Urban Scale Building Energy & Materials Use
Cities are responsible for 67% of global energy use. Building sector accounts for considerable amount of energy consumption, contributing to adverse environmental and climate change impacts. As the consequences of climate change increase, there is an increased focus on the building sector, a major negative contributor and a major positive solution provider. Currently, the knowledge regarding building energy use at the urban scale in accordance to climate change projections is scarce. Understanding urban building energy use provides perspective about a city’s energy demand which is beneficial for climate change reduction goals and policies, design of distributed energy resources, and utility planning.
In addition to use phase energy, new materials entering into urban building stocks because of maintenance and renovation, and wastes generated as result of demolition have significant environmental impacts. Therefore, there is an expressed need to analyze material stock and material flows of existing buildings at an urban scale. Knowledge of where and what types of building materials existing in a city may allow for more sustainable maintenance, demolition, and waste management.
Energy Use & Climate Change
Approaches to predicting building energy use range from linear regression to advanced machine learning approaches such as random forest and extreme gradient boosting. However, disparate methods, temporal resolutions, results, and recommendations have emerged. Furthermore, integrating climate change models into building energy modeling at urban scale is limited. The goal of this research is to select a machine learning model that provides good performance on predicting annual energy use of commercial buildings. Moreover, the predictive power of this model is used in concert with climate change scenarios.
Four prediction models (multiple linear regression, regression tree, random forest, and extreme gradient boosting) were developed based on the Commercial Building Energy Consumption Survey. Comparing the mean absolute error and root mean square error of these models shows that random forest model provided better predicting performance.
In the Commercial Building Energy Consumption Survey, climate regions have low spatial resolution. Therefore, to implement the random forest model for predicting commercial buildings energy use, new geographic regions with higher spatial resolution were created. Under the highest scenario established by the Intergovernmental Panel for Climate Change (RCP8.5), the majority of geographic regions within U.S. will experience increase in energy use intensity compared to 2012. However, the magnitude of increase varies for different regions which shows the importance of regional mitigation policies and plans for commercial buildings.
Mohammadiziazi, R., Copeland, S.+, Bilec, M.M.* (2021). “Urban building energy model: Database development, validation, and application for commercial building stock.” Energy and Buildings, 248:111175. https://doi.org/10.1016/j.enbuild.2021.111175
Mohammadiziazi, R., Bilec, M.M.* (2020). "Application of Machine Learning for Predicting Building Energy Use at Different Temporal and Spatial Resolution under Climate Change in USA." Buildings, 10 (8), 139. https://doi.org/10.3390/buildings10080139
Mohammadiziazi, R., Bilec, M.M.* (2019). "Developing a framework for urban building life cycle energy map with a focus on rapid visual inspection and image processing." Procedia CIRP, 80, 464-469. https://doi.org/10.1016/j.procir.2019.01.048