A Statistical Study On Fault Arc Based On Wavelet Analysis And SVM

As the fault arc phenomenon is difficult to be detected and harmful,Researchers in many fields have studied it.As a common method in the field of fault detection,wavelet analysis still has some problems in fault arc identification problem,it can not complete the identification of the fault arc alone.As a new statistical learning method,Support vector machine,has great advantages over traditional algorithms,but it has not been fully applied in the problem of fault arc identification.Therefore,this paper combined wavelet analysis with support vector machine to study the problem of fault arc.This is very important for the application of support vector machine and the construction of faulty arc prevention system.In order to understand and study the fault arc,this paper is based on the standard arc data.The problem of fault arc is described by using statistical language,and complicated features are transformed into simple statistical indicators.Based on these statistical indicators,descriptive statistical analysis of fault arc characteristics is carried out.Secondly,this paper analyzes the fault arc current signal by wavelet analysis,and excavates the characteristics of fault arc from different angles.On the basis of the above two kinds of analysis,this paper applies the support vector machine algorithm and the wavelet analysis method to the fault arc identification problem for the first time,and constructs a new fault arc identification model through a large number of experiments to select the kernel function and parameters.The results show that compared with the traditional fault arc identification method,the model constructed in this paper has more rigorous theoretical basis and more extensive applicability.In the further promotion and improvement of statistical learning theory in the field of fault arc identification applications,an efficient,innovative and reliable fault arc identification model is presented.