Individual Research Project 05 

HiFi flow solvers for flexible walls, running on GPUs, and Big Data Analysis for aneurysm studies

Andreas Chrysolouris – ESR 05
Since CFD analyses making assumption of rigid walls cannot provide information about stresses within the aneurysm wall and their effect on its growth and rupture, the GPU code of ESR04 will be extended to accommodate moving walls by incorporating a generic wall thickness model and Fluid-Structure Interaction (FSI) techniques. The resulting unsteady CFD runs will be very expensive memory-wise. Thus, a method based on Deep Neural Networks (DNN) will be considered to replicate the timesteps of the CFD run. The DNN will be also used to reconstruct solution at the intermediate timesteps based on the CFD runsThe structural problem within the solid region will be treated using the finite element method (FEM). Both domains are coupled at the interface, i.e. at the luminal surface, where data (fluid pressure and wall shear stresses, WSS, and wall displacements) will be exchanged. In the absence of patient-specific data for the wall thickness, simulations will be based on the mean value obtained after processing available data for patients. In the FEM, inlets and outlets of the solid domain will be fixed and only the intermediate surface will be free to change. Streamlines and WSS will be visualized to understand the hemodynamic flow patterns in aneurysms. Also, the temporal-averaged stress distribution inside the aneurysm wall will be visualized over the inner and outer surface. All available patient data will be considered as Big Data and analytics will be studied using and improving the data analytics options of the ESI Mineset proprietary software and other techniques like Regression Trees, Decision Trees, Random Forest etc. (contribution by ESI). The output will assist the understanding and the validity of the simulations results presented above. To employ accurate low-cost mechanisms to adapt of the computational grids of both domains to the changing interface and the communication of interfacial data in a conservative manner, volumetric NURBS and RBF methods will be used and compared. Simple Fluid-Solid-Growth models of cerebral aneurysm evolution, according to the existing literature, which combine fluid and solid mechanics analyses of the vascular wall with the kinetics of biologic growth and re-modelling, will be used. The model can further be extended by using the available patient data. A DNN will be trained based on the real data to predict the cerebral aneurysm evolution and validate the existing Fluid-Solid-Growth models

Expected Results

  • A CFD s/w running on GPUs for the aneurysm studies (unsteady simulation; flexible walls), as part of the Digital Twin.
  • Know-how on the coupling of CFD and FEM (FSI) for the aneurysm study..
  • Grid adaptation and morphing techniques. Comparative studies..
  • Useful conclusions on the rupture risk (with respect to ESR04).
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