Building Occupancy and Thermal Modelling in the Wild

• Author / Creator

  Zhang, Tianyu

• Occupancy and thermal modelling is the foundation of several smart building applications, such as intelligent control of residential and commercial buildings' Heating, Ventilation and Air Conditioning (HVAC) system to improve energy efficiency and overall occupant experience in the built environment. And yet developing occupancy and thermal models is quite challenging due to the lack of information about the building's layout and structure, and reliable means for collecting ground truth about occupant presence and actions. Thus, most HVAC systems are presently operated without incorporating these models, a practice that has been proven to be extremely inefficient and inflexible. This thesis aims to address the challenges of building suitable models using data collected by sensors that are normally installed in buildings. We propose black-box models to estimate the occupancy state of the building, and grey-box lumped parameter thermal models to explain how the indoor temperature changes over time. The black-box occupancy models are built using the data acquired through the Building Management System (BMS), while the grey-box thermal models are built using the data reported by smart thermostats. We present a methodology for transferring occupancy estimation models from a controlled environment, where training data is abundant, to another similar environment, where training data is sparse or non-existent. Experiments run using real data collected from a large number of buildings corroborate that the transferred models can achieve high accuracy. Moreover, the thesis presents an open-source toolkit, which enables the development and evaluation of various occupancy estimation models across different buildings. Several case studies are presented to show the usefulness and extensibility of this toolkit.

• Subjects / Keywords

  • Occupancy Modelling
  • Smart Buildings
  • Thermal Modelling
  • Non-intrusive Monitoring
  • Supervised Learning
  • Transfer Learning
  • Domain Adaptation

• Graduation date

  Fall 2019

• Type of Item

  Thesis

• Degree

  Master of Science

• DOI

  https://doi.org/10.7939/r3-zm45-7x26

• License

  Permission is hereby granted to the University of Alberta Libraries to reproduce single copies of this thesis and to lend or sell such copies for private, scholarly or scientific research purposes only. Where the thesis is converted to, or otherwise made available in digital form, the University of Alberta will advise potential users of the thesis of these terms. The author reserves all other publication and other rights in association with the copyright in the thesis and, except as herein before provided, neither the thesis nor any substantial portion thereof may be printed or otherwise reproduced in any material form whatsoever without the author's prior written permission.