Accuracy improvement of remaining capacity estimation for energy storage batteries

Scheduling lithium‐ion batteries for energy storage applications in power systems requires accurate estimation of their remaining capacity. Due to the varying discharge rate during a cycle caused by complex operating conditions, conventional estimation methods suffer greatly from poor estimation accuracy. Therefore, this study proposes an improved Ah‐counting method for remaining capacity estimation, in which a neural network‐based model is introduced to consider the rate capability of batteries. Moreover, to avoid the inherent error caused by Ah‐counting, a straightforward method is applied to the estimation. Simulations show that the estimation accuracy can be effectively improved by the proposed method.