I am interested in mathematical modeling, PDEs and numerical analysis. I apply these tools to solve complex, real-world problems. You can find more about my research here.
I am interested in mathematical modeling, PDEs and numerical analysis. I apply these tools to solve complex, real-world problems. You can find more about my research here.
Regularity problem of the axisymmetric Navier–Stokes equations
The motion of fluids like water, air and blood is captured by a set of equations called the Navier–Stokes equations. Starting from a smooth initial state, the equations promise to tell you the future. But there’s a catch: in three dimensions, nobody knows whether that promise holds. The regularity problem asks a deceptively simple question: if you start with a perfectly smooth flow, can the equations ever spontaneously produce a singularity (an infinite spike in speed, or a loss of smoothness) in finite time?
My research focuses on the axisymmetric case, where the flow looks the same from every angle around a central axis. You see flows like this every day: a tornado, the whirlpool above a bathtub drain, or coffee swirling in a cup after you stir it. Such flows are sometimes called “2.5-dimensional”: they are simpler than a fully three-dimensional flow, but the swirl around the axis can still stretch and intensify the vortex, as a tornado tightens and speeds up when it narrows. Whether this intensification can run away to infinity in finite time is still an open problem.
What I work on
My recent work considers a fluid in a finite cylinder (think of the stirred cup) and asks how the fluid’s behaviour at the walls affects regularity.
NHL boundary condition. I proved that the flow stays smooth for all time whenever the initial data satisfy a critical smallness condition. (arXiv:2605.18011)
Total-slip boundary condition. Here the fluid slides along the walls. I obtained a similar global regularity result under a smallness condition on the initial data. (arXiv:2611.13263)
Navier slip with friction (in progress, joint with Zijin Li). The walls now drag on the fluid as it slides, which is closer to how real surfaces behave.
“Critical” means the condition is scaling invariant: it does not depend on the units used to measure length and time, which is the natural threshold for the Navier–Stokes equations. The condition is also explicit, with computable constants, so one can check from the initial state whether a given flow is guaranteed to stay smooth forever.
Mathematical modelling
Mathematical modelling applies ODEs and PDEs to model the spread of a pandemic. Ordinary differential equations (ODEs) are the backbone of the classic SIR model—they divide people into compartments and track the compartments across time, assuming everyone mixes evenly. In real life, people also move, commute, and cluster in cities. That’s where partial differential equations (PDEs) step in. They add the dimensions of space and even age, turning the model into a dynamic map where the disease ripples across neighbourhoods, diffuses along travel routes, and hits some places harder than others.
My current research aims to incorporate game theory into the models and to study the economic and social impact. Previously, we have studied a COVID-19 model where the population exhibit different behavior based on some economic factors. You can find out a simulation of this model on the right. We can see that waves and waves of infection population.