Prognostics of Lithium-Ion Batteries Under Uncertainty Using Multiple Capacity Degradation Information

Prognostics of lithium-ion batteries play an important role in the intelligent battery management systems. The State of health (SOH) estimation for batteries often needs to be implemented under uncertain situations for the complicated operating conditions. In this work, a novel integrated approach based on a mixture of Gaussian process (MGP) model is presented for lithium-ion battery SOH estimation under uncertain conditions, where the degradation process parameters distribution can be learnt from the multiple available capacity measurements. To capture the time-varying degradation behavior, the distribution information of the degradation model parameter is extracted by fusing the training data from different battery conditions with the MGP model. Moreover, by exploiting the degradation model parameter distribution information, the PF algorithm is employed to predict the battery SOH. The correlation case study and comparison analysis are provided to show the efficiency and effectiveness of the proposed prediction method.