粒子群优化
均方误差
储能
健康状况
模拟退火
非线性系统
电池(电)
计算机科学
控制理论(社会学)
功率(物理)
算法
数学
统计
人工智能
物理
控制(管理)
量子力学
作者
Ran Xiong,Shunli Wang,Qi Huang,Chunmei Yu,Carlos Fernandez,Wei Xiao,Jinping Jia,Josep M. Guerrero
出处
期刊:Energy
[Elsevier]
日期:2024-04-01
卷期号:292: 130594-130594
标识
DOI:10.1016/j.energy.2024.130594
摘要
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 °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.
科研通智能强力驱动
Strongly Powered by AbleSci AI