A New Hybrid Islanding Detection Scheme Using Discrete Wavelet Transform and Artificial Neural Networks

Abstract

In today’s world, sources of renewable energy (RES) using PV arrays are the most extensively used. When RES is connected to the grid, there will be some issues owing to the unexpected circuit breakers connected to the grid trip, which creates islanding. This islanding condition should be detected within two seconds, as per IEEE standards. This paper presents frequency disturbance triggered hybrid islanding detection using artificial neural network (ANN) and discrete wavelet transformation (DWT). The ANN model is trained by these WT features and this approach calculates the detection time for various loading and non-islanding conditions. DWT analysis is performed up to level 4 in this work which is fed to an ANN model to predict islanding detection time. Simulation of frequency triggered hybrid islanding detection approach is implemented on the MATLAB 2018b platform. Python 3.9.5 is used for the discrete wavelet transform and ANN.