Hybrid GA-SVR technique for contingency screening in power system

Contingency analysis is essential to operate power system in secure state. Normally Contingency analysis is performed offline since it is a time consuming task. Due to continuously changing nature of power system it is imperative to have a fast contingency analysis method. This work proposed a very efficient, fast and accurate method for online contingency screening using the hybrid Genetic Algorithm (GA) and Support Vector Regression (SVR), hence named GA-SVR. SVR is used to build a model which takes a network topology and load profile as input and maps them into contingency performance indices. It is desirable to find the optimal value of SVR parameters to make it fast and accurate. Therefore GA is applied for searching global optimal parameters for SVR model. Standard IEEE-30 bus system is tested to illustrate the effectiveness of presented GA-SVR technique.