Source - InterFlex Simris Microgrid (2018)

IEEE journal paper (2018, doi: 10.1109/MELE.2018.2871295) — “Microgrid Field Trials in Sweden – The village of SIMRIS.” Authors: Henning Wilms, Dominik Mildt, Sebastian Schwarz, Marco Cupelli, Antonello Monti (E.ON ERC / RWTH Aachen University) + Peder Kjellen, Thomas Fischer, Demijan Panic, Michael Hirst, Eugenio Scionti, Paul Kessler, Luis Hernández (E.ON). Funded by EU H2020 InterFlex project (grant agreement no. 731289).

Document

• File: raw/Rapporten-INTERFELX-H2020.pdf
• Pages: 22
• Publication: IEEE Electrification Magazine 2018
• Project: InterFlex H2020 (grant 731289); RWTH Aachen E.ON Energy Research Center

The Simris project

E.ON decided in 2015 to build a grid-connected Local Energy System (LES) in the village of Simris in southern Sweden — a pilot to demonstrate that a distribution grid segment can host 100% renewable supply while being able to seamlessly disconnect from and reconnect to the main grid.

Grid topology: Existing 10 kV distribution system with seven 10/0.4 kV substations, five of which supply 150 customers. Feed-in point at a 20/10 kV substation from a 10 kV bay.

Asset inventory (Table 1)

LES taxonomy

The paper defines four LES types (grid-connected private/public; off-grid private/public). Simris is a grid-connected public LES operated directly by E.ON as a service. This taxonomy is relevant for distinguishing energy communities and microgrids under EU regulation.

EMS operating modes

Virtual Island mode: BESS controls active/reactive power set points so that no exchange occurs over the Point of Common Coupling (PCC). The PCS operates in Current-source-inverter (CSI) mode as a grid-following unit. Net exchange can be set to any value — enabling LES provision of ancillary services to the overlying grid.

Intentional Island mode: PCS runs in Voltage-source-inverter (VSI) mode as the grid-forming unit. EMS measures real/reactive power at PCC, sets voltage+frequency targets until exchange is near zero, then triggers the circuit breaker to open. EMS holds 10.7 kV / 50 Hz while islanded along BESS droop curves. To reconnect, EMS synchronises voltage+frequency before closing the breaker.

Islanding test results (12 April 2018)

System operated in island mode from 8:00 a.m. to 8:00 p.m. (12 hours). During this period:

• Frequency: lower variation than the Nordic synchronous grid (tighter band than grid-connected)
• Voltage: more stable than grid-connected reference
• THD (Total Harmonic Distortion): lower than grid-connected reference

Conclusion: when islanded, the LES achieved better power quality than when grid-connected. This is a counter-intuitive result arising from the EMS’s tight control of BESS droop curves.

Short-circuit testing also performed (3-phase fault, 63 A fuses): BESS terminal voltage dropped to 86% during fault; nominal voltage restored within 200 milliseconds.

Key cost finding: BESS vs grid upgrade

BESS + Power Conversion System (PCS) was found to be up to 4× less costly than a conventional grid upgrade (conductor/substation) for mitigating voltage deviations caused by newly connected RES plants in a rural area. The Simris analysis was motivated by a planning case where two new PV farms on a 24 km MV line would have required a grid upgrade costing 3× the total PV investment value — a cost that would have been distributed across all customers in the area under existing regulation.

This is one of the strongest quantified DSO value cases for distributed storage as an alternative to copper investment, predating and complementing later Swedish evidence (FlexAbility 2025, Ellevio case studies).

DSR platform: heat pump flexibility

Heat pumps are the primary DSR asset. Key operational findings:

Flexibility depends on outdoor temperature (Table 2):

DSR signal: percentage-based (−100% to +100% of maximum power), derived from main battery SoC. Comfort boundary: ±1 °C from customer-set temperature range — if breached, DSR is automatically disabled.

Hot water boilers: only controlled during excess energy generation (no reduction of boiler heating, to avoid comfort disruption). Pre-heating optimized based on consumption pattern monitoring.

EV charging (EVSE): throttled in 3 steps based on main battery SoC — no charging at SoC <20%; 33% (10 A) at 20–50%; 66% (21 A) at 50–80%; 100% (32 A) at >80%.

Residential batteries: set point sent every 5 minutes based on central battery SoC; 30–70% SoC deadband — within this range, residential batteries operate freely per customer preference.

RWTH Aachen University: MPC and forecasting

RWTH Aachen developed an optimization-based Energy Management System using Model Predictive Control (MPC) with a rolling-horizon approach (1-hour timestep; re-optimized at each step using updated state information).

Three operational objectives were compared:

• Case A (Economic): minimize operational costs + line losses
• Case B (Local use): minimize energy exchange with main grid
• Case C (PIT): maximize Potential Islanding Time — prioritize BESS state of charge for islanding readiness

Finding: PIT maximization (Case C) achieves similar energy exchange reduction as pure cost optimization, but at significantly higher line losses — because the BESS charges and discharges rapidly to maintain high SoC readiness, generating excess current flow. Islanding time optimization thus conflicts with efficiency.

Forecasting: LSTM encoder-decoder RNN for 24h ahead. Inputs: historical load/generation time series + meteorological variables (temperature, pressure, irradiation, wind speed/direction, cloud cover) + calendar variables (weekday, holiday, month). Architecture outperforms standard encoder-decoder RNN baseline. Combined MPC + LSTM approach → >10% improvement in self-consumption vs no control (simulation results on Simris synthetic load data).

Project context and funding

InterFlex was an EU H2020 project (2016–2019, grant 731289) exploring Local Energy Systems and their role in enabling community-level renewable integration. The Simris field trial was one of several European demonstration sites. Key partners: E.ON (lead operator), E.ON ERC at RWTH Aachen University (research/simulation), Encorp (EMS hardware), Loccioni (PCS), Holtab (substations).

Relevance to wiki topics