The picture of the PIV setup is shown in Figure 3.9 and the photograph of important hardware of PIV is shown in Figure 3.10. In the present experiments, PIV measurements were carried out at selected planes perpendicular and parallel to the cylinder axis. A double pulsed Nd:YAG laser of wavelength and with a maximum repetition rate of per laser head was used. The light sheet had a maximum scan area of 10 10 . The sheet thickness was about 1 mm to minimize the effect of the out-of-plane velocity component. The assembly of Peltier-cooled 12 bit CCD camera and frame grabber with a frame speed of was used for acquisition of PIV images. Figure 3.11 shows geometric diagram of PIV measurements. A cross section of the flow is illuminated with a thin light sheet, and the tracer particles in the light sheet are projected onto a recording medium (CCD) in the image plane of a lens as shown in Figure 3.11.The intensity of the light sheet thickness is assumed to changes only in the direction. The magnification of particle image depends upon the position of the imaging lens. The CCD consisted of an array of 1280 1024 pixels. A Nikon 50 mm manual lens with was attached to the CCD camera for covering the field of interest. Both the camera and laser were synchronized with a synchronizer controlled by a dual processor PC. The field of view employed in the present set of PIV measurements was 40 mm by 35 mm. velocity vectors were calculated from particle traces by the adaptive cross-correlation method. The final interrogation size was 16 16 pixels starting from an initial size of 64 64. Thus, 5561 velocity vectors were obtained in the imaging area with a spatial resolution of 0:5 mm. Inconsistent velocity vectors were eliminated by local median filtering and subsequently replaced by interpolated data from adjacent vectors. The laser pulse width was 20 and the time delay between two successive pulses was varied from 40 to 200 depending on the fluid velocity (Keane and Adrian, 1990). The time-averaged velocity field was obtained by averaging a sequence of 200 velocity vector images, corresponding to a total time duration of 50 seconds. Laskin nozzles were used to produce seeding particles from corn oil. The mean diameter of oil particles was estimated to be 2 . Data generated from PIV carries superimposed noise. Noise is introduced during recording of PIV images (optical distortion, light sheet non-homogeneity, transfer function of the CCD, non-spherical particles, and speckle) and during data processing (peak fitting algorithm, image interpolation and peak deformation). The validation of the PIV technique was carried out by comparing velocities with pitot static tube and hotwire anemometry, as discussed in later sections.