Long short-term memory-based forecasting of uncertain parameters in an islanded hybrid microgrid and its energy management using improved grey wolf optimization algorithm

Abstract

An islanded hybrid AC-DC microgrid interconnects renewable energy sources, distributed generators, and energy storage, primarily for remote areas without grid access. Its reliability depends on variable renewable output and load demand, while an energy management system optimizes power scheduling and reduces costs. In the first phase of this paper, uncertainty parameters like day-ahead power from renewable energy sources (RES) and load demand (LD) are forecasted using the long short-term memory (LSTM) deep learning algorithm. The LSTM outperforms the artificial neural network (ANN) model in terms of mean square error (MSE) and prediction accuracy (R²) for both training and testing datasets. In the second phase, the forecasted RES power and LD are used for optimal distributed generator (DG) scheduling using the improved grey wolf optimization (IGWO) algorithm. The objective of energy management in an islanded hybrid microgrid (HMG) is to minimize daily operating costs by considering load demand and the bidding costs of energy sources and storage devices. Two operational scenarios are evaluated to minimize the operating costs and optimize battery life. The proposed method, validated with IEEE standard test systems, is compared against several metaheuristic techniques. Results demonstrate that the improved grey wolf optimization (IGWO) algorithm is more effective at reducing costs and provides faster optimal solutions.