Occupant-Centric Digital Twin for Advancing Thermal Comfort in Shared Indoor Spaces

The importance of ensuring optimal thermal comfort in built environments is becoming more important and also more complex, especially as office spaces continue to adopt new shared and flexible layouts. Recent developments in sensing technologies, including Internet of Things (IoT) devices and occupant-based control strategies, have made it possible to manage indoor environments in a more dynamic and responsive way. Emerging developments of the new generation of sensing technologies, such as machine learning algorithms, IoT devices, and control strategies based on occupants, layout new opportunities for dynamic control of indoor climates. In comparison, more traditional methods by employing Predicted mean vote (PMV) / Predicted percentage of dissatisfied (PPD) models, adaptive models, and manual system operation have often been plagued by shortages in being able to provide solutions to the diverse thermal tastes and habits of occupants, hence causing them to feel discomfort and reduced productivity. However, using these technologies widely is not without challenges. There are still issues related to integrating them with older building systems, gaining acceptance from users, protecting personal data, and making sure they work well in different types of indoor spaces and layouts. This research aims to develop and validate an AI-driven Digital Twin framework capable of real-time seating suggestions to occupants in shared workspaces to achieve thermal comfort. Three objectives guide this investigation: (1) to assess the combined effects of environmental and personal factors on thermal comfort, productivity, and concentration; (2) to develop a Digital Twin system that integrates real-time environmental sensing, AI-based prediction, and occupant feedback using a modular architecture comprising digital master creation, digital shadow development, and intelligent data analytics; and (3) to validate the system's performance, adaptability, and predictive accuracy through a real-world case study in a structurally distinct open-plan office setting. Personalized comfort information was visualized through an interface, and a proximity-based seat recommendation algorithm was introduced to support data-driven occupant decision-making. The system was assessed through a pilot study and validated in a separate office environment to test its generalizability and applicability. The goal of this work is to create a responsive and sustainable built environment for occupants that leverages smart systems to enhance both occupant well-being and building performance.