High Impedance Fault Detection In Radial Feeder Under Different Micro Climateic Conditions Using Optimised Support Vector Machine Algorithms

Abstract

Detection of high impedance faults (HIF) in radial feeders is a challenging task. HIF faults occurs mostly due to single line to ground faults in rural and semi urban areas in Andhra Pradesh, India of about 80%. Existing methods detect the single line to ground faults (SLGF) and not classifies the subtypes of SLGF faults such as small-impedance faults (SIFs), medium-impedance faults (MIFs), high-impedance faults (HIFs), arc grounding faults (AGFs), intermittent grounding faults (IGFs), and transient grounding faults (TGFs). SLGF subtype classification helps for selective tripping, Minimizes Outages, Limits Prolonged Fault Operation, Prevents Cascading Failures and improves the system monitoring. In this paper, optimized Support Vector Machines (SVM) algorithms such as (i)cubic-SVM, (ii) PSO-SVM and (iii) BO-SVM are proposed for the detection and classification of the HIF under different environmental i.e. microclimatic conditions. From the results, BO-SVM provides the better accuracy of 97% in HIF detection and subtype classification for 4 - node and 13 -node IEEE radial test feeders. The proposed algorithms are compared with the traditional algorithms and evaluated the performance analysis.