Hybrid unscented particle filter based state-of-charge determination for lead-acid batteries

Accurate prediction of cell SOC (state of charge) is important for the safety and functional capabilities of the battery energy storage application system. This paper presents a hybrid UPF (unscented particle filter) based SOC determination combined model for batteries. To simulate the entire dynamic electrical characteristics of batteries, a novel combined state space model, which takes current as a control input and let SOC and two constructed parameters as state variables, is advanced to represent cell behavior. Besides that, an improved UPF method is used to evaluate cell SOC. Taking lead-acid batteries for example, we apply the established model for test. Results show that the evolved combined state space cell model simulates battery dynamics robustly with high accuracy and the prediction value based on the improved UPF method converges to the real SOC very quickly within the error of±2%. •This paper introduces a hybrid UPF based SOC determination model for batteries.•The evolved model takes SOC and two constructed parameters as state variables.•The combined state space cell model simulates battery dynamics robustly.•NLMS based method is employed to lessen search space and fasten convergence process.•Novel model converges to the real SOC robustly and quickly with fewer particles.