| Summary: | 碩士 === 淡江大學 === 機械工程學系 === 88 === This thesis develops a genetic algorithm (GA) based optimization program written in C language. The features includes several improvements, such as roulette wheel selection, multi-point crossover, mutation and expression. This program then combines an imposed-on penalty approach and a weighting method for solving a mix of continuous, integer, and discrete variables in multi-objective functions. Various test examples verify the developed program is efficient and correct. Consequently, the proposed program is then applied to the optimum designs of the micro-channel heat sink (MCHS) with square fins.
According to the theory of heat transfer, structural characteristics of MCHS and previous efforts of other researchers, the engineering analysis of MCHS with square fins in silicon was executed at first. The optimum designs of MCHS contain the constraints of pressure drop and fin numbers. The mathematical design model of minimization of the total thermal resistance (or maximization of the total heat transfer) in laminar flow is then constructed consequently in order to find optimum fin height, fin distance, fin thickness, fin numbers (integer variables) and fluid mass flow rate (continuously real variable). The design results obtained in the study then are compared with the MCHS performance in other research work. The design examples show the accuracy and practicability of the proposed GA based optimization program and the engineering design model of MCHS.
|
|---|
