Machine Learning based Detection and Classification of Power Quality Disturbances in Smart Grid System

Abstract

Identification and Classification of common/ naturally occurring Power quality disturbances has been a challenge for electrical engineers/researchers worldwide. Power Quality Disturbances can be caused by various factors, including natural phenomena (Lighting, Thunder, Heat), Heavy loads/resistors, Switching of Capacitor Banks, etc. The research paper proposes a practical machine-learning-based methodology to address the issue. The approach is based on mathematical modeling/simulation of the PQ Disturbances and training of the obtained dataset using suitable Machine Learning Models. The parameters of the Machine Learning Models are tuned using a brute-force approach; thus, the obtained model parameters claim to have the highest possible achievable accuracy. Noise in signals is common, and hence, the research tries to incorporate the noises. The proposed methodology's robustness is examined in the presence of a Gaussian noise-contaminated dataset. © 2022 IEEE.