A synthetic building operation dataset

Area of Science:

• Building Science
• Energy Systems
• Data Science

Background:

• Accurate building operation data is essential for energy efficiency research.
• Existing datasets may lack detail or realism in simulating diverse operational conditions.
• Synthetic data generation offers a scalable solution to data scarcity.

Purpose of the Study:

• To present a large-scale, high-resolution synthetic building operation dataset.
• To provide data for benchmarking, calibration, and algorithm development in building energy analysis.
• To enable realistic simulations considering occupant behavior and environmental factors.

Main Methods:

• 1395 annual simulations of a U.S. DOE reference office building.
• Inclusion of HVAC, lighting, and miscellaneous electric loads (MELs) data at 10-minute intervals.
• Utilized 30 years of historical weather data across three climates and incorporated realistic occupant behavior assumptions.

Main Results:

• A 1.2 TB HDF5 dataset containing detailed building operation parameters.
• Data covers energy consumption, environmental conditions, and system operations.
• Dataset includes metadata structured using the BRICK schema for interoperability.

Conclusions:

• The synthetic dataset offers a valuable resource for advancing building energy research.
• It supports applications such as energy model calibration, load prediction, and control strategy development.
• The dataset facilitates the evaluation of occupant and weather impacts on building performance.