Project Name

Project (long title)Vorticity Annihilition at high Reynolds Numbers
Campaign Title (name data folder)17ANNI
Lead AuthorJochen KRIEGSEIS
ContributorFrieder KAISER, Joel Sommeria, Samuel Viboud
Date Campaign Start09/01/2017
Date Campaign End27/01/2017

0 - Publications, reports from the project:

Preliminary studies were published at the 18th International Symposium on Application of Laser and Imaging Techniques to Fluid Mechanics in Lisbon, Portugal (2016): Kaiser F, Wahl T, Gatti D, Rival DE and Kriegseis J “Vorticity propagation for spin-up and spin-down in a rotating tank” 18th International Symposium on the Application of Laser and Imaging Techniques to Fluid Mechanics, Lisbon, Portugal (2016)

Furthermore, an abstract was submitted for the 12th International Symposium on Particle Image Velocimetry in Busan, Korea (2017): Kaiser F, von der Burg M, Viboud S, Sommeria J, Rival DE, Kriegseis J, "High Reynolds Number Measurements of Vorticity Generation and Annihilation with Rapidly Changing Boundary Conditions", 12th International Symposium on Particle Image Velocimetry in Busan, Korea (2017)

1 - Objectives

The objectives of this study focus on the influence of centrifugal forces, Coriolis forces and pressure gradients onto instabilities in boundary layers and on turbulence in general. The project aims at understanding how rapid vorticity annihilation takes place across a sharp vortical interface over a wide range of Reynolds numbers. In order to exclude any influences of (convective) length scales from the flow scenarios under consideration, all investigations center around boundary layer development at the outer wall of rotating facilities. To achieve different initial conditions the rotating tanks were either spun-up from rest ($\Omega_0=0$) or spun-down from a solid body rotation (SBR).

The experiments performed during 17ANNI at the Coriolis platform are covering the high Reynolds-Number ($Re=\Omega R^2/\nu$) regime and are complemented by small scale experiments at Karlsruhe Institute of Technology (KIT) and direct numerical simulations (DNS). This allows the investigation of the occurring structures from $Re=3000$ (DNS) up to $Re \approx 4 000 000$ (17ANNI).

During the different stages of the boundary layer formation multiple flow features and their dependence on the initial conditions are of interest

  • Transition: Onset of the transition, instability mechanism responsible for the transition (shear instability or Goertler instability, respectively). Shape, size and orientation of the first vortices occurring during transition.
  • Boundary layer thickness: Speed of boundary layer growth
  • Turbulence: After transition to turbulence, the turbulent statistics, its anisotropy, potential large scale structures and the decay of turbulence is of interest.
  • Spatially averaged azimuthal velocity profile: Preliminary studies showed a region with no vorticity in axial direction $<\omega_z>=0$ during spin-down. It is examined if this still holds for the experiments performed during 17ANNI.

As only the side wall boundary layer is of interest, bottom wall effects are unwanted and should to be reduced to a minimum. However, due to Ekman pumping, additional convection is expected after $\Omega t = 1$ (Duck, 2001). To reduce this influence some of the performed experiments during 17ANNI used a saturated salt water layer at the bottom of the CORIOLIS platform.

2 - Experimental setup

2.1 General description

The Coriolis platform was used without any inner installations. Optical access was possible through glass windows at the lateral walls and the bottom wall as well as through the free surface of the fluid. Two laser sheets were introduced through different lateral windows. One horizontal sheet allowed measurements in the $R-\phi$-plane while a vertical sheet was used to perform measurements in the $R-z$-plane. Multiple cameras were used to capture large parts of the flow in high resolution.

2.2 Definition of the coordinate system

$x$ radial horizontal, origin at the inner wall of the tank, radius R=6.5 m.

$y$ horizontal perpendicular to x, directed opposite to the tank rotation (such that x,y,z is a direct basis). Origin on the reference radial line.

$z$ vertical upward, origin at position of the laser sheet, 61 cm from tank bottom.

2.3 Relevant fixed parameters:

Notation DefinitionValueremarks
$H$total water height at rest$H$=100 cm .....
$R$tank radius$R$=650 cmaccuracy +-1mm. However: Windows in side walls lead to +3mm
$h_p$measurement plane$h_p=61cm$Level was chosen to avoid free surface and bottom wall effects as good as possible

2.4 Definition of the variable control parameters

Notation DefinitionUnitRemarks
$\Omega_0$initial rotationrpm
$D\Omega$change of rotation, >0 for spin-up, <0 for spin-downrpm
$Dta$acceleration times
$h_s$salt heightcm
$T$temperature of water (for viscosity $\nu$)$!^0$C
$Dt_{Dalsa}$Delay of data acquisition for Falcon,PCO1,PCO2,PCOpixelfly
$Dt_{Miro}$Delay of data acquisition for Phantom Miro 12475&12491
$Dt_{HS}$Delay of data acquisition for PCO.hs1200

2.5 Definition of the relevant non-dimensional numbers

3 Instrumentation and data acquisition

3.1 Instruments

Orgasol particles (diam. 30 micrometers) used as PIV tracer.

