Dissertation Defence: Evaluation of Energy Retrofits for Residential Buildings in Canada: An Integrated Modelling Approach
March 26 at 8:30 am - 12:30 pm
Haonan Zhang, supervised by Dr. Kasun Hewage, will defend their dissertation titled “Evaluation of Energy Retrofits for Residential Buildings in Canada: An Integrated Modelling Approach” in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Civil Engineering.
An abstract for Haonan Zhang’s dissertation is included below.
Examinations are open to all members of the campus community as well as the general public. Please email Kasun.Hewage@ubc.ca to receive the Zoom link for this exam.
Abstract
Energy retrofits play a critical role in enhancing buildings’ indoor thermal comfort, reducing energy consumption, and mitigating greenhouse gas (GHG) emissions. However, identifying optimal retrofit strategies remains challenging due to the diverse building characteristics, occupant behaviours, and climate variability. Furthermore, conventional physics-based building energy modelling (BEM) used for energy retrofit evaluation often requires detailed building-specific information and involves complex modelling procedures, while the computational demand of physics-based BEM poses additional limitations. This research aims to develop an integrated approach to evaluating building energy retrofit strategies by combining physics-based BEM with data-driven approaches. A multi-stage approach was proposed to address key challenges and bridge existing research gaps.
In the first stage, a systematic literature review was conducted to examine current practices in energy retrofit evaluation and identify uncertainty sources in BEM and retrofit assessment. In the second stage, a life cycle thinking-based energy retrofit evaluation framework was formulated. This framework enables a comprehensive assessment of life cycle GHG emissions and life cycle costs for various energy retrofit packages, facilitating holistic retrofit decision-making. The third stage introduced an integrated approach that combines physics-based BEM and interpretable machine learning techniques to quantify uncertainties in retrofit evaluation and identify optimal energy retrofit packages. This approach significantly improves the computational efficiency of conventional physics-based energy modelling and enhances the transparency of data-driven techniques. In the fourth stage, a data-driven approach was developed to analyze post-retrofit building energy load profiles and generate synthetic energy data using state-of-the-art deep generative models (DGMs). The results demonstrate that DGMs are effective in synthesizing fine-grained energy data while addressing challenges related to data scarcity and privacy concerns. Finally, this research provided building energy modelling practices for energy retrofit practitioners and policy recommendations to promote the penetration of energy retrofit programs.
The outcomes of this research provide overall methodological and practical contributions to the field of building energy research. The proposed approach supports multiple stakeholders, including energy researchers, retrofit practitioners, homeowners, utility providers, and municipalities, in evaluating retrofit impacts and identifying energy-efficient, cost-effective, and low-carbon retrofit strategies for existing residential buildings in Canada.
