Comparative Evaluation of Power-Law and Viscoelastic Hemolysis Models in the FDA Nozzle Benchmark Geometry

Abstract Purpose The mechanical damage of red blood cells (hemolysis) remains a key challenge in the design of blood-contacting medical devices. This study aims to compare the predictive performance of the empirical power-law model and the physiologically motivated viscoelastic model in the FDA nozzle benchmark geometry, with a focus on how model parameters and shear stress definitions influence hemolysis predictions. Methods Computational fluid dynamics (CFD) simulations were performed using a finite volume framework to resolve flow fields and shear stresses in the FDA nozzle. Both hemolysis models were implemented under identical geometry, mesh and boundary conditions. A parametric analysis was conducted by varying key model parameters, including the viscoelastic constants A , $$T_1$$ T 1 , $$T_2$$ T 2 , the shear stress threshold $$\tau _S$$ τ S , and the power-law parameters. Predictions were evaluated using both total shear stress and turbulent viscous shear stress (TVSS) formulations, and compared against experimental data. Conclusion The viscoelastic model demonstrated strong sensitivity to the relaxation time $$T_2$$ T 2 and to the choice of stress definition. A quasi-linear dependence on parameter A was identified, enabling calibration with experimental results. In contrast, the power-law model was less sensitive to stress definitions but strongly dependent on parameter fitting. The effect of $$\tau _S$$ τ S differed between the two models: it reduced hemolysis in the power-law case but increased it in the viscoelastic formulation. Background The viscoelastic model offers a more mechanistic description of hemolysis but requires careful parameter calibration and is highly sensitive to stress definitions. The power-law model remains computationally efficient and easier to implement, though empirical by nature. Future research should refine parameter estimation and extend validation to more complex flow regimes, considering feedback effects of hemolysis on viscosity. Ultimately, CFD prediction of hemolysis remains a challenging unsolved problem.