THE EFFECT OF SAMPLE SIZE , TURBULENCE INTENSITY AND THE VELOCITY FIELD ON THE EXPERIMENTAL ACCURACY OF ENSEMBLE AVERAGED PIV MEASUREMENTS

A statistical analysis of the PIV data, obtained inside the near wake of a staggered 2-row array of circular cylinders, is performed for the investigation of the effect of sample size, turbulence intensity and the velocity field on the ensemble-averaged data. The PIV measurements are performed at two diameters downstream inside the wake. The ensemble averaged velocity field is obtained using 700 vector maps. The experiments are conducted for a Reynolds number of 10000 based on the inlet velocity and the cylinder diameter. For statistical analysis purposes, 2750 vector maps are collected separately. Variations of the statistical means and variances with sample size for the two components of the velocity vector are obtained for a line of measuring points along y axis at the mid-section of the domain. The turbulence intensity levels vary between 30-70 % along this line. The x and y components of the velocity vector vary between 0.88-5.4 m/s and 0-0.6 m/s, respectively. Using the total number of 2750 samples, 100 randomly selected statistically independent averages of 5, 10, 25, 50, 100, 250, 500, 750 and 1000 ensembles are calculated for each measuring point. The scattering around the mean value for each sample size depends on both the local turbulence intensity level and the local magnitude of the velocity. It is concluded that the errors in the ensemble averaged data increase not only with the increasing turbulence intensity levels but with the decreasing velocity magnitude as well.