Battery Health Monitoring under Multi-dynamic Operating Conditions Based on Fused-model

Accurate monitoring of the state of health (SOH) of lithium-ion batteries is vital for electric vehicles. However, it faces considerable challenges owing to the uncertainty of operating conditions and limited computational power available on board. An efficient method is introduced for estimating the SOH of a battery under multi-dynamic operating conditions using a fused-model approach. First, this method is based on the electrochemical theory and a simplified electrochemical model is established to characterize the physicochemical reactions in the battery. Second, the areas of voltage and state of charge errors under different aging cycles are adopted as the aging features, whereas the model parameters remained constant. Integrating these features with those that responded to the dynamic operating conditions resulted in more stable fused features under multi-dynamic operating conditions. Finally, the multiple fused features are used as inputs to a support vector regression model, enabling precise SOH estimation under diverse dynamic operating conditions. This framework is highly generalizable without the need to update the model parameters or use complex algorithms, and the computational effort is sufficiently small to achieve accurate estimation results, with a root mean square error and mean absolute error consistently within 1%.