Research
Multiphysics simulation has become integral to scientific discovery, complementing both theory and experiment. This integration is driven by advancements in computing hardware, software, and the growing role of data-driven and machine-learning approaches. As a researcher specializing in mathematical modeling and scientific computing, my work centers around on computational fluid dynamics, numerical methods, and open-source software for interdisciplinary collaborative applications across biology, physics, and engineering. I also aspire to bridge art and science through scientific visualization, making complex concepts more accessible and fostering science communication.
My research focuses on developing fast and accurate computational tools that meet the dual needs of discovery and design. In close collaboration with experimentalists and artists, I develop multiphysics simulation of fluid–structure interaction from three aspects: numerical, experimental, and artistic. These projects span from developing general numerical methods with broad applications to creating task-specific simulations that digitally recreate experiments and artistic processes.
Numerical
Experimental
Artistic
Computational methods, from reduced-order models to data-driven approaches, are increasingly pivotal in advancing scientific discovery. My vision is to leverage computational tools as a bridge between theory and experiment, fostering collaborative efforts to uncover novel insights in fluid dynamics and beyond. These tools can discover previously unknown phenomena, and then design new systems or materials based on those findings. This “discover–design” cycle will guide my research.
Numerical methods for fluid–structure interaction
Interactions between solids and fluids underpin the life and locomotion of living systems, yet many scientific puzzles remain unsolved regarding this seemingly simple mechanism. Efficient and accurate simulations have become mainstream tools to complement experimental approaches to study biophysical systems, for example, in intracellular flow, red blood cells, and cardiovascular dynamics. However, developing FSI methods is a nontrivial task. The main challenge stems from the intrinsic dichotomy in the preferred framework: Because solid stress comes from strain, solid simulations often use Lagrangian approaches; while fluid simulations favor Eulerian methods because fluid stress comes from strain rate. My research focuses on developing Eulerian FSI methods, which avoid the extra computational cost of remeshing solids or managing communication between Lagrangian and Eulerian frameworks.
[More information]
Data-driven methods for experimental digital twins
Digital twins offer a powerful complement to experimental research, providing access to data that are unattainable through physical experiments only. By simulating experimental environments on computers, this approach reveals parameters and variables that traditional methods cannot measure or track. I collaborate with experimentalists from physics, biology, and materials science to develop data-driven methods that optimize experiments and improve designs. Recently, I developed 3D simulations of sample vitrification in cryo-plunging, a technique for rapid cooling biological samples for cryo-EM analysis. I am also part of an inter-university project on designing redox flow batteries, combining experimental electrode fabrication with simulations to visualize and optimize electrochemical processes.
[More information]
My fascination with fluid dynamics stems not only from its significance as a research field but also from the everyday beauty it reveals. Simple actions, like pouring cream into coffee, can display mesmerizing phenomena such as the Rayleigh–Taylor instability. My goal is to highlight the scientific merit and aesthetic appeal of fluid motion through scientific visualization of my work, workshops on generative art, and contributions to the Gallery of Fluid Motion (GFM). I studied the hydrodynamics of paper marbling, an ancient art form known for its visually mesmerizing yet unexpected fluid behavior. Additionally, I developed a pipeline to visualize binary simulation in animation software for improved lighting, staging, texturing, and rendering.
[More information]