Categorizing Power System Stability Using Clustering Based Support Vector Machines
B. Dora Arul Selvi
The current deregulation trend and the participation of many players are contributing to the decrease in security margin. This seeks the development of reliable and faster security monitoring methods. Support Vector Machines, a Neural Network Technology has been as presented an important contributor for reaching the goals of online Transient stability assessment. The training complexity of SVM is highly dependent on the size of data set. Since the power systems are of high dimensionality, feature extraction techniques must be implemented to make the application feasible. This study presents a new Clustering Based SVM to identify the stability status of power system. Here we have applied an exclusive clustering algorithm and an overlapping clustering algorithm, which scan the entire data set only once to provide SVM with high quality samples that carry the statistical summaries of the data such that the summaries maximize the benefit of learning the SVM. Transient stability of New England 39 Bus system is assessed by SVM trained with complete input feature set. The aspects of training time and classification accuracy are compared to the results obtained from CB-SVM.This shows that CB-SVM is highly useful for very large data sets while also generating high degree of classification accuracy.