摘要
为改善传统方法对监控设备典型缺陷分析结果的不理想,提出了基于数据挖掘技术的集中监控设备缺陷预测方法。预测集中监控设备在未来时间内的相关监控数据,并将其与历史典型缺陷发生期间内的监控数据相似度进行比较,以量化指标的形式对集中监控设备的典型缺陷进行预测。实验结果表明,利用所提方法对实际发生的油温异常典型缺陷进行预测,计算出的相似度指标形成了一个峰值,而缺陷未发生时的相似度指标较低,证明本文方法能够较好地体现缺陷发生的可能性,整体应用性较高。
In order to improve the traditional method to analyze the typical defect analysis results of monitoring equipment,a method based on data mining technology was proposed to predict the defects of centralized monitoring equipment.Which predicts relevant monitoring data of the centralized monitoring equipment with the results are compared with the similarity of monitoring data during the historical typical defect occurrence period to predict the typical defects of the centralized monitoring equipment in a form of quantitative indicators.The experimental results show that the proposed method can be used to predict the typical oil temperature anomalies with a calculated similarity index forms a peak while the similarity index is lower when the defect does not occur,which proves that the method with a high overall application can be used to reflect the possibility of defects.
作者
肖飞
冷喜武
徐元直
何忠
XIAO Fei;LENG Xi-wu;XU Yuan-zhi;HE Zhong(State Grid Shanghai Electric Power Company,Shanghai 200122,China;National Electric Power Dispatching and Control Center,State Grid Corporation of China,Beijing 100031,China;Tellhow Software Software co., LTD,Nanchang 330096,China)
出处
《科学技术与工程》
北大核心
2020年第16期6522-6526,共5页
Science Technology and Engineering
关键词
数据挖掘
设备缺陷
集中监控
趋势预测
data mining
equipment defect
centralized monitoring
tendency projection
