Advanced Control Technique for Optimal Power Management of a Prosumer-Centric Residential Microgrid

作     者：Gbadega, Peter Anuoluwapo Sun, Yanxia Balogun, Olufunke Abolaji 

作者机构：Univ Johannesburg Dept Elect & Elect Engn Sci Auckland Pk Kingsway Campus Johannesburg 2006 South Africa 

出 版 物：《IEEE ACCESS》 (IEEE Access)

年 卷 期：2024年第12卷

页      面：163819-163855页

核心收录：

基　　金：South African National Research Foundation [PSTD2204285206  141951  137951  AJCR230704126719120106] 

主　　题：Microgrids Renewable energy sources Energy storage Optimization Power system management Power system stability Photovoltaic systems Distributed power generation Reliability Vehicle-to-grid Distributed energy resources (DERs) home energy management system (HEMS) residential microgrid prosumers teaching-learning-based optimization (TLBO) 

摘      要：Advanced control approaches are needed for prosumer-centric residential microgrids to manage power optimally due to the shift to sustainable energy systems In residential contexts where prosumers are important, this study investigates novel approaches to improve renewable energy sources efficiency, dependability, and integration. Prosumers who produce and consume energy introduce unique challenges and opportunities for managing power flows and maintaining grid stability. This study proposes a smart Home Energy Management System (HEMS) designed to reduce daily energy costs for prosumers while offering ancillary services to the grid. The prosumer s residence includes an electric vehicle (EV), a home battery pack, a photovoltaic (PV) system, a mix of flexible and critical electrical loads, and heating and cooling loads. The HEMS can operate independently by disconnecting from the main grid. Additionally, it supports grid frequency management by considering the operational flexibility of the controlled power sources and loads. To solve the multi-objective power management optimization problem, the Teaching-Learning-Based Optimization (TLBO) algorithm was implemented. This algorithm optimizes day-ahead power scheduling by considering temperature forecasts, energy prices, solar irradiance, system constraints, and user preferences. A residential load forecasting system, which combines k-means clustering and artificial neural networks (ANNs), was developed and implemented to predict residential loads efficiently. Various simulated scenarios illustrate how the HEMS may save costs while guaranteeing that all power component operating limitations are reached, and user comfort and satisfaction are maintained. To demonstrate the capabilities and cost-effectiveness of the suggested HEMS, several scenarios are simulated and the outcomes of the day-ahead optimization carried out by the HEMS for the prosumer residence are examined. Furthermore, it is verified that the system can su