Maximizing Parameter Identifiability of a Combined Thermal and Electrochemical Battery Model Via Periodic Current Input Optimization

This paper develops a framework for estimating the parameters of a combined thermal and electrochemical model of a lithium-ion battery using a single, non-invasive, dynamic experiment. The paper is motivated by the low parameter identifiability suffered by lithium-ion battery models. The existing literature contains a substantial number of studies on battery parameter estimation. However, the existing research focuses either on thermal or electrochemical model parameter estimation. The main goal of this paper, in contrast, is to examine the feasibility of a single, current cycling experiment to identify the parameters of the two models. This study uses equivalent circuit models to simulate both the thermal and electrochemical dynamics of a lithium-ion battery. However, the underlying theoretical framework is extendible to higher order battery representations. The proposed input current trajectory in this study is optimized to maximize the Fisher identifiability of the combined set of parameters (i.e., thermal and electrochemical). We validate the optimal current input trajectory using a Monte Carlo simulation study. The results from the simulation study demonstrate that the optimized current trajectory allows the simultaneous estimation of the combined set of parameters using a single, current cycling, experiment over a 12-hour period of time. The results in this paper are encouraging because they demonstrate that a current input trajectory can be designed to optimize the identifiability of all the parameters of a thermal and electrochemical battery model.