🤖 AI Summary
Existing building thermal datasets lack sufficient diversity in building types and operational conditions, limiting advances in thermal dynamic modeling, energy efficiency control, and fault diagnosis. This work addresses this gap by constructing and releasing the ThermBuild dataset, which uniquely integrates 15 months of high-resolution (15-minute) measured data from two real residential buildings with three years of TRNSYS-simulated data from 958 residential units. The dataset encompasses diverse heat pump systems, building characteristics, and climatic conditions, and includes multidimensional variables such as heat pump operation, indoor environmental parameters, and weather data. ThermBuild enables research in cross-domain transfer learning, simulation-to-reality generalization, and model benchmarking, significantly enhancing the applicability, robustness, and reproducibility of data-driven approaches in building energy systems.
📝 Abstract
This paper presents the ThermBuild dataset, which comprises real-world measurements from two single-family homes and simulations of 958 TRNSYS building models. The buildings cover diverse combinations of air-source heat pump systems, numbers of thermal zones, occupancy profiles, building ages, thermal masses, sizes, orientations, window glazings, five European climates, and ventilation configurations. The dataset contains 15-minute-resolution operational data spanning 15 months for the real-world buildings and 3 years for the simulated buildings. Each building time series includes detailed measurements of heat pump operation, the heating distribution system, the domestic hot water system, weather conditions, and zone-level indoor climate variables. The ThermBuild dataset is designed for data-driven thermal dynamics modeling, thereby supporting the deployment of energy-efficient control, as well as fault detection and diagnosis in buildings. It is particularly suited for transfer learning, generalization modeling, benchmarking, simulation-to-reality transfer, and reproducible thermal modeling research.