Estimation and reduction of peak-locking errors in PIV measurements

Abstract

In PIV, the systematic tendency of the measured sub-pixel displacement to be biased towards the integral pixel values is called peak-locking. This occurs when the particle image diameter is less than a pixel. The bias error causes inaccuracy in the measured PIV data which does not reduces with increase in the sample size. Recently developed LaVision’s optical diffuser was investigated to determine the reduction in peak-locking. The point-spread-function width of diffuser was examined to calculate the change in the shape and size of the point source under the influence of different parameters. Planar-PIV experiment were carried out in uniform, low-speed and high-speed flow conditions to analyse the effectiveness of the optical diffuser in reducing the bias error and change in the random error. Also, a comparative assessment was done between the use of conventional defocusing and optical diffusers during image acquisition. The use of optical diffusers reduces the bias error and random error by a factor of three. The reduction in the measurement error is similar to the best defocused optical position of the lens which is very difficult to determine. Additionally, an experimental analysis was done with three different camera-lens combination to determine the best relative aperture size for keeping the measurement error minimal for the large scale 2D PIV. With the help of optical diffusers, experimentalists can have more accurate PIV measurements which would lead to more realistic capturing of the flow phenomenon. With the use of optical diffusers, it would also help the CFD and theoretical experts to compare their predictions with better experimental benchmark results.