Numerical simulation of pulsatile blood flow across a tilting–disk mechanical heart valve

Thumbnail Image
Issue Date
Nejadmalayeri, Alireza
Hoffmann, Klaus A.

To overcome several clinical challenges involving mechanical heart valves, accurate numerical simulation of blood flow through these devices has been of interest. Since heart disease is the leading cause of death around the world, the hemodynamic study of the heart is of extraordinary interest in the field of bio fluid dynamics. Recent numerical/experimental investigations have shown that mechanical heart valves inherit the “production of sufficiently large shear and turbulent stresses to cause clinical problems such as hemolysis.” Due to several parameters including the non-Newtonian behavior of blood, pulsatile waveform, strong blood and tissue interactions, clinical difficulties, etc., experimental examination of blood flow in the heart and its valves is a very difficult task. Therefore, comprehensive numerical analysis of this complex fluid-structure system is essential. However, precise experimental investigations are still imperative for developing appropriate and accurate turbulence models and for validating numerical techniques.In the current computational effort, the first set of objectives was numerical investigations of vortex shedding behind a two-dimensional tilting-disk mechanical heart valve in a straight channel, which is a simplified representation of the mitral position, using first- and several higher-order finite volume schemes as well as examination of the non-Newtonian viscosity effects on shedding frequency and amplitude. For the low Reynolds number (low inflow), both first- and higher-order schemes resulted in identical shedding frequencies; however, higher-order schemes improved the shedding amplitude. For pulsatile inflow, the first-order scheme was found to be in better qualitative and quantitative agreement with previous investigations. Careful scrutiny of the findings revealed that higher-order schemes implemented in the code produced too much dispersion error for applications in the present study. In addition, non-Newtonian viscosity did not affect the overall flow structure, although it caused significant shear-thinning for both types of inflow. Furthermore, numerical simulations of laminar Newtonian and non-Newtonian blood flow across 2D and 3D models of a Björk-Shiley tilting-disk mechanical heart valve in the aortic position with sinuses of valsalva (valve replacements are more common in the aortic position) under steady and physiological pulsatile inflows were performed to investigate the three dimensional effects as well as the influences of pulsatile waveform and non-Newtonian nature of blood on the overall flow structure. The results of grid and time-step independency tests as well as validation with previous investigations were satisfactory. Obvious differences of various flow parameters between 2D and 3D analyses clarified noticeable breakup of symmetry by the third dimension. Similar to the 2D case in a straight channel, it was observed that the non-Newtonian viscosity did not affect significantly the maximal velocity components and maximal vorticity magnitude, although it caused substantial shear-thinning and altered the overall flow structure, particularly during the regurgitation phase. It was found that three-dimensional effects were much smaller for the non-Newtonian viscosity model than the Newtonian model at all time levels. Indeed, the non-Newtonian behavior resulted in a “less complex vortical flow” and consequently “less probability of blood cell damage” compared to the Newtonian fluid. Since the longer the blood cells are trapped in the vortices, the more chance of hemodynamic damage, it is essential to consider the non-Newtonian behavior of blood in the analysis.

Table of Content
Thesis (M.S.)--Wichita State University, College of Engineering, Dept. of Aerospace Engineering
"July 2007."