Volatility-Aware Sizing and Seasonally Adaptive Control of Hybrid Energy Storage Using Multi-Horizon Forecasting in a Nordic Campus Microgrid
Lopullinen julkaistu versio - 7.01 MB
Shikdar, T. A. & Laaksonen, H. (2026). Volatility-Aware Sizing and Seasonally Adaptive Control of Hybrid Energy Storage Using Multi-Horizon Forecasting in a Nordic Campus Microgrid. IEEE Access, 14, [78757-78781]. https://doi.org/10.1109/ACCESS.2026.3695607
© 2026 The Authors. This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/
Lataukset17
Pysyvä osoite
Kuvaus
The electrification of heating, transportation, and buildings is transforming modern campuses into multi-energy environments where photovoltaic (PV) generation, electric-vehicle (EV) fast charging, and heat-pump (HP) loads interact across multiple timescales. These interactions introduce high volatility that conventional battery energy storage systems (BESS) and short-horizon forecasting techniques cannot manage effectively, particularly in Nordic climates with extreme seasonal irradiance and heating-dominated load patterns. This study develops a unified analytical–AI–optimization framework for sizing and controlling a hybrid battery energy storage system (HBSS) that integrates a fast-response supercapacitor/lithium titanate (SC/LTO) layer with a Li-ion energy storage layer. Volatility is quantified using three indicators: the PV Volatility Index (PVI), EV Fast-Charging Impact Factor (EFI), and Thermal Load Flexibility Index (TLFI) derived from a 3.5-year, 15-minute dataset from a Nordic university campus microgrid. A multi-horizon deep-learning model (HERA-4C), combining convolutional blocks, BiLSTM layers, and attention mechanisms, produces 1-, 4-, 24-, and 96-step forecasts and is embedded into a predictive energy-management and multi-objective optimization scheme using NSGA-II. The five-objective formulation minimizes annualized cost, CO2 emissions, grid import, Li-ion degradation, and unmet load. Results indicate that the optimal HBSS configuration ( ≈ 260 kW SC/LTO and 200 kWh Li-ion) mitigates over 90% of PV–EV–HVAC volatility, reduces peak grid import by 72%, eliminates curtailment and unmet load, and achieves 15–22% annual cost reduction alongside a 14–18% decrease in CO2 emissions. Seasonal evaluation confirms enhanced winter reliability through adaptive state-of-charge management. The proposed framework provides a reproducible methodology for resilient HBSS design in operational campus microgrids and supports scalable smart-building applications under high-volatility conditions.
Emojulkaisu
ISBN
ISSN
2169-3536
Aihealue
Kausijulkaisu
IEEE access|14
OKM-julkaisutyyppi
A1 Alkuperäisartikkeli tieteellisessä aikakauslehdessä (vertaisarvioitu)