Research on SOC Estimation Based on Firefly Algorithm Optimization Particle Filter Algorithm

Various characteristic states inside battery and data for function management system can be reflected by state of charge (SOC). Accurate SOC estimation plays a decisive role in safe and stable operation of battery management systems and electric vehicles, but the measurement error, battery nonlinearity, electrochemical reaction complexity, performance aging, temperature change and other factors affect the SOC estimation accuracy. For complex nonlinear system of battery, particle filter (PF) is used to predict SOC because of its advantages in nonlinear system, but particle filter algorithm also has problems of particle degradation and depletion. In this paper, the firefly algorithm (FA) is used to optimize the particle filter, and the position of the individual reflectivity and attraction is updated continuously to achieve particle optimization. To simulate the online state, a battery is selected to simulate the UDDS working condition, and the maximum estimated error of SOC is 0.43%.