Feature Optimization Base On Wavelet Transform Effects On Power Quality Disturbance Classification

Energy is an energy pattern which is economic, practical, clean and easy to control. It is important significance that the high-quality power ensure the security of electrical equipment and power grids, economicoperation, to improve product quality and to protect the broad masses of the people live a normal life. Its application has become one of the main indicators of the country’s development level, as science and technology and the development of the national economy, Power Quality issues have become more and more attention by the community, the establishment of power quality testing and analysis system, its correct detection, assessment and classification systems not only for electricity but also for users, becomes very important.This paper systematically introduces phenomenon and theories foundation of the power quality first, and expounds the classification, definition, characteristic of the various power quality disturbance and the reason of its creation, and the synopsis introduces our country power quality standard.In the right to conduct a detailed study on the basis of power quality, this paper propose improved methods based on feature optimization of wavelet transform. Due to the recognition rate of power quality disturbance using the conventional wavelet transform is not precise, proposed wavelet transform to extract the difference of the disturbance and the original signal’ energy as feature, and the pre-classification of the voltage interruption improve classification accuracy significantly. Removed from the insensitive classification feature, reduce the number of features and operating costs of the system. The proposed method to achieve the classification and recognition of the multiple signals include the sine wave signal, voltage swells, voltage sags, voltage interruptions, harmonics, impulse, switching transient, flicker, and composite disturbance, and has a relatively strong anti-noise performance, show that the algorithm is practical and feasible.