Application of a Distributed Genetic Algorithm to the Improvement of Propeller Design

• Main Authors: Chuan-Hung Hung, 洪銓鴻
• Other Authors: Ching-Yeh Hsin
• Format: Others
• Language: zh-TW
• Published: 2002
• Online Access: http://ndltd.ncl.edu.tw/handle/68083999073019326062

碩士 === 國立海洋大學 === 系統工程暨造船學系 === 90 === In this thesis, a propeller design tool is developed using a Genetic Algorithm. This design tool is developed for the assistance of propeller hydrodynamic designs, and used to design skew, rake and chord-length distributions for reducing cavitation on the blade surface. All these geometric parameters are traditionally designed by experience or existing data. In this method, Bèzier curves are adopted for defining these geometric parameters, and a Genetic Algorithm is used for searching the optimum solutions. A parallel version of this design method is developed on a distributed computing system for computationally efficiency. The computational results show that the skew distribution is critical to the cavitation behavior, and the rake distribution is not. A good chord-length distribution can improve the blade cavitaion, however, it should be cooperated with the structure design. In order to obtain a practical and usable skew distribution, some constrains should be added to the design conditions. We have tried six different constrains, and discussed the design results in this thesis. Different fitness functions are tested, and it is found that “cavitation volume / propeller thrust” is the most appropriate function. Using different number of generations and different number of parents are also investigated in the thesis. Design example are shown in the thesis, and suggestions are made for improving this design method.