Click for next page ( 71


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 70
4 APPLICATION OF THE METHODOLOGY T he methodology developed by the committee was applied to evaluate two 150,000deadweight ton (DWT) tanker designs-- one single- and one double-hull--and two 40,000-DWT tankers--one single- and one double-hull. The application involved structural damage and outflow calculations for each tanker, oil-spill fate and transport sim- ulations in four geographic locations, combining of the outflow and spill simulation results into a consequence measure, and design comparisons. The application and its results are described in this chapter: the selection of vessels and collision and grounding scenarios, the collision and grounding analyses and their results, the selection of hypothetical spill scenarios and consequence measures, limitations of the consequence analysis, and finally the design comparison. It is important to keep in mind that the work pre- sented is only an illustration of the use of the methodology. As discussed in Chapter 3, further refinement and testing of the methodology are rec- ommended, particularly if the alternative design in question includes in- novative design features. SELECTION OFVESSELS AND COLLISION AND GROUNDING SCENARIOS The selection of vessels for use in testing the methodology was based on the following criteria: The vessels to be compared had to have the same cargo capacity. Detailed structural information about the vessels had to be available (to demonstrate how the methodology models structural resistance during accident scenarios). Vessel drawings and other relevant information had to be available to the committee. 70

OCR for page 70
APPLICATION OF THE METHODOLOGY 71 Because of the above requirements, alternative designs lacking sufficient design detail had to be excluded from consideration. The pro- files and midship sections of the 150,000-DWT and 40,000-DWT vessels included in the study and their subdivisions into cargo and ballast spaces are shown in Figures 4-1 and 4-2, respectively. It is important to note that the double-hull ships have a deeper draft than the single-hull ships, which is typical for existing single- and double-hull ships of the same capacity. The draft and subdivision of the reference double-hull vessels are important for the outflow results and must be carefully evaluated when selecting the standard double-hull ships. The programs SIMCOL and DAMAGE were selected for the col- lision and grounding analyses, respectively. These programs use a sim- plified approach to calculate resistance and therefore are suitable for analyzing a large number of collision and grounding events. Oil outflow from damaged cargo tanks is a part of the SIMCOL output, whereas DAMAGE provides information on structural damage only. A program was written to allow batch runs of a large number of grounding cases and to add an outflow calculation based on the structural damage output from DAMAGE. These damage modeling programs were used in the com- mittee's illustration of the methodology, but other programs could be in- corporated into the methodology as well. As discussed in Chapter 3, SIMCOL and DAMAGE each use a sim- plified approach to analyze collision and grounding damage, and they include assumptions that limit their application. The main limitations of SIMCOL are as follows: The bow of the striking vessel is assumed to be rigid. The energy absorbed by the deformation of the striking bow is neglected. Analytical models in SIMCOL are based on empirical data, and they may not be applicable to analysis of innovative structural designs. SIMCOL models collisions between two vessels. Collisions with solid objects, such as bridges or docks, cannot be modeled. SIMCOL has limitations in modeling raking damage. The main limitations of DAMAGE are as follows: The structural model includes only the cargo block. The effect of the bow and stern on structural behavior in the damage region is neglected. The model is built with conventional structural members, and the ma- terial used is limited to that which can be described with the stressstrain

OCR for page 70
41.6 m 41.6 m 41.6 m 41.6 m 41.6 m 14.8 m 31.2 m 31.2 m 31.2 m 31.2 m 31.2 m 31.2 m 41.6 m 41.6 m 41.6 m 41.6 m 41.6 m 20 m 29 m 20 m 14.8 m 31.2 m 31.2 m 31.2 m 31.2 m 31.2 m 31.2 m 20 m 20 m Cargo VCG = 2.42 m Cargo VCG = 3.78 m below WL below WL All ballast tanks are Note that DB hull has of `J' type DB height = 3.34 m .34-m greater draft FIGURE 4-1 Profile, plan, and midship section for 150,000-DWT ships (VCG vertical location of the center of gravity; WL waterline).

OCR for page 70
23.3 m 23.3 m 23.3 m 23.3 m 23.3 m 23.3 m 17.5 m 17.5 m 17.5 m 17.5 m 17.5 m 17.5 m 17.5 m 17.5 m 23.3 m 23.3 m 23.3 m 23.3 m 23.3 m 23.3 m 8.1 m 11.1 m 8.1 m 17.5 m 17.5 m 17.5 m 17.5 m 17.5 m 17.5 m 17.5 m 17.5 m 12.6 m 12.6 m Cargo VCG = 2.48 m Cargo VCG = 3.3 m DS width = 2.44 m below WL below WL All ballast tanks are Note that DB hull has of 'L' type DB height = 2.1m .57-m greater draft FIGURE 4-2 Profile, plan, and midship section for 40,000-DWT ships ( VCG vertical location of the center of gravity; WL waterline).

