Power Quality Recognition And Analysis Based On Wavelet And SVM

With the development of science technology and the establishment of numerous industrial plants, the quality of electricity power has become increasingly demanding. Due to the widely use of a large number of nonlinear and impact loads, the quality of electricity is getting worse and worse. Some high precision electronic equipment controlled by the computer require high quality of electricity power, so decline in power quality may result in these electronic devices to stop working or produce false action. Some serious situation may cause damage to equipment and operators. Many power users as well as the electricity sector of electricity to concern the quality of the problem by the traditional sense of power supply reliability, power supply voltage stability, power supply frequency and waveform stability transition to a voltage oscillation and pulse transient aspects. Therefore, electricity power quality problem is an important subject in the research of modern power system.In order to improve the electricity power quality of the power system, it is needed to find a way to accurately detect and analyze the electricity power quality in which we can classify the existing problems and analyze them accurately and effectively. In this paper, the analysis method of power quality based on wavelet analysis is proposed and the following three aspects are included in this paper:1. The principle and algorithm of wavelet entropy are introduced, and some kinds of common disturbance types are presented, and their mathematical expressions and waveforms are also presented. By adding a noise signal to the voltage interruption of the transient power disturbance waveform first, then uses noise reduction principle of wavelet entropy to simulate the noise reduction.2. Through the analysis of the correlation principle of wavelet transform, it is sure that the wavelet transform modulus maxima can be used to detect the mutation point in electricity power quality. So it can be combined with the relevant problems of electricity power quality, such as voltage sag, voltage swell voltage interruption, voltage pulse transient and voltage oscillation problem. By using this method, we can accurately locate and analyze these faults.3. Based on wavelet analysis and support vector machine(SVM), a method of electricity power quality disturbance signal classification is proposed. By wavelet decomposing the characteristic value of electricity power quality disturbance signal and calculating the corresponding sample entropy, then the sample entropy is put into the SVM classifier and RBF classifier for training and test. Finally through the simulation experiment, the simulation diagrams are obtained based on SVM and RBF classifier. The feasibility of the method of classifying electricity power quality disturbance signals based on wavelet analysis and SVM is verified by comparison.