A computer model for preliminary design and economics of container ships
This thesis is concerned with the development of a computer algorithm for determining the principal dimensions of a container ship at the preliminary design stage. The algorithm was devised to aid a Naval Architect to design the most economical ship, given the ship owner's requirements. The emphasis has been on developing an algorithm which acts as an aid in the design process. There are basically four models of the computer aided ship design which can be used in stages. The first model or algorithm is based on a deterministic approach with parametric variation of principal dimensions to locate the optimum design with minimum required freight rate. The second model incorporates optimisation techniques to arrive at the optimum ship. Though the optimisation technique is very powerful in the search of an optimum both in computer time and computing cost, the parametric method is preferred where a designer has little faith in the optimisation process or as an aid to check the answer arrived at in the optimisation process. The third model of the computer aided design can be used once the optimum has been found. A new approach to carry out sensitivity analysis is introduced. This approach overcomes the deficiencies of the past approach, in the sense that sensitivity analysis is carried out for achievable variation in variables rather than an arbitrary variation. The third model of computer aided design may be used once the designer has identified the variables, the variation of which, influences the required freight rate most. The use of the third model of the ship design may be adequate in identifying the total risk of the project. Together with sensitivity analysis, the designer can evaluate the total risk involved in an investment since the third model also incorporates a simple approach to risk analysis. However three estimates are required in the third model compared to single estimates of variables in the first and the second model. The fourth model incorporates the risk analysis by Monte Carlo method of simulation. In this model the designer can assess the xvii total risk of the project by generating the risk profile of the Required Freight Rate. The designer must either subjectively or objectively input the probability distribution of each of the influencing variables before using the fourth model. The four computer aided design models form a complete suite of computer programs, which can either be used in a deterministic mode, (first and second model), or in a probabilistic mode, (third and fourth model). Compared to previous ship design algorithms developed solely to deal with deterministic phase, this thesis incorporates ideas on how to incorporate uncertainty and assess risk in capital investment in a shipping venture. The designer can either use these computer models in stages, from deterministic phase to probabilistic phase or the models can be used on their own.