论文部分内容阅读
电力网络中信息系统与物理系统的深度融合,导致现代电力系统易受异常数据的影响。现有的电力数据异常检测方法未能充分挖掘数据特征,存在计算复杂、灵活性差、精度较低等缺点。提出一种基于时间序列提取和维诺图的异常数据检测方法,利用重要点分段的时间序列提取方法,将高维数据进行降维处理,并将其映射到二维平面上,构造维诺图分区,进而检测出异常数据。该方法可降低数据维度和算法复杂度,能根据序列特征灵活设定异常阈值,实现异常数据的准确检测,仿真实验证明所提方法的有效性。
The in-depth integration of information systems and physical systems in power networks has led modern power systems to be susceptible to abnormal data. The existing power data anomaly detection method fails to fully exploit the data features, which has the disadvantages of complex computation, poor flexibility and low accuracy. This paper presents a new method for detecting abnormal data based on time series and Venn diagrams. Using the method of time series extraction of important points, the high dimensional data is reduced dimensionally and mapped onto a two-dimensional plane. Figure partition, and then detect abnormal data. The method can reduce the data dimension and the complexity of the algorithm, and can flexibly set the abnormal threshold according to the sequence features to achieve the accurate detection of abnormal data. The simulation results show the effectiveness of the proposed method.