OCR for page 70
ENVIRONMENTAL PERFORMANCE OF TANKER DESIGNS 74 curve and the assumed failure modes. Innovative structural designs using new materials would require extensions to the current program. DAMAGE is limited to modeling powered head-on grounding on a single pinnacle. Other types of obstruction or grounding (e.g., grounding on a reef or soft soil, drift grounding) currently cannot be modeled in DAMAGE. A more detailed description of the programs and their validation can be found in the Alternative Tanker Designs Collision Analysis and Alternative Tanker Designs Grounding Analysis reports included on the accompa- nying CD. The collision and grounding incidents to be used were defined by accident factor distributions, which were selected to sample conditions in U.S. waters with a high density of tanker traffic. Data were collected for the selected hypothetical spill locations, and it was assumed that the vessels had an equal likelihood of being in each of the geographic loca- tions. Available data were inadequate for many of the variables, and further refinement of the accident-factor distributions is recommended. Four hypothetical spill locations were used in the application (their selection is discussed later in this chapter): Big Stone Anchorage, Delaware Bay; Galveston lightering area, Gulf of Mexico; Carquinez Strait Bridge, San Francisco Bay; and Farallon Islands, offshore San Francisco. Accident factors for grounding were collected from the following sources: The speed distribution was based on information received from pilots and operating personnel on typical speeds in the Galveston ligh- tering area, in San Francisco Bay near Carquinez Strait Bridge, and outside of San Francisco Bay. No speed data were available for Delaware, but speeds were assumed to be similar to those in San Francisco Bay. The tidal distribution was based on information on the four loca- tions obtained from NOAA (tidesonline.nos.noaa.gov). Obstruction depths were based on data for Galveston, Delaware, and San Francisco Bay. These data were received from USCG vessel traffic service (VTS) centers and NOAA charts. No data were available on the shape of the obstructions in these locations.

OCR for page 70
APPLICATION OF THE METHODOLOGY 75 The distribution for the obstruction tip radius was taken from work by Rawson et al. (1998). The minimum and maximum apex angles were based on the limitations of the theory in DAMAGE, and a uniform distri- bution was assumed. Inert gas pressure1 distribution was based on data for a typical range of inert gas pressures provided by INTERTANKO. The capture2 distribution was selected on the basis of model test re- sults presented in an IMO (1992) comparative study. The accident factors and their distributions for grounding are given in Table 4-1. Brown provided the committee with accident factors for collision scenarios. His primary data sources were a report by Sandia National Lab- oratories (1998) and 1993 Lloyd's Worldwide Ship Data provided by the U.S. Maritime Administration. A discussion of the data can be found in the Alternative Tanker Designs Collision Analysis report on the accom- panying CD. Table 4-2 presents the accident factors and their distribu- tions used in the collision analysis. COLLISION AND GROUNDING ANALYSES The accident factors described above were sampled using Monte Carlo simulation to generate 10,000 collision and grounding events. The struc- tural damage and oil outflow were analyzed for each vessel using the same 20,000 scenarios. This allowed direct design comparisons to be made. The collision calculations were contracted to Brown, the de- veloper of the SIMCOL program. The grounding calculations were ob- tained from an ongoing Society of Naval Architects and Marine Engineers/ Ship Structures Committee (SNAME/SSC) project on Prediction of Struc- tural Response in Grounding. The ships were assumed to be in a fully loaded condition, ad- justed so that each vessel carried the same quantity of cargo and main- tained an even-keel condition. The cargo was crude oil with a density of 0.84 grams per cubic centimeter (g/cm3), which corresponded to one of the oil types used in the oil fate and transport simulations.3 The param- 1Inert gas pressure is maintained in cargo tanks to avoid explosive conditions. 2Capture refers to the amount of oil that is captured in ballast spaces adjacent to damaged cargo tanks. 3The density of the other oil type (North Cape No. 2 fuel oil) is 0.86. Since the densities of the two oil types are similar, the outflow results will be close as well.

OCR for page 70
ENVIRONMENTAL PERFORMANCE OF TANKER DESIGNS 76 TABLE 4-1 Grounding Accident Factors and Their Distributions Factor Minimum Maximum Distribution Speed (knots) 0 20 Probability that the speed is in the range 0 to 5 knots--25% 5 to 8 knots--45% 8 to 15 knots--8% 15 to 16 knots--20% 16 to 20 knots--2% Obstruction depth 0 19 Probability of depth ranges from mean low 0 to 5 m--11% water (meters) 5 to 10 m--28% 10 to 15 m--31% Larger than 15 m--30% Obstruction apex 15 50 Truncated normal distribution. Strong angle (degrees) positive correlation with tip radius (0.80). Large apex angles correspond to large tip radii Obstruction tip 0 10 Truncated normal distribution radius (meters) Nondimensional 0 1 Uniform distribution rock eccentricity, e/(beam/2), from centerline Tidal variation 0 2.5 Probability that the ride is in the range (meters from mean 0 to 0.7 m--50% low water) 0.7 to 1.7 m--35% Greater than 1.7 m--15% Inert tank pressure 400 1000a Uniform distribution (millimeters water gauge) Capture in ballast 0 50 Uniform distribution tanks (as percent of tank volume) Minimum outflow 0.5 1.5 Uniform distribution. Moderate positive (as percent of correlation with speed (0.50). Higher ruptured tank speeds are more likely to have higher volume) minimum outflow aPressure valves preset at 1500 mm water gauge (WG), but to represent industry practice and allow the use of a uniform distribution, the range of 400 to 1000 mm WG was applied. eters defining the vessel condition and the liquids carried are shown in Table 4-3. Both the collision and grounding calculations included several simplifying assumptions. These assumptions are discussed in detail in the reports on the accompanying CD; the main assumptions are sum- marized below.

