AI and Google Street View Forecast Large-Scale Energy Costs
Low-income households in the United States are facing a significant energy burden that is three times higher than the national average, according to the U.S. Department of Energy. With over 46 million households struggling to pay more than 6 percent of their gross income on basic energy expenses such as heating and cooling, the need for innovative solutions is urgent.
Passive design elements, like natural ventilation, offer a promising way to reduce energy consumption and create a more comfortable living environment without incurring high costs. However, the lack of data on passive design makes it challenging to evaluate the potential energy savings on a large scale.
To address this gap, a multidisciplinary team of experts from the University of Notre Dame, in collaboration with faculty from the University of Maryland and University of Utah, has leveraged artificial intelligence to analyze passive design features in households. By studying a household’s passive design characteristics, the researchers were able to predict its energy expenses with more than 74 percent accuracy.
By combining this predictive model with demographic data, including poverty levels, the team has developed a comprehensive framework for forecasting energy burden in the Chicago metropolitan area across 1,402 census tracts and nearly 300,000 households. Their research, published in the journal Building and Environment, provides valuable insights for policymakers and urban planners to identify vulnerable neighborhoods and pave the way for smart, sustainable cities.
The research team focused on key passive design factors such as window size, type (operable or fixed), and shading levels of buildings in Chicago. By using a convolutional neural network to analyze Google Street View images, the researchers found that these characteristics are closely linked to the average energy burden in households and are crucial for accurate prediction models.
Ming Hu, associate dean for research at the School of Architecture, emphasizes the importance of addressing energy burden issues, especially for low-income families, to mitigate health risks and adapt to climate change challenges. The team’s innovative approach, which combines everyday tools like Google Street View with machine learning, presents a scalable and efficient way to assess energy burden disparities and work towards energy justice in the United States.
Looking ahead, the researchers plan to expand their analysis to include additional passive design elements like insulation, cool roofs, and green roofs, with the goal of addressing energy burden disparities at a national level. By leveraging AI and machine learning technologies for the common good, the University of Notre Dame exemplifies its commitment to sustainability and making a positive impact on society.
This groundbreaking research not only sheds light on the energy challenges faced by low-income households but also offers practical solutions to reduce energy burden and promote environmental justice. By harnessing the power of AI and data analytics, the study paves the way for a more equitable and sustainable future, where all families can thrive regardless of their socio-economic status.
