Degree Type

Thesis

Date of Award

2016

Degree Name

Master of Science

Department

Mechanical Engineering

Major

Mechanical Engineering

First Advisor

James Michael

Second Advisor

Terrence Meyer

Abstract

Particle image velocimetry (PIV) is robust means of making accurate velocity measurements in dynamic fluid environments. However, limitations in temporal resolution and dynamic velocity range have always resulted from insufficient data acquisition systems. As laser technology continues to advance, so do the potential capabilities of PIV. Today, burst-mode laser systems are capable of MHz repetition rates. These advancements in temporal resolution allow new insight into the dynamics of highly turbulent flows. Optimization of the pulse-burst laser for PIV will allow for unparalleled measurement capabilities.

In this work, a method for high temporal resolution, high dynamic velocity range PIV is proposed by means of multi-frame processing of particle images. This was accomplished using a triple-pulsed, burst-mode laser system and high-speed, CMOS sensor. Pulse triplets, with time spacings tailored to the flow velocity, were repeated at 10 kHz which resulted in three particle images per instance. Correlation between each combination of these three images produced three separate vector fields. Vectors were then selected based on comparisons of correlation plane data. This method of vector selection was tested using synthetically generated particle image sequences. Resulting composite vector fields increased the dynamic velocity range while simultaneously decreasing processing based uncertainty when compared to standard, two-frame PIV processing.

An experimental study of a fully developed turbulent jet over a range of Reynolds numbers was performed. PIV data was collected both free jet and cylindrical bluff-body test cases. Results of multi-frame processing supported the synthetic results, showing decreased uncertainty over the entire velocity range of the flow when compared to standard processing.

Copyright Owner

Mark Jeffrey Johnson

Language

en

File Format

application/pdf

File Size

128 pages

Share

COinS