基于标准差偏离倍数的暂态事件检测算法  

Transient event detection algorithm based on standard deviation multiple

作　　者：张广龙 任建文[1] 周明[1] Zhang Guanglong;Ren Jianwen;Zhou Ming(North China Electric Power University,Baoding 071003,Hebei,China)

机构地区：[1]华北电力大学,河北保定071003

出　　处：《电测与仪表》2023年第2期138-146,共9页Electrical Measurement & Instrumentation

基　　金：国家重点研发计划项目(2016YFB0901100)。

摘　　要：暂态事件检测是非侵入式负荷监测的重要研究部分。然而已有的暂态事件检测方法需要已知投切负荷的有功功率跃变幅值,并依此设定功率跃变前后功率差的检测阈值,导致无法应用于多个有功功率跃变幅度不同的负荷同时投切的情况。为此,提出了一种基于标准差偏离倍数的暂态事件检测算法。该算法通过有功功率偏离均值相对于标准差的倍数来检测变点是否出现,并通过偏离倍数最近一次穿越零点的时刻来精确定位变点位置。通过仿真对比了偏离倍数法和滑动窗算法的检测速度和精度,用BLUED数据集验证了算法在多个不同特性负荷投切情况下的有效性。该算法相比已有的检测算法,不仅不需要设定功率跃变前后功率差的检测阈值,还能应用于多个不同特性负荷投切的情况。经测试该算法具有更好的检测速度、检测精度,较低的误检率和漏检率。Transient event detection is an important part of non-invasive load monitoring. However, the existing transient event detection methods need to know the active power jump amplitude of the load switching, and set the detection threshold of the power difference before and after the power jump, which makes it unable to be applied to multiple loads with different active power jump amplitude. Therefore, a transient event detection algorithm based on standard deviation multiple is proposed in this paper. The algorithm detects whether the change point occurs by the multiple of the deviation of active power from the mean value relative to the standard deviation, and accurately locates the change point by the time when the deviation multiple crosses the zero point most recently. The detection speed and accuracy of the deviation multiple method and sliding window algorithm are compared through simulation. The validity of the algorithm is verified by BLUED data set in the case of multiple load switching with different characteristics. Compared with the existing detection algorithms, this algorithm not only does not need to set the detection threshold of power difference before and after power jump, but also can be applied to the case of multiple load switching with different characteristics. The testing results show that the algorithm has better detection speed, detection accuracy, lower false detection rate and missing detection rate.

关 键 词：变点检测 非侵入式负荷监测 BLUED数据集 算法统计特性 

分 类 号：TM933[电气工程—电力电子与电力传动]