Seeding optimization for instantaneous volumetric velocimetry. Application to a jet in crossflow

Abstract

Every volumetric velocimetry measurements based on tracer (particles, bubbles, etc.) detection can be strongly influenced by the optical screening phenomenon. It has to be taken into account when the the statistical properties associated to the performances of the particle detection and tracking algorithms are significantly affected. It leads to a maximum concentration of particles in the images thus limiting the final spatial resolution of the instantaneous three-dimensional three-components (3D3C) velocity fields. A volumetric velocimetry system based on Defocused Digital Particle Image Velocimetry (DDPIV), named Volumetric 3-Components Velocimetry (V3V), is used to show that above a critical visual concentration of particles in the images, the concentration and accuracy of the final instantaneous raw velocity vector field drop. The critical concentration depends on physical parameters as well as on the processing algorithms. Three distinct regimes are identified. In the first regime, the concentration is well adapted to volumetric velocimetry, the largest concentration being optimal to maximize the number of valid velocity vectors. In the second regime, the performances of the detection and tracking algorithms are no longer optimal. Finally, the third regime is strongly influenced by optical screening. A rigorous methodology is proposed to optimize volumetric measurements taking into account the system specifications (pixel size, focal length, system magnification, etc.). An optimal particle concentration of 4.5×10 −2 particles.pixel −1 (ppp) is found. A relationship between the imaged concentration and the optimal final spatial resolution for fully interpolated instantaneous velocity field is obtained.The present study can be used as a guideline to achieve the measurement of instantaneous three-dimensional velocity fields with a good spatial resolution. It is finally successfully applied to a complex 3D flow test case: a jet in cross-flow.

Conditional averaging on volumetric velocity fields for analysis of the pseudo-periodic organization of jet-in-crossflow vortices

“…Because most optical volumetric velocimetry techniques use tracers to measure the flow velocity field, the achievable tracer concentration is inherently limited by the optical screening of the tracers. This phenomenon induces an optimal resolution of the final velocity fields [2]. As a result, while it is theoretically always possible to increase the resolution of numerical velocity fields to observe smaller flow structures (for a computational cost), the resolution of instantaneous 3D velocity fields is intrinsically limited.…”

“…The geometrical and physical parameters of the experiment presented in this article are the following : the bulk channel velocity U inf = 9 cm.s −1 , the mean jet velocity over the jet exit V jet = 15 cm.s −1 , the velocity ratio V R = V jet /U inf = 1.67, the jet diameter d jet = 8 mm, the boundary-layer thickness δ = 15 mm, the crossflow Reynolds number Re cf = U inf · δ/ν = 1350, the jet Reynolds number Re jet = V jet ·d jet /ν = 1200. The flow is seeded with 50 µm particles, with a visual concentration 4.5 × 10 −2 particles per pixel [2]. The flow is illuminated through the upper wall and the particles are tracked in the volume using three cameras facing the side wall.…”

“…More details can be found in [2,3,7]. Using the work of Cambonie [2], the set-up was designed and the physical (particle concentration, measurement volume dimensions...) and numerical parameters were chosen to optimize the quality and resolution of the instantaneous velocity fields while avoiding the optical screening of the particles in the back of the measurement volume by the particles in the front layers. Indeed, a relationship exists between the imaged concentration and an optimal final spatial resolution of the interpolated instantaneous velocity fields.…”

“…9a,b for instance) have a resolution of 1 velocity vector/1.2 mm. This resolution of interpolation is the best one achievable on this setup taking into account the local particle concentration in the water, the screening effect of particles between each other, and the conditions to have adequately and quasi-entirely interpolated (>99%) velocity fields [2]. By keeping the same resolution (1.2 vec/mm), the voxels keep the same volume than the volume of the instantaneous voxels.…”

“…Nevertheless, the resolutions of the experimental instantaneous 3D velocity fields are still low (1 to 4 vectors per mm) in comparison of what can be achieved numerically. Indeed, the experimental resolutions depends strongly on the tracer concentration in the flow (bubbles, particles...), concentration which is itself limited by the visual screening phenomenon [2] which occurs between the tracers. As a result, it limits the size of the structures which can be observed.…”