Deep Learning Based Optimal Scheduling of Battery Energy Storage Systems Using Real-Time Market Signals and AI Data Center Load Forecasts
Energy, Environment, and Economy · 01.01.2024 · 2 (1) · 1-8
DOI: 10.25163/energy.2110482 ↗Kurz gefasst
Fachartikel in der Zeitschrift Energy, Environment, and Economy, online erschienen am 01.01.2024.
Abstract (Originalsprache)
Background
The integration of renewable energy and the growing demand for reliable data center operations have increased the need for efficient energy management. Battery Energy Storage Systems (BESS) function as adaptable systems which handle supply and demand variations to decrease expenses and boost operational stability.
Methods
Data collection involved structured questionnaires to measure demographic information and operational variables which included energy price awareness and renewable energy. A Bidirectional Long Short-Term Memory (Bi-LSTM) model was developed to optimize BESS charge-discharge schedules and forecast load. Model performance was evaluated using RMSE, MAPE, forecast accuracy, and energy cost reduction. Operational and behavioral variables were analyzed for their relationships through Correlation (r) and Chi-square statistical methods.
Results
The Bi-LSTM system reached 96.7% accuracy in forecasting while achieving a 25% decrease in energy expenses and producing results with an RMSE of 0.038 and MAPE of 2.85% which exceeded traditional model outcomes. The study discovered a strong positive relationship between operational awareness and behavioral readiness because the correlation coefficient reached 0.68 with statistical significance at p=0.01. The Chi-square test revealed three significant relationships between energy price awareness and AI tool usage (X² = 12.45, p = 0.01) and renewable energy integration with automated load control (X² = 9.82, p = 0.03) and grid price sensitivity with cost reduction (X² = 7.36, p = 0.05).
Conclusion
The study shows that AI-based BESS optimization systems which work together with human operators create improved energy efficiency and system reliability and lower operational costs.
Zitieren
APA
(2024). Deep Learning Based Optimal Scheduling of Battery Energy Storage Systems Using Real-Time Market Signals and AI Data Center Load Forecasts. Energy, Environment, and Economy, 2(1), 1-8. https://doi.org/10.25163/energy.2110482
Vancouver
Deep Learning Based Optimal Scheduling of Battery Energy Storage Systems Using Real-Time Market Signals and AI Data Center Load Forecasts. Energy, Environment, and Economy. 2024;2(1):1-8. doi:10.25163/energy.2110482
BibTeX
@article{euvimed2024,
title = {Deep Learning Based Optimal Scheduling of Battery Energy Storage Systems Using Real-Time Market Signals and AI Data Center Load Forecasts},
journal = {Energy, Environment, and Economy},
year = {2024},
volume = {2},
number = {1},
pages = {1--8},
doi = {10.25163/energy.2110482}
}
EUVIMED