Online state and parameter estimation of the Li-ion battery in a Bayesian framework

Due to an ever-growing role of lithium-ion batteries in industry, particularly automotive industry, an effective battery management system (BMS) is of critical importance. A reliable battery state estimation scheme is an integral part of such a BMS. Complicated nature of battery dynamics, weak observability, lack of knowledge about the degradation mechanisms of these batteries, etc has made their state estimation a challenging task. Among the published works on Li-ion battery estimation, a subject that has not received a great deal of attention is parameter estimation of the battery. Parameter estimation has a direct impact on both state of charge and state of health estimation of the battery. Most of the works in the field of battery estimation are built upon the known parameters of the battery whereas in reality these parameters change over-time and may not be known a priori, particularly for aged batteries. This work tackles the problem of parameter and state estimation of lithium-ion batteries from a model-based perspective using a multi-rate particle filter. This filter is applicable to the full-electrochemical of the battery without any restrictive assumption or simplification of the model equations. The filter is proposed in a multi-rate structure in order to address the run-time of the process and computational load of the algorithm. The simulation studies demonstrate the effectiveness of the proposed algorithm.