Zinc-Air Battery Dynamics’ Identification Using Transfer Functions and Hammerstein-Wiener Models

This work aims to identify the dynamic behavior of a prototype tubular zinc-air battery by using a behavioral modeling approach. The class of models used to predict the behavior of the battery are as follows: linear transfer function and nonlinear Hammerstein-Wiener (HW) model. Model parameter estimation was implemented in MATLAB and the response data test was categorized into three sections including single-step, multi-step and random-step responses. With respect to single-step response, both the transfer function and HW model were able to predict the response curve. As expected, the linear model exhibited lower accuracy than the nonlinear model. For multi-step and random-step responses with various step-size, the linear model cannot cope with the experimental behavior and is not able to provide accurate prediction of battery dynamics. One of the first conclusions of the study was that the nonlinear characteristic of the battery cannot be ignored, and the linear model might be inappropriate to predict the dynamic behavior of the battery. The second result was that the nonlinear model with a classic structure was able to predict the response with various step size precisely. Results implied that the HW model proved to be a versatile choice for use within a zinc-air battery system.