BRITE Pivot: Quantum Computing and Machine Learning for Fluid-Structure Interaction Problems

Texas A&M University, College Station TX

Investigators

Abstract

This Boosting Research Ideas for Transformative and Equitable Advances in Engineering (BRITE) Pivot award will fund research that accelerates the discovery of solutions to grand challenges involving interactions between fluids and solids, with applications to aquatic habitat restoration, optimized energy efficiency of turbine generators, and the study and treatment of heart disease, thereby promoting the progress of science and advancing the national prosperity, welfare, and health. Understanding the physics of fluid-structure interactions is a critical prerequisite for such progress. Advanced computational tools play an important role but are limited by resource and time constraints, even on existing supercomputers. This project will address these limitations by developing new computational tools that rely on state-of-the-art machine learning and quantum computing techniques and that also anticipate implementation on future generations of quantum computers. Quantum computing is still five to ten years away from a practical gate-based digital quantum computer but has shown promise for handling probabilistic rather than deterministic problems already on available quantum computers and by inspiring new algorithms for classical computers. Machine learning has revolutionized many industries such as image/speech recognition and is expected to have a great impact on scientific computing. This research is integrated with activities that aim to increase interest in science, technology, engineering, and math among the public and students, including through accessible online videos, new curricular content, and integration of undergraduate students in interdisciplinary research. This research aims to make fundamental contributions to the computational study of turbulent fluid-structure interactions by pivoting from classical approaches to new computing concepts, namely quantum computing and machine learning, enabling significant improvements in computational speed and efficiency for both forward and inverse problems. Consistent with the probabilistic perspectives that underpin both quantum computing and machine learning, a key concept of this research is reformulating traditional fluid-structure interaction problems into probabilistic ones. To this end, this project will extend the filtered density function approach for turbulence modeling to large-eddy simulations of particle-turbulence interactions, develop scientific machine learning techniques for inverse discovery of closure terms and forward and inverse prediction of fluid-structure interactions for energy harvesting applications, and develop quantum-ready and quantum-inspired algorithms for mixing applications. These developments will help the principal investigator gain expertise in novel research tools that have the potential to lead to significant advancement of fundamental knowledge and enable application to problems as diverse as cardiovascular flows, bioinspired flow control and sensing, or control of biomimetic aquatic robots. 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.