Analysis And Recognition Of Ship Power Quality Transient Disturbances Based On Wavelet Transform

In order to analyze the problem of ship power quality, find the cause of power quality issues and take appropriate solutions, it is very significant to classify power quality disturbances correctly. First this paper introduces the issue of power quality, analyzes the existing classification methods of power quality disturbances in depth. Then classifies power quality disturbances using wavelet transform and support vector machines. The work of this paper mainly includes three aspects:First, an improved wavelet denoising of threshold function method is proposed. For the shortcomings, not completely de-noising for hard threshold function and easy to lose useful information for soft threshold function, of traditional methods, a compromise approach is proposed. Then five kinds of mathematical models of transient power quality disturbances are set up. After adding WGN to them, denoises these disturbance signal using three methods. Simulation results show that this method has the better performance than traditional methods.Then, gets the feature vectors of transient power quality disturbances using wavelet. Through the analysis of feature vector extraction, signal power variations at each scale are extracted as feature vectors. The advantage of this method is the less number of feature vectors, calculated convenient and high classification accuracy. By analyzing the impact of mother wavelet and decomposition scale to the feature vectors, uses the Db4 wavelet for 10-scale decomposition. The simulation results show that under these conditions, the feature vectors work on better classification results.Last, an improved multi-class support vector machine classifier method is proposed. For shortcomings of general multi-class support vector machine classifier, an improved method is proposed. This method not only inherits the advantages of fast and accurate of original method, but also achieves a classification of complex power quality transient disturbance. The simulation results show that the method has the better performance than General artificial neural network classifier and the general multi-class support vector machine classifier method.In this paper, the above mentioned method of power quality transient disturbances classification make improvements respectively in both denoising and Multi-class support vector machine for the limitations of previous methods. Simulation results verify the feasibility and effectiveness of the method in power quality transient disturbance classification.