A5.2 - Three-Dimensional, Three-Component Flow Field Measurements with a Single Camera

SENSOR+TEST Conferences 2011
2011-06-07 - 2011-06-09
Proceedings SENSOR 2011
A5 - Flow Sensors II
C. Cierpka, C. Kähler, M. Rossi, R. Segura - Universität der Bundeswehr München, Neubiberg (Germany)
118 - 122


In many flow devices complex three-dimensional velocity fields are present and need to be characterized for further optimization of the devices. Optical particle based techniques are very advanced and a reliable tool in fluid dynamics. 2D2C PIV (particle image velocimetry) can give velocity information of high spatial resolution of the 2D velocity vector in the light sheet plane. To measure the third velocity component, multi camera approaches, e.g., a stereoscopic or tomographic setup can be used. The major difficulty for multi camera techniques is the need for a very precise calibration and optical setup. Small errors in mapping the camera images to the position in real space can lead to large measurement uncertainties. Another problem arises by the fact, that in many flows, e.g. in micromechanical devices, engines or other internal flows, the optical access is limited.
To overcome both problems a 3D PTV (particle tracking velocimetry) technique was developed. The main idea is the coding of the depth position of small tracer particles by astigmatic image aberrations. This is accomplished by introducing a cylindrical lens into the observation optics. According to the particle’s depth position their images appear elliptically deformed. An advanced algorithm based on autocorrelation and 2D wavelet transform determines the particle’s 3D position in the volume.
For the calibration, an intrinsic calibration procedure, based on the particle images, was developed. One of the major advantaged of this procedure is, that the calibration can be perfomred prior to the experiment outside of the device (microchannel, combustion chamber, etc.) and no complex calibration routine is necessary. If the 3D position is known at two different time instants, the 3D velocity vector can be estimated.
Since just one camera is used, no mapping errors between different cameras will decrease the accuracy. Using fast cameras astigmatism PTV can be used to resolve particle trajectories in space and time. Due to the intrinsic calibration the system is able to handle unwanted image aberrations and can be used without expert knowledge.