Improved cooperative competitive particle swarm optimization and nonlinear coefficient temperature decreasing simulated annealing-back propagation methods for state of health estimation of energy storage batteries

At present, the accurate establishment of the battery model and the effective state of health (SOH) estimation under actual energy storage conditions have become the main problems in new energy storage stations. Therefore, a SOH estimation method based on cooperative competitive particle swarm optimization (CCPSO) and nonlinear coefficient temperature decreasing simulated annealing-back propagation (NSA-BP) is proposed. The novelty of this research mainly includes the design of extraction methods in different health indicators (HIs) and the construction of developed NSA-BP network for SOH estimation. In this research, the contributions of SOH estimation are mainly to assist in battery replacement and provide relevant economic reference. Low-rate constant current energy storage degradation experiments and a variable-rate energy storage degradation experiment are performed for different battery packs at 25 degrees C. The experimental results indicate that the root mean square error (RMSE) and the mean absolute error (MAE) of the proposed method are 0.00588 and 0.00481 under the 0.5 rate condition, and the corresponding values are 0.00732 and 0.00639 under the variable-rate condition. Under the same condition, the proposed SOH estimation method is superior to the methods before improvement in RMSE and MAE, which can provide a basis for efficient monitoring of energy storage batteries.