Research
The laboratory builds computational models of blood flow and of the tissue it supplies, based on each patient's anatomy taken from CT and MR images.
These models estimate quantities that are difficult to measure directly: the flow carried by a narrowed artery, the volume of tissue at risk after a stroke, the stress in an aneurysm wall. The work is organised in two parts — the diseases we model with clinical collaborators, and the computational methods those models rest on.
Clinical applications
Diseases the laboratory models directly, in collaboration with clinical partners.
Coronary artery disease
Estimating whether a narrowed coronary artery limits flow, from a CT scan rather than a pressure wire.
Read more →Cerebrovascular disease and stroke
How much brain tissue is at risk when a vessel is occluded, and what compensates for the loss.
Read more →Aortic aneurysm
A mechanical criterion for rupture risk, where clinical practice still relies mainly on diameter.
Read more →Venous disease and compression therapy
How the settings of a compression device translate into flow inside the leg.
Read more →Computational methods
The methods the applications rest on. Most are developed because an existing approach is too slow, too coarse, or stops where the imaging stops.
Patient-specific modeling
Turning a clinical scan into a solvable model, including the boundary conditions the scan does not contain.
Read more →Reduced-order models
Recovering three-dimensional accuracy at a fraction of the cost, so results arrive within a clinical timeframe.
Read more →Fluid–structure interaction
Solving flow and deformation together at one to two orders of magnitude lower cost.
Read more →Synthetic vascular trees
Generating the vessels that imaging cannot resolve, under physiological constraints.
Read more →Multiscale transport
Heat and mass exchange between a vascular network and the tissue it runs through.
Read more →Uncertainty quantification
Which assumptions a computed diagnostic index actually depends on.
Read more →Methods and tools
Members come from mechanical engineering, bioengineering and applied mathematics backgrounds. The work involves developing and running solvers, and comparing their predictions against clinical and experimental measurements.
Finite element simulation
Solving the Navier–Stokes equations and solid mechanics on unstructured meshes.
Image-based modeling
Turning CT and MR data into vessel geometry, and generating the branches imaging cannot resolve.
Multiscale coupling
Joining one-dimensional networks to three-dimensional domains so both stay conservative.
Reduced-order modeling
Calibrating fast approximate models against a small number of full simulations.
Uncertainty quantification
Sensitivity analysis to find which physiological parameters actually matter.
Programming and HPC
Python, C++ and MATLAB; running solvers on high-performance computing resources.
Ongoing funded projects
| Project | Agency | Role |
|---|---|---|
| Cardiovascular and cerebrovascular M3DT based on patient-specific diagnostic technology | Ministry of Food and Drug Safety · KHIDI | Lead |
| Multiscale brain mechanics model for digital-twin-based brain monitoring | Samsung Research Funding & Incubation Center | Lead |
| Research center for a precision medicine platform based on smart hemodynamic indices | National Research Foundation of Korea | Joint |
| CNS-focused fluid-shift driven diseases: diagnosis and treatment | Ministry of Health and Welfare | Joint |
| Sustainable nuclear energy using spent-fuel chlorination and molten-salt fast reactors | KAIST | Joint |