Scribd Logo
Upload

Welcome to Scribd!

  • 92 views

    Moody Steel Pipes

    Uploaded by

    Stanislav Sviderek
    The document summarizes research on turbulence in pipes at high Reynolds numbers. Experiments were conducted using a specialized facility capable of achieving Reynolds numbers up to 35 million. The results provide new data in the smooth and rough pipe regimes, requiring updates to existing diagrams like Moody's. Computational simulations are also discussed as a promising tool for further studies, especially with emerging petascale resources, though challenges remain in simulating the highest Reynolds numbers and modeling roughness.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd

    Moody Steel Pipes

    Uploaded by

    Stanislav Sviderek
    92 views57 pages
    The document summarizes research on turbulence in pipes at high Reynolds numbers. Experiments were conducted using a specialized facility capable of achieving Reynolds numbers up to 35 million. The results provide new data in the smooth and rough pipe regimes, requiring updates to existing diagrams like Moody's. Computational simulations are also discussed as a promising tool for further studies, especially with emerging petascale resources, though challenges remain in simulating the highest Reynolds numbers and modeling roughness.

    Original Description:

    Pipe friction, pressure loss

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    The document summarizes research on turbulence in pipes at high Reynolds numbers. Experiments were conducted using a specialized facility capable of achieving Reynolds numbers up to 35 million. The results provide new data in the smooth and rough pipe regimes, requiring updates to existing diagrams like Moody's. Computational simulations are also discussed as a promising tool for further studies, especially with emerging petascale resources, though challenges remain in simulating the highest Reynolds numbers and modeling roughness.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    92 views57 pages

    Moody Steel Pipes

    Uploaded by

    Stanislav Sviderek
    The document summarizes research on turbulence in pipes at high Reynolds numbers. Experiments were conducted using a specialized facility capable of achieving Reynolds numbers up to 35 million. The results provide new data in the smooth and rough pipe regimes, requiring updates to existing diagrams like Moody's. Computational simulations are also discussed as a promising tool for further studies, especially with emerging petascale resources, though challenges remain in simulating the highest Reynolds numbers and modeling roughness.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    You are on page 1of 57

    Turbulence in Pipes:

    The Moody Diagram and Moores Law


    Alexander Smits
    Princeton University
    ASME Fluids Engineering Conference
    San Diego, July 30-August 2, 2007

    Thank you
    Mark Zagarola, Beverley McKeon, Rongrong
    Zhao, Michael Shockling, Richard Pepe, Leif
    Langelandsvik, Marcus Hultmark, Juan
    Jimenez
    James Allen, Gary Kunkel, Sean Bailey
    Tony Perry, Peter Joubert, Peter Bradshaw,
    Steve Orszag, Jonathan Morrison, Mike
    Schultz

    Lewis Ferry Moody 1880-1953


    Professor of Hydraulic Engineering
    at Princeton University, 1930-1948
    Friction factors for pipe flow, Trans.
    of the ASME, 66, 671-684, 1944

    (from Glenn Brown, OSU)

    The Moody Diagram


    Laminar
    Increasing roughness k/D

    Smooth pipe
    (Blasius)
    Smooth pipe
    (Prandtl)

    Outline
    Smooth pipe experiments
    ReD = 31 x 103 to 35 x 106

    Rough pipe experiments


    Smooth to fully rough in one pipe
    Honed surface roughness
    Commercial steel pipe roughness

    A new Moody diagram(s)?


    Predicting arbitrary roughness behavior
    Theory
    Petascale computing

    High Reynolds number facilities


    National Transonic Facility, NASA-LaRC

    80 x 120
    NASA-Ames

    High Reynolds Number in the lab:


    Compressed air up to 200 atm as the working fluid
    34 m

    Flow

    Diffuser section

    Primary test section


    with test pipe shown

    Test leg

    Flow conditioning section


    1.5 m

    To
    motor

    Flow

    Pumping section

    Heat exchanger

    Princeton/DARPA/ONR Superpipe:
    Fully-developed pipe flow
    ReD = 31 x 103 to 35 x 106

