A Battery Management System With a Lebesgue-Sampling-Based Extended Kalman Filter

The estimation and prediction of state-of-health (SOH) and state-of-charge (SOC) of Lithium-ion batteries are two main functions of the battery management system (BMS). In order to reduce the computation cost and enable deployment of the BMS on the low-cost hardware, a Lebesgue-sampling-based extended Kalman filter (LS-EKF) is developed to estimate the SOH and SOC. An LS-EKF is able to eliminate unnecessary computations, especially when the states change slowly. In this paper, the SOH is first estimated and the remaining useful life is predicted by the LS-EKF. Then, the estimated SOH is used as the initial battery capacity for SOC estimation and prediction. The SOH and SOC estimation and prediction are calculated repeatedly in the whole battery service life. The proposed method is verified with the application to the capacity degradation of the Lithium-ion battery. The results show that the LS-EKF-based algorithm has a good performance in SOH and SOC estimation and prediction in terms of accuracy and computation cost.

I. Introduction

Batteries are safety critical components that provide power to system functions including command, control, communications, and intelligence  [1]. It is important to improve the performance of the battery management system (BMS) to make battery operation safe, reliable, and cost efficient [2], [3]. A BMS is designed to evaluate the instantaneous state-of-charge (SOC) in the battery, and at the same time, to monitor the slowly varying state-of-health (SOH).