The lithium battery state of charge estimation based on the fractional order unscented kalman filter algorithm

In the assessment of the state of charge (SOC) of lithium batteries, it has been observed that the conventional integer-order second-order resistance-capacitance (RC) network falls short in accurately replicating the non-linear attributes and the precision of SOC estimation influenced by the internal polarization reaction within the battery. Thus, we present the incorporation of constant-phase element (CPE) and Warburg components to achieve a model with superior recognition accuracy compared to the conventional second-order equivalent circuit model(ECM). The parameterization is carried out using the particle swarm optimization algorithm, resulting in enhanced robustness of parameter identification. Later on, when we mixed the regular Unscented Kalman Filter (UKF) algorithm with the fractional-order model (FOM), we obtained an algorithm to estimate the SOC of lithium battery in real time, named FOUKF, and compared FOUKF with the integer-order Extended Kalman Filter (EKF) and UKF algorithms. This confirms that the model is reliable and the algorithm’s estimation is accurate.