An open-source framework for the multi-objective optimization of passive micromixer based on RSM and MOEA/D

Abstract

Balancing the mixing index, pressure drop and mixing energy consumption is crucial for enhancing the performance of passive micromixers. In this paper, a T-shaped micromixer scheme integrating cylindrical obstacles and contraction-expansion parts is proposed and the geometric parameters are studied in order to achieve multi-objective optimization. The geometry of the micromixer is optimized by selecting three design variables (L₁, L₂ and D_cy), where L₁ and L₂ are used to adjust the changes in the cross section of the flow channel and D_cy controls the diameter of the cylindrical obstacle. For different Reynolds numbers (0.1 and 100), COMSOL simulation and Box-Behnken experiment design are used to evaluate the effects of each design variable on the mixing evaluation indicators. The findings indicate that L₂ exerts the most substantial influence on the performance of the micromixer. It can effectively modify the width of the flow channel, thereby optimizing the fluid velocity and concentration gradient, and consequently having a significant impact on mixing evaluation indicators. Regarding multi-objective optimization, the decomposition based multi-objective evolutionary algorithm (MOEA/D) is used to construct the Pareto optimal solution. This algorithm offers a balanced scheme between different mixing evaluation indicators. The experimental results demonstrate that the optimized micromixer can not only enhances the mixing index but also remarkably reduces the pressure drop. It can achieve a complete mixing effect while reducing the pressure drop by nearly 41% compared to the passive micromixer using two-layer serpentine crossing channels when Re = 100, which is especially suitable for microfluidic applications such as biochemical detection. Furthermore, this study makes the MATLAB program source code used in the calculation process publicly available, which provides a strong support for the subsequent research and application.