Contingency ranking in static security assessment using teaching learning based optimization enhanced support vector regression

Abstract

Static security assessment is one of the important aspects of power system. The traditional techniques used for the contingency ranking suffer from longer and complex calculations. It led to move towards pattern recognition approach for contingency ranking. In this paper, ranking module based on Support Vector Regression is designed to rank contingencies based on the severity. A new combined performance index is formulated to know the severity of contingency. The combined performance index is combination of voltage deviation performance index and line overload performance index. The parameters of SVR are selected using Teaching Learning Based Optimization. The proposed pattern recognition approach is validated on IEEE standard test systems.