The Study Of Dection And Identification Of Voltage Sag Based On Generalized S-transform And SVM

With the development of modern industrialization, digital automatic control technology has been applied in large scale in industrial production, such as frequency control devices, programmable logic controllers and so on. The extensive use of high precision electronic equipment put a higher demand on power quality of the power supply system. Among various types of power quality problems, voltage sag is more common and influential. The accurate identification of voltage sag disturbance source has great significance to prevent and govern voltage sag. And it also can provide a proof to resolve disputes between the power supply system and users.This paper analyzes the different types of voltage sag disturbance sources firstly. The reasons of causing voltage sag include single voltage sag disturbance source and composite voltage sag disturbance source. Then the reasons of causing voltage sag and the characteristics of sag waveform are studied. S-transform has good time-frequency characteristics, and it is suitable for the analysis of transient power quality disturbances. On the basis of the standard S-transform, the paper proposes that using generalized S-transform to detect the amplitude of voltage sag and the beginning and ending time of disturbance. Considering voltage sag characteristics, the features indices of identification of voltage sag are extracted based on generalized S-transform.Support vector machine has an outstanding performance of classification under the small sample case.And the training time is short. In order to identify the disturbance accurately, the paper uses the method of one-against-one to build support vector machine, and the parameters of multiple classifiers are optimized by particle swarm algorithm.Finally through using the different samples of voltage sag which are obtained by voltage sag simulation system in MATLAB/Simulink, the proposed method is verified. The simulation results show that the proposed method can correctly identify single and composite disturbance source of voltage sag under different noisy conditions. S-transform can accurately detect the voltage sags. The extracted feature indices of voltage sag have great anti-noise performance. The proposed method has good adaptability and can be applied to identify the voltage sag disturbance sources.