Session T21: Experimental Techniques: Multiphase Flow

Non-Intrusive Calibration Target for Flow Velocimetry

A non-intrusive calibration technique for 3D velocity measurements is developed. Techniques such as Particle Image Velocimetry (PIV) require calibration to map particle images to physical space. Experimental setups that preclude access to the measurement volume make the use of a physical calibration target impossible. Here, we describe a non-intrusive, virtual calibration target, whereby a translating system of two lasers is used to create a calibration point array with precisely known point locations. This low-cost tabletop-sized system is agnostic to the index of refraction in the measurement domain, and the size of the virtual calibration target can be scaled arbitrarily. Performance of this technique is tested for both planar and stereo PIV.   

Unbalanced Optimal Transport for Particle Tracking in PTV

Non-invasive flow measurement techniques, such as particle tracking velocimetry, resolve 3D velocity fields by pairing tracer particle positions in successive time steps. These trajectories are crucial for evaluating physical quantities like vorticity, shear stress, pressure, and coherent structures. However, reliable track estimation is challenging due to measurement noise caused by high particle density, particle image overlap, and falsely reconstructed 3D positions. To overcome this, we employ a probabilistic approach that combines reconstructed 3D position uncertainty with unbalanced optimal transport theory, yielding a robust particle tracking method. Our algorithm utilizes Bayesian reconstruction, producing Gaussian posterior distributions of particle positions, allowing for uncertainty quantification. This method accounts for epistemic uncertainty, unlike a standard PTV process that deterministically reconstructs particle positions. Subsequently, we optimize a transport plan by moving all particle position distributions between time frames to achieve effective particle tracking. We validate our method using synthetic datasets, compare the results with FlowFit, and demonstrate the performance for an experimental aneurysm flow.   

Lagrangian Coherent Structure Detection using Natural Tracers in Particle-Laden Surface Flows

Surface flows exist in a variety contexts, with modern examples including the motion of drifting buoys, ice floes, and plastic pollution in the ocean, and the advection of debris in rivers. An open problem in the study of such flows is the characterization of coherent particle behavior without direct access to underlying velocity fields or a dense distribution of tracers. This work proposes the finite-time Q criterion as an objective metric for detecting Lagrangian coherent structures and develops a trajectory-based algorithm for its sparse approximation. The tool will be demonstrated on analytical and experimental data for flows representative of real-world applications.   

Rapid meso-scale 3D PIV and PTV with plenoptic macro photography

Describing fluid flows and particle motion at mesoscale with multi-camera setups presents numerous challenges, including limited illumination, constrained space, shallow depth of field, and particular calibration requirements. We propose a novel optical train design with a feature-based calibration method tailored for commercial macro lenses. By incorporating a microlens array (MLA) into the intermediate plane between the macro lens and camera sensor, we achieve effortless calibration of a singular camera cage system without any additional operations on the camera system, effectively minimizing experimental noise. We conducted a ray tracing-based simulation employing a realistic macro lens model to assess the uncertainty associated with GPU-accelerated ray intersection and cloud point classification methods. We performed a Lagrangian particle tracking around a jet flow as a proof-of-concept.   

Experimental study on wake characterization of tidal stream generators with multiple deployments

This study visualized the flow speed distribution generated between two tidal stream power generators. A vortex with regularity was generated in the perpendicular direction of the blade rotation direction. It was observed that this could affect up to 10 to 12 times of blade diameter. On the other hand, in the case of the support structure and foundation, the vortex strength was lowered by more than 60%, and it was confirmed that both units had the same effect. In the case of multiple turbines, we chose a row and triangular arrangement. Using the shape parameter, we could see the resulting Gaussian distribution and boundary layer downstream of the tidal turbine. In the future, these results will be used to develop underwater robots for operation and maintenance.   

