Hybrid Genetic Algorithm - Neural Network - Genetic Algorithm based Multi-Objective Optimization Technique for Design of Heat Fins

摘要

Heat transfer optimization is very important for commercial design of heat fins. Conventionally, heat transfer optimization studies for design of heat fins are carried out by direct numerical analysis using CFD code. But a computational bottleneck is seen in running the numerical simulations. This is because the number of solution abruptly increases with increase in the number of parameters. To minimize the number of simulations, a learning procedure using Artificial Neural Network (ANN) is used. To reduce the effort in determining suitable parameters for ANN, Genetic algorithm is used. ANN is trained based on a minimal set of numerical/CFD results. The trained ANN can then predict the base plate temperature for any other untrained set of parameters in the same range. Genetic algorithm is again used with ANN as the fitness function to obtain the multi-objective optimal. The effectiveness of this GA-ANN-GA hybrid for heat fin optimization is thus presented.