Optimization of thermoelectric modules with the method of genetic algorithms

In this doctoral dissertation, 6 different programs have been developed by employing genetic algorithm method in order to design low-cost thermoelectric modules which maximize cooling capacity, heat difference and the coefficient of performance. The software of these programs has been developed by Java programming language in NetBeans 7.6.1. environment. The initial population selected in this study is 300 thermoelectric modules. Optimization of the module has been carried out depending on different operations regimes of thermoelectric module defined by the heat of the hot surface, the quantity of the semiconductor modules for constant but different volumes modules, the total volume of the space, the arm area, the arm height and the current passing through the arms. As a result of the process of optimization, the optimal values of TE module the coefficient of performance, maximum cooling capacity, and minimum cost have been obtained. The application of the program was performed taking into the geometric parameters for each of the 36 thermoelectric modules produced by Melcor Company and maximum cooling capacities were obtained. Results obtained have been compared with the cooling capacity projected by Melcor Company. The results of the comparison of the cooling capacity of modules produced by Melcor Company and results obtained by optimization have been found out to be very close.