Classification of static security status using multi-class support vector machines

Abstract EN

This paper presents a Multi-class Support Vector Machine (SVM) based Pattern Recognition (PR) approach for static security assessment in power systems.

The multi-class SVM classifier design is based on the calculation of a numeric index called the static security index.

The proposed multi-class SVM based pattern recognition approach is tested on IEEE 57 Bus, 118 Bus and 300 Bus benchmark systems.

The simulation results of the SVM classifier are compared to a Multilayer Perceptron (MLP) network and the Method of Least Squares (MLS).

The SVM classifier was found to give high classification accuracy and a smaller misclassification rate compared to the other classifier techniques.