Hemodynamic Design Optimization of a Blood-wetted Medical Device associated to treatment oCardiovascular Diseases applying Computational Fluid Dynamics: Case study of Kazakhstan

Abstract

This project aims to the improvement and validation of a multiphase fluid dynamics numerical model of blood flow to provide more accurate modeling of blood damage occurring within blood-wetted medical devices. Heart diseases are responsible for a significant portion of deaths worldwide and particularly in Kazakhstan, where blood-wetted medical devices are increasing in their use as a vital tool in the treatment of these diseases, supporting patient’s lives or as a bridge to needed transplants. However, these devices can introduce problems associated with biocompatibility of the materials or due to prolonged large shear stresses acting on blood cells. Within the family of blood-wetted devices used for cardiovascular diseases, the Ventricular Assist Devices (VAD) are of profound consideration, especially nowadays in Kazakhstan. Computational Fluid Dynamics (CFD) is proven to be a very useful and promising tool to perform blood flow simulations, as it is less expensive and time-saving compared with in-vitro experiments with real blood. Nevertheless, there is still lack of agreement between scientists worldwide about accurate modeling blood damage using CFD tools. This project aims to recover a current state-of-the-art multiphase CFD model of blood and tackle some of its potential weaknesses to improve its predictability characteristics. In particular, an Eulerian-Eulerian multiphase model, developed by the advisor of this investigation and his graduate student team, with moderate accuracy on prediction of blood damage, will be subject to modifications to include a new sub-model based on the principles of Granular Kinetic Theory (GKT) aiming to a better capture of blood cells segregation, which is blamed as the main cause of inaccuracies of that model. The existing model improved the blood damage prediction from a predecessor homogeneous model by just incorporating a particle-like nature to red blood cells (RBC) and platelets (PLT) transported in the blood stream. However, at the beginning of this investigation there was still large space to improve the damage prediction. The major inaccuracy of the base model was hypothesized as its incapability of predicting the correct peak of concentration of platelets that is expected to occur very near the vessel wall, also called the Fahraeus-Lindqvist (F-L) effect. Indeed the base model designed to simulate a three-phase plasma-RBC-PLT fluid flow, considering interfacial interaction between PLT and RBC only with plasma in a single-particle-interaction approach, was capable to predict a peak of platelet concentration near the wall, but with intensity much lower than expected.