Rickard Bensow - Chalmersweb.student.chalmers.se/groups/ofw5/Presentations/... · =0.71 and !=4.455. Bensow and Liefvendahl !7" previously re-ported a validation study for this propeller

Shipping and Marine TechnologyHydrodynamics

Assessing Cavitation Nuisance using LES in OpenFOAMRickard Bensow

[email protected]

Wednesday, June 23, 2010


Ways towards erosive cavitation

1. Traveling cavity 2. Shedding 3. Upstream moving collapse

4. Upstream desinence 5. Secondary cavities



NACA0015

• 6° angle of attack

• Re = 1.08 x 106

• σ = 1.0

• Cavitation mechanism:– Shed cloud

– Upstream moving collapse

– Secondary cavitation



• LES– Incompressible

– Implicit subgrid modeling

–Wall modeling based on Spalding

• VoF– Source terms for

condensation and vaporization

– Consistent with LES

!t "v( ) +# $ "v% v( ) = &#p +# $ S & " v% v & v% v( )( )# $ v = 1

"v& 1

"l( ) !m p,' ; pv( )

!t' +# $ 'v( ) = !m p,' ; pv( )"v

Modeling approach

• Two phase mixture with finite rate mass transfer



Numerics

• Time– backward

– Co ~0.4

• Convective terms– limitedLinear ~0.2

• Vapor fraction transport

– standard solve– no compression

– limitedLinear ~0.5

• PISO– Partially implicit treatment of pressure dependent mass transfer

terms



ValidationHemispherical head

• Full 3D domain

• Transient

• 2.2 & 4.5 M cells

• Kunz & Sauer



NACA0015 specifics

• ~2.7 M cells

• Hexahedral– conforming

• Sauer



P1477 Propeller

Frame 420 Frame 430 Frame 440 Frame 450



Frame rate: 30000 fps, J =1.0, σ =2.1σ



NACA0015

• Upstream moving collapse and generation of 2ry cavitation



Summary

• Capture many mechanisms initiating for cavitation nuisance– Internal jets

– Secondary cavitation

– Leading edge desinence

– Cavity flow interaction

• Good, but what about...– Balance between them?

• Influencing cavity extent, shedding frequencies, etc.

– Compressibility effects?• Acoustic interaction

– Non-condensable gases?

– Effect of under-resolved structures?

– Collapse pressure pulses

– ...



In the pipe

• Compressible low mach solver for cavitating flows– Thermodynamic equilibrium model

• New PISO for incompressible flows– Advection-diffusion pressure equation

• Adaptivity– Load balancing challenge

• Large relative motions– GGI, Overset or deforming?



Recent papers - more details

Rickard E. Bensow

Göran Bark

Shipping and Marine Technology,Chalmers University of Technology,

412 96 Gothenburg, Sweden

Implicit LES Predictions of theCavitating Flow on a PropellerWe describe an approach to simulate dynamic cavitation behavior based on large eddysimulation of the governing flow, using an implicit approach for the subgrid terms to-gether with a wall model and a single fluid, two-phase mixture description of the cavi-tation combined with a finite rate mass transfer model. The pressure-velocity coupling ishandled using a PISO algorithm with a modified pressure equation for improved stabilitywhen the mass transfer terms are active. The computational model is first applied to apropeller flow in homogeneous inflow in both wetted and cavitating conditions and thentested in an artificial wake condition yielding a dynamic cavitation behavior. Althoughthe predicted cavity extent shows discrepancy with the experimental data, the most im-portant cavitation mechanisms are present in the simulation, including internal jets andleading edge desinence. Based on the ability of the model to predict these mechanisms,we believe that numerical assessment of the risk of cavitation nuisance, such as erosionor noise, is tangible in the near future. !DOI: 10.1115/1.4001342"

1 IntroductionCavitation is responsible for most major constraints in propeller

design, related to noise, vibration, and erosion but is a complexphenomenon not yet neither reliably assessable nor fully under-stood. Standard design tools typically include potential flow solv-ers, lifting surface or boundary element approaches, with stricttheoretical limits on cavitation modeling that only in the hands ofan experienced designer may give satisfactory propeller designs.This situation, together with fairly short design cycles, makes itdifficult to advance the design toward smaller safety margins tocavitation nuisance and thus hinder efficiency improvements.There is thus a need both for better understanding of the physicalmechanisms leading to cavitation nuisance as well as for im-proved prediction and analysis tools.

