LSTM Neural Network for Current Prediction and Harmonic Analysis in Solar Power Systems

Abstract

Distributed energy resources (DERs) are increasingly integrated with the power grid, affecting its operation, stability, and reliability. This decentralized generation influences the performance of traditional power plants and dispatch centres. In this study, we use a multi-input Long Short-Term Memory (LSTM) neural network to monitor power quality and process data. MATLAB software is employed to predict waveforms, and the results demonstrate the LSTM model's ability to accurately predict current and harmonic variations in solar power generation across varying weather conditions. © The Institution of Engineering & Technology 2024.