Prediction of loadability margin of power system using Support Vector Machine

This paper presents the applications of Support Vector Machine (SVM) for estimating the Loadability Margin of a power system. Conventional methods of performing continuous load flow and stability studies are highly time consuming and infeasible for on-line application. It necessitates a requirement of a on line tool to calculate the loadability margin under normal condition as well as under contingency cases. SVM is a powerful and promising data classification and function estimation tool. In SVM the input vector in the form of real and reactive power load and the target or output vector is in the form of lambda (loading margin) are considered. To minimize the training time and improve accuracy of the SVM, the Kernel type and its parameter are selected cautiously. Two test systems were considered for simulation studies namely sample-6 bus and IEEE 30 bus systems. To verify the effectiveness of the proposed SVM method, the performance is compared with the Continuation Power Flow (CPF). Studies show that the SVM gives faster and more accurate results compared with CPF.