Characterization Of Piezoelectric Actuator With Physics-Informed Neural Networks

This paper introduces for the first time the application of physics-informed neural networks (PINNs) in characterizing microelectromechanical systems (MEMS), particularly for piezoelectric micro-actuator. The capability of the constructed PINNs model is demonstrated through inversely determining piezoelectric coefficient from measured deformation data of an electrically actuated beam. This work serves as a starting point of the merging between MEMS design and machine learning techniques.