Neural Topology Optimization Via Active Learning for Efficient Channel Design in Turbulent Mass Transfer

The design of fluid channel structures of reactors or separators of chemical processes is a key to enhancing the mass transfer processes inside the devices. However, systematically designing channel topological structures remains challenging for complex turbulent flows. Here, this challenge is addressed by developing a machine learning framework to efficiently perform topology optimization of channel structures for turbulent mass transfer. A topological structure is represented using a neural network (referred to as "neural topology"), which is optimized by employing pre-trained neural operators combined with a fine-tuning strategy with active data augmentation. The optimization is performed targeting two objectives: maximization of mass transfer efficiency and minimization of energy consumption, for the possible considerations of compromise between the two in real-world designs. The developed neural operator with active learning is data efficient in network training and demonstrates superior computational efficiency compared with traditional methods in obtaining optimal structures across a large design space. The optimization results are experimentally validated, proving that the optimized channel improves concentration uniformity by 37% compared to the original channel. The variation of the optimal structures with changes in inlet velocity conditions is also demonstrated, offering insights for designing turbulent mass transfer devices under various operating conditions.