A flexible method for state-of-health estimation of lithium battery energy storage system

Prognostics of battery health conditions are regarded as significant tools for ensuring safety and stability of battery energy storage systems. Meanwhile, flexible and practical techniques are promising and innovative for achieving accurate and cost-effective state estimation. Herein, a novel curvature-based method is proposed to track battery degradation conditions. The proposed technique can ingeniously capture the health features of an aged battery from incremental capacity curves. Specifically, an advanced filter is employed to smooth the incremental capacity curves. Secondly, a self-definition sine function is applied to fit the smoothed incremental capacity curves. The curvature-based technique provides an essential manner for extracting the health features including the peak value and position from the fitted incremental capacity curves. On this basis, a flexible cost-effective numerical aging model is constructed by mapping the relationship between the feature indicators and battery capacity for battery health prognostic. Finally, for verification of the generalization and accuracy, the developed model is to estimate battery health conditions with the same type of battery under different experimental conditions. Experimental results of two types of batteries manifest that the proposed algorithm can provide accurate and credible battery health conditions with maximum relative errors of less than 4%.