为了提高锂电池荷电状态(State of Charge,SOC)的估计精度,提出一种采用两级核极限学习机(Kernel Extreme Learning Machine,KELM)模型的SOC估计方法.该模型的前后两级分别为SOC预估计模型和误差补偿模型,其内核均采用KELM算法.前一级的SOC预估计模型以电池的工作电压和电流作为输入变量,SOC预估计值作为输出变量;后一级的误差补偿模型以预估计模型的输入和对应的输出作为输入变量,SOC预测误差作为输出变量;然后,采用后一级的误差预测结果对前一级的SOC预估计值进行补偿,得到最终的SOC估计值.为验证该方法的有效性和先进性,分别在锂电池恒流放电工况测试和动态应力测试两种情况下开展了实验研究,对比分析了 ELM、KELM和两级KELM三种方法的估计精度.实验结果表明,本文所提的两级KELM模型估计精度有显著提高,最大误差不超过1.36％.
State of Charge Estimation of Lithium Battery Based on Two-stage KELM Model
In order to improve the state of charge(SOC)estimation accuracy of lithium batteries,a two-stage kernel extreme learning machine(KELM)model is proposed.The two stages of the model are SOC pre-estimation model and error compensation model respectively,and the kernel of the model adopts KELM algorithm.The previous SOC estimation model takes the operating voltage and current of the battery as input variables,and the SOC estimation as output variables.The latter level of error compensation model takes the input and output of the pre-estimation model as the input variables,and the SOC prediction error as the output.Finally,the error prediction results of the latter stage are applied to compensate the SOC estimates of the previous stage,and the final SOC estimates are obtained.In order to verify the effectiveness and superiority of the proposed method,two experimental studies were carried out under the conditions of constant current exile testing and dynamic stress testing(DST)of lithium batteries,and the estimation accuracy of ELM,KELM and two-stage KELM methods were compared and analyzed.The experimental results show that the accuracy of the proposed two-stage KELM model is significantly improved,and the maximum error is less than 1.36％.
lithium batterystate of chargetwo-stage KELMerror compensation
国家科技期刊平台
NSTL
万方数据