Continuous Nd:YAG Laser (Spectra-Physics Millennia - 25 Watt) used for horizontal sheet, positioned at a lateral window. Vertical laser sheet pulsed laser ( 'Evergreen 70 mJ) on another vertical window.

In total 8 cameras were used. In the following and in the Data structure the listed abbreviations are used:

  • PCO1: PCO.Edge 5.5
  • PCO2: PCO.Edge 5.5
  • Falcon(Dalsa): Dalsa Falcon2 4M
  • PCO_HS: PCO.hs1200
  • MIRO_12475: Phantom Miro 310 (Top Camera of stereo setup)
  • MIRO_12491: Phantom Miro 310 (Bottom Camera of stereo setup)
  • PCO_px1: PCO.pixelfly usb
  • PCO_px2: PCO.pixelfly usb

For the majority of the experiments (EXP1-EXP14 and EXP27-EXP35) 4 cameras were positioned for a top view along a radial line: PCO1, PCO2, Falcon(Dalsa), PCO_HS. 2 more cameras (MIRO) were placed on the side for stereoscopic PIV. 2 cameras pixelfly for stereo at a different radial location, using a vertical radial laser sheet.

For some experiments (EXP14-EXP26) PCO1 was mounted on the bottom of the tank. This avoided free surface effects, which introduced errors in experiments with large $\Delta \Omega$.

Different objective lenses were used for the different cameras:

  • PCO1: Samyang ED AS UMC 35mm (EXP1-EXP13,EXP27-EXP35), Samyang 20mm (EXP14-EXP26)
  • PCO2: Samyang ED AS UMC 35mm
  • Falcon(Dalsa): f=28mm (name?)
  • PCO_HS: f=50mm (name?)
  • MIRO_12475: f=60mm Macro lense (name?)
  • MIRO_12491: f=60mm Macro lense (name?)
  • PCO_px1: Nikon N 1.4 f50
  • PCO_px2: Nikon N 1.4 f50

This leads to the following field of views at full resolution:

  • PCO1: 21cm x 25cm (EXP1-EXP13,EXP27-EXP35), 40cm x 50cm (EXP14-EXP26)
  • PCO2: 21cm x 25cm
  • Falcon(Dalsa): 25cm x 35cm
  • PCO_HS: 20cm x 25 cm
  • MIRO_12475: 6cm x 13cm
  • MIRO_12491: 6cm x 13cm
  • PCO_px1: 6cm x 8cm
  • PCO_px2: 6cm x 8cm

3.2 Definition of time origin and instrument synchronisation

The time origin is set to the time, where the actual acceleration of the rotating platform started. Simultaneously to the onset of the tanks acceleration the software RDvision was triggered. RDvision triggered PCO1, PCO2 and Falcon for all experiments. The first pulse of the RDvision system also triggered two ILA-sychronizers. One of those synchronizers (Sync1) was controlling the PCO.pixelfly system. The other syncronizer (Sync2) controlled the PCO_HS.

The synchronization and trigger of the MIRO system was varied multiple times during the experiments. While the first configuration worked fine, it was only possible to manually delay the onset of the acquisition. To overcome this issue the system was changed multiple times.

  • EXP1-EXP13: RDivision triggered the MIRO. For 2 experiments the onset of Miro acquisition was delayed manually.
  • EXP14-EXP18: Sync2 gave a 25Hz signal to the MIROs, which triggered bursts of 5 pictures. Overexposure of some frames lead to the decision to change the system.
  • EXP19-EXP22: Sync2 synchronized the MIROs. An exact internal delay was able to delay the onset of acquisition. However, problems occurred with Miro_12475 (overframed).
  • EXP23-EXP35: Sync2 triggered the acquisition of the MIRO with an exact delay (between 5s and 15s). An MIRO internal Master-Slave synchronization was used to overcome the problems of EXP19-EXP22.

Each image series and corresponding PIV is labelled by its time starting at t=0 for the first image of the series.

A time delay of this origin with respect to the start of the tank acceleration is given by Dt_Dalsa, Dt_Miro, and Dt_HS specified in the table of the experiments.

