Battery State of Charge estimation using Extended Kalman Filter

Battery State of Charge (SOC) evaluation in electrical vehicles (EVs) environment is a key point for widely acceptance of this new technology. As SOC cannot be measured it is imperative to use some method that can accurately make its estimation using measurable variables, such as battery terminal voltage and current. As Kalman Filters are commonly accepted as the best estimator for linear systems, and a good one for non-linear ones, it is expected that they can be successfully applied to battery SOC estimation. So, in this paper, it is presented an Extended Kalman Filter (EKF) based technique for SOC estimation. As EKF needs a process model it is used a new one-dimensional, physics based, battery model that solves Fick's diffusion equations subjected to its boundary conditions. Used Batteries are Lithium-polymer ones due to its promising application as EVs energy source. A set of experiments, with different current profiles, were made and its results confronted with simulated ones enabling to conclude about the quality of proposed battery model and EKF SOC estimation.