摘要
变电设备作为电网中的重要组成部分,其及时、准确的监测对于保障电网安全具有重要意义。目前对于变电设备的监测主要是通过日常巡视结合例行试验的方式,对设备运行状态监测存在延迟性和不全面性。本文通过大数据分析技术,依托变电设备的大量参数数据、运行数据、检修数据、试验数据等构建特征数据集,采用非参数检验方法对各参数进行校验,利用主成分分析对特征集进行降维处理,得到新的表征设备状态的特征数据集进行状态评价分析。通过采用梯度提升树构建设备状态评价模型,并结合交叉验证分析评估训练误差和泛化误差,通过与神经网络、支持向量机等其他分类模型进行比较分析,经测试数据验证,其预测结果具有较高的准确性。通过大数据分析技术与设备监测相结合,提高了设备状态监测的准确率,具有较好的推广应用价值。
As an important part of the power grid,the timely and accurate monitoring of substation equipment is of great significance to ensure the safety of the power grid.At present,the monitoring of substation equipment is mainly through daily patrol combined with routine test.There are delays and incompleteness in the monitoring of equipment operation status.By combining the big data analysis technology and relying on a large number of parameter data,operation data,maintenance data and test data of substation equipment,this paper constructs the characteristic data set,uses the nonparametric test method to verify each parameter,uses the principal component analysis to reduce the dimension of the characteristic set,and obtains a new characteristic data set representing the state of equipment for state evaluation and analysis.The equipment status evaluation model was constructed by using gradient lifting tree,and the training error and generalization error were evaluated by cross-validation analysis,and compared with other classification models such as neural networks,support vector machine,etc.,the test data show that the prediction result has high accuracy.Through the combination of big data analysis technology and equipment monitoring,the accuracy of equipment condition monitoring is improved,which is worth popularizing and applying.
作者
马洪斌
王文峰
石峰
杨飞
郇帅
MA Hongbin;WANG Wenfeng;SHI Feng;YANG Fei;HUAN Shuai(State Grid Zaozhuang Power Supply Company,Zaozhuang 277000,Shandong,China)
出处
《电力大数据》
2022年第4期48-55,共8页
Power Systems and Big Data
关键词
电力大数据
变电设备
主成分分析
交叉验证
梯度提升树
power big data
substation equipment
principal component analysis
cross validation
gradient lifting tree
