Product Detail

About the Speaker

Vigor Yang is presently Ralph N. Read Chair and Regents’ Professor at the Georgia Institute of Technology. He has published extensively in aerospace propulsion, combustion, and data science. He was the recipient of the American Institute of Aeronautics and Astronautics (AIAA) Air-Breathing Propulsion Award (2005), the Pendray Aerospace Literature Award (2008), the Propellants and Combustion Award (2009), and the von Kármán Lectureship in Astronautics Award (2016). He was awarded the Worcester Reed Warner Medal (2014) by the American Society of Mechanical Engineers (ASME), and the Lifetime Achievement Award (2014) by the Joint U.S. Army, Navy, NASA, and Air Force (JANNAF) Interagency Propulsion Committee. He also received the Statistics in Physical Engineering Sciences Award (2019) from the American Statistical Association (JSA). A member of the U.S. National Academy of Engineering, an Academician of Academia Sinica, and a foreign member of the Chinese Academy of Engineering, Dr. Yang is a fellow of the AIAA, ASME, Royal Aeronautical Society, and Combustion Institute.

Abstract

This lecture will address a multi-fidelity modeling strategy to facilitate data-enabled design of cimbustion systems. As a specific example, the issue of combustion dynamics (unsteady flow motions in combustion chambers) will be discussed. An interdisciplinary research program is under way, focused on the development of an efficient and robust capability to understand, analyze, and predict combustion dynamics in contemporary and future combustion systems. This effort involves work in supercritical combustion, combustion instability, reduced-order modeling (emulation), statistics, uncertainty quantification, and machine learning. Recent breakthroughs in modeling and data analytics techniques have been utilized to substantially improve modeling capabilities. New techniques address issues specific to physics extraction of complex systems, such as propulsion and power-generation engines. This research will enable efficient (with practical turn-around times) design space surveys, investigating how design attributes and operating conditions affect system stability behavior. The integrated approach described here starts with large eddy simulation (LES)-based high fidelity modeling and simulations of combustion dynamics in engines. Reduced-basis models and emulation then leverage the established database for physics-based data assimilation. Stochastic-based extraction of physics from complex flowfields provides faithful and interpretable representations of the underlying mechanisms. Feature extraction techniques are incorporated into a spatio-temporal surrogate model built on machine-learning techniques such as Gaussian process (GP) regression. Combined with statistical methodologies and control theories, these techniques allow for an efficient survey of flow evolution and combustion dynamics, with special attention to the identification of combustion response and gasdynamic driving mechanisms. Data-driven and physics-based quantification of the transfer function for the identified mechanisms is achieved. Finally, a system-level model is developed for effective assessment of combustion stability behaviors in a practical system.