Parallel Numerical Simulation of Blood Flows in Patient-specific Aortic Dissection

Abstract

Aortic dissection is the separation of the inner layers of the aortic wall, which allows the blood to flow into it. The computational fluid dynamics allows a better understanding of its pathology and treatment. However, it is time-consuming due to the computationally expensive. In this work, we introduce a parallel algorithm for accurate and fast three-dimensional blood flow simulation of a patient-specific full-size aorta with dissections. Specifically, the unsteady Navier-Stokes equations are discretized by a unstructured finite element method in space and a fully implicit finite difference method in time, and the fully coupled nonlinear system at each time step is solved by a domain decomposition method based parallel scalable iterative algorithm. The numerical results are carefully analyzed and it shows that the simulated pressure, velocity, and wall shear stress are within a reasonable range. In addition, the algorithm achieves a parallel efficiency of 40% when using 3840 processor cores on the Tianhe-2A supercomputer, which shows the potential to do fast and high fidelity blood flow simulations of aortic dissection.