基于大数据挖掘技术的输变电设备故障诊断方法

发布时间：2018-12-17 03:31

【摘要】：传统输变电设备故障诊断方法具有易受专家意见主观影响以及模型固化的缺点。为此,提出了基于大数据挖掘技术的设备故障诊断方法,介绍了设备故障模式聚类算法、状态参量相关性分析算法以及基于相关性矩阵的故障诊断方法等基于大数据分析的设备故障诊断关键技术,并采用某电网公司500 kV电压等级油浸式变压器套管近10 a故障记录数据作为数据挖掘案例进行了实证。研究结果表明:k-means聚类算法配合轮廓系数能准确得得出故障分类模式;Apriori关联算法配合Tanimoto系数可用于评估状态参量之间的强弱关系;基于皮尔逊相关系数构造故障诊断矩阵能够准确判断出与实际运行维护试验结果相符合的设备故障模式。因此,基于大数据挖掘技术的输变电设备故障诊断方法能够有效挖掘出设备状态记录数据内在的规律,实现具有数据自适应性的、更加准确的设备故障诊断。
[Abstract]:The traditional fault diagnosis method of transmission and transformation equipment has the disadvantages of subjective influence of expert opinion and solidification of model. In this paper, a fault diagnosis method of equipment based on big data mining technology is proposed, and the clustering algorithm of equipment fault pattern is introduced. The key techniques of equipment fault diagnosis based on big data's analysis are the correlation analysis algorithm of state parameter and the fault diagnosis method based on correlation matrix. A 500 kV voltage grade oil-immersed transformer bushing fault record data of a power grid company is used as a data mining case. The results show that the fault classification mode can be obtained accurately by using k-means clustering algorithm combined with contour coefficient, and the Apriori association algorithm with Tanimoto coefficient can be used to evaluate the strong and weak relationship between state parameters. The fault diagnosis matrix based on Pearson correlation coefficient can accurately judge the equipment fault mode which is consistent with the actual operation and maintenance test results. Therefore, the fault diagnosis method of transmission and transformer equipment based on big data mining technology can effectively excavate the inherent rule of equipment state record data, and realize more accurate equipment fault diagnosis with data adaptation.
【作者单位】：电力系统发电设备控制和仿真国家重点实验室(清华大学电机工程与应用电子技术系);广西电网公司电力科学研究院;
【基金】：国家自然科学基金(51429701) 南方电网公司科技项目(GX2014-2-0025)~~
【分类号】：TM507

【相似文献】