基于小波分析的电力信息通信过程不良数据辨识研究  

Research on Bad Data Identification in Electric Power Information Communication Process Based on Wavelet Analysis

作　　者：王渭 马梦轩 王圣杰 温兴贤 陈峥 WANG Wei;MA Meng-xuan;WANG Sheng-jie;WEN Xing-xian;CHEN Zheng(State Grid Ningxia Electric Power Co.,Ltd.,Information and Communication Company,Yinchuan 750001 China)

机构地区：[1]国网宁夏电力有限公司信息通信公司,宁夏银川750001

出　　处：《自动化技术与应用》2022年第10期129-132,共4页Techniques of Automation and Applications

摘　　要：传统的电力通信过程不良数据辨识方法的消噪能力较差,导致辨识效率较低。为此,本研究基于小波分析设计了新的电力信息通信过程不良数据辨识方法。根据小波变换的奇异性对电力信息通信过程进行局部奇异性检测,根据检测结果,结合神经网络算法区分正常数据和不良数据。然后采用软阈值和硬阈值去噪方法消除不良数据中的含噪信号,在计算噪声强度后,将某一尺度内的小波变换系数的平方由小到大排列,并计算似然估计向量,再根据向量中的最小值和最大值删除信号中的噪声部分。实验结果表明:该方法能够有效提高消噪能力、增强辨识效率。The traditional identification method of bad data in power communication process has poor denoising ability, which leads to low identification efficiency. Therefore, a new identification method of bad data in power information communication process is designed based on wavelet analysis. According to the singularity of the wavelet transform, the local singularity detection of the power information communication process is carried out. According to the detection results, the normal data and bad data are distinguished by the neural network algorithm. Then the soft threshold and hard threshold denoising methods are used to eliminate the noisy signals in the bad data. After calculating the noise intensity, the square of the wavelet transform coefficient in a certain scale is arranged from small to large, and the likelihood estimation vector is calculated. Then the noise part in the signal is deleted according to the minimum value and maximum value in the vector. Experimental results show that the proposed method can effectively improve the ability of noise elimination and enhance the identification efficiency.

关 键 词：小波分析 电力通信 不良数据 数据辨识 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]