面向非技术性损失的用电异常检测方法分析

Analysis of electricity anomaly detection method for non technical loss

在非技术性损失条件下进行用电异常检测,提高用电监测和管理能力,保障供电稳定。提出一种基于谱特征提取和衰减特征分量融合的用电异常检测方法,提取用电输出流量序列,对电力输出序列进行频谱特征分解和离散性调度,采用高斯随机频谱检测方法进行用电输出流量序列的特征提取,采用小波尺度分解进行非技术性损失下的用电异常流量的多元尺度分解和衰减特征分量融合,建立检测统计量,在最佳判决条件下进行用电异常特征提取,实现非技术性损失下的用电异常检测。仿真结果表明,采用该方法进行用电异常检测的准确检测概率较高,降低了电能衰减,减小了用电的非技术性损失,提高用电监测和管理能力。

Electric anomaly detection in the condition of non technical losses,improve the power monitoring and management ability,ensure a stable supply of electricity. This paper puts forward a new extraction and the attenuation of electric anomaly detection method of feature component fusion feature spectrum based on the extraction of electricity output flow sequence,spectral feature decomposition and discrete scheduling on power output the detection method of random sequence,Gauss spectrum was extracted by the characteristics of the electrical output flow sequence,using the wavelet scale decomposition using multiple scale electric anomaly decomposition and attenuation characteristics of component non technical losses under the established test statistics,was extracted by electric anomalies at the best decision conditions,implementation of electric anomaly detection non technical losses. The simulation results show that by using this method is accurate with high detection probability electric anomaly detection,reduce the power attenuation,reduced the non technical loss of electricity can be used to improve the power monitoring and management ability.