Developing blood-contacting medical devices like ventricular assist pumps demands rigorous testing. Not only must the device deliver the right flow and pressure, but it also needs to minimize red blood cell damage (hemolysis). Predicting hemolysis early in the design phase can save time and reduce costly prototype testing. One of the most reliable ways to check the accuracy of computational fluid dynamics (CFD) models for this purpose is to use standardized test cases. The FDA Nozzle Benchmark — introduced by the U.S. Food and Drug Administration — is the most widely used of these.

What Is the FDA Nozzle Benchmark?

This benchmark consists of a simple nozzle geometry with a narrow throat and two downstream configurations: a sudden expansion and a conical diffuser. Because its geometry is fully defined and experimental data are available for both flow and hemolysis, it has become a ‘wind tunnel’ for blood-pump CFD models. If a model can match the FDA’s nozzle results, it is much more trustworthy when applied to complex devices.

How We Conducted the Study

In our project, we reproduced both nozzle configurations in 3D and performed:
– Experimental testing with a blood-analog fluid (sodium iodide) to measure hydrodynamic parameters such as axial velocity and pressure drop.
– Hemolysis testing with bovine blood in the same setup to quantify the Modified Index of Hemolysis (MIH) under real biological conditions.
– CFD simulations using Reynolds-Averaged Navier–Stokes (RANS) equations and three turbulence models, including k-ε, k-ω, and k-ω SST, to replicate the same flow conditions numerically.

We also created several computational meshes ranging from 0.5 million to 2 million cells, with the final 1.5-million-cell grid selected for its balance of accuracy and computational cost.

Results: Hemolysis Prediction

Using the validated flow fields, we then calculated hemolysis indexes with three different definitions of the equivalent stress. We studied the effect of scaling extensional stresses by adding a coefficient C to the equivalent stress (following Faghih & Sharp). Key findings:
– The hemolysis index falls into a realistic range when 18 < C < 30 for the nozzle test cases.
– Higher C values predict significantly more hemolysis, which illustrates the sensitivity of the power-law model to extensional stress.
– Our CFD-based MIH values matched the FDA’s experimental hemolysis data across different flow conditions (CD6, SC5, SE6), demonstrating that our numerical methodology can reliably forecast hemolysis.

Why This Matters

By successfully reproducing both the hydrodynamic and hemolysis results of the FDA nozzle benchmark, we have validated our CFD approach for blood-pump design. This means we can confidently apply the same methodology to more complex geometries, such as full pump assemblies, to predict and reduce hemolysis before physical testing.

Looking Ahead

The FDA Nozzle Benchmark is just the first step. With a validated CFD model, designers and engineers can identify critical regions of high stress inside pumps, adjust blade or diffuser shapes, and test ‘what-if’ scenarios long before building prototypes. This reduces risk, accelerates development, and ultimately leads to safer, more efficient medical devices for patients.