Predicting turbulent multi-phase flows with high fidelity: a physics-based approach
Abstract
We propose an integrated computational, theoretical and experimental program todevelop physical modeling and computational methods" for multiphase flows. Theproposed research spans the multiple scales from the dynamics/thermodynamics of watermolecules to the macroscopic behavior of turbulent bubbly and cavitating flows. Ouroverall vision is to develop a new predictive capability for turbul"ent cavitating flows, thatsignificantly reduces/eliminates empiricism and uncertainty. Specifically, we aim todevelop next-generat""ion homogenous mixture and Lagrangian physical models, as wellas a multi-scale hybrid simulation methodology that will be applicabl"e across the variousregimes of bubbly flows and cavitation.The proposal is submitted to ONR in response to FY17 MURI Topic #14 tit"led, ~High-Fidelity Simulation Methodologies for Multi-Phase Flows~. The proposal is titled,~Predicting turbulent multi-phase flow""s with high fidelity: a physics-based approach~,and requests $4,500,000 (3 years) + $3,000,000 (2 years) = $7,500,000 (5 years).Ou""r work plan considers the micro-physics that occur at the various length and timescales,uses a combination of simulation and experi""ment to study them, then uses formalmachine-learning and model-reduction techniques to develop next-generation reducedordermodels," and a framework to integrate them across the various scales to develop anovel multi-scale hybrid methodology.We will use molecular dynamics and Monte-Carlo simulations to study nucleation andphase change; direct numerical simulation using highly innovative lev"el set and volumeof fluid approaches to study coalescence, break-up and wall contact; machine-learningmethodologies to develop red"uced-order models; dynamic Lagrangian models thataccount for bubble-bubble interactions and collapse; a pdf-based polydispersemeth"odology for bubble transport; stability and sensitivity analyses for the homogeneousmixture equations; extended mixture models, inn"ovative numerical methods and subgridmodels for cavitating flows; and four unique experiments that provide highly resolvedunsteady data on the micro-physics.Prof. Krishnan Mahesh of the University of Minnesota will be the Principal Investigatorand will serve a"s the technical point of contact for the project. A diverse team ofresearchers from the U. Minnesota (Mahesh, Ilja Siepmann), Calte""ch (Tim Colonius), UCSanta Barbara (Fredric Gibou), U. Iowa (Pablo Carrica), U. Michigan (Steve Ceccio),MIT (Themis Sapsis) and Jo"hns Hopkins (Joe Katz) will work together on the program.Current Reynolds-averaged methodologies are unable to reliably predict phe"nomena suchas cavitation inception, sheet to cloud transition, bubble entrainment, and bubble-inducedacoustics. The proposed work"" will develop high-fidelity multiphase models, subgridmodels and numerical methods which will allow applications involving these ph"enomenato be predicted with unprecedented fidelity. The proposed multi-scale hybridmethodology will provide a new conceptual framework for multiphase flow simulations.The experimental and DNS data will yield novel physical insight into flows of Navyrelevance." The proposed work will significantly enhance DoD predictive capability, andopen the door to next-generation vessels that are uniqu"ely optimized for the multi-phaseenvironment they operate in.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Jul 07, 2017
- Source ID
- N000141712676
Entities
People
- Krishnan Mahesh
Organizations
- Office of Naval Research
- Regents of the University of Minnesota
- United States Navy