Bubble Plumes and their use for Sound Mitigation
Simon Beelen is a PhD student in the department Physics of Fluids. (Co)Promotors are prof.dr. D. Lohse and dr. D.J. Krug from the faculty of Science & Technology.
This thesis aims at understanding the sound mitigating properties of bubble curtains. We start our investigation by developing a measurement system in chapter 1 for the acoustically relevant hydrodynamic properties. This system combines the use of an array of electrical probes with an underwater camera. The reliable and cheap electrical probes distinguish air based on the difference in conductance between air and water. The underwater camera images are analysed using an image analysis algorithm allowing for the deconstruction of overlapping bubbles. The combination of these two methods allows us to calibrate our measurements system in-situ. We describe a correction for the void fraction measured by the electrical probes based on the length of the exposed tip of the needle sensor. Using a similar correction and a statistical approach we are able to deduce the bubble size distribution from a set of two closely spaced needles with a small difference in length. We show that the bubble size distribution measured by the probes and the camera correspond closely if the calibration of the probes uses information regarding the shape of the bubbles measured by the camera. The developed measurement technique is consequently applied to bubble curtains generated by different manifolds. With these measurements we noticed the need to correct them for temporal changes in the bubble curtain. Our post processing steps are adapted to do just that. We present some initial results of this measurement process, out of which the most striking observation is the maintained influence of the manifold on the development of the bubble curtain. The angle at which a bubble curtain widens, remains larger when a porous hose is employed as compared to an array of nozzles.
In chapter 2 we employ our measurement system to a planar bubble plume originating from an array of nozzles. The bubble plume is placed in a large basin (31x40 m wide and 10 m deep), yet, using the newly developed measurement system, we are still able to provide detailed descriptions of the local void fraction and bubble size distribution. Based on this gathered data we successfully calibrated an integral plume model, initially developed for round bubble plumes. We adapted this model to be applicable for planar plumes originating from nozzles. We opted for this model since it is able to describe the void fraction distribution but importantly also the local bubble size distribution. The latter being instrumental for future use in acoustical models. As input to our model we use an entrainment relation consistent with our data. This entrainment relation depends on the airflow rate, and was so far missing for planar bubble plumes. For the calibration of the model we require a high bubble slip velocity, far exceeding the rising velocity of a single bubble. We argue for the effect of the collective behaviour of the bubbles. The bubbles rise in cloud like structures possibly increasing their apparent slip velocity. From our data we found a bubble size distribution that is independent of the supplied airflow rate and translated that into our modelling.
Finally, in chapter 3 we measure the performance of the bubble curtain(s) by the insertion loss. Firstly, we notice that the performance of a single bubble curtain hardly depends on the airflow rate. For that reason, we propose splitting up the airflow rate into two distinct bubble curtains to increase the performance of the supplied air. The experiments indeed reveal a significant increase in insertion loss when two manifolds are employed. The probable reason for the drastic performance increase is that the reflective properties of the bubble curtain hardly changes when the airflow rate is halved, essentially resulting in a “free" extra reflective boundary. We also modelled this effect using the commonly used equivalent fluid method combined with finite element simulations. Qualitatively, the model follows the measurements, however, quantitatively there is a relatively large discrepancy. We also measured the insertion loss for the bubble curtain with different shots of an airgun. This revealed that variations in the insertion loss can largely be attributed to the temporal changes of the local air distribution in the bubble curtain.