A Risk-Averse Energy Management System for Optimal Heat and Power Scheduling in Local Energy Communities

Local energy communities (LECs) facilitate energy distribution, supply, consumption, storage, and trading for the communities and their members. This paper proposes a risk-averse energy management system (EMS) for optimal heat and power scheduling in LECs. Three approaches namely high accuracy forecast models, advanced optimization models, and providing flexibility sources are followed to handle uncertainties of photovoltaic power and load. To this end, the load demand and photovoltaic power as uncertain variables are predicted using machine learning methods and the problem is modeled under uncertainties by information-gap decision theory (IGDT). This method doesn't require probability distribution functions of uncertain variables which makes it valuable in cases with high levels of uncertainties or lack of sufficient historical data. The advantage of flexibility in increasing robustness is studied by adjusting desired indoor and hot water temperatures. The effectiveness and efficiency of the proposed model are evaluated on the LEC at Chalmers University of Technology campus, Gothenburg, Sweden.