    Return leg

    Standard velocity profile


    U+ = U/u

    Inner variables

    Outer

    Overlap region
    Inner

    y+ = yu /

    Similarity analysis for pipe flow


    Incomplete similarity (in Re) for inner & outer region
    U
    yu Ru
    +
    = U = f
    ,
    = f(y+ , R + )

    u

    Inner scaling

    y Ru
    U CL U
    = g(, R + )
    = g ,
    R

    uo

    Outer scaling

    Complete similarity (in Re) for inner & outer region


    yu
    U = f
    = f(y+ )

    U CL U
    y
    = g = g()
    R
    uo
    +

    Inner
    Outer

    Overlap analysis: two velocity scales


    At low Re:
    uo
    = h(R + )
    u
    > Match velocities and velocity gradients power law

    At high Re:
    uo
    = constant
    u
    > Match velocities and velocity gradients power law
    > Match velocity gradients log law

    Superpipe results

    Pipe flow inner scaling


    30

    25

    20
    U = 8.70(y
    +

    U =

    + 0.137

    15

    1
    +
    ln y + 6.15
    0.436

    10

    5
    U + = y+
    0

    100

    101

    102

    y+

    103

    104

    105

    Smooth pipe summary


    Log law only appears at sufficiently high Reynolds number
    New log law constants: =0.421, B=5.60 (cf. 0.41 and 5.0)
    Spalart: = 0.01, gives CD=1% at flight Reynolds
    numbers
    New outer layer scaling velocity for low Reynolds number
    What about the friction factor? Need to integrate velocity
    profile.

    Gottingen, Germany

    Ludwig Prandtl

    Theodor von Krmn

    Johann Nikuradse, 1933


    (from Glenn Brown)

    Prandtls Universal Friction Law

    Prandtl:

    Nikuradses data

    Prandtl

    Superpipe results
    Prandtl (1935)
    McKeon et. al. (2004)
    Blasius
    (1911)
    McKeon et al. (2004)

    Prandtl
    (1935)

    Two complementary experiments


    10

    Princeton Superpipe
    Oregon

    0.1

    0.01
    1

    10

    10

    10

    10

    10

    Re

    10

    10

    10

    Roughness
    How do we know the smooth pipe was really smooth at
    all Reynolds numbers?
    Were the higher friction factors at high Reynolds numbers
    evidence of roughness?
    What is k?
    rms roughness height: krms
    equivalent sandgrain roughness: ks

    Nikuradses rough pipe experiments (sandgrain roughness)


    ks+ < 5, smooth
    5 < ks+ < 70, transitionally rough
    ks+ > 70, fully rough

    Nikuradse's sandgrain experiments


    Transition from smooth to fully rough included inflection
    fully rough

    smooth

    transitional

    "Quadratic Resistance" in fully rough regime - Reynolds


    number independence

    Colebrook and the Moody Diagram


    Data from Colebrook & White (1938), Colebrook (1939)
    Tested various roughness types
    Large and small elements
    Sparse and dense distributions

    Studied many different pipes with commercial roughness


    Nikuradse sand-grain trend with inflection deemed irregular
    Focus on the behavior in the transitional roughness regime

    Cyril F. Colebrook

    (from Glenn Brown)

    Lewis Moody

    The Moody Diagram

    Colebrook transitional
    roughness function

    Where did Colebrooks function come from?

    Colebrook (1939)

    Colebrook & White boundary layer results

    New experiments on roughness


    Use Superpipe apparatus to study different roughness types
    by installing different rough pipes
    Advantage: able to cover regime from smooth to fully
    rough with one pipe
    Honed surface roughness
    To help establish where Superpipe data becomes rough
    (10 x 106, or 28 x 106, or 35 x 106, or what?)
    To help characterize an important roughness type
    (honed and polished finish) (ks = 6krms, ks = 3krms?)
    Commercial steel pipe roughness
    Most important surface for industrial applications

    Honed surface finish


    "Smooth pipe, 6in

    Honed rough pipe, 98in

    Results in smooth regime

    Results: transitional/rough regime

    Inner scaling - all profiles

    Hama roughness function

    Rough pipe: Inner scaling

    rough pipe
    Moody

    ks+
    Therefore smooth Superpipe was smooth for ReD </= 21 x 106

    Friction factor results for rough pipe

    Nikuradse

    Monotonic
    (Moody)

