Random Microfluidic Chip Design with Diagonal Channels Using K-Means Clustering for Fluid Dilutions

The utilization of microfluidic biochips has seen continuous improvement due to their rapid analytical capabilities and cost-effectiveness. However, optimizing their efficiency requires modifications to the chip design. In literature, we found that a large number of COMSOL simulations are required to create the dataset of random chip designs, which is a time-consuming task. In this paper, we propose a method to enhance the efficiency of randomly designed chips by integrating diagonal channels within the random grid structure. Moreover, we propose a methodology based on
$K$
-means clustering technique to facilitate the attainment of desired concentrations, thereby improving user accessibility and usability. Additionally, the proposed method reduces the reliance on COMSOL simulations and aims to contribute to the ongoing advancement and practical application of microfluidic biochips. By incorporating these diagonal channels, we observe a 45 % increase in the number of unique concentrations and velocities, a 20% reduction in the error between user-required and achieved values, and a 15 % reduction in the total channel length, resulting in increased precision and accuracy of the design for fluid dilutions.