Evaluation of Energy Sharing on a Local Energy Market Through Comparison of Energy Management Techniques

Abstract

Local energy markets (LEMs) are a promising way to solve the challenges of the increasing extension of decentralized energy systems and to promote the further integration of renewable energy sources. LEMs enable costumers with distributed energy resources to trade and share their electrical energy with each other. In the existing literature, the research focus is mostly on the development and evaluation of specific elements of LEMs, such as bidding strategies or market designs. The paper contributes a comprehensive evaluation of a LEM and the quantification of its benefit regarding the market-based device operation. For the evaluation in terms of financial outcome and local energy exchange, a centralized and a decentralized operation optimization serve as upper and lower references. In centralized optimization, the system boundary comprises the entire neighbourhood. In decentralized optimization, each building is balanced separately. For the LEM, we introduce a distributed market design with the involvement of an auctioneer. We focus there on the implementation of learning bidding strategies and a double-sided auction with non-iterative market clearing rules. For all three energy management techniques, the operating schedules of the devices are determined using mixed-integer linear programming. In several case studies we investigate different neighbourhoods in order to evaluate the influence of different technologies and their penetrations as well as the impact of the building stock in terms of building type and construction year. We evaluate the market outcome with multiple key performance indicators (KPIs) such as the supply-and demand-cover-factor, the total operation costs and the peak load. The results show that total energy costs can be reduced by up to 6.4 %. For the energy exchange, it is shown that the electricity surplus is up to 72 % and the electricity demand of the QUartier decreases by up to 6.8 % compared to the decentralized optimization and increases by up to 14.3 % compared to the centralized optimization. Further, we noted up to 46.2 % higher peak loads.