INCLUSION OF A CREW SAFETY NODE INTO THE PRELIMINARY DESIGN OF FISHING VESSELS By Ayhan Akinturk BSc. , Istanbul Technical University, Istanbul, Turkey 1986; MSc. , University of Newcastle Upon Tyne, U.K. 1990 A T H E S I S S U B M I T T E D I N P A R T I A L F U L F I L L M E N T O F T H E R E Q U I R E M E N T S F O R T H E D E G R E E O F D O C T O R O F P H I L O S O P H Y in T H E F A C U L T Y O F G R A D U A T E S T U D I E S D E P A R T M E N T O F M E C H A N I C A L E N G I N E E R I N G We accept this thesis as conforming to the required standard T H E U N I V E R S I T Y ' O F B R I T I S H C O L U M B I A May 1997 © Ayhan Akinturk, 1997 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of Mechanical Engineering The University of British Columbia 2324 Main Mall Vancouver, B.C., Canada V6T 1Z4 Abstract The working and living conditions on-board of fishing vessels affect the crew's perfor-mance and well being. They contribute to occupational accidents on-board, which cost to the health and lives of crew members. There is also a cost of lost opportunity, when crew members can not perform their duties on-board and have to halt fishing due to deteriorated conditions at sea. In this thesis, a methodology that will allow the inclusion of crew comfort and safety considerations into the preliminary design of fishing vessels has been developed. This new methodology and the traditional preliminary design spired for monohull and SWATH (Small Waterplane Area Twin Hull) vessels have been implemented in Echidna, which is a logic programming environment that supports constraint based reasoning. In the literature, ship motions have been reported to be the most prominent contrib-utory factor to the occupational accidents and crew's performance on-board. Hence, two sets of rules suitable to a knowledge-based environment and based on different engineer-ing concepts have been developed and implemented for the preliminary design of fishing vessels to improve their seakeeping characteristics. Unlike some ship motions' calculation programs, for example SHIPMO, these rules do not require detailed hull form definition. Hence, they are used as guidelines (or heuristic rules) during the initial stages of ship design. Considering the nonlinear nature of ship design, the procedure developed was able to find a solution for a given design sea state and owner requirements. The effects of the rules on ship size, cost and improved seakeeping qualities are presented in this thesis. Additionally, monohull and SWATH vessels for the same owner requirements are ii compared in terms of vessel size and cost. Finally, the knowledge-based system described in this thesis provides a tool to map crew comfort levels and a design sea state to the vessel parameters. Hence the cost difference due to the crew comfort and safety considerations can be quantified. The methodology described here can easily be applied for small craft with small changes to the knowledge base. Keywords : ship design, crew safety, crew comfort, knowledge-based design, seakeeping. iii Table of Contents Abstract ii List of Tables viii List of Figures xii Acknowledgements xix Chapter 1 INTRODUCTION 1 1.1 Objective of This Study 6 1.2 Proposed Method 7 Chapter 2 OCCUPATIONAL SAFETY ON-BOARD FISHING VESSELS 9 2.1 Statistics of Accidents On-board : -. . 9 2.1.1 Canada 12 2.1.2 France 16 2.1.3 United Kingdom . 20 2.1.4 Norway 21 2.1.5 Spain 27 2.1.6 Japan 32 2.2 Some notes on accident statistics 32 2.3 Impact of Ship Motions on Fishing 33 iv 2.4 Design considerations for improved working conditions on-board based on warships 39 2.5 Methods for reduced motions 43 2.6 Design margins 45 Chapter 3 A N OVERVIEW OF SHIP DESIGN 47 3.1 Overview of Literature for Ship Design 49 3.2 Overview of the Ship Design Process (Design Spiral) 55 3.3 Summary 56 Chapter 4 DESIGN, KNOWLEDGE-BASED SYSTEMS AND ECHIDNA 59 4.1 Design in a Knowledge-Based Environment 61 4.2 Echidna Expert System Shell 63 Chapter 5 MONOHULL VESSEL DESIGN KNOWLEDGE BASE (UBC-MONO) 68 5.1 Remarks 70 5.2 Summary 76 Chapter 6 SWATH PRELIMINARY DESIGN AND UBC-SWATH 84 6.1 Multi-hull vessel for fishing role 84 6.2 General Features of Multi-hull Design . 85 6.3 SWATH Design Algorithms and Discussion 86 6.4 Remarks 90 Chapter 7 SEAKEEPING CONSIDERATIONS 97 7.1 Rule Set I 99 7.2 Evaluation of the Example Designs by Rule Set I 114 7.2.1 Remarks about the ship dimensions and displacement 120 7.3 Rule Set II 134 7.4 Comparison of the outcomes of Rule Sets I and II 149 7.5 The Criterion Suggested by Kimura et al. [40] 156 7.5.1 Summary 160 Chapter 8 SWATH MONOHULL COMPARISON 174 Chapter 9 CONCLUSIONS 179 9.1 Suggestions for Future Work 181 Nomenclature 184 Bibliography 186 Appendix A Echidna - Knowledge Base Environment 193 A.l Run-time user commands 196 A.2 Knowledge base for the barge example 197 A. 3 A sample output for the barge example 202 Appendix B Monohull Design Algorithms 206 B. l Linear Dimensions 206 vi B.2 Weight Estimation 208 B.3 Resistance and Powering 209 B.4 Stability 211 B. 5 Cost Estimation 212 Appendix C SWATH design algorithms 214 C. l Linear Dimensions 214 C.2 Strut(S) Linear Dimensions 217 C.3 Weight Estimation 217 C.4 Resistance and Powering 221 C.5 Cost Estimation 224 Appendix D Prediction of a vessel's heave and pitch motions 226 Appendix E Frequencies of Peak Heave and Pitch Responses 228 Appendix F RMS Heave and Pitch Responses 229 vii List of Tables 1.1 Deaths and accident rates for 1985 per 1000 employees in some industries [26] 3 1.2 Types of Casualties in Fishing 3 2.1 Accidents Aboard Canadian Registered Ships by Vessel Type (Source : [66, pp. 36]) 13 2.2 Fatalities Aboard Canadian Registered Ships by Vessel Type (Source : [66, pp. 36]) 13 2.3 Injuries Aboard Canadian Registered Ships by Vessel Type (Source : [66, PP.36]) . . . 15 2.4 Based on Tables 2.1, 2.2, and 2.3, the percentages of aboard casualties occurred on Canadian Registered fishing vessels 15 2.5 10 year averages of percentage values of primary contributing factors in casualties aboard all vessel types 16 2.6 Number of Accidents Aboard Canadian Registered Ships by Primary Con-tributing Factor. All vessel types mentioned in Tables 2.1, 2.2, and 2.3 are included. (Source : [66, pp. 38]) 17 2.7 Fatalities Aboard Canadian Registered Ships by Primary Contributing Factor. All vessel types mentioned in Tables 2.1, 2.2, and 2.3 are included. (Source : [66, pp. 38]) 17 viii 2.8 Injuries Aboard Canadian Registered Ships by Primary Contributing Fac-tor. All vessel types mentioned in Tables 2.1, 2.2, and 2.3 are included. (Source : [66, pp. 38]) 18 2.9 Deaths in United Kingdom Fishing Vessels during 1980 - 1988 (Source : [26]) 22 2.10 Ratio of number of fishers with respect to the length of the vessels in the Norwegian fishing fleet [41] 22 2.11 Contributing factors to on-board casualties in the Norwegian fishing fleet ([34] in [41]) 23 2.12 The most important causes of injury in the Norwegian fishing fleet [36]. Values are % of the total disabilities or less serious injuries 24 2.13 Survey on noise levels on 17 Norwegian fishing vessels ([1] from [41]) . . . 26 2.14 Number of casualties for different fishing types in the Norwegian fishing vessels ([12] in [41]) 27 2.15 % number of accidents on deck by type of crew member in Spanish fishing vessels [22] 29 2.16 % number of accidents by type of crew member in Spanish fishing vessels (Source [22]) 29 2.17 Places of accidents in Spanish fishing vessels [22] 30 2.18 Work agents causing the most serious injuries in Spanish fishing vessels [22] 31 2.19 Motion Correlations given by Walden and Grundmann [68] 44 5.1 Input to Echidna for Kynoc. "[ ]" indicates an input as an interval. . . . 71 5.2 Comparison of Echidna produced design with real Kynoc 72 5.3 Example for the intervals after Echidna reaches a solution 73 ix 5.4 Example for the parameters' intervals after Echidna completes design and further arbitrarily refines them 73 5.5 Input to Echidna for the second phase of validation for a fleet of varying hold capacities 74 6.1 Input parameters for SWATH vessel design expert system 87 6.2 Statistical analysis of % variations for A after excluding the pathological cases (4 cases out of 41 total number of cases). The 4 existing designs are as follows: a 305 ton fishing vessel, a 225 ton ferry, a 225 ton crew boat and a car ferry with a 1250 ton displacement 88 7.1 Input values to Echidna knowledge base for monohull fishing vessel. . . . 102 7.2 The averages and standard deviations (over hold capacities) of the % Changes obtained in rms heave and pitch motion amplitudes. The ves-sel is assumed to be operating in Sea State 5 117 7.3 The averages and standard deviations (over hold capacities) of the % Changes obtained in rms heave and pitch accelerations. The vessel is assumed to be operating in Sea State 5 118 9.1 Principal particulars of the SWATH M.V. Frederick G. Creed (from [28]). 175 9.2 Principal particulars of the monohull F.P.V. Louisbourg (from [28]). . . 175 9.3 Comparison of some of the significant motion characteristics between M.V. Frederick G. Creed and F.P.V. Louisbourg (from [28]) 176 A. l Refinement of intervals during the example barge design in Echidna . . . 195 B. l Coefficients for UBC Series Resistance Algorithm for = 0.615 210 C. l Residual drag coefficient as a function of volumetric Froude number. . . . 223 F . l Sea state information. [Sabuncu, 1983] 229 F.2 Scaling factors for the raw regression data obtained from the software SHIPMO 229 x i List of Figures 1.1 Deaths per 10,000 man-years in some occupations. (Source [36]) 2 2.1 Crew members trying to transfer an empty net from the factory ship. . . 10 2.2 A crew member working at the stern to secure one of the doors used to keep the mouth of the trawl net open while trawling 11 2.3 Tasks performed on-board after taking the catch is in (from [8, p 88]). . 14 2.4 Position of the fishing gear at the bow (from [32]) 20 2.5 Position of the fishing gear in the middle (from [32]) 21 3.1 Design Spiral 49 3.2 Representation of design process by Mistree et al. [50] 50 3.3 The process of basic design of ships in Akagi and Fujita's expert system (from [3]) 58 4.1 Generative and interpretive knowledge in defining spaces of designs (from [27]) 64 5.1 Comparison of Echidna generated aluminum and steel designs' lengths (vessel type : seiner). The jump could be because of alternative estimation rules embedded for some variables in the knowledge base 77 5.2 Comparison of Echidna generated aluminum and steel designs' beams (ves-sel type : seiner). The jump could be because of alternative estimation rules embedded for some variables in the knowledge base 78 xii 5.3 Comparison of Echidna generated aluminum and steel designs' drafts (ves-sel type : seiner). The jump could be because of alternative estimation rules embedded for some variables in the knowledge base 79 5.4 Comparison of Echidna generated aluminum and steel designs' hull weights (vessel type : seiner) 80 5.5 Comparison of Echidna generated aluminum and steel designs' required powers (vessel type : seiner) 81 5.6 Comparison of Echidna generated aluminum and steel designs' displace-ments (vessel type : seiner) 82 5.7 Comparison of Echidna generated aluminum and steel designs' costs (vessel type : seiner) 83 6.1 Comparison of U B C - S W A T H and existing design A's . The horizontal axis represents existing designs' displacement values 92 6.2 Comparison of U B C - S W A T H and existing design L0AS 93 6.3 Comparison of U B C - S W A T H and existing design Seam's 94 6.4 Comparison of U B C - S W A T H and existing design Draffs 95 6.5 Comparison of U B C - S W A T H and existing design Installed Powers. . . 96 7.1 Frequencies that maximize sea spectrum and ship response spectrum. . 107 7.2 The variation of the ratios (Equation 7.1). Control case shows the ratios without any seakeeping rules included in the design. Design-sea-state is 5; operational-sea-state is 5 108 7.3 The effects of Rule Set I on the lengths. Control case shows the lengths without any seakeeping rules included in the design. Design-sea-state is 5; operational-sea-state is 5 109 xiii 7.4 The effects of Rule Set I on the beams. Control case shows the beams without any seakeeping rules included in the design. Design-sea-state is 5; operational-sea-state is 5 110 7.5 The effects of Rule Set I on the drafts. Control case shows the drafts without any seakeeping rules included in the design. Design-sea-state is 5; operational-sea-state is 5 I l l 7.6 The effects of Rule Set I on the displacements. Control case shows the displacements without any seakeeping rules included in the design. Design-sea-state is 5; operational-sea-state is 5 112 7.7 The effects of Rule Set I on the costs. Control case shows the costs with-out any seakeeping rules included in the design. Design-sea-state is 5; operational-sea-state is 5 113 7.8 Values of rms for heave 122 7.9 % change in rms heave values 122 7.10 Values of rms heave 123 7.11 % change in rms heave values 123 7.12 Values of rms for heave 124 7.13 % change in rms heave values 124 7.14 Values of rms for pitch 125 7.15 % change in rms pitch values 125 7.16 Values of rms for pitch 126 7.17 % change in rms pitch values 126 7.18 Values of rms for pitch 127 7.19 % change in rms pitch values 127 8.20 Values of rms for heave accelerations (Rule Set I) 128 xiv 8.21 % change in rms heave accelerations (Rule Set I) 128 8.22 Values of rms for heave accelerations (Rule Set I) 129 8.23 % change in rms heave accelerations (Rule Set I) 129 8.24 Values of rms for heave accelerations (Rule Set I) 130 8.25 % change in rms heave accelerations (Rule Set I) 130 8.26 Values of rms for pitch accelerations (Rule Set I) 131 8.27 % change in rms pitch accelerations (Rule Set I) 131 8.28 Values of rms for pitch accelerations (Rule Set I) 132 8.29 % change in rms pitch accelerations (Rule Set I) 132 8.30 Values of rms for pitch accelerations (Rule Set I) 133 8.31 % change in rms pitch accelerations (Rule Set I) 133 8.32 The effects of Rule Set II on rms heave amplitudes. Design and operational sea states are 5 140 8.33 The effects of Rule Set II on rms pitch amplitudes. Design and operational sea states are 5 140 8.34 The effects of Rule Set II on lengths. Design-sea-state is 5 141 8.35 The effects of Rule Set II on beams. Design and operational sea states are 5.141 8.36 % Change in the lengths of the example designs. Design-sea-state is 5. . 142 8.37 % Change in the beams of the example designs. Design-sea-state is 5. . . 143 8.38 The effects of Rule Set II on drafts. Design and operational sea states are 5.144 8.39 The effects of Rule Set II on displacements. Design-sea-state is 5 144 8.40 % Change in the drafts of the example designs. Design-sea-state is 5. . . 145 8.41 % Change in the displacements of the example designs. Design-sea-state is 5 146 8.42 The effects of Rule Set II on costs. Design-sea-state is 5 147 8.43 % Change in the costs of the example designs. Design-sea-state is 5. . . . 148 xv 8.44 % Changes of the lengths of example designs obtained using Rule Sets I and II. Design-sea-state is 5 151 8.45 % Changes of the beams of example designs obtained using Rule Sets I and II. Design-sea-state is 5 152 8.46 % Changes of the drafts of example designs obtained using Rule Sets I and II. Design-sea-state is 5 153 8.47 % Changes of the displacements of example designs obtained using Rule Sets I and II. Design-sea-state is 5. . 154 8.48 % Changes of the costs of example designs obtained using Rule Sets I and II. Design-sea-state is 5 155 8.49 A crew member working at the stern 157 8.50 Values of Tm (Equation 8.15), where T m measures conditions to assess the ability of a crew member to maintain his balance (see page 38). Design-sea-state is 5 and operational-sea-state is 2; ship speed = 0.01 [fen]; midpoints of Echidna intervals were used 162 8.51 % Change in the values of Tm with respect to "Control case", in which there was no seakeeping considerations, (for the data in Figure 8.50). . . 163 8.52 Values of T m (Equation 8.15), where T m measures conditions to assess the ability of a crew member to maintain his balance (see page 38). Design-sea-state is 5 and operational-sea-state is 5; ship speed = 0.01[fenj; midpoints of Echidna intervals were used 164 8.53 % Change in the values of Tm with respect to "Control case", in which there was no seakeeping considerations, (for the data in Figure 8.52). . . 165 xvi 8.54 Values of T m (Equation 8.15), where T m measures conditions to assess the •ability of a crew member to maintain his balance (see page 38). Design-sea-state is 5 and operational-sea-state is 2; ship speed = 5[fcn]; midpoints of Echidna intervals were used 166 8.55 % Change in the values of Tm with respect to "Control case", in which there was no seakeeping considerations, (for the data in Figure 8.54). . . 167 8.56 Values of T m (Equation 8.15), where Tm measures conditions to assess the ability of a crew member to maintain his balance (see page 38). Design-sea-state is 5 and operational-sea-state is 5; ship speed = 5[An]; midpoints of Echidna intervals were used 168 8.57 % Change in the values of T m with respect to "Control case", in which there was no seakeeping considerations, (for the data in Figure 8.56). . . 169 8.58 Values of T m (Equation 8.15), where T m measures conditions to assess the ability of a crew member to maintain his balance (see page 38). Design-sea-state is 5 and operational-sea-state is 2; ship speed = 10[A;TI]; midpoints of Echidna intervals were used 170 8.59 % Change in the values of T m with respect to "Control case", in which there was no seakeeping considerations, (for the data in Figure 8.58). . . 171 8.60 Values of T m (Equation 8.15), where Tm measures conditions to assess the ability of a crew member to maintain his balance (see page 38). Design-sea-state is 5 and operational-sea-state is 5; ship speed = 10[fcra]; midpoints of Echidna intervals were used 172 8.61 % Change in the values of Tm with respect to "Control case", in which there was no seakeeping considerations, (for the data in Figure 8.60). . . 173 xvii 9.1 Comparison of the displacements of SWATH and monohull type of fishing vessels obtained using Echidna. For monohulls, Control case corresponds to the case in which there was no seakeeping considerations in the design. For SWATH designs, no seakeeping considerations were used during the design 177 9.2 For the mid points of Echidna intervals, the values of the ratios (SWATH displacements over monohull displacements, for data given in Figure 9.1). Control case is the one, in which there was no seakeeping considerations in the design 178 A . l An example class definition for rectangles (the object in this example). . 195 C. l Schematic of a SWATH vessel 225 xviii Acknowledgements I would like to express my sincere gratitude to Dr. S. M. Cah§al for his patience steadfast encouragement, guidance and support throughout my research. I would also like to thank the members of my supervisory committee, Dr. A. B. Dunwoody, Dr. F. Sassani, Dr. W. Havens and Ms. Judy Village for their invaluable comments and suggestions. Furthermore, I would especially thank Dr. Havens for allowing us to use Echidna, the expert system shell developed at the Expert Systems Laboratory, in Simon Fraser University. Without Echidna, the research would not have reached where it is today. Ms. Susan Sidebottom had been very patient with my questions while learning how to use Echidna. She and the other members of the laboratory deserve special thanks from me for their valuable answers and suggestions. I would also like to thank Ms. Judy Village, for her extremely helpful suggestions and commentary on the ergonomic aspects of this research. Dr. Mehmet Atlar has contributed a lot to this work. He has been ever helpful in any kind of difficulty during this research. I would like to express my indebtedness to him. Dr. Erol Varoglu showed me many interesting points related to research. Our stim-ulating talks have contributed greatly to my understanding of knowledge-based systems and problem solving strategies and have been times of enjoyment for me. I thank him sincerely for all of these. I would also like to thank the Ministry of Education of Turkey, without their spon-sorship I would not be able to undertake such an academic journey. Special thanks to Mr. Andrew Duthie from the Department of Fisheries and Oceans for his support in the earlier parts of this project. xix During my researh, my friends and colleagues had contributed to this work and my life through social events, technical discussions and immeasurable technical support at times. I am greatly indebted to them. Last but not least, my gratitude also goes to my family for their perpetual support and encouragement. x x Chapter 1 INTRODUCTION Ship design is one of the oldest endeavors of mankind. Historically, it has been a repro-duction of successful ships mostly. There have been a number of novel ideas introduced into different aspects of ship design, such as the introduction of steam engines as an al-ternative power source to sailing, using different hull materials e.g. steel, aluminum, fiber glass, etc., or different hull forms e.g. hydrofoils, catamarans, SWATH (Small Water-plane Area Twin Hull), etc. to name a few. However, life at sea for humans still continue to be hazardous. Statistics related to the risk levels in different occupations (see Figure 1.1) are in agreement with this fact and reveal that particularly fishing is among the most dangerous occupations. According to Hansen in [36] fishing with 13.7 deaths per 10000 man-years due to accidents is the most hazardous occupation among the others in Figure 1.1. A similar study done in England [26] reveals similar findings. Chaplin and Burney [26] reported the following comparison between the fishing industry and the some other industries (see Table 1.1). For the period of 1981 to 1988 for all vessels in the fishing industry, death and non-fatal accident rates per 1000 employees were 1.05 and 6.42 respectively. These rates are much higher than the rates for other industries reported in Table 1.1. It should be noted that the numbers given in Figure 1.1 and Table 1.1 are total numbers related to casualties in fishing. The numbers related to each type of categories of casualties in fishing given in Table 1.2 are not given separately. However, Dorval [32] mentions a statistical report done in France that had disturbing facts for life at sea. According to the statistics, the probability of a fisherman dying at sea is 3%, 1 Chapter 1. INTRODUCTION 2 while the probability of dying due to an occupational accident during his/her career is 50%. He reports that the situation is not much different for other European countries as well. 1 4 Figure 1.1: Deaths per 10,000 man-years in some occupations. (Source [36]) As described by K u o , "Safety is a perceived quality that determines to what extent the management, engineering and operation of a system is free of danger to life, property and the environment" [43]. According to this definition, these three categories make a different contribution to the overall safety in different circumstances. It is also implied that there is no absolute level of safety in performing a task. Perhaps, the best that can be expected is to assess priorities for problem areas, which wil l produce the greatest return for an Chapter 1. INTRODUCTION 3 Table 1.1: Deaths and accident rates for 1985 per 1000 employees in some industries [26] Industry Deaths Major accidents Ore and mineral extraction 0.54 2.97 Coke ovens 0.29 5.71 Oil and gas extraction 0.26 3.39 Coal mining 0.17 4.17 Construction 0.11 2.26 Meted manufacturing 0.06 2.42 Fishing industry 1.05 6.42 improved safety. It is important to ensure an acceptable level of safety. The general safety problem in fishing can be studied in two categories in terms of casualties related : 1. to vessel 2. to occupational accidents on-board. Each category might involve different types of casualties as given in Table 1.2. Table 1.2: Types of Casualties in Fishing Related to vessels Related to occupational accidents Grounding Falls Capsizing Crushing Collision Blows Fire and explosion Cuts, pricks, wounds Leaks Falling objects Foundering Caught by gear etc. etc. A safer vessel from the vessel casualties point of view may mean increased stability Chapter 1. INTRODUCTION 4 while the other group requires introduction of ergonomic design concepts, i.e. design of a fishing vessel as a work place. For example, in the first category, improvements in a vessel's safety could mean improved design features, e.g. increased initial stability, wa-tertight deckhouses or a better understanding of the factors that degrade safety in the vessel's lifetime. For the second category, occupational accidents, improved safety may mean improved working conditions on-board of vessels. This could involve reduction in levels of noise, as well as vibration and motions i.e motion amplitudes and acceleration levels, better deck and bridge design, training and education of fishermen, etc. It is also interesting to note that requirements for each category towards a safer vessel may dic-tate conflicting goals. For example, increased GM, which is a parameter related to initial stability, could be desirable from a vessel's safety point of view, though it may prove otherwise for the crew on-board as the frequency of roll motion may fall in the uncom-fortable region for humans. With regard to occupational casualties in fishing, Chaplin and Burney [26] believe that this is partly because a fishing vessel is a moving platform to work on and expose fishermen to the elements which are not generally experienced in other industries. Additionally, they suspect that unrestricted working hours in fishing, unlike other workers ashore, may impose higher levels of fatigue on the crew. There are two relevant factors for greater safety: • Identification of the main problem areas • Knowledge of remedial actions which will be effective. Then prevention from accidents may be considered in three stages: • Design Stage • Construction Stage • Operation Stage Chapter 1. INTRODUCTION 5 The identification of problem areas and the understanding of how casualties occur will help to improve the measures necessary to be taken. In the design stage, it is important to check whether any safety regulation is violated or not, and determine what kind of recommendations are provided by the rules of classification societies and regulatory bodies. It is also necessary to design an adequate protective system to withstand the effects of the accidents. Thus, every possible means of eliminating hazards can be taken into proper consideration. In the construction stage, safety is mostly related to the supervision of whether those safety features specified in the design stage were properly constructed and provided, such as hand-rails, non-skid floors, etc. In the operation stage, the training of crews and a good maintenance policy will influence safety problems, in particular the avoidance of hazards, listed in Table 1.2. More discussion can be found on the casualties related to occupational accidents on-board fishing vessels in Chapter 2. However, the following numbers taken from [66] with regard to accidents on-board of fishing vessel in Canada illustrate the severity of the safety problem: Year Injured Died 1988 109 12 1989 111 16 1990 80 11 The consequences of occupational accidents or danger of having casualties can be considered in two categories: • Cost of injured and dead crew to society and the economy. This cost can be expressed in monetary terms as well as the impact on those left behind. Chapter 1. INTRODUCTION 6 • Cost of lost opportunity to fish. As reported by Tupper [67], the New England fishing fleet in the U S A operates at 60% of its annual natural production capacity because they are forced to halt fishing due to deteriorated sea conditions. 1.1 Objective of This Study Both human health and life and monetary aspects of the losses due to occupational ac-cidents on-board fishing vessels continue to be a matter of concern. Wi th this in mind, in this study the main interest area is the identification of the most important factors affecting occupational safety, crew comfort and performance aboard fishing vessels. Af-ter the definition of the factors, the objective is then to develop a design tool for the preliminary design of fishing vessels, which will include some ergonomic rules in order to improve living and working conditions hence crew safety on-board fishing vessels. A de-sign modification including a step to minimize crew risk related to occupational accidents is also one of the objectives of this study. In order to develop such a design tool, one needs to establish some kind of mapping between the parameters that define a vessel and the factors related to a human being's performance or well-being on-board of a vessel: for example, above 0.2g R M S vertical acceleration, the performance of personnel on-board degrades for certain tasks. In order to be able to use such information in ship design we need to establish a relationship that maps a given sea conditions and a vessel to the acceleration levels on-board of the vessel under consideration. 1 1 According to the terminology used in Chapter 4, this is a mapping between the design characteristics and the performances. Chapter 1. INTRODUCTION 7 1.2 Proposed Method In earlier studies within the Mechanical Engineering Department in the University of British Columbia , a general formulation technique following a classical design spiral and using a nonlinear optimization technique was implemented for the preliminary design of fishing vessels. The procedure based on an objective function and a series of equalities and inequalities was solved with existing computer codes such as Coupler Optimization Technique [14]. As a follow-up an Expert System shell (PC Consultant) was used to simulate a classical ship design spiral [24]. This study is based on the use of a knowledge-based system developed at Simon Fraser University called Echidna [60], [37] which has features that make it more suitable for design. While in principle any nonlinear problem solver that handles inequalities and equalities could be used to design a ship, Echidna with its built-in logic to handle constraints, and the ability to reduce the design space using constraint propagation, offered a superior programming environment. The main difference between an approach based on a design spiral or a sequential solution and a knowledge-based solution is that in the latter there was no assigned solution path. The solution path is internally selected by the program based on the available information and constraints supplied to the knowledge base. The Echidna system enables the user to define constraints, ranges, tolerances and relationships for vessel design parameters such as length, beam, draft, displacement, etc. The built-in constraint processing facilitates the information flow between the parameters during each iterative cycle of the design, thus narrowing the feasible domains of the parameters. For the purpose of this study, the preliminary design of two ship concepts - namely monohull and SWATH, have been implemented in an Echidna environment. The SWATH concept was selected as an alternative to a monohull because of its superior seakeeping Chapter 1. INTRODUCTION 8 qualities. The implementation and validation of these concepts in Echidna are further discussed in Chapters 5 and 6. The results of these'two implementations, which had no ergonomic design considerations during the design, are reported in [25] and [4] for monohull and SWATH concepts respectively. From the existing information in the literature, vessel motions are reported to be the most important contributing factors to occupational accidents on-board of fishing vessels [67], [56], [31]. Two different rule sets, referred to as Rule Set I and Rule Set II in this thesis, are used during the design for improving the conditions on-board in terms of vessel motions. The effects of these two set of rules on the size of the designs as well as building costs are given in Chapter 7. Chapter 2 OCCUPATIONAL SAFETY ON-BOARD FISHING VESSELS Crew members perform different tasks on-board fishing vessels. As shown in Figures 2.1 and 2.2, the tasks could be related to handling of net and doors. Depending on the fishing method, e.g. gilnetting, trawling, etc., and whether or not a factory ship is used, crew members may also perform such tasks as gutting, transferring the fish along the deck, cleaning the catch after the catch is taken aboard as illustrated in Figure 2.3. During the course of this study, a fishing trip was made on-board Eastward-Ho; this is one of the fishing vessels operating in the West Coast of Canada. Two of the photos taken during that trip is given in Figures 2.1 and 2.2. These photos show two example work postures for this particular fishing method. In this fishing trip, the catch (handled-in net) was immediately transferred to a factory ship nearby. After this, the net was returned to the fishing vessel. Figure 2.1 shows some of the crew members trying to snag and hand the empty net on-board. Figure 2.2 shows a crew member securing one of the doors, also called as otter boards, after the fishing had been completed. These metal doors, act as lifting surfaces, which keep the front of the net open during trawling are quite heavy. In rough sea conditions, securing the doors may endanger the crew member. 2.1 Statistics of Accidents On-board In an attempt to understand the working conditions aboard fishing vessels, some of the accident statistics in the literature are reviewed in the present section. Statistics on 9 Figure 2.1: Crew members trying to transfer an empty net from the factory ship. Figure 2.2: A crew member working at the stern to secure one of the doors used to keep the mouth of the trawl net open while trawling. Chapter 2. OCCUPATIONAL SAFETY ON-BOARD FISHING VESSELS 12 accidents and related factors are useful in exposing the problem areas and detecting the patterns leading to an accident. The following is the review of some countries reported in the literature. 