Estimation bending deflection in an Ionic Polymer Metal Composite (IPMC) material using an artificial neural network model

The ipmc bending characteristic depends on the accuracy of the manufacturing process for an ipmc specimen and the working conditions such as humidity,temperature,and applied electrical field. so,ipmc behavior is significantly nonlinear and uncertain. in the present paper,we propose an accurate nonlinear neural network black-box model (nbbm) to predict the bending motion of ipmc taking into consideration the applied electrical voltage characteristics and the working conditions (specimen dimensions,temperature and humidity of working environment. an experimental setup and testing program is used to test several ipmc specimens and measure the bending motion at different working conditions and applying electrical voltage signals. the nbbm for the ipmc is designed with suitable input and output parameters to estimate the ipmc specimen tip deflection. the optimal brain surgeon (obs) pruning algorithm is used to capture the optimal network size and to solve the overfitting problem among the training patterns. modeling results show that the optimized nbbm model can best describe the bending behavior of the ipmc specimen according to the applied electrical power signal and the working environment without using any measuring sensor and the proposed model can be used for modeling and controlling the ipmc bending motion in a single segment form. � 2016 jordan journal of mechanical and industrial engineering. all rights reserved.