Integrated machine learning and data-driven retrofit pathways for equitable urban building energy performance

Friday, September 27^th, 2024

2-3 pm PT

In-person: Y2E2 room 299

Virtual: Zoom

Abstract

In recent years, the conversations around climate action plans and environmental justice have become increasingly intertwined, showing a growing understanding of how our built environment impacts the planet and the people around us. Cities have begun addressing equity more seriously in their climate action plans, often creating specific plans and initiatives to address historical inequities in urban development. These social inequities are particularly apparent in buildings and urban morphology; in disadvantaged communities, buildings are less efficient and there are fewer public green spaces. Any conversations about developing or redeveloping the urban environment must now consider how the people and the planet are impacted. The overarching motivation for this dissertation is to lay the foundations for frameworks and methods for evaluating equitable decarbonization pathways at building- and urban-scales.

Overall, the research in this dissertation contributes to frameworks and methods for evaluating building- and district-scale retrofit opportunities for equitable decarbonization. Data-driven methods built off open-source data streams are combined with energy modeling and machine learning methods to create comprehensive analyses of urban building energy performance at multiple scales. By performing these analyses through the lens of energy equity, I gain a deeper understanding of how differences in the structure of the built environment can influence energy performance. The results and methods I introduce have theoretical and practical implications that enable a more equitable approach to urban building decarbonization.