Experimental observations alone can show many of the phe-nomena occurring but suffers from limitations in the measurementtechniques. One example is in measuring reentrant jets and inter-nal flow, flow features hidden for optical measurement techniquesby the cavity itself but often important to study in the develop-ment of erosive cavitation !1". The access to the complete flowfield through a numerical simulation would thus be a welcomecomplement to experimental data. The numerical simulation ofcavitation does, however, include many complications since someof the physics of the fluid and the mass transfer are unknown.Moreover, the cavitation dynamics is governed by medium tosmall flow scales, both in time and space !1", necessitating largecomputational grids and small time steps.

In this paper, we study the feasibility to use incompressiblelarge eddy simulation #LES$ techniques, using an implicit model-ing approach for the subgrid term, for the prediction of cavitationon a propeller, based on considering the flow as a single fluid,two-phase mixture. A model transport equation for the local vol-ume fraction of vapor is solved together with the LES equationsand a finite rate mass transfer model is used for the vaporizationand condensation processes.

The propeller we use is the four-bladed INSEAN E779A. Al-though an old design, the experimental database is extensive, in-cluding both PIV and LDV wake measurements !2–4" and cavi-tation observation in homogeneous !5" and inhomogeneous !6"flow conditions, which makes it a good validation case. The con-

ditions studied computationally in this work is #i$ uniform inflowat an advance number J=0.71 and cavitation number !=1.76 and#ii$ nonuniform inflow, formed by an artificial wake inflow, at J=0.71 and !=4.455. Bensow and Liefvendahl !7" previously re-ported a validation study for this propeller in noncavitating con-ditions, using the same LES technique as in this work.

We start by describing the modeling approaches in some detailand discuss the basic validation of the computational model bothfor noncavitating as well as for cavitating flows. We then continueby summarizing the results of Bensow and Liefvendahl !7" for thenoncavitating conditions at J=0.88 as a validation for wetted pro-peller flow. Next, follows an analysis of the simulations results forthe cavitating flow, starting in uniform inflow followed by thepropeller in the artificial wake. Finally, we summarize our find-ings and give some outlooks for future work.

2 Modeling TechniquesThe simulations presented in this paper are based on applying

an incompressible LES approach, based on cell-centered, unstruc-tured finite volume #FV$ technique, solved using a segregatedpressure implicit with splitting of operators #PISO$ algorithm forthe velocity-pressure coupling with a modified pressure equationto increase the stability for cavitation simulations. To model thecavitating flow, we adopt a two-phase mixture assumption by in-troducing the vapor volume fraction and solving an additionaltransport equation, incorporating finite rate mass transfer modelsfor the vaporization and condensation processes. Before givingsome details of the different modeling components listed above,we start by discussing and motivating our choices. References aregiven to recent work that complement or support our ideas but arenot intended as a review of the current status of the research field.For comparisons between different mass transfer modeling ap-proaches based on a transport equation we refer to, e.g., Refs. !8"or 9, where the latter also gives a brief outlook on other ap-proaches.

In LES, the large, energy-containing structures are resolved onthe computational grid, whereas the smaller, more isotropic, sub-grid structures are modeled; this separation of scales within theflow is accomplished by a #implicit$ low-pass filtering of theNavier–Stokes equations, see, e.g., the book by Sagaut !10". Incontrast with RANS approaches #e.g., Wilcox !11"$, which arebased on the average flow description, LES naturally and consis-tently allows for medium- to small-scale, transient flow structures.When simulating unsteady, cavitating flows, we believe this is animportant property in order to be able to capture the mechanisms

Contributed by the Fluids Engineering Division of ASME for publication in theJOURNAL OF FLUIDS ENGINEERING. Manuscript received July 1, 2009; final manuscriptreceived February 9, 2010; published online April 13, 2010. Assoc. Editor: OlivierCoutier-Delgosha.

Journal of Fluids Engineering APRIL 2010, Vol. 132 / 041302-1Copyright © 2010 by ASME

Downloaded 13 Apr 2010 to 129.16.81.136. Redistribution subject to ASME license or copyright; see http://www.asme.org/terms/Terms_Use.cfm



!

Thanks for your attention!

NaiXian [email protected]

Göran [email protected]

Rickard [email protected]


Rickard Bensow - Chalmersweb.student.chalmers.se/groups/ofw5/Presentations/... · =0.71 and !=4.455. Bensow and Liefvendahl !7" previously re-ported a validation study for this propeller

Documents