摘要
运行条件下气体绝缘组合电器(gasinsulatedswitchgears,GIS)的局部放电检测已取得了大量的应用,但对于检测到局部放电信号的严重程度评估仍然是亟待解决的难题。利用变电站现场GIS的局部放电检测数据,结合长短时记忆网络(long short-term memory network,LSTM)和Bagging集成学习方法,提出一种运行条件下GIS局部放电严重程度评估方法。先明确了用GIS设备未来1个月内的故障概率来定义局部放电严重程度,并基于该定义对大量的变电站现场检测数据确定了数据标签,建立了数据集。针对数据样本不均衡,利用Bagging集成学习方法将N个LSTM深度网络构建成适用于局部放电严重程度评估的集成学习模型。通过对由局部放电数据特征值、局部放电技术影响因素、设备运行信息等组成的特征向量进行分析,模型最终可以输出局部放电严重程度评估结果。通过与普通LSTM网络、反向传播神经网络(backpropagationneuralnetwork,BPNN)以及Bagging-BPNN方法的对比,以及变电站现场检测案例分析,结果表明所提方法可以有效地对运行条件下GIS局部放电进行严重程度评估,易于实施,与普通LSTM、BPNN和Bagging-BPNN相比评估结果的可信度更高。
Partial discharge(PD) detection for the gas-insulated switchgear(GIS) in service has get got a large number of promotions, but the assessment of the severity of detected partial discharge signals is still an open question. In this paper, based on the partial discharge detection data from substation GIS on site, combined with the long short-term memory(LSTM) network and Bagging ensemble learning, a method for assessing the severity of partial discharge for GIS in service was proposed. First, the definition of partial discharge severity for GIS in service was defined by the failure probability, a large set of site data were established and data labels were determined based on the definition. Second, aiming at the problem of imbalanced data, an ensemble learning model for assessing the severity of partial discharge was constructed using N individual LSTM depth networks via Bagging ensemble learning. By analyzing the eigenvectors composed of the characteristic of partial discharge data, technical impact factors, and equipment operation information, the model can output the severity assessment result. The comparative experiments with common LSTM network, backpropagation neural network(BPNN) and Bagging-BPNN were undertaken,as well as on site detection case analysis, the results show that the proposed method can effectively assess the severity of GIS partial discharge under operating conditions and is easily to implement. Compared with common LSTM, BPNN and Bagging-BPNN, the proposed method has a higher credibility.
作者
宋辉
代杰杰
李喆
罗林根
盛戈皞
江秀臣
SONG Hui;DAI Jiejie;LI Zhe;LUO Lingen;SHENG Gehao;JIANG Xiuchen(Department of Electrical Engineering,Shanghai Jiaotong University,Minhang District,Shanghai 200240,China)
出处
《中国电机工程学报》
EI
CSCD
北大核心
2019年第4期1231-1241,共11页
Proceedings of the CSEE
基金
国家重点研发计划项目(2017YFB0902705)
国家电网公司科技项目~~