AI-augmented geothermal model for scalable energy uncertainties in buildings
A. Markowitz, R. Abuaamoud, S. Ben Ayed, A. Rupe, R. Yang, T. Kliphuis, V.V. Vesselinov
Scientific Reports2026DOI 10.1038/s41598-026-40837-4
Summary
This work presents a scalable software tool for predicting geothermal energy use in residential buildings under conditions that demand fast, physically informed simulations. Using a reduced-order model benchmarked against EnergyPlus and datasets generated with Latin Hypercube Sampling, Saltelli, and eFast methods, the study applies XGBoost to predict energy usage with high accuracy while preserving computational efficiency.