Genetic Algorithm as a Design Tool for Freeform MEMS Device

This study introduces an innovative design approach that employs a genetic algorithm (GA) in conjunction with freeform geometries for microelectromechanical systems (MEMS) at both the device and system levels. At the device level, an optimization process is performed solely within the mechanical domain. An accelerometer incorporating a freeform mechanical motion amplifier has been developed to showcase the efficacy of the proposed design methodology. Experimental findings indicate a 100% enhancement in the figure of merit (calculated as the product of sensitivity and bandwidth frequency) and a 141% increase in sensitivity when compared to a conventionally designed device. Subsequently, a system level electro-mechanical co-optimization is carried out in the electrical and mechanical domain. A closed-loop accelerometer, also incorporating a freeform mechanical motion amplifier, has been developed to illustrate the performance of the design method. The optimization procedure results in a 195% enhancement of the primary figure of merit (FOM). Measurements indicate that, in the closed-loop system, the accelerometer's displacement is reduced by 86% when a feedback voltage of 4.85 V is applied for 1 g acceleration at 100 Hz, in contrast to the open-loop system.