2.1.1 Canada In Canada, there are more than 42300 vessels according to Transport Canada, Fisheries and Oceans Canada [66]. For the fishing industry, it is estimated that the number of fishing vessels operating in Canadian waters are approximately 40000, of which some 20300 are on register. The rest of the fleet is believed to be consist of unregistered small fishing vessels. In a report published by the Transportation Safety Board of Canada [66], accident reports in the Canadian fleet, including all types of vessels, during the period of 1981 to 1990 were studied and some of the statistical results were summarized. Tables 2.1 through 2.3 from [66] are important in showing the present situation in fishing vessels in comparison to other vessel types in Canada. In the tables, the category "other" represents vessels such as research and survey, ice breaking, laying and repair of sea-bed cables, search and rescue, pilotage, dredging, patrol, naval service, pleasure, etc. According to the tables, fishing vessels seem to be one of the most dangerous vessel types. However, a simple comparison of the number of accidents among the different types of vessels does not reveal much in terms of exposure to risk. In Table 2.4, the percentages of the number of casualties occurred in fishing vessels are given. As 10 year averages, calculated from Table 2.4, 36% of the aboard accidents, 40.18% of the total number of fatalities and 35.57% of the total number of injuries occurred in fishing vessels. This is followed by cargo vessels with 28.57%, 24.38% and 29.18% in accidents, fatalities and injuries respectively. Chapter 2. OCCUPATIONAL SAFETY ON-BOARD FISHING VESSELS 13 Table 2.1: Accidents Aboard Canadian Registered Ships by Vessel Type (Source : [66, pp. 36]) Vessel Type 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 Cargo 37 47 65 63 65 46 61 82 80 83 0B0 1 3 0 4 0 0 2 0 2 2 Tanker 7 7 5 9 9 4 5 13 21 16. Tug 11 6 2 8 14 9 5 9 9 6 Barge 3 1 4 2 3 2 1 3 5 6 Offshore 3 5 7 19 14 10 3 3 8 3 Fishing 59 66 66 43 75 83 105 109 112 80 Passenger 6 2 2 4 10 5 6 5 8 6 Ferry 7 8 3 6 5 17 9 4 13 7 Other 12 14 11 16 29 33 45 37 101 115 Table 2.2: Fatalities Aboard Canadian Registered Ships by Vessel Type (Source : [66, PP- 36]) Vessel Type 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 Cargo 5 14 11 6 10 4 6 7 2 2 OBO 2 0 0 0 0 0 0 0 1 0 Tanker 0 1 2 1 1 2 0 1 1 0 Tug 1 1 0 2 1 1 1 3 1 1 Barge 2 0 3 2 1 1 0 0 3 0 Offshore 1 2 0 1 2 0 0 0 0 0 Fishing 10 16 8 7 9 4 13 12 16 11 Passenger 2 1 1 0 3 1 0 0 1 2 Ferry 5 4 1 3 2 2 2 2 1 4 Other 4 2 1 0 3 2 1 0 0 1 Chapter 2. OCCUPATIONAL SAFETY ON-BOARD FISHING VESSELS Figure 2.3: Tasks performed on-board after taking the catch is in (from [8, p Chapter 2. OCCUPATIONAL SAFETY ON-BOARD FISHING VESSELS 15 Table 2.3: Injuries Aboard Canadian Registered Ships by Vessel Type (Source : [66, pp. 36]) Vessel Type 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 Cargo 35 49 59 58 61 43 58 78 80 88 O B O 1 3 0 5 0 0 2 0 1 2 Tanker 7 6 5 9 8 2 5 12 22 18 Tug 10 5 2 6 15 8 4 6 8 5 Barge 2 3 3 2 3 1 1 3 2 6 Offshore 2 4 7 19 16 10 3 3 8 3 Fishing 49 54 65 38 69 80 101 109 110 76 Passenger 7 1 1 5 8 4 6 5 7 4 Ferry 2 7 2 3 3 15 7 2 12 6 Other 11 12 10 18 30 32 46 38 105 116 Table 2.4: Based on Tables 2.1, 2.2, and 2.3, the percentages of aboard casualties occurred on Canadian Registered fishing vessels. Year 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 Accidents 40.41 41.51 40.00 24.72 33.48 39.71 43.39 40.91 31.20 24.69 Fatalities 31.25 39.02 29.63 31.82 28.13 23.53 56.52 48.00 61.54 52.38 Injuries 38.89 37.50 42.21 23.31 32.39 41.02 43.35 42.58 30.98 23.46 Tables 2.6, 2.7 and 2.8 examine the casualties in terms of the primary contributing factors [66]. In these tables, "human factor" corresponds to operational errors on the part of the people involved in fishing. "Environmental conditions" include atmospheric conditions, sea state and ice conditions. The general status of the vessel and its cargo are represented in the category of "vessel condition". " Unfortunately, the values in these tables are for all types of vessels given in Table 2.1, and not for fishing vessels only. 10 year averages of the contributing factors are given in Table 2.5. According to the tables, human factors, with 10 year averages of 61.58%, 70.45% and 59.02%, is the Chapter 2. OCCUPATIONAL SAFETY ON-BOARD FISHING VESSELS 16 Table 2.5: 10 year averages of percentage values of primary contributing factors in casu-alties aboard all vessel types Factor Accidents Fatalities Injuries Human factor 61.58% 70.45% 59.02% Environ, cond. 4.31% 3.61% 5.07% Vessel cond. 1.85% 2.77% 1.79% Equip./struc. 3.55% 3.55% 6.33% Other 0.38% 0.00% 0.46% Unknown 26.93% 19.60% 27.31% leading primary contributing factor in casualties aboard Canadian vessels. 2.1.2 France In [32], Dorval mentions a statistical report prepared in France that had disturbing facts for life at sea. According to the statistics, the probability of a fisherman dying at sea is 3%, while the probability of dying due to an occupational accident during his/her career is 50%. He reports that the situation is not much different for other European countries. Dorval in [32] reported the results of four trawlers that were examined in terms of occupational safety. In that paper the integration of safety and working conditions on-board was discussed in the following categories : movements, accommodation, fishing gear, work stations and noise. The author used this to suggest some improvements in each category. Firstly, the movements of crew on-board during different stages of fishing, such as during the hauling of the net, were observed, and how the present layout designs of each vessel affected these movements were reported. Accessibility to different work stations on-board and communication between crew members were the main areas of concern for occupational crew safety. Furthermore, it was suggested that in the design, human Chapter 2. OCCUPATIONAL SAFETY ON-BOARD FISHING VESSELS 17 Table 2.6: Number of Accidents Aboard Canadian Registered Ships by Primary Con-tributing Factor. All vessel types mentioned in Tables 2.1, 2.2, and 2.3 are included. (Source : [66, pp. 38]) Factor 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 Human factor 98 93 87 57 46 92 194 215 324 286 Environment al conditions 13 14 9 7 1 6 11 8 6 11 Vessel conditions 3 2 5 2 4 4 7 2 2 10 Equipment /Structural 11 13 18 13 5 9 7 6 9 4 Other 0 1 3 0 0 0 1 0 1 . 2 Unknown 21 36 43 95 168 98 22 33 17 11 Table 2.7: Fatalities Aboard Canadian Registered Ships by Primary Contributing Factor. All vessel types mentioned in Tables 2.1, 2.2, and 2.3 are included. (Source : [66, pp. 38]) Factor 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 Human factor 26 32 18 20 17 8 17 18 17 16 Environmental conditions 3 1 1 0 0 0 0 3 1 1 Vessel conditions 0 3 2 0 1 1 0 1 0 0 Equipment /Structural 1 1 4 0 1 0 1 0 2 0 Other 0 0 0 0 0 0 0 0 0 0 Unknown 2 4 2 2 13 8 5 3 6 4 Chapter 2. OCCUPATIONAL SAFETY ON-BOARD FISHING VESSELS 18 Table 2.8: Injuries Aboard Canadian Registered Ships by Primary Contributing Factor. All vessel types mentioned in Tables 2.1, 2.2, and 2.3 are included. (Source : [66, pp. 38]) Factor 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 Human factor 80 75 76 42 38 85 181 205 328 285 Environmental conditions 12 15 10 7 2 6 17 10 5 11 Vessel conditions 3 0 3 3 3 4 7 1 3 13 Equipment /Structural 12 20 19 17 6 9 7 9 7 4 Other 0 1 4 0 0 0 1 0 1 2 Unknown 19 33 42 94 164 91 20 31 11 9 factors' recommendations should be applied, such as the angle for a step ladder must be between 60° and 65°, unless there is an overriding need to do otherwise. Secondly, in the category of accommodation the suggestions for some of the trawlers examined were mainly related to the ventilation and living quarters having their own sink or washrooms. Thirdly source of potential accidents were arrangement and relative positions of the fishing gear and related systems, e.g. hoists and winches. The attention was drawn to the position of the fishing gear on the deck, (e.g at the bow or amidships), the way it is hauled or towed, the interference between the fishing gear and the other deck equipment. It has been suggested that fishing gear at the bow for stern trawlers (see Figure 2.4) is more advantageous than one amidships (see Figure 2.5). Fishing gear at the bow allows the net to be hauled on-board in a single operation, thereby reducing the physical effort required by the crew and the amount of time spent near the stern ramp. Forthly, on the fishing deck as a work platform, crew are exposed to the weather Chapter 2. OCCUPATIONAL SAFETY ON-BOARD FISHING VESSELS 19 conditions in all seasons, and depending on the sea conditions, spray and rushing sea water coming onto and clearing from the deck. Vessel motions are reported to be particularly important in terms of individual crew members' keeping their balance and the extra strain on the fishing gear. The broad suggestions listed for the general features of the work stations are : the working area should be uncluttered, have protection against falling overboard e.g. hand rail, have non-skid floors to minimize the probability of slipping and falling down; han-dling areas and work postures must follow the general ergonomic principles [32]. Fifthly, as far as the noise levels are concerned, two levels are mentioned. Firstly, a warning level , beyond which there is a danger of a hearing loss, is an exposure to 85 DB of steady noise for 8 hours a day and 40 hours a week. Secondly, when anyone is exposed to 90 db noise or greater, a significantly increased danger of a hearing loss and actions must be taken to change the exposure conditions. It was also noted that unlike land based work places, exposure to noise is continuous for the duration of a fishing trip, whether it is rest or working time. Their sample measurements in the 8 fishing vessels reveals that for one fishing trip: • 54% of the time the level of exposure was between 64 and 76 db • 39% of the time it was at 82 db • 7% of the time it was more than 90 DB with 3% of this time was above 100 db. Chapter 2. OCCUPATIONAL SAFETY ON-BOARD FISHING VESSELS 20 Fishing Net Winches Figure 2.4: Position of the fishing gear at the bow (from [32]). 2.1.3 United Kingdom Chaplin and Burney [26] reported the casualties in United Kingdom Fishing Fleet for the period of 1980 - 1988. Table 2.9 divides the data with respect to deaths due to vessel casualties and personal accidents, which might also be called occupational accidents. As far as the total numbers are concerned, there is not much difference in the two categories. The authors also divided the data with respect to vessel size, that is vessels less than 24 meters in length and vessels longer than 24 meters. The total number of deaths for vessels longer than 24 meters in both casualty categories (vessel casualties and personal accidents) are small compared to smaller vessels. In the category of deaths due to personal accidents, the authors examined the death rates by taking account of the number of men at risk. For vessels below 24 meters number of deaths per 1000 at risk is 0.41, which is almost one third of the rate 1.48 for vessels above 24 meters [26]. As far as non-fatal accidents are concerned, these authors reported that during 1986 Chapter 2. OCCUPATIONAL SAFETY ON-BOARD FISHING VESSELS 21 Winches Figure 2.5: Position of the fishing gear in the middle (from [32]). to 1988, rates were 4.7 men per 1000 for vessels between 10 to 24 meters, and about 20 men per 1000 for vessels greater than 24 meters. According to this, the probability of a non-fatal accident happening on a vessel larger than 24 meters is almost 4 times greater than the probability on a smaller vessel. However, these authors suspect that there is significant under-reporting of non-fatal accidents in vessels below 24 meters. This may unfairly bias the data against vessels above 24 meters. As for the causes of accidents, they report that between 50% and 60% of non-fatal accidents on vessels above 10 meters occur during handling or stowing of fishing gear, while 40% occur through different causes such as vessel movements, sea washing inboard or negligence. The authors caution about the distinction that an accident might occur while someone is working with fishing gear, yet the real reason might be the movement of the vessel. 2.1.4 Norway According to Kjerstad [41], the mean length of the Norwegian fishing fleet is 21.9roer.ers. Chapter 2. OCCUPATIONAL SAFETY ON-BOARD FISHING VESSELS 22 Table 2.9: Deaths in United Kingdom Fishing Vessels during 1980 - 1988 (Source : [26]) Casualty group Length < 24 [m] Length > 24 [m] Total Personnel accidents 70 29 99 Vessel casualties 87 7 94 Both groups 157 36 193 Table 2.10: Ratio of number of fishers with respect to the length of the vessels in the Norwegian fishing fleet [41] Vessel L e n g t h , meters R a t i o Less than 12.1 Between 12.1 and 24.3 Greater than 24.3 1/3 1/3 1/3 The distribution of number of vessels in different length ranges is given in Table 2.10. In another classification, the Norwegian fishing vessels are divided into three groups, namely inshore, coastal and deep sea fishing vessels. The vessels smaller than 8 meters in length, which are mainly inshore fishing vessels, are known to have very high roll frequencies (15 to 20 roll periods per minute [41]). The most conspicuous characteristics of the Norwegian fishing fleet is that in vessels with an L O A up to 15 meters, the main ergonomic concerns are due to lack of space on the working deck and little shelter in the work area, and the vessels' motions. However, as the sizes of the vessels increase, noise in the working-deck and the accommodation area cause some problems [41]. According to a statistical study of accidents on-board, the injury rate is mentioned as 58 per 1000 man-years in [36]. In relation to the age group, the probability that the youngest crewmen are injured is double the average. This is attributed to the less work experience. Chapter 2. OCCUPATIONAL SAFETY ON-BOARD FISHING VESSELS 23 Table 2.11: Contributing factors to on-board casualties in the Norwegian fishing fleet ([34] in [41]) Factor Number of occurrence Vessel motions 10 Slippery/icy deck 10 Water on deck 5 Slippery/icy ladder 4 Missing gangway 2 Low temperature 3 Vibration 2 Exposure to unshielded equipment 2 Heavy work load 5 Numerous working hours/lack of sleep 3 Monotonous routines 2 Bad weather 8 Other 19 From the investigation of accident reports, Hansen [36], and Kjerstad and Grinde [41] give the following as the most important the contributing factors to accidents: • Large vessel motions • Slippery/ icy decks • Water on deck • Slippery/icy ladder • Heavy work load In Table 2.11 from [41], some additional factors, as well as those listed above are given with their number of occurrences. Chapter 2. OCCUPATIONAL SAFETY ON-BOARD FISHING VESSELS 24 Table 2.12: The most important causes of injury in the Norwegian fishing fleet [36]. Values are % of the total disabilities or less serious injuries. Resulting in Cause Disability (%) Less serious injury (%) Falls 38 30 Crushing 36 20 Cuts, rips, pricks 16 15 Total 90 65 In [36], Hansen lists falls, crushing and cuts as the most important causal factors. Table 2.12 shows the types of injuries caused by these factors. According to the table 90% of the disabilities in fishing is caused by these three groups. For less serious injuries, they constitute 65% of the whole. Hansen considers large vessel motions to be a very important contributory factor in all casual groups mentioned in the table. The estimated total annual cost of accidents in fishing is approximately 32.3 millions Pounds according to Hansen [36]. The following presents the cost categories involved. Present value of human lives lost, injuries, and losses in production Medical and institutional Cost of founderings Cost of casualties Total annual cost Nkr. 187 Millions Nkr. 40 Millions Nkr. 48 Millions Nkr. 80 Millions Nkr. 355 Millions (Pounds 32.3 Million) Chapter 2. OCCUPATIONAL SAFETY ON-BOARD FISHING VESSELS 25 Although the above analysis does not distinguish explicitly between the cost asso-ciated with vessel casualties and the occupational/personal accidents, Hansen [36] es-timated the annual cost of each category to the society around 16 millions Pounds by assuming it is equally divided. In [41], Kjerstad and Grinde also focused on the effect of the working environment on accidents and health issues on-board vessels in the Norwegian fishing fleet. They highlighted the fact that high standards of living quarters on-board help to ease some of the stress during rest times. In their observations, since the off-duty living conditions on-board have reached a satisfactory high level in recently built vessels, more attention should be given to improve the working environment, e.g. the handling of gear and the catch. Kjerstad and Grinde [41] describe the working conditions on-board Norwegian fishing vessels as crowded with equipment, some of which are dangerous, in addition to the vessel's continuous motions. They also report a tendency, among the crew who work long hours, to complain of many ergonomic and mental strains. In their findings, the largest and newest vessels seem to have more comfortable work environments, in contrast to the smaller and older coastal fishing vessels, which are exposed to more severe working conditions, especially due to the long working hours and the excessive motions of the vessels. Also in these vessels the most frequently reported [41] factors are noise, strong vibrations and dangers while operating equipment on the deck. The sample measurements of noise levels reported in [1] (from [41]) are given in Table 2.13. According to the samples given in the table, in almost all of the areas, in which measurements were made, there is a concern about the high noise levels. After finding some complaints of the more strenuous working conditions on vessels that were 20 years and older, the authors compared the new and the old vessels in the fleet in terms of hard climate, use of out-of-date equipment, ergonomic strain and work stress. The conclusion of the authors was that the quality of working conditions were approximately the same, Chapter 2. OCCUPATIONAL SAFETY ON-BOARD FISHING VESSELS 26 Table 2.13: Survey on noise levels on 17 Norwegian fishing vessels ([1] from [41]) Area Full Fishing Max speed condition recommended (min, max, middle) (min, max, middle) level DB (A) Engine room 99, 110, 104 110 (unmanned) Bridge 64, 90, 79 58, 72, 68 65 Accommodation aft 65, 85, 78 65,70 Accommodation forward 70, 90, 76 68, 77, 72 60 - 65 Working deck 66, 90, 76 65, 81, 72 70 - 85 although exposure to excess motions in rough seas, too much vibration and impure air was less in the newer vessels compared to the older ones. The authors also compared the casualty statistics for different fishing types (see Table 2.14). Despite the high number of casualties, gillnet and longlining was not regarded as the most dangerous type of fishing due to the lighter levels of injuries. However, trawling is considered as a more dangerous fishing method as reflected in the serious nature of the injuries which occurred. The authors [41] claim that there would be a 10% reduction in the mortality rate if working conditions were improved for the 14% of the fishing crew who work in the most dangerous conditions. According to the interviews with the fishing crew in the Norwegian fishing fleet, the suggestions made to improve the working conditions on-board were [41]: • a reduction in the amount of time spent fishing, • reducing the work related stresses on-board. Chapter 2. OCCUPATIONAL SAFETY ON-BOARD FISHING VESSELS 27 Table 2.14: Number of casualties for different fishing types in the Norwegian fishing vessels ([12] in [41]) Type of gear Casualties per 1000 units Handline 18 Gillnet 26 Longline 27 Mechanized longline 69 Seine 14 Purse seine 34 Trawl 38 2.1.5 Spain A survey team from CETERA-Spain (Technical-Maritime Study Center) investigated and reported the living and working conditions on several Spanish fishing vessels [22]. According to the interviews done with approximately 400 Spanish fishermen, 82% com-plained about disturbance from ambient noise, 65% about exposure to vibrations and 42% about high risk of suffering accidents. The team also focused on : • the technical characteristics of the fishing vessels' design, • the layout and the type of equipment used on-board, • the work procedures on-board, • the living conditions on-board, • the existence and efficacy of safety measures on-board. The most noticeable deficiencies observed are given as follows: • High risk of falls into the sea and on deck, Chapter 2. OCCUPATIONAL SAFETY ON-BOARD FISHING VESSELS 28 • Excess effort due to arduous operations in confined spaces, • Deficient lighting , especially for high risk tasks, • High noise levels, • Uncomfortable accommodation, lacking in privacy, • Insufficient means of fire extinction. According to their survey the highest work accident rate corresponds to deck staff (see also Tables 2.15 and 2.16). This is explained, by first, a higher representation of drifter and trawler vessels in the vessels examined and the higher percentage of deck staff in the whole fishermen population. Secondly, the deck area is the most dangerous work place on the vessel. Table 2.17 shows the places where accidents occurred in Spanish fishing vessels. In parallel to the above, the deck area and fish bay are the places where the highest number of accidents happen. The engine room and gangway follow that as second and third respectively. In respect to the gangway, an area that includes the means of access to the vessel, the possible causes of accidents given by the team are as follows: • insufficient protection of steps or planks, • lack or scant of handrail, • dirty or slippery rungs, • large height difference between vessel and dock, • insufficient lighting of the gangway. Having established some statistical picture of the conditions on-board Spanish fishing vessel, the team studied the agents (e.g. robes, cables, etc.) which were involved in work Chapter 2. OCCUPATIONAL SAFETY ON-BOARD FISHING VESSELS 29 Table 2.15: % number of accidents on deck by type of crew member in Spanish fishing vessels [22] Category % number of injuries Sailors(working on fishing gear, handle the catch, etc.) 65.97 Skippers 20.14 Petty officers 6.94 Refrigeration 2.08 Deckhands 2.08 Officers 1.39 Mastmen 1.39 Table 2.16: % number of accidents by type of crew member in Spanish fishing vessels (Source [22]) Category % number of injuries Total deck Total machines Total catering 75. 19.79 5.21 accidents. The list of agents and the types of consequences are given in Table 2.18. Falls, three categories mentioned in Table 2.18, is the most numerous work accident in the Spanish fishing vessels. Accidents involving winches and blows against miscellaneous objects are among the other most common types. Some of the possible causes of the accidents are given : • Slippery / dirty decks or floors, • Failure to wear suitable foot wear with non-slip soles, • Movements of vessel due to the state of the sea, wave shock, etc., Chapter 2. OCCUPATIONAL SAFETY ON-BOARD FISHING VESSELS 30 Table 2.17: Places of accidents in Spanish fishing vessels [22] Place Number of accidents % number of accidents On board 192 97 Going to/coming from the vessel 6 3 Vessel at sea 157 84 Vessel in port 30 16 Deck and fish bay 108 69 Engine room 15 10 On gangplank 12 8 Holds 7 4 Steps 4 3 Freezer rooms 3 2 Accommodation 3 2 Bridge 3 2 Kitchen 2 1 • Objects, fixed or not fixed, which hinder free, movement and are capable of causing trip or slip, • Insufficient protection at edges of vessel, low height of gangway etc., • Insufficient protection from moving parts, • Failure to use protective gloves or other elements, • Lack of coordination between the person who controls, the winch and the one who carries out other tasks. The survey team [22] also reported that falls into sea are the most serious accidents, in which 15 accidents with 14 deaths and 1 crushing between vessel and dock occurred. Apparently, they occurred during the course of work on deck. As for the falls at different levels, there were some miscellaneous bruising and fractures reported. Chapter 2. OCCUPATIONAL SAFETY ON-BOARD FISHING VESSELS 31 Table 2.18: Work agents causing the most serious injuries in Spanish fishing vessels [22] Agent Amputation Crush Bruise Injuries Fracture Wound % Robes, cables 2 1 2 1 5.00 Boxes 2 1 2.50 Falls from height 6 7 10.83 Falls into sea 1 0.08 Falls same level 14 2 1 14.17 Cuts 5 4.17 Straps, belts 2 1 2.50 Struck by fixed objects 6 4 2 10.00 Struck by mobile items 5 4 1 8.33 Power tools 1 1 1.67 Hand tools 1 3 3.33 Winches 11 1 3 1 13.33 Haulage doors 2 1 1 3.33 Fish handling 1 0.08 Pulley, etc. 2 2 2 5.00 Doors 1 2 2.50 Chemical products 1 0.08 Check stoppers, break 1 0.08 Breakage: ropes, cables 2 2 1 4.17 Net 1 5 1 5.83 In the accidents involving winches, the possible reasons mentioned are: • lack of communication between the person in control and the others, • lack of sufficient protection from moving objects, e.g. cables, ropes, fish, etc., • failure to use protective gloves. Chapter 2. OCCUPATIONAL SAFETY ON-BOARD FISHING VESSELS 32 Main consequences in these accidents are amputations of fingers, broken fingers, bruising and wounds. In relation to blows against miscellaneous objects, the authors [22] report that in one third of the accidents, the effects could have been reduced if the corresponding protective elements, such as gloves and helmet had been used. Approximately, 5% of the total accidents occurred when dealing with nets or seines on deck during fishing. The common injury types in this group are amputations of fingers and fracture. 2.1.6 Japan In a different survey, Amagi and Kimura [5] analyzed the injuries to fishermen in scallop beam trawlers in Japan. According to their observations, the most common accidents are as follows : • Two dredge nets on the deck move and strike fisherman • Fishermen are caught between dredge and hull • Fishermen accidentally run into the hull, bulwark, hatch or dredge net. 2.2 Some notes on accident statistics Kose [42] also studied the fishing vessel casualties in the world. According to his analysis, human error was found to be the most important cause of the total loss of the vessel. It was then suggested that in order to reduce human error, motion and noise levels on-board fishing vessels should be reduced [42, page 107]. In most of the cases reviewed above, there was a general agreement that the work-ing conditions on-board fishing vessels needed improvement. The primary contributing Chapter 2. OCCUPATIONAL SAFETY ON-BOARD FISHING VESSELS 33 factors to the on-board accidents are vessel motions, slippery floors and ladders, noise, fatigue due to long and arduous working hours and insufficient work space. Vessel mo-tions are also believed to be a contributory factor in most of the accidents. In fact, in determining the how an accident occurred, there might be more than one contributing factor involving at the time. In the prevention of these types of accidents, some of the contributing factors, such as slippery floors, can be dealt with more easily than some others such as vessel motions. 2.3 Impact of Ship Motions on Fishing Vessel motions can impact on fishing in different ways. As reported by some authors above, vessel motions can contribute to accidents on-board. This means lose of life or injuries, and lost working hours. In another way, vessel motions interfere with fishing operations and causing it to be halted as explained in the next paragraph. In either way, Tupper [67] reported that it has undesirable consequences in monetary terms. In [67], Tupper examined the New England Groundfish industry. After studying the good weather fishing performance in terms of number of fishing trips and fish landings throughout one year, he reported a good correlation among the three. Based on this correlation, he concluded that the amount of catch per unit effort is approximately con-stant. However, dockside fish prices vary almost by 50% between a high during winter and a low during summer. This was explained by the varying number of fishing trips due to weather conditions. His conclusion was that due to deteriorated weather conditions, which also imply larger vessel motions, the New England fishing industry operates at 50% of its potential efficiency and effectiveness. Based on his interviews with fishermen and a survey on the motions' effect on fishing, he reported six ways in which vessel motions interfere with fishing: Chapter 2. OCCUPATIONAL SAFETY ON-BOARD FISHING VESSELS 34 • danger to vessel, • gear does not fish, • vessel can not stay on gear, • loose gear on deck poses threat, • water on deck poses threat, • motion impact on crew. Danger to the vessel, includes the danger of swamping, capsize, loss of directional control and structural damage due to slamming. Heave and pitch mainly affect fishing and "gear does not fish" because the net and doors experience a pulsing motion. This reduces the mouth of the net and disturbs fish before they go into the net, hence the catch rate is reduced dramatically. Loose gear on deck can cause a threat to the vessel, crew or itself. Similarly, water on deck can be a danger to the crew, fish that is being processed on deck, and to the gear. The effects of ship motions on crew are seasickness, slipping and sliding, jerking around and exhaustion from fighting motion. By the frequency of occurrence, Tupper ranked the above items in halting fishing. The most frequently reported reasons are • gear does not fish • fishing vessel can not stay on gear • jerking around • exhaustion from trying to counteract the motion of the boat. Chapter 2. OCCUPATIONAL SAFETY ON-BOARD FISHING VESSELS 35 Slips and slides, and water on deck are cited as moderately frequent reasons. Perhaps The motion interference with the crew could be in two ways. The immediate effects are to impair crew members' balances and thus they may have to simultaneously adjust their position. In the second case, the effect is cumulative and in the longer term showing up as fatigue and exhaustion. In general, fishermen are exposed to different types of body motions and accelerations due to ship motions in different sea conditions. The injuries related to the excessive motion can be grouped as sudden acceleration impact injuries and injuries due to motion sickness. The symptoms of motion sickness vary between individuals and depend on the envi-ronmental conditions, e.g. surrounding scenery, smell, etc. The development of symp-toms of motion sickness can take many minutes in ships in contrast to the exposure in laboratory conditions which may result in vomiting in a few seconds or in a few minutes To study motion sickness phenomena and its relation to motion characteristics O'Hanlon and McCauley [54] subjected a number of students to platform motion in a laboratory conditions. They were exposed to motion frequency in the range of 0.083 Hz to 0.5 Hz. and acceleration levels in the range of 0.03 g to 0.4 g, where g is the acceleration due to gravity. The results showed that the highest motion sickness incidents occur around 0.2 Hz. frequency and number of incidents increases with acceleration. They proposed a formula for predicting motion sickness incidence (MSI) as a function of the acceleration magnitude and frequency. surprisingly, danger to vessel was the least frequent reason listed. [35]. (2.1) Chapter 2. OCCUPATIONAL SAFETY ON-BOARD FISHING VESSELS 36 where MSI is the percentage of subjects in the experiment who experienced vomiting within two hours; x is a variable of a log-normal distribution with the following definition for p and cr. p = 0.659 + 3.840 l o g / + 2.467(log/) 2 cr = 0.40 log a p corresponds to the log a value associated with a 50% MSI value for a particular level. Lower p values indicate that less acceleration is required to produce the same MSI . a is given as a = 0 .637a m a x — 0.901a r m , where a r m 5 is the root mean square of acceleration in each half-wave cycle, and a m a x is the absolute value of peak acceleration. For motion sickness incidents, ISO 2631 part 3 ( International Organization for Stan-dardization, 1985c) suggests the magnitudes of vertical oscillation expected to produce a 10% incidence of motion sickness in sitting or standing over the 0.1 - 0.63 Hz . frequency range as follows: a t = constant where a is the rms (root mean square) acceleration and t is the exposure time. British Standards 6841 (British Standard Institution, 1987a) also defines a realizable frequency weighting (Wg) to be used for assessing low-frequency motion with respect to motion sickness. Realizable weighting implies greatest sensitivity in the range 0.16 - 0.2 Hz . Having defined some criteria for motion sickness, Aboulazm [2], and Andrew and Lloyd [6] proposed methods for designing ships with reduced number of motion sickness incidences. Chapter 2. OCCUPATIONAL SAFETY ON-BOARD FISHING VESSELS 37 In [2], Aboulazm describes the ship motion characteristics causing seasickness and presents the human tolerance, based on the human physiological factors, and the severity of ship motions. In a different study, Andrew and Lloyd [6] suggested a method of assessing the sub-jective motion magnitude (SM) of ship motion to be used in motion sickness evaluation as follows : SM = A 1.43 (2.2) where A defines a frequency weighting (A = 30 4- 13.53(ln / ) 2 ) , a is the amplitude of the vertical acceleration (m/s2) , g is the acceleration due to gravity (9.81ra/.s2), and f is frequency in Hz . To extend this relation to random motion, it is assumed that a = s(rri4.)1/2 and where 7714 and are the variance of the vertical acceleration and the rate of change respectively. Andrew and Lloyd [6] proposed a limiting value of 12 for S M for a 12 hour exposure. In this study discomfort were considered only in the frequency range of 0.25 Hz . and 4.0 Hz . When the criteria mentioned above for motion sickness are compared, Equation 2.1 (O'Hanlon and McCauley [54]) relates motion sickness incidence to the magnitude and frequency of the acceleration. However, it is mathematically involved compared to the other criteria mentioned above. The ISO's criterion seems simpler to use and relates the magnitude of the acceleration to the exposure time. It is interesting to note that the frequency ranges for greater sensitivity to motion sickness incidence are approximately the same for the three criteria (around 0.2 Hz for the criterion of O'Hanlon and McCauley, 0.1 to 0.315 Hz . for ISO criterion [35, p 308] and 0.16 to 0.2 Hz . for British criterion). Chapter 2. OCCUPATIONAL SAFETY ON-BOARD FISHING VESSELS 38 The criterion for motion sickness suggested by Andrew and Lloyd [6] is not intended to predict motion sickness incidence. However, it gives a measure of merit to evaluate subjective discomfort due to the oscillatory motion in 0.25 to 4.0 Hz . frequency range. For the other effects of ship motions i.e. other than motion sickness, Amagi and Kimura [5] suggested a method to reduce the number of human casualties in which ship motions are believed to be an important factor. In order to reduce the effect of ship motions the authors proposed a forebridge type vessel with the working deck astern of the bridge. In this way the effect of pitching on the working deck area is to be reduced. Kimura et al developed the following formula in [40] for evaluating the conditions on-board of fishing vessels in regard to maintaining one's balance. A T (threshold) value of 0.22, above which a crew member can no longer maintain his balance, was suggested after their experimental analysis. T is given by Equation 2.3. T = Mlx + 0.1659 Mly + 0.1133 Mlz (2.3) where M l is the first moment of power spectrum of the acceleration on deck, x,y and z subscripts represent the coordinate axis x,y and z. /•oo M l = / u S{u)dw (2.4) Jo In their formulation, the acceleration components due to g (gravity) are neglected. From their experience, they claim that for small angles of pitch and roll motions g sin9 and g cos^t are negligible. In a different study, Pingree [56] reported that for personnel performing tasks of a continuous nature the upper limits quoted for tolerable root mean square (rms) vertical acceleration as a proportion of g -acceleration due to gravity, is • 0.2 (at bridge) in U S A , Chapter 2. OCCUPATIONAL SAFETY ON-BOARD FISHING VESSELS 39 • 0.16 (at 0.2L) in Netherlands, • 0.18 (at 0.0L) in Germany, • 0.14 (weighted average) in the U.K., • 0.20 (at 0.2L from the forward perpendicular) in Canada. Furthermore, it was noted that there is an agreement of a seakeeping design criteria applicable to personnel for continuous exposure in rms values of 4.5° roll, 3.0° pitch and 0.18g vertical acceleration. The author also remarked that since many human motor processes are sensitive to the orientation in relation to the predominating motion, the personnel work stations should be designed with this in mind. The favorable orientation for an operator depends on the tasks being undertaken. 