| |
By Yves-Marie Lefebvre and Philippe Corrignan, Sirehna, Nantes, France
View the pdf of this Supplement
Research in the marine industry usually requires instantaneous and time-averaged velocity data for complex flow conditions. Particle image velocimetry (PIV) is a widely used experimental technique for simultaneously collecting 2D sets of velocity components. It makes use of neutrally buoyant tracer particles uniformly distributed throughout the flow. Illuminated by a planar sheet of light and photographed repeatedly at normal incidence, the velocities are extracted from the displacements of the particles from one image to the next. At Sirehna, a new 3D PIV test facility has recently been developed in collaboration with Marin, the Maritime Research Institute in the Netherlands. The new system, which uses multiple cameras and light sheets, is capable of generating full, 3D data sets for the analysis of marine, automotive, and aeronautic applications.
Flow measurement at the leading edge of a propeller, made using Sirehna's 3D PIV system
Sirehna provides research and industrial applications in naval hydrodynamics, fluid mechanics, and structural dynamics through numerical simulation and testing. It operates test facilities in France and elsewhere in Europe for a wide range of naval applications. Sirehna has been using FLUENT for CFD simulation since 1990, and it has adopted the CFD postprocessor FIELDVIEW (from Intelligent Light), both to explore complex CFD results and to process PIV experimental data. Because 3D PIV data can easily require 200GB of memory, FIELDVIEW, with such 3D functionality as vortex core detection and pathline creation, helps make experimental results more easily understood. In particular, Sirehna is taking advantage of the unique capabilities of FIELDVIEW to compare unsteady CFD data and experimental results. The methodology was developed for 2D cases some years ago and was extended recently to handle 3D cases as well.
The vorticity difference (between experimental and CFD values) in the water underneath a rolling barge; the discrepancies (red and blue) correspond to a moving vortex with a changing orientation that was observed during the experiments, but could not be captured by the 2D techniques used
As an example, Sirehna was tasked with studying roll motion and mooring effects on oil production barges by comparing CFD results with experimental data. Three PIV cameras were mounted on the bottom of the barge and then submerged in a large test tank. Data from the three cameras were converted into three separate 2D, structured PLOT3D data sets. The dynamic mesh model in FLUENT was used for the 2D simulation. A user-defined function (UDF) was used to impose the same forced roll motion on the CFD model of the barge that was imposed during the experiment.
The 3D PIV equipment.
In order to do a visual comparison of the rocking motion of the barge in FLUENT and the three PIV cameras moving at the same time, an unsteady animation was created. The vorticity was computed from the unstructured FLUENT data and from the structured PIV data. Sirehna used FVX, the FIELDVIEW programming language, to numerically compare the structured and unstructured vorticity fields, and to compute the difference for each time step during the motion. The engineers expected that the animation would show little difference in the experimental and computed vorticity fields. They found, however, two regions where the difference was significant. The regions corresponded to a vortex, which in the experiments was observed to change its orientation as it moved along the length of the barge during the rolling motion. Because this phenomenon could neither be captured with a 2D CFD simulation nor a 2D PIV, a discrepancy resulted. While this simple
test case gave Sirehna's engineers a better understanding of and confidence in the new tools, it also illustrated the need to use 3D simulations combined with 3D PIV in the future.
|
|
|
