Highly Resolved Simulations of Hemodynamics in Intracranial Aneurysms
University of Siegen
Local Project ID:
HPC Platform used:
SuperMUC of LRZ
A significant part of modern mortality is contributed by strokes, caused by the rupture of intracranial aneurysms (IA). Nearly 4-5% of the world population is reported to be suffering from IA. The deployment of a flow diverter stent in the parent artery of an aneurysm is a novel and minimally invasive treatment procedure, which can cause complete obliteration of the aneurysm by thrombosis.
Numerical simulations of the complex dynamic flow can help to better understand important effects and to optimize the design of such stents. Due to different spatial scales in the complete artery (order of cm) and the thin wires of the flow diverter stent (order of µm), extremely fine spatial resolutions are required. Moreover, several cardiac cycles are required for the simulation of the complex biological process of thrombosis. The necessary computational resources for this kind of detailed simulations are only available on modern HPC systems.
The main goal of this project is the quantification of blood flow changes due to intubation of a flow diverting stent. For the simulations the open source lattice Boltzmann solver Musubi is executed on 16,000 cores of SuperMUC for 48 hours each. The aneurysm is discretized with ~500 million fluid elements, where the mesh around the stent is locally refined to represent its geometry better.
Ongoing efforts aim to model the complex biological process in more detail, which includes simulation of moving artery walls. In addition, the efforts of the research team also target the assessment of high frequency flow fluctuations in bifurcation aneurysms, where computational meshes usually consist of up to 1.5 billion cells, capturing intricate hemodynamic phenomena.
The scientific team consists of the following researchers:
Kartik Jain1, Simon Zimny1, 2, Harald Klimach1, Jörg Bernsdorf2, Sabine Roller1 (PI)
1 Simulation Techniques and Scientific Computing, University of Siegen, Siegen, Germany
2 German Research School for Simulation Sciences and RWTH Aachen University, Aachen, Germany
The presented research work was performed within the THROMBUS project, funded by the European Commission in the seventh framework program in the area of Virtual Physiological Human (ICT-2009.5.3, PR 269966).
The simulations were performed on the GCS supercomputer SuperMUC installed at the Leibniz Supercomputing Center, Garching near Munich (Grant – pr45du).
Prof. Dr.-Ing. Sabine Roller
Simulation Techniques and Scientific Computing, University of Siegen
Hölderlinstr. 3, D-57076 Siegen/Germany