OCR for page 70
APPLICATION OF THE METHODOLOGY 77 TABLE 4-2 Collision Accident Factors and Their Distributions Factor Minimum Maximum Distribution Ship type N/A N/A Distribution of ship type Tankers--25.2% Bulk cargo--17.6% Freighters--42.4% Passenger--1.4% Container--13.5% Probability by ship type Displacement Tankers--700 Tankers--274,000 Tankers--Weibull (0.84, 11.2) (metric tons) Bulk--1,800 Bulk--130,000 Bulk cargo--Weibull (1.2, 21.0) Freighters--500 Freighters--42,000 Freighters--Weibull (2.0, 11.0) Passenger--1,000 Passenger--76,000 Passenger--Weibull (0.92, 12.0) Container--1,100 Container--59,000 Container--Weibull (0.67, 15.0) Speed of 0 20 Distribution based on historical striking ship data, approximately (knots) Weibull (2.2, 6.5) Collision angle 0 180 Distribution based on historical (degrees) data, approximately truncated normal (90.0, 29.0) Strike location 0 1 Beta (1.25, 1.45) relative position from bow Speed of 0 20 Distribution based on historical struck ship data, approximately (knots) exponential (0.584) NOTE: N/A = not available. The collision program SIMCOL solves the external ship dynamics and the internal deformation mechanics simultaneously in the time domain. The external dynamics model in SIMCOL takes into account the yaw mo- tions of the striking and struck vessels, as well as the relative horizontal rotation between the two vessels. Principal dimensions and displacements define the ships. The striking ship's bow is assumed to be wedge-shaped, and the deformation is considered only for the struck ship. The internal mechanics model uses analytical modeling to de- termine reaction forces for vertical members, but the reaction forces and absorbed energy for horizontal members are based on empirical results.4 The structural model of the struck ship includes vertical and horizontal 4Horizontal members refer to those structural members for which the collision force is applied in the plane of the structure (e.g., the crushing of a deck), and vertical members refers to those structural members for which the collision force is applied normal to the plane of the structure (e.g., side shell) or across the axis of a member (e.g., side-shell stiffeners).

OCR for page 70
ENVIRONMENTAL PERFORMANCE OF TANKER DESIGNS 78 TABLE 4-3 Vessel Condition Prior to Collision and Grounding Single-Hull, Double-Hull, Single-Hull, Double-Hull, Parameter 150,000 DWT 150,000 DWT 40,000 DWT 40,000 DWT Displacement (metric tons) 175,907 175,940 47,448 49,410 Cargo oil (metric tons) 149,635 149,635 35,949 35,922 Draft at MS [meters (m)] 16.78 17.12 10.58 11.17 Draft at fore perpendicular (m) 16.78 17.12 10.58 11.17 Draft at aft perpendicular (m) 16.78 17.12 10.58 11.17 Summer load line (m) 16.785 17.205 10.614 11.303 VCG (m) 13.35 14.71 7.526 9.26 Waterplane area [square 11,506 11,513 4,800 5,014 meters (m2)] Transverse metacentric height 6.96 5.15 3.243 2.91 (GMt) (m) Longitudinal metacentric 306.65 299.75 276.54 263.19 height (GMl) (m) Distance from MS to LCF 0.85 fwd 0.58 fwd 0.131 aft 2.81 aft (LCF relative to MS) (m) Density [grams per cubic centimeter (g/cm3)] Cargo 0.84 0.84 0.84 0.84 Fuel oil 0.98 0.98 0.98 0.90 Diesel oil 0.90 0.90 0.90 N/A Lube oil 0.85 0.85 0.85 0.92 Fresh water 1.00 1.00 1.00 1.00 Salt water 1.025 1.025 1.025 1.025 Tanks (% full) Cargo tanks (ex slops) 98 98 98 98 Slop tanks 66 89 N/A N/A FO 96 96 96/95 96 FO settling, service, overflow 20 20 N/A N/A DO 96 96 N/A N/A DO service 20 20 N/A N/A Fresh water 98 98 98 100 Forepeak ballast tank 3.5 11 0 100/11.2 Other ballast tanks 0 0 0 0 NOTE: DO = diesel oil, FO = fuel oil, LCF = longitudinal center of flotation, MS = midships, N/A = not applicable, VCG = vertical center of gravity. members on the side. Horizontal members, side shells, and bulkheads (transverse and longitudinal) are modeled by smeared thickness (which includes the effect of stiffeners). Web frames are modeled with addi- tional detail. Once collision damage calculations have been completed, SIMCOL determines which cargo tanks have ruptured. The outflow calculation as- sumes that all oil is lost from a damaged cargo tank. No hydrostatic or hydrodynamic effects are taken into account.

