User-Centric Personal Thermal Comfort System: Automatic Control and Fault Diagnosis for Enhanced energy efficiency, Comfort, Productivity, and Health

Personal Environmental Control Systems (PECSs). It addresses two main objectives. The first objective is to develop a robust software infrastructure to support continuous development and experimentation with the PECS. The second objective is to leverage the developed PECS software platform for Automatic Fault Detection and Diagnosis (AFDD). These developments enable the ability of PECSs to enhance occupant comfort and improve the energy efficiency of HVAC systems in buildings.
This thesis first designed the PECS platform in a centralized, modular, and scalable manner to support the integration of subjective user feedback, objective sensor data, and machine learning algorithms for intelligent decision making and autonomous control. This platform has features, including real-time data acquisition, gamification elements to improve the user experience, a user-friendly interface, automated intelligent control of PECS, and Building Management System (BMS) integration. These features support different machine learning algorithms, dynamic survey configurations, and the addition of new hardware devices.
Subsequently, this PECS platform was used and tested via two experiments for an AFDD and performance evaluations of PECS operating modes (manual and autopilot modes). The AFDD experiment showed the sensitivity of thermal preference as a metric for fault detection. However,
this sensitivity varied across varying environmental conditions. The performance evaluation experiment highlighted the requirement of a manual override option in autonomous mode in PECS and body-zone-specific control strategies. This indicates that both software and hardware
designs are equally important for the effectiveness of occupant-centric PECS. In summary, the PECS software platform developed in this thesis provides a promising foundation to support continuous the development of PECS and a potential application in the AFDD.