Battery Health Status Prediction Based on Incremental Capacity Curve

Machine learning-based methods are widely used for estimating the state of health of lithium-ion batteries. However, during the training of machine learning models, the health features used as inputs significantly impact the training effectiveness and the accuracy of SOH prediction. In practical scenarios, due to the varying effects of different aging stresses, the different degradation trajectories of batteries make it challenging to extract health features. This paper introduces a feature extraction method based on selecting values at different voltage points on the incremental capacity curve during the discharge process. Validation with battery experimental data from the University of Michigan and the Massachusetts Institute of Technology demonstrates that this method can effectively predict the state of health of batteries under different aging stresses.