Inverse design of microreactor flow fields through anisotropic porous media optimization and dehomogenization

Abstract

Microreactors are widely used in energy storage and conversion systems, where fluid flow fields play a significant role in reaction-fluid performance. While conventional and bio-inspired flow fields have been investigated in literature, they are often limited to a forward design framework. Forward-designed flow fields depend on an initial layout selection with limited design freedom. In this paper, an inverse design and dehomogenization framework is proposed to discover innovative microreactor flow field designs. A gradient-based method is applied to optimize the spatially varying orientation of anisotropic porous media. A dehomogenization method is adopted to recover the optimized porous media performance by way of intricate microchannel structures in a post-processing step. Through a numerical example, we find trade-offs between the reaction performance and fluid flow performance for multiple optimized microreactor flow fields. Parallel flow fields are confirmed as low pressure drop designs. Branching flow fields with primary and secondary flow paths are discovered for enhanced reaction performance. Optimized designs with balanced reaction-fluid performance can also be obtained with input design requirements. The three-dimensional (3D) performance of each flow field design is verified by additional numerical simulation. The optimized flow fields are shown to outperform conventional parallel and serpentine benchmark designs. The findings of the paper will be useful in novel reactor flow field designs for enhanced performance in biomedical, pharmaceutical, and energy applications.