Experimental validation of a system based on Particle Tracking Velocimetry (PTV) Federico Ibáñez 1 , Pablo González-Altozano 1* , María Gasque 2 , Pau Martí 1 Universitat Politècnica de València. Camino de Vera s/n. 46022 Valencia (España) 1 Dept. Ingeniería Rural y Agroalimentaria. 2 Dept. Física Aplicada * E-mail: pgaltozano@agf.upv.es Abstract This work describes the starting up of an experimental PTV system, a technique which allows the determination and characterization of the velocity field in a hydraulic device. It has been determined the number of images required to achieve the convergence of velocity to a mean value, a key issue to determine the minimum duration of the test. The curve of spectral density of energy vs frequency in a logarithmic scale has allowed the adjustment of the image acquisition frequency with the PTV system. The relative experimental error in the instant velocity detection was lower than 3-6% (p > 0.01) and lower than 2.01% (p > 0.01) in the mean velocity estimation. Key words: PTV, velocity field, instantaneous velocity vector, experimental validation. 1. Introduction The advances in instrumentation and computation of the last decades have contributed to the development of experimental techniques, allowing a more precise characterization of the hydrodynamic performance. These techniques provide complementary information to classic approaches, where head losses are assessed globally through the device using pressure and flow sensors (Palau-Salvador et al., 2006, 2008). Among these, it is important to enhance those based on snapshot of consecutive flow images, which allow the visualization of the velocity field (Adrian, 1991). Partickle Tracking Velocimetry, PTV, (Brevis et al. 2006, 2007) is a non-intrusive technique which allows instantly the determination of the velocity fields with high spatial resolution (Raffel et al., 1998; Bardera, 2007). In PTV tracer particles are added to the flow. A two-dimensional slice of the flow is illuminated by a laser sheet and the light scattered by the particles is recorded showing their paths. Particle snapshots are taken at a specific frequency, frames per second, FPS. The images are processed to digitize and identify each position of the particles, which are compared in successive images. Considering two images taken in a time interval, the velocity vector of the particles is determined from the distance travelled. A suitable choice of particle size, lighting parameters and camera-particle distance is crucial to get images which may allow a proper visualization of particles. The correspondence between the particles of two consecutive images is obtained from a correlation coefficient, calculated through the cross-correlation algorithm (Dill et al., 1995). The value of this coefficient is calculated by analyzing the distribution of certain particles taken as reference in the first image (in a region called window of interrogation), and possible candidates in the second image that has been moved so that the centroids of the potential pairs match. A comparison of these algorithms and other methods of correlation can be found in Ohmi and Li (2000). The particle tracking program provides an instantaneous velocity vector for particles from an initial image to another taken later. The PTV system uses a Lagrangian framework, and allows to characterize the velocity field based on the determination of the displacement of each tracer particle which is individually identified and tracked (Cenedese & Querzoli, 1997).