Embedded Files
Respiratory Flows
Previous research focused on developing computational methods for the simulation, analysis, and treatment planning of respiratory flows.
An end-to-end pipeline was developed that combines medical imaging, machine learning, and computational fluid dynamics.
Super-resolution of CT data
A super-resolution pipeline for CT data was developed to reconstruct high-resolution anatomical information from coarse CT scans, enabling reliable CFD simulations even when only low-resolution imaging data are available.
CNN-based surface extraction
For the automated reconstruction of 3D models, CNN-based methods for extracting the nasal cavity surface directly from CT data were developed, reducing the manual effort required for patient-specific geometry generation.
Validation of LB simulations
To ensure reliable predictions, we performed systematic validation of lattice-Boltzmann simulations for respiratory flows, demonstrating their capability to reproduce relevant flow characteristics in patient-specific airways.
Hemodynamic Flows
Patient-specific MRV measurements were combined with high-resolution lattice-Boltzmann CFD to establish a validated framework for simulating cerebral aneurysm hemodynamics with realistic, measurement-based inflow conditions. Instead of prescribing an idealized or analytical inlet profile, the experimentally measured velocity field is interpolated onto the CFD grid and used directly as the simulation inflow.
The study also compares Newtonian and non-Newtonian blood models. The non-Newtonian Carreau–Yasuda model produces different velocity, pressure, and wall-shear-stress characteristics, highlighting that the choice of blood rheology can matter when evaluating aneurysm hemodynamics.
Page updated
Google Sites
Report abuse