CDS&E: Integrated Molecular Dynamics and Machine Learning Investigation of Gas-Liquid Interface in Transcritical Reacting Flows

Propulsion systems, such as rockets or jet engines, are more efficient when they work at high pressures. But when liquid fuel is injected in these very hot and pressurized conditions, the usual boundary between gas and liquid disappears. How this change happens is not fully understood, because it is not easy to visualize or measure what’s happening at small scales near the gas/liquid interface. The objective of this project is to use integrated molecular dynamics simulations and active machine learning to fill these knowledge gaps. Molecular simulations can mimic the behavior of atoms in these extreme conditions without needing any assumptions other than a potential model that describes the interactions between atoms. Machine learning can help make these simulations faster and more effective. The project aims to understand the underlying physics of transcritical reacting flows. In particular, it will address the question of how to model the non-equilibrium interfaces with thermodynamic jumps and how to handle the thickened transcritical interfaces. It employs an integrated approach, combining molecular dynamics simulations and active machine learning techniques, to address these questions. Its goal is to uncover previously unexplored physics governing interfaces, phase changes, and molecular transport within transcritical high-pressure flows. The incorporation of machine learning not only substantially diminishes the computational burden associated with molecular simulations but also possesses the potential to generate innovative, more precise models without prior knowledge of the underlying physical laws. The new knowledge will be instrumental in the development of more efficient and environmentally friendly liquid-fueled energy and propulsion systems. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.