Rajeev Jaiman, UBC

Rajeev K. Jaiman is an Associate Professor and NSERC/Seaspan Industrial Chair in the Department of Mechanical Engineering at the University of British Columbia (UBC), Vancouver, Canada. Prior to his current appointment at UBC, he was an assistant professor in the Department of Mechanical Engineering at the National University of Singapore (NUS). Before joining NUS, he was the Director of Computational Fluid Dynamics (CFD) Development at Altair Engineering, Inc., Mountain View, California. The CFD technologies that Dr. Jaiman has developed are routinely used in the marine/offshore, wind turbine, nuclear reactor, automotive and aerospace industries. Dr. Jaiman earned his first degree in Aerospace Engineering from the Indian Institute of Technology, Mumbai. He received his master’s and doctorate degrees from the University of Illinois at Urbana-Champaign (UIUC). His research interests include fluid-structure interaction, computational fluid dynamics, finite element analysis, data-driven modeling and physics-based machine learning. He is currently an Associate Editor of ASME-OMAE Journal, a senior member of AIAA and a member of ASME, SNAME, USACM, APS, AAM, and SIAM.

Fluid-Structure Interaction: From High-Fidelity to Data-Driven Modeling

Advances in high-performance computing have empowered us to perform large-scale finite element analysis (FEA) routinely for millions of variables in multiphysics and coupled fluid-structure systems. These high-fidelity simulations via nonlinear partial differential equations can provide invaluable physical insight for the development of new designs and devices in marine/offshore and aerospace engineering. Despite efficient numerical algorithms and powerful supercomputers, the state-of-the-art computational fluid dynamics (CFD) and coupled fluid-structure simulations are somewhat inefficient and hence less attractive concerning downstream tasks such as parameter space exploration, design optimization and real-time control and monitoring strategies. On a similar note, two other pillars: namely theoretical analysis and experimental testing, can suffer serious limitations with regard to the scaling to realistic geometry and physical situations. The emergence of data-driven methods and machine learning has been recently recognized as a powerful alternative and can provide a fourth pillar as a unifying force to combine the three pillars of science and engineering.

In this talk, I will present some of our recent lab efforts to integrate and complement the HPC-based high-fidelity computations with data-driven modeling of multiphysics interfaces, with a particular emphasis on unsteady fluid flow and fluid-structure interaction. The primary focus of this talk is: (i) to demonstrate the capability of in-house multiphase FSI framework using Eulerian-Lagrangian and fully Eulerian formalisms, and (ii) to develop efficient reduced-order and deep learning models for the physical modeling of fluid-structure systems. I will present validation of our FSI methods and tools for increasing complexity of problems along with the demonstration of a full-scale flying bat, flexible propeller-blade cavitation, and an ice-going ship in open water. A series of canonical academic test cases will be covered to elucidate the integration of CFD/FEA datasets with model reduction and deep learning techniques for unsteady flow and fluid-structure interaction. Our in-house hybrid high-fidelity CFD/FEA and data-driven framework is precisely aligned with the aerospace and marine industry’s need for structural life prediction, control and monitoring via physics-based digital twin.