A method for SOC estimation for lead-acid battery based on Multi-Model Adaptive Extended Kalman Filtering estimation

The State-of-charge (SOC) is a critical parameter for the battery, which is a hot research in areas such as new energy vehicles, but the research on lead-acid batteries as backup power in Uninterrupted Power System in data center has rarely been done. In this paper, for the first time, a model named Poly-Nernst, which is a combination of polynomial model and Nernst battery model, is proposed to estimate the SOC of the battery working in data center. Extended Kalman Filtering method is adopted for SOC online simulation, whose estimating results are compared with traditional Ah method. The simulation on the real UPS shows that the EKF estimating error is much smaller than that of Ah method with no accumulated integral error. To improve the adaptive capability of established the Poly-Nernst model, multi-model adaptive estimation based on EKF algorithm is designed and the experimental results show the MMAE model estimation is much better than any one of a single EKF estimation. To balance the model complexity and estimating accuracy, the experimental results show that a 3-model EKF estimation model is suggested to be appropriate. This SOC estimating method provides technical instructions for the maintenance of the battery system in data center.