantlings (see Appendix MS Excel Spreadsheet). The
47、requirements are based on the moment equation above, but with 10 in the denominator because the girders can not be considered to be rigidly supported, but rather a condition between simple and rigid support which yields a maximum moment greater than pbL2/12 but less than pbL2/8.Improved Design of No 1 Hatch Covers?Faulkner (Ref. 1) suggests an improved design of the hatch covers where the three girders are removed, and the l0 longitudinal stiffeners are replaced with 14 similarly