OCR for page 70
APPLICATION OF THE METHODOLOGY 79 The friction coefficient in DAMAGE was assumed to be 0.3.5 The ship model includes the cargo block without the bow and stern of the ship. Structural resistance is analyzed in a stepwise manner by moving the ship forward and, at each time step, finding the rock penetration and static equilibrium of the ship. Ship motions, excluding sway and yaw, are taken into account. Heave, roll, and pitch motions are calculated on the basis of static equilibrium using a simplified model. Surge motion is based on energy balance. Validation of DAMAGE can be found in work by Simonsen (1998) and Tikka (1998). The DAMAGE program outputs the vertical, horizontal, and lon- gitudinal extent of damage, and this information was used to determine the damaged tanks. The oil outflow from damaged cargo tanks was cal- culated according to the principle of hydrostatic balance. The pressure balance was calculated at the lowest point of the damaged tank. Similar to the assumptions in the IMO guidelines (IMO 1996), some oil was assumed to be captured in the ballast tanks, and a minimum outflow was assumed from damaged cargo tanks adjacent to seawater as a result of dynamic effects. The IMO guidelines assume a constant value for capture and minimum outflow, whereas the calculations used by the committee were based on a range of values sampled from the initial distributions for these variables to account for the uncertainty involved. The outflow calculation was performed in the initial condition (conceptual analysis), and it did not include a damage stability (surviv- ability) analysis.6 Tikka (1998) compared the conceptual analysis with a survivability analysis for double-hull tankers with a range of tank arrange- ments in four size ranges. The error in the mean outflow was small for con- ventional tank arrangements (typically less than 6 percent), and the error in the zero-outflow probability was insignificant (less than 0.1 percent) with no tide when conceptual analysis was used. At lower tides, the error percentages were smaller. Large errors were found for tank arrangements without centerline bulkheads, which can no longer be built according to MARPOL regulations. Finally, if one of the designs includes an active system, that sys- tem's effect should be taken into account in the outflow analysis. As dis- cussed in Chapter 3, a proposal involving an active system must include a reliability analysis. Results of the reliability analysis provide distribu- 5Friction coefficient values of 0.3 to 0.4 are typically used in grounding analyses (Simonsen 1998). 6Conceptual analysis assumes that the ship is aground on a shelf at a draft equal to the initial intact draft. Survivability analysis takes into account changes in the ship's condition and includes a damage stability analysis. If the vessel does not meet the MARPOL damage stability criteria, all oil is assumed lost.

OCR for page 70
ENVIRONMENTAL PERFORMANCE OF TANKER DESIGNS 80 tions for the variables that define the operation of the active system, and these variables and distributions are taken into account in the generation of collision and grounding scenarios. In other words, an active system adds another layer of uncertainty analysis to the methodology, and this is taken into account in the definition of the accident scenarios. Once the accident scenarios have been defined, determination of the outflow volume in each scenario must be adapted to the particular active system, but the basic principles of the methodology apply in the same way to analysis of active and passive systems. RESULTS OF COLLISION AND GROUNDING ANALYSES Collision and grounding analyses provided oil outflow for each of the 10,000 collision and grounding events. A summary of the results is pro- vided in Tables 4-4 and 4-5, respectively. The results indicate that a double hull is effective in reducing the number of spills due to collision and grounding. For collision, there were cases in which the double-hull design had a spill, but damage to the single-hull design occurred only to the side ballast tanks and resulted in no spill. For grounding, there were no cases in which the double-hull design spilled and the single-hull did not. The double-hull designs reduced the number of spills (over the single-hulls) by 54 and 67 percent for the 150,000- and 40,000-DWT tankers, respec- tively. In grounding, the smaller tankers had significantly fewer spills than the larger ones (1,833 versus 5,911 scenarios) because of their shallower drafts. The obstruction depths were the same in both analyses.7 For col- lision, the double-hull vessels had a larger average spill size (given a spill) than the single-hulls, but the single-hulls had a larger maximum spill. For the grounding scenarios, in comparing average spill size given a spill, the single-hull vessel had a larger average spill than the double-hull in the 150,000-DWT size, but the reverse was true for the 40,000-DWT size. The double-hull designs had a larger maximum spill than the single-hulls. The tank subdivision in the transverse direction had a strong impact on the outflow results, given a breach of the cargo block, in both collision and grounding. The analyzed single-hull vessels had three cargo tanks across, whereas the double-hull vessels had two tanks. The side tanks in the single-hulls were smaller than those in the double-hulls. On 7Because the depth and size of the obstacles were kept the same for both the 40,000- and 150,000-DWT tanker analyses, comparison across the two tanker sizes could be performed.