    Inflectional

    Revised resistance diagram for honed surfaces

    "Rough"

    "Smooth"

    Commercial steel surface roughness


    Honed rough pipe, 98in

    Commercial steel rough pipe, 195in

    3.82

    5.0m

    Sample 2: non-rust spot

    Sample 2: rust spot

    Velocity profiles: inner scaling

    Velocity profiles: inner scaling

    Velocity profiles: inner scaling

    Hama roughness function

    Colebrook
    transitonal
    roughness
    commercial
    steel pipe

    honed
    surface
    roughness

    Commercial steel pipe friction factor

    ks = 1.5krms

    Moody diagram for commercial steel pipe

    Rough pipe summary


    Honed surface roughness
    Smooth
    transitional
    fully
    rough
    krms/D = 19 x 10-6
    Pipe L/D = 200
    ReD = 57 x 103 to 21 x 106
    Smooth for ks+ < 3.5
    ks = 3.0krms
    Inflectional friction factor not
    monotonic (Nikuradse not
    Colebrook)

    Commercial steel pipe roughness


    Smooth
    transitional
    fully
    rough
    krms/D = 38 x 10-6
    Pipe L/D = 200
    ReD = 93 x 103 to 20 x 106
    Smooth for ks+ < 3.1
    ks = 1.5krms (instead of 3.5krms)!
    Friction factor monotonic (but not
    Colebrook)

    Why the Moody Diagram needs updating


    Prandtls universal friction factor relation is not universal
    (breaks down at higher Reynolds numbers: >3 x 106)
    Transitional roughness regime is represented by
    Colebrooks transitional roughness function using an
    equivalent sandgrain roughness, which takes no account of
    individual roughness types
    Honed surfaces are inflectional not monotonic
    Commercial steel pipe monotonic but not Colebrook
    The limitations of the Moody Diagram were well-known
    (e.g., Hama), but no match for text book orthodoxy

    Where do we go from here?


    More experiments, more data analysis?
    Schultz and Flack

    A predictive theory?
    Gioia and Chakraborty

    Petascale computing?
    Moser, Jimenez, Yeung

    Goia and Chakrabortys (2006) model

    Model the energy spectrum in the inertial and dissipative ranges


    Use the energy spectrum to estimate the speed of eddies of size s
    Model the shear stress on roughness element of size s as
    Hence
    , then integrate across all scales to find

    Prospects for Computation: Moores Law


    Intel co-founder Gordon Moore

    2005
    April, 1965

    Petascale computing
    Earth Simulator (2004): 36 x 1012 flops peak
    DNS of 40963 isotropic turbulence

    Petascale computing (2007): Blue Gene/P 3 x 1015


    flops peak
    Remarkable resource, but what questions can it answer?

    Example: DNS of channel flow


    Bob Moser, UT Austin
    Re = uR/ (approx = ReD/40)

    Channel flow simulations

    Domain L2000

    Resources for L2000 (ReD approx 80,000)

    How much higher can we go?


    How much higher need we go?

    Extended log-law (Moser)

    Comparison with DNS channel data

    From Re = 2000 to 5000

    Resource requirements

    Are we done with channels at Re = 5000?


    Will give about an octave of log-law
    Will display true inner and outer regions
    Inadequate for high Reynolds number scaling (need
    Re > 50,000)
    What about roughness?
    With a 10 Petaflop machine
    Re = 5000 is cheap enough to do experiments
    Roughness studies?

    Maybe we can do roughness with a teraflop machine


    (if we are clever)

    Conclusions
    Moody diagram should be revised, or used with caution
    Colebrook is pessimistic (makes us look good)
    Transitional roughness behavior not universal: depends on
    roughness
    Gioia model combined with better surface characterization may
    lead to predictive theory
    Petascale computing will provide powerful resource for fluids
    engineering, but maybe well solve roughness without it
    A Golden Age in the study of wall-bounded turbulence?

    Questions??

    Osborne Reynolds

    You might also like

    pFad - Phonifier reborn

    Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

    Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


    Alternative Proxies:

    Alternative Proxy

    pFad Proxy

    pFad v3 Proxy

    pFad v4 Proxy