A Method for Sequential Detection and Estimation of Battery Energy Storage Operational Data Based on the Integration of Equivalent Circuit Models and Data-Driven Models

The State of Charge (SOC) and State of Health (SOH) collectively provide a comprehensive reflection of a battery's operational condition. This paper proposes a joint estimation method that couples equivalent circuit models with data-driven models based on this premise. Aimed at simultaneously estimating SOC and SOH to achieve sequential detection of battery operational data. By establishing a second-order RC equivalent circuit model and employing a recursive least squares method with a forgetting factor, this method can online identify battery parameters based on different SOC and SOH states. Furthermore, the SOH of the battery is estimated using Convolutional Neural Networks (CNN), and the estimated SOH values are combined with the battery's rated capacity to calculate its actual capacity. Following this, the second-order RC state-space equation is updated, and the battery's SOC is estimated using the Unscented Kalman Filter algorithm, thus effectively coupling the estimation of SOC and SOH. And the results shows that the proposed joint estimation method can accurately track the real values of lithium-ion battery SOC and SOH throughout the battery's lifecycle.