2.4 Design considerations for improved working conditions on-board based on warships Towards improved working and living environment on-board fishing vessels, there have been number of suggestions made in the literature. Tupper [67] categorized design considerations in three groups: • designing vessels with reduced motions • arrangement of tasks on-board with the motions of the vessel in mind (e.g. amidship is being the most preferable place on the deck in terms of vessel motions) • arrangement of details of body supports at the work stations. Reducing the vessel motions pitch and roll implies an increase in the effectiveness of fins, bilge keels, chines and paravanes [67]. Tupper draws attention to the complaints that Chapter 2. OCCUPATIONAL SAFETY ON-BOARD FISHING VESSELS 40 vessels with sharp chines that have too jerky a motion. He also suggests G M (metacentric height) could be another control parameter. Larger G M values are needed for the stability of the vessel. Also, larger G M s tend to reduce roll motion amplitude, thereby reducing the acceleration and jerk levels. However, care should be taken in that increased G M values also mean higher roll frequency, which leads to higher acceleration and jerk. As a note for simple sinusoidal periodic motion, acceleration and jerk amplitudes are (amplitude * frequency2) and (amplitude * frequency3) respectively. Another design parameter mentioned by the author is the vessels' windage. From a station keeping in rough weather stand point, windage should be kept down. Vessels with high bow and low stern might have problems in staying on course when working over the gear at low speed and in high winds. For arrangements where tasks are carried out, Tupper reports that the positioning the pilot house very high and close to bow is very common. Although from a motions point of view, such a place is one of the worst locations on board, it can give a good all-round visibility from the pilot house. For body supports at work stations, the author suggest that handhold/handrails and body supports require more attention. In designing the tasks, consideration should be given to reduce the amount of distance a crew member required to traverse on a pitching and/or rolling deck, and to eliminate the handling of heavy objects. Towards a safer ship in terms of occupational accidents, Stoop [63] suggested a node called "working conditions (safety) " be included into the design spiral. He analyzed some of the drawbacks in Dutch beam trawler design and gave some guidelines for what the properties of a so called "Beamer 2000", a newly designed beam trawler including safety aspects, should be. Similarly, in [8] types of tasks performed on-board fishing vessels were examined. Suggestions were made for a better working environment such as design of equipment etc, with respect to fishermen's safety. Chapter 2. OCCUPATIONAL SAFETY ON-BOARD FISHING VESSELS 41 In [55], Payne defined a method to quantify ride quality. He suggested that in the presence of a vehicle's motions with induced shocks or impulsive velocity changes, rms (root mean square) acceleration is not a good criteria to quantify ride comfort. Some useful evidence of experience in the design and operation of warships is given by Brown [18], who examined the adverse effects of ship motions on the degradation of performance. Due to bad weather, some of the consequences are course altering, reduction in speed, weapon operations becomes impossible and human performance degradation. For a given sea state, ship speed and heading, ship motions will vary depending on the angle between the ship's heading and the mean direction of the waves in the sea. Due to excessive ship motions, it may become necessary to alter the course of the ship. However, it may not always be possible as reported by Brown, for example during mine hunting. This kind of voluntary course altering is considered as a restriction of the operational capability of a warship. Another influence of bad weather is on speed. Generally, there will be involuntary speed reductions due to the increased resistance in waves, however, Brown reports that in most cases, there is a voluntary reduction in speed in order to avoid excessive motions, wetness, slamming and possible structural damage. The following are some of the limiting cases that he listed in [18]. • In a 15 minute period, the peak slam induced whipping acceleration is 0.09g at the bridge. • Subjective motion amplitude equals to 14. • The frequency of deck wetness is around one in every 40 seconds For helicopter operations in warships, Brown gives the following limiting values on some of the motion characteristics. • Roll (single amplitude) up to 3^° , unrestricted, impossible over 5° . Chapter 2. OCCUPATIONAL SAFETY ON-BOARD FISHING VESSELS 42 • Combined heave and pitch vertical velocity is unrestricted up to lm/s, and the limit is at 2m/s. • Maximum transverse velocity is lm/s. Brown also mentions two aspects of ship motions induced effects on humans, these are when tasks on-board become more difficult to perform and when the human performance degrades over the period of time. The first is related to the amplitudes and accelerations of the motions, while the second depends on the cumulative effects of fatigue, loss of sleep, and nausea. For a 3000 ton frigate, the loss of fighting effectiveness is 10% in a sea state of 5, 30% for a sea state of 6 and 95% for sea states of 7 and over. Brown [18, pp 44] estimates that "...the cost of bad weather to the assumed ship (Leander - 3000 tons, 108 m in length) at 1985 values is Sterling 1.29 Million per year". On the basis of model tests on a similar size hulls, he concludes that "...It is unlikely that changes in shape will lead to further improvement, and only increased size or a change to a novel configuration such as SWATH remains". Furthermore, he assumed that in a ship fitted with active roll fins, the loss of effec-tiveness will be mainly due to heave and pitch motions. He then compared a 108 m ship and a 125 m ship in terms of loss of effectiveness. In the example he gives, the 108 m ship loses 11.7 days a year, whereas 125 m ship losses 7 days a year. In monetary terms, choosing the larger ship (125 m), there will be 404,200.00 Sterling cash savings per annum, that also amounts to Sterling 4,000,000.00 discounted @ 5% over 20 years. The author argues that although the figure given for the savings has an error margin of ± 50%, there is a clear large savings possible. According the his estimations, the additional cost of opting for a larger, more spacious ship will be paid off in three years. Chapter 2. OCCUPATIONAL SAFETY ON-BOARD FISHING VESSELS 43 2.5 Methods for reduced motions Walden and Grundmann [68] suggested methods to be used in the early design stages in order to investigate seakeeping characteristics of frigate and destroyer type hull forms, and estimate freeboard. They improved the method of the seakeeping rank factor orig-inally given by Bales [11]. Bales' calculated seakeeping rank factor R as the average of the normalized rms values of eight ship motion related quantities, namely pitch, heave, relative motion at station 0 (forward perpendicular), absolute acceleration at station 0, slamming parameter at station 3 i.e. 15% of length from forward perpendicular, heave acceleration, absolute motion at station 20 (aft perpendicular), relative motion at station 20. Bales gives the estimation equation for the normalized hulls with respect to a 4300 ton hull as follows (also see [68]): R = 8.42 + 45.1CW + 10.1CW - 378- + 1.27-=- - 2Z.5CVpF - 15.9CVPA (2.5) L L In Equation 2.5 : R : Estimated seakeeping rank factor CWF '• Waterplane area coefficient forward of midships CWA '• Waterplane area coefficient aft of midships T : Mean draft L : Length c : Longitudinal location of cut-up, aft or forward perpendicular CVPF '• Vertical prismatic coefficient forward of midships CVPA '• Vertical prismatic coefficient aft of midships Chapter 2. OCCUPATIONAL SAFETY ON-BOARD FISHING VESSELS 44 Walden and Grundmann report that the parameter (c/L) in Equation 2.5 is difficult to define and has negligible influence on the estimation. They also computed a correlation matrix between the reciprocals of the 8 ship motion related quantities mentioned above and the calculated seakeeping estimation factor R (see Table 2.19). According to this table, except for the quantity corresponding to the relative motion at station 20, all other quantities show a good correlation with the R value. Table 2.19: Motion Correlations given by Walden and Grundmann [68] R 1/0 1/z 1/io c. 1/5 1/aao 1/^20 R 1 1/0 0.97 1 V * 0.96 0.94 1 1/i'o 0.84 0.75 0.81 1 l/*o 0.98 0.98 0.92 0.82 1 c. 0.68 0.62 0.56 0.42 0.67 1 1/5 0.98 0.96 0.98 0.79 0.95 0.62 1 1/-S20 0.92 0.87 0.84 0.86 0.90 0.47 0.87 1 1/^ *20 0.05 -0.11 -0.10 0.31 -0.01 -0.08 -0.07 0.34 1 In Table 2.19 : R : Calculated seakeeping rank factor 1/6 : Pitch 1/z : Heave at the longitudinal center of gravity l/r*o : Relative motion at the forward perpendicular 1/i'o : Absolute acceleration at the forward perpendicular Cs : Slamming parameter at station 3 1/5 : Heave acceleration at the longitudinal center of gravity Chapter 2. OCCUPATIONAL SAFETY ON-BOARD FISHING VESSELS 45 1/^ 20 I/V20 : Absolute motion at the aft perpendicular : Relative motion at the aft perpendicular 2.6 Design margins Generally, margins are used to increase the probability of success in an event or process with uncertainty in its outcome. However, there is also a cost associated in providing margins. Ship design and construction is no exception. Walden [68] mentions three cat-egories of margins in warship design: Design and Construction Margins, Future Growth Margins and Assurance Margins. Design and Construction Margins involve three categories of uncertainties : 1. Prediction errors associated with estimating techniques, 2. The unknowns during the prediction, 3. Anticipated minor changes in the design specifications during the design process. Future Growth Margins are defined as the allowances made upon the request of the customer in anticipation of future installations in a ship. These type of margins enhance the ship's flexibility for successful adaptation to future needs. Assurance Margins are accommodated in the design for three reasons: 1. In order to maintain a required level of operating capability even under certain adverse circumstances 2. To offset the progressive and predictable degradation of equipment and subsystems in a ship Chapter 2. OCCUPATIONAL SAFETY ON-BOARD FISHING VESSELS 46 3. To account for the uncertainties involved in the loading and the capabilities of the systems being designed. In order to reduce ship size and cost associated with the margins, the author explored the feasibility of reducing design margins. For the example naval ship studied by Walden and Grundmann [68], it is reported that weight and KG margins resulted in excess area and volume. They concluded that the Future Growth Margins can be excluded, and Assurance Margins nearly so, if the reductions in flexibility and performance can be accepted. On the other hand, Design Margins prove to be a more difficult case as the feasibility of producing a ship to meet design requirements is in question. Some of the Design Margins may be essential, and unless they are initially provided, the ship may have to be enlarged and the reiteration of the design may become necessary. Chapter 3 AN OVERVIEW OF SHIP DESIGN In the past each ship design was mostly based on one or more existing successful ships. That is if a ship was observed to complete the delivery of cargo and return safely to the home port she was considered to be a "successful ship". Shipyards or the designers then included the features of such a "successful ship" in their own designs. Over the years ships grew in size, new materials were used and new production meth-ods were developed and the installed power increased. During recent decades the devel-opments in "Advanced Marine Vehicles" brought new practical hull configurations such Multi-hulls, hydrofoils, S W A T H s , SES etc. The design procedure for such ships is now somewhat better understood and is aided by extensive use of computing systems and ap-plication programs. Once the geometry of a ship is known as a result of the preliminary design, there exist computer packages for commercial applications that can take the ship design from the lines drawing to the material ordering and assembly stages. That is the design and manufacturing processes are now computerized and interconnected. While the advances listed above took place, the physical laws pertaining to the design and development of ships remained the same. Similarly the preliminary design cycle of a ship consists of almost the same number of steps, almost in the same sequence. There is a mathematical reason for this. In general the equations representing the owner requirements and the physical laws for ship resistance, stability, seakeeping, cost etc. are nonlinear. That is the quantities that naval architects must establish, such as the length, beam etc. of the ship, influence ship stability and ship resistance in a complex 47 Chapter 3. AN OVERVIEW OF SHIP DESIGN 48 way usually involving at least powers of the principal dimensions. In particular the linear theoretical wave resistance depends on the square of the beam and in a very complex way on ship speed. Similarly the value of metacentric radius depends on the square of the beam. While it is desirable for efficient powering to reduce the beam, the stability requirements might dictate otherwise. Essentially a number of the equations a naval architect solves, one way or other, are nonlinear in terms of the basic ship particulars. There is, in general, no simple procedure to solve such a set of nonlinear equations. In view of the above, the general ship design procedure can be outlined as follows: A naval architect collects information from successful ships, designed in the past. This forms the basis of a trial or a first order solution. A method such as a perturbation procedure or a trial and error procedure suitable for the solution of nonlinear equations is then followed. The solution is expected to satisfy the owner requirements and the physical laws for flotation, stability etc. The nonlinear nature of the equations suggests that the existence or the uniqueness of the solution are never guaranteed. In fact the wide variety of ships that essentially satisfy the same set of user requirements is a good proof that solutions are not unique. In another real life example, a press extract (Meek [49]) related to the bidding of 3 very large ro-ro vessels for a consortium reports that the consortium received 10 different tenders that prompted a ship-owner from the consortium as saying "we have now got 10 technical solutions and we have to pick out the best points and get them into one grand specification". We can conjecture that there may be some room to increase the nodes in the design cycle or the number of equations in the design process. That is more conditions could be added as owner requirements or the number of equations in the set of equations to be solved can be increased. Of course the main questions are whether there is a need for this and what type of benefits one can have if more equations or conditions are added to the definition of the design problem. Chapter 3. AN OVERVIEW OF SHIP DESIGN 3.1 Overview of Literature for Ship Design 49 Height Groups Occupational Safety Economics iQuiDment Stability Resistance k Powering Seakeeping Figure 3.1: Design Spiral. Traditionally, the ship design process is represented by a design spired after Evans [33] in 1959. A similar ship design spiral to Evans's spiral is shown in Figure 3.1. The spiral representation depicts the very iterative nature of the process. Bremdal [16] describes the ship design process as a goal directed, iterative and creative process. The design starts with the mission requirements and proceeds through the nodes of the spiral towards a detail design. Buxton [21] in 1972 introduced economics into the design spiral. Chapter 3. AN OVERVIEW OF SHIP DESIGN 50 In his representation the spiral is divergent rather than convergent as shown in Figure 3.1. This shows the amount of increase in the knowledge acquired about the final design as the design process proceeds round the spiral. Later Andrew represented the ship design process as a helical corkscrew in an attempt to account for the open nature of the design process [7] in 1981. He incorporated into the spiral type approach some additional constraints related to the design process plus some constraints originating from the design environment. In these spiral type approaches, the characteristic of the ship design process is that it has a sequential and iterative nature. Mistree et al. [50] criticized the design spiral approaches mentioned above as the process of design they represented is sequential and difficult to incorporate life cycle considerations into it. Hence, they suggested so called decision-based design in 1990. They represented the design process by a funnel or the frustum of a cone. In their representation, the design process takes place inside this frustum in contrast to the other spiral approaches of [33], [21], and [7]. Each phase of the design such as preliminary design is depicted as an irregular disc-like shape inside the frustum (see Figure 3.2). In this representation, converging to a solution means the irregular shaped disc becomes circular. Weight Estimation / Stability / Powering Ring of interaction Figure 3.2: Representation of design process by Mistree et al. [50]. Chapter 3. AN OVERVIEW OF SHIP DESIGN 51 There are uncertainties involved in ship design. Some of them related to the environ-ment that the ship will operate in such as market conditions, sea conditions, etc. Another group of uncertainties involve the estimation of ship characteristics such as weight and costs. Sen [59] gives an overview of some of the techniques to take uncertainty into account in preliminary ship design. Ray et al. [57] used a probabilistic modeling of uncertainty. In their simulation for a bulk carrier, the principal uncertain factors that have the greatest influence on NPV (net present value) are the load factor and the freight rate. Speed of the vessel, cost of fuel oil, crew wages and port charges are, of course, important and influential on NPV. According to Taggart [64], there are four steps involved in the sequence of ship design. These four steps, embodied in the spiral, are conceptual design, preliminary design, contract design and detail design . The concept design includes technical feasibility studies and creation of alternative solutions, e.g. different ship arrangements. In this stage, some of the main ship charac-teristics such as length, beam and power of the candidates are determined approximately. In the preliminary design phase, these and other ship characteristics are further refined. Some of the ship parameters e.g. length and beam are not expected to change after the completion of the preliminary design. At the contract design stage further the preliminary design definition is refined and extended e.g. hull form is faired, power estimation is more precise perhaps based on model testing. Structural schema and some details are denned in this stage. Final general arrangement is then developed. In the detail design as the final stage of ship design, detailed work plans are produced. These include constructional details, installation plans, instruction to welders, outfitters, etc. One important feature of this stage is that at this point the final product of the design is passed to the production teams. Chapter 3. AN OVERVIEW OF SHIP DESIGN 52 In the past, there have been different methods developed for the preliminary design part of the ship design. One of them, Mistree [50], is called design-through-enumeration. All design variables are expressed in terms of vessel length and the length is varied until a feasible design is reached. Another approach mentioned in [50] is conceptualizing the preliminary design as an optimization problem. In this context, Murphy et al. [51], Nowacki et al. [53] and Smith and Woodhead [62] are cited as examples of preliminary ship design modeled and solved as a single objective optimization problem. For the mul-tiobjective formulation of the preliminary ship design, Lyon and Mistree [46] introduced some generic algorithms. Bower [14] developed a nonlinear optimization tool to be used during the preliminary design stage of fishing vessels for an operational scenario reflecting the conditions off the West Coast of Canada. Another ship design tool using optimization techniques is reported by Ivanov and Apollinariev [38]. They developed a program for the design of fishing vessels. They used steepest descent method for the optimization process. For the mathematical model, they used statistical information based on existing vessels. In the second part of the design package, the user is provided with a simulation tool for fishing vessel operations in order to evaluate the design further. Zanic in [71] introduced a different approach and developed a multi-attribute decision making system based on the random generation of nondominated solutions for fishing vessel design. However, this system requires fast computers. In fact, the program was run in a Prime EXL 7330 RISC workstation and during an example run for the design of a trawler, it is reported that a total of 159000 different alternatives were produced. Lee et al. [44] report an interactive computer system for the design of merchant ships. The program is integrated by using a relational database management system. It also has graphic user interface features. For a given set of owner requirements, it computes the Chapter 3. AN OVERVIEW OF SHIP DESIGN 53 principal particulars, plus the shape of midship section and hull form. Furthermore, it performs structural design, strength analysis and evaluates hydrodynamic performances. The implemented design process utilizes the information on a similar ship, which is selected by the designer from the available database, i.e. the design from a base ship. In the absence of a base ship, the program performs preliminary design calculations based on the empirical formulae provided in the code. Once the hull form definition is completed, it can perform structural design and analysis, vibration analysis (i.e. natural frequencies and mode shapes of hull girder vibration based on beam theory), resistance and powering based on the statistical analysis of a database consisting of 397 model test data, propeller design and performance analysis and maneuverability prediction. In parallel to the foregoing, there have been developments in artificial intelligence based paradigms. Some applications of these new developments applied to ship design are reported in [3], [30], [72], [24], [25] and [4]. The examples listed here are by no means exhaustive. Akagi and Fujita [3] developed an expert system for engineering design. The system is based on an object-oriented knowledge representation. In [3], they report a case study applied to the basic design of ships. The domain specific knowledge was implemented for cargo ships, bulk carriers, and container ships. Initially, the system uses a trial and error method to determine the appropriate ship parameters that satisfy the design re-quirements, e.g. owner requirements. If the requirements are not satisfied, the designer selects the design variables to be modified by the system. By making a unit change in the selected design variables, the system computes the amount of changes in all other parameters. According to the effect of the unit change, the designer then applies appro-priate amounts of change to the selected variables. In cases of failure to meet some of the design requirements and as an alternative to make unit changes they implemented optimization by sequential linear programming [3]. After a few trial and error searches Chapter 3. AN OVERVIEW OF SHIP DESIGN 54 for the solution, at the point reached they assume that the current position is in the neighbourhood of the solution. The gradient of the objective function (i.e. one of the design objectives such as the required hold capacity by the ship-owner), which has not been satisfied by then, is assumed to be not very steep in that region. Hence, in order to find a solution an optimization is performed. In their system, the design is implemented as shown in Figure 3.3. In [30], Daizhong and Forgie report another expert system developed for a general engineering design procedure. In their methodology, basic design units are used as build-ing blocks of different concepts. Then different concepts formed by the basic units are evaluated by allocating marks to concepts which can then be ranked. At this stage, the designer can modify existing concepts or generate new ones. Once a concept is selected, this stage of the design is followed by the detail design. In this stage, the design param-eters are determined. It is followed by initial layout design, individual component design and final layout design. For the layout design, they utilize some interactive graphical aids during the design. Zhou et al. [72] describe an artificial intelligence based system, named CLEER, developed to assist equipment arrangements on warships. The system is aimed to help improve electromagnetic compatibility among on-board systems and other equipments. It consists of the necessary databases of systems, equipments, etc. as well as knowledge bases for constraint definitions, information on existing arrangements on other warships and heuristic knowledge accumulated from previous experiences. Welsh et al. [69] developed an expert system environment for preliminary ship design. The system also has graphic display capabilities in order to visualize the final design product. Dai et al [29] describes a hybrid system integrating a knowledge-based system and a numerical optimization technique for ship propeller design so that propeller induced Chapter 3. AN OVERVIEW OF SHIP DESIGN 55 vibration is minimized. The design process is iterative and requires the experience of a propeller designer. In this hybrid system, the knowledge-based part of the system quickly produces a potentially good design point. Then the numerical optimization part of the hybrid system searches for the optimum in the neighbouring design space. The authors report reduced preliminary design time for propellers in comparison with other techniques. Calisal and McGreer [24] developed an expert system for the preliminary design of monohull type fishing vessels based on information on the series developed at the University of British Columbia (UBC series). Later, the same system was converted into a different Al environment based on constraint propagation [25]. In [4], a system for preliminary design of SWATH ships is reported. 3.2 Overview of the Ship Design Process (Design Spiral) Figure 3.1 on page 49 shows a design spiral. This figure is important in illustrating the iterative nature of the design, although more recent representations of ship design are mentioned earlier, a cork screw [7] and the frustum of a cone [50], which are also described as spiral type. Neither the number of nodes nor the order of nodes are standard in general. Each node of the spiral is associated with different aspects of the design, such as weight estimation or stability of the vessel. At each node in the spiral new information becomes available to the designer. As the design proceeds through the spiral, an abstract artifact, becomes more concrete. Most of the parameters that define a ship are nonlinearly related to each other. A change in one of the parameters may induce a snowball effect of variations in other parameters. In fact, this aspect of the design is well illustrated by Burcher and Rydill [20]. They compare a good design with a jigsaw, in which all the pieces are arranged to form the whole picture. If one of the Chapter 3. AN OVERVIEW OF SHIP DESIGN 56 pieces is altered in shape some of the adjoining pieces have to be modified in order it to complete the whole picture again. These consequential modifications may cause more pieces to be displaced, perhaps until almost all of the pieces may undergo modifications to generate a picture again. The amount of the modifications will depend upon the type of picture as well as the initial piece itself. Likewise, inappropriate choices made in previous design stages may have undesirable consequences in terms of cost of the design. Furthermore, freedom to make a choice is reduced as the designed artifact, in this case a ship, is realized. Design teams include specialists in different fields such as structures, hydrodynamics, fish handling, bridge design, electronics, etc. The communication or technical manage-ment aspect of the process becomes very important. The problem of conflicting design requirements especially in the detail design is a very familiar problem to most designers. From the stability point of view a beamer vessel might be desirable, while resistance and human comfort considerations may dictate otherwise. It is also not uncommon to observe the conflict between a structural designer trying to keep the hull structure as continuous as possible, while piping and electrical wiring designers request holes and openings in the design. The detection of such design conflicts at an early stage will reduce the cost of production and ensure a better designed vessel. 3 . 