An adaptive hybrid approach for online battery state of charge estimation

With the widespread adoption of electric vehicles (EVs) and energy storage in renewable energy systems, the use of lithium-ion batteries has increased significantly, making the battery safety performance a primary concern. The accurate state of charge (SOC) estimation can help mitigate the safety risks for the utilisation of EVs and renewable energy systems. Due to the dynamic and non-linear properties of batteries, an adaptive online SOC estimation is proposed in this paper by combining the online parameters estimation using equivalent circuit model (ECM) and the improved particle filter (PF) algorithm. It firstly deduces ECM parameters equations using bilinear transformation with the elimination of the variation caused by the ambient temperature. Then, the seeker optimization algorithm (SOA)-based fixed-length weighted least square (LS) algorithm is introduced to online estimate the battery parameters accurately. With the established ECM, the battery SOC can be estimated by the improved genetic algorithm (IGA) resampling-based PF algorithm, which effectively alleviates the particle degeneracy problem during the estimation, consequently, offering a better performance in SOC estimation. Both simulations and experiments have been conducted to validate the effectiveness of the proposed method. Compared with other existing algorithms, it shows that the proposed algorithm can accurately model the battery with the root mean squared error (RMSE) <0.1 % and achieve the real-time SOC estimation with less computation burden and high accuracy.