An identification method for disturbance signal basing on support vector machine with optimized multi-universe algorithm

Aiming at the problems of low accuracy and poor stability of disturbance signal of power quality classification and identification in distribution network, a power quality disturbance identification method based on multi universe algorithm optimized support vector machine is proposed. Aiming at the scene of diversified noise of power quality disturbance source signal, the adaptive empirical mode decomposition method is used to process the noise, and a series of inherent mode function components are obtained. Then, using the natural mode function component, the sample entropy of the first seven natural mode components IMF of disturbance signal of power quality is calculated, and a set of eigenvectors are constructed. Aiming at the problem that it is difficult to select the parameters of support vector machine, which is easy to lead to great deviation of classification and identification results, an improved multiuniverse optimization algorithm is proposed to optimize the penalty factor and kernel function parameters of support vector machine. Based on the optimized penalty factor and kernel function parameters of support vector machine, the classifier of support vector machine based on multi universe algorithm is constructed, and the classification and identification method of disturbance signal of power quality in multi noise scene is proposed. 15 kinds of disturbance signals of power quality are tested by adding 0dB, 20dB and 60dB Gaussian white noise, and compared with five common classifiers. The simulation results show the efficiency and accuracy of the proposed method.