3 Summary The preceding sections provided an overview of some of the characteristics of ship design. The following points can be made. • There is no unique way of designing ships. Historically, ship design has been the reproduction of successful ships. One way of designing ships has been to use a very similar ship as a basic ship/design and then to apply some modifications to the Chapter 3. AN OVERVIEW OF SHIP DESIGN 57 basic design in order to satisfy the current design requirements. In this category, ship designs are based on stereotyped arrangements of compartments, disposition of equipments to form operational systems, etc. These are generally resized. Another way is to use the information on existing ships in the form of empirical formulae. This way is more involved compared to the first one. In the case of a relatively new ship type such as a SWATH, the design is initially even more involved because of the lack of accumulated design experience in the new concept. On the other hand, SWATH type of configuration provides new opportunities for new internal arrangements, deck layouts, etc. • There is no unique solution for a set of given design requirements. Some of the parameters involved in ship design are nonlinearly related to each other in general. Therefore, not only a solution to the ship design problem in general but also, if there is any solution, its uniqueness is not guaranteed. • Ship design is a nonlinear and iterative process. There are conflicting requirements involved in ship design. Earlier detection of these conflicting requirements will reduce the cost of design. Furthermore, it may even prevent undesirable costs induced during production. Chapter 3. AN OVERVIEW OF SHIP DESIGN 58 Ship Owner's Requirements Determination of Principal Particulars Estimation of Ship's Performance Selection of Main Engine Lines I Hull Resistance and Propulsion i General Arrangement and Capacity Plan \ Hydrostat ic Curves 1 Trim and Stability Simulation I Loading Plan Figure 3.3: The process of basic design of ships in Akagi and Fujita's expert system (from [3])-Chapter 4 DESIGN, KNOWLEDGE-BASED SYSTEMS AND ECHIDNA Brown and Chandrasekaran [17, p 1] compare artificial intelligence (Al) type computa-tions (e.g. expert systems) with traditional computational methods such as finite element stress analysis programs. They propose a view to characterize intelligent behavior such that "a collection of general strategies that use knowledge in such a way that the complex-ity of computation inherent in certain tasks is minimized" [17, p 3]. Furthermore, they use the definition (originally described by Newell [52] as the Problem Space Hypothesis) to distinguish some types of intelligent algorithms (methods), "... explore a problem space, implicitly defined by a problem representation, using general search strategies which exploit typically qualitative heuristic knowledge about the problem domain" [17, p 4]. For example, an algorithm designed to find the greatest common denominator is not qualified as an intelligent algorithm in the above sense based on the definition given. The authors [17] are of the opinion that Al based techniques such as expert systems are more appropriate when the underlying solution spaces are very large and solution algorithms of restricted complexity are not available as in diagnosis and design problems in general. In a broad sense, Taylor [65] defines expert systems as "the collected rules of thumb-human experience in a computer". Another definition is given by Buchanan [19] (in [27]), which is that "knowledge-based systems are computer systems in which operable human knowledge about some domain is captured and rendered". One of the characteristics of a knowledge-based system is that the knowledge it contains must be explicit in such a way that it can be inspected and understood independently of the way in which it 59 Chapter 4. DESIGN, KNOWLEDGE-BASED SYSTEMS AND ECHIDNA 60 is controlled. Knowledge exists as a discrete entity and can be processed in different ways [27, p 35]. The goal in knowledge-based systems is to represent knowledge in such a way that it is comprehensible to both humans and computers. A knowledge-based system is different from a traditional computer program as generally understood, in which the domain knowledge under consideration is coalesced with control statements such as variable declarations, loops, conditionals, etc. According to Bratko the distinction between an expert system and a knowledge-based system is that an expert system has the capability of explaining its decisions and the underlying reasoning [15, p 332]. One can describe expert systems (or knowledge-based systems in general) as merely procedural programs with a sufficient number of if-then-else structures. Ideally, one answer to such a criticism based on the definition given above is that an expert system differs from a if-then-else procedural program in that the control structure is separated from the knowledge base. Modifying the knowledge base does not affect the control structure. However, once the domain specific knowledge representation is formalized it may, in most - if not all, cases require some recoding so as to be able to use the same representation in another shell. Expert systems as new computer tools will eventually change the role of computers. One of the outcomes according to Mistree : "designers . . . will be able to use a computer not just as a tool, but as an advisor, a critic, and ultimately as a partner in the process of design". This will also affect the role of designers. "They will be involved primarily with the unstructured or partially structured parts of problems (that is, with establishing system goals, partitioning the system in terms of its functional subsystems and planning the design process itself) rather than with the structured part (that is the design of components of the decision process) which will be automated" [50]. Chapter 4. DESIGN, KNOWLEDGE-BASED SYSTEMS AND ECHIDNA 61 4.1 Design in a Knowledge-Based Environment According to Brown [17], a design problem is a search problem in a very large space [17]. Fundamentally, each element of the space could be a possible candidate for the design problem under consideration. Simon [61, p 151] suggests that in the design of complex systems it is important to represent the highly desirable variants in the design process rather than to create systems that will optimize some hypothesized utility function. He argues that in real life examples, it is usually the case to find a satisfactory solution ,as Simon calls satisficing, rather than to choose between satisfactory and optimal solutions as rarely a method of finding the optimum is available. Artificial intelligence based tools may offer a useful tool for the design of complex systems. In the following paragraphs, first a knowledge-based model of design will be introduced. Then, Echidna, the knowledge-based system used in this study will be described. Coyne et al. [27] consider three important concepts in a knowledge-based model of design: • Representation - how information is represented in a computer. Representation involves facts, knowledge and control. In this context, knowledge is defined as statements about mappings between facts. Then, control is characterized in terms of mappings between items of knowledge. • Reasoning - involves what design reasoning is about. Deduction, induction and abduction (see the table below, from [27]) could be named as different modes of reasoning. Deduction (fact, knowledge and inferred fact) is associated with interpretation of databases to derive attributes of the design not explicitly defined in the description. For example, given a ship we might be interested in knowing the resistance of the ship in certain speeds (performance issue). However, the designers' primary aim is the production of design descriptions (e.g. ship particulars e.g. Chapter 4. DESIGN, KNOWLEDGE-BASED SYSTEMS AND ECHIDNA 62 principal dimensions, hull geometry, etc.). Unlike the previous case, they start with some rules and performances to arrive at a description. This is considered analogous to abduction. Coyne et al [27] give the following example for abduction. If the rule "all houses are buildings" and the conclusion "this is a building" are known, but the original statement "this is a house" is not known, the step that one might decide "this is a house" is called abduction. Deduction : case + rule —> result Induction : case + result —» rule Abduction : rule -f result —> case • Syntax - the role of syntactic knowledge. Analogy is made to the natural language, i.e. vocabulary and grammar. An example set for vocabulary elements for a ship might be bridge, engine, propeller, shaft, hull, etc. Grammar then provides a formed way to represent knowledge about composition, i.e. how these elements are put together. The authors categorize two separable tasks in the design process: • Interpretation - the mapping between the design descriptions and their perfor-mances. • Generation - the composition of designs Coyne et al [27] suggested a linguistic model of design. In the analogy, vocabulary of a language i.e. words, related to the elements that compose a design. Hence, grammar (rules of syntax) in language helps determine whether a sentence is legal. It could also be expressed as actions for constructing a sentence. The analogies they considered are given in the following table ( [27]). Chapter 4. DESIGN, KNOWLEDGE-BASED SYSTEMS AND ECHIDNA 63 Language Design Vocabulary : words parts Syntax : grammar actions for configuration Utterances : sentences designs Semantics : meaning interpretation of designs as performances, e.g. propulsive performance They characterize design as a search within a space defined by the knowledge con-cerned with interpretation and design syntax (generation). Given a set of performances, the aim of designer is to produce designs that will be within the intersection of the two design spaces (see Figure 4.1). Design as process (or activity) is concerned with the definition of the spaces being searched and the search process itself. 4.2 Echidna Expert System Shell Echidna is an expert system developed at Simon Fraser University's Expert Systems Laboratory. It is briefly described as a logic programming language embedded in an object-oriented framework [60]. It supports the following features: 1. Constraint-based reasoning - In general, any choice made for some aspect of the design constrains the choices available for remaining aspects of the design. 2. Model-based reasoning - Models of sub-systems and components play an important role in the design of an artifact. The artifact can be seen as a complex system with an inherent hierarchical structure: The top-level represents the entire system, the bottom level represents the parts and pieces which make up the various components. Chapter 4. DESIGN, KNOWLEDGE-BASED SYSTEMS AND ECHIDNA 64 Space of interpretations Generative Interpretive knowledge knowledge Generation Abduction Spaces of designs Figure 4.1: Generative and interpretive knowledge in defining spaces of designs (from [27]). Instead of representing the top level object as a mere collection of bottom level objects, it is more useful to model the subsystems and components as such: Each model describes the relationships between the input and output parameters of the object modeled. A corresponding representation must be supported in a powerful design system. 3. Logic programming - Logic programming allows a declarative representation of the task. Declarative representations are common in design. For example, the declara-tion that "the preferred length to beam ratio for the UBC Series fishing vessels is between 2.6 and 4.0". Declarative representations are concerned with the relations Chapter 4. DESIGN, KNOWLEDGE-BASED SYSTEMS AND ECHIDNA 65 denned in the program, hence the outcome. Whereas the procedural aspects of a program are concerned with how these relations are evaluated by the program. In other words the two. approaches deal with the two different questions respectively : " What are the relations/goals in the program, and what is the output?\ and "How is the output obtained? ". The programmer , in the former, is encouraged not to think about the executional details and to concentrate on the meaning of the program. Therefore, this declarative approach generally makes the programming easier [15, pp. 25 - 26]. The Echidna expert system incorporates a new type of constraint logic programming. Traditional logic programming systems allow only the representation of one special type of constraint: equality. For example, a tradi-tional system may return a result such as "length = 5 OR length = 6 OR length = 7". A more powerful system also allows the representation of more general types of constraints such as inequalities. Here a representation corresponding to the above result could be "5 < length < 7". Echidna permits the definition of constraints such as ranges, and weight functions for vessel parameters such as length, beam, and hold volume. The system is very suitable for design tasks in general since the built-in constraint processing easily facilitates the narrowing of parameters during the progress of the design. For example, constraints such as "5.2 < length < 20.1" and "12.4 < length < 30.0" could appear at different times during processing and appropriately combined to "12.4 < length < 20.1". 4. Hypothetical reasoning - Design tasks require search. Partial designs must be first proposed, then integrated into the evolving solution and their suitability evalu-ated under design constraints. The system must be able to investigate proposed solutions, and return to an earlier solution if a proposal proves infeasible. The inte-gration of partial designs is controlled by rules which specify both design strategies Chapter 4. DESIGN, KNOWLEDGE-BASED SYSTEMS AND ECHIDNA 66 preferred by expert designers and constraints inherent in the model. 5. Intelligent backtracking - This is essentially an issue of the performance of Echidna. In the traditional chronological backtracking approach, the system searches for a solution to the (design-) problem and upon failure backtracks to the most recent choice point and selects a different alternative. This alternative may not resolve the problem, i.e., the failure persists and the process repeats until no more alternatives are available at that choice point. The system then chronologically backtracks to the next-most recent choice point and selects a different alternative there, etc. The choice that causes a failure is often made many steps before the failure can actually be detected. Chronological backtracking results in an unacceptable amount of obsolete processing and is inadequate for any problem of realistic complexity. To make the product commercially viable, intelligent backtracking is essential. Here the system keeps track of the dependencies between choices and their effects. Thus, when a failure occurs, all most-recent choices that are irrelevant to the particular failure are skipped and the system backtracks directly to a choice point that is likely to resolve the problem. 6. Mixed initiative - This is required to support a cooperative problem solving between the expert system and the human designer. It would be unrealistic to expect that all design tasks and subtasks can be properly identified, analyzed and formally represented in an efficient way. Therefore any adequate system will aim at assisting the user in solving the design problem, not at replacing him. A peer relationship is essential in such an approach: The user must be able to take the initiative and to interact with the system in order to actively take part in the (computer-assisted) design process. Likewise the system must be able to take the initiative, for example, if the user breaks important design rules and needs to be warned. The integration of Chapter 4. DESIGN, KNOWLEDGE-BASED SYSTEMS AND ECHIDNA 67 constraint processing and intelligent backtracking within Echidna has the potential for a more efficient interactive design system than traditional rule-based expert system shells. The Echidna expert system also takes advantage of object-oriented programming to easily model a design. Objects are used to represent both physical objects, such as the vessel itself and its components (such as hold volume), as well as intelligent agents, each with domain knowledge in a particular aspect of design (such as vessel stability, crew safety, and comfort). Other design applications for the Echidna Expert System shell include heat exchanger design, Stirling engine design, and residential log home design. In the development of the Echidna expert system at UBC for fishing vessel design our interest is to use the Echidna Expert System for all of its features listed above. Computer-aided fishing vessel design and small boat design need all of the facilities supported by the Echidna system. The purpose of the application of the Echidna expert system is to synthesize the main characteristics of a marine system, that is able to float with sufficient stability in calm water, determine the resistance of the vessel so that a sufficient propulsive system required to go fishing, search, fish and return in possibly rough seas. Furthermore, prediction of ship motions and assessment of sea-kindliness in the preliminary design stage is fully incorporated into the Echidna Expert System. Chapter 5 MONOHULL VESSEL DESIGN KNOWLEDGE BASE (UBC-MONO) The design knowledge representation in general is discussed in Chapter 4. In this chapter, the knowledge base developed for monohull fishing vessel design is described. The design algorithms used in it are given in Appendix B. The approach employed in UBC-MONO preliminary monohull fishing vessel design knowledge base is similar to the design spiral. Perhaps, one of the reasons for this is the nature of the process. Each node on the spiral requires some of the information obtained in the previously visited nodes. Hence, in order to help the program reach a solution faster, the same model of the design process has been adopted here. First of all, a schema1 "Monohull" representing monohull vessels has been written. This schema contains all the necessary information related to the vessel characteristics, from geometrical dimensions to its resistance characteristics, etc. However, the code that is necessary to compute these characteristics have been coded in different schemas. One can consider these schemas as task performing schemas. The tasks are basically to estimate vessel characteristics at different stages of the design. For example, this could be the estimation of design waterline length based on some other parameters. Analogous to the design spiral, the task performing schemas are for the estimation of geometric dimensions, weight groups, stability checking, resistance, required horse power and cost estimation. Additionally, there are also schemas to estimate some of the motion characteristics (related to heave and pitch amplitudes, see Chapter 7) of the current •"^ See Appendix A for the definition of a schema. 68 Chapter 5. MONOHULL VESSEL DESIGN KNOWLEDGE BASE (UBC-MONO) 69 design. The above approach enables the user to do partial designs, as well as evaluating existing ones. By performing partial designs means taking the design process up to a certain node in the design spiral, or running only a desired schema , e.g. schema for power estimation, on a given schema "Monohull". Another important aspect of the Echidna knowledge based system, and for that mat-ter knowledge based systems in general, is that alternative design methods can co-exist in the system. For example, in [47, p 18] the overall length of a SWATH vessel is given in Equation C . l . Macgregor [47] reported that although the coefficient of the equation was obtained as "5.33" after a regression analysis, there was a ± 2 variance in the coeffi-cient. In fact, the ratio of LOA/A* in the existing, as-built designs varied in the range of [3.33,7.33]. Hence, in the Echidna knowledge base, it is possible to have two alternative design rules in the estimation of LOA-orderchooaeLengthOverAll. (5-1) chooseLengthOverAll(LoA) '• — LOA — 5.33 * A s . (5.2) chooseLengthOverAll(LoA) : - L0A = [3.33,7.33] * A * . (5.3) In the above design methods, the first line ensures that Echidna will choose the methods in the order of appearance. In other words, it will first attempt to create some designs where LOA/A' ratio satisfies Equation 5.2. In case no solution is possible by using this method of estimation, it will then try the alternative, i.e. Equation 5.3. If the line 5.1 is omitted, Echidna randomly chooses a method. Therefore, this feature of Echidna allows implementation of the kind of actions to be taken in case of a failure to produce a design. Chapter 5. MONOHULL VESSEL DESIGN KNOWLEDGE BASE (UBC-MONO) 70 5.1 Remarks This section reports the results of the investigation on the validity of the knowledge base for preliminary monohull fishing vessel design. The aim of the investigation was to compare the performance of Echidna with another knowledge-based system called PC Plus Consultant, in which the same fishing vessel design algorithms, excluding seakeeping considerations, were implemented by Calisal and McGreer [24]. Furthermore, an existing vessel, namely Kynoc, was redesigned in Echidna system and the outcome was compared with the original Kynoc. It should be noted that in obtaining the results given below, no ergonomic design considerations (i.e. ship motion criteria for crew safety and comfort) were included in the knowledge base. In the investigation, there were two main points : • Echidna was required to demonstrate a similar trend, which is reported by Calisal and McGeer in [24] (Figure 5), for the comparison of steel and aluminum vessels. ("Can Echidna reproduce similar results obtained in another design environment and especially by conventional design practices?1'') • After being provided with sufficient information related to an existing fishing ves-sel, Echidna was required to conclude if this was a valid design for the Echidna knowledge base as well. ("Can Echidna reproduce an existing designf) A valid design in Echidna is a set of parameters that characterize a vessel, and each of which Echidna is able to assign a value such that all of the constraints of the fishing vessel design in the knowledge base are satisfied. The most important constraint in the knowledge base is that displacement of the vessel should be equal (or greater than by a predetermined amount) to the total weight of the vessel, i.e. Archimedes's principle. Some of the other main constraints worth mentioning are the vessel should have sufficient Chapter 5. MONOHULL VESSEL DESIGN KNOWLEDGE BASE (UBC-MONO) 71 Table 5.1: Input to Echidna for Kynoc. "[ ]" indicates an input as an interval. Parameter Input Value Fishing method seiner Hull material aluminum Design speed 10.3 [knots] Displacement [80.0, 85.0] [LT] Length 56.7 [ft] Beam 20.0 [ft] Draft 4.92 [ft] Block coeff. 0.519 Mid ship coeff. 0.728 Hold volume [23.0, 27.0] [LT] initial stability, power and cost limitations. For the first part of the investigation, an existing monohull seiner type vessel, namely Kynoc, was chosen. Kynoc operates off the Pacific coast of Canada. Table 5.1 shows the main particulars of Kynoc used as input to Echidna. The results of the comparison are presented in Table 5.2, which proves that Kynoc is a valid design in Echidna's knowledge base. This conclusion stems from the fact that Echidna was able to produce a design and this design is very close to Kynoc. The Echidna version of Kynoc satisfies the constraints of the ship design, which were implemented in the knowledge base, i.e., the vessel floats, has sufficient initial stability, and the power requirements sire comparable between the existing one and the Echidna version as shown in Table 5.2. In the second phase of the knowledge base validation, in order to reproduce the trend reported by Calisal and McGeer in [24] (Figure 5), a number of steel and aluminum vessels have been designed for the same hold volume and other input parameters. Hold volume has been varied between 5 [LT] and 300 [LT]. For this investigation the input to Echidna is given in Table 5.5. During this phase, the intervals given as solutions by Chapter 5. MONOHULL VESSEL DESIGN KNOWLEDGE BASE (UBC-MONO) 72 Table 5.2: Comparison of Echidna produced design with real Kynoc. Parameter Kynoc Echidna Fishing method seiner seiner Hull material aluminum aluminum Design speed 10.3 [knots] 10.3 [knots] Displacement 82.0 [LT] [82.52, 85.0] [LT] Length 56.7 [ft] 56.7 [ft] Beam 20.0 [ft] 20.0 [ft] Draft 4.92 [ft] 4.92 [ft] Depth N/A [5.08, 6.02] [ft] Block coeff. 0.519 0.519 Mid ship coefF. 0.728 0.728 Prismatic coeff. 0.712 [0.711, 0.717] Hold volume 25.0 [LT] [23.0, 25.5] [LT] Hull weight N/A [12.38, 14.86] [LT] Machinery weight N/A [3.09, 3.91] [LT] Engine 420 [HP] [134.31, 378.35] [HP] Hull Cost N/A [137042, 164440] [Can.$] Machinery Cost N/A [29842, 84062] [Can.$] Total Cost N/A [1106250, 1275000] [Can.$] Echidna were sometimes very large, therefore, Echidna was forced to select arbitrarily smaller intervals (subintervals) from these initial solutions. In order to clarify this point, a hypothetical example is given as follows: For the sake of simplicity, assume that there is only one relationship among the parameters given in Table 5.3, that is Equation 5.1. Table 5.3 shows the intervals given as a solution. Both in Table 5.3 and Table 5.4, the relationship given above holds for the intervals given. Displacement volume = Block coefficient x Length x Beam x Draft Chapter 5. MONOHULL VESSEL DESIGN KNOWLEDGE BASE (UBC-MONO) 73 Table 5.3: Example for the intervals after Echidna reaches a solution. Parameter Solution intervals Length Beam Draft Block coefficient Displacement volume [60.0, 65.0] [20.0, 25.0] [6.0, 9.0] [0.61, 0.62] [4392, 9067.5] Table 5.4: Example for the parameters' intervals after Echidna completes design and further arbitrarily refines them. Theoretically speaking, the intervals given in Table 5.3 may contain an infinite number of solutions. However, any randomly chosen Displacement volume, Block coefficient, Length, Beam, and Draft values from the intervals in Table 5.3 may not necessarily construct a solution, since a solution implies that the new sub-intervals (e.g. the ones given in Table 5.4), should satisfy Equation 5.1. Similarly, for the second phase of the validation process, the solutions given below (Figures 5.1 to 5.7) are arbitrarily selected by Echidna from the original, much larger intervals given as solutions. A given interval as a solution, may contain an infinite number of smaller intervals (or discreet values) that might correspond a solution. However, the design constraints (rules) will determine sets of smaller intervals that form a solution. Parameter A solution Another solution Length [60, 60.01953] [60, 60.01953] Beam [20, 20.01953] [24.49219, 24.51172] Draft [6, 6.011719] [6, 6.011719] Block coefficient [0.61, 0.6100391] [0.61, 0.6100391] Displacement volume [4392, 4406.697] [5378.381, 5395.503] Chapter 5. MONOHULL VESSEL DESIGN KNOWLEDGE BASE (UBC-MONO) 74 Table 5.5: Input to Echidna for the second phase of validation for a fleet of varying hold capacities. Parameter Input Value Fishing method seiner Hull material aluminum and steel Design speed 10.3 [knots] Hold volume from 5 [LT] to 300 [LT] The effects of different hull material type on some of the vessel parameters as well as the total cost of the vessel are presented in Figures 5.1 to 5.7. Figure 5.1 illustrates the change in the length of the designs due to different hull material. The figure shows no particular difference in the length of steel and aluminum fishing vessels, except in the region of 25 [LT] hold volume capacity. This discontinuity could be associated with the alternative estimation rules incorporated in the Echidna knowledge base. Methods used to estimate initial ship dimensions order length. (5.5) Method I : Length = 3.3(28.5 Hold-volume)03669 (5.6) Method II : Length = [1.0, 200] J^engthlnFeet (5.7) The examination of the length estimation methods used in the knowledge base (Equa-tions 5.6 to 5.7), reveals that there are two distinct methods to estimate length. The first one, i.e."Method I", is based on the hold volume capacity required for the vessel. This formula has been derived from the information on existing designs [24]. "Method II" merely defines an initial interval in which any solution, if it exists, should occur. In this case, the other constraints of ship design will determine the subintervals that include a solution. The effect of preferring one method for a given hold capacity requirement Chapter 5. MONOHULL VESSEL DESIGN KNOWLEDGE BASE (UBC-MONO) 75 and the other method for a similar hold capacity requirement during the design can be observed by an abrupt change in the trend in the figures presented. As for the beams of the designs, steel and aluminum designs have almost the same values in general. However, in the regions of 25 [LT] and from 45 [LT] to 80 [LT] hold vol-ume capacity (see Figure 5.2), steel vessels have up to 20% larger beams than aluminum ones. The jump between 50 to 75 [LT] hold capacities in the figure, could be related to the alternative estimation rules used for a design variable in the knowledge base. The effects of material type on draft is shown in Figure 5.3. In the whole hold capacity range considered, aluminum vessels have shallower drafts than steel vessels. This is in confirmation with the present picture of existing fishing vessels. In Figure 5.4, the hull weights of the aluminum and steel vessels are compared. As the figure depicts, steel vessels have heavier hulls throughout the hold capacity range under consideration. In order to compare different vessels in terms of their sizes, displacements of the vessels are good indicators. The results obtained in this investigation suggest as expected that steel vessels are heavier (or larger) than aluminum vessels designed for the same design speed and hold capacity (see Figure 5.6). This result can be observed among existing vessels as well. In the investigation, power requirements of steel and aluminum vessels have also been compared (see Figure 5.5). Up to 75 [LT] of hold volume capacity, power requirements of the two groups are approximately the same, although the mid points of the intervals for steel designs seem greater than those for aluminum designs. For hold volume capacities larger than 75 [LT], steel designs clearly require more power than aluminum designs. Finally, for this investigation some cost terms have been compared. Figures 5.7 illustrates the variations of total costs with respect to hold volume and building material. Up to 75 [LT] of hold capacity, there are cross overs between the mean lines of the Chapter 5. MONOHULL VESSEL DESIGN KNOWLEDGE BASE (UBC-MONO) 76 intervals. This result suggests that region of up to 75 [LT] hold capacity, aluminum vessels might offer a less expensive alternative to steel vessels, as also reported by Calisal and McGeer [24] (Figure 5). 5.2 Summary In this part of the study, the aim was to see if the knowledge base in Echidna produces reasonable designs before incorporating some ship motions criteria (for crew comfort and safety reasons) into it. The above results can be seen as a validation of the knowledge base in Echidna, UBC-MONO, for monohull vessel design. From the above results, it can be concluded that • By using UBC-MONO, it was possible to obtain similar results reported earlier [24]. • Echidna was able to reproduce an existing vessel within reasonable differences from the existing design. • The validation for UBC-MONO was at an acceptable level of confidence. Hence, the UBC-MONO knowledge base could be used in our investigation as a design tool. Chapter 5. MONOHULL VESSEL DESIGN KNOWLEDGE BASE (UBC-MONO) 77 Figure 5.1: Comparison of Echidna generated aluminum and steel designs' lengths (vessel type : seiner). The jump could be because of alternative estimation rules embedded for some variables in the knowledge base. Chapter 5. MONOHULL VESSEL DESIGN KNOWLEDGE BASE (UBC-MONO) 78 Figure 5.2: Comparison of Echidna generated aluminum and steel designs' beams (vessel type : seiner). The jump could be because of alternative estimation rules embedded for some variables in the knowledge base. Chapter 5. MONOHULL VESSEL DESIGN KNOWLEDGE BASE (UBC-MONO) 79 Figure 5.3: Comparison of Echidna generated aluminum and steel designs' drafts (vessel type : seiner). The jump could be because of alternative estimation rules embedded for some variables in the knowledge base. Chapter 5. MONOHULL VESSEL DESIGN KNOWLEDGE BASE (UBC-MONO) 80 1.4E02 1.2E02 1.0E02 " 8.0E01 3 6.0E01 3 4.0E01 2.0E01 O.OEO Steelj Vessels \y\ j yf \ S / : / s / Alumiiium Ves sels j J 1 I 1_ 50 100 150 200 250 300 Hold Capacity [LT] Figure 5.4: Comparison of Echidna generated aluminum and steel designs' hull weights (vessel type : seiner). Chapter 5. MONOHULL VESSEL DESIGN KNOWLEDGE BASE (UBC-MONO) 81 1E03 9E02 8E02 ST 7E02 X 6E02 '3 £ 5E02 4E02 3E02 2E02 0 50 100 150 200 250 300 Hold Capacity [LT] Figure 5.5: Comparison of Echidna generated aluminum and steel designs' required pow-ers (vessel type : seiner). Chapter 5. MONOHULL VESSEL DESIGN KNOWLEDGE BASE (UBC-MONO) 82 6E02 5E02 4E02 e g 3E02 o CO 2E02 1E02 0E0 • I N / / / / / i ; / / 1 Y / / s V ! /• / ../' y A lumin / / ; <• / / < y / A / ! am Vesf els , / ,.s / / / i : : i , i i • 50 100 150 200 250 Hold Capacity [LT] 300 Figure 5.6: Comparison of Echidna generated aluminum and steel designs' displacements (vessel type : seiner). Chapter 5. MONOHULL VESSEL DESIGN KNOWLEDGE BASE (UBC-MONO) 83 3.5E06 3.0E06 2.5E06 ^2.0E06 •*-» o U o 1.5E06 H 1.0E06 5.0E05 0.0E0 -Aluminum Vessels zy- Steel Vessels <^ i 1 > -VS j i ; A/ -i , i i 50 100 150 200 250 Hold Capacity [LT] 300 Figure 5 . 7 : Comparison of Echidna generated aluminum and steel designs' costs (vessel type : seiner). Chapter 6 SWATH PRELIMINARY DESIGN AND UBC-SWATH 6.1 Multi-hull vessel for fishing role In nowadays' ever-changing economic and environmental conditions, the traditional spe-cialized one-technique fishing vessel is rapidly disappearing. Ideally, a modern fishing vessel should be flexible in its approach, enabling any number of techniques such as, deep or shallow trawl and long line fishing. This versatility, through the use of a multi-purpose vessel, would mean greater employment security through the vessel's ability to adapt to different species, new sustainable fisheries management programs, or better adaptation to changes in climatic conditions or selective harvesting techniques. The adaptability in variety of deck configurations, in a multi-hull vessel's large stable box-shaped upper hull, could provide an effective deployment of existing fishing equip-ment. As a multi hull vessel, SWATHs have been receiving considerable attention due to their superior seakeeping behaviour. Together with the other features as reported by Atlar et al in [10] such as Vehicle Motions, Safety, Operations, Habitability, Hydrody-namic Characteristics, etc., the SWATH concept offers advantageous features for fishing operations. It has also been reported by Kennel in [39] that in 1992, 10 of worldwide built 25 SWATH ships with 40,000 tons of cumulative tonnage accomplish their missions as both multi-purpose work boats and fishing vessels. 84 Chapter 6. SWATH PRELIMINARY DESIGN AND UBC-SWATH 85 6.2 General Features of Multi-hull Design The design process of a SWATH vessel is considered to be very similar to a conventional monohull design. The novelty in the SWATH concept is in its hull form. Hence, its design process follows the traditional form i.e. the owner requirements are at the outside the design spiral and a completed design at the center obtained through successive steps of iterations. The multi-hull design as being a weight sensitive design represents a difficult case requiring greater attention to its design. In conventional monohull ship design, the main design parameters are usually decoupled or less dependent upon each other. However, for a multi-hull vessel, this is not the case due to the increased number of design parameters and the higher order of inter-dependence of these parameters. SWATH vessels even represent the most difficult case of a multi-hull design, mainly due to its unconventional underwater hull form, combined with small waterplane area. Moreover, the lack of a historical SWATH design database and the immaturity of the validated empirical design methods require the use of sophisticated design analysis tools even in the early design process. The Echidna expert system environment provides the necessary tools for high speed evaluation and iteration of the design through its constraint propagation techniques and intelligent backtracking. However, the current version of the Echidna shell does not have any design analysis tools. In Atlar [9] and MacGregor [47] SWATH designers have cautioned on the special features of SWATH vessels and of the high interdependence between many of the basic SWATH design parameters. Two of the main characteristics reported in comparison to a monohull of the same displacement are reduced length and increased beam in SWATH vessels. However, in terms of the size of the weather deck area SWATH vessels are similar, though with more usable area due to their rectangular shape, compared to monohull vessels. Despite the fact that the large depth, draft, and freeboard are of importance Chapter 6. SWATH PRELIMINARY DESIGN AND UBC-SWATH 86 mainly to the operation of the ship, to a certain extent, these parameters will also affect, the design of some of the subsystems (e.g. the larger the draft, the better the propulsion characteristics while the higher the deck height, the poorer the steering and maneuvering and structural integrity due to greater structural loads). The larger wetted surface area affects the powering requirements, hull painting systems, and cathodic protection system. As far as the hydrostatic properties are concerned, the trim and weight related issues for SWATH ships take the first priority in design decision-making to the SWATH designer in contrast to the monohull design. Unanticipated weight growths due to estimation errors during the design can jeopardize SWATH's effective use and even in extreme cases might lead to the design of an unacceptable ship. Lastly, since a SWATH vessel has the same heel sensitivity as the monohull designed for the same application, the selection of the appropriate beam and waterplane area is an important key factor in the determination of the adequate transverse stability and seakeeping capabilities of the vessel. 