Home > Research > Underwater Noise
Underwater Noise
Turbulence, interacting with solid boundaries, generates noise. Underwater, sound is the only means of detection, and it propagates very efficiently in water, so for submarines, any source of noise is a huge problem. Turbulence also causes structural vibrations that generate additional noise. Engineers do not have reliable tools to predict noise generation due to turbulence; most of what we know is based on simplified formulations derived from empirical relationships. Unlike other applications of turbulent flow, studying noise requires full characterization of turbulence near the vicinity of the boundaries, precisely where most simplified models break down. Because of limitations in resolution, Direct Numerical Simulation, or DNS, (see the Computational Engineering area), the only accurate method available, cannot be used to compute the full turbulent flow around bodies in the foreseeable future. Consequently, the Navy has invested substantial effort to develop approximate techniques for modeling the turbulence and its dynamics in the vicinity of boundaries. Large Eddy Simulation (LES) is one such approach. In LES, the equations of motion are solved explicitly for all scales larger than some given threshold (the grid-scale). Motions smaller than these (the sub-grid scale) are parameterized by a set of models that depend on various simplifying assumptions about the small-scale dynamics. Profs. Meneveau and Katz are conducting 2D and 3D flow visualization experiments to modify and improve current LES modeling techniques.
Typical LES models treat turbulence as isotropic, meaning that statistical properties of the flow are equal in all directions. In the vicinity of a boundary, however, turbulence undergoes rapid straining and rotation. This in turn alters the physics of the turbulence, making it anisotropic. To account for this, modifications must be made to the sub-grid scale models used in LES. (See the Computational Engineering area, for details on LES methodology.)
To create a 3D velocity distribution of the flow in the vicinity of a boundary, Prof. Katz draws upon his experience in holography. In the laboratory, he “seeds” a turbulent flow with particles, and then records multiple-exposure holograms. The holograms are reconstructed and scanned, and he obtains the 3D velocity distribution from the displacement of the particles. The data are then spatially filtered, giving the filtered velocity field and the subgrid stresses. This three-dimensional version of the Particle Image Velocimetry technique (see Turbomachinery, above), known as Holographic PIV, gives instantaneous vector maps of 130 x 130 x 130 vectors—an unprecedented level of resolution. This highly resolved data set is the only one of its kind in the world, and promises to lead to very important advances in our understanding of turbulence. From the more practical viewpoint of the Navy, someday this may help solve some serious underwater noise problems.
Profs. Katz and Meneveau have joined forces with Professor Shiyi Chen to obtain a second grant, this time from the NSF, to develop a system that can generate instantaneous vector maps of 500 x 500 x 500 vectors. This research requires extensive (not to mention expensive) new equipment and will take the state of the art in flow visualization to a completely new level. Stay tuned!