Two Argonne researchers will discuss their Laboratory-Directed Research and Development (LDRD) sponsored work at the LDRD Seminar Series presentation Tuesday, March 12, 2019, at 12:30 p.m. in Building 212, Room A157. All are welcome to attend.
Visit the LDRD website to view upcoming seminars.
Yuki Hamada
“Unmanned Aerial Systems (UASs) for Environmental Characterization and Monitoring,” by Biophysical Scientist Yuki Hamada (EVS)
Abstract
Argonne’s Environmental Science, Decision & Infrastructure Sciences and Strategic Security Sciences divisions came together and explored the capability of unmanned aerial systems (UASs) for environmental characterization and monitoring. For this first effort, we aimed to characterize land surface and subsurface, and to detect land cover change using UAS imaging. We collected imagery using UAS-mounted RGB- and thermal-infrared cameras. With the help of domain experts such as archeologists and ecologists, we successfully extracted information from the imagery including the indication of subsurface artifacts, land cover change, vegetation types and road damage. During the project, we faced various challenges and learned valuable lessons with regards to improving the odds of a successful mission. In addition to a summary of our upcoming UAS activities, this talk will conclude with tips for how to develop UAS tasks effectively and efficiently for your future research and projects.
Feng Qiu
“Machine Learning for Mathematical Optimization: Using ML to Help Solve Large-Scale Security- Constrained Unit Commitment (SCUC) in Power Systems,” by Computational Scientist Feng Qiu (ES)
Abstract
Complex operational and planning decisions often rely on mathematical optimization to come up with an optimal (or satisfying) solution in a reasonable time. In most of the application settings, those decisions are made routinely and useful information from solving the optimization problems could be accumulated and used for learning better strategies to make better decisions faster. In this LDRD swift project, we developed a machine learning framework to expedite the solution of a fundamental operational decision-making problem in power systems, i.e., security constrained unit commitment (SCUC), and the results show that our machine learning framework can speed up the SCUC by 10 times on average.
Biography
Feng Qiu received his Ph.D. from the School of Industrial and Systems Engineering at the Georgia Institute of Technology. He is a principal computational scientist with the Energy Systems division. His current research interests include optimization in power system operations, electricity markets and power grid resilience.