6.3 SWATH Design Algorithms and Discussion The SWATH preliminary design expert system, developed at UBC, performs its prelimi-nary design computations from empirical formulae originally developed at the University of Newcastle upon Tyne [9] and Glasgow University, United Kingdom [47]. These algo-rithms are given in Appendix C. It is important to note that the current expert system produces satisfactory solutions. To elaborate more, the intervals given as solutions satisfy the relevant constraints in the knowledge base. However, it should be noted that these solutions may not necessarily be the optimum solutions in the sense of a conventional optimization process (e.g. Simplex method). However, the convergence to the solution obtained by a conventional optimiza-tion routine might be achieved, applying further constraints, such as on weight and/or Chapter 6. SWATH PRELIMINARY DESIGN AND UBC-SWATH 87 cost, once Echidna reaches a solution. Another approach could have been to start to design from an existing vessel. However, this has not been implemented in the present version of the knowledge base. For the validation of our SWATH vessel preliminary design expert system, 41 existing designs from [47] were used as test cases. The aim was to reproduce each existing design in Echidna. Table 6.1 shows the input to the expert system. The choice of some of the parameters such as hull material may be left unspecified at the input if they are not known initially. Table 6.1: Input parameters for SWATH vessel design expert system. Parameter Domain Hull material {mild steel, aluminum, hybrid } Lower hull type {circular, noncircular } Strut type {long, short } Machinery type {high speed diesel, medium speed diesel, gas turbine } Displacement Cargo Maximum speed Cruise speed Range at maximum speed Range at cruise speed Number of passengers Number of crew In the formulation of the SWATH vessel design problem the main constraint is floata-tion i.e. in UBC-SWATH (displacement > [1.0,1.03] * total-weight). Some of the results obtained from this analysis are given in Figures 6.1 to 6.5. The expert system developed here is planned to be used in designing fishing vessels up to approximately 1000 tons of displacement. Therefore, whenever possible the performance of the expert system in the 0 to 1000 ton displacement range is also given. Chapter 6. SWATH PRELIMINARY DESIGN AND UBC-SWATH 88 Table 6.2: Statistical analysis of % variations for A after excluding the pathological cases (4 cases out of 41 total number of cases). The 4 existing designs are as follows: a 305 ton fishing vessel, a 225 ton ferry, a 225 ton crew boat and a car ferry with a 1250 ton displacement. % Variations All Data A = [0,1000] Lower Upper Lower Upper Average 20.58 28.98 18.90 24.60 Standard deviation 42.73 55.73 38.12 50.01 Minimum -28.57 -28.10 0.0 0.0 Maximum 179.82 219.88 116.92 162.30 In Table 6.2, the percentage variations was calculated by %Vaviation = 100 * PaTameterUBC-SWATH - ParameterExUting D e , i g n Parameter Existing Design Generally speaking, the variation of the design parameters given by Echidna follows a trend similar to that of the design parameters of existing designs. Figure 6.1 shows how Echidna performed with respect to the displacement. In this figure, it seems that UBC-SWATH produces a similar trend in the displacement values of the Echidna designs. During the validation process, it has been noticed that for some particular designs such as passenger ferries and fast attack boats, UBC-SWATH was not able to perform as expected. As also discussed in [47], the current lower hull form design algorithms are only for simple type lower hulls. Therefore, for the existing SWATH designs with different type of lower hull forms, UBC-SWATH experienced difficulties in obtaining similar values of design parameters to those of existing designs. It is interesting to note that usually the same designs have caused higher standard deviations. Almost the same pathological pattern repeats itself in the study of other parameters such as LOA-Figure 6.1 shows the comparison of displacement values by Echidna vs existing designs Chapter 6. SWATH PRELIMINARY DESIGN AND UBC-SWATH 89 in different displacement ranges. In this figure, except 3 cases, two of which with 225 tons of displacements, the third one with 305 tons of displacement, the variation in [0, 1000] ton displacement range seems reasonable. These three cases are a ferry, a crew boat and a fishing boat. The exclusion of any design by Echidna having three or more times larger displacement than that of the existing design, which are the above three cases and a car ferry with a 1250 ton displacement, improves the average percentage variation in A, and its standard deviation considerably. The number of cases excluded is only 4 cases out of 41 total number of different design cases. Table 6.2 gives the new statistical results for the design parameter A only. In Figure 6.2 LOA values are compared. It is noticeable that much of the discrepancies occurred in the same existing design cases mentioned in the previous paragraph. As seen from Figure 6.3, average variation between the "Beam" of the existing and Echidna designs is within a reasonable range. However, there still exists some discrepan-cies. The parameter "Draft" values also exhibit a similar variation to the beam parameter, as in Figure 6.4. The last design parameter obtained by Echidna compared with existing designs' is the "Installed Powei}'1 (see Figure 6.5). This is the parameter that cannot be said to have been estimated with confidence. The main reasons for the discrepancies between the expert system given and existing design parameters are believed to be : • Not in all of the cases, the mission of the vessel to be designed i.e. the speed, endurance, amount of cargo, and number of passengers and crew, was available. In these cases, these variables were guessed from similar existing design cases. How-ever, as a SWATH is a weight sensitive design, these somehow arbitrary assignments affected the procedure. Chapter 6. SWATH PRELIMINARY DESIGN AND UBC-SWATH 90 • It has been observed that final solutions are usually very sensitive to changes in the coefficients of the empirical formulae, especially related to weight groups estimation. For example, the weight of the passengers (and crew) with effects is assumed to be 0.143 tons, i.e. 7 people with their effects weigh 1 ton. However, any change in this assumption seems to dramatically affect the solution produced for passenger ferries, and navy vessels, where a large number of people are on-board. • The empirical formulae used here were developed for all types of SWATH vessels. It is believed that these formulae fail to capture type specific (e.g. pertaining to passenger ferries, or patrol boats, etc.) features of SWATH vessels. However, this was again due to the lack of historical data or well developed "knowledge base" in SWATH vessel design. • Regarding the power estimation, the current system (UBC-SWATH) fails to predict the power requirements with confidence. One of the reasons for this might be the resistance estimation. Figure 4.11 in [47, pp. 126 ], shows excessive scatter in the residual resistance data. Therefore, this is also believed to be one of the contributing factors. 6.4 Remarks As mentioned earlier, when compared to a monohull, a SWATH design is weight sensitive. In the absence of well defined input parameters, it is not surprising that the procedure in Echidna converged to a solution with different designs, usually larger SWATH vessels. In other cases, Echidna was able to produce similar designs to the existing designs. In the following figures, the x axis represents the displacement values of the existing designs. Chapter 6. SWATH PRELIMINARY DESIGN AND UBC-SWATH 91 It should also be noted that since in Echidna the calculations are done in real intervals thus the results are given as real intervals. Therefore, in the figures, "Echidna lower" and "Echidna upper*' represent the lower and upper bound of the interval. The line in the figures shows the trend in the existing design parameter. Chapter 6. SWATH PRELIMINARY DESIGN AND UBC-SWATH 92 1 0 2 1 0 3 1 0 4 Displacements of existing designs [tons] Figure 6.1: Comparison of UBC-SWATH and existing design A's. The horizontal axis represents existing designs' displacement values. Chapter 6. SWATH PRELIMINARY DESIGN AND UBC-SWATH 93 03 O 160 140 120 100 80 60 40 20 0 • Echidna designs - Q — Existing designs I I I I I 1 I I I I I I I 1 I I I I I l I I I L 10 100 1000 10000 Displacements of existing designs [tons] (log scale) Figure 6.2: Comparison of U B C - S W A T H and existing design £ O A ' S . Chapter 6. SWATH PRELIMINARY DESIGN AND UBC-SWATH 94 70.0 60.0 50.0 40.0 a 03 PQ 30.0 — 20.0 10.0 0.0 y/ Existing designs O Echidna designs t -o I I I I 111ll I I I I 11 III I I I I I I I 10 100 1000 10000 Displacements of existing designs [tons] (loglO scale) Figure 6.3: Comparison of U B C - S W A T H and existing design Beam' s . Chapter 6. SWATH PRELIMINARY DESIGN AND UBC-SWATH 95 16.0 - • Echidna designs 14.0 — — e - - Existing designs 12.0 10.0 as Q 8.0 6.0 4.0 2.0 j i i m ? i i i i i i 1 1 1 i n i 10 100 1000 10000 Displacements o f existing designs [tons] ( log 10 scale) Figure 6.4: Comparison of U B C - S W A T H and existing design Draffs. Chapter 6. SWATH PRELIMINARY DESIGN AND UBC-SWATH 100000 "8 10000 1000 ((log 10 - log 10) scale) I T -0—Existing designs • Echidna designs i I 1 • J I I I hi Mil I I I l i l l M i l l I I I I 10 100 1000 10000 Displacements of existing designs [tons] Figure 6.5: Comparison of U B C - S W A T H and existing design Installed Powers. C h a p t e r 7 S E A K E E P I N G C O N S I D E R A T I O N S In the previous chapters, the work related accidents on board fishing vessels and their possible contributing factors have been discussed from the point of view of the effects on human performance and safety. Among the factors, vessel motions were mentioned as one of the most important contributing factors to accidents by various authors in the literature (e.g. [67], [56], [31]). It was suggested that reduced motions would mean a better living and working environment on board fishing vessels. However, the cost of modifications to existing vessels to obtain reduced levels of acceleration and amplitudes to achieve a worthwhile improvement in the quality of living and working conditions on board or incorporating into the new designs needs to be investigated. This must also be done within the relevant constraints imposed by the physics of the problem (e.g. stability) and owner related monetary constraints that are life cycle costings and benefits. There are different strategies that can be used to reduce vessel motions in practice. These are: 1. Active and passive roll and pitch stabilizers for example fins, bilge keels, bossings, gyroscopic stabilizers, tank stabilizers, jet flaps and rudders are mentioned in [13] and [58]. Each of these systems has advantages and disadvantages. Their efficiency, added cost and added weight, the space required for the stabilizing system are factors in selecting one system or another. Naturally, there is a limitation on the amount of improvement that these systems would offer. 97 Chapter 7. SEAKEEPING CONSIDERATIONS 98 2. Another strategy of reducing the motions is to design larger vessels. The amount of increase in the vessel size and associated first building and operating costs are important to the owner, although they need to be traded off against the change in earnings throughout the expected lifetime of the ship. 3. In some cases, it may not even be possible to increase the size of the vessel due to some regulations such as the Canadian fishing vessel replacement rules. A more dramatic strategy would be the use of a different concept altogether in the design, for instance to use a SWATH (Small Waterplane Area Twin Hull); this has inherent advantages over monohulls especially in regard to the vessel motions ([9] and [47]). In this thesis, a methodology, more along the same lines of the second and third strategies mentioned above has been developed, especially applied to monohull vessels. Among the reasons for the development of these design strategies include the criticisms that • the appendages extended outward of the hull e.g. fins, are not really favored by fishermen, • the active or passive stabilizing systems are usually considered for one of the vessel motions only such as roll motion, • the systems mentioned are fairly well understood in terms of their efficiency, costs and benefits. In the strategies used in this thesis, the goal of reducing motions was not achieved by explicitly asking the Echidna knowledge base to design a larger vessel, though at the end, sometimes larger designs may have been obtained as solutions. Rather, the approach was to incorporate some rules implied from the relationship between the vessel motions and the human comfort on board, and let Echidna find a solution. Chapter 7. SEAKEEPING CONSIDERATIONS 99 In an attempt to reduce motions, two different sets of rules have been developed and used for monohull vessel design in this thesis. These sets of rules are described in the following sections. The comparison of the two sets of rules in terms of their effects on the size and costs of the monohull vessels can be found at the end of this chapter. Another presentation in this chapter is the comparison of the monohull designs, ob-tained with these two sets of rules included in their design, to SWATH vessels with the the same hold capacity. 7.1 Rule Set I In general when a dynamic system is excited near one of its natural frequencies the amplitude of motion is (expected to be) larger. Similarly under normal circumstances, for a given sea state, a ship will be most vulnerable to such large amplitude motions when the frequencies of the high energy waves coincide with the peak natural frequencies of the ship (see Figure 7.1 on page 107). Hence, a way of reducing large vessel motions and related discomfort would be to design vessels in such a way that their natural frequencies will not coincide with the frequencies of the high energy waves in the sea for a given sea state. However, the amount of shift in the peak resonance and the peak energy is an important factor and should be determined in conjunction with cost, improvements in the ship motion levels for crew safety and comfort, and other physical rules of design such as the ones related to stability. For this approach, a design parameter (Equation 7.1) has been defined to quantify the amount of ratio of the frequency of the sea state's highest energy waves and the natural frequency of the vessel. „ . , , Vessel's response peak frequency Ratio of peaks = — £ % . (7.1) Sea states peak frequency Chapter 7. SEAKEEPING CONSIDERATIONS 100 In this equation, peak frequency for a sea state corresponds to the frequency of highest energy waves, whereas for the vessel it is the peak natural frequency for a particular motion, e.g. pitch motion. . In order to implement this strategy in the Echidna knowledge base, the following set of rules have been used. Rule Set I t order ratio.of-peaks. (7-2) 1st rule : — Ratio > 3.0. (7-3) 2nd rule : - Ratio > 2.4. (7.4) 3rd rule : - Ratio > 1.8. (7.5) As mentioned earlier in the thesis (see page 69 in Chapter 5), the first rule (7.2) in the knowledge base, i.e. "order ratio_of_peaks.", ensures that the rules are exhausted in the order of appearance. Hence, Echidna will first try to produce designs whose "Ratio" values are greater or equal to 3.0. In the above rules, shifts of the ship responses peaks towards lower frequencies (e.g. "Ratio < 1.0") are not implemented, simply because the natural frequencies of smaller vessels tend to be in the higher frequency range compared to larger vessels. Furthermore, as the sea state becomes more severe, the peak of the sea states energy spectrum shifts to lower frequencies. Therefore, at first a shift towards higher frequencies might seem more preferable. As far as the frequency of the exciting forces on a vessel is concerned, in the case of motions due to waves at the sea, it is not the frequencies of the waves, but the encounter frequencies that define the excitation frequency. The encounter frequency, which is given Chapter 7. SEAKEEPING CONSIDERATIONS 101 by Equation 7.6, depends on the speed and the heading with respect to waves encountered by the vessel, as well as the frequency of the waves. u>e = u> Vcosx (7'6) 9 Where u> is the frequency of the waves in the ocean to a fixed observer, uje is the encounter frequency, g is the acceleration due to gravity, V is the speed of the vessel and x is the heading of the vessel with respect to waves encountered. One of the implications of Equation 7.6 is that for a given sea state and vessel's heading, it is possible to reduce the vessel's excitation at its natural frequency by reducing or increasing the speed of the vessel. However, unlike reducing the speed, it is not always possible to increase the speed because of the rapidly growing power requirements that cannot be matched by the engine in the vessel, and/or the increased probability of increasing structural damage to the hull. Since the sea state, the speed and the heading were all constant in all of the example designs, a shift of the vessels natural frequency to higher frequencies means smaller vessels in general. It is also interesting to note that the higher the sea states are, the easier it is for the design process to get higher values of the ratio, as there is a shift of high energy waves to lower frequencies for increased sea states. In regard to the determination of the limits used in the above rules, 3.0 is considered to be a fairly satisfactory value for the ratio of the frequencies in Equation 7.1. Whereas after selecting 3.0, it was desirable to keep the number of rules to a minimum in order not to increase the solution space unnecessarily. Although in larger solution spaces, the chance of finding a solution is greater, it may take longer time to reach a solution. Thus, it was decided to include two more alternative rules i.e. 7.4 and 7.5, after some numerical design experimentation on the values for these limits and the number of the rules that a designer can use in the design process. Chapter 7. SEAKEEPING CONSIDERATIONS 102 Table 7.1: Input values to Echidna knowledge base for monohull fishing vessel. Parameter Value Gear type Material type Design speed [kn] Ballast (% of V) Hold capacities [LT] seiner aluminum 10.0 0 - 10 10, 25, 50, 75, 100, 125, 150, 200, 250, 300 In order to examine the effects of introducing Rules 7.2 to 7.5 into the knowledge base, for 10 different hold capacities 10 different designs were obtained without the rules in effect. Except for the hold volume capacity, the input to Echidna was the same as given in Table 7.1. Subsequently, this process was repeated for 5 different sea states, namely for 1, 2, 3, 4, and 5, and with the rules in effect resulting in 60 example designs in total, including the case without the rules. In obtaining the examples, the knowledge base (UBC-MONO) was the same for all cases except for the values of the required hold capacity were used as inputs. In the Echidna environment, it is possible to further reduce an interval solution by either applying additional constraints on it i.e. in this case the interval range for the ratio or using a built-in function, split, which randomly selects a subset of the original interval, specified by one of the rules of 7.3 to 7.5, that satisfies all the constraints of these rules. However, this option was not exercised because randomly shrinking the intervals in different examples can have different effects, and this would not conform with using exactly the same set of rules for all of the examples. Figures 7.3 to 7.7 depict the comparison of using and not using Rule Set I (Rules 7.2 to 7.5). Since the output values given by Echidna are expressed as real intervals for each ship design parameter, the interval solutions are represented as vertical bars in the Chapter 7. SEAKEEPING CONSIDERATIONS 103 figures. The length of the bar indicates the size of a corresponding interval. In order to show the trend in the figures, lines are passed through the midpoints of the intervals. In principle, there could be several solutions enclosed in the interval, and none of them may be at the midpoint. However, if a uniform probability distribution is assumed for the values included in an Echidna interval to represent a solution, choosing the midpoint in the interval as a solution would minimize the error associated. Therefore, the midpoints have been selected as the representative values of the solutions within the intervals. Also, there are two types of lines in the figures, solid and dashed lines. The dashed lines in the figures correspond to the case in which Rules 7.2 to 7.5 were not incorporated during the design, i.e. they correspond to "no sea state criteria" case. Whereas the cases with the rules imposed during the design were represented by solid lines, i.e. "Sea state 5" case. The following results are presented for the comparison of the example designs with and without (control case) the new rules for a design-sea-state 5 only. Figure 7.2 on page 108 shows the values of ratios, defined by Equation 7.1, with and without Rule Set I in effect. The general conclusion from the Figure 7.2 is that Echidna was able to shift the natural frequencies of the designs. By studying the figure, it seems that even before the activation of the rules in Rule Set I, the majority of example designs for their minimum values in the intervals satisfy Rule 7.5. By taking the midpoints of the intervals into account, most of the example designs with Rule Set I in effect, satisfy Rule 7.3 and Rule 7.2 for heave and pitch motions respectively. The effects of Rule Set I on the ship size were investigated through Figures 7.3 to 7.6. Figure 7.3 on page 109 displays the effect of the shift of the vessel's natural frequency on the length. The general implication from this figure is that Rule Set I caused an increase of up to 20% in the length, (excluding 10 [LT] hold capacity with 50% increase) if the mid points of the length intervals of the example designs are considered. The Chapter 7. SEAKEEPING CONSIDERATIONS 104 lengths show an increasing trend with increasing hold capacity for both cases, whether the rules are in effect or not, except in the interval between the hold capacity of 50 to 150 [LT] for cases with rules in effect and 100 to 300 [LT] for cases without the rules in effect. The lengths of the vessels are approximately the same for the example designs for the hold capacities contained in these regions. The reason of this irregularity in the trend is due to the methods used in the estimation of length in the knowledge base. Rule 7.7 and 7.9 shows the rules in Echidna for vessel's length estimation. Vessel's length estimation : order estimate length. (7-7) 1st rule : — Length estimated < Licenselength. (7.8) 2nd rule : — Use estimated Length as it is. (7-9) These rules initially cause Echidna to produce designs whose lengths are less than or equal to the license length (see Equation 7.8). In the knowledge base, the method used in the estimation of license length is based on the design speed alone (see Equation B.2 in Appendix B on page 206). In our investigation the parameters that were held constant were design speed and the license length for different hold volume capacities. In effect, Echidna attempted initially to generate designs whose length satisfy Rule 7.8. For the cases without Rules 7.2 to 7.5 incorporated, Echidna was able to generate designs by only using Rule 7.8. However, when Rule Set I was in effect, two features can be observed in the figure. Firstly, License length limitation was reached earlier and secondly, it was no longer possible to produce designs which satisfy Rule 7.8, for hold capacities greater than 150 [LT]. The general tendency in the lengths of the example designs can be summarized as: the example designs are lengthened in the cases when Rules 7.2 to 7.5 were activated. Figure 7.4 illustrates the changes in the beams of the example designs due to activa-tion of Rule Set I (Rules 7.2 to 7.5). Up to 75 [LT] hold capacity, beams of both example Chapter 7. SEAKEEPING CONSIDERATIONS 105 designs have an increasing trend with hold capacity, however, with different slopes. At 75 [LT] hold capacity the increase in beam due to Rule Set I is around 24%. After 75 [LT], the beam values of the different example designs, one with the new rules and the other without (control case), first converge to each other between 125 and 150 [LT] and then diverge as hold capacity increases. In the case of "No sea state criteria" example design for 300 [LT] hold capacity, the initial interval prior to the design was preserved during the design process, thus causing a very large interval as a solution value for the beam. For this specific hold capacity, Figure 7.4 suggests an approximately 37% decrease in the beam due to the new rules in the design if the mid points are to be considered. However, this percentage would be reduced if additional constraints or built in "split" function were used. In connection with the changes in the lengths, for 150 [LT] hold capacities Echidna solutions for sea state 5 case opted for larger L/B ratios, since the variation in lengths seem to be larger than the variation in beams. Overall, the set of example designs with Rule Set I, present beamer vessels. Increases to the beam may mean increased displacement. Hence, the larger added mass and damping coefficients are expected to reduce motions. However, wave excitation forces increases due to the increased waterplane area. Generally, a large (beam/length) ratio is reported to reduce absolute vertical acceleration while relative motions are increased ( [45, p 471]). Figure 7.5 suggests that vessels with shallower drafts emerge after incorporating Rules 7.2 to 7.5 into the design process. Contrary to the changes in length and beam, the changes in draft presents a clearer picture. Similar to Figure 7.4, 300 [LT] example design of the group, Control case, in Figure 7.5 shows a very large interval. In Figure 7.6, displacements of the example designs with and without the rules are compared. Except the two cases of hold capacities at 10 and 300 [LT], the example designs with the rules in effect during the design have larger displacements in general. Although the solution intervals in both cases overlap each other to some extend, if the mid points Chapter 7. SEAKEEPING CONSIDERATIONS 106 of the intervals are considered, an increase in the displacement values is suggested in the figure. One of the most important parameters from an owner's point of view is the vessel's cost. Unfortunately, most of the solution intervals overlap heavily in Figure 7.7, which illustrates the variation of the cost between the designs with and without Rules of Method I, 7.2 to 7.5, thereby making it more difficult to reach a clear conclusion. However, up to 100 [LT] of hold capacity the average percentage increase in the cost is 24.4% with a standard deviation of 31.9% when the mid points of the intervals are considered. However, the overall values of percentage increase for average and standard deviation are 16.4% and 26.4%, respectively, for the mid points of the intervals. Chapter 7. SEAKEEPING CONSIDERATIONS 107 S e a spectrum \ S e a sea Frequency Ship response Frequency • ship Figure 7.1: Frequencies that maximize sea spectrum and ship response spectrum. Chapter 7. SEAKEEPING CONSIDERATIONS 108 3.5 3.0 -•£3 «© 2.0 I 15 1.0 0.5 6.0 5.0 With Rule Set I I -Control case _L _L ,4.0 -2 2.0 -1.0 With Rule Set I J_ _L 50 100 150 200 250 300 350 Hold capacity [LT] Control case _ J I . I L 50 100 150 200 250 Hold capacity [LT] 300 350 Figure 7.2: The variation of the ratios (Equation 7.1). Control case shows the ra-tios without any seakeeping rules included in the design. Design-sea-state is 5; oper-ational-sea-state is 5. Chapter 7. SEAKEEPING CONSIDERATIONS 109 0 50 100 150 200 250 300 350 Hold capacity [LT] Figure 7.3: The effects of Rule Set I on the lengths. Control case shows the lengths without any seakeeping rules included in the design. Design-sea-state is 5; opera-tional-sea-state is 5. Chapter 7. SEAKEEPING CONSIDERATIONS 110 Figure 7.4: The effects of Rule Set I on the beams. Control case shows the beams without any seakeeping rules included in the design. Design-sea-state is 5; operational-sea-state is 5. Chapter 7. SEAKEEPING CONSIDERATIONS 111 Figure 7.5: The effects of Rule Set I on the drafts. Control case shows the drafts without any seakeeping rules included in the design. Design-sea-state is 5; operational-sea-state is 5. Chapter 7. SEAKEEPING CONSIDERATIONS 112 Figure 7.6: The effects of Rule Set I on the displacements. Control case shows the displacements without any seakeeping rules included in the design. Design-sea-state is 5; operational-sea-state is 5. Chapter 7. SEAKEEPING CONSIDERATIONS 113 3.0 Figure 7.7: The effects of Rule Set I on the costs. Control case shows the costs without any seakeeping rules included in the design. Design-sea-state is 5; operational-sea-state is 5. Chapter 7. SEAKEEPING CONSIDERATIONS 114 7.2 Evaluation of the Example Designs by Rule Set I In the next step, the example designs obtained by the shifts in the frequencies corre-sponding to the peak motions are evaluated in terms of rms motion characteristics. This is in line with reported studies in the publicly available literature discussing discomfort due to ship motions - all are related to rms characteristics. In order to calculate rms characteristics for amplitudes or accelerations, an algorithm developed by Howard [23] is used. The algorithm is presented in detail in Appendix D. In obtaining the rms values using Howard's algorithm some issues had to be resolved. Since Echidna is not a suitable platform for intense computations, it was necessary to compute those values outside of the Echidna system. However, in choosing a system other than Echidna, there is a difficulty in representing real interval values in mathematical computations such as numerical integrals. In order to overcome this problem, three values for each interval were selected: the minimum value, the mid point and the maximum value of each interval. In doing this, the following assumptions have been made: • By arbitrarily selecting values from the intervals, we are not guaranteed that the set of selections represents a valid design, which satisfies the design constraints. However, if the intervals are sufficiently small and the nonlinearity in the math-ematical model is low, it may be assumed that the set of selections is close to a design. • Secondly, using a set of arbitrarily selected values from the intervals one may cause the resultant intervals to shrink. The nonlinearity in the model and the size of the interval is the determining factor on the amount of shrinkage. Chapter 7. SEAKEEPING CONSIDERATIONS 115 With these assumptions a FORTRAN code was written by the author to compute rms heave and pitch motion characteristics by the algorithm given in [23]. For the exam-ple designs obtained by Rule Set I, rms motion characteristics (motion amplitudes and accelerations) were computed for the three values (minimum, mid point and maximum) selected from the intervals that define an example design (e.g. length, beam, etc.). In Figures 7.8 to 8.31, the legend " (Echidna interval, lower bounds)" means that in obtaining the results presented in the figure, the minimum values of the related in-tervals were used. Similarly, "(Echidna interval, mid points)" and "(Echidna interval, upper bounds)" correspond to computations using mid points and maximum values of the intervals. In the figures showing "% Change" due to Rule Set I in rms heave and pitch motions (among Figures 7.8 to 8.31), "% Change" was calculated with respect to the control case, in which no seakeeping rules were incorporated in the design (see Equation 7.10). 