An image-based computational fluid dynamic method for haemodynamic simulation

Blood vessel in circulatory system often shows a rich variety in geometry as well as in morphology and image-based anatomic modeling of such vessels for haemodynamic simulation is usually of great time-consumption. In this paper we proposed a new computational fluid dynamic method that is capable to directly utilize the pixel and/or voxel dataset based on the threshold for the raw medical images of MRI, CT and etc. to define the computational domain in a Cartesian coordinate system. Boundary of the domain is determined in a manner of VOF (fractional Volume Of Fluid) and the flow is computed using a non-staggered finite difference method, in which treatment of the boundary conditions is conducted with the neighboring point local collocation method (NPLC). Results are presented of two pulsatile flows in a two-dimensional stenosed blood vessel model. Comparison with other reliable results shows that the present method is able to reasonably predict complicated vortical flow in blood vessel very well.