基于内阻预测的阀控铅酸蓄电池故障预警方法  

Warning method of fault for valve-regulated lead-acid batteries based on resistance prediction

作　　者：张菲菲 王寅超 陆忠心 黄冬 王月强 ZHANG Fei-fei;WANG Yin-chao;LU Zhong-xin;HUANG Dong;WANG Yue-qiang(State Grid Shanghai Changxin Eeerie Power Company,Shanghai 201913,China)

机构地区：[1]国网上海市电力公司长兴供电公司,上海201913

出　　处：《电源技术》2020年第6期860-862,898,共4页Chinese Journal of Power Sources

基　　金：国网上海市电力公司科学技术资助项目(5209KZ-18001K)。

摘　　要：阀控铅酸(VRLA)蓄电池的故障影响电源的供电可靠性,而内阻是反映蓄电池运行状态的重要参数.提出基于蓄电池预测内阻及其变化率的故障预警方法,分析电池内阻变化原因,量化蓄电池内阻的影响因素,采用长短时法神经网络(LSTM),以充放电状态、电压、电流、温度作为输入量,构建蓄电池内阻预测模型,并根据电池内阻预测值及其变化率预判电池故障原因,发出蓄电池预警信号.试验结果验证了所提电池内阻预测模型误差在3%以内,基于电池内阻预测模型的蓄电池故障预警方法有利于蓄电池安全稳定运行.The fault of the valve-regulated lead-acid(VRLA)battery affects the reliability of the power supply,and its resistance is one of the important parameters reflecting the operating state of the battery.A fault warning method based on the battery resistance and its change rate was proposed.The change cause of the battery resistance was classified,and the influencing factors of battery internal resistance were quantified.The prediction model of the battery resistance was constructed by the long-and-short-time memory neural network(LSTM)with the inputs of charge and discharge state,voltage,current,temperature of the battery.According to the predicted value of the resistance and its change rate of the battery,the fault reasons of the battery were determined to issue the battery warning signal.The results of the example show that the error of the battery resistance prediction method is within 3%.The fault warning method of the battery based on the prediction model of the battery resistance is beneficial to the safe operation of the VRLA battery.

关 键 词：阀控铅酸蓄电池 长短时法神经网络 电池内阻 故障预警 

分 类 号：TM912[电气工程—电力电子与电力传动]