Parameter estimation of lithium-ion batteries and noise reduction using an H∞ filter

Lithium-ion batteries are widely used in conventional hybrid vehicles and in some electrical devices. A lumped parameter model of lithium-ion battery is constructed and system parameters are identified by using the autoregressive moving average (ARMA) and a genetic algorithm (GA). The precise information of state-of-charge (SOC) and terminal voltage are required to prolong the battery life and to increase the battery performance, reliability, and economics. By assuming a priori knowledge of the process and measurement noise covariance values, Kalman filter or extended Kalman filter has been used to estimate the SOC and terminal voltage. However, the main drawbacks of the Kalman filter is to use correct a priori covariance values, otherwise, the estimation errors can be lager or even divergent. These estimation errors can be relaxed by using the H-a filter, which does not make any assumptions about the noise, and it minimizes the worst case estimation error. In this paper, H-a filter is used to estimate the SOC and terminal voltage. The H-a filter can reduce SOC estimation error, making it more reliable than using a priori process and measurement noise covariance values.