OCR for page 70
ENVIRONMENTAL PERFORMANCE OF TANKER DESIGNS 102 (single-hull and double-hull) and two different ship sizes (40,000-DWT and 150,000-DWT). Because identical accident scenarios across designs were used, the committee was able to compare outflow performance at the individual scenario level; it was not necessary to resort to comparison of the distributions of outflows. A fraction of the 80,000 scenarios resulted in outflow, which was converted to the consequence measure, the 500,000- gallon spill equivalent. By calculating and analyzing the differences in this measure across the scenarios, it was possible to determine whether an alternative design was equivalent to the target double-hull design. Transforming the outflow gallons into 500,000-gallon spill equiva- lents made small spills look relatively more damaging than large spills. This point becomes important when one is examining the difference in outflow consequences: differences in large spills will appear less important than equal-sized differences in small spills. Table 4-9 provides an example of this effect. Even though the difference in spill amounts between the two de- signs is a constant 100,000 gallons, this difference has decreasingly less con- sequence as the absolute spill size increases. A 100,000-gallon difference in outflow when one design does not leak is equivalent to 0.563 times the consequence of a 500,000-gallon spill, whereas a 100,000-gallon difference when both designs spill about 20,000,000 gallons is equal to 0.007 times the consequence of a 500,000-gallon spill. Table 4-10 provides a summary of the outflow and consequence measures for each of the design comparisons. In this table, the 80,000 ac- cident scenarios are broken down by accident type, ship size, and perfor- mance measures (i.e., which design spilled more oil). For example, in comparing the 40,000-DWT designs, neither design spilled oil in 8,167 of 10,000 grounding scenarios; of the remaining scenarios, only the single-hull design spilled in 1,221 cases, and both spilled in 612 cases. In no case did the double-hull spill and the single-hull not spill. For the 612 cases in TABLE 4-9 Demonstration of the Consequence Function Outflow in Gallons 500,000-Gallon Spill Equivalents Design 1 Design 2 Difference Design 1 Design 2 Difference 0 100,000 100,000 0.000 0.563 0.563 500,000 600,000 100,000 1.000 1.067 0.067 1,000,000 1,100,000 100,000 1.281 1.325 0.044 5,000,000 5,100,000 100,000 2.275 2.291 0.016 10,000,000 10,100,000 100,000 2.914 2.924 0.010 20,000,000 20,100,000 100,000 3.732 3.739 0.007

OCR for page 70
APPLICATION OF THE METHODOLOGY 103 TABLE 4-10 Characteristics of the 80,000 Accident Analyses Both Designs Spill One Design Spills SH > DH DH > SH Event Measuresa No Spill Only SH Only DH SH DH SH DH Collision 40,000-DWT Countb 5,711 2,885 244 219 941 Avg. mil. gal. N/A 0.72 0.74 1.62 0.92 0.73 0.99 Avg. 500k equiv. N/A 1.13 1.14 1.50 1.22 1.14 1.26 150,000-DWT Countb 6,567 2,407 250 365 411 Avg. mil. gal. N/A 3.10 4.20 5.79 3.86 2.77 4.83 Avg. 500k equiv. N/A 1.89 2.12 2.32 2.07 1.80 2.19 Grounding 40,000-DWT Countb 8,167 1,221 0 273 339 Avg. mil. gal. N/A 0.62 0.00 0.90 0.68 0.60 0.96 Avg. 500k equiv. N/A 1.00 0.00 1.15 1.02 1.00 1.18 150,000-DWT Countb 4,089 3,187 0 1,622 1,102 Avg. mil. gal. N/A 4.06 0.00 5.42 3.24 3.22 4.88 Avg. 500k equiv. N/A 2.00 0.00 2.25 1.83 1.82 2.06 NOTE: N/A = not applicable. aMeasures shown are count of scenarios in each spill category (count), average outflow in millions of gallons (avg. mil. gal.), and average outflow in 500,000-gallon spill equivalent units (avg. 500k equiv.). bEach count represents a comparison between the results of two accident scenarios, one for a double-hull and one for a single-hull design. which both spilled, the single-hull spilled more in 273 cases, and the double-hull spilled more in 339.14 Also shown in the table are the average spill sizes (in millions of gallons) and average consequence (in 500,000- gallon spill equivalents). Note that average outflow for the case in which only the single-hull tanker spilled was 0.62 million gallons and that the values for both designs' spills were generally larger (0.60 million to 1.18 million gallons). The fact that when both designs spilled, the outflow numbers were large but relatively close to each other means that when these outflow values are expressed in terms of 500,000-gallon spill equiva- lents, the average equivalent units are very close (1.00 to 1.15). This pattern of relationships is generally repeated across the remainder of the table. The only major difference is that with the collision analysis, there was a small number of accidents that caused the double-hull design to spill but not the single-hull (244 for the 40,000-DWT design and 250 for the 150,000-DWT design). 14These numbers are specific to the selected designs and cannot be used to draw general conclusions on all double- and single-hull designs. The choice of the reference ship to be used in the evaluation of equivalency is important.

