Individual Research Project 04

HiFi flow solvers for fixed walls, running on GPUs, and ROMs for aneurysm studies

Bharghav Krishna Chitneedi – ESR 04
The objective is to build a CFD model to simulate the flow in aneurysm, assuming rigid nonflexible walls with infinite resistance, running on GPU clusters; the NTUA, GPU-enabled, finite-volume URANS code version for incompressible fluids will be used as the background tool. Modifications will be performed to accommodate varying fluid properties within the CFD simulation part of the Digital Twin. The ESR will also be responsible for (a) processing data from computed tomography (CT), grid generation and post-processing (incl. metrics computation) of results for visualization purposes. Emphasis will be given to the memory handling, so as to minimize MPI-based inter-board communications. Next to this, ROM techniques will be examined for the HiFi CFD computations and, in particular, the concepts of HyperReduction, POD, PGD and Artificial Neural Networks (contribution by ESI). Studies with Newtonian fluid model and variable viscosity (Carreu’s model) will be carried out. Geometric quantities (aneurysm length, maximum diameters, etc.) and the associated non-dimensional metrics or indices (asymmetry metric, saccular index, deformation diameter rate, tortuosity index, etc.) of the aneurysm will be correlated with the hemodynamic loads and the rupture risk. The appropriateness of wall shear stress models will be evaluated. Studies will focus on estimates of the aneurysm rupture risk; from a biomechanical point of view, an aneurysm ruptures when the stresses acting on the arterial wall exceed its failure strength (material failure). The rupture risk will be examined within suitable HiFi arterial wall simulations using the EWK model of the VPS s/w by ESI. An artificial neural network will be created to represent the detailed rupture within a global FE model. By doing so, CFD can improve the understanding of factors determining the origin and progression of aneurysms. Moreover, Deep Neural Networks will be used to predict the failure points of the arterial wall, the origin and the progress of the aneurysms by avoiding the expensive CFD process. The DNN training will be based on HiFi CFD simulations done in the previous stages of the research.

Expected Results

  • A CFD s/w running on GPUs for the aneurysm studies (steady simulation; fixed walls), as a possible component of the Digital Twin..
  • Pre- and post-processors, connection with CT tools.
  • ROM for the HiFi computations for aneurysm modelling, an alternative model to be built in the Digital Twin.
  • Useful conclusions on the rupture risk.
This project has received funding from the European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement No 859836
Privacy Settings
We use cookies to enhance your experience while using our website. If you are using our Services via a browser you can restrict, block or remove cookies through your web browser settings. We also use content and scripts from third parties that may use tracking technologies. You can selectively provide your consent below to allow such third party embeds. For complete information about the cookies we use, data we collect and how we process them, please check our Privacy Policy
Consent to display content from - Youtube
Consent to display content from - Vimeo
Google Maps
Consent to display content from - Google