Benchmark RadarRESCAST-100K
RESCAST-100K evaluates cross-domain residential load and indoor temperature forecasting. It provides ~100,000 EnergyPlus-simulated U.S. homes with 15-minute time series for total load, HVAC load, and indoor temperature, plus weather, setpoints, and static covariates. It includes configurable domain axes and integrates five real-world datasets for sim-to-real evaluation.
- Released
- 2026-06-01
- Readiness
- Paper only
- Primary field
- General AI
Why it matters
Existing residential forecasting datasets are narrow and lack structured cross-domain evaluation. RESCAST-100K enables systematic assessment of transfer learning and domain adaptation under controlled shifts, supporting better generalization in home energy management and grid-scale applications.
Motivation
Accurate short-term forecasting of residential energy load and indoor temperature is essential for home energy management systems, grid-level demand response, and community energy efficiency efforts.
Primary resources
Benchmark Radar records only publicly supported details and links back to primary sources for verification.