Improving flow-induced hemolysis prediction models.

hemolysis;prediction models

Mohammad Mohaghegh Faghih

University:University of Louisville

Department:Mechanical Engineering

关键词: hemolysis;    prediction models;   

Others  :  https://ir.library.louisville.edu/cgi/viewcontent.cgi?article=4466&context=etd

美国|英语

来源: The Universite of Louisville's Institutional Repository

【 摘 要 】

Partial or complete failure of red blood cell membrane, also known as hemolysis, is a persistent issue with almost all blood contacting devices. Many experimental and theoretical contributions over the last few decades have increased insight into the mechanisms of mechanical hemolysis in both laminar and turbulent flow regimes, with the ultimate goal of developing a comprehensive, mechanistic and universal hemolysis prediction model. My research is broadly divided into two sections: theoretical/analytical/Computational Fluid Dynamics (CFD) analyses and experimental tests. The first part of my research revolved entirely around analyzing the simplest and most popular hemolysis model, commonly called as the power-law model. This model was developed only for laminar pure shear flow within a limited range of exposure time. Subsequently, modified versions of this model have been developed to be used for more complex flows. Many of these modified models assume that hemolysis scales with a resultant, scalar stress representing all components of the fluid stress tensor. The most common representative stress used in the power-law model is a von-Mises-like stress. However, using membrane tension models for pure shear and pure extension in both laminar and turbulent flows, for some simple example cases, we have shown that scalar stress alone is inadequate for scaling hemolysis. Alternatively, the rate of viscous energy dissipation rate has also been proposed as the parameter to scale hemolysis with. Applying the same order-of-magnitude estimate as vi mentioned above, we have found that dissipation rate even behaves worse than the resultant scalar stress for hemolysis prediction. It is therefore concluded that energy dissipation rate alone is also not sufficient to universally scale blood damage across complex flows. These show that a realistic model of hemolysis must take into account different responses of the viscoelastic cell membrane to different stress type. Various discretized version of the power-law model has also been introduced for post-processing of the CFD results. The power law can be either discretized in space, Eulerian treatment, or in time, Lagrangian treatment. Our study on the Eulerian approach revealed that the current equations used in the literature has a missing term, and thus incorrect. We also examined the mathematical stability of the discretized power-law model, and found that it may introduce significant error in red cell damage prediction for certain pathlines with specific stress history. Experimental results on deformation of red cell in pure shear flow is present for a relatively wide range of shear