Multi-objective optimization of a bidirectional-ribbed microchannel based on CFD and NSGA-II genetic algorithm

Abstract

The bidirectional rib (BR) has been recognized as one effective heat transfer enhancement technique for the microchannel heat sink (MCHS) since it can realize the heat transfer improvement in both vertical and spanwise directions. However, the utilizing of BR causes a large pressure drop penalty in the microchannel due to its broader obstruction area. For the engineering application, the heat transfer and pressure drop penalty of MCHS are two key metrics to evaluate its overall performance. To address this optimization problem, 144 sample points of the microchannels with different BRs are generated by CFD simulation, and the NSGA-II has been utilized to search for the optimal solutions in-terms of both objectives. For most optimal solutions, the length of SR and rib pitch are 200 μm and 800 μm, respectively, while the other two parameters, the width and height of SR, almost cover the entire value range. Subsequently, the best compromise solutions have been determined from the Pareto fronts by the TOPSIS and LINMAP methods according to the decision maker's preference. Finally, the selected optimum solutions are verified by numerical and experimental methods.