ThermoLLM: Thermodynamics-Aware HVAC Control with Spatial-Semantic Knowledge Graph
2026-06-22 • Artificial Intelligence
Artificial IntelligenceMachine Learning
AI summaryⓘ
The authors explore controlling heating and cooling in a building with multiple rooms by using a smart system that knows how heat moves between connected spaces. They create a method that uses a physics-based map of the building combined with recent temperature and control data to help a language model decide how to set temperatures. This approach helps the system make better heating and cooling choices than other methods, balancing comfort and energy use more effectively. Their tests show it works well in managing both energy efficiency and indoor comfort.
multi-zone HVACthermal dynamicsspatial knowledge graphEnergyPlus simulationLLM controlBrick schemathermal comfortpredictive controlPMV violationenergy efficiency
Authors
Kirtan Bhatt, Xiachong Lin, Matthew Amos, Flora D. Salim, Wen Hu
Abstract
Multi-zone HVAC control is a spatial decision problem in which indoor thermal evolution and control decisions depend not only on outdoor conditions and internal heat gains but also on zone layout, physical adjacency, and delayed thermal interactions across the building. Recent LLM-based HVAC controllers have shown that prompt-based control is feasible. However, these methods typically rely on task descriptions, observation values, short textual feedback, or unstructured retrieval, which limits their ability to reason about zone coupling, thermal response, and building dynamics. This paper presents a thermodynamics-aware LLM control framework for a five-zone EnergyPlus building simulation. The controller is grounded in a physics-informed spatial knowledge graph derived from Brick-style building semantics and linked with recent interaction history. At each control step, the model receives the current building state, graph-structured spatial context, and recent environment-controller history, enabling it to make decisions that reflect both building structure and short-term thermal evolution. We evaluate the framework against standard control baselines and several LLM-based alternatives. Results show that the proposed approach achieves the best overall energy-comfort trade-off and the lowest PMV violation while maintaining energy-efficient operation.