Optimal location of three heat sources on the wall of a square cavity using genetic algorithms integrated with artificial neural networks

In this work, optimization of location of heat sources in a square enclosure with natural convection is performed to maximize the global conductance in the enclosure. For this study we have taken a square enclosure with three adiabatic walls, one isothermal wall opposing the wall having three heat sources. Numerical simulations are done by changing positions of heat sources for different Rayleigh numbers using Fluent 6.3(2d, double precision). And for some configurations maximum temperature inside the enclosure is noted. Optimization is done using genetic algorithms (GA) combined with artificial neural networks (ANN). An ANN is trained using the above data obtained from numerical solutions. The trained ANN will be the simulation tool, whenever required by the GA for optimization. It is shown that at high Rayleigh number the spacing between the heat sources should be zero for optimum heat transfer. Variation in optimum solution for unequal heat fluxes are also studied.