Development of a new bus zone identification algorithm using support vector machine

Abstract

This study presents a new support vector machine (SVM)-based fault zone identification scheme, which correctly identifies faults occurring inside and outside the protection zone of a busbar. A 220 kV substation with double busbar scheme present in an existing Indian system is considered for validation of the proposed fault zone identification algorithm for busbar protection. Fault simulations are performed using PSCAD/EMTDC software package and the outcome is then interfaced with MATLAB, where fault zone identification algorithm is implemented. A total of 44 650 internal and external fault cases with varying fault resistances, fault inception angles, fault locations, types of faults and source impedances have been used for validation of the proposed scheme. Even though the proposed scheme utilises only 29.63% of the total simulation cases (63 450) for the training of SVMs, it gives an overall fault zone identification accuracy of more than 90% when tested with using remaining 70.37% simulation cases for validation. The proposed scheme successfully discriminates between in-zone and out-of-zone faults with high fault zone identification accuracy for different fault and system conditions. Moreover, the proposed scheme remains stable during cross-country faults and severe current transformer (CT) saturation condition.