Artificial Intelligence in Performance-Driven Design

2Machine Learning in Urban Building Energy Modeling

Narjes Abbasabadi¹ and Mehdi Ashayeri²

¹ Department of Architecture, School of Architecture, College of Built Environments, University of Washington, Seattle, WA, USA

² School of Architecture, College of Arts and Media, Southern Illinois University Carbondale, Carbondale, IL, USA

Introduction

Addressing climate change targets and fostering low‐carbon urban areas is crucial, as cities are responsible for approximately two‐thirds of worldwide energy consumption and greenhouse gas emissions (IPCC 2014). An accurate representation of urban building energy dynamics is key for informed energy‐driven planning, design, optimization, and policy evaluation (Reinhart and Davila 2016; Kontokosta and Tull 2017; Mostafavi, Farzinmoghadam, and Hoque 2017; Happle, Fonseca, and Schlueter 2018; Schiefelbein et al. 2019). Urban Building Energy Modeling (UBEM) stands as a key element in the domain of urban energy management, offering invaluable insights into the energy dynamics of cities and the multifaceted aspects of the built environments. Urban energy modeling endeavors to bridge the gap in data by generating essential quantitative energy information and strategies for energy reduction using various methods, including physics‐based or data‐driven models (Reinhart and Davila 2016; Keirstead, Jennings, and Sivakumar 2012). However, despite significant efforts, existing urban energy modeling tools are constrained in presenting a realistic energy ...