Blood flow simulations with application to cerebral aneurysms

Abstract

Cerebral aneurysms are dilations of arterial walls that can grow and rupture. We present a novel simulation system of the blood flow through intracranial aneurysms including its interaction with the surrounding vessel. It enables physicians to estimate rupture risks by calculating the distribution of blood pressure, velocity and wall stresses, in order to support the planning of clinical interventions. For the numerical simulation, the computational domain is extracted from medical image data of the patient's cerebrovascular system. The blood is modeled as an incompressible Newtonian fluid, and the surrounding vessel wall as an isotropic linear elastic material. Both the Navier-Stokes equations for the fluid domain and the Lame equations for the solid domain are handled with a finite element method, and the resulting linear equation systems are solved via an algebraic multigrid algorithm. Implicit coupling between blood flow and wall elasticity is achieved using an iterative fluid-structure interaction technique deforming the fluid mesh according to the wall displacement in each step. Boundary conditions are applied by prescribing measured waveforms of blood velocity and pressure at inlet and outlet areas. Due to the time-critical nature of the application, we exploit state-of-the-art numerical methods on heterogeneous CPU-GPU extremely parallel architectures.