Application of Pattern Recognition Method in Classifying Power System Transient Disturbance

Power system transient can cause serious damage to main power system apparatus and sensitive loads. There are many causes of power system transient including capacitor bank switching, switching of large inductive loads and lightning. This paper discusses the application of pattern recognition method, namely Support Vector Machine (SVM) to classify the cause of transient disturbance in power system. Two types of feature extractions are applied to provide the inputs to the SVM, i.e. the minimum and maximum peak voltage values and the wavelet energy level of the transients. The IEEE 30 bus system is modeled using the Power System Computer Aided Design (PSCAD) software to generate different type of transient data caused by capacitor switching and lightning. Feature extraction is performed using discrete wavelet transform (DWT) analysis. The results showed that the performance of the feature extraction using maximum and minimum peak voltage values is superior (80%) as compared to the wavelet energy (54%) to classify the cause of the transient. Key-Words: Power quality, transient, discrete wavelet transform, support vector machine, radial basis function