Investigation of Dynamic Behavior of Smart Piezoelectric Actuators Using Artificial Neural Networks

Abstract

The purpose of this study is to investigate microelectromechanical behavior of smart piezoelectric actuators using Artificial Neural Networks due to simple, multi harmonic and dynamic pulse excitations. Regarding to complexity and time-consuming analyses of vibration of smart structures, existing classical models are often insufficient. Nowadays, artificial intelligence tools are used for modeling such complex phenomena. The theoretical model is a three-layer piezoelectric composite beam that behaves as an axial actuating mechanism. This actuator consists of an elastic core sandwiched between two piezoelectric active outer layers. The piezoelectric layers are polarized transversely, i.e., the polarization vector is parallel to the applied electric field intensity vector. For initializing the electromechanical effect, an electric field is applied to the piezoelectric layers. The finite element modeling is constructed using ANSYS. Then, harmonic and dynamic vibration analyses are performed and the responses of smart beam are calculated. The required data used for artificial intelligence were collected from vibration analyses. Obtained results demonstrate that artificial neural network is in good agreement with observed values.