Real-Time Optimal Energy Management for Microgrids Under Uncertainty Using Deep Neural Networks

The adoption of electric vehicles (EVs) and renewable energy sources in the power grid has substantially accelerated in recent years. In this paper, a novel energy management system (EMS) using the deep neural network (DNN) is presented for a grid-connected microgrid. Additionally, we put forth the particle swarm optimization (PSO) algorithm to fine-tune the number of hidden layer nodes in the DNN. The overarching goal is to minimize the total cost of energy consumption and charging capacity while evaluating EV discharge and load demand and suggesting energy storage charging requirements for buildings in real time. Furthermore, we conducted simulations over 24 hours using MATLAB/Simulink, and the outcomes substantiate the real-time adaptability of the proposed algorithm. In addition, experimental studies conducted on a laboratory-based HMG demonstrate the practical viability of the proposed Home Energy Management System (HEMS). The experimental results reveal a 9.4% and 8.8% improvement in the detection rate compared to conventional DNN methods.