OCR for page 70
ENVIRONMENTAL PERFORMANCE OF TANKER DESIGNS 104 To determine the equivalency of two designs, differences mea- sured in consequence units taken at the accident scenario level must be analyzed. Table 4-11 provides summary data for these differences. This table depicts benefit measures for use of the different designs. For ex- ample, for 40,000-DWT vessels involved in groundings, the double-hull design spilled less oil than the single-hull in 1,494 scenarios (when at least one design spilled), and the single-hull outperformed (spilled less than) the double-hull in 339 scenarios (in the remaining 8,167 scenarios, neither design spilled oil; see Table 4-10). In the 1,494 scenarios in which the double-hull design performed better:15 On average, the benefit of having a double hull was 0.55 million gallons per scenario. The total amount not spilled because of using a double hull was 816 million gallons. On average, the benefit of using a double hull was 0.84 500,000- gallon spill equivalents per scenario. The total number of 500,000-gallon spill equivalents saved because of using the double hull was 1,250. In the 339 scenarios in which the single-hull design performed better: On average, the benefit of having a single hull was 0.36 million gallons per scenario. The total amount not spilled because of using a single hull was 124 million gallons. On average, the benefit of using a single hull was 0.18 500,000- gallon spill equivalents per scenario. The total number of 500,000-gallon spill equivalents saved because of using the single hull was 61. Note that the gallon measures are included only for reference. It is the committee's strong recommendation that all design comparisons be made using the 500,000-gallon spill equivalent units. Thus in comparing the two designs, the double hull had better performance under some scenarios and the single hull under others. To 15The "benefits" shown cannot be used to calculate amounts of oil that would not have been spilled in pre- vious real-world accidents had a particular design been in use. These benefits pertain only to the hypothetical accident scenarios used in the simulation.

OCR for page 70
APPLICATION OF THE METHODOLOGY 105 TABLE 4-11 Design Comparison Using Difference in Outflows Double Single Net Performance Hull Hull Advantage Advantage Performs Performs to Double of Double Event Measuresa Betterb Betterb Hullb Hullc Collision 40,000-DWT Count 3,104 1,185 1,919 2.62 Avg. mil. gal. 0.72 0.36 0.36 1.99 Sum mil. gal. 2,237 429 1,808 5.22 Avg. 500k equiv. 1.07 0.33 0.74 3.22 Sum 500k equiv. 3,326 394 2,932 8.44 150,000-DWT Count 2,772 661 2,111 4.19 Avg. mil. gal. 2.94 2.89 0.05 1.02 Sum mil. gal. 8,144 1,910 6,234 4.26 Avg. 500k equiv. 1.67 1.04 0.63 1.61 Sum 500k equiv. 4,627 687 3,940 6.73 Grounding 40,000-DWT Count 1,494 339 1,155 4.41 Avg. mil. gal. 0.55 0.36 0.18 1.50 Sum mil. gal. 816 124 692 6.59 Avg. 500k equiv. 0.84 0.18 0.66 4.65 Sum 500k equiv. 1,250 61 1,189 20.50 150,000-DWT Count 4,809 1,102 3,707 4.36 Avg. mil. gal. 3.43 1.67 1.76 2.06 Sum mil. gal. 16,489 1,835 14,654 8.99 Avg. 500k equiv. 1.46 0.26 1.20 5.62 Sum 500k equiv. 7,033 287 6,746 24.54 a Measures shown include number of scenarios in each spill category (count), average benefit in millions of gallons (avg. mil. gal.), sum of benefits in millions of gallons (sum mil. gal.), average benefit in 500,000-gallon spill equivalent units (avg. 500k equiv.), and sum of benefits in 500,000-gallon spill equivalent units (sum 500k equiv.). b Values given indicate benefits of the design. cValues given indicate how many times better the performance of the double-hull design is relative to that of the single-hull overall. determine which design is superior overall (i.e., has the smaller total con- sequence), the difference between the measures can be calculated. For the designs analyzed in this report, the double hull was clearly superior. Table 4-11 indicates that, for the 40,000-DWT tankers involved in grounding incidents, use of a double-hull design in the 10,000 scenarios saved 1,189 500,000-gallon spill equivalent units over use of a single-hull design. Note that all of the other measures in the table also show a considerable ad- vantage for the double-hull design. This will not necessarily be the result if the methodology is applied in other comparisons. For example, a mid- deck design could have a positive benefit relative to a double-hull design with regard to total gallons spilled but be inferior according to the 500,000- gallon spill equivalent measure. To demonstrate how this could happen,

OCR for page 70
ENVIRONMENTAL PERFORMANCE OF TANKER DESIGNS 106 the last column in Table 4-11 shows the relative benefit of the double-hull design across the different measures and scenarios. Referring back to the 40,000-DWT tankers in grounding incidents, the double-hull tanker was superior to the single-hull in 4.41 times more scenarios, spilled 6.59 times less oil, and caused 20.50 times less damage according to the 500,000- gallon spill equivalent measure. These differing ratios show how different measures could lead to apparently conflicting results. This is precisely why the committee believes the consequence measure should be the final deter- minant of design equivalency with the double hull. Figures 4-24 through 4-27 depict the 500,000-gallon spill equiv- alent results shown in Table 4-10. They also demonstrate how sensitive the results are to assumptions about the consequence function. Figure 4-24a shows the results for the 40,000-DWT tankers involved in grounding sce- narios using the best-fit consequence function. The light gray region in the lower left corner of the figure represents the 339 scenarios in which the single-hull design outperformed the double-hull. The height of the region indicates the number of 500,000-gallon spill equivalent units saved by using the single-hull design in each scenario. The larger black area rep- resents the 1,494 scenarios in which the double-hull design outperformed the single-hull. The area of the two regions represents the sum of 500,000- gallon spill equivalent units (1,250 for the black region and 61 for the gray). If the designs were equivalent, the two regions would be equal in size. Figures 4-24b and c show the same results using the lower- and upper-bound consequence functions from Figure 4-22, respectively. The lower-bound function is even more concave, increasing the relative impact of smaller versus bigger spills, while the upper-bound function is closer to a linear relationship between spill size and consequence. The effect of these different functions is not significant in determining the equivalency of the two designs: the double hull is clearly superior in all figures. However, the changes in the shaded areas highlight the impor- tance the functional form could have for future comparisons. In Figure 4-24b it is easy to identify the 1,221 scenarios from Table 4-10 in which only the single-hull design spills (the steep "cliff" in the black region just beyond 1,200). This cliff disappears when the flatter upper-bound conse- quence function is applied. This is an example of the effect demonstrated in Table 4-9. Figure 4-25 shows the same analysis for the 150,000-DWT designs. Figures 4-26 and 4-27 show the results of the collision analysis for the 40,000-DWT and 150,000-DWT vessels, respectively. Note that in a few 150,000-DWT scenarios the single-hull has a larger advantage over

