Computational fluid dynamic investigation of blood flow through heart valve prostheses

Tools

Wilson, P., 1997. Computational fluid dynamic investigation of blood flow through heart valve prostheses. PhD, Nottingham Trent University.

Preview

Text
10290273.pdf - Published version
Download (36MB) | Preview

Official URL: http://gateway.proquest.com/openurl?url_ver=Z39.88...

Abstract

This thesis describes a computational fluid dynamic (CFD) analysis, and associated experimental validations, relating to a prototype conduit valve prosthesis under development at Nottingham Trent University. Current commercially available valved conduits are little more than short lengths of graft tubing into which are sewn conventional cardiac valve prostheses. These valves rely for their efficient closure, at least in part, on the vortex flow generated as the blood passes from one chamber of the heart to another (atrium to ventricle or ventricle to aortic root). They are therefore not ideally suited to controlling flow in a straight tube outside the heart and so the development programme was initiated to produce a purpose built mechanical valved conduit with the blood guided round a ball occluder, delivering not only superior flow characteristics but also the prospect of good haemodynamics.

The initial stages of the work described here concentrated on the comparison of steady non- Newtonian and Newtonian blood flow through the prosthesis at varying Reynolds numbers. The conclusion was that as differences between the two types of flow were small, neither could be shown to be more applicable to the analysis of the valve conduit. Therefore both flow situations were modelled in the final time dependent analysis to fully assess any differences between the two blood models under more realistic flow conditions.

The analysis was conducted using the CFD package PHOENICS. The model was required to simulate time varying inlet flow velocity, determine fluid forces acting on the occluder, predict occluder displacement and redefine the finite volume grid when necessary.

Experimental validation consisted of video recording the movement of the ball occluder in a model of a conduit valve under cyclic flow conditions to determine the occluder's position with respect to time. Comparison between the experimental and computational results showed that the overall performance of the CFD model was good but some detailed predictions were unreliable due to previously undetected errors in PHOENICS. Representation of field values using MPEG video bit streams proved to be a particularly efficient method of analysing the fluid properties over the entire flow cycle.

The main conclusions are:

The fact that only small differences were observed between the non-Newtonian and Newtonian blood models indicate that the use of a blood analogue is quite acceptable in experimentally determining the occluder's dynamic performance, but cannot fully predict the true haemolytic and thrombogenic potentials.

The haemolytic and thrombogenic potentials of the conduit were low because of the low shear stresses and shear rates which apply for most of the flow cycle; higher values occur, but only for a small period of time, during the initial opening phase of the occluder.

CFD is the only current method available to allow access to field values at any point and time within the flow cycle and therefore it provides a powerful tool in the pre-clinical trials of new cardiac prostheses.