Vahraz Zamani and Travis Tippens were recently selected as one of the top six posters submitted to the SAS Analytics 2013 Student Poster Contest.
Zamani is a PHD candidate in Electrical Engineering. Tippens is an undergraduate in Electrical Engineering with a concentration in Renewable Electric Energy Systems and a 2013 FREEDM Undergraduate Research Scholar. Both students are members of the FREEDM Center working under the direction of Dr. Mesut Baran.
Their research focuses on developing a statistical model to predict the output of the Photovoltaic (PV) panels -- active power -- for the next hour, or even a day in advance. The available data for this work include next day weather prediction data as well as historical PV data with actual weather data. This model is essential for utilities to have an acceptable prediction about the installed PV panel output for customers over their short and long-term operation in addition to the utilities control decisions.
"One of the greatest challenges associated with solar power is its unpredictability. We are developing a statistical model that will provide grid operators with a
reliable estimate of available power on a minute-by-minute basis," explains Tippens. According to Zamani, "This work shows how Predictive Analytics can help utilities to manage the intermittent energy resources like solar PV in the distribution grid."
The student poster contest winners will be featured at the Analytics 2013 conference to be held October 21-22 in Orlando, FL. The Analytics 2013 conference is an industry-leading conference for the latest ideas, research and best practices in the field of Analytics. This year's program features keynote speakers Jim Goodnight, CEO of SAS, Will Hakes, Co-founder and CEO of Link Analytics, Sven Crone, Deputy Director of the Lancaster Centre for Forecasting, and Ed Gaffin, Director of Marketing Analytics and Optimization for The Walt Disney Company. In addition to the keynotes and the more than 65 breakout sessions, the conference include two special offerings, one of which is the Student Poster Contest.
Performance of the proposed model for one week (Actual vs Prediction)
Applying to the student poster contest was a two-step process. For the first step, a 250 word abstract that included a description of how the author had used analytics to improve their process and/or analyze their work. If the author's abstract was accepted, they then had to submit a poster to be judged. Over 36 posters were judged as a part of the student poster content.
Tippens will receive an all-inclusive trip to the conference for being one of the six student poster contest winners where he will present their research. A complete list of the posters submitted to the conference can be found on the Analytics 2013 website.