"Machine Learning Approaches for Computational Fluid Dynamics of Supercritical Fluid Flows"
Vigor Yang
Prof. Vigor Yang
Petro Junior Milan
Petro Junior Milan

A team of current and former School of Aerospace Engineering faculty and graduate students has won the Best Paper Award from the 21st Annual Conference on Liquid Atomization and Spray Systems, held remotely in China, October 23-26.

The paper, "Machine Learning  Approaches for Computational Fluid Dynamics of Supercritical Fluid Flows," was co-authored by doctoral student Petro Junior Milan, his advisor Regents' Professor Vigor Yang, and two former AE grad students (now professors), Prof. Xingjian Wang and Dr. Yixing Li.

The team will now be invited to submit extension of its presented work for expedited peer review and consideration to the journal of Atomization and Sprays.

The paper investigated the challenges associated with the simulation of high-pressure supercritical fluid flows in practical devices, such as diesel engines, gas turbines and liquid rockets. It also presents a review of recent machine learning (ML) methods developed for simulation speed-up, showing their performance, main results, limitations, and opportunities. The team also discussed ways of using deep neural network to reduce computational burden of high-fidelity simulations.

"We also discussed the ability of machine learning to produce efficient surrogate models that provide fast and reliable statistical approximations of the computationally expensive simulators, hence, allowing for rapid exploration of the design space," said Milan.

The Institute for Liquid Atomization and Spray Systems, Asia, was established in 1991 as an outgrowth of the International Conference on Liquid Atomization and Spray Systems (ICLASS). It was built for the mutual exchange of scientific and technical ideas in all field related to atomization and spray at Asia area.