Experimental and Numerical Study of an Optically Accessible Cavitating Venturi
Abstract
Cavitating venturis are flow metering devices that exploit an inherently unstable phenomenon, which is challenging to model accurately. To improve understanding of venturi performance, it is essential to visualize the cavitation structure. In a graduate-level course titled “Propulsion Design Build Test” at Purdue University, students were tasked with designing optically accessible components for a hypergolic rocket engine experiment. We designed a pressure vessel rated for 500 psig and incorporated a transparent body manufactured via stereolithography using Formlabs V4.1 resin. This printed body was secured between the vessel’s endcaps and featured an axisymmetric conical converging-diverging flow path. We analyzed the venturi’s surface roughness using optical profilometry and found it significantly higher than typical machining roughness for metal, which influences cavitation behavior. To capture the cavitation phenomenon, we employed high-speed cameras with microsecond exposure, achieving high-resolution visualization across a range of inlet pressures. We also developed computational fluid dynamics (CFD) models to simulate cavitation behavior and compared these predictions to experimental results. Standard cavitation models provided estimates of cavitation length and discharge coefficients, enabling direct comparison with measured data. Analysis of high-speed video revealed nucleation points at the venturi throat, which initiated cavitation at lower pressures than required elsewhere in the body. We further applied doublet image subtraction to spatially characterize regions of net vaporization and condensation under steady-state conditions. This approach allowed qualitative assessment of cavitation stability both at the throat and throughout the cavitation cloud length.
Fields of Study & Subjects
Cavitation Phenomena in Pumps
Hydraulic flow and structures
Flow Measurement and Analysis
Publication details
- Journal / source
- AIAA SciTech Forum
- Publication date
- 8 January 2026