Advances in Time-Resolved Tomographic Particle Image Velocimetry

Abstract

This thesis details advanced developments in 3-D particle image velocimetry (PIV) based on the tomographic PIV technique, with an emphasis on time-resolved experiments. Tomographic PIV is a technique introduced in 2006 to measure the flow velocity in a three-dimensional volume. When measurements are performed at a rate high enough to sample the dynamical evolution of the flow, the measurement is considered time-resolved. The present work begins with a description of developments in tomographic PIV since its inception and outlines the working principles. A detailed explanation of the reconstruction and correlation procedures is given. The primary limitations of the technique are identified, and it is hypothesized that new analysis methods exploiting time-resolved image sequences can overcome these limitations. This research characterizes the baseline performance of tomographic PIV, explores this hypothesis for both reconstruction and correlation, and demonstrates a novel system for acquiring time-resolved data in high-speed flows. Even though tomographic PIV has been a topic of research for nearly a decade, its accuracy remains an active area of inquiry. Few studies exist which approach the topic in a parameteric manner using experiments and varying the primary variables including the number of cameras and seeding concentration. Chapter 3 presents a novel experimental assessment of tomographic PIV accuracy, enabled by a 12-camera system as a reference measurement. The reconstruction quality, signal-to-noise ratio, and variance are determined, and the velocity field accuracy is assessed. These analyses provide a set of guidelines for experimentalists in the set-up phase of a measurement campaign. Time-resolved tomographic reconstruction is the focus of chapter 4, where the sequential motion-tracking enhanced (SMTE) reconstruction technique is developed. SMTE is based on a time-marching predictor of the reconstructed intensity field, extending the concept of motion-based predictors from the motion-tracking enhanced (MTE) reconstruction of Novara et al. The novel application to time-resolved data yields significant improvements in reconstruction quality and velocity field accuracy with little or no increase in computational burden. The improvements also allow for operation at higher seeding densities or with fewer cameras. Improvements to the correlation procedure for time-resolved data are introduced in chapter 5. A novel evolution of the image deformation method is proposed to extend correlation-based measurement to account for non-linear fluid trajectories and incorporate information from multiple snapshots. The fluid trajectory correlation (FTC) allows the measurement interval to be enlarged compared to single-pair or ensemble-averaging approaches, and is particularly well-suited for tomographic PIV sequences where no out-of-plane motion allows fluid parcels to be tracked in 3D space over a number of frames. The application of the technique to both synthetic and experimental data sets shows an improvement in measurement dynamic range without compromising the measurement by truncation error. The robustness and accuracy afforded by the aforementioned algorithms are realized only for time-resolved data sets. The use of standard high-speed tomographic PIV equipment in high-speed flows is not possible due to the limited laser power and camera frame rate. In chapter 6 a burst-mode tomographic system able to acquire up 4 pulses within microseconds is developed and demonstrated for the measurement of the instantaneous velocity and material acceleration which is applicable for transonic flows to attempt non-intrusive measurements of unsteady pressure fluctuations. Furthermore, a connection is made to the FTC method of chapter 5 to show the synthesis of novel hardware and software to measure the material acceleration. The work concludes with a summary of the main findings and a perspective on future research directions potentially leading to a broader adoption of the time-resolved tomographic PIV technique.