Optimizing Sorting of Micro-Sized Bio-Cells in Symmetric Serpentine Microchannel using Machine Learning

Efficient sorting of target cells is crucial for advancing cellular research in biology and medical diagnostics. Inertial microfluidics, an emerging technology, offers a promising approach for label-free particle sorting with high throughput. This paper presents a comprehensive study employing numerical computational fluid dynamics (CFD) simulations to investigate particle migration and sorting within a symmetric serpentine microchannel. By adopting a Eulerian approach to solve fluid dynamics and a Lagrangian framework to track particles, the research explores the impact of flow Reynolds number and the number of loops in the serpentine channel on sorting efficiency. To generate a robust data-driven model, the authors performed CFD simulations for 200 combinations of randomly generated data points. The study leverages the collected data to develop a data-centric machine learning model capable of accurately predicting flow parameters for specific sorting efficiencies. Remarkably, the developed model achieved a 92% accuracy in predicting the Channel Reynolds Number during testing. However, it is worth noting that the model currently faces challenges in accurately predicting the required number of loops for efficient sorting.