0 / / - r z . ->nn r m S Valuewith RuleSet I ~ TmS ValueControl case 1 a N 7o Change — 100 (7.10) rms ValueControl case In Figures 7.8 to 8.31, rms heave and pitch values and the % Change in them are plotted for different design-sea-states and for an operational-sea-state of 5. The different curves in these figures correspond to different design sea states. For example, "design-sea-state : 5" means that during the design, Rule Set I ( 7.2 to 7.5) were in effect for Sea State 5. In other words, in a sea state of 5, the frequencies corresponding to the ship's peak response (for heave and pitch motions) of the vessel being designed should be greater than the frequencies of the high energy waves in the sea state 5 by the amounts indicated by Rule Set I ( 7.2 to 7.5). Figures labeled "Operational-sea-state : 5", display how different designs, which are designed for different design-sea-states, would perform in sea conditions of state 5. Chapter 7. SEAKEEPING CONSIDERATIONS 116 For the investigation of the effects of Rule Set I, operational-sea-states of 3, 4 and 5 were used. After examining the results obtained, the largest effects of introducing Rule Set I into the design were for operational-sea-state of 5. Therefore, Figures 7.8 to 8.31 shows only the results for operational-sea-state of 5. Discussion of the information in Tables 7.2 and 7.3; Tables 7.2 and 7.3 (on pages 117 and 118 respectively) summarize the information given in Figures 7.8 to 8.31. In these tables, the averages of the changes in motion amplitudes or accelerations for different hold capacities, i.e. for each design-sea-state, sum of the % changes in the rms motion amplitudes divided by 10, which is the number of different hold capacities used. The negative signs imply a reduction in the rms values. For example, the first row in Table 7.2 shows that if the minimum values in the Echidna intervals are used, there will be 2.647% increase on the average in the rms heave ampli-tude. For any given design-sea-state, as shown in Tables 7.2 and 7.3, selecting larger values from the Echidna interval e.g. maximum value (upper bound), seems to reduce the rms motion amplitudes more than any other value from the interval. Perhaps, this is because selecting larger values from the Echidna intervals also imply larger vessels, thus the motion amplitudes are smaller. Furthermore, for the same set of values of in-tervals, e.g. minimum, midpoint or maximum values, and the same sea state conditions, e.g. sea state 5 as in the figures, as the design-sea-state becomes larger, so does the improvement in general, that is the reduction in motion amplitude. Although there is not always a reduction in the average rms motion amplitude, see first row of Table 7.2, as the design-sea-state becomes larger the change in the average rms amplitude is in the right direction, i.e. decreasing. Chapter 7. SEAKEEPING CONSIDERATIONS 117 Table 7.2: The averages and standard deviations (over hold capacities) of the % Changes obtained in rms heave and pitch motion amplitudes. The vessel is assumed to be oper-ating in Sea State 5. Design Value selected from % Change Motion sea state the interval Average Stand, deviation Heave 3 lower bound 2.647 2.524 3 mid point -0.842 1.246 3 upper bound -2.259 0.500 4 lower bound -0.982 4.321 4 mid point -6.621 4.236 4 upper bound -9.167 4.992 5 lower bound -1.398 4.580 5 mid point -6.423 3.734 5 upper bound -8.811 4.226 Pitch 3 lower bound 4.813 1.859 3 mid point 2.785 1.713 3 upper bound -1.371 1.717 4 lower bound -6.001 10.763 4 mid point -11.411 9.530 4 upper bound -16.661 8.818 5 lower bound -6.081 10.662 5 mid point -11.504 9.645 5 upper bound -16.428 8.865 Discussion of the information in Figures 7.8 to 8.31; Some of the general points worth mentioning about Figures 7.8 to 8.31 here are: • For design-sea-state 3, Echidna was able to find only two solutions, i.e. for hold capacities of 10 and 25 [LT]. For other hold capacities, Echidna failed to find a solution, which could have meant there is no solution under the circumstances. However, there was a final message of "System error, exiting" upon the termination Chapter 7. SEAKEEPING CONSIDERATIONS 118 of Echidna. Therefore, this implies that for those particular cases of hold capacities, the answers are inconclusive, i.e. there could be a solution but "system error" messages from Echidna prevented the normal conclusion. Hence, in the figures missing data points correspond to such inconclusive cases. Table 7.3: The averages and standard deviations (over hold capacities) of the % Changes obtained in rms heave and pitch accelerations. The vessel is assumed to be operating in Sea State 5. Design Value selected from % Change Motion sea state the interval Average Stand, deviation Heave 3 lower bound 4.328 3.016 3 mid point -0.176 2.124 3 upper bound -3.901 1.287 4 lower bound ' -3.588 7.779 4 mid point -10.199 6.657 4 upper bound -14.813 6.662 5 lower bound -3.856 7.034 5 mid point -10.192 5.559 5 upper bound -14.595 5.300 Pitch 3 lower bound 5.073 0.258 3 mid point 5.850 0.666 3 upper bound 1.717 1.702 4 lower bound -9.031 14.600 4 mid point -11.970 12.967 4 upper bound -16.540 11.053 5 lower bound -9.038 15.963 5 mid point -12.502 15.336 5 upper bound -16.929 13.333 • The lines in the figures join the data points given by Echidna for corresponding hold capacities. In the absence of any symbol e.g. solid rectangle, hollow triangle, Chapter 7. SEAKEEPING CONSIDERATIONS 119 etc., at the data point, this implies that Echidna was not able to find a solution for the corresponding hold capacity. When the minimum values of Echidna intervals are considered (see Figures 7.8 and 7.9), there is a slight improvement in the rms heave amplitude, which is 1.4% reduction on the average for the design-sea-state 5. When the mid points of Echidna intervals were used in the rms calculations, there is an overall improvement for the all hold capacity range (see Figures 7.10 and 7.11). For the maximum values of intervals (see 7.12 and 7.13), the difference in % Change increased. Figures 7.14 to 7.19, show a similar case study for rms pitch motion amplitudes. Figures 7.14, 7.16 and 7.18 display rms pitch motion amplitudes for the cases in which minimum, midpoint and maximum values of Echidna intervals were used in the rms calculations, respectively. The largest reductions in rms pitch amplitudes occur in the hold capacity ranges of 50 to 100 [LT] and 250 [LT] and greater. Irrespective of the which values of Echidna intervals i.e. minimum, mid point or maximum, are used in the rms calculations. The decreases in rms pitch amplitudes are greater in design-sea-state 5 than in the case of design-sea-state 4. Similar results are presented for rms heave and pitch accelerations in Figures 8.20 to 8.31. In Figures 8.20, up to 125 [LT] of hold capacity, the rms heave acceleration seems to be reduced for design-sea-states 4 and 5. A similar case is depicted in Figure 8.21, which shows % changes rms heave accelerations. In the 0 - 125 [LT] hold capacity range, % changes are negative for the most part, thereby showing an up to 10% reduction in rms heave acceleration. Whereas for hold capacities larger than 125 [LT] % changes are positive, implying an undesirable increase in the rms values. However, the situation improves dramatically when midpoints (Figure 8.23) and maximum values (Figure 8.25) of the Echidna intervals were used. Nevertheless, the same pattern of larger reductions as Chapter 7. SEAKEEPING CONSIDERATIONS 120 in the rms heave amplitudes below 125 [LT] hold capacities, has emerged in Figures 8.23 and 8.25. Figures 8.26 to 8.31 display the information related to rms pitch acceleration. The hold capacities of 25 and 50 [LT] appear to be the most difficult cases to improve. The overall pattern is also repeated in the figures related to rms pitch accelerations. Using maximum values in the rms calculations results in the largest improvement. As the design-sea-state becomes higher, the improvement in the rms pitch accelerations decrease considerably. 7.2.1 Remarks about the ship dimensions and displacement With the introduction of Rule Set I ( 7.2 to 7.5) into the design, some changes in the particulars of the example designs as well as their costs were obtained. These changes due to Rule Set I are as follows (Figures 7.3 to 7.7): • Length; There is up to approximately 20% increase in the lengths (see Figure 7.3). • Beam; For the beams of the smaller example designs (less than 100 [LT] hold capacity), there is up to 25% increase in the beams. Whereas for larger hold capacities (greater than 100 [LT]), there is up to approximately 35% decrease in the beam values (see Figure 7.4). • Draft; Drafts of the example designs were reduced by up to approximately 30% (see Figure 7.5). • Displacement; If mid points of the intervals are considered, the designs with Rule Set I in effect have up to around 25% larger displacements (see Figure 7.6). • Cost: An up to 25% increase in costs of the designs were observed due to Rule Set I (see Figure 7.7). Chapter 7. SEAKEEPING CONSIDERATIONS 121 In return, the following changes were obtained in the rms heave and pitch motion characteristics of the example designs. In the results reported below, mid points of the Echidna intervals were used. Additionally, the sea state in which the vessel is assumed to be operating is 5 (see Tables 7.2 and 7.3 for more details). • For design-sea-state 3: — Heave: The above mentioned changes in the vessels' particulars due to Rule Set I, resulted in an average of c. 0.8% decrease in the rms heave amplitudes of the example designs. Similarly, there was a decrease in the rms heave accelerations of the example designs due to the rules used. — Pitch: There were 2.78% and 5.85% (on the average) increases in rms pitch amplitudes and accelerations respectively, when compared to the control case (no seakeeping considerations). • For design-sea-state 4 : — Heave: There were 6.6% and 10.2% (on the average) reductions in the rms heave motion amplitudes and accelerations respectively due to the rules used. — Pitch: For rms pitch motion amplitudes and accelerations 11.4% and 11.97% reductions were obtained respectively by using the rules during the design. • For design-sea-state 5: — Heave: c. 6.4% and c. 10.19% (average) reductions in rms heave motion amplitudes and accelerations were achieved. — Pitch: For rms pitch motion amplitudes and accelerations, the reductions were 11.5% and 12.5% respectively on the average. Chapter 7. SEAKEEPING CONSIDERATIONS 1.11 1.08 1.05 v 1.02 T 3 f t 0.99 s \ : \ \ : . . . I - • - Design-searstate : 3 \—m--- Design-sea-state : 4 ; . L--^-. Design-searstate..; .5.-.. 0 50 100 150 200 Hold Capacity [LT] 250 300 Figure 7.17: % change in rms pitch values. Chapter 7. SEAKEEPING CONSIDERATIONS 13 12 11 10 9 8 Operational-sea-state : 5 :. (Echidna interval; upper bounds) o Design-sea-tstate : none - - Q - - Design-sea-state : 3 ; — m-- Design-searstate : 4 i Desigri-seafstate : 5 • 50 100 150 200 Hold Capacity [LT] 250 300 Figure 7.18: Values of rms for pitch. 0 -5 -10 -15 -20 -25 -30 VV Operational-sea-state : 5 \ (Echidna interval^ upper bounds) t i L....I.. / . t •• i t •/•/• V / - E 3 - Design-sea-state : 3 Jmrr.-. Pesign-sea-state ; 4 -ky- - Design-sea-stiate : 5 _i i i , i , ^ 50 100 150 200 Hold Capacity [LT] 250 300 Figure 7.19: % change in rms pitch values. Chapter 7. SEAKEEPING CONSIDERATIONS 1.5 1.4 1.3 1.2 1.1 1.0 Qperational-sea-state : 5 (Echidna interval,Tower bounds) — Design-sea-state : none _i -iq--Design-sea-state : 3 ..Am--- Design-sea-state : 4 - 4A- - Design-sea-state : 5 , i . i , i , i , 50 100 150 200 250 300 Hold Capacity [LT] Figure 7.20: Values of rms for heave accelerations (Rule Set I). 10 5 0 -5 -10 -15 -20 — • -v I t 1 1 • •• I / / I ' " / I I I •m /• -y f • / i \ / : / • \ ' i_ -/r-. \.//... Operational-sea-state : 5 i X !• (Echidna interval^ lower bounds) N : - - Q - - Design-sea-state : 3 I l.-.-.-.«-.r..-.Design-seatstete..: 4. ; Design-sea-state : 5 : j i i , i i . 50 100 150 200 Hold Capacity [LT] 250 300 Figure 7.21: % change in rms heave accelerations (Rule Set I). Chapter 7. SEAKEEPING CONSIDERATIONS 1.5 2 1.4 2 CD 8 o < > CD 1.3 1.2 1.1 1.0 Operational-sea-state : 5 (Echidna iriterval, mid points) • - Q - -Design-sea-state Pesign-sea-state : 3 ..-m--- Design-sea-state : 4 Design-sea-state : 5 i i i i , i _ none 50 100 150 200 Hold Capacity [LT] 250 300 Figure 7.22: Values of rms for heave accelerations (Rule Set I). C D o o 0 3 C D a 0 -5 -10 -15 L. -20 \ s : : „ —4... ./A..v, \ // Operationai-sea-state : 5 Y (Echidna interval, mid points) - -o f - Design-sea-state : 3 . . . p L . . Design-sea-state : 4 - - A j - Desigrii-sea-statp : 5 50 100 150 200 Hold Capacity [LT] 250 300 Figure 7.23: % change in rms heave accelerations (Rule Set I). Chapter 7. SEAKEEPING CONSIDERATIONS 1.5 1.0 -e j— Design-sea-state : none - Q - - Design-sea-state : 3 .Desigh-sea-state : 4 - Desigh-sea-state : 5 dperational-sea-state : 5 (Echidna interval, upper; bounds) 1 1 1 I i I i 50 100 150 200 Hold Capacity [LT] 250 300 Figure 7.24: Values of rms for heave accelerations (Rule Set I). o o 0 3 > 0 3 CD C 0 3 6 6S o -5 -10 .S -15 -20 --25 i Operational-sea-state : 5 \ ; (Echidna interval, upper bounds) A / / > - - - - - . .... •A : * / / : / : V //..I / w n / \ . \ / - : - C 3 - - Design-sea-state : 3 'A : \\ . - 1 \ ; U :.t :: Design-sea-state : 4 Design-sea-state : 5 50 100 150 200 Hold Capacity [LT] 250 300 Figure 7.25: % change in rms heave accelerations (Rule Set I). Chapter 7. SEAKEEPING CONSIDERATIONS Operational-sea-state : 5 (Echidna interval, lower bounds) — — Design-sea-state : none - H Q - - Design-sea-state : 3 Design-sea-state : 4 Design-sea-state ':" 5 100 150 200 Hold Capacity [LT] 250 300 Figure 7.26: Values of rms for pitch accelerations (Rule Set I). 20 10 0 -10 -20 P--30 - / -40 I \ I \ 7 v •; / x • T\ X i: \ : \ /; \ ; \ I /.' V \ A . ....; / / . 1 >4 /Operational-sea-state : 5 . ^ i . _ ^ (Echidna interval, lower bounds) — Q - - Design-sea-state : 3 ;..| | ... Design-sea-state : 4 • - A - -_ i Design -sea-state : 5 50 100 150 200 Hold Capacity [LT] 250 300 Figure 7.27: % change in rms pitch accelerations (Rule Set I). Chapter 7. SEAKEEPING CONSIDERATIONS 0.55 FT 0.50 U V * . 0 3 ^ 0.45 k § 0 .40 - < 0 3 £ 0.35^ o < 0.30 | 0.25 k 0.20 iOperational-sea-sitate : 5 (Echidna interval,; mid points) I - oDesign-sea-jstate : none — - Q - - Design-sea-state : 3 j L-i-fli.-:Design-seaistate••: 4- \ Design-sea-jstate : 5 ; - A . _ l 50 100 150 200 Hold Capacity [LT] 250 300 Figure 7.28: Values of rms for pitch accelerations (Rule Set I). 20 'i 10 s , I -10 .5 & -20 C3 o -30 r_ -40 : Operational-sea-:state : 5 l_ , . . .^(Echidna.interval, mid points). B~ Jm i \ ,p ^ \ \ /' : ^ A ; > ^ -\ \ s • : \ \ / : v - - Q - - Design-sea-state : 3 , „.m— Design-sea-state 4 - - Design-sea-state : 5 , i , i i > 50 100 150 200 Hold Capacity [LT] 250 300 Figure 7.29: % change in rms pitch accelerations (Rule Set I). Chapter 7. SEAKEEPING CONSIDERATIONS 0.50 r1 0.45 -±L 0.40 s o '3 0.35 § 0.30 < S 0.25 O H 0.20 iOperatidnal-sea-sitate : 5 •(Echidna interval^ upper bounds) I—e— Dessign-sea-istate : none L . - G - - Design-sea-state : 3 : ; ~ - H - ~ Design-sea-state : 4 ! - A - - Design-sea4state : 5 : 50 100 150 200 Hold Capacity [LT] 250 300 Figure 7.30: Values of rms for pitch accelerations (Rule Set I). \ : ne. v.. / ; \ // ' // — H i: i: i: •( ; / / Operational-sea-state : 5-(Echidna interval, upper bounds) . - - Q - - Design-seja-state : 3 ---m--- Design-sea-state : 4 Design-sea-state : 5 A 7 V 11 50 100 150 200 Hold Capacity [LT] 250 300 Figure 7.31: % change in rms pitch accelerations (Rule Set I). Chapter 7. SEAKEEPING CONSIDERATIONS 134 7.3 Rule Set II In the previous section, a way to incorporate sea keeping considerations in the preliminary design process is described. In that approach, achieving the goal of lesser discomfort due to vessel motions was attempted by shifting the vessel's natural frequencies away from the frequencies of the high energy waves in the sea (Rule Set I, on page 100). However, in this section a different approach is described. Instead of shifting the peak frequencies, some of the rms values of the vessels' motions were constrained during the preliminary design. A new set of rules was specified, in order to implement these new rules during the design in a knowledge-based system without extensive calculations, the rms values of different motions under consideration, e.g. pitch motion, need to be estimated. For heave and pitch rms motion amplitudes, two estimation formulae are developed. The details of these algorithms can be found in Appendix F. The rules incorporated in the Echidna knowledge base are given in Rules 7.12. The first rule is related to rms heave motion amplitude. The value of upper limit has been decided upon after examining the rms values of example designs without rules. Although the acceleration levels are mentioned in the literature as one of the parameters that affect human performance, some of the suggested constraints (e.g. [67] and [56]) were on some of the ship motions amplitudes. Therefore, in this case study constraints on the amplitudes were used. Rule Set II rms heaveamplitude < 0.7 [m] (7.11) rms pitchamplitude < 5.0° (7.12) These two rules we added to the knowledge base and used during the preliminary design, both Rules 7.11 and 7.12 were activated simultaneously. However, Rule Set I Chapter 7. SEAKEEPING CONSIDERATIONS 135 ( 7.2 to 7.5) in the knowledge base were switched off while Rule Set II ( 7.11 and 7.12) was active. For each of the hold capacities mentioned in Table 7.1, with the other inputs to Echidna being the same as these used in Rule Set I then, 10 new example designs were obtained while Rule Set II was in effect during the design process with Echidna. In Figures 7.32 to 7.43, the effects of incorporating Rule Set II ( 7.11 and 7.12) in the preliminary design are presented. The lines in the figures join the midpoints of Echidna intervals for the same case study, e.g. the knowledge base with Rules 7.12 in effect, similar to the figures in the previous sections. The dotted lines represent the case, in which Rules 7.12 were not activated during the design. Whereas the solid lines are for the case study with the rules in effect. All were applied for sea state 5. In these series of figures, the first two figures, Figures 7.32 and 7.33, display the amount of changes in rms heave and pitch motion amplitudes respectively, after Rule Set II was added in the knowledge base. In the case of heave motion (Figure 7.32), prior to the addition of Rule Set II to the knowledge base, rms heave motion amplitudes were greater than the assigned limit values in almost all of the example designs. The example designs obtained with Rule Set II show reasonable reduction in rms heave mo-tion amplitude particularly in the lower range of hold capacities below approximately 75 [LT]. As shown in Figure 7.33, Echidna intervals for rms pitch motion amplitudes overlap significantly over the range of hold capacities, thereby suggesting Rule Set II have almost no effect on the current example designs, except in the small hold capacities up to approximately 75 [LT]. The effects of Rule Set II on the cost and the vessel's sizes, are illustrated in Figures 7.34 to 7.43. In the figures, the percentage change is calculated as follows: Chapter 7. SEAKEEPING CONSIDERATIONS 136 % Change = 100 Design Parameteri — Design Parameter (7.13) Design Parameter^ In the above equation, Design Parameteri represents the case where Rule Set II was included during the design, Design Parameter^ is for the case without any seakeeping Figure 7.34 shows the change in the lengths of example designs due to the addition of Rule Set II. Whereas, Figure 7.36 displays this change in the length as percentage of the lengths of the designs without the rules included in their design. The percentage change in the length has its largest values in the small hold capacities up to 75 [LT], after which it remains approximately constant. If the mid points of the intervals are considered in Figure 7.36, the line is in the positive side of Y-axis, thereby implying an increase in length for all hold capacities. In Figure 7.36 there is a jump in the current trend of % change in the length for 250 [LT] hold capacity. However, the size of the % change interval remains almost the same as those neighbouring it. Perhaps, one explanation for this might be that this region is a transient one where the license length limitation had to be violated during the design. This part of the curve is another case that could be investigated in detail for its deviation from the earlier trend in length depicted in the figure. It is interesting to note that the same situation also occurs for the beam, draft and displacement values of the example designs. Another interesting point related to this figure is that there is no observable difference in the intervals of the two cases, with and without the rules, for 300 [LT] hold capacity. If some extreme points such as the one for 250 [LT] hold capacity are neglected, considering the mid points in Figure 7.36, there is nearly a 3% increase in the lengths of the example designs larger than 75 [LT] hold capacity. For smaller hold capacities, the increase in length reaches as much as c.9% for mid points of the intervals. rules (control case). Chapter 7. SEAKEEPING CONSIDERATIONS 137 Figures 7.35 and 7.37 displays the effects of Rule Set II on the beams of the example designs. Rule Set II tends to increase both beams and lengths. Almost the same patterns as in Figures 7.34 and 7.36 emerge in the figures for the beams of the example designs. As mentioned earlier, the jump for the 250 [LT] hold capacity is also present in Figure 7.37. If the mid points of the intervals are considered, there is approximately 5% increase in the beams. Figures 7.38 and 7.40, are difficult cases as there is no clear trend in the changes of the drafts due to Rules 7.12. Generally speaking, the sizes of the intervals of % changes are comparatively larger for smaller hold capacities of up to 75 [LT] in Figure 7.37 as these are relatively small boats and they are significantly improved by increasing their lengths. Again, the intervals in both cases (with and without the rules) remained the same for 300 [LT] hold capacity. Up to 50 [LT] hold capacity, % changes imply an increase in the draft, whereas, between 50 to 100 [LT] hold capacities, a reduction in draft seems more plausible from the figure. Between 100 to 200 [LT] hold capacities, the midpoints of the intervals are very close to 0% change, however, for greater hold capacities, there may be shallower drafts suggested in the figure, except 300 [LT] hold capacity, in which there is no change in the draft intervals of the example designs (with and without the rules). Displacements of the vessels are more appropriate to compare the sizes of different vessels. Hence, Figures 7.39 and 7.41 reveal the changes in displacements of the example designs before and after the introduction of Rules 7.12 in the design. Figure 7.39 shows the displacements of example designs in long tons. The two lines in the figure, which join the mid points of the intervals, are nearly parallel to each other. This suggests an increase due to Rules 7.12 in the sizes of the vessels, which is generally expected. However, the important factor in this investigation is the amount of increase to satisfy the constraints on rms heave and pitch motions. Figure 7.41, in which the change is given as % of Chapter 7. SEAKEEPING CONSIDERATIONS 138 the displacements of the example designs obtained without Rules 7.12 present in the knowledge base, is helpful in determining the necessary increase. If the mid points of the intervals are taken into account, the trend in displacement is remarkably similar to the trends in lengths and beams of the example designs, as shown previously in Figures 7.36 and 7.37 respectively. For small hold capacities up to 75 [LT], a larger increase in displacement is suggested, with the largest being 20% for 10 [LT] hold capacity. Above 75 [LT] hold capacities, the mid points of the % change intervals oscillate around 5%. On the other hand, if the whole intervals are considered, the sizes of intervals are relatively greater up to 150 [LT] hold capacity, thereby increasing the uncertainty in the % change. Additionally, the intervals of the % changes extend below X-axis, implying a possible reduction in the displacements, thus in the sizes of the example designs. However, the sections of the intervals extending below the X-axis are smaller than those remain above the axis. Therefore, an increase in the displacements is assumed to be more probable. The example designs for 300 [LT] hold capacity presents a particular case, in which the introduction of Rules 7.12 changes neither the size of the displacement interval nor its location in the figure. Hence, the mid point of the interval for this hold capacity in Figure 7.41 is located at 0%, suggesting this large vessel may already satisfy Rules 7.12. From the owners' point of view, the costs of the vessels are possibly of more importance than the other parameters examined above. Figures 7.42 and 7.43, which are similar in content to the previous figures mentioned above, reveal the cost of including Rules 7.12 in to the preliminary design. Figure 7.42 suggests an increase in the cost of example designs after the addition of the rules. In the figure, the Echidna intervals of the two cases, that are with and without the rules, overlap for most of the hold capacities. When mid points of the intervals are considered, however, there is a clear increase in the costs as illustrated in the next figure (Figure 7.43). In percentage, the mid points of the % change intervals are distributed in close neighbourhood of 10% for hold capacities Chapter 7. SEAKEEPING CONSIDERATIONS 139 between 75 to 250 long tons. Up to 75 [LT] hold capacity, satisfying Rules 7.12 may cost as much as approximately 30% more, if the mid points of the intervals are considered. However, within this hold capacity range, the line has a negative slope, thus as the vessels become larger the amount of cost increase due to the rules decreases for the mid points of the intervals. Similar to the case of percentage change in the displacements, the sizes of the % change intervals are larger for the smaller hold capacities. Hence, up to 250 [LT] hold capacity, the intervals extend into the negative region, implying a reduction in the cost, as in the previous case. However, approximately 2/3's of the intervals are in the positive region in Figure 7.43. Therefore, an increase in the costs of the example designs is believed to be the case. There is no change in either the cost or the % change intervals of the example design for 300 [LT] hold capacity vessel (in Figures 7.42 and 7.43 respectively). Considering the fact that the size of the example design for this hold capacity also remained the same after introducing Rules 7.12, one may regard 300 [LT] hold capacity as the upper limit, beyond which Rules 7.12 may no longer be effective. Chapter 7. SEAKEEPING CONSIDERATIONS 140 1.4 1.2 1.0 0.8 M 0.6 | C D eav 0.4 a 0.2 0.0 o " \ - 1 \ With Rule Set II — Control case I I I I L 50 100 150 200 Hold Capacity [LT] 250 300 Figure 7.32: The effects of Rule Set II on rms heave amplitudes. Design and operational sea states are 5. 20 i — '15 10 i M 5 0 OH 0 With Rule Set II Control case 50 100 150 200 Hold Capacity [LT] 250 300 Figure 7.33: The effects of Rule Set II on rms pitch amplitudes. Design and operational sea states are 5. Chapter 7. SEAKEEPING CONSIDERATIONS 141 3 9 0 8 0 — 7 0 6 0 5 0 4 0 3 0 0 — — s ~ /k— 1^ :::::^ - r x .Afc. .... / JL **SJgPr With Rule Set II -\/Y b' ' ^ ^ ^ ^ ^ ^ • / — : /• Control case _.± i 1 i i i 1 , 1 , i 5 0 1 0 0 1 5 0 2 0 0 Hold Capacity [LT] 2 5 0 3 0 0 Figure 7.34: The effects of Rule Set II on lengths. Design-sea-state is 5. 2 8 2 6 2 4 B 2 2 d C D PQ 2 0 1 8 1 6 0 /A With Rule Set II — Control case 5 0 1 0 0 1 5 0 2 0 0 Hold Capacity [LT] 2 5 0 3 0 0 Figure 7.35: The effects of Rule Set II on beams. Design and operational sea states are 5. J Chapter 7. SEAKEEPING CONSIDERATIONS 142 20 0 50 100 150 200 250 300 H o l d Capaci ty [ L T ] Figure 7.36: % Change in the lengths of the example designs. Design-sea-state is 5. Chapter 7. SEAKEEPING CONSIDERATIONS 143 Figure 7.37: % Change in the beams of the example designs. Design-sea-state is 5. Chapter 7. SEAKEEPING CONSIDERATIONS 144 1 2 r— 1 1 — 1 0 — o3 S 1 9 Q 8 7 6 A -/ _ / With Rule Set II Control case J I i I i o 5 0 1 0 0 1 5 0 2 0 0 Hold Capacity [LT] 2 5 0 3 0 0 Figure 7.38: The effects of Rule Set II on drafts. Design and operational sea states are 5. 5 0 0 4 0 0 _ 3 0 0 C D s C D -fL 2 0 0 G/2 1 0 0 0 r J L With Rule Set II Control case j I 5 0 1 0 0 1 5 0 2 0 0 Hold Capacity [LT] 2 5 0 3 0 0 Figure 7.39: The effects of Rule Set II on displacements. Design-sea-state is 5. Chapter 7. SEAKEEPING CONSIDERATIONS 145 Figure 7.40: % Change in the drafts of the example designs. Design-sea-state is 5. Chapter 7. SEAKEEPING CONSIDERATIONS 146 Figure 7.41: % Change in the displacements of the example designs. Design-sea-state is 5. Chapter 7. SEAKEEPING CONSIDERATIONS 147 2 .5E06 r — 2 .0E06 d O 1.5E06 o 1.0E06 5 .0E05 With Rule Set H j L Control case _L 5 0 100 150 200 Hold Capacity p_T] 250 300 Figure 7.42: The effects of Rule Set II on costs. Design-sea-state is 5. Chapter 7. SEAKEEPING CONSIDERATIONS 148 100 Figure 7.43: % Change in the costs of the example designs. Design-sea-state is 5. Chapter 7. SEAKEEPING CONSIDERATIONS 149 7.4 Comparison of the outcomes of Rule Sets I and II In the previous sections two strategies (Rule Sets I and II) are described to try and improve the working and living conditions on board fishing vessels by reducing some vessel motions amplitudes. In Rule Set I, the peak frequencies of a vessel's response spectrum and the given sea state spectrum are considered to be the important design factors, hence, the concept of shifting the vessel's peak frequency away from the frequencies of the sea state. A fundamentally different strategy was employed in Rule Set II, in which rms values of heave and pitch motions were constrained. In both of the strategies (Rule Sets I and II), the same vessel motions, namely heave and pitch motions, are used as basis for the design criteria at the case studies. The following discussion compares the two rule sets in terms of the amount of changes they have caused to the control cases, when neither Rule Sets I and II were used during their design (control case). The difference in the ship costs predicted involved in the two rule sets is also examined. In Figures 7.44 to 7.48, the percentage change is the change in parameter values used in the two rule sets (Rule Sets I and II), and calculated by Equation 7.14. wnu , n n ParameterRule S e t n - ParameterRule Set I / r r , ToCnange — 100 («-14) ParameteTRuie Set I One characteristic feature of Figures 7.44 to 7.48 is that the difference between the two rule sets converges to zero as the hold capacity increases. This is in close agreement with the previous findings that for each rule set the amount of change from the original example design (control case) diminishes as hold capacity increases. This could be ex-plained as larger vessels experience less severe motions than smaller ones for a given sea state. Otherwise, they need to be designed for a higher sea state. The following differences in the design parameters due to Rule Sets I and II can be Chapter 7. SEAKEEPING CONSIDERATIONS 150 seen in Figures 7.44 to 7.48 : • Length : In general, Rule Set II resulted in longer designs in comparison to the designs obtained by Rule Set I (see Figure 7.44). • Beam : When the mid points of the intervals are considered, Rule Set I caused wider designs than those obtained using Rule Set II for hold capacities larger than 50 [LT] (see Figure 7.45). • Draft : Rule Set II resulted in deeper draft designs in comparison to Rule Set I designs, with diminishing difference as hold capacity increases. • Displacement : For larger hold capacities than 50 [LT], the designs obtained using Rule Set I have larger displacements than those obtained by Rule Set II (see Figure 7.47). • Cost : It is difficult to say with confidence from Figure 7.48) that one rule set produced less expensive vessels than the other. % Change values in this figure, start from ±50% at the lowest hold capacity, oscillates around the 0 line and converges to 0 as the hold capacity increases. Chapter 7. SEAKEEPING CONSIDERATIONS 151 150 100 c c o 50 0 3 0 -50 0 50 100 150 200 Hold Capacity [LT] 250 300 Figure 7.44: % Changes of the lengths of example designs obtained using Rule Sets I and II. Design-sea-state is 5. Chapter 7. SEAKEEPING CONSIDERATIONS 152 IP 50 40 30 20 10 0 •10 -20 -30 -40 0 50 100 150 200 Hold Capacity [LT] 250 300 Figure 7.45: % Changes of the beams of example designs obtained using Rule Sets I and II. Design-sea-state is 5. Chapter 7. SEAKEEPING CONSIDERATIONS 153 200 150 C O 100 50 0 -50 0 50 100 150 200 250 Hold Capacity [LT] 300 Figure 7.46: % Changes of the drafts of example designs obtained using Rule Sets I and II. Design-sea-state is 5. Chapter 7. SEAKEEPING CONSIDERATIONS 154 120 90 — «s 60 I o O S ^ 30 .s J 3 u 0 -30 -60 0 50 100 150 200 Hold Capacity [LT] 250 300 Figure 7.47: % Changes of the displacements of example designs obtained using Rule Sets I and II. Design-sea-state is 5. Chapter 7. SEAKEEPING CONSIDERATIONS 155 150 -S3 O o CD 100 50 0 -50 A / 100 0 J i I i I i L _ 50 100 150 200 Hold Capacity [LT] 250 300 Figure 7.48: % Changes of the costs of example designs obtained using Rule Sets I and II. Design-sea-state is 5. Chapter 7. SEAKEEPING CONSIDERATIONS 156 7.5 The Criterion Suggested by Kimura et al. [40] Earlier in Section 2.3, it was mentioned that Kimura et al. [40] suggested a criterion (Equation 2.3 on page 38) for quantifying the relationship between maintaining the human balance and the ship motions. In this section, the example designs obtained above by the two rule sets have been evaluated further using Equation 7.15. This equation is slightly different from the original equation (Equation 2.3) such that 0.1659 M l y term has been omitted. The omitted term correspond to the accelerations due to roll, sway and yaw motions. Considering the fact that roll motion is very much dependent on the center of gravity of the vessel and therefore loading conditions, and the example designs are assumed to be in head seas conditions (although one can not obtain a pure head sea condition in a real sea), the contribution to the acceleration due to roll motion has been neglected. Sway and yaw motions are considered to be small, thus they are neglected in Equation 7.15. Tm = M l . + 0.1133 M l , [m2/*5] (7.15) where M l is the first moment of power spectrum of the acceleration on deck, x,y and z subscripts represent the coordinate axes x (longitudinal), y (transverse) and z (vertical). T m values (Equation 7.15) were evaluated at three locations for each of the example designs, namely on the deck at the stern, at midships and at the bow. The stern has been chosen as the place where crew members perform some of the tasks such as handling the net or doors (see Figure 7.49). The midship area is generally the location of bridge and accommodation. At the bow, crew may have to handle the anchor or lines. In calculating of the acceleration terms, the following equations have been used: Chapter 7. SEAKEEPING CONSIDERATIONS 157 Figure 7.49: A crew member working at the stern. Vertical acceleration — < zcosO + ^ 6 at the bow ZCOS& at the midship (7.16) zcosO — at the stern Horizontal acceleration = < zsinO at the bow zsinO at the midship (7-17) zsin8 at the stern In these equations, 8 is the pitch amplitude, 6 corresponds to the amplitude of the pitch acceleration, z terms are due to the heave acceleration. In these equations, cen-trifugal acceleration due to pitch motion have been neglected as it is usually very small compared to the other terms ([13], [58]). The vertical and horizontal accelerations in the above formulation are the accelerations of the ship. When considering the compound Chapter 7. SEAKEEPING CONSIDERATIONS 158 accelerations (for that matter any other ship motions related terms such as velocities) in vertical, transverse or longitudinal directions the phase differences between the differ-ent type of ship motions, for example between heave and pitch motions, are important. However, it is considered a good approximation to assume that heave and pitch motions are in phase at the bow and out of phase at the stern of a ship. This assumption has been used in establishing Equations 7.16 and 7.17. For amidships, the effect of pitch in the horizontal accelerations has been neglected, this leaves only a heave related term in the horizontal acceleration in Equation 7.17. In Figures 7.50 to 7.61, T m values have been computed for sea states of 2 and 5 for the three groups of example designs (Control case and with Rule Sets I and II). It was therefore possible to quantify the effects of Rule Sets I and II on the conditions on board of the fishing vessels with respect to the criterion (Equation 7.15). In the evaluation 3 vessel speeds have been used, which are 0.01, 5 and 10 knots. 