电容型设备介损在线监测数据趋势分析

Trend Analysis of On-line Monitoring Data for Dielectric Loss of Capacitive-type Equipment

介质损耗角正切值是评价电容型变电设备绝缘状况的重要参数，经过预处理后的在线监测数据能够更加准确地显示设备状况的变化规律，但要检测出设备绝缘的长期发展趋势，还需对数据进行趋势分析。介绍了差分法和斜率法的原理，并利用其原理分别对阶跃增长和持续增长两种不同的故障模式进行了趋势分析。结果表明，差分法能有效地检测出数据的阶跃变化，斜率法能判断缓慢而持续增长数据的整体变化趋势。差分法和斜率法结合起来，就既能鉴别出数据的奇异突变又能检测出数据的整体变化趋势。

Dielectric loss factor is a very important parameter for estimating the insulation condition of capacitive-type substation equipment. The on-line monitoring data after being pretreated can reveal the change law of power equipment more accurately. However, the trend analysis of on-line data is needed to find out the long-term change trend of the apparatus. In the paper, the principles of both the slope coeffcient method and the difference method are presented, by which the trend analysis is carried out for two different fault patterns arising from bouncing increase and sustaining respectively. The study shows that the difference method can detect the bouncing increase in the data, while the coefficient method can determine the general trend of the data which changes in a slow and sustaining manner. The bouncing increase and the sustaining increase can be found if two methods are used together.