Sensitivity analysis of FDA´s benchmark nozzle regarding in vitro imperfections - Do we need asymmetric CFD benchmarks?

Abstract

Modern technologies and methods such as computer simulation, so-called in silico methods, foster the development of medical devices. For accelerating the uptake of computer simulations and to increase credibility and reliability the U.S. Food and Drug Administration organized an inter-laboratory round robin study of a generic nozzle geometry. In preparation of own bench testing experiment using Particle Image Velocimetry, a custom made silicone nozzle was manufactured. By using in silico computational fluid dynamics method the influence of in vitro imperfections, such as inflow variations and geometrical deviations, on the flow field were evaluated. Based on literature the throat Reynolds number was varied Rethroat = 500 ± 50. It could be shown that the flow field errors resulted from variations of inlet conditions can be largely eliminated by normalizing if the Reynolds number is known. Furthermore, a symmetric imperfection of the silicone model within manufacturing tolerance does not affect the flow as much as an asymmetric failure such as an unintended curvature of the nozzle. In brief, we can conclude that geometrical imperfection of the reference experiment should be considered accordingly to in silico modelling. The question arises, if an asymmetric benchmark for biofluid analysis needs to be established. An eccentric nozzle benchmark could be a suitable case and will be further investigated.

Keywords

• computational fluid dynamics (cfd)
• fda nozzle
• geometrical imperfection
• asymmetry