基于高适用性特征和MIPOA-DHKELM的锂电池SOH估计  

Lithium battery SOH estimation based on high applicability features and MIPOA-DHKELM

作　　者：张宇[1] 胡朝朝 吴铁洲[1] ZHANG Yu;HU Zhaozhao;WU Tiezhou(Hubei Key Laboratory for High-efficiency Utilization of Solar Energy and Operation Control of Energy Storage System,Hubei University of Technology,Wuhan Hubei 430068,China)

机构地区：[1]湖北工业大学太阳能高效利用及储能运行控制湖北省重点实验室,湖北武汉430068

出　　处：《电源技术》2024年第12期2419-2425,共7页Chinese Journal of Power Sources

基　　金：国家自然科学基金(51677058)。

摘　　要：现有研究注重于提高锂电池健康状态(state of health,SOH)的估算精度,而缺乏实际应用性。针对这一问题,从恒流阶段电压在4.0~4.2 V之间的数据和恒压阶段电流在1.5~0.3 A之间的数据中各提取了两个具有高适用性的健康特征,通过这两个充电数据片段均可准确估计锂电池的SOH。此外,通过多种策略改进鹈鹕优化算法(pelican optimization algorithm,POA),使其收敛速度更快,种群分布更均匀。最后,采用改进的POA来优化多层极限学习机和混合核极限学习机融合成的深度混合核极限学习机模型。经实验验证,该方法无需大量充电数据即可提取健康特征,并且能够很好地追踪容量再生现象。在所有对比模型中,该模型预测精度最高,误差分布最稳定。The existing research focuses on improving the estimation accuracy of the state of health(SOH)of lithium-ion batteries but lacks practical applicability.To address this issue,this study extracted two highly applicable health features from the data during the constant current stage with voltages between 4.0-4.2 V and the data during the constant voltage stage with currents between 1.5-0.3 A.Both charging data segments can accurately estimate the SOH of lithium-ion batteries.Additionally,the pelican optimization algorithm(POA)was improved through multiple strategies to enhance its convergence speed and population distribution.Finally,the improved POA was used to optimize the deep hybrid kernel extreme learning machine model(DHKELM),which combines multiple layers of extreme learning machines and hybrid kernel extreme learning machines.The experimental results demonstrate that this method can extract health features without requiring a large amount of charging data,and it can effectively track capacity regeneration phenomena.Among all the compared models,this model exhibits the highest prediction accuracy and the most stable error distribution.

关 键 词：SOH 高适用性特征 多策略改进POA 深度混合核极限学习机 

分 类 号：TM912[电气工程—电力电子与电力传动]