PSC Consulting
Identifying and Estimating Behind the Meter PV Capacity on a Feeder Using Machine Learning
The amount of Behind the Meter (BTM) generation may not be known to utilities. With increased penetration of rooftop solar, can PSC Consulting come up with estimates of solar power generation in a neighborhood or on a feeder? Recognizing solar panels from satellite imagery seems to be well-understood and well-defined. This student team will work to go through a learning exercise of combining the knowledge of recognizing solar panels from satellite imagery with other publicly available or utility provided information (e.g. address of the unit, approximate KW, connected feeder, scrape information from permits of the city or county or utility). This student team will also work to understand current regulations, data challenges, etc. - This student team will work to find and download Puget Sound metro area aerial images of roof-tops, process them with a program, and identify houses with solar panels. - This student team will also work to estimate the capacity of those panels. - This student team will work to determine the nearest feeder - This student team will work to aggregate solar generation capacity on those feeder(s). The outcomes this student team is working toward are 1) a technical paper submission and acceptance by IEEE Power & Energy Society or similar entity and 2) the usual university/department requirements.
Faculty Adviser
Daniel Kirschen,
Donald W. and Ruth Mary Close Professor,
Electrical & Computer Engineering
Students
AJ Solano
Ciel (Jiacheng) Sun
Larry Tran
Marcel Ramirez
Riley Estes
Victoria Nguyen