Acoustical effects of bubble assemblages, and considerations for accurately measuring bubble size distributions

Open Access
Mcelhone, Dale Ian
Graduate Program:
Master of Science
Document Type:
Master Thesis
Date of Defense:
November 08, 2013
Committee Members:
  • Richard Lee Culver, Thesis Advisor
  • mcelhone
  • bubbles
  • culver
  • acoustics
  • bubble
  • scatter
  • scattering
  • backscatter
  • backscattering
  • attenuation
Assemblages of bubbles, specifically those found in ship wakes, have been found to scatter mid-frequency acoustics (1-5 kHz) more efficiently than individual-bubble models suggest. Acoustical bubble dynamics are thoroughly examined in the literature, but few controlled laboratory experiments have measured the below-resonance scattering from bubble ensembles. In order to predict, measure, and fully understand the scattering properties of bubbly media, the bubble size distribution and its spatial distribution must be known with great detail. Mathematical models capture the physical behavior of individual bubbles, and various theories describe how bubbles will collectively respond to an incident acoustic wave. An "effective medium" theory is used as the foundation of a Matlab model for predicting mid-frequency backscatter from bubbly wakes and plumes, which is discussed in this thesis. In order to experimentally verify effective medium theory, a bubble distribution must first be created that has very few large bubbles (with a radius greater than 300 microns). Plumes of bubbles were generated first by a commercially available ceramic oxygen diffuser, then by a Venturi water oxygenation system. The resulting bubble distributions are presented in this thesis. Unfortunately, the large quantities of bubbles with radii greater than 300 microns made it impractical to measure backscattering in the Applied Research Laboratory Reverberation Tank at a frequency that was below the resonance of all individual bubbles. The experimental and computational methods used in estimating bubble distributions are analyzed in this thesis and suggestions are made for increasing the accuracy of bubble distribution estimates.