Particularised Kalman filter for the state-of-charge estimation of second-life lithium-ion batteries and experimental validation

Abstract

A critical issue for a proper energy management of a lithium-ion (Li-ion) battery is the estimation of its state-of-charge (SOC). There are various methods available for the SOC estimation, being some of them robust and accurate, but requiring high computational power for its applicability, which is inconvenient for their use with the usual low-cost microcontrollers that build a typical BMS. This contribution proposes an SOC estimation algorithm based on a simplified Kalman Filter, that combines a high accuracy with reduced computational requirements. The proposed simplifications result from a careful analysis of the Li-ion battery performance and linearization of processes that entail negligible loss of accuracy. The proposed algorithm is used to estimate the SOC of a second-life Li-ion battery operating in an experimental PV self-consumption facility. Its performance, in terms of accuracy, robustness and computational requirement, is compared with an Extended Kalman Filter (EKF), a Particle Filter (PF) and other low-performance estimation algorithms, proving its tradeoff between accuracy and computational cost.