Multi-Objective Optimization for Rectangular Microchannel Using Response Surface Model Coupled With Particle Swarm Algorithm

With the development of microelectronics and micro-electromechanical systems, the performance requirements for microchannels are becoming increasingly higher and more complex. This study aims to improve the overall performance of rectangular microchannels using the multi-objective particle swarm optimization algorithm (MOPSOA). First, the response surface methodology (RSM) was adopted to fit the thermal resistance function. Three-dimensional contour plots and response surface plots were created to analyze the interaction between fin thickness, channel width, and channel depth, aiming to understand their impact on thermal resistance values. Second, a mathematical model for the MOPSOA was developed with the objective functions being thermal resistance and pressure drop. Next, the Pareto optimal solution set for thermal resistance and pressure drop was determined by conducting simulations, and the K-mean clustering method was employed to identifying the four representative solutions. The results indicate a high level of accuracy in the thermal resistance function fitted by the RSM, with correlation coefficients R² = 0.9981 and adjusted correlation coefficient adj R² = 0.9961 respectively. Finally, the performance of a microchannel heat sink was assessed using the computational fluid dynamics method, and the optimized heating surface has a maximum temperature of 11 °C and a maximum pressure drop of 5.292 kPa lower than the non-optimized one. Additionally, the temperature distribution on the substrate is more uniform. This revealed a superior heat transfer capability and lower pressure drop, resulting in a more comprehensive and efficient performance.