基于模糊神经网络的锂离子电池组均衡策略研究

Balancing Strategy of Lithium Ion Battery Pack Based on Fuzzy Neural Network

针对串联锂离子电池组在使用过程中的不一致性问题,提出模糊神经网络均衡策略。该策略以电池荷电状态为均衡变量,采用模糊神经网络算法,训练隶属度函数,得到隶属函数的最佳参数,使隶属度函数的设置不依赖专家经验,能够有效提高均衡精度,更好地改善电池组不一致性。使用Matlab/Simulink软件进行模型搭建并仿真,实验结果表明,与传统的模糊逻辑控制算法相比,在相同的静置、充电和放电均衡条件下,使用ANFIS算法均衡时间减少了约16%,能量利用率提高约1.31%,均衡后单体电池的离散度下降了约47%,验证了该均衡方案的可行性。

For the inconsistency problem of series connected lithium-ion battery pack in the process of use,this paper proposes a fuzzy neural network equalization strategy.The strategy takes State of Charge as the equalization variable,designs the Adaptive Network-based Fuzzy Inference System,trains the affiliation function,and obtains the optimal parameters of the affiliation function,so that the setting of the affiliation function does not depend on expert.It can effectively improve the equalization accuracy and better improve the battery pack inconsistency.The experimental results show that compared with the traditional Fuzzy Logic Control(FLC)algorithm,the equalization time is reduced by about 16%,the energy utilization is increased by about 1.31%,and the dispersion of a single cell after equalization is reduced by about 1.31%under the same static,charging and discharging equalization conditions using ANFIS algorithm.The dispersion of the single cell is reduced by about 47%.The feasibility of this equalization scheme is verified.