基于EMD的异常用电检测方法

A detection method of abnormal electricity consumption based on EMD

异常用电行为的时频特性往往具有强随机不确定性,而固定参数相关的分析方法无法有效处理此类数据。为此,本文提出了一种基于经验模式分解(EMD)的异常用电检测方法。首先,针对用电数据的不同特点进行初步筛选,进而采用EMD方法对用户用电量和线损电量序列进行自适应分解,提取EMD分解所得高频分量,通过对其变化趋势和相关性进行分析,标定异常用电行为。结合实际案例的分析比对,验证了该方法的有效性。

Time-frequency characteristics of abnormal electricity consumption data have strong randomness and uncertainty. So the methods related to fixed parameters are not good at handling such data. To solve the problem, a detection method of abnormal electricity consumption based on empirical mode decomposition(EMD) is proposed in this paper. Preliminary screening for different characteristics of electricity consumption data is done first. Then the electricity consumption and line loss power sequences are decomposed self-adaptively by using the EMD method. The change trend and correlation of high-frequency components obtained by EMD decomposition are analyzed, and thus the abnormal electricity consumption is calibrated. Comparative analysis with actual cases verifies that the proposed method is proved to be effective in detecting the abnormal electricity consumption.