Dynamic analysis and structural optimization study of porous FG-MEE microplate with impact and hygrothermal loads

Abstract

Functionally graded magnetic-electro-elastic (FG-MEE) materials with their remarkable multifunctional energy conversion capabilities are widely used in several components such as biosensors, micro imaging devices, micro robotics and micro actuators. This paper presents vibration analysis and optimization studies of FG-MEE microplate resting on nonlinear visco-elastic foundation under hygrothermal loads. The equations of motion for a two-dimensional (2D) functionally graded MEE porous microplate are first obtained using four variable higher order shear deformation plate theory in conjunction with the modified couple stress model. By considering the material grading and porosity distribution functions, the dynamic formulation of the microplate is derived using variational approach. The general analytical procedure is employed to obtain the stiffness and mass matrices for different boundary conditions. Initially, the model is validated using free vibration analysis results. Further, dynamic analysis of FG-MEE microplate subjected to different impulse excitations is conducted to obtain the response amplitudes. Parametric studies are performed to investigate the effects of grading indices, porosity factor, length-scale parameter, foundation parameters, multi-physical loads on the fundamental frequency and dynamic response. Using design of experiments, it is noticed that five factors only have vital influence on the dynamic response of the microplate. Accordingly, a multilayer perceptron neural network model is trained with parametric data acting as a surrogate model. Finally, the optimal parameters are predicted to improve the stability by solving a constrained optimization problem using firefly optimization scheme with the trained neural network model. From optimization results it is noticed that the fundamental frequency has improved by 63.66% with enormous saving of computational time using the neural network surrogate model.