断路器控制系统中的异常数据检测

The Abnormal Data Detection in the Control System of Circuit Breaker

针对断路器控制系统中异常数据检测存在的准确度低、时间复杂度高等问题,引入数据流挖掘技术,提出了一种基于滑动窗口和K—近邻距离的数据检测算法。在该算法中,利用断路器系统中的电流或电压信号的有效值对当前滑动窗口中的所有数据进行剪枝,筛选出绝大部分的正常数据,再利用K—近邻距离的数据检测方法对剩下的可能是异常的数据进行进一步的筛选,从而可以较快且较准确地检测出数据流中的异常数据。通过实验证明,在对同一数据流进行检测时,与其它数据检测算法相比较,该算法具有更好的执行效率和准确度。因此文中提出的算法能很好的运用到断路器控制系统中的异常数据检测。

Aiming at the inaccuracy and high time complexity of outlier data in the control system of circuit breaker, the date stream mining technology was introduced and a new algorithm of date detection which was based on K-close distance in the sliding windows was proposed. In this algorithm, using the rms value of the Current or voltage signal of the system of circuit breaker in the sliding windows to purn all the data, and screen out the most normal data, then using the detection method of the K-close distance to detect the remaining da-ta, thus detect the abnormal data in the stream data quickly and accurately. The results of the experiments show that using this algorithm and other algorithms to detect the same stream data, this algorithm has the better execution efficiency and accuracy. So this algorithm will be good in the detection of abnormal stream data in the control system of circuit breaker.