Affiliations: Dept. of System Design Engineering, Keio University,
3-14-1 Hiyoshi, Kohoku-ku, Yokohama 223-8522, Japan | Dept. of Aerospace Engineering & Mechanics,
University of Minnesota, 110 Union St., SE, Minneapolis, MN 55455, USA
Abstract: A stereo PIV (SPIV) acquisition and analysis system was developed to
measure three velocity components in planar flow fields. The analysis software
is based on a third order mapping function method. The system was calibrated by
imaging a square grid in three measurement planes with two Kodak Megaplus
cameras oriented at 30 · to the bisector between them. The camera images
were dewarped into real coordinates by employing a set of transform matrices
computed for each calibration plane. Bias and rms errors were determined by
comparing displacements measured directly with displacements estimated from the
dewarping and recombination algorithm. The bias errors in the directions
parallel with the measurement plane were negligible while the bias in the z
direction was about 0.6 pixel. The rms errors, 0.2-0.3 pixels, were largest in
the z direction. These errors were thought to result from limitations in the
calibration method. The SPIV system was tested in a two-dimensional oblique jet
with Reynolds number of 1800. The three dimensional results were taken in a
vertical (x, y) plane parallel with the jet span. The SPIV results were
compared with LDV data and two-dimensional PIV data obtained in a vertical (y,
z) plane of the same jet. The SPIV measurements yielded accurate values for the
in-plane mean and rms velocity components. The measured out-of-plane mean
component was underestimated due to the bias error mentioned above. The rms
component was accurate in part of the field but overestimated in another part
due to local variations in rms error. It is expected that in the future, the
out-of-plane errors can be minimized by improving the calibration and
transformation procedures.