OCR for page 70
2 107 Measure 1.75 Double-Hull Advantage Single-Hull Advantage equivalent) 1.5 1.25 spill 1 Consequence 0.75 in 0.5 0.25 (500,000-gallon 0 Difference 0 200 400 600 800 1,000 1,200 1,400 Number of Scenarios (a) 1.50 Double-Hull Advantage Measure 1.25 Single-Hull Advantage equivalent) 1.00 spill 0.75 Consequence 0.50 in 0.25 0.00 (500,000-gallon Difference 0 200 400 600 800 1,000 1,200 1,400 Number of Scenarios (b) 4.00 Measure 3.50 Double-Hull Advantage equivalent) 3.00 Single-Hull Advantage 2.50 spill 2.00 Consequence 1.50 in 1.00 0.50 (500,000-gallon0.00 Difference 0 200 400 600 800 1,000 1,200 1,400 Number of Scenarios (c) FIGURE 4-24 Comparison of 40,000-DWT vessels in grounding scenarios using three models: (a) best fit, (b) lower bound, and (c) upper bound.

OCR for page 70
ENVIRONMENTAL PERFORMANCE OF TANKER DESIGNS 3.5 108 Double-Hull Advantage Measure 3.0 Single-Hull Advantage equivalent) 2.5 spill 2.0 1.5 Consequence in 1.0 0.5 (500,000-gallon 0.0 Difference 0 500 1,000 1,500 2,000 2,500 3,000 3,500 4,000 4,500 Number of Scenarios (a) 1.8 Double-Hull Advantage Measure 1.5 Single-Hull Advantage equivalent) 1.2 spill 0.9 Consequence 0.6 in 0.3 (500,000-gallon 0.0 Difference 0 500 1,000 1,500 2,000 2,500 3,000 3,500 4,000 4,500 Number of Scenarios (b) 9.0 Double-Hull Advantage Measure Single-Hull Advantage 7.5 equivalent) 6.0 spill 4.5 Consequence 3.0 in 1.5 0.0 (500,000-gallon Difference 0 500 1,000 1,500 2,000 2,500 3,000 3,500 4,000 4,500 Number of Scenarios (c) FIGURE 4-25 Comparison of 150,000-DWT vessels in grounding scenarios using three models: (a) best fit, (b) lower bound, and (c) upper bound.

OCR for page 70
APPLICATION OF THE METHODOLOGY 1.6 109 Double-Hull Advantage Measure 1.4 Single-Hull Advantage equivalent) 1.2 1.0 spill 0.8 Consequence 0.6 in 0.4 0.2 (500,000-gallon 0.0 Difference 0 250 500 750 1,000 1,250 1,500 1,750 2,000 2,250 2,500 2,750 3,000 Number of Scenarios (a) 1.4 Double-Hull Advantage Measure 1.2 Single-Hull Advantage equivalent) 1.0 spill 0.8 0.6 Consequence in 0.4 0.2 (500,000-gallon 0.0 Difference 0 250 500 750 1,000 1,250 1,500 1,750 2,000 2,250 2,500 2,750 3,000 Number of Scenarios (b) 2.0 Double-Hull Advantage Measure Single-Hull Advantage equivalent) 1.5 spill 1.0 Consequence in 0.5 (500,000-gallon Difference 0.0 0 250 500 750 1,000 1,250 1,500 1,750 2,000 2,250 2,500 2,750 3,000 Number of Scenarios (c) FIGURE 4-26 Comparison of 40,000-DWT vessels in collision scenarios using three models: (a) best fit, (b) lower bound, and (c) upper bound.

