Results of the joint DLR-ONERA DGV measurement campaign in the F2 wind tunnel – Part I

Two Doppler global velocimetry systems were applied in parallel to assess their performance in wind tunnel environments. Both DGV systems were mounted on a common traverse surrounding the glass-walled 1.4 x 1.8 m2 test section of the ONERA F2 wind tunnel. The traverse normally supports a three-component forward-scatter laser Doppler velocimetry system. The reproducible vortical flow field generated by the blunt tip of an airfoil (300 mm chord) was chosen for this investigation and was precisely surveyed by LDA just prior to this investigation. Both DGV systems shared the same continuous wave laser light source (argon ion, 514 nm), laser frequency stabilization and fiber optic light sheet delivery system. The principle differences between the DGV implementations are with regard to the imaging configuration. The ONERA Configuration relied on a single camera view that observed three successively operated light sheets. The DLR configuration used three camera views simultaneously to observe a single light sheet using a four-branch fiber imaging bundle. In both configurations off-normal viewing arrangements were present, the DGV system of ONERA observed two light sheets in forward scattering and one in back-scattering mode which had a roughly ten times lower signal level. The DLR system had all viewing directions at a similar forward scattering angle, such that all views had approximately the same signal levels at the exit of the fiber imaging bundle. The closed-circuit wind tunnel was seeded globally with sub-micron oil droplets using an oil-based smoke generator. Stream-tube seeding was used in addition to improve the back-scatter signal levels. CCD camera exposures varied from 0.25−2.0 seconds in forward scattering views to 2−5 seconds in back scattering views. The subsequent processing of the multi-view DGV data resulted in improvements of DLR’s DGV software package to deal, for instance, with artifacts introduced by the image fiber bundles. The availability of reproducible reference data from LDA allows a further quantification of error sources in the DGV measurements. Overall the agreement between the DGV and the LDA measurements is on the order of 1 m/s for the DLR system. Residual bias is believed to be caused by scene illumination stemming from the luminous light sheet within the seeded flow which cannot be easily quantified in absence of seeding (i.e. by background image acquisition). Also the spatial averaging imposed by the optical system dampens out velocity gradients such that vortex peaks are underestimated in comparison to the ONERA DGV system which has a higher optical fidelity (see Part II). Nonetheless the DLR system was shown to acquire reliable 3-C velocimetry data using single recordings of the composite views. In principle the presented configuration would be well suited for pulsed DGV measurements using a single light sheet