Applications Of Wavelet-probability Neural Network In Power Quality Detection

Power is a widely used energy. It is important significance that the high-quality power ensure the security of electrical equipment and power grids,economic operation,to improve product quality and to protect the broad masses of the people live a normal life.The recent years, With the vigorous development of China's national economy,the power load increases sharply and the load capacity is also growing with a variety of rectifiers,frequency converter,electric arc furnace,electrification railway. These new non-linear,impact and imbalance of electricity,make the power grid voltage and current waveform distortion,or cause the voltage fluctuation and flicker, three-phase unbalanced and even give rise to fluctuations in the system frequency,It cause seriously pollution to the quality of power supply; On the other hand, the rapid development of modern science and technology promote that industrial production processes is highly automated and intelligent, these high-tech industry need to supply high quality requirements, many of the electronics, power electronics equipment are particularly sensitive to the quality of power supply, power supply work properly only be to achieve very high quality conditions in order to work properly.For a long time, Fourier transform and Short-time Fourier transform are the dominant position in signal analysis, however because of its inherent defects, They can not be effectively applied to process non-steady-state's disturbances of power quality. Wavelet transform can provide a variable time-frequency's window and it has characteristic of Multi- resolution Analysis, which makes wavelet transform to have Incomparable advantages in analysis the non-steady-state signal than the traditional method.This article details the basic principles of the wavelet transform and wavelet packet transform; Probabilistic neural network's basic theory, it's main contents include the basic theory of neural networks, probabilistic neural network structure and learning algorithm; In the simulation, analysis some of the major state,dynamic disturbance power quality using wavelet transform, Accord to the Parseval theorem, extract of energy as a feature vector input the probabilistic neural network, use the probabilistic neural network to detect the type of power quality disturbance, improve the problems in simulation; use the Matlab language, design of a dialog-based user interface, by pressing the button in the interface to realized the signal input,wavelet transform,the wave transform's output,extract the characteristics of the energy,output the automatic classification of power quality by probabilistic neural network.