小波神经网络预测变电站VRLA电池工作寿命  

Prediction of operating life of substation VRLA battery by wavelet neutral network

作　　者：梅成林 王文强[2] 刘波峰[2] 李涛[2] 

机构地区：[1]广东电网责任有限公司电力科学研究院,广东广州510080 [2]湖南大学电气与信息工程学院,湖南长沙410012

出　　处：《电池》2014年第6期351-353,共3页Battery Bimonthly

摘　　要：研究预测变电站阀控式密封铅酸(VRLA)电池工作寿命的方法。引入小波神经网络(WNN),建立电池工作寿命的WNN模型,再通过实验数据对WNN模型进行训练,得到用于电池运行寿命预测的WNN模型,最后将WNN模型的预测结果与实际结果进行对比。预测的平均相对误差为1.49%,最大相对误差为2.35%。WNN模型可以快速、准确地得到电池的工作寿命,可用于变电站电池工作寿命的预测。A method for predicting substation battery operating life was researched. Wavelet neural network （WNN） was intro- duced, a WNN model of battery operating life was established, then the WNN model was trained with experimental data to get WNN model for battery operating life prediction. Finally, the predicted results of the WNN model and actual data were compared. The average relative error of prediction was 1.49% , the maximum relative error was 2. 35%. The WNN model could get the battery operating life quickly and accurately, was able to use to predict the substation battery operating life.

关 键 词：阀控式密封铅酸(VRLA)电池 工作寿命 小波神经网络 

分 类 号：TM912.1[电气工程—电力电子与电力传动]