ChatTwin: Enabling Natural Language Interactions with Infrastructure Digital Twins

Infrastructure lifecycle management requires interactions with dynamic datasets. Traditional interfaces often hinder users’ ability to rapidly locate lifecycle-specific information they need. Our work proposes ChatTwin, a system that employs Large Language Models (LLMs) to enable natural language queries related to various lifecycle stages of infrastructure, with a focus on operations and maintenance. Simulated scenarios were constructed to test the system. The results demonstrate that the system can effectively categorise interactions, fetch relevant information, and produce human-friendly outputs. With this LLM-based approach, we present an improvement in user-centricity in infrastructure lifecycle management, streamlining interactions and decision-making throughout the entire infrastructure lifecycle.