Constructal design of a T-shaped porous fin adopting minimum composite function, ANN and NSGA-II

Based on constructal theory, a T-shaped porous-fin model is considered in this paper. Constructal design for the T-shaped porous-fin is conducted with minimum composite function, which is composed of the maximum temperature-difference and pumping-power consumption. The influences of air inlet temperature, velocity, porosity, fin fraction and heat flux on the optimum constructs are analyzed, respectively. The results show that when taking fin length and height-width ratio as variables, composite function can reach double minimum at the value of 0.878, and the corresponding optimal fin length and height-width ratio are 5.45 mm and 0.2, respectively. Comparing to original design, composite function and maximum temperature-difference after twice optimization are decreased by 12.20 % and 42.69 %, respectively. Thus, comprehensive performance of porousfin is improved by simultaneously optimizing the fin length and height-width ratio. Artificial neural network is adopted to predict fin performance, and NSGA-II method is adopted to conduct multi-objective optimization. The optimal porous fin structure is gained by LINMAP decision-making approach. The results gained herein can provide some theoretical guidelines for the porous-fin heat dissipation designs of electronic devices.