0.01[fcn] corresponds to a stationary vessel, which could be taking the catch in. 5[kn] speed is chosen in regard to trawling speed. This speed may vary from 3 to 5 knots. Finally, a speed of 10[fcn] may not be achievable in especially smaller vessels in a sea state of 5, however, it is done as an extreme case study. The Tm values obtained in this thesis are dramatically larger than T (Equation 2.3) values given in [40] especially for the ship speeds of 5 and 10 knots. However, the details of the calculations in the original paper were not clear in [40]. Therefore, Equation 7.15 is used as a merit of ranking for the example designs rather than the way it is suggested in [40]. For the following paragraphs (see also Figures 7.50 and 7.61) , % Change have been calculated by /w s~i i ., n n I -*m iRule Set I or II K-'-m/Control case /n 1 r ) \ % Change = 100 —T=-r (7.18) \-l-m)Control case Chapter 7. SEAKEEPING CONSIDERATIONS 159 Figures 7.50 and 7.53 are for operational-sea-states of 2 and 5 respectively. These two figures corresponds to a stationary vessel case. Up to approximately 75 [LT] hold capacities, there is not any clear difference between the three cases (Control case, Rule Sets I and II). However, for larger hold capacities (hence larger vessels) than 75[LT], Rule Set I has a visible improvement in Tm values. For sea state 2 (Figures 7.50 and 7.51), including Rule Set I ( 7.2 to 7.5), into the design resulted in approximately 20% reduction in Tm values on the average. Whereas Rule Set II ( 7.11 and 7.12) produced around a 10% reduction for the larger hold capacities, between 150 and 300 [LT]. For the designs obtained using Rule Set I, at the same vessel speed but an operational-sea-state of 5, T m values were decreased by approximately 15% at the bow, 20% at the amidship and 15% at the stern on the average. On the other hand, Rule Set II produced 2%, 5% and 0.7% (average) reductions on Tm values at the bow, amidship and stern respectively. Figures 7.54 and 7.57 are for operational-sea-states of 2 and 5 respectively but for a vessel at the speed of 5 [kn]. As in Figures 7.50 and 7.53, the larger reductions in T m values have been obtained for an operational-sea-state of 5. In the operational-sea-state 2 (see Figures 7.54 and 7.55), there were reductions of approximately 8% and 45% on the average at the bow and amidship respectively for the designs obtained using Rule Set I. For those obtained using Rule Set II, the reductions in Tm at the bow and amidship are 3.7 and 16% respectively. For Tm values at the stern, there are 10% and 1.8% average increases for Rule Sets I and II respectively. However, this situation has been improved for an operational-sea-state of 5, in which there are approximately 6% and 5% reductions for the stern values for Rule Sets I and II respectively. At the bow and amidship, Tm values have been reduced by 17% and 43% for Rule Set I and by 7% and 16% for Rule Set II. As mentioned earlier in this section, a 10 [kn] speed in an operational-sea-state of Chapter 7. SEAKEEPING CONSIDERATIONS 160 5 especially for small vessels is unrealistic to achieve. In any case, the analysis has been carried out for an operational-sea-state of 5 and a vessel speed of 10 [kn]. Figures 7.58 and 7.59 is for the operational-sea-state of 2 and 10 [kn] vessel speed and for the operational-sea-state of 5 the results are presented in Figures 7.56 and 7.57. As it appears in Figure 7.58 (operational-sea-state of 2), three cases (Control case, Rule Sets I and II) have very close values to each other after hold capacity of 75 [LT]. As far as the average values of Tm are concerned, reductions in Tm for an operational-sea-state of 2 was possible for amidships only (by 45% and 10% for Rule Sets I and II respectively). At the bow, Rule Sets I and II have 20% and 28% increases on the average in Tm, although at some hold capacities there were reductions. The average increases at the stern were 5% and 17% for Rule Sets I and II respectively. The situation seems to be improved for the operational-sea-state of 5 (Figures 7.60 and 7.61). On the average, there were reductions for the designs obtained using Rule Sets I and II. The largest average reductions were at the amidship with 48% and 18% for Rule Sets I and II respectively. For the bow, T m values have been reduced by 5% and 4% for Rule Sets I and II respectively. At the stern, Rule Set I produced approximately 5% increase unlike Rule Set II which has approximately 1% reduction on the average. However, if the results for hold capacities less than 75 [LT] are neglected, the new averages are around 25% and 5% reductions for stern for Rule Sets I and II respectively. 7.5.1 Summary In summary, the following are the outcome of the evaluation of the example designs obtained using Rule Sets I ( 7.2 to 7.5) and II ( 7.11 and 7.12) in the design in terms of Tm values, Equation 7.15 on page 156, (see Figures 7.50 to 7.61). Chapter 7. SEAKEEPING CONSIDERATIONS 161 • For Rule Sets I and II, most of the reductions in Tm values were achieved for larger hold capacities. • Rule Sets I caused lower Tm values than Rule Set II in general. • In using Rule Set I in the design, approximately 20% reductions in the Tm values of all three locations on the deck (stern, amidship and bow) was achieved for a vessel speed of 0.01 [kn] and an operational-sea-state of 2. Whereas Rule Set II caused an approximately 10% reduction on the average for the same speed and sea conditions. For the same speed (0.01 [fen]) but an operational-sea-state of 5, the average reductions in T m values were around 15% for Rule Set I and 2.5% for Rule Set II for all of the three locations on deck. • In the case of 5[kn] vessel speed, the largest reduction in Tm values occurred at the mid ship with approximately 40% and 15% reduction for Rule Sets I and II for both of the operational-sea-states. This is important because amidships is usually the place where the bridge is located and some other work activities of the sort of cleaning and the transferring the of the catch occur. Chapter 7. SEAKEEPING CONSIDERATIONS 162 Figure 7.50: Values of Tm (Equation 7.15), where Tm measures conditions to assess the ability of a crew member to maintain his balance (see page 38). Design-sea-state is 5 and operational-sea-state is 2; ship speed = 0.01[fcra]; midpoints of Echidna intervals were used. Chapter 7. SEAKEEPING CONSIDERATIONS S CD gP c « S CD f E E-CD 4 0 . 0 2 0 . 0 0 . 0 - 2 0 . 0 - 4 0 . 0 - 6 0 . 0 - 8 0 . 0 - 1 0 0 . 0 4 0 . 0 2 0 . 0 0 . 0 - 2 0 . 0 - 4 0 . 0 - 6 0 . 0 - 8 0 . 0 - 1 0 0 . 0 S T E R N / - A With Rule Set I With Rule Set H _L 0 . 0 5 0 . 0 1 0 0 . 0 1 5 0 . 0 2 0 0 . 0 2 5 0 . 0 3 0 0 . 0 Hold Capacity [LT] Figure 7.51: % Change in the values of T m with respect to "Control case", in which was no seakeeping considerations, (for the data in Figure 7.50). Chapter 7. SEAKEEPING CONSIDERATIONS 164 0.8 ^ 0.6 < s 0.2 0.0 Kimura et al's threshold value (0.22) = 8 = -Control case W i t h Rule Set 1 Wijth Rule ,Set LT -6 t o !^ 0.06 E < e 0.10 0.05 |— 0.00 0 Kimura et al's threshold value (0.22) X X- -50 100 150 200 Hold Capacity [LT] 250 300 Figure 7.52: Values of T m (Equation 7.15), where Tm measures conditions to assess the ability of a crew member to maintain his balance (see page 38). Design-sea-state is 5 and operational-sea-state is 5; ship speed — 0.01 [kn]; midpoints of Echidna intervals were used. Chapter 7. SEAKEEPING CONSIDERATIONS Figure 7.53: % Change in the values of Tm with respect to "Control case", in which was no seakeeping considerations, (for the data in Figure 7.52). Chapter 7. SEAKEEPING CONSIDERATIONS 166 o.o 0 50 100 150 200 H o l d C a p a c i t y [ L T ] 250 300 Figure 7.54: Values of Tm (Equation 7.15), where Tm measures conditions to assess the ability of a crew member to maintain his balance (see page 38). Design-sea-state is 5 and operational-sea-state is 2; ship speed = 5[fen]; midpoints of Echidna intervals were used. Chapter 7. SEAKEEPING CONSIDERATIONS 167 Figure 7.55: % Change in the values of Tm with respect to "Control case", in which there was no seakeeping considerations, (for the data in Figure 7.54). Chapter 7. SEAKEEPING CONSIDERATIONS 168 Figure 7.56: Values of Tm (Equation 7.15), where Tm measures conditions to assess the ability of a crew member to maintain his balance (see page 38). Design-sea-state is 5 and operational-sea-state is 5; ship speed = 5[fen]; midpoints of Echidna intervals were used. Chapter 7. SEAKEEPING CONSIDERATIONS Figure 7.57: % Change in the values of Tm with respect to "Control case", in which was no seakeeping considerations, (for the data in Figure 7.56). Chapter 7. SEAKEEPING CONSIDERATIONS 170 a E-1 Figure 7.58: Values of T m (Equation 7.15), where Tm measures conditions to assess the ability of a crew member to maintain his balance (see page 38). Design-sea-state is 5 and operational-sea-state is 2; ship speed = 10[fcra]; midpoints of Echidna intervals were used. Chapter 7. SEAKEEPING CONSIDERATIONS Figure 7.59: % Change in the values of Tm with respect to "Control case", in which was no seakeeping considerations, (for the data in Figure 7.58). Chapter 7. SEAKEEPING CONSIDERATIONS 172 Control case W i t h Rule Set I W i t h Rule Set H Figure 7.60: Values of Tm (Equation 7.15), where Tm measures conditions to assess the ability of a crew member to maintain his balance (see page 38). Design-sea-state is 5 and operational-sea-state is 5; ship speed = lOffera]; midpoints of Echidna intervals were used. Chapter 7. SEAKEEPING CONSIDERATIONS B c CD s .s CD a fr-> a gP C O -a 100 150 200 Hold Capacity [LT] 250 300 Figure 7.61: % Change in the values of Tm with respect to "Control case", in which was no seakeeping considerations, (for the data in Figure 7.60). Chapter 8 SWATH MONOHULL COMPARISON In this chapter, a comparison of monohull and SWATH type designs for the same hold capacity is presented. The unit costs per unit of displacement are assumed to be the same for both a monohull and a SWATH vessel [9]. Figure 8.1 displays the comparison of displacements, hence total costs are implied with the above assumption. In Figure 8.2, ratio of SWATH displacements versus monohull displacements are shown. Macgregor [47, p 5] reports that a SWATH and a monohull combatant displacement values differ by 30 to 60% (SWATH designs being the heavier one), and a SWATH would cost about 20% more than a monohull. These difference are lower than the ones found in the figures mentioned above. Bhattacharya [13, p 398] gives a comparison of heave and pitch motion amplitudes between a monohull and a SWATH vessel. Up to the ratio of (ship length/wave length) c. 1.8, the SWATH vessel's response is almost an order of magnitude less than that of the monohull. Considering the amount of improvement in monohull example designs reported above, Bhattacharya's findings suggest that SWATH concept has superior seakeeping characteristics. In another study, Cumming [28] reports the finding from a comparative seakeeping trial between a SWATH vessel M.V. Frederick G. Creed and a monohull vessel F.P.V. Louisbourg. The principal particulars of the two vessels are given in Tables 8.1 and 8.2. Some of the results from the sea trial are given in Table 8.3. Although the monohull F.P.V. Louisbourg is at least 3 times larger than the SWATH M.V. Frederick G. Creed 174 Chapter 8. SWATH MONOHULL COMPARISON 175 Table 8.1: Principal particulars of the SWATH M.V. Frederick G. Creed (from [28]). Length over all 20 [m] Pontoon length 18.3 [m] Beam 9.91 [m] Full load draft 2.44 [m] Full load displacement 74.2 [tons] Light ship displacement 55.9 [tons] Maximum forward speed 25.5 [kn] Complement 5 Table 8.2: Principal particulars of the monohull F.P.V. Louisbourg (from [28]). Length over all 38.1 [m] Length between perpendiculars 31.7 [m] Beam (moulded) 8.23 [m] Depth 3.40 [m] Draft (maximum) 2.49 [m] Displacement 247.06 [tons] Maximum forward speed 12 [kn] Complement 11 in terms of their displacements (see Tables 8.1 and 8.2), generally the SWATH vessel shows comparable motion characteristic to the monohull F.P.V. Louisbourg (see Table 8.3). According to Cumming, "... there was a general consensus among the observers that the 'Creed' was the superior seakeeping platform relative to the monohull". This is an important observation in relation to fishing, as a SWATH vessel can offer a more comfortable platform for the crew. However, as Figure 8.1 suggest, a SWATH fishing vessel is almost as twice large as a monohull for the same hold capacity requirement. It should be noted that in monohull design in Echidna the aim was not to achieve the same seakeeping characteristics as a SWATH. If one is still to continue with the Chapter 8. SWATH MONOHULL COMPARISON 176 Table 8.3: Comparison of some of the significant motion characteristics between M . V . Frederick G . Creed and F . P . V . Louisbourg (from [28]). Wave Forward M.V. Frederick G Creed F.P.V. Louisbourg heading speed Roll Pitch Heave Roll Pitch Heave (degrees) (knots) (degrees) (degrees) accel. (g) (degrees) (degrees) accel. (g) 0 12 3.2201 4.7178 0.0668 3.9391 3.1698 0.0405 45 12 8.2371 7.4474 0.1236 12.5545 4.8151 0.1437 90 12 5.0258 3.2415 0.1607 18.5344 3.8132 0.2176 135 12 3.7182 3.9646 0.2217 18.1182 5.6987 0.3634 180 12 1.6312 3.0755 0.2316 7.9494 7.0737 0.3914 0 5 6.3533 6.5355 0.2048 11.9796 6.4274 0.1903 90 5 7.3999 5.4146 0.2485 20.0605 6.7927 0.2738 180 5 7.1862 6.1317 0.1791 16.6082 6.9605 0.2490 comparison, as Figures 8.1 and 8.2 suggest S W A T H might still cost c. 1.5 to 2.2 times more than a monohull. Chapter 8. SWATH MONOHULL COMPARISON 177 1000 — Monohull - Control case -O Monohull - with Rule Set I 800 5 S3 s jo 600 Q 400 Monohull - with Rule Set LT / SWATH 200 -0 50 100 150 200 250 300 Hold Capacity [LT] Figure 8.1: Comparison of the displacements of SWATH and monohull type of fishing vessels obtained using Echidna. For monohulls, Control case corresponds to the case in which there was no seakeeping considerations in the design. For SWATH designs, no seakeeping considerations were used during the design. Chapter 8. SWATH MONOHULL COMPARISON 178 - SWATH/(Monohull - control case) • -1 -e-- SWATH / (Monohull - with Rule Set I) i -1 SWATH / (Monohull - with Rule Set II) 0 50 100 150 200 250 300 Hold Capacity [LTJ Figure 8.2: For the mid points of Echidna intervals, the values of the ratios (SWATH displacements over monohull displacements, for data given in Figure 8.1). Control case is the one, in which there was no seakeeping considerations in the design. Chapter 9 CONCLUSIONS In this study, the objective was to develop a design tool, which will allow some ergonomic considerations to be incorporated into the preliminary design of fishing vessels. Considering the nonlinear nature of ship design, at the beginning of the author's research it was not known if it was possible to obtain a solution by increasing the number of criteria (equations and/or constraints) used in ship design. The study showed that by the inclusion of crew safety/comfort considerations in the ship design cycle and for an assigned design-sea-state, the design tool is able to find a solution. While the solutions might not be unique and depend on the criteria used in the crew safety considerations the results (see Chapter 7) suggest that: • Such crew safety considerations can be included in the preliminary ship design procedure. • For a specific design-sea-state and a set of owner requirements one can establish preliminary principal ship dimensions that will increase the crew safety/comfort. • 2 simple rule sets (Rule Set I on page 100 ( 7.2 to 7.5) and Rule Set II on page 134 ( 7.11 and 7.12)) that were developed and tested suggest that similar simple rules might be used to improve seakeeping. The results obtained so far suggest that a solution is obtainable using the procedure developed (algorithms and criteria developed for the UBC series hull forms) and as ex-plained in thesis. This however does not limit the application of the process as UBC 179 Chapter 9. CONCLUSIONS 180 Series hull forms can be applied to any small craft design. In the detailed results of incorporating some seakeeping considerations for crew com-fort and safety (Rule Sets I and II), into ship design is presented in Chapter 7. The following conclusions are from the results presented in that chapter (considering the mid-points of the Echidna intervals): • The effects of introducing Rule Set I into the design are increases in displacement and costs of the designs. An approximately 25% on the average increase in displace-ments suggested up to 100 [LT] hold capacities. For the same hold capacity range the average increase in cost is also around 25%. The overall average increases that covers all hold capacity ranges are approximately 22% and 17% for displacements and costs respectively. • In the second group, Rule Set II produced an increase in displacements by up to 20% on the average up to 75 [LT] hold capacity. For larger hold capacities the average is around 5% for increase in displacements. The effects in the cost were similar to the case of displacements with roughly 30% and 10% average increases for 0 to 75 [LT] and 75 to 250 [LT] hold capacity ranges. • The example designs of SWATH type for the same hold capacity as monohull versions cost between 1.5 to 2.2 times more than an equivalent monohull design with some crew comfort criteria. However, the motion levels of the same SWATH vessel for a given sea conditions can be assumed superior to that of monohulls based on the information given by Bhattacharya [13] and Cumming [28]. Chapter 9. CONCLUSIONS 181 9.1 Suggestions for Future Work The rules used in Chapter 7 to improve seakeeping qualities are important in terms of demonstrating the concept. " What rules to use?" and "How many rules to use?" might be an interesting follow-up research in this area. A plausible extension to Rule Set I ( 7.2 to 7.5) might be to include the ratio between the amplitudes of the sea state's energy spectrum and the ship response spectrum at the peaks. Rule Set I define a frequency shift to higher frequencies. An alternative might be to attempt shifting the ship peak response frequencies to lower frequencies. This might be a more difficult case to obtain a solution, since there is a shift of the wave spectral peak towards lower frequency range for higher sea states. In Rule Set II ( 7.11 and 7.12), rms motion amplitudes were used. In these rules, only heave and pitch motion amplitudes were considered. The contributions of the roll motion to the overall motion characteristics of a ship (i.e. compound amplitudes, accelerations etc.) are also very important. However, roll motion is strongly dependent on the weight distribution in the vessel. In the thesis, the effects of roll motion were neglected. This was because the vessel is assumed to be heading into waves. In real life situations, there is not a pure head seas condition. A future research extension might be incorporation of roll motion considerations into the knowledge base for different scenarios (i.e. different headings and loading conditions). Furthermore, the following improvements to the current system might also be con-sidered. • The empirical formulae used in the preliminary design of monohull and SWATH vessels could be improved. Especially for SWATH designs there is a need for some revisions. Chapter 9. CONCLUSIONS 182 • If the system is intended to be used in real design cases, clearly the cost information incorporated in the knowledge base may soon become obsolete. Therefore, the cost terms, e.g. unit material costs etc., in the knowledge base must be updated regularly. The current cost model is very simple and could be improved. • Currently, for comparison of the designs total building cost is used. However, NPV (net present value) could be a better term to compare two designs. This could be implemented in the knowledge base. However, the implementation of NPV has its own difficulties such as the prediction of income throughout the lifetime of the vessel. There are some techniques suggested in the literature for dealing with such uncertainties [59], [70]. However, one can postulate plausible income/cost streams into the future to see to what extend NPV's affect the design parameters derived with Echidna. • Development of possibly other seakeeping rules similar to the ones developed in the thesis. The number of crew comfort/safety rules should be increased. There could be additional rules related to motion accelerations and jerk levels. However, one problem with such rules is that their highly nonlinear nature makes it very difficult to come up with an estimating formula that is reasonably simple yet accurate enough. Including several such rules simultaneously may not be necessary as some of the rules might become redundant. • Layout design of ships seems the next logical step towards safer vessels. Detailed layout design could be a follow up research in this context. Need to develop a guidance manual for use by designers. • In view of the comments by Dorval [32] that a crew member is 15 times more likely to die due to an occupational accident, one future extension of this thesis might Chapter 9. CONCLUSIONS 183 be devoted estimating the life cycle savings obtained by reducing motion levels at sea and comparing them to the increased cost predicted for these boats. This will allow to quantify the crew comfort/safety in monetary terms from the owners point of view. • Extend the strategy developed in this thesis to SWATH ships and evaluate them economically. Nomenclature B Ship's beam Cb Block coefficient Cp Prismatic coefficient D Ship's depth Fn Froude number (Fn = V/Jg/L) g Acceleration due to gravity (9.801 [m/s2] or 32.2 [ft/52] H Average wave height i f( i/3) Significant wave height # ( 1 / 3 ) 5 Significant wave height for sea state 5 k Wave number ( = 2ir/(WaveJength)) Lbp Ship's length (between perpendiculars) Loa Ship's length (over all) LT Long tons (1016 [kg]) Lwi Ship's length (water line) NHi/z Non-dimensional significant wave height {NHi/3 = # 1 / 3 / ^ ( 1 / 3 ) 5 ) NPV Net Present Value OBO Ore/Bulk/Oil carrier rms Root mean square SES Surface Effect Ship T Ship's mean draft T Average wave period for a given sea state 184 Nomenclature Tmax The wave period that maximizes wave energy spectrum for a given sea state V Ship speed z Heave displacement z r m s rms heave amplitude z Heave acceleration 6 Pitch displacement 6 Pitch acceleration BTm, rms pitch amplitude A Wave length A Average wave length for a given sea state w Encounter frequency [rad/s] ojn Non-dimensional encounter frequency (u>n = u>/^ Jg/T) Umax Frequency that maximizes the sea state's energy spectrum [rad/s] ijJnzp Non-dimensional frequency that corresponds to the ship's peak heave response (see Appendix E) u)nep Non-dimensioned frequency that corresponds to the ship's peak pitch response (see Appendix E) Bibliography [1] H. Aasjord. And^ ya-prosjektet, - tiltaksplaner for 15 fart<£ yer. Technical report, FTFI Report, September 1984. [2] A. F. Aboulazm. Ship motion and crew safety with application to fishing vessel. In International Symposium, Safety and Working Conditions aboard Fishing Vessels, Actes Proceedings, pages 228 - 235, Quebec - Canada, August 1989. [3] S. Akagi and K. Fujita. Building an expert system for engineering design based on the object-oriented knowledge representation concept. In Proceeding of the 1988 ASME Design Technology Conference - the Design Automation Conference, num-ber 14, pages 287 - 296, 1988. [4] A. Akinturk, M.Atlar, and S. M. Cah§al. Preliminary design of multi-hull fishing vessels using an expert system environment. In The Sixth International Symposium on Practical Design of Ships and Mobile Units, volume 2, pages 2.1288 - 2.1299, South Korea, September 1995. [5] K. Amagi and N. Kimura. A proposal to prevent occupational accidents in scal-lop beam trawlers. In In International Symposium Safety and Working Conditions Aboard Fishing Vessels, pages 300 - 309, Quebec, Canada, 1989. [6] R. N. Andrew and A. R. J. M. Lloyd. Full-scale comparative measurements of the behaviour of two frigates in severe head seas. Transactions of the Royal Institution of Naval Architects, 123:1 - 31, 1981. [7] David Andrews. Creative ship design. Transactions of the Royal Institution of Naval Architects, 123:447 - 471, 1981. [8] M. Andro, P. Dorval, G. Le Bouar, Y. Le Roy, G. Longchamps, and C. Roullot. Securite et conditions de travail a la peche, petite peche - peche cotiere - peche au large. Technical report, IFREMER - Centre de Brest, B.P. 337 - 29273 Brest Cedex - France, 1987. [9] M. Atlar and M. Guner. Methodology for initial design of a swath fishing vessel. Internal Report, Department of Marine Technology, University of Newcastle upon Tyne - U.K., October 1, 1994. 186 Bibliography 187 [10] M. Atlar and A. Scott. Application of the swath concept to fishing vessels, using de-sign analysis package. In International Workshop on Fisheries Energy Optimization, Vancouver, British Columbia, Canada, August 1989. [11] N. K. Bales. Optimizing the seakeeping performance of destroyer-type hulls. In 13th Symposium on Naval Hydrodynamics, Tokyo, Japan, 1980. [12] J. A. Basberg and K. O. $ Degard. Forprosjekt for sikkerhet i fiskeflaten. Technical report, FTFI Report, 1987. 40p. [13] Rameswar Bhattacharyya. Dynamics of Marine Vehicles. Ocean Engineering, a Wiley Series, A Wiley-Interscience publications. John Wiley & Sons, Toronto, 1978. [14] D.C. Bower. Fishing vessel optimization - a design tool. Master's thesis, The Uni-versity of British Columbia, Vancouver, B.C., Canada, April 1985. [15] Ivan Bratko. Prolog Programming for Artificial Intelligence. Addison-Wesley Pub-lishing Company, 2nd edition edition, 1990. [16] B. A. Bremdal and S. Kristiansen. Concept definition in marine system design. In Proceedings, First International Conference on Industrial and Engineering Applica-tions of Artificial Intelligence and Expert Systems, IEA/AIE, pages 403 - 421. ACM Press, 1988. [17] David C. Brown and B. Chandrasekaran. Design Problem Solving: Knowledge Struc-tures and Control Strategies. Research Notes in Artificial Intelligence. Morgan Kauf-mann Publishers, Inc., 1989. [18] David K. Brown. The value of reducing ship motions. Naval Engineers Journal, 97(3):41 - 46, March 1985. [19] B. G. Buchanan. New research on expert systems. In J. E. Hayes and D. Michie, editors, Machine Intelligence 10, pages 269 - 299. Wiley, London, 1982. [20] Roy Burcher and Louis Rydill. Concepts in Submarine Design. Cambridge Ocean Technology Series 2. Cambridge University Press, The University Press, Cambridge, Great Britain, 1994. [21] I. L. Buxton. Engineering economics applied to ship design. Transactions of the Royal Institute of Naval Architects, 114:409 - 428, 1972. [22] J, Carbjosa. Work accidents aboard Spanish fishing vessels. In International Sym-posium Safety and Working Conditions Aboard Fishing Vessels, 1989. Canada. Bibliography 188 [23] S. M. Cali§al, J. D. Howard, and J. Mikkelsen. Prediction of a vessel's motion in a given sea state. Marine Technology and SNAME News, 34(1):10 - 23, January 1997. [24] S. M. Cali§al and D. E. McGreer. An expert system for fishing vessel design. In J^th International Marine Systems Design Conference, Proceedings, pages 131 - 140. [25] S. M. Cah§al, J. Mikkelsen, D. McGreer, A. Aki nturk, W. Havens, and S. Joseph. Fishing vessel design with a constraint reasoning system. In T. K. S. Murthy and J. A. Alaez, editors, Fourth International Conference on Computer Aided Design, Manufacture and Operation in the Marine and Offshore Industries, Design of Marine and Offshore Structures, pages 183 - 200. Computational Mechanics Publications / Southampton, Boston and Elsevier Science Publishers / London, New York, 1992. [26] P. D. Chaplin and S. Burney. Casualties to fishing vessels and accidents to fishermen to december 1988. Technical report, England, December 1989. [27] R. D. Coyne, M. A. Rosenman, A. D. Radford, M. Balachandran, and J. S. Gero. Knowledge-Based Design Systems. Addison - Wesley Publishing Company, Univer-sity of Sydney, 1990. [28] D. Cumming. Results of a comparative seakeeping trial between a swath and a monohull. In First Canadian Marine Dynamics Conference, pages 83 - 89, 1991. St. John's, Canada. [29] C. Dai and S. Hambric. Propeller design optimization technique for the minimiza-tion of propeller induced vibration using artificial intelligence and numerical opti-mization. In H. Kim and J. W. Lee, editors, PRADS'95 The Sixth International Symposium on Practical Design of Ship and Mobile Units, Seaoul, Korea, volume 1, pages 1.194 - 1.206, 1995. [30] S. Doizhong and R. J. Forgie. Development of an expert system approach to an engineering design procedure, pages 975 - 989. [31] P. Dorval. Safety and working conditions in ocean fishing through a study of occu-pational accidents. In Jean-Louis Chaumel, editor, Proceedings of an International Symposium, Labor Developments in the Fishing Industry, pages 10 - 22, Rimouski, Quebec, CANADA, November 1983. GERMA (Groupe d'etude des ressources mar-itimes) and CIRAST (Centre d'intervention et de recherche pour l'amelioration des situations de travail, Department of Fisheries and Oceans - Ottawa 1984. [32] Patrick Dorval. The concept of integrated safety applied to work situations in sea fishing. In Proceedings World Symposium on Fishing Gear and Fishing Vessel De-sign, pages 547 - 554. Marine Institute, St. John's, Newfoundland, CANADA, 1988. Bibliography 189 J. H. Evans. Basic design concepts. Naval Engineers Journal, pages 671 - 678, Nov 1959. A. Farstad and E. Kjerstad. Produktivitet og arbeidsmilj^ > pet tralere, fangstbehan-dlingssytem for direkte sl<^ ying. Technical Report B 07, FTFI Report on working environment and catch-processing, 1987. 64p. M. J. Griffin. Handbook of Human Vibration. Academic Press Limited, San Diego, California, 1990. pp 988. A. Hansen and et al. Safety in norwegian fishing fleet. Technical report, Norway. W. Havens, S. Sidebottom, G. Sidebottom, J. Jones, and R. Ovans. Echidna : A constraint logic programming shell. In The Proceedings of the 1992 Pacific Rim International Conference on Artificial Intelligence, Seoul, 1992. (To be appeared). Victor E. Ivanov and Alexander V. Apollinariev. Dialogue - based software package of request for proposal for fishing vessel design, pages 2.1461 - 2.1474. C. Kennel. Swath ships. TNR Bulletin 7-5, The Society of Naval Architects and Marine Engineers, Jan. 1992. N. Kimura, K. Amagi, and Y. Inaba. On the relationship between the maintenance of human posture and the ship oscillatory motions. In In Proceedings World Symposium on Fishing Gear and Fishing Vessel Design, pages 564 - 568, Newfoundland, Canada, 1988. Einar Kjerstad and Jorgen Grinde. Working environment and health aspects in fishing vessels. In Proceedings World Symposium on Fishing Gear and Fishing Vessel Design. Marine Institute, St. John's, Newfoundland, CANADA, 1988. Ercan Kose. Risk assessment of fishing vessels. Master's thesis, Department of Marine Technology, The University Of Newcastle Upon Tyne, Newcastle - U.K., 1991. C. Kuo. The impact of safety requirements on shipbuilding. In H. Kim and J. W. Lee, editors, PRADS'95 The Sixth International Symposium on Practical Design of Ship and Mobile Units, Seaoul, Korea, volume 1, pages 2.1240 - 2.1251, 1995. Kyu Yeul Lee, Seung Wan Suh, and Soon-Hung Han. On the development of com-puter integrated basic ship design and performance analysis system. In J. B. Cald-well and G. Ward, editors, Practical Design of Ships and Mobile Units, Proceedings, pages 2.1269 - 2.1283. Elsevier applied Science, May 1992. Bibliography 190 [45] A. R. Lloyd. SEAKEEPING Ship Behaviour in Rough Weather. Ellis Horwood Series in Marine Technology. Ellis Horwood Limited, Publishers, Chichester. [46] T. D. Lyon and F. Mistree. A computer-based method for preliminary design of ships. Journal of Ship Research, 29(4):251 - 269, 1985. [47] James R. MacGregor. A Computer Aided Method For Preliminary Design of SWATH Ships. PhD thesis, Dept. of Naval Arch, and Ocean Eng., University of Glasgow, U.K., May 1989. [48] G. Masini, A. Napoli, D. Colnet, D. Leonard, and K. Tombre. Object Oriented Languages. The A. P. I. C. Series No 34. Academic Press, 1991. [49] M. Meek. The effect of operational experience on ship design. In Proceedings: The-ory and Practice of Marine Design Conference, volume 7, pages 79 - 91, Piccadilly Hotel, London, April 1982. [50] F. Mistree, W.F. Smith, B. A. Bras, J. K. Allen, and D. Muster. Decision-based design: A contemporary paradigm for ship design. SNAME Transactions, 98:565 -597, 1990. [51] R. Murphy and D. J. Sabat R. J. Taylor. Least cost ship characteristics by computer techniques. Marine Technology, 2(2):174 - 202, 1965. [52] A. Newell. Reasoning, problem solving and decision process: The problem space as a fundamental category. In L. Erlbaum, editor, Attention and Performance, VIII, pages 693 - 718, 1980. [53] H. Nowacki, F. Brusis, and P. M. Swift. Tanker preliminary design - an optimiza-tion problem with constarints. Transactions of the Society of Naval Architects and Marine Engineers, 78:357 - 390, 1970. [54] J. F. u. M. E. McCauley O'Hanlon. Motion sickness incidence as a function of the frequency and acceleration of vertical sinusoidal motion. Aerospace Med., 45(4):366 - 369, 1974. [55] Peter R. Payne. Method to quantify ride comfort and allowable accelerations. Avi-ation, Space and Environmental Medicine, 49(1 - 6):226 - 269, 1978. [56] B. J. W. Pingree. A review of human performance in a ship motion environment. Warship Technology, 39:73 - 76, May 1988. [57] T. Ray, R. P. Gokarn, and 0. P. Sha. Analysis of ship design under uncertainty in the multiple criteria decision making framework. In H. Kim and J. W. Lee, editors, PRADS'95 The Sixth International Symposium on Practical Design of Ship and Mobile Units, Seaoul, Korea, volume 1, pages 2.1264 - 2.1275, 1995. Bibliography 191 T. Sabuncu. Ship Motions. Number 1248. T.C. Istanbul Teknik Universitesi Kutiiphanesi, Istanbul, Turkey, 1983. (in Turkish). Pratyush Sen. Methods for incorporating uncertainty in preliminary ship ship de-sign. North East Coast Institution of Engineers and Shipbuilders, 102(6):59 - 70, November 1985. Susan Sidebottom, William Havens, and Sumo Kindersley. Echidna Constraint Rea-soning System (Version 1) : Programming Manual. Marine Media Management Ltd., Expert Systems Laboratory, Centre for Systems Science, Simon Fraser Univer-sity, British Columbia, CANADA V5A 1S6, manual version 2.0 edition, December 1992. Herbert A. Simon. The Sciences of the Artificial. MIT Press Series in Logic Pro-gramming. The MIT Press, 2nd edition, 1992. G. K. Smith and R. G. Woodhead. A design scheme for structures. Transactions of the Royal Institute of Naval Architecs, 116:47 - 66, 1974. J. Stoop. Safety and Design Process. PhD thesis, Universiteitsdrukkerij Technische, Universiteit Delf, Netherland, 1990. Robert Taggart, editor. Ship Design and Construction. The Society of Naval Archi-tects and Marine Engineers, May 1986. 