Water flow through a converging-diverging glass nozzle experiences a pressure drop and its velocity increases as it flows through the converging section. For an inviscid fluid, the pressure minimum occurs at the nozzle throat, where the cross-sectional area is minimum. If the minimum pressure is below the water vapor pressure, cavitation may occur. The actual minimum pressure through a converging-diverging nozzle depends on many factors and may not occur at the nozzle throat. Additionally, fluid through the nozzle may be driven into the metastable region and subsequently cavitate at a lower pressure than the vapor pressure. All of these factors combine to create a complex and unsteady flow pattern.
The precise conditions leading to the onset of cavitation in water flowing in a converging-diverging nozzle are not well understood. Utilization of a clear glass converging-diverging nozzle enabled Particle Image Velocimetry (PIV) measurements of the velocity vector field inside the nozzle without significantly promoting premature cavitation formation. Glass spheres of 10 μm diameter were selected as seed particles for use in the PIV measurements. These seed particles did not significantly affect the formation (or onset) of cavitation in the nozzle; however, larger seed particles (120 μm diameter) provided nucleation sights and promoted cavitation prematurely. The seed particles were injected into the flow significantly upstream from the nozzle to prevent disrupting the flow entering the nozzle. High seed density was needed to supply enough seed particles to interrogate small regions near the nozzle wall; however, high seed density could also cause speckling and reduce the ability to produce meaningful PIV measurements. A Nd:YAG laser provided illumination of the seed particles in the nozzle. Laser reflections off of the nozzle exterior had to be minimized to avoid saturating the PIV camera. A polarizing filter was installed on the camera to reduce reflections. An enclosure that surrounded the nozzle was also designed and utilized. The enclosure was filled with water to reduce laser reflections off of the nozzle exterior wall. The time elapsed between frames had to be adjusted for each section of the nozzle interrogated with PIV. For accurate velocity measurements, particles needed to travel at least two particles diameters but less than 25% of each interrogation cell. The large variation in velocities present in the nozzle prevented one time interval from satisfying the seed particles displacement requirements. The time interval between frames had to be tailored to each section of the nozzle, depending upon the range of velocities seen in that section. Detailed measurement of the velocity profile near the nozzle throat required precise control over all timing parameters and pushed the available hardware to its smallest possible time interval. Detailed PIV measurements near the wall in regions of recirculation and at the cavitation front required the use of a long-distance microscope. This limited the field of view and necessitated a high seed particle density, which presented problems due to the lack of control over the flow of the seed particles in the near wall region.
PIV allowed for the measurement of the velocity vector field inside a converging-diverging nozzle without disrupting the flow. These measurements provided detailed velocity and flow pattern information throughout the nozzle, particularly in the regions near the cavitation front where boundary layer separation was observed along with regions of recirculating flow. These detailed velocity profiles were compiled to present a complete PIV analysis of the converging-diverging glass nozzle. Measurements of the velocity field near cavitation onset allowed for a better understanding of the conditions triggering cavitation and the degree to which the water flow was able to be driven into the metastable region.