3D Particle Tracking Velocimetry of Radial Outflow Between Two Parallel Plates

A neutrally buoyant and contactless grooming device utilizes confined radial outflow of water at a constant flow rate to detach biofouling organisms using peak wall shear in addition to securely levitating parallel to the hull at some equilibrium gap height, where the overall force generated by the fluid field is zero. Particle Tracking Velocimetry (PTV) measurements were taken on the flow field at a stable equilibrium gap configuration. This Lagrangian measurement approach uses an iterative particle reconstruction technique in combination with high seeding density to reconstruct a 4D (x, y, z, t) flow field. This 4D reconstruction allows the pressure field to be reconstructed using the Navier-Stokes equations. Various experiments have been conducted on confined radial outflow but PTV measurements are presented here for the first time and was used to validate a Computational Fluid Dynamics (CFD) model.   

High-resolution volumetric flow measurements in a transverse shear layer between two angled streams, using Colored Tomographic Particle Shadow Velocimetry.

We measure the full volumetric velocity field in a transverse shear layer between two angled streams.  The flow is set up in two in-house designed rectangular 3D-printed channels, having the same flow rate driven by independent centrifugal pumps, but angled in the horizontal direction and separated by a splitter plate.  Polyamide particles are fed into the flow through holes embedded near the edge of the splitter plate.  We use four high-resolution 12k (Blackmagic) video cameras from different angles to capture particle shadows using the Colored Tomographic Particle Shadow Velocimetry technique [1][2].  By pulsing a set of three differently colored LEDs, red – green – blue, during a single frame exposure, we can triple the base recording frame rate without significantly sacrificing spatial sensor resolution by encoding the temporal information in the color of the particle shadows. Therefore, we can increase our base frame rate from 24 fps to 72 fps. The timing between pulses is controlled with a digital delay generator. We preprocess the recorded images to separate the three different color fields by using hue values masks and minimizing noise. The resulting images are then processed with commercial software to obtain the full 3D-3C flow velocimetry with LaVision Shake-The-box and Tomographic-PIV algorithms.   

Preliminary experimental study of liquid droplets explosively induced

The explosive release of liquids is a subject studied in the field of defense, in particular for fuel air charges [1,2], but also for civilian applications such as fighting forest fires [3].

During the detonation of a charge in contact with a liquid, a cloud of droplets expands very quickly. It comes from the growth of instabilities at the interface with the detonation products and the liquid : this phenomenon, which is quite difficult to map, still remains a challenge today.

A way to investigate it is to access to physical quantities of interest which are the size and the concentration of the droplets in the cloud. There is no off-the-shelf device for accessing these measurements. This is why a study has been carried out for several years at CEA GRAMAT to test different measurement methods

The preliminary approach consisted in studying low droplet density cloud where classical optical method was relevant.

For that a campaign during which an explosive charge (1.3 g eq TNT) was immersed in a cylindrical container filled with 1 kg of water. A high-speed camera fitted with a Cassegrain-type telescope was implemented and associated with a Fresnel lens powered by a pulsed coherent light source.

The aim was to film by ombroscopy in a control volume of 3 cm³ on the outskirts of the cloud in order to record the shadows of the droplets. An image processing algorithm specially designed for this purpose was applied and then made it possible to discriminate between droplets outside the area of interest and then to measure their size.

The results of this campaign provided very interesting data for the continuation of the study, attesting feasibility of a measurement by transmission in the near infrared for denser clouds.

[1] D. R. GARDNER, «Near-field dispersal modeling for liquid fuel-air explosives,» Sandia National Laboratory, Sandia, 1990.

[2] S. SINGH et V. SINGH, «Extended Near-Field Modelling and Droplet Size Distribution for Fuel-Air Explosive Warhead,» Defence Science Journal, vol. 51, n°13, pp. 303-314, 2001.

[3] K. STEFANSKI, «Explosive Formation and Spreading of Water-Spray Cloud – Experimental Development and Model Analyses,» Central European Journal of Energetic Materials,, vol. 6, n°13-4, pp. 291-302, 2009.