Rajat Mittal
Professor of Mechanical Engineering
Johns Hopkins University
Presenting in Track 11: Fluids Engineering
Presentation Title: From Vortices to Forces: Physics-Enabled Human Learning in the Age of Data and AI
Abstract: Understanding the origins of unsteady aerodynamic forces is central to problems ranging from the performance of aircraft, drones, rotors, and flow-energy devices to the propulsion of fish, birds, and insects. These forces arise from the combined effects of vorticity, added-mass response, and viscous diffusion, yet disentangling their relative contributions remains a fundamental challenge. In the age of Bigdata and AI, where massive data sets from high-fidelity simulations and experiments are increasingly driving new black-box machine learning methods, there is an increasing need for interpretable and generalizable physics-grounded methods that allow humans to continue to learn from these large datasets rather than be overwhelmed by them.
The Force Partitioning Method (FPM) addresses this need by decomposing pressure forces into vortex-induced, acceleration-induced, and viscous components. Unlike purely data-driven black-box models, FPM leverages first principles to transform complex flow fields into interpretable causal attributions. By applying FPM to diverse vortex-dominated systems including dynamic stall on foils, vortex-induced vibrations, schooling fish, and rough-wall turbulence, we demonstrate how data can be interpreted into meaningful physical insights. In this way, FPM serves not only as a diagnostic tool but also as a framework for enhancing human learning from large-scale flow data. Extensions of FPM to experimental datasets highlight its utility as a bridge between numerical, laboratory, and field measurements. Overall, FPM exemplifies how data-enabled, interpretable methods can continue to empower fluid dynamicists to extract knowledge, guide design, and accelerate discovery in the era of big data and AI.
Biography: Rajat Mittal is a fluid dynamicist and a professor of mechanical engineering at Johns Hopkins University (JHU). He also holds a secondary appointment in the JHU School of Medicine. Mittal earned his bachelor's degree in Aeronautical Engineering from the Indian Institute of Technology, Kanpur, in 1989. He then pursued a MS in Aerospace Engineering from the University of Florida, followed by a Ph.D. in Applied Mechanics from the University of Illinois at Urbana-Champaign in 1995. Mittal's academic journey includes postdoctoral research at the Center for Turbulence Research at Stanford University, where he focused on large-eddy simulation. He began his teaching career at the University of Florida's in 1996 and from 2001 to 2009, he was a faculty member at George Washington University. Since 2009, he has been a professor at Johns Hopkins University. Mittal is recognized for his work on immersed boundary methods and their applications in fluid flow problems. He leads the Flow Physics and Computation Lab at JHU, focusing on computational fluid dynamics, vortex dominated flows, biofluid mechanics, bioinspired engineering, and flow control. His research has contributed to fluid-structure interaction, cardiology, biolocomotion, bioacoustics, COVID biophysics, gastric digestion, active flow control and turbulent flows. Mittal is also the founder and CTO of HeartMetrics, Inc., a startup developing computational tools for diagnosing coronary artery disease. He is the recipient of the 1996 Francois Frenkiel and the 2022 Stanley Corrsin Awards from the Division of Fluid Dynamics of the American Physical Society (APS), and the 2006 Lewis Moody as well as 2021 Freeman Scholar Awards from the American Society of Mechanical Engineers (ASME). He is a Fellow of ASME, APS, and AIMBE, and an Associate Fellow of AIAA. He is an associate editor of several journals including Journal of Computational Physics and Physics of Fluids.