LLM Agent for Fire Dynamics Simulations

Significant advances have been achieved in leveraging foundation models, such as large language models (LLMs), to accelerate complex scientific workflows. In this work we introduce FoamPilot, an LLM agent designed to enhance the usability of FireFOAM, a specialized solver for fire dynamics and fire suppression simulations within OpenFOAM, a popular open-source toolbox for computational fluid dynamics (CFD). FoamPilot provides three core functionalities: code insight, case configuration, and simulation execution. Code Insight is an alternative to traditional keyword searching leveraging retrieval-augmented generation (RAG), enabling efficient navigation and summarization of the FireFOAM source code. For case configuration, the agent interprets user requests in natural language to create and modify simulation setups with accuracy and repeatability. FoamPilot's job submission functionality manages the submission and execution of simulations in high-performance computing (HPC) environments and provides preliminary analysis of simulation results. The integration of these functionalities into a single agent aims to accelerate the simulation workflow for engineers and scientists employing FireFOAM for complex simulations critical for improving fire safety.