Classification of Fault Disturbances in a DG Integrated Hybrid Power System Using Support Vector Machine and Decision Tree

Abstract

AbstractTo fulfill the increasing load demand, distributed generations are becoming increasingly crucial in the electrical power system. The connection of a distributed generator (DG) to the grid raises a number of challenges related to power system structural protection and control. The consequence of dispersed generations on a grid is that the fault current level fluctuates, which complicates fault investigations. Even after years of in-depth research, categorization of fault remains one of the most important difficulties that arise in a distributed generation integrated power system. The fault disturbances are classified by the pattern recognition techniques like decision tree (DT) and support vector machines (SVM). The study is carried out both graphically as well as in terms of performance indices like standard deviation (STD) and entropy. The objective of this article is to use DT and SVM to categorize fault-related disturbances in a DG-infiltrated hybrid power system. To categorize the Line-Line fault and Line-Line-Ground fault, an unique technique based on SVM and DT is proposed. Based on these analyzes, it is observed that DT Gives the highest viable accuracy as compared to other methods, which proves its robustness in various working situations such as load variability, in the scheme’s parameters, there are solar insolation, noise, and harmonics.KeywordsDistributed generator DG)Support vector machine (SVM)Decision tree (DT)Point of common coupling (PCC)