基于等压升充电时间的镍镉蓄电池SOH预测  

SOH Prediction of nickel-cadmium battery based on the charging time under constant voltage rise

作　　者：孙宁 王升晖 于天剑[2] 代毅 SUN Ning;WANG Shenghui;YU Tianjian;DAI Yi(CRRC Qingdao Sifang Rolling Stock Co.,Ltd.,Qingdao,Shandong 26611l,China;School of Traffic&Transportation Engineering,Central South University,Changsha,Hunan 410075,China)

机构地区：[1]中车青岛四方机车车辆股份有限公司,山东青岛266111 [2]中南大学交通运输工程学院,湖南长沙410075

出　　处：《机车电传动》2022年第5期103-108,共6页Electric Drive for Locomotives

基　　金：湖南省自然科学基金项目(2020JJ5757)。

摘　　要：动车组镍镉蓄电池健康状态会影响列车运行安全,但由于动车组运行工况复杂,导致现有方法无法较好地在线监测电池健康状态。为研究蓄电池健康状态变化趋势和实现在线预测,假设蓄电池在通过充电机充电的过程中无放电过程。为此,提出基于等压升充电时间的蓄电池健康状态在线预测方法,该方法通过对基于移动电压窗口的等压升充电时间与蓄电池健康状态综合相关性分析以确定最佳的等压升电压区间(即充电起始电压与截止电压之间),再通过等压升电压区间提取最佳等压升充电时间作为长短期记忆网络模型输入,利用麻雀搜索算法对长短期记忆网络参数寻优,建立蓄电池健康状态预测模型,实现了蓄电池健康状态的在线预测。试验结果表明,相较于传统长短期记忆网络和反向传播神经网络,基于麻雀搜索算法优化长短期记忆网络的蓄电池健康状态预测模型具有更高的预测精度。The state of health(SOH)of nickel-cadmium battery used on EMUs influences the operation safety of trains.Due to the complex working conditions of EMUs,the existing SOH monitoring methods are unable to realize good online monitoring.To study the change tendency of battery SOH and realize online prediction,this paper proposed an online SOH prediction method based on the charging time under constant voltage rise,with the assumption that there is no discharge when using charger to charge the battery.This method determined the optimal constant voltage rising range(i.e.the voltage difference between the start and end of charging)through a comprehensive correlation analysis of charging time under constant voltage rise and battery SOH based on moving voltage window,and then extracted the optimal charging time under constant voltage rise from the voltage range as the input for the long short-term memory(LSTM)model.The sparrow search algorithm(SSA)was used to optimize the LSTM parameters,establishing a battery SOH prediction model,and realizing the online prediction of battery SOH.The test results show that,compared with traditional LSTM and back propagation(BP)neural network,the battery SOH prediction model based on SSA-LSTM has higher prediction accuracy.

关 键 词：动车组 镍镉蓄电池 SOH预测 等压升充电时间 SSA-LSTM 

分 类 号：TM912[电气工程—电力电子与电力传动]