Nikolaos Koltsaklis

Smart Home Energy Management Systems can manifest energy consumption reduction targets in the residential sector and can be viewed as an approach to transform the consumer into an active prosumer. The present paper presents a smart home energy management system that includes flexible appliances, electric vehicles, and energy storage units. Efficient forecasting algorithms support the robust operation of the smart home energy management system. Specifically, the smart home energy management system receives as inputs forecasts of demand, renewable energy sources including photovoltaics and Wind Turbine generations, and real-time prices. In order to minimize energy costs, a variety of algorithms is compared to provide highly accurate forecasts. (C) 2022 The Author(s). Published by Elsevier Ltd.

Smart homes that contain renewable energy sources, storage systems, and controllable loads will be key components of the future smart grid. In this article, we develop a reinforcement-learning (RL)-based scheme for the real-time energy management of a smart home that contains a photovoltaic system, a storage device, and a heating, ventilation, and air conditioning (HVAC) system. The objective of the proposed scheme is to minimize the smart home's electricity cost and the residents' thermal discomfort by appropriately scheduling the storage device and the HVAC system on a daily basis. The problem is formulated as a Markov decision process, which is solved using the deep deterministic policy gradient (DDPG) algorithm. The main contribution of our study compared to the existing literature on RL-based energy management is the development of a clustering process that partitions the training data set into more homogeneous training subsets. Different DDPG agents are trained based on the data included in the derived subsets, while in real time, the test days are assigned to the appropriate agent, which is able to achieve more efficient energy schedules when compared to a single DDPG agent that is trained based on a unified training data set.

This work presents an optimization framework for the optimal scheduling of a multi-energy microgrid based on Mixed-Integer Programming techniques and consisting of a number of aggregated end-users. With the objective of satisfying the microgrid's electricity and heat demands in a cost-optimal way, the microgrid includes several technologies such as micro combined heat and power units, gas turbines, heat pumps, renewable energy sources, exchanges with the main grid, as well as flexibility providers such as energy storage systems, electric vehicles, and demand response. One salient feature of the proposed framework is that it considers the contribution of power generating sources in the ancillary services provision, both up and down, providing an additional income for their operation and enhancing the grid operation. An illustrative case study has been used to test the applicability of the proposed approach in both economic and operational terms. The results underscore the significance of including the ancillary services market as a revenues source to the MES as well as the fact that the participation of various resources in both energy and ancillary services markets affects the operational scheduling of the microgrid, and the services provided by the flexibility providers play a major role in the overall cost reduction. System operators, aggregators, and market participants can utilize the proposed optimization framework to determine their operational and investment strategies for optimal resource utilization and portfolio selection. Copyright © 2021, AIDIC Servizi S.r.l.