3-d particle tracking velocimetry (PTV) is an established technique in the field of fluid mechanics to obtain 3-d velocity fields up to large Lagrangian trajectories (Figure). Such information helps to understand the development of flow instabilities in turbomachines. After examining turbulences (e.g. rotating stall) in the liquid phase, it is now important to understand the birth of instabilities in the gas phase. For this purpose a 3-d PTV system with a high temporal as well as spatial resolution is being developed based on a three-camera setup. This ensures the ability to investigate gas flows at relatively high speeds and involving small eddies. In that case it is clear that PTV can only be successful when there is a high concentration of seeded tracer particles.

Trajectories at winglet arrangement.

This point is known as a major issue for PTV, since the correspondence analysis between the different camera views becomes extremely difficult. To establish the spatial correspondence to determine 3d positions (by photogrammetric methods), each tracer has to be found in all pictures of the involved cameras during a timestep. This central problem can be partially solved for the correspondence both from a temporal and from a spatial point of view using coloured tracer particles. In that case, classifying in all images the initial dense cloud regarding to particle colours results in different particle subsets. Analysing these subsets for correspondences separately is much easier than analysing all particles as a whole because the number of ambiguities is reduced dramatically. Using colour classes diminishes the apparent particle density without inducing any a-priori restriction for the measuring accuracy. Only the tracer particles belonging to a specified colour group have to be recognized at successive timesteps.

Contact: Christian Bendicks , +49 391 67 11473

Videos

	Computed trajectories from winglet arrangement setup.
	Image postprocessing. There is an apperatus to create spiral shaped trajectories.		Reconstructed spiral shaped trajectories.