Real-World Model Predictive Control for Home Battery Systems: Towards Closing the Simulation-to-Reality Gap
ACM SIGEnergy Energy Informatics Review · 01.09.2026 · 6 (3) · 130-137
DOI: 10.1145/3861130.3861140 ↗Kurz gefasst
Fachartikel in der Zeitschrift ACM SIGEnergy Energy Informatics Review, online erschienen am 11.11.2026. Verfasst von Lukas Moosbrugger und 4 weiteren Autor:innen. Die Arbeit nennt 17 Literaturangaben.
Abstract (Originalsprache)
The growing share of residential photovoltaic systems is leading to increasing periods of surplus solar generation in the grid and calls for prosumers to shift feed-in and maximize local energy self-consumption. This requires advanced battery management strategies, for which Model Predictive Control (MPC) has proven to be a reliable and effective solution. However, current research largely relies on simulation, with few demonstrated real-world MPC deployments, leaving the simulation-to-reality gap insufficiently addressed. Therefore, this study incrementally enhances simulation details and benchmarks against real-world performance. Beyond quantifying the simulation-to-reality gap, this work identifies its underlying causes and evaluates how incremental modeling refinements reduce it in real-world deployment. Starting from a simplified mixed-integer linear programming (MILP) formulation, we observe that this model significantly overestimates economic potential—by approximately 30% over the tested seven-day period. By contrast, a more detailed Python-based simulation incorporating additional system dynamics and constraints reduces this discrepancy, yielding an economic simulation-to-reality gap of 10%. Our analysis identifies forecast errors as the primary driver of performance loss and demonstrates how a two-stage control strategy can mitigate their impact. We further identify controller-induced oscillations around zero grid power as a previously underreported source of economic loss in deployed MPC systems. This effect reveals additional optimization potential. By combining real-world validation with systematic analysis, this work provides actionable insights and a reproducible framework for bridging the simulation-to-reality gap in MPC-based residential energy systems.
Zitieren
APA
Moosbrugger, L., Seiler, V., Wohlgenannt, P., Kepplinger, P., & Ristov, S. (2026). Real-World Model Predictive Control for Home Battery Systems: Towards Closing the Simulation-to-Reality Gap. ACM SIGEnergy Energy Informatics Review, 6(3), 130-137. https://doi.org/10.1145/3861130.3861140
Vancouver
Moosbrugger L, Seiler V, Wohlgenannt P, Kepplinger P, Ristov S. Real-World Model Predictive Control for Home Battery Systems: Towards Closing the Simulation-to-Reality Gap. ACM SIGEnergy Energy Informatics Review. 2026;6(3):130-137. doi:10.1145/3861130.3861140
BibTeX
@article{Moosbrugger2026,
author = {Moosbrugger, Lukas and Seiler, Valentin and Wohlgenannt, Philipp and Kepplinger, Peter and Ristov, Sashko},
title = {Real-World Model Predictive Control for Home Battery Systems: Towards Closing the Simulation-to-Reality Gap},
journal = {ACM SIGEnergy Energy Informatics Review},
year = {2026},
volume = {6},
number = {3},
pages = {130--137},
doi = {10.1145/3861130.3861140}
}
EUVIMED