奇异值分解和最小二乘支持向量机在电能质量扰动识别中的应用

Application of SVD and LS-SVM in Power Quality Disturbances Classification

基于奇异值分解(singular value decomposition,SVD)和最小二乘支持向量机(least square support vector machine,LS-SVM)提出电能质量扰动类型识别的新方法。通过对电能质量扰动信号的小波包变换系数矩阵进行奇异值分解,将基频、扰动频率分量、噪声分解到不同的正交特征子空间。再与正常电压信号的奇异值作比值以抵消噪声能量的影响,最大限度地体现出扰动类型间的细微差别,以此作为扰动特征向量,作为最小二乘支持向量机分类器的输入参数,来实现电能质量扰动类型的识别。仿真结果表明,该方法识别准确率高,受噪声影响小,算法稳定性好。

Based on singular value decomposition （SVD） and least square support vector machines （LS-SVM）, the paper proposed a new method of identifying the type of power quality disturbance. The method decomposed the voltage signal with components of fundamental frequency, the faulty frequency and the noises into different orthogonal characteristics subspace by SVD of the coefficient matrix from the wavelet package transformation; then used them to make devision with singular value of normal voltage to cancle the effect of niose energy, displayed the tiny differences among different disturbance types as much as possible. And then it used them as the input parameters of the LS-SVM to realize the identification of the power quality disturbance type. Simulation results indicate that the method has good performance of accuracy and stability, and is immune to noises.