Non-linear kalman filters for battery state of charge estimation and control

In this paper, two different non-linear Kalman Filters for lithium-ion battery state of charge estimation are presented and compared. Nowadays, lithium-ion batteries are extensively used for hybrid and electric vehicles; in such applications, cells are assembled in module and pack to achieve high performance. At this scope, a Battery Management Systems BMS is required to control each cell and improve the battery pack performance, safety, reliability, and lifecycle. One of the major tasks a BMS must fulfill is an accurate online estimation of the State Of Charge (SOC) of the battery pack. In this paper, the Extended Kalman Filter and Sigma Points Kalman filter are developed and compared. A battery equivalent circuit model has been chosen to have a good compromise between complexity and accuracy and model parameters have been identified from Hybrid Pulse Power Characterization (HPPC) tests carried out at different temperatures and current rates to obtain a model valid for a wide range of operating conditions. The SOC estimation strategies are developed starting from the experimental results and it is validated through different driving cycling simulations. The results show that the Sigma Points Kalman filter produces a better estimate of SOC with respect to the Extended Kalman Filter, due to its better capability to deal with system non-linearities, with comparable computational complexity.