A Semanti Digital Twin for Occupant-Centric Workplace Management

Providing high-quality indoor workplaces that meet occupants’ needs and support their activities is essential for employee health, well-being, and productivity. Established Facility Management (FM) approaches—focused on monitoring environmental parameters and viewing occupants as passive recipient of averaged ideal conditions—have shown limited success in fully meeting these expectations. Additionally, conventional Post-Occupancy Evaluation (POE) surveys, which often lack the contextual data necessary for a nuanced understanding of occupant experiences, constrain the assessment of workplace performance and thereby limit opportunities for improvement. To overcome these limitations, occupant-centric approaches has recently gained momentum, emphasizing dynamic monitoring of occupant-environment interactions to generate insights that inform workplace management decisions and optimise both physical and organisational factors. However, despite promising progress, conceptual and technical solutions for occupant-centric workplace management remain in their infancy. This thesis contributes to address this gap by developing a framework for a Workplace Semantic Digital Twin (WSDT) and testing and evaluating a prototype implementation in a case study. The framework comprises the collection and integration of static building information, dynamic sensor observations, and subjective occupant feedback obtained through micro-surveys using an Ecological Momentary Assessment (EMA) approach, along with their processing to extract meaningful insights that support workplace management decisions. Following an extensive literature review on factors characterizing indoor workplace ecosystems and current Digital Twin (DT) solutions for Facility Management (FM), the use cases, as well as functional and technical requirements of the proposed WSDT system, were identified, guiding the definition of its conceptual and technical architecture. The system integrates BIM, IoT, and Semantic Web technologies using a Linked Data approach, and is designed to store, interlink, and process heterogeneous workplace data within information containers compliant with the Information Container for linked Document Delivery standard according to predefined ontologies. To build and structure the workplace semantic knowledge base, the Occupant-centric Workplace Management Ontology (WOMO) was developed, characterizing worker-activity-workspace concepts and their interrelationships. A dedicated web platform, previously developed at the Ruhr University Bochum (RUB), was adopted to support the creation and management of the ICDD containers and customized to provide the system’s data querying and visualization functions within user-friendly dashboards. Time series data from sensors and occupant feedback are stored in a time series database, allowing platform services to retrieve data on demand and separating bulk data storage from the consolidated workplace knowledge base, which enhances efficiency. The system was implemented and tested in two stages. The first prototype was realized to monitor a single workspace at RUB, deploying a lightweight sensor network based on commercial smart home devices and using a developed web application to collect voluntary feedback from four employees. This prototype confirmed the capability of the deployed system components and functions to gather and integrate workspace and worker data, supporting a workplace performance assessment use case. Insights gained from this initial room-scale implementation informed the further development of an extended case study encompassing 14 workspaces within the same department, aiming to assess the system’s feasibility in a setting that more closely reflects real-world demands. Upscaling the system involved expanding the sensor network with additional sensors and routers to enhance signal coverage and adapting the workplace knowledge base configuration to a multi-containerized setup, enabling efficient data management at both workspace-specific and workplace-wide levels. Moreover, to address further system use cases, additional features were conceptualized for the web feedback application to capture more worker features, and four new platform dashboards were designed, although the technical implementation and evaluation of these enhancements is left for future works. This thesis contributes to enabling the adoption of occupant-centric approaches in workplace management by proposing a WSDT framework and demonstrating its feasibility through the implementation of BIM, IoT, and Semantic Web technologies to semantically integrate heterogeneous building and occupant data in a containerized solution. This integration allows for the extraction of meaningful insights to enhance workplace quality, thereby improving employee well-being and productivity, and paves the way for applying advanced reasoning and learning capabilities to proactively optimise workplace environments for their occupants.