OCR for page 70
ENVIRONMENTAL PERFORMANCE OF TANKER DESIGNS 3.0 110 Measure Double-Hull Advantage 2.5 Single-Hull Advantage equivalent) 2.0 spill 1.5 Consequence 1.0 in 0.5 0.0 (500,000-gallon Difference 0 250 500 750 1,000 1,250 1,500 1,750 2,000 2,250 2,500 2,750 Number of Scenarios (a) 1.8 Double-Hull Advantage Measure 1.5 Single-Hull Advantage equivalent) 1.2 spill 0.9 Consequence 0.6 in 0.3 (500,000-gallon 0.0 Difference 0 250 500 750 1,000 1,250 1,500 1,750 2,000 2,250 2,500 2,750 Number of Scenarios (b) 8.0 Double-Hull Advantage Measure 7.0 Single-Hull Advantage 6.0 equivalent) 5.0 spill 4.0 Consequence 3.0 in 2.0 1.0 (500,000-gallon0.0 Difference 0 250 500 750 1,000 1,250 1,500 1,750 2,000 2,250 2,500 2,750 Number of Scenarios (c) FIGURE 4-27 Comparison of 150,000-DWT vessels in collision scenarios using three models: (a) best fit, (b) lower bound, and (c) upper bound.

OCR for page 70
APPLICATION OF THE METHODOLOGY 111 the double-hull than the double-hull has over the single-hull in any sce- nario. Nonetheless, a comparison of the two shaded regions shows the overall benefits of using the double-hull design. Since the committee did not attempt to determine the likelihood of any of the scenarios actually occurring, the above analysis cannot be used to determine the actual savings that might be incurred by using a fleet of double-hull designs in a given time period. This is a very different and more difficult problem that would require a detailed risk analysis of a specific port area with defined operations and traffic patterns--the type of analysis that would be required if one wanted to determine the relative costs and values of a double-hulled fleet. SUMMARY By performing the applications described in this chapter, the committee has demonstrated that its methodology can be used to compare tanker de- signs and determine their relative environmental performance. The ap- plicability of available computational tools to the prediction of structural damage and resulting oil outflow in multiple accident scenarios has also been shown. The applications involved single-hull and double-hull de- signs of two different sizes. The committee checked the outflow results for two vessel sizes to show that the methodology provides consistent re- sults for a range of possible conditions, as well as reasonable distributions of the differences when two designs are compared. The committee then tested its approach to the development of a consequence metric that can be used to modify the outflow results and represent a rational measure of environmental consequence differences in a design comparison. The com- mittee also conducted sensitivity analyses to test the rigor of the conse- quence function. Finally, the committee illustrated its methodology for comparing designs for a range of both collision and grounding events. The final design comparisons showed that the methodology can be applied to designs with different features, and the results can be depicted graphically to determine which design exhibits superior performance. The committee understands that its methodology will lead to un- ambiguous results only when the factors used in the comparison of designs, taken together, show clearly that one design has superior performance over another. In the illustration in this report, that appears to be the case. There could be instances, however, in which the relative performance of one design versus another is not obvious using this methodology alone. In those cases the regulatory authority must exercise its judgment, take into account other factors in the comparison, and make a determination that

OCR for page 70
ENVIRONMENTAL PERFORMANCE OF TANKER DESIGNS 112 is consistent with its public mission. The committee believes that any proposed methodology would require a similar process. REFERENCES ABBREVIATIONS IMO International Maritime Organization OTA Office of Technology Assessment, U.S. Congress SNAME Society of Naval Architects and Marine Engineers Carls, M. G., S. D. Rice, and J. E. Hose. 1999. Sensitivity of Fish Embryos to Weathered Crude Oil: Part I. Low-Level Exposure During Incubation Causes Malformation, Genetic Damage, and Mortality in Larva Pacific Herring (Clupea pallasi). Environmental Toxicology and Chemistry, Vol. 18, No. 3, pp. 484493. Henitz, R. A., J. W. Short, and S. D. Rice. 1999. Sensitivity of Fish Embryos to Weathered Crude Oil: Part II. Increased Mortality of Pink Salmon (Oncorhynchus gorbuscha) Embryos Incubating Downstream from Weathered Exxon Valdez Crude Oil. Environmental Toxicology and Chemistry, Vol. 18, No. 3, pp. 494503. IMO. 1992. Report on IMO Comparative Study on Oil Tanker Design. Report MEPC 32/7/15. London, Feb. IMO. 1996. Interim Guidelines for the Approval of Alternative Methods of Design and Construction of Oil Tankers Under Regulation 13F of Annex I of MARPOL 73/78. MARPOL 73/78 1994 and 1995 Amendments. London. OTA. 1990. Coping with an Oiled Sea: An Analysis of Oil Spill Response Tech- nologies. OTA-BP-O-63. U.S. Government Printing Office, Washington, D.C., March. Rawson, C., K. Crake, and A. Brown. 1998. Assessing the Environmental Perfor- mance of Tankers in Accidental Grounding and Collision. Presented at the SNAME Annual Meeting, San Diego, Calif. Sandia National Laboratories. 1998. Data and Methods for the Assessment of the Risks Associated with the Maritime Transport of Radioactive Materials: Results of the SeaRAM Program Studies. Report SAND98-1171/2. Albuquerque, N.Mex., May. Simonsen, B. C. 1998. DAMAGE Theory Validation. Report 63. Joint MIT-Industry Program on Tanker Safety, Boston, Mass., May. Tikka, K. K. 1998. Review and Improvement of the IMO Probabilistic Methodology for Evaluating Alternative Tanker Designs. Webb Project 1812-351. American Bureau of Shipping, April.