基于可去间断点的连续采样飞点检测算法

电力系统电气量采样中的飞点数据是引起保护误动的主要原因之一。目前,国内外对于飞点等异常采样的研究较少,且这些异常采样检测算法存在着针对性不强、灵敏度不高等不足。深入研究飞点产生的原因及其特征,提出一种基于连续函数可去间断点的飞点检测算法,并给出了灵敏且自适应的检测门槛,运用本算法能够迅速、准确地检测出飞点数据。另外,还提出基于正弦函数的预测修复算法,检出飞点后运用该修复算法可将飞点修复到正常值,不需闭锁保护,完全消除了飞点对保护的影响。仿真实验结果表明,所提算法是可行的。

A flying spot detection algorithm for continuous sampling based on removable discontinuous points

Flying spot results in a lot of protection malfunctions. Few researches are focusing on the flying spot. In practice, most abnormal samples detection algorithms are with ambiguous pertinence, insensitivity, ambiguous threshold and easy wrong ruling. Some detection algorithms only carry out simple latch-up protection, which decrease the malfunction rate while delay the protection action time. To solve these difficulties, a new sampling flying spot algorithm is proposed, which is based on removable discontinuous spot in continuous function. At the same time, a remedy algorithm based on predicting sine curve is proposed. The proposed algorithm can detect the flying spot effectively, and the remedy sampling algorithm can correct the flying spot without latch-up protection. The proposed algorithms can remove the malfunctions for flying spot. The simulation results show the proposed algorithms are effective.