Five key takeaways from Duke University’s Energy Data Analytics Symposium

Dr. Colin Parris during the Energy Data Analytics Symposium 2020
Keynote speaker Dr. Colin Parris (Senior VP, GE and CTO, GE Digital) discussed how digital models are helping transform GE’s offerings. 

At Duke University’s Energy Data Analytics Symposium in December 2020, experts talked shop about how data science techniques can transform energy systems to become more accessible, affordable, reliable, and clean.

More than 200 researchers, energy professionals, and students from 19 countries and nearly 100 organizations took part in the two-day convening hosted by Duke University’s Energy Data Analytics Lab. The event featured 14 presentations by researchers and industry experts, several virtual networking opportunities, and a “lightning talks” competition that drew entries from 21 emerging scholars and early-career professionals. (View all presentations.)  

Here are five key takeaways from the Symposium:  

1. Digital models of energy systems—fueled by data monitoring and machine learning techniques—can lead to better planning and management of those systems. Check out: Keynote speaker Dr. Colin Parris (GE Digital) on “digital twins” for predicting equipment failure and optimizing power plant operation; Dr. Dylan Harrison-Atlas (National Renewable Energy Laboratory) on similar methods for increasing production and minimizing wake effect at wind farms; and Dr. Zoltan Nagy (University of Texas at Austin) on a deep neural network for urban energy simulation models.
 
  
2. Remote sensing and machine learning techniques can automatically monitor energy systems in novel ways. Check out: Dr. Kyle Bradbury (Duke University) on estimating distributed generation and consumption; Martha Morrissey (Development Seed) on identifying transmission infrastructure; Dr. Brian Min (University of Michigan) on estimating energy access and reliability; and Dr. Heather Couture (Pixel Scientia Labs) on monitoring power plant emissions and capacity factors.  
 
3. Advanced statistical methods, combined with machine learning, can help assess and predict the effectiveness of energy system improvementsCheck out: Dr. Brian Prest (Resources for the Future) on a novel method for evaluating time-varied electricity prices and other efforts to reduce energy demand; Dr. Ed Rubin (University of Oregon) on analyzing 600M natural gas bills to estimate consumer responses to price changes; and Dr. Cynthia Rudin (Duke University) on using sophisticated matching methods to evaluate the impact of new practices (e.g., inspection of manholes to prevent explosions and fires); and Mario Bergés (Carnegie Mellon) on energy-efficient autonomous buildings.
 
4. New platforms for collecting, visualizing, and organizing data can yield transformative insights about energy systems. Check out: Dr. Elisabeth Moyer (University of Chicago) on how archival data from 200+ years of U.S. energy history can aid our understanding of the current energy transition; John Pressley and Dylan Lustig (Duke Energy) on how an open analytics platform is helping transform their utility’s offerings, planning, and operations; and Dr. Dimitris Mentis (World Resources Institute) on an online mapping platform to support energy access and electrification planning.  
 
5. Are you an early-career professionals seeking to apply data science tools to energy problems? Heed this advice. Dr. Colin Parris and other speakers emphasized the importance of understanding the broader context of energy problems, so you can “connect the dots” for people and avoid focusing too narrowly on technical details. Practice an agile workflow: get your idea or product out there as early as possible, then continue to improve it along the way. Commit to continuous learning and network like crazy: make friends in the field who know things you don’t, help them whenever you can, and reach out when you need advice.   

Related:  President Price Statement on Protests: 'A Direct Result of a Campaign to Sow Mistrust in Our Democracy'

The 2020 Energy Data Analytics Symposium was organized by Duke faculty affiliated with the university’s interdisciplinary Energy Data Analytics Lab: Dr. Kyle Bradbury, Dr. Jordan Malof, Dr. Brian Murray, Dr. Billy Pizer, and Dr. Cynthia Rudin. The Lab is a collaborative effort of the Duke University Energy Initiative (which houses it), the Rhodes Information Initiative at Duke, and the Social Science Research Institute

“The past decade has brought phenomenal advances in artificial intelligence, machine learning, and other data science techniques,” reflected Dr. Brian Murray, director of the Energy Initiative. “These methods, when combined with increasingly larger energy and geospatial data sets, hold great promise for transforming energy systems and helping decision makers achieve ambitious climate and energy access goals. Timely knowledge-sharing across industry, NGOs, and academia is critical, which is why we had robust interest in this event from such diverse quarters.”   
 
Funding support was provided by a grant from the Alfred P. Sloan Foundation. Conclusions reached or positions taken by researchers or other grantees represent the views of the grantees themselves and not those of the Alfred P. Sloan Foundation or its trustees, officers, or staff.  

For updates on energy events, news, and opportunities at Duke, sign up for the Energy Initiative’s email list


More from: | Category: University News