服务于用电设备的快速辨识边沿检测方法研究

Study of Fast Identification Edge Detection Method for Electrical Equipment

针对当前用户侧用电设备非侵入式辨识中负荷边沿检测方法准确率不高的问题,该文在研究阈值检测算法（threshold detection algorithm,TDA）、暂态能量启动算法（transient energy to start the algorithm,TEA）、微分算子（method of differential algorithm,MDA）以及拟合方法的基础上,提出一种基于高斯滤波器和工业检测累加求和（cumulative sum,CUSUM）算法的边沿检测方法,解决传统算法考虑因素单一、精度不高等问题。该方法采用自适应高斯滤波器能有效过滤噪声,同时保留突变点信息的特点,通过检测去噪后的突变点波动,提升检测的准确性。结合CUSUM算法对设备状态与工作模式变化检测灵敏的特点,提升设备模态的检测速度和准确性。通过搭建的非侵入式负荷辨识平台,对提出的方法进行仿真和实验验证,显示所提出的方法能有效提高用电设备的检测速度和准确率。

In view of the current method for load of user side electric equipment in edge detection performance problems, based on the study of threshold detection algorithm（TDA）, transient energy startup algorithm（TEA） method of differential operator algorithm（MDA） and the fitting process algorithm, a method of edge Gauss filter and industrial detection based on cumulative sum（CUSUM） algorithm were proposed, which solve the problem of single and low precision of traditional algorithm. The method adopt the adaptive Gaussian filter to filter the noise effectively, while retaining the characteristics of the mutated point information, and the accuracy of the detection is improved by detecting the fluctuation of the mutation point after denoising. Combined with CUSUM algorithm, the detection speed and accuracy of the device mode are improved. Through the construction of non-invasive load identification platform, the proposed method was simulated and verified, and the proposed method can effectively improve the detection speed and accuracy of the electrical equipment.