Residential energy management with flexible and forecast uncertainties

This study proposes the hybrid topology of residential micro-grid (MG) energy management, taking into account prosumers, flexibile service opportunities, and projected uncertainties. The proposed hybrid technique is the combination of the Dilated Residual Convolutional Neural Network (DRCNN) and the Archerfish Hunting Optimizer (AHO) and is usually referred to as the DRCNN-AHO strategy. The main goal of the proposed strategy is to strengthen the framework for energy management, which works to effectively address difficulties brought on by regional weather fluctuations in the environment. Accurate forecasting is essential for future residential MGs. It uses a DRCNN-based forecaster to collect past utility price-based energy consumption data to estimate day-ahead pricing signals and energy use. The AHO is used to optimize the micro-grid's operating costs and grid energy consumption while satisfying the generation-demand balance and the accompanying limitations. The involvement of prosumers in the energy management system is crucial for optimizing energy consumption and integrating renewable sources. The proposed topology is implemented in MATLAB, and its performance is compared to existing approaches such as the Seagull Optimization Algorithm (SOA), Giza Pyramids Construction (GPC), and Color Harmony Algorithm (CHA). The proposed DRCNN-AHO method attains a higher profit (keuro) of 195 in case 1 and 240 in case 2. Also, the existing methods such as SOA, GPC, and CHA attain a higher profit of 170, 160, and 170, respectively. Additionally, the DRCNN-AHO approach achieves a larger profit of 195,000. The proposed strategy generates a better profit when compared to existing techniques.