NeuralBES: A Differentiable, Control-Aware Emulator for Scalable Building Energy Modeling

2026-10-07

AI Summary

This study addresses the tension between the poor scalability of physics-based simulations and the limited trustworthiness of data-driven approaches in building energy modeling by proposing a differentiable simulator that integrates physical priors with deep learning. The method employs a single encoder to map heterogeneous building metadata into resistance-capacitance (RC) thermal model parameters, solves linear recurrences via parallel scans in log space, and incorporates a predictor-corrector mechanism to close the loop on nonlinear thermostat control, thereby enabling full gradient flow and large-scale parallel training. Experiments on the ResStock dataset demonstrate that, while strictly preserving physical consistency, the proposed approach achieves accuracy comparable to the best baselines, reduces parameter count by an order of magnitude, and decreases MAPE by 50% relative to gray-box models of similar scale.

Abstract

Demand-side flexibility i.e. forecasting, shifting, and curtailing residential energy loads, depends on thermal models trusted across millions of heterogeneous buildings. Existing tools force a hard tradeoff: high-fidelity physics simulators such as EnergyPlus are accurate but sequential and require per-building calibration, while purely data-driven sequence models scale but abandon the physical structure that makes their predictions trustworthy. We introduce NeuralBES (Building Energy Simulation), a differentiable emulator that resolves this tradeoff by parameterizing a resistance--capacitance (RC) based thermal model with a shared neural encoder: static building metadata such as floor area, vintage, and HVAC type is mapped to physically bounded capacitances, conductances, and equipment coefficients, which become the coefficients of a scalar linear recurrence solved via a log-space parallel scan, and a predictor--corrector loop closes the thermostat--temperature nonlinearity while preserving full-horizon gradient flow. Trained on the ResStock dataset across three climate zones, NeuralBES handles heterogeneous building archetypes, vintages, and climate zones within a single trained encoder, while black-box baselines produce statistically plausible but physically inconsistent trajectories. On the annual full-year rollout, NeuralBES is the only data-conditioned model that is simultaneously physics-valid and accurate to within 4 MAPE points of the strongest raw-error baseline, while operating at roughly an order of magnitude fewer parameters than the transformer and recurrent baselines; among physics-valid baselines at parameter parity it more than halves the MAPE of the grey-box RC alternative.

Problem

• Building Energy Modeling
• Demand-side Flexibility
• Thermal Models
• Physics-informed Simulation
• Scalability

Innovation

• Differentiable Emulator
• Resistance-Capacitance Thermal Model
• Parallel Scan
• Predictor-Corrector Loop
• Building Energy Simulation