Particle Image Velocimetry Used to Qualitatively Validate Computational Fluid Dynamic Simulations in an Oxygenator: A Proof of Concept.

Abstract

Computational fluid dynamics (CFD) is used to simulate blood flow inside the fiber bundles of oxygenators. The results are interpreted in terms of flow distribution, e.g., stagnation and shunt areas. However, experimental measurements that provide such information on the local flow between the fibers are missing. A transparent model of an oxygenator was built to perform particle image velocimetry (PIV), to perform the experimental validation. The similitude theory was used to adjust the size of the PIV model to the minimal resolution of the PIV system used (scale factor 3.3). A standard flow of 80 mL/min was simulated with CFD for the real oxygenator and the equivalent flow of 711 mL/min, according to the similitude theory, was investigated with PIV. CFD predicts the global size of stagnation and shunt areas well, but underestimates the streamline length and changes in velocities due to the meandering flow around the real fibers in the PIV model. Symmetrical CFD simulation cannot consider asymmetries in the flow, due to manufacturing-related asymmetries in the fiber bundle. PIV could be useful for validation of CFD simulations; measurement quality however must be improved for a quantitative validation of CFD results and the investigation of flow effects such as tortuosity and anisotropic flow behavior.