Lithium-Ion Battery Pack On-Line Degradation State Prognosis Based on the Empirical Model

Lithium-ion battery cells are connected in series and parallel to meet the demand for voltage and capacity of the battery energy storage system (BESS). However, with the battery cell degrading gradually, the performance of the battery pack also degrades. Therefore, it has to be focused on battery pack degradation state prognosis to ensure the system working safely and reliably. On the other hand, accurate degradation state prediction of battery pack is also helpful for conditional based maintenance which decreases the maintenance cost. The terminal voltage of the battery pack is utilized as the character presenting the battery degradation in this paper. The relationship between cycle number and the terminal voltage is described by an empirical double exponential function. Particle filter (PF) algorithm is applied to identify and update the parameters in the model. The battery degradation state can be directly predicted with the predefined cycle number. The battery remaining useful life (RUL) can also be estimated with the given failure threshold of the degradation feature. The experimental results indicate that the proposed prediction framework can get high prediction accuracy. At the same time, for the different operating conditions, this approach is of great robustness.