Page 82 - Thermal Hydraulics Aspects of Liquid Metal Cooled Nuclear Reactors

Rod bundle and pool-type experiments in water serving liquid metal reactors 57

seeding particles dispersed in the flow and registered by a charged couple device
(CCD camera) in two subsequent images (or CMOS sensors for high-speed PIV mea-
surements). When one camera is used, only the in-plane displacements of the particles
within the measurement region of the laser sheet can be registered. With two cameras
oriented along different directions, also the out-of-plane displacements can be catched
and the system is referred to as stereoscopic PIV; this arrangement requires also the
Scheimpflug optical condition be fulfilled (the prolongation of image plane, imaging
optics plane, and object plane meet on the same line). An extensive review of PIV can
be found in Adrian and Westerweel (2011) and Raffel et al. (2007). The images are
divided into several interrogation windows; here the cross-correlation of the light
intensity is computed to detect the particle displacements, and thus the particle speed.
The final result will be a quantitative (and also qualitative) map of the instantaneous
velocity field; more images can be acquired in order to calculate the average values
and velocity root mean square. Fig. 3.1.3 illustrates the sketch of a typical PIV system.
The particle scattering condition can be described by the Mye’s theory: the amount
of light scattered forward is larger than what is reflected backward or laterally. Even
so the images are usually recorded at 90 degrees with respect to the laser direction,
rather than from a position aligned with the light source (forward scatter), because
the optical access to the measured region is often limited.
A particular case of PIV is the PTV. This is a low-image density PIV referred to as
low-particle number density PIV by Adrian (1991).

Illumination requirements
The two laser pulses have a short-time duration δt (even of the order of ns) and they are
delayed by a separation time Δt. In order to have a good cross-correlation analysis, δt
has to be short enough such that, in each photograph, the images of the moving par-
ticles appear as “frozen” rather than as streaks of light. This constraint is expressed by
the following condition:

ϕ
δt≪ p ,
uM
where ϕ p is the particle diameter in the image, u is the particle velocity, and M is the
magnification factor defined as the ratio between the object size in the image and its
real size. The acceptable particle displacement between the two images is also a crit-
ical parameter: 2–3 pixels displacement is commonly accepted as an optimum for the
maximum imaged displacement. With larger displacements the correlation calcula-
tion does not work well. The latter condition can be achieved by adjusting the sepa-
ration time Δt accordingly. The laser source can be continuous (CW) or pulsed; the
first type is usually cheaper and the separation time Δt is dictated by the closing fre-
quency of the camera shutter: it is suitable for PIV measurements in low-speed flows.
A pulsed laser source condenses the power in a single pulse being thus able to deliver
more power per unit of surface than the CW laser. Generally a pulsed light source
consists of two Nd:YAG lasers (neodymium-doped yttrium aluminum garnet) emit-
ting two pulses independently (dual-head system); 10 ns pulse duration and 10–30 Hz