3.3 Requested final output and statistics

4 - Methods of calibration and data processing:

4.1- Calibration

Each of the cameras for top view PCO1, PCO2, Falcon(Dalsa), PCO_HS are calibrated with a linear transform between phys and image coordinates (rescaling, rotation, translation, no projection deformation). This is done in three steps:

  1. A grid is used for scaling
  2. the image of a radial rope is used to control the origin y=0 (rotation and translation of the reference points coordinates). Since the rope position is not very precise, we adjust the position with the array of dots used for the MIRO camera: the marked dot is taken at x=13.5 cm, y=0.
  3. The image of a ruler is used to adjust the x origin with x=0 at the outer wall of the tank.

All calibrations are done in water. Images with reduced width are made from the calibration images (using the function merge_proj with transform ima_crop in uvmat), to get calibration for these cases. For PCO1 and 2 the image is cropped on each side keeping the middle band, while the upper part is kept for Falcon and PCO_HS.

The PCO1 has been used at the top, then in the bottom (EXP14-26), then again at the top, so three sets of calibration are needed. Note that the image is reversed in x when the image is viewed from bottom. This is taken into account by flipping the calibration points in x. The absolute position in x has been obtained by looking at the laser sheet edge visible in EXP24 (its width is equal to the window width 19 cm at x=0). The image edge is in fact close to x=0.

The PCO2 in principle did not move during the whole campaign. Calibration image at different dates show the same scaling and x position but shifts with respect to the rope by typically 1 cm. But the position of the rope is not known with a better precision. So we keep the same calibration for PCO2 for the whole series, as determined from Calib_24-01_top. Its position in y has been precisely adjusted so that the ruler coincide with PCO1, which itself has y=0 (and x=13.5) on the marked dot visible in Calib_24-01_top-evening/PCO1/Calib_PCO1_forMIRO. Application of the calibration of 24/01 gives excellent result for the calibration image of 12/01, which confirms that PCO2 did not move.

A similar thing is done for Falcon but the coincidence in y with PCO2 is not so clear, there are different images with different rulers! Looking at the times of file creation, the coincidence is done by Calib-water1. There is also some discrepancy with the ref image of 12/01, but the dates has been lost, so it may not be reliable.

For MIRO, the array of dots is used, positioned at the level of the laser sheet, considered z=0 (50 cm ??? above the bottom). The upper face of the dotted plate is at z=0.23 cm (seen by the upper camera MIRO_12471) while the lower face is at z=-0.23 cm (seen by the lower camera MIRO_12491).

All final reference .xml files are stored in 0_REF/XmlCalibFinal, classified by cameras, band selection, and range of experiments (distinguishing 1-26 and 27-end).

5 - Organization of data files:

All data related to the project are in Coriolis /.fsnet/project/coriolis/2014/'project name'

  • 0_DOC: miscellaneous documentation and reports
  • 0_MATLAB_FCT: specific matlab functions
  • 0_PHOTOS: photos of set-up
  • 0_REF_FILES: files of general use (calibration data, grids ...)
  • 0_RESULTS: processed data (time series, statistics...) and figures.
  • EXP01, EXP02, folder for each experiment with names given in the table below.

For each experiment we have six image folders:

  • Falcon: image series labelled by a number, or by two numbers (bursts)
  • PCO1.png: image series labelled by a number, or by two numbers (bursts)
  • PCO2.png: image series labelled by a number, or by two numbers (bursts)
  • PCO_HS.png: image series labelled by a number, and an index j=1,2 in each image pair
  • MIRO_12475: images in a single binary file .cine
  • MIRO_12491: images in a single binary file .cine

An xml file Falcon.xml ... is associated to each folder. It determines the image timing (except for MIRO for which the time is included in the file .cine), and the geometric calibration.

PIV data are stored in netcdf files (extension .nc) in folders derived from the image folder by the extention .civ, for instance Falcon.civ. ... They are stored as pixel displacement, and translation to physical velocities is done thanks to the xml files.

The final velocity fields are interpolated on a regular cartesian grid in phys space, concneting the data of different cameras. Each component is stored as a 2D matrix in a netcdf file. The files are in a folder named Falcon.civ-PCO1.png.civ-PCO1.png.civ marking the source of the data.

6 - Table of experiments

List of parameter, Param1... , denoted by names defined in section 2.4.

EXP0213/010.5-0.50000NaNdata useless, experiment was not in SBR
EXP0517/011-10020017.3manual Dt_miro
EXP0817/010.00.50-0017.0Burst mode: PCO1/2, Falcon, Miro
EXP0918/010.5-0.50010016.7manual Dt_miro
EXP1319/0100.25000016.4Miros in burst mode
EXP1419/0101000016.2Miros in burst mode
EXP1520/011-1000016.0Miros in burst mode
EXP1620/010.75-0.75000016.0Miros in burst mode
EXP1720/0100.75000015.9Miros in burst mode
EXP1820/0101000015.8Miros in burst mode