Zhecheng Wang shared his work on tracking solar installations during the Bits & Watts Community Forum on November 15, a project that has already gained media attention.
In order to measure the potential power available through solar in the U.S., Wang and collaborator Jiafan Yu are making use of high-resolution satellite imagery available through Google Earth. Wang’s project, dubbed DeepSolar, uses deep learning to identify solar installations within satellite images. It then calculates their size and estimates their capacity.
One of the challenges of expanding the use of solar energy, he explained, is the variability of how much power is produced at any given moment. By developing a deep-learning model to automatically detect solar installation and estimate their sizes, Wang et al. are creating a national database of installations within the contiguous United States. Once a capacity estimate has been established, the information can be combined with other data to obtain a deeper understanding of the underlying factors affecting solar adoption. For example, the location of solar panels could be correlated with environmental and socioeconomic factors. In addition, since Google Maps images are updated yearly, DeepSolar’s database could track changes over time.
Wang, who is in the early stages of a PhD in Civil and Environmental Engineering, responded throughout the presentation to questions from Stanford students and staff who were curious about how the program distinguishes between commercial and residential installations as well as the mechanisms for identifying solar panels.
To learn more about Wang’s research, visit http://web.stanford.edu/group/deepsolar/home or read the Stanford News story at https://news.stanford.edu/2018/12/19/inventory-indicates-goes-solar/. To use the Interactive Map, Analytics tool, and Comparitive tool visit the project website and select one of the options in the top right http://web.stanford.edu/group/deepsolar/home.
You can read past stories about Bits & Watts Community Forum speakers and Research Showcases here.
The Bits & Watts Community Forum is builds a community of Stanford students, faculty, and researchers doing work on the future electric grid. Engineers, lawyers, policy experts, economists, and psychologists alike are invited to participate. The gathering provides a regular opportunity for the Stanford research and industry communities to network and discuss the latest ideas and news about grid research and innovations. The Community Forum takes place on Thursdays during Stanford term time.
To learn more about the Bits & Watts Initiative, contact email@example.com or visit the initiative’s website at https://bitsandwatts.stanford.edu/ .