AI-driven Digital Twins for Predictive Maintenance

Abstract: An AI-driven Digital Twin (DT) framework was developed to integrate real-time sensor data, facility management systems, and AI-driven predictive analytics to enhance building operations. Implemented in a three-story academic building, the DT serves as a dynamic common data environment (CDE) that supports real-time monitoring, asset management, and predictive maintenance. By integrating IoT data streams with an LSTM-based machine learning (ML) module, the framework provides actionable insights for proactive decision-making, helping to streamline facility management (FM) workflows and enhance collaboration among stakeholders. Future work will broaden the scope of ML techniques, expand use cases, and promote wider adoption across similar facilities.