Repository logo

Defining and developing an efficient design tool for ship and yacht design

No Thumbnail Available



Change log


Tan, Kok Thong 


Both the yacht design and shipbuilding industries are going through an era of fierce competition. Among others, the quality of designs would make a significant contribution to their success. Under this compelling situation, an efficient design tool, which assists designers to achieve high quality designs, is highly desired. Acknowledging that there are many factors, which affect the survivability of these industries, this thesis sets out to develop an efficient design tool for ship design, with yacht design as an example. Prior to the development of this tool, there is a fundamental question which needs to be answered first, i.e. what are the requirements for an efficient design tool? In order to identify these requirements, the thesis studies what ship designers do in the design process, and how they work out their solutions. The thesis reveals that the design of ships, by its very nature, might deter the exploration of high quality designs. This primarily arises from the sequential nature of the Ship Design Spiral. Ironically, the spiral model, which was first visualised by Evans in 1959, has become the iconic representation of ship design. Its major characteristic is that the design process is essentially sequential and iterative. While the design model is repeatedly altered by means of consecutive iterations, the spiral model simply provides guidance on the sequential order in which the design process should be carried out. Due to its simplicity, the model is widely accepted as a classic model for ship design. This popular, simple model, however, does not spell out how design problems are clarified and defined, and how design solutions are generated. By studying the design methodologies of software development and architectural design, the thesis hypothesises that ship designers employ an iterative, solution-oriented approach because they are dealing with ill-defined problems. These hypotheses are substantiated by a case study of the design of a 14m sport-catamaran. This case study also shows that the designers, without having an overview of the design, adversely altered other design characteristics while concentrating on a particular aspect. The above findings extend the current knowledge of how ship designers work. The review of present design tools shows that many of them do bring about an improvement in terms of productivity by cutting down the overall design and production cost and time. The thesis asserts that new opportunities have opened for intellectual exploration. However, the present tools do not specifically solve the problems of the Ship Design Spiral, so the new opportunities cannot be fully utilised in an efficient manner. In view of the problems of ship design and the current state of design tools, the thesis articulates five requirements for an efficient design tool. The tool must be highly interactive, must provide an overview of the design by relating all the relevant issues in a fully integrated, and possibly concurrent, manner. This will enable designers to recognise the overall effects of modifications quickly, and thus to avoid adversely altering other design characteristics. In order to do that, the tool must be able to break the sequential nature of the spiral model, and to deduce assumptions from past experience and known theories. Moreover, the tool must be able to track the design changes. This allows designers to visualise potential design trends, which help them to conjecture further improvements. This thesis demonstrates that a yacht design process could be compressed by up to 40% without having to affect its overall behaviour, and there is actually very little room for changing the order of the steps. In order to break the sequential nature of the spiral model, and achieve a higher degree of parallelism, preliminary assumptions have to be made for the inherently sequential components. The thesis, therefore, proposes a novel concept of concurrent integration for a sequential design process. It integrates all the relevant programs concurrently, and generates assumptions, which are based on past experience and known theories, for partial computation in advance. This facilitates the parallel execution of design analyses regardless of the state of the design model. As parts of the computation involve generating assumptions, the outcome of the analyses is constantly compared with known domains in order to allow designers to estimate their reliability. This concept introduces new learning in process decomposition, integration and compression. This concept was developed into a Concurrent Integrated (CICAD) System for a particular type of sailing yacht design. With this system, a designer can interact with his design model freely, and gets a prompt feedback for any modification he makes, even for the subtle ones. The thesis shows that in a matter of seconds, instead of many hours or even days, the designers will be able to obtain the analyses output of hydrostatics, static stability, velocity prediction, maximum shearing force and bending moment. Moreover, he will be provided with certainty factors, which indicate the degree of confidence in the outcome of the analyses, and a design history, which helps him to visualise his design trends for further improvements. This thesis demonstrates that it is possible to break the sequential nature of the Ship Design Spiral, and at the same time allows designers to explore many more alternatives in a very efficient, highly interactive manner. CICAD is verified in terms of correctness and consistency by using Standfast 43 and Pen Duick V as test cases. For validation, nineteen expert designers participated in a series of 4-hours workshops to test-drive and review CICAD. They confirmed that CICAD was a highly interactive and responsive tool, which enabled them to improve their ability to recognise the effects of modifications very rapidly. With the overview provided by CICAD, the designers agreed that this would help them to minimise adverse modifications. They found that the way CICAD tracked the design changes was useful, and they strongly believed that it would help them to conjecture further improvements. The degree of realism of the preliminary assumptions was acceptable. There is a need to incorporate the past experience of each individual designer. In conclusion, they overwhelmingly agreed that CICAD, as an efficient tool, would enable them to conduct more iterations in a shorter period. Although there are many factors that constitute high quality designs, the designers suggested that CICAD would play a very significant role in helping them to achieve that goal.


This thesis was digitised by the British Library from microfilm. You can acquire a single copy of this thesis for research purposes by clicking on the padlock icon on the thesis file. Please be aware that the text in the supplied thesis pdf file may not be as clear as text in a thesis that was born digital or digitised directly from paper due to the conversion in format. However, all of the theses in Apollo that were digitised from microfilm are readable and have been processed by optical character recognition (OCR) technology which means the reader can search and find text within the document. If you are the author of this thesis and would like to make your work openly available, please contact us:





Doctor of Philosophy (PhD)

Awarding Institution

University of Cambridge