Massively Scalable Mixture Model for Small-scale Sand Ripples

Abstract

Award Number: N0001410WX20018 http://www7400.nrlssc.navy.mil/index.htm LONG-TERM GOALSOur long-term goal is to develop a modeling framework to predict sediment transport, the evolution of seafloor roughness, and acoustic propagation through the seafloor in the nearshore and littoral battlespace environment. OBJECTIVESOur primary objective is to develop and validate a massively parallel version of an existing mixture model, SedMix3D (Penko et al., in press), for simulating small-scale ripple dynamics in shallow littoral environments. Applicability of the serial version of SedMix3D is severely hampered by physical limitations (memory and CPU speed) of typical desktop workstations. The scalable version of SedMix3D developed here will be able to simulate prototype size domains as found in the center of a laboratory U-tube (up to 1 m in length), for example. The parallel version of SedMix3D is a powerful research tool that will be used to study the details of small-scale sand ripple dynamics including (1) the effects of suspended sediment concentration on turbulence modulation, (2) the dynamics of ripple transitions from 2D to 3D (and back to 2D) under changing forcing conditions, and (3) the role of terminations and bifurcations on ripple migration and growth rates. APPROACHSedMix3D solves the unfiltered Navier-Stokes equations for the fluid-sediment mixture with the addition of a sediment flux equation that considers the balance between gravitational settling, advection, and diffusion. The capability of SedMix3D to simulate small-scale sand ripple dynamics has been illustrated both qualitatively and quantitatively (Penko and Slinn, 2006;Penko et al., 2008;