Modelling of Harris Hawks optimization with deep learning-assisted microgrid energy management approach

Abstract

Microgrid (MG) is a potential decentralized energy distribution and generation technology that is resilient, reliable, and efficient. This small-scale power system is reliable and resilient since it connects to the grid or runs independently. Renewable energy is difficult to integrate into MG due to variable load and unreliable electricity. MG operation relies on an energy management system (EMS) to balance electricity demand and supply, reduce operational costs, and maximize renewable energy use. Intelligent control systems, optimization methods, and machine learning algorithms were used for MG EMS. The Harris Hawks optimization with deep learning-assisted microgrid energy management (HHODL-MGEM) technique is developed in this work. HHODL-MGEM comprises two main stages. In the first step, the HHODL-MGEM approach uses the Harris Hawks optimization or HHO algorithm to meet load power demands at a low cost while maintaining DC bus voltage and protecting the battery from overcharging and depletion. In the second step, long short-term memory (LSTM) networks can predict power costs. The HHODL-MGEM approach is evaluated using multiple methods. The experimental results showed that HHODL-MGEM outperforms other methods.