This paper describes a first effort to investigate the feasibility of droplet size and shape characterization by direct laser sheet imaging using time resolved Digital Particle Image Velocimetry. A 60-degree conical, high-pressure spray generated a poly-dispersed droplet distribution. Measurements were preformed for seven planes parallel to the spray axis, and separated by 4mm. A CMOS camera recorded the DPIV images at sampling rate of 10 KHz. Advanced image processing techniques were employed to identify the droplets and individually resolve their velocity using a hybrid cross-correlation particle-tracking algorithm. Subsequently, the size distribution of each droplet was quantified using geometric optics theory to convert the droplet image information to the true droplet size. Finally, the entire volume of the spray velocity and size distributions was reconstructed in a time-averaged sense. The droplet sizes from our direct imaging DPIV system were validated using a Phase Doppler Particle Analyzer (PDPA). The calculated sizes from the direct imaging methodology were found to agree with the measured PDPA results for droplets images larger than the diffraction limited diameter. Resolution limitations introduced inaccuracy for smaller droplets. In addition, the shedding frequency of the spray ligament was observed to be on the order of 1KHz, demonstrating the feasibility of using a high speed, direct imaging system in the characterization of unsteady, liquid sheet breakup properties. This preliminary effort illustrates the potential of performing global time resolved velocity and size measurements using a simple DPIV configuration based on CMOS imaging technology.

This paper represents a continuation of our effort to develop a velocity evaluation scheme optimized to resolve multiphase flows. An improved adaptive hybrid scheme that integrates the dynamically adaptive cross-correlation method with a particle tracking velocimetry algorithm is developed, presented and evaluated in this paper. A detailed description of the methodology, error analysis using Monte-Carlo simulations and elaborate comparisons with established schemes and robust commercial packages are presented. Improvements were guided towards increased accuracy for resolving vortical and poly-dispersed multi-phase flows. We introduce a novel iterative scheme that localizes the cross-correlation. We incorporate state of the art elaborate image processing techniques that allow increased particle densities. A new particle pairing method based on an adaptive cross-correlation masking is introduced. Finally, a refined gaussian estimation scheme that involves only four non-saturated pixels for the particle centroid detection is proposed. Overall, the dynamically adaptive hybrid velocity evaluation scheme presented here allows superior resolution of high velocity gradients, minimizes the loss of the rotational motion of the particles, and eliminates the spatial averaging effects inherent from the cross-correlation.

A unique, super spatio-temporal resolution Digital Particle Image Velocimetry (DPIV) system for the analysis of time-dependent multiphase flows has been developed. The system delivers a sampling frequency between 1KHz and 10KHz, with continuous total acquisition time up to 4 secs and resolution 1Kx1K pixels down to 256×256 pixels. The hardware is integrated with sophisticated image processing algorithms that allow direct image segmentation in order to resolve the multiple phases present in the flow and provides quantitative information about the shape and size of droplets or bubbles present. Finally, the in-plane velocities are measured by a super-resolution, dynamically-adaptive cross-correlation algorithm which is coupled with a particle-tracking scheme. Each individual phase present in the flow is resolved with mean spatial resolution in the order of 3–4 pixels, and accuracy in the order of 0.01–0.1 pixels, while the spatial averaging effects of cross correlation are eliminated.

Digital Particle Image Velocimetry (DPIV) currently represents the state of the art for non-invasive global flow velocity measurements. The instantaneous velocities are determined by cross-correlating patterns of particles between consecutive images, thus mapping in space and time the velocity distribution for thousands of points in the flow field simultaneously.