A smart building digital twin prediction framework integrating fiber optic sensing, IoT, and BIM

This study proposes a smart building digital twin prediction framework that integrates fiber optic sensing, Internet of Things (IoT), and Building Information Modeling (BIM) to address energy prediction and resource optimization challenges. By leveraging fiber optic sensing and IoT-based data collection, the system monitors key parameters such as energy consumption, indoor temperature, and occupancy in real time, while dynamically updating BIM models for centralized analysis and visualization. The Long Short-Term Memory (LSTM) network was employed to effectively capture non-linear dependencies and temporal patterns in building performance data. Experimental validation was conducted in a smart campus with diverse building types. The LSTM model achieved a mean absolute error (MAE) of 2.3%, significantly outperforming traditional methods such as ARIMA and linear regression. Integrating predictive analytics into BIM enables actionable insights, such as energy usage heatmaps, helping facility managers identify inefficiencies and optimize resource allocation. Resource optimization experiments demonstrated the system’s utility, reducing peak energy consumption by 15% and overall consumption by 9.8%. The results underscore the potential of combining fiber optic sensing, predictive analytics, and intuitive visualization to enhance operational efficiency and sustainability, offering a robust solution for smart city and smart building management.