a thorough review of the development of PIV.
became PIV. PIV is distinguished from LSV and
This very brief history of PIV glosses over its
PTV in the concentration of seed particles it
many intricate aspects, but it emphasizes an
requires (a much greater concentration is needed
important point, namely, that PIV is essentially a
for LSV, but far less for PTV) and, relatedly, in the
pattern-recognition procedure. Changes in seed-
method used to interpret or interrogate (the tech-
particle position create Young's fringe patterns that
nical term used for this purpose) images.
can be interpreted to yield velocities in so-called
Recent developments in instrumentation,
interrogation spots, or subareas, of flow fields. The
together with computer hardware and software,
interrogation methods currently in use are prima-
have made PIV practicable for obtaining accurate,
rily statistical. They infer the proper pairings of
detailed measurements of instantaneous and
particles and measure the average displacement
averaged velocity fields for diverse situations of
of particles in or between interrogation spots. This
flow and fluid-transport processes. The technique
process entails either auto- or cross-correlation
is also capable of yielding distributions of flow and
analysis to establish the most probable magnitude
particle velocities at sections across flows and
and direction of particle displacements within the
determining time-histories of velocities at points
interrogation spot. The correlation algorithm used
within flows. With appropriate particle seeding,
PIV also can be used to obtain data on variations
for the present application is illustrated and
explained in Figure 6.
in density and temperature within flows. The
Within an interrogation area, an interrogation
interested reader might refer to Adrian (1991) for
Search region
Interrogation area
t = t0 + dt
t = t0
Set an interrogation area
aij at P on the first image
Choose an interrogation area bij in the
search region on the second image
and calculate cross-section coefficient Rab;
find maximum Rab and
→
the displacement vector PQmax
=
Estimate sub-pixel movement
Calculate velocity vector at P:
r →
U = PQmax / dt
Post-processing to correct erroneous vectors
End
Figure 6. Correlation algorithm used for determining ice velocities.
9
Previous Page
