CIRED Workshop - Ljubljana, 7-8 June 2018 Paper 0476 Paper No 0476 Page 1 / 4 INTERFLEX – SIMRIS – TECHNICAL MANAGEMENT OF A GRID-CONNECTED MICROGRID THAT CAN RUN IN AN ISLANDED MODE WITH 100% RENEWABLE GENERATION Milica BOGDANOVIC Henning WILMS Marco CUPELLI RWTH Aachen – Germany RWTH Aachen – Germany RWTH Aachen - Germany mbogdanovic@eonerc.rwth-aachen.de hwilms@eonerc.rwth-aachen.de mcupelli@eonerc.rwth-aachen.de Michael HIRST Luis ARTURO HERNANDEZ SALMERON Antonello MONTI E.ON – UK E.ON – Germany RWTH Aachen - Germany mike.hirst@eon.com luis.hernandez@eon.com amonti@eonerc.rwth-aachen.de ABSTRACT This paper demonstrates a model predictive control (MPC) approach for microgrids with a significant proportion of renewable energy sources (RES). Forecasts are derived using recurrent neural networks and are included within the MPC. The demonstration is based on E.ON’s microgrid site in Simris, Sweden. E.ON plans to deploy a microgrid that can run with 100% renewable generation. Distributed household PV and battery systems are exploited for the required demand flexibility. A central battery system is used as grid forming unit to balance the microgrid. Simulation results show the effectiveness of the proposed approach. INTRODUCTION To realize the potential of renewable energy sources as distributed generation we must take a system approach which views generation and associated loads as subsystems or “microgrids” [1]. A microgrid is part of the distribution grid, which comprises generation assets, storage devices, and controllable loads, operating as a single controllable system in grid connected mode or isolated from the utility grid in islanded mode. Using residential flexibility for grid services has already been demonstrated in various projects [2]. While being islanded, the microgrid has to cope with relatively high power fluctuations on both the production (wind and PV) and consumption sides. Microgrids should be capable of managing and coordinating their own distributed generation, storages, and loads without a centralized coordination and management from the main grid. Therefore, one of the main ideas behind the microgrid concept is to decentralize the operation of electrical grids [3]. The focus of this paper is on the optimal operation planning of a microgrid. Our aim is to minimize the energy exchange between the microgrid and the main grid while satisfying complex operational constraints, such as the energy balance. Our research proves that the algorithm implemented and the test site used provide meaningful results about the feasibility of the chosen approach. That results support future field-testing at the test site. DESCRIPTION OF TEST SITE E.ON has built up and put into operation a microgrid in the South of Sweden in 2017 in Simris. The microgrid consists of a wind turbine as main generation, supported by a ground-mounted PV array and a battery system. Further, a back-up generator is used for test periods in case there is not enough power available from the microgrid’s RES when running in islanded mode The microgrid comprises about 130 E.ON customers of which some have household PV and batteries installed. The test site includes roughly 5 km of 10 kV cables, which feed eight secondary substations as well as about 11 km of low voltages cables connecting the customers. In island operation, the system is purely fed by devices that are grid-connected by power electronics and without rotating masses. Fig. 1 Schematic of the field test microgrid in Simris An energy storage system in virtual island mode measures and controls the active and reactive power flows within the microgrid. The central battery system is in charge of the almost instantaneous balancing of the micro grid. In islanded mode voltage and frequency control are provided by the main battery system. The main battery takes on the function of a grid-forming unit during islanded mode. The total on-site renewable generation is about 1MW.