利用信息融合技术的储能锂离子电池组SOC估算

Study of SOC Estimation Algorithm for Energy Storage Lithium Battery Pack Based on Information Fusion Technology

在电池储能系统的实际工程中,电池组荷电状态(state of charge,SOC)估算精度越来越受重视。电池组容量、运行环境、循环时间和充放电倍率等都将影响电池组的SOC估算精度,采用单一的电池模型和数据模型很难获得准确的SOC。提出了一种基于信息融合技术的锂离子电池SOC估算方法,主要基于开路电压(open circuit voltage,OCV)-SOC曲线进行。根据锂离子电池运行特性,把OCV-SOC曲线空间划分为锂电池稳定运行区间、识别校正区间、过充区间和过放区间,并据此重新定义锂离子电池运行模式。然后根据其运行模式,在不同运行区间内对锂电池的估算模型进行切换和优化。采取基于信息融合的SOC估算方法,不断修正消除估算模型在运行状态下产生的各种误差,得到较为精确的SOC估算值。最后搭建实验平台,以某储能电站的实际储能工况对该算法进行实验验证,结果表明,上述SOC估算算法在实际锂电池储能系统应用中具有较强的可行性和实用性。

1According to practical engineering problems of battery energy storage system（BESS）, accuracy and robustness of state of charge（SOC） estimation becomes increasingly important. Battery pack capacity, operation condition, cycle times, environment temperature, charge and discharge rate will affect accuracy of SOC estimation, and it is difficult to obtain accurate SOC with single battery model and data model. In this paper, a new method for SOC estimation of lithium battery is proposed based on multi-model-driven and data-driven information fusion technology. Different operation sections are created on open circuit voltage（OCV）-SOC curve, i.e. flat voltage section, identification-correction section, over charge section and discharge section. In order to utilize estimation model according to different operation sections, the algorithm is switched to optimized estimation model to acquire better performance. Then the algorithm eliminates model error introduced by working state. Finally, an experimental platform is built, and experiments and numerical computation are conducted with actual operation data. Test result shows that this SOC estimation algorithm based on information fusion technology has better robustness in practical application of lithium battery energy storage system.