Experimental and Numerical Investigation of the Arms Displacement in a New Electrothermal MEMS Actuator

Microgrippers can be effectively applied for handling, positioning and assembling of the micro components. In the present study, a new design of a U-shape electrothermal microgripper was fabricated and developed with the voltages correspond between ۱ to ۱۰ volts. The microgripper was made of silicone with thickness of ۲۵ microns, and pieces between ۴۶۰ to ۴۸۰ microns. The proposed microgripper has a simpler design and more facile fabrication comparing to most reported electrothermal microgripper. The behavior of the microgripper was simulated in COMSOL software to measure the displacement of the arms which hold and heat generations during the voltage changes. The present microgripper has more thermal and voltage tolerance comparing to other electrothermal microgripper. Furthermore, the obtained amount of tip displacement for voltage changes is acceptable. Another simulation method based on a three layer artificial neural network model (ANN) was carried out. Feed forward back propagation algorithm was employed as training algorithm to predict the displacement. The obtained results from both models proved that ANN model had better estimation due to the mean absolute percentage error of ۱.۰۲۴% and determination coefficient of ۰.۹۹۹۵. Moreover, they confirm higher capability and accuracy of ANN in prediction of arms displacement compared to FEM.

Keywords:

Artificial Neural Network , Electrothermal actuator , Finite Element Method , Micro electro mechanical system