2nd printing. William A. Taylor. What Every Engineer Should Know About Artificial Intelligence. The MIT Press, Massachusetts Institute of Technology, Cambridge, Massachusetts 02142, 1989. 2nd printing. Transportation Safety Board of Canada. Transportation safety board statistical summary - marine occurences. Technical report, 1990. Cat. No. TU1-1/1990. C. N. Tupper. Fishing and ship motions - design considerations based on observa-tions of operations. Ocean Research Corporation, Kennebunk, Maine, USA. D. A. Walden and P. Grundmann. Methods for designing hull forms with reduced motions and dry decks. Naval Engineers Journal, 97(4):214 - 223, May 1985. M. Welsh, I. L. Buxton, and W. Hills. The application of an expert system to ship concept design investigations. Transactions of the Royal Institute of Naval Architects, 133:99 - 122, May 1991. J. Wolfram. Uncertainty in engineering economics and ship design. North East Coast Institution of Engineers and Shipbuilders, 96(2):77 - 94, Feb 1980. Bibliography 192 [71] Vedran Zanic, Izvor Grubisic, and Giorgio Trincas. Multi-attribute decision making system based on random generation of nondominated solutions: An application to fishing vessel design, pages 2.1443 - 2.1459. [72] H. H. Zhou, B. G. Silverman, and J. Simkol. Cleer : An ai system developed to assist equipment arrangements on warships. Naval Engineers Journal, pages 127 -137, May 1989. Appendix A Echidna - Knowledge Base Environment Echidna is an expert system developed at Simon Fraser University's Expert Systems Laboratory. It is described as a logic programming language embedded in an object oriented framework [60]. Some of its general features have been described earlier. In this appendix, programming aspects is summarized. The Echidna shell can perform computations on variables defined as real intervals (e.g. length = [50.0, 60.0]). Hence most of the solutions in this study were in form of real intervals. Echidna refines the intervals of variables by propagating constraints on them. Once a solution is reached, one can further refine a particular interval given by Echidna. This could be done either by additional constraints or asking the Echidna to refine specific variables randomly. The following hypothetical example is given to illustrate this point clearly. Example: Lets design a hypothetical simple barge that can carry 25 [tons] of material. For simplicity, we are only concerned if the barge floats at its design draft reading when loaded. Hence, the total weight of the barge should be equal to its displacement. The following equations are used to calculate relevant terms. Hull weight PlateJthickness x Total-surf ace-area (A . l ) Displacement 1.025Length * Beam * Draft (A.2) Total weight Cargo + Hull weight (A.3) 193 Appendix A. Echidna - Knowledge Base Environment 194 The following constraints are also used in the example. 0.97 Displacement < Total weight < Displacement (A.4) 2.0 < (Length/Beam) < 5.0 (A.5) 2.0 < (Beam/Draft) < 5.0 (A.6) The knowledge base written for this example, run time commands issued in the Echidna shell and the corresponding example output are given at the end of this ap-pendix. In Table A . l , the second column shows the initial intervals assigned to the design parameters, "unbound" means the parameter defined as a real interval variable, however, the initial interval is not defined explicitly. In the third column, when Echidna reached a solution, some of the parameter intervals shrank, some did not change and the previously "unbound" ones were assigned an explicit interval. However, these intervals are too large to define a barge explicitly. One can impose more constraints. The effect will be some (or all) of the intervals will be further shrunk (see the fourth column in the table). If the intervals given by Echidna are still not sufficiently small, Echidna can randomly selects "a" set of smaller parameter intervals that satisfy the design constraints given above, as is the case in the fifth column in Table A . l . In the following texts, the rest of the line after a % character is a comment. A schema is a definition of a class. It contains the description of the structure and composition of every possible instances of the class. A class is defined as "... a family of objects having the same structure and behaviour. It describes a set of data and a set of procedures or functions." [48, page 16]. A class then has two parts, a static part, which are the fields holding the data, and a dynamic part, which represents the common behaviour of the objects, i.e. the procedures or the methods. The following example shows a class definition for rectangles. Appendix A. Echidna - Knowledge Base Environment 195 Table A . l : Refinement of intervals during the example barge design in Echidna Parameter Initial 1st solution For (L/B) > 4.95 Randomly refined Length [m] [1, 200] [1, 200] [4.1, 200] [11.445, 11.448] Beam [m] [1, 200] [1, 100.5] [1, 41.42] [2.360, 2.363] Draft [m] [1, 200] [1, 50.75] [1, 21.21] [1.030, 1.033] Depth [m] [1, 200] [1, 56.968] [1, 24.32] [1.14, 1.15] Freeboard [m] unbound [0.115, 5.83] [0.115, 2.44] [0.123, 0.123] A [tons] unbound [1.02, 10052.4] [1.02, 3862.06] [29.19, 29.26] Hull weight [tons] unbound [0.26, 9698.4] [0.26, 3689.4] [4.19, 4.20] Class { Rectangle Fields Length Beam Methods Area = LengthxBeam Circumference = 2 (Length + Beam) Figure A . l : An example class definition for rectangles (the object in this example). Appendix A. Echidna - Knowledge Base Environment 196 The syntax of a schema is given as follows: schema name:superclass_name •c the code that define schema instance variables, method definitions, etc. } A . l Run-time user commands '/, Load knowledge base f i l e "barge.kb" load barge.kb */, Define a schema variable of the type "weights"} weights Vessell. '/. Instantiate the schema variable defined above} Vessell i s a weights([25.0, 25.001] _Hold_l, [0.97, 1.0] B a l l a s t _ l ) . '/, Print the schema variable after instantiation} print (Vessell). '/, Access some of the schema instance variables} Vessell:length(interval LWL). Vessell:beam_overall(interval Beam). Vessell:depth_of_vessel(interval Depth). Vessell:draft_of.vessel(interval Draft). Vessell:fboard(interval Freeboard). Vessell:ltob(interval Lwl_to_Beam) . Vessell:btot(interval Beam_to_Draft). Vessell:btod(interval Beam_to_Depth). Appendix A. Echidna - Knowledge Base Environment Vessell:ttod(interval Draft_to_depth). Vessell:disp(Displacement). */, Additional user defined run-time constraint on '/, the (length/beam) ratio Lwl\_to\_Beam >= 4.95. '/, Print the schema variable i n order to see the effects of '/, the constraint of the (length/beam) ratio print Vessell '/, Ask the Echidna to randomly refine length, beam, draft '/, and depth parameters split(LWL, Beam, Draft, Depth), precision(16). '/, Print the schema variable i n order to see the effects of '/, random refinement of the parameters intervals print Vessell '/, Show the queries issued so far queries A.2 Knowledge base for the barge example '/, General constants definitions ftdefine density 7.8 '/, Density of steel 7.8 [ton/nT3] #define p i 3.141592654 '/, Sea water density ft3/longTon #define seadensft3LT 35.00458671 Appendix A. Echidna - Knowledge Base Environment '/, User defined types hu l l .mater ia l = {aluminum, steel}, gear.type = {seiner, trawler}. range_for.length = [1.0, 200.0]. range_for_beam = [1.0, 200.0]. range.for.depth = [1.0, 200.0]. range. for .draft = [1.0, 200.0]. schema barge i 7, Persistant instance v a r i a b l e s . . . range_for_length Length. '/, Length in m range_for_beam Beam. '/, Beam i n m range_for_depth Depth. '/, Depth i n m range_for_draft Draft . '/, Draft i n m in terva l Freeboard. '/, Freeboard in m '/, Non-dimensional parameters.. . in terva l LB. in terva l BD. in terva l BT. in terva l TD. '/, The following are i n long tonnes. in terva l Hold.capacity. in terva l HullWeight. in terva l TotalWeight. in terva l Displacement. '/, Length i n m '/, Beam in m '/, Depth in m % Draft i n m Appendix A. Echidna - Knowledge Base Environment interval Ballast. '/, As percentage of displacement '/, Accessors. .. hold_capacity(Hold_capacity). length(Length). beam_overall(Beam). depth_of.vessel(Depth). draft_of.vessel(Draft). fboard(Freeboard). ltob(LB). btot(BT). btod(BD). ttod(TD). disp(Displacement). hullWeight(HullWeight). bal l a s t ( B a l l a s t ) . } schema particulars:barge •C '/, Instance i n i t i a l i z a t i o n methods. . . particulars. particulars(Hold_capacity) :-length(Length), beam_overall(Beam), Appendix A. Echidna - Knowledge Base Environment depth_of.vessel(Depth), draft_of .vessel(Draft) , freeboard(Freeboard), Draft =:= Depth - Freeboard, lb (LB) , beam_to_draft(BT), beam_to_depth(BD), draft_to_depth(TD). '/, Methods used to estimate i n i t i a l ship dimensions y, order l b . lb(LB) : -LB =:= Length / Beam, LB >= 2.0, LB =< 5.0. lb(LB) : -LB =:= Length / Beam, order beam_to_draft. beam_to_draft(BT) : -BT =:= Beam / Draft , BT >= 2.0, BT =< 5.0. beam_to_draft(BT) : -BT =:= Beam / Draft . beam_to_depth(Beam_to_depth) : -Beam_to_depth =:= Beam / Depth, draft_to_depth(TD) : -Appendix A. Echidna - Knowledge Base Environment 201 TD =:= Draft / Depth, order freeboard, freeboard(Freeboard) :-Freeboard =:= 0.115 * Draft. '/, from UBC series models, freeboard(Freeboard). > schema weights:particulars i '/, Instance i n i t i a l i z a t i o n methods . . . weights(Hold_capacity, Ballast) :-particulars(Hold_capacity), displacement(Displacement), hullWeight(HullWeight), totalWeight(TotalWeight), bal l a s t ( B a l l a s t ) . '/, DISPLACEMENT displacement(Displacement) :-Displacement > 0, Displacement =:= 1.025*Length*Beam*Draft. '/. WEIGHT GROUPS '/. HULL WEIGHT Appendix A. Echidna - Knowledge Base Environment hullWeight(HullWeight) : -in terva l Thickness =:= (0.056*Length + 5.5)/1000.0, in terva l Surface_area =:= 2.0*(Length*Beam+Beam*Depth+Length*Depth), HullWeight =:= 7.8*Thickness*Surface.area, HullWeight > 0. '/. TOTAL WEIGHT '/. order totalWeight. totalWeight(TotalWeight) : -in terva l Sum_of.Weights =:= HullWeight + Hold_capacity, in terva l LowerLimit =:= Ballast * Displacement, in terva l UpperLimit =:= Displacement, TotalWeight =< UpperLimit, TotalWeight >= LowerLimit, TotalWeight =:= Sum_of.Weights. } A.3 A sample output for the barge example Echidna Version 0.947 beta Compiled: Thu Nov 25 14:42:16 PST 1993 (c) Copyright Intel l igent Systems Lab. Simon Fraser Univers i ty , 1991, 1993 A l l r ights reserved Appendix A. Echidna - Knowledge Base Environment loading data base f i l e "rina.db" . . . weights.O = { range_for_length Length = [1, 200]. range_for_beam Beam = [ i , 100.5]. range_for.depth Depth = [1, 56.96875]. range_for_draft Draft = [1, 50.75], in terva l Freeboard = [0.115, 5.836948]. in terva l LB = [1.994697, 5.003655]. in terva l BD = [0.01755348, 100.5]. in terva l BT = [1.998462, 5.000922]. in terva l TD = [0.01755348, 50.75]. in terva l Hold_capacity = [25, 25.001]. in terva l HullWeight = [0.260004, 9698.486]. in terva l TotalWeight = [25.25999, 9723.487]. in terva l Displacement = [1.024969, 10052.48]. in terva l Bal last = [0.97, 1], }weights.0 = { range_for.length Length = [4.109375, 200]. range.for_beam Beam = [1, 41.42188]. range_for.depth Depth = [1, 24.32031]. range_for_draft Draft = [1, 21.21094]. in terva l Freeboard = [0.115, 2.439956]. in terva l LB = [4.948725, 5.003655]. in terva l BD = [0.03901885, 41.42425]. in terva l BT = [1.998462, 5.000922]. in terva l TD = [0.04077695, 21.21129]. Appendix A. Echidna - Knowledge Base Environment 204 in terva l Hold_capacity = [25, 25.001]. in terva l HullWeight = [0.260004, 3689.481]. i n t e r v a l TotalWeight = [25.25999, 3715.073]. in terva l Displacement = [1.024969, 3862.061]. in terva l Bal last = [0.97, 1]. } weights.O = { range_for_length Length = [11.68544, 11.68848]. range_for.beam Beam = [2.360352, 2.363388]. range_for_depth Depth = [1.148788, 1.151825]. range_for_draft Draft = [1.030365, 1.033401]. in terva l Freeboard = [0.118492, 0.1188412]. in terva l LB = [4.95, 4.952007]. in terva l BD = [2.049228, 2.057288]. in terva l BT = [2.28406, 2.293739]. in terva l TD = [0.89455, 0.8995577]. in terva l Hold_capacity = [25, 25.001]. in terva l HullWeight = [4.197202, 4.206248], in terva l TotalWeight = [29.19718, 29.20726]. in terva l Displacement = [29.19546, 29.26168]. in terva l Bal last = [0.97, 1]. done #0 . done #1 Vessel l i s a weights, Vessell:weights(_Hold_l, B a l l a s t _ l ) . Appendix A. Echidna - Knowledge Base Environment done #2 pr in t (Vesse l l ) . done #3 Vessell:length(LWL). done #4 Vessell:beam.overall(Beam). done #5 Vessell:depth_of.vessel(Depth). done #6 Vessel l :draft_of .vessel(Draft) . done #7 Vessell:fboard(Freeboard). done #8 Vessell:ltob(Lwl_to_Beam). done #9 Vessell:btot(Beam_to_Draft). done #10 Vessell:btod(Beam_to_Depth). done #11 Vessell:ttod(Draft_to_depth). done #12 Vessell:disp(Displacement). done #13 Lwl_to_Beam >= 4.95. done #14 sp l i t (LUL, Beam, Draft , Depth), done #15 precision(16). inside echidna 33> Appendix B Monohull Design Algorithms The following design formulae are originally developed by Calisal and McGreer [24]. The units for linear dimensions and weight groups are feet and [LT] respectively. B. l Linear Dimensions Cubic Number (Nc): Nc = (Hold Capacity + 28.496)/l. 1313 (if hull material is aluminum) Nc = (Hold Capacity + 15.395)/0.6606 (if hull material is steel) (B.l) License Length (Llicence): Licence = Speed,2/1A (B.2) Length estimate based on hold capacity (Li): Xi = 3.3 (28.5 Holdcapacity)03669 (B.3) Length estimate based on (L/B), (B/D) and Nc (L2): L 2 = (100 Nc (4)2 ^ ) 0 3 3 3 3 3 3 + 5 (B.4) B D Waterline length (LWL)'-LWL = smaller of (Liicence,L2) (B-5) 206 Appendix B. Monohull Design Algorithms 207 Length between perpendiculars (LBP)I LBP i-'WL 1.066 (B.6) Length over all (LOA)'-LQA — 1.066 LWL (B.7) Beam over all (B): B > B < LLW JLW 0 . 0 1 7 7 W + 1.690323 (B.8) Depth (D): D < -^-^ (if hull material is aluminum) D < —— (if hull material is steel) ~ 2.2 w ' (B.9) Draft (T): T < D-Freeboard (B.10) Freeboard (based on U B C series model data): Freeboard = 0.1157/ ( B . l l ) § (L = LWL)-Coefficient C„ — = 0.017742 LX + 1.690323 B Cm — LWL B D 100 (B.12) (B.13) Appendix B. Monohull Design Algorithms 208 B.2 Weight Estimation A more detailed description of the formulae given below can be found in [25]. Hull Weight W, (New Brunswick Department of Fisheries): These equations determine the weight of hull, deck and deckhouse in [LT] for both steel and aluminum construction. Wa — k 0.00002813(x3 - 173.52a;2 + 41880a; - 0.0000158) (B.14) where k = 1 for steel, and k= 0.55 for aluminum. a_10O(MB±m i (B.15) V 3000 / v ; Outfit Weight W0: This weight includes the joiner work, interior, piping, and miscellaneous equipment. W0 = 0.25 L W L B (B.16) Machinery Weight Wm: This weight includes the main engine, gear box, shafting, and the propeller. where Cn is the cubic number Cn — ^ f ^ p -Stores Weight Ww: The stores weight includes the fuel, water, water ballast and provisions. Ww = 3.618L - 39.03 (B.18) Appendix B. Monohull Design Algorithms 209 Gear Weight Wa: For a seiner vessel the gear weight includes the net and the skiff, Wg = 7.23 and for a trawler the gear weight includes the net and wire rope. Wg = 3.05 Hold capacity: Fish, ice and storage containers in hold. The Echidna program balances the displacement of the vessel with the sum of the various weight groups. The constraint that the displacement must be greater than the weight determines the ranges for the principal dimensions of the vessel. B.3 Resistance and Powering The next step in the Echidna system is the determination of the calm water resistance for the vessel. The Echidna program utilizes a resistance algorithm developed for the U B C Series. This algorithm is based on regression analysis of model test data of the series. The algorithm is based on an equation developed by Oortmerssen for small vessels. Cr = d e 9 + C2e + C3e * sin F n 2 + C4e 'S cos F 2 (B.19) where Ci = di,0 + ditlji + di>2 (I) + 4 , 3 | + diA ( | ) (B.20) and Appendix B. Monohull Design Algorithms 210 Table B.l: Coefficients for UBC Series Resistance Algorithm for C\> = 0.615 i 1 2 3 4 difl 0.074654 0.076958 -0.000162 -0.018642 di,i 0.001879 7.7 x 10-5 7.4 x 10~5 1.05 x 10-5 di,2 -0.000701 1.8 x 10-5 0.000907 0.000119 di,3 -0.05158 -0.005247 -0.001137 -0.001944 di,i 0.009871 0.000572 -0.000661 0.001169 m = 0.14347C"21976 (B.21) with Cp defined as the prismatic coefficient of the vessel. The coefficients, diwere determined from regression analysis of model test data for the UBC Series. An example set of coefficients for one block coefficient, Cb=0.615 are given in Table B. l . Once the wave resistance coefficient, C P , is determined, the total resistance coefficient of the vessel, Ct, is determined by the addition of the frictional resistance coefficient, Cj, which is based on the Reynolds number, ft, of the vessel, and an allowance of 0.0004 to account for the roughness of the vessel. The frictional resistance is given by: 0.075 (log ft-2)2 (B.22) and Ct = Cr + Cf + 0.0004 (B.23) The total resistance of the vessel, R, is estimated from the standard equation: Appendix B. Monohull Design Algorithms 211 R = Ct\pSV2 (B.24) where p is the water density, V is the vessel speed, and S is the wetted surface area of the vessel. B .4 Stability The center of gravity of the vessel is calculated and compared to maximum center of gravity criteria developed for the U B C series. The stability criteria used in the knowledge base was developed for the U B C Series [25]. To calculate the center of gravity of the vessel, the center of gravity of each weight group is estimated by the following equations. The value ^jj- is the ratio of the height of the center of gravity above the baseline of the vessel to the overall depth of the vessel. The equations for hull weight and outfit were adjusted by adding 0.05 and 0.15 respectively to give a lightship center of gravity closer to those measured by the inclining tests of 10 typical vessels [25]. The center of gravities for the weight groups as follows: Hull Weight: f = ^°- m Outfit Weight: 4? = + 1-19 (B.26) D 8820 v ' Machinery Weight: D 1309000 1 ' Appendix B. Monohull Design Algorithms 212 Stores Weight: KG — = 0.75 (B.28) Gear Weight: KG =jf- = 1.66 (B.29) Hold Weight: = 0.60 (B.30) B.5 Cost Estimation Hull Cost [Can. $]: HulLcost = 11064 W, (if hull material is aluminum) Hull-cost = 4155 Ws (if hull material is steel) (B.31) Outfit Cost [Can. $]: Outfit-cost = 120-OL^ (B.32) Shaft horse power [HP]: SHP = EHP/(0.97 * 0.58) (B.33) Machinery cost [Can. $]: Machinery-cost =:= 250 SHP (B.34) Appendix B. Monohull Design Algorithms 213 Fuel consumption [Gallon/hr]: Fuel cost [Can. $]: GalJir = ^SHP (B.35) Fuel-cost = 2.5 J 5°°, GalJir (B.36) Speed Gear cost [Can. $]: Gear-cost = 200000 (B.37) Capital cost [Can. $]: C apital-cost = 1.3221 (Hull-cost + Outfit-cost + Mach-cost + Gear-cost) (B.38) Morgage cost [Can. $]: Interest rate Capital cost Mortgage-cost = — — - (B.39) Depreciation cost [Can. $]: Depreciation-cost = 0.045 Capital-cost (B.40) Insurance cost [Can. $]: Insurance-cost = 0.020 Capital-cost (if Capital-cost < 500000) Insurance-cost = 0.015 Capital-cost (if 500000 < Capital-cost < 2000000) Insurance-cost = 0.010 Capital-cost (if Capital-cost > 2000000) (B-41) Total cost [Can. $]: Total-cost = C apital-cost + Mortgage-cost + Depreciation-cost + Insurance-cost (B.42) \ Appendix C SWATH design algorithms In the preliminary S W A T H vessel design knowledge base the algorithms given in [47] and [9] have been used. The algorithms given in this appendix are derived from the available data published in the literature by Macgregor[47] and Atlar and Guner [9]. The data given in [47] suggest that a variety of S W A T H vessels' (from fishing vessels to patrol boats) principal particulars were included in the data. Therefore, the accuracy and range of applicability of the following algorithms depend upon the reliability of the data from which these algorithms were derived. C . l Linear Dimensions The first step in the S W A T H fishing vessel design is to estimate the principal dimensions. The following estimating formulae are given in detail in [47, 9] : Overall length (LQA) : LQA = 5.33A3 (C . l ) Length between perpendiculars ( L B P ) : L B P = 0.886LQA - 0.47 (C.2) Overall beam (BOA) : 214 Appendix C. SWATH design algorithms 215 Draft (T) Depth (D) BOA = 3.05[A3]°- 9 7 2 with 1.1 < ^ < 3.3 A3 T = 0.588[Ai]0-9 7 2 Circular hulls T = 0.583[A3]0-9 1 7 Non - circular hulls (C.3) Also, Or, D — B c + T (C.4) DWD = 0.833A 3 to wetdeck DMD = 1.167A3" to maindeck (C.5) (C.6) DWD = 1.55T to wetdeck DMD = 2.10T to maindeck (C.7) (C.8) Box clearance (Be) : Freeboard Be = 0.041 + 0.038Fn (C.9) freeboard = 0.134 + 0.593As (CIO) Appendix C. SWATH design algorithms 216 Lower Hull (LH) Dimensions Lower hull cross section characteristics : Circular hull : Diameter DIALH = 0.134 + 0.593As ( C . l l ) Non — circular hull : Non — circular detph DJJ = 0 . 8 3 3 A ' (C.12) Non — circular hull : Non — circular beam Bg = 1 . 1 6 7 A 5 (C.13) Length of lower hull L H : Long strutted LH = 0 . 9 3 1 I O j 4 - 0.91 ( C 1 4 ) Short strutted LH — LOA (C.15) Lower hull cross sectional area A H : Circular hull AH = 0 . 0 1 2 4 L ^ 6 7 6 (C.16) Non- circular hull AH = 0 . 0 1 8 2 L # 6 5 4 (C.17) Lower hull prismatic coefficient C p H : CPH = 1 - 0 . 3 3 3 ^ - 0 . 4 6 7 ^ (C.18) LH LH Where LH is lower hull length, LT is tail length, and Xjy is nose length (Lg = LT + Lparallel-body + ). Appendix C. SWATH design algorithms 217 C.2 Strut(S) Linear Dimensions Length Ls : Long strutted Ls = 0.21 + 0.903L O x (C.19) Short strutted Ls = 0.75LOA (C.20) Thickness ts : For single struts : ts = 0.149As (C.21) Strut coefficients : Waterplane area coefficient, (Cw) : C w = 1 . 0 2 1 5 ^ - 0 . 3 3 3 ^ (C.22) Slenderness coefficient (^ -) : Long strutted = 0.02 • • • 0.04 (C.23) Short strutted = 0.03 • • • 0.05 (C.24) Strut setback ratio (f^8-) : v L O A ' S s ~ 0 . 0 7 (C.25) •>OA C .3 Weight Estimation Structural weight (Ws) '• In [47] gives following formulae for the overall structural weight (Ws) versus displacement Appendix C. SWATH design algorithms 218 ratio. Mild steel = 0.425 - 1.75 * 10~6A (C.26) Hybrid ^ = 0.417 - 8.5 * 10~6A (C.27) Aluminium ^ = 0.388 - 3.11 * 10 _ 5A (C.28) where Ws and A are in tons. It has also been noted that aluminum and hybrid designs are limited to vessels of less than 6000 tons, which is not of any particular concern in our case. Machinery weight (WM) '• High speed diesel machinery WM = 0.007Pjif [tons] (C.29) Medium speed diesel machinery WM = 0.00113Pflf [tons] (C.30) Diesel electric machinery WM = 0.543-P^679 [tons] (C.31) where PM is the installed power at 900 fixed RPM and is in metric HP. Auxiliary system weight (WAUX) • WAUX = 0.147A- 0 0 0 2 [tons] (C.32) Outfit weight (Wo) : = 0.1890.147A"0119 [tons] (C.33) Fuel weight (WF) : WF = (WFM + WFC + WFG)(m + 1) [tons] (C.34) where m is the margin for extra time and the trapped fuel in the tanks. WFM is the fuel required at the design speed and given by Appendix C. SWATH design algorithms 219 WFM = ^-{PMSFCM)10-6 [tons] (C.35) VM Wpc is the fuel required at the cruise speed and given by P-C , „ r,^,-, X 1 „ _ 6 WFC = -^{PcSFCc^O'6 [tons] (C.36) vc WFG is the fuel required for generator and given by WFG = [^ + ^](PGSFCG)lO-6 [tons] (C.37) VM VC In the above equations : VM,VC are design and cruise speeds in knots respectively, P-MiRc are operational requirements of range in nautical miles for the design and cruise conditions respectively, PMIPC are required propulsive power in [kW] in design and cruise conditions respectively, PQ is the average generating load required by the vessel services and given by PG = 0.25(1.002A°- 9 2 4) [kW] (C.38) In the following equations (C.39,C.40), SFCM,C,G are specific fuel consumptions in [ Ballon 1 In the case of • High speed diesel (direct and electric drive) : SFCM.C = 200 if P > 1810 [Kw] (C.39) if P s 1 « m TI.M71 905 2 SFCM,c = 250 - if P < 1810 [kW] Medium speed diesels Appendix C. SWATH design algorithms 220 SFCMC = 178 if P > 11190 [Kw] L J (C.40) SFCM,c = 230 - if P < 11190 [kW] In equations C.39 and C.40, P is the total required mover power in [kW] at cruise or design speed. And, for SFCG '• SFCG = 210 (C.41) Stores Weight (WSto) : Wsto = WProv + Was + WFW [tons] (C.42) Provisions weight (Wprov) : WProv = 0.005 {NC) + ^ ] [tons] (C.43) Where NC is number of crew and 0.005 is the required provisions per person per day in tons. General stores weight (WGS) '• WGS = 0.0011 (NC) + ^ ] [tons] (C.44) VM VC Fresh water (WFVT) : WFW = 0.125 {NC) + 7 ^ ] [tons] (C.45) VM VC Equation C.45 is valid if (WFW < 50 [tons]). In the case of (Wpw > 50 [tons]), a distillation plant will be necessary. Crew and effect (WCREW) '• WCREW = 0.143 (NC) [tons] (C.46) That is 7 crew member with effect represent 1 ton. Appendix C. SWATH design algorithms 221 Design margin (W m a r f l t n ) : Wmargin = 0.0755 WUg%5 [tons] (C.47) Where WugHt = Ws + WM + WAUX + Wo [tons] C . 4 Resistance and Powering Installed Power ( P M ) : In the preliminary design stage the installed power is estimated [47] by PM = {SM + 1) [kW] (C.48) n Where RT : total resistance [kN] VM : design speed [m/s] SM : percentage service margin •q : overall propulsion system efficiency For preliminary calculations SM = 0.15 and n = 0.70 are assumed. Total Resistance (RT) : Total resistance is given as the summation of frictional (RF), residuary (RR), appendage (RAP), and aerodynamic drag (RAA), all in [kN]. RT = RF + RR + RAP + RAA [kN] (C.49) Appendix C. SWATH design algorithms 222 Frictional Resistance (RF) : The following formula is given for frictional resistance [9] : R F - = \ P (CF + CA) S VM [kN] (C.50) Where density of water (for sea water, it is approximately 1025 [j^]) frictional resistance coefficient correlation allowance wetted surface area [m2] P CF CA S CF is calculated using ITTC formula of 0.075 C f ~ ( l o g l o R n - 2 ) 2 ( C - 5 1 ) And Reynolds number is VM LOA Kn = v with kinematic viscosity u = 1.19a;10_6 \ ~ \ for 15°C sea water. Correlation allowance coefficient CA is approximately 0.0005 [47, 9]. Wetted surface area is given by Equation C.52, and if lowerhull slenderness ratio is approximated to |j w 14, Equation C.53 can then be used instead [47, 9]. S = V» [13.6 - 0.31(20 - ^)] [m2] (C.52) S = 11.74 Vf [m2] (C.53) Residuary Resistance (RR) : RR = \p (GFF + Cw) S VM [kN] (C.54) Appendix C. SWATH design algorithms 223 Where p : density of water (for sea water, it is approximately 1025 [^]) CFF ' form resistance coefficient and approximated to 0.0005 [47, 9] Cw '• wave making drag coefficient S : wetted surface area [m2] However, a new term CR, which is the summation of CFF and Cw, is introduced into Equation C.54 and calculated by the regression formulae (See Table C.l) based on the analysis of the available SWATH data [47, 9]. Table C.l : Residual drag coefficient as a function of volumetric Froude number. C R = Cpp + Cw 0 < Fn < 0.688 0.003 rp 0.688 • r " 0.688 1.808 0.002 Appendage Resistance (RAP) : RAP = 0.15 {RF + RR) [kN] (C.55) Appendix C. SWATH design algorithms 224 Aerodynamic Resistance (RAA) : RAA = \ PairCDAFVM [kN] (C.56) Where Air drag coefficient : CD — 0.7 Frontal area : AF — QMLOA [m2] Density of air : pair = 1.293 [^ ] C.5 Cost Estimation For comparison purposes, it has been suggested that the unit cost per unit displacement is the same in both monohull and SWATH vessels [9]. Appendix C. SWATH design algorithms 225 Main deck Super sturcture II. Super sturcture I Wet deck . ±iauQcn_regjarL. Strut Lower hull H OA Super sturcture II Super sturcture I Main deck BOX (cross structure) Wet deck THaunch i l Strut Lower hull K A Strut _SZ_ Lower hull A Figure C.l: Schematic of a SWATH vessel Appendix D Prediction of a vessel's heave and pitch motions Howard [23] analyzed data obtained from the model testing of UBC series models at the Ocean Engineering Centre at BCR Corp. in British Columbia. After a statistical analysis, the following equation is given for a UBC series type of vessel's response spectrum. 1 1 + e ( a - * « . ) 8 1 + e^-i"') ^D'1^ Where n is the non-dimensional heave or pitch amplitude, za/£a and 9a/{k£a) respec-tively. The coefficients in the above equation are as follows: For heave : A = 0 . 8 5 9 1 ^ ^ + 1.5804^ (D.2) a = 1.816 + 31 .42^ (D.3) S = 16.48-^- 0.7485-^- (D.4) Lpp t>0B 8 = 6 . 4 3 6 ^ ^ + 1 6 3 . 1 ^ ^ (D.5) Bz B 7 = 2 . 8 7 0 ^ ^ + 4 0 . 3 6 ^ (D.6) For pitch A = i . i 3 i 3 ^ H ? ! + o.9264^| (D.7) C T a = 0.7904 + 45.47-^- (D.8) Lpp 226 Appendix D. Prediction of a vessel's heave and pitch motions 227 5 = 16.82-^- - 0.8621-^ — (D.9) Lpp Cblj 8 = 1 8 . 3 4 ^ ^ + 3 7 . 3 5 ^ | ^ (D.10) 7 = 7 . 2 1 1 ^ ^ + 3 . 0 3 9 ^ (D.ll) if D In the above equations the following notation is used: Lpp Length between perpendiculars B Beam Cb Block coefficient Fn Froude number -- V/ J*L^g g Gravity k Wave number (2ir/(Wave Jength)) T Mean draft V Velocity za Heave amplitude 6a Pitch amplitude £ a Wave amplitude ue Nondimensional encounter frequency scaling factor for u>e is j'Lpp/g Appendix E Frequencies of Peak Heave and Pitch Responses Based on the formulation given by Howard (see also Appendix D), a number of heave and pitch response spectrums have been generated by systematically varying the parameters used in Equation D.l (Appendix D). Using the data obtained in this analysis, the following empirical formulae have been developed for the Echidna knowledge base. More complex estimation models (such as ones involving exponential etc.) have been avoided for the sake of simplicity and ease of use in the Echidna knowledge base. ^ 2 (38.8668 - 11.5286C6 - 5.2002^ - 1.1912§) 2 (0.8652 - 0.2816F„)2(49.6354 - 20.4321Cb - 6.8076^ + 0.644lf )2 (E.l) 2 (2.4541 - 0.7280Cb - 0.3283^ - 0.0752f f W n $ p ~ (0.9481 - 0.3086Fn)2(2.8599 - 1.1773C* - 0.3922^ + 0.037lf )2 Where u>n2p and ujnep are non-dimensional circular frequencies that maximize ship re-sponse spectrum for heave and pitch motions respectively. a;- or we u n z p or ujn6p - p (E.3) In Equation E.3, u>Zp or u>ep are dimensional [radians/second] circular frequencies that maximize ship response spectrum for heave and pitch motions respectively, g is the gravitational acceleration, and T is the mean draft reading. 228 Appendix F R M S Heave and Pitch Responses In order to be able to estimate some of the rms motion characteristics of UBC series models, SHIPMO - a ship motion estimation program, was used. Table F . l : Sea state information. [Sabuncu, 1983] Sea State H [m] # 1 / 3 [m] T[s] Tmax [s] A[m] 1 0.18 0.30 2.4 3.4 6.1 2 0.55 0.88 3.9 5.4 15.9 3 0.88 1.40 4.6 6.5 21.6 4 1.55 2.45 5.7 8.1 34.0 5 2.40 3.65 6.8 9.7 49.0 Table F.2: Scaling factors for the raw regression data obtained from the software SHIPMO Motion Non-dimensionalizing scaling factor Heave # 1 / 3 Pitch u>2H1/3/g After a regression analysis on the data obtained from the runs of SHIPMO for 13 UBC series models the following empirical formulae have been developed. The same ranges as the ones used in the Echidna knowledge base are valid for the parameters used in Equations F . l and F.2. 229 Appendix F. RMS Heave and Pitch Responses 230 Fzl = 0.17244257 - 0.10902408-?- - 0.062345772(-B-)2 Lbp Lbp T T Fz2 = 0.92116235 + 9.26182543- 17.01219503(—)2 Lbp Lbp Fz3 = 1.25445302 + 0.015068303C(, Fzi = 0.37459542 + 0.33475563Fn + 0.69543449F,2 - 0.62937284F3 Fz5 = 0.51156287 + 1.67664716-^ - - 0.60058403(-^-)2 + 0.065781060(-^-)3 Lbp Lbp Lbp Fz6 = 1.27331482 - 1.02535654(7V#1/3) + 0.56388486(/v"#1/3)2 TP TP TP TP TP TP TP r z — rz\r z2r z3r zif z5rzQ = Fz + 0.74616424FZ2 - 0.46955309F3 (F.l) Hi/ Br Hx/^maxlg y l / 3 y l / 3 F0i = 0.465143 -2.343704-—+3.482758(-—)2 Lbp Lbp F02 = 1.323205 + 13.947877-^ - - 26.389156(-^-)2 Lbp Lbp Fez = 0.346817 - 24.620275-^ - + 37.557687(-^)2 Lbp Lbp F6A = 0.044729 + 1.003151C(, F65 = -0.182212 - 0.296472Fn + 0.059477Fn2 - 0.245965F, Fee = -0.181446 + 4.924184-^ - - 0.834965(-^-)2 + 0.074877(-^-)3 Lbp Lbp Lbp Fe7 = 2.16213525 - 2.78020823(^/3)+ 1.54756589(/Vif1/3)2 rp TP TP TP TP TP TP TP •• F9 - 0.46697339F,2 + 0.56576473F/ (F.2) 3