The MeDiTATe project will be participating in the 14th European Fluid Mechanics Conference (EFMC14), that will take place in Athens, Greece on September 13-16, 2022.

Two of our Early Stage Researchers, Bhargav Krishna Chitneedi and Christos Karliampas, will be presenting their activities at this Conference. A short summary of their work is reported in the following lines.

Bhargav Krishna Chitneedi – ESR 04: Numerical Prediction of Blood Flow in Arteries, Interacting with Walls, on GPUs. In this article, a framework for the Fluid-Structure Interactions (FSI) simulations of vascular blood flows, running on Graphics Processing Units (GPUs) is presented. The FSI simulations are performed, based on the partitioned approach, with the GPU-enhanced Finite Volume CFD code, PUMA (developed by the PCOpt/NTUA) for fluids, coupled with the open-source finite element solver, CalculiX, for the vessel walls. The flow is assumed to be laminar, and the non-Newtonian behaviour of the blood is represented using the Carreau model. The isotropic linear elastic model of CalculiX is used for the arterial wall computations. The GPUs are used for the flow simulation, which is expensive because of using fine mesh with boundary layers at wall vicinity. The structural analysis uses much coarser meshes and runs inexpensively on the CPU. The Precice coupling framework is used which also enables using non-matching timesteps and non-matching meshes to perform FSI simulations. The temporal velocity profile at the inlet and the 3-element Windkessel pressure model at the outlet are imposed. The FSI simulations are performed on patient-specific geometries of thoracic aorta with realistic flow conditions. The hemodynamic metrics like wall shear stresses are computed to predict the rupture prone areas which helps medical experts to take an informed decision to improve healthcare.

Christos Karliampas – ESR 06: Hemodynamic simulation of the thoracic aorta in the presence of uncertainties, using a reduced-order polynomial chaos expansion. The purpose of this paper is to investigate the sensitivity of the inlet flow profile in cardiovascular simulations of thoracic aorta. Due to limited, poor quality, patients’ data, it is a compromise to resort to idealized spatially distributed velocity profiles as the inlet boundary condition in undertaken simulations. This assumption introduces uncertainties, affecting the numerical solution in the entire fluid domain. Here, Womersley number is chosen to parameterize the imposed inlet velocity and a fast reconstruction method is proposed, based on the proper orthogonal decomposition polynomial chaos expansion, to analyze the velocity profile uncertainty propagation on the hemodynamic metrics.”