otc-1071-ms offshore pipelines

Upload: anil-chanana

Post on 02-Jun-2018

226 views

Category:

Documents


0 download

TRANSCRIPT

  • 8/10/2019 OTC-1071-MS Offshore Pipelines

    1/14

    OFFSHORE

    TECHNOLOGY

    CONFERENCE

    6200

    North

    Central Expressway

    D allas Texas 75206

    THIS

    IS

    A PRE

    PRINT

    SUBJECT

    TO

    CORRECTION

    R e

    7

    Stress nalysis

    uring

    of

    Offshore

    Pipelines

    Installation

    By

    T. Powers Humble Pipe

    Line Co.

    an d L. D. Finn Esso Production Research Co.

    Copyright 1969

    Offshore TechnologyConfe-rence

    orr behalf

    o f

    American

    I ns t i t ut e o f Mining Metallur gical

    and

    Petroleum Engineers I n c . The American

    Association

    o f Petroleum

    Geologists

    American I ns t i t ut e of

    Chemical Engineers American Society o f

    Civil

    Engineers The American Society o f

    Mechanical

    Engineers The

    I ns t i t ut e

    o f E l e c t r i c a l

    an d E lectr onics Engtneers I n c .

    Marine

    Technology Society

    Society

    o f Expl orat i on G e ophys i c is t s

    arid Socie-cy

    o f Naval

    Architects

    Marine Engineers.

    This paper

    was

    prepared fo r

    presentation

    a t the Firs t Annual Offshore Technology Conference

    to

    be

    held

    in Houston Tex.

    on

    May 18-21 1969. Permission to

    copy i s

    r e s t r i c t e d to an a bstra c t o f

    no t

    more than 300

    words.

    I l l us t r a t i ons may not

    be copied.

    Such use o f

    an

    a bstra c t should contain

    conspicuous

    acknowledgment o f where

    an d

    by_whom

    the paper

    i s

    presented.

    ABSTRACT

    A finite-beam-element

    i n i t i l ~ v l u e

    analysis

    procedure d e t e = i n e s s tr es se s i n

    a

    subsea

    pipeline

    suspended between

    th e

    ocean

    fl o o r an d a laybarge or--stinger. The b as i c

    theory i t s a dv an ta ge s o ve r

    other

    the orie s an d

    a comparison o f r e s u l t s

    with

    th e

    r e s u l t s

    o f

    a na lytic a l

    p r oc ed u re s b as ed

    on

    other

    t h eo ri es

    are

    included. The finite-element

    theory is

    applicable over a wide

    range

    o f

    marine

    pipelay

    in g problems an d compares

    favorably

    with

    other

    accepted theories

    in

    the-

    ranges

    o f thei r a p p ~ i ~

    c a b i l i t y .

    INTRODUCTION

    Thick-wall piRe an d o n r e t e ~ t i n g s axe_

    used to weight marine p i p el i n es to insure t h a t

    the pipeline wil1.remain

    in

    place a ft er i ns ta l

    l a t i o n . During inst i l lat ion i n deep- water this

    weight

    imposes

    high stre sse s i n

    th e

    pipe

    sus-

    pended

    from th e laybarge and may

    cause

    t h e i l l ~

    to fai l Deepwater

    re pa ir

    may

    nof

    be

    possible;

    as a r e s ul t a f a i l u r e may r e ~ u i r e

    the

    entire

    l i n e to

    be replaced. o ~ e l e c t the

    proper

    pipe weight an d

    grade anclt;o prevent

    c os t l y

    f a i l u r e s

    th e

    stre sse s

    imposed

    i n

    th e

    pipe

    during laying must be defined an d construction

    procedures

    an d

    e9.uipment must be accurately

    analyzed.

    This paper

    presents a

    procedure

    fo r

    making t he s e a n al ys e s. The c a lc u la ti o n t ec h

    nique

    may

    be used to

    analyze

    an y construction

    method which ins ta l l s

    th e

    l i n e from

    th e

    surface

    References

    an d

    i ll u st ra ti o ns a t

    en d

    o f paper.

    in a continuous s t r i n g .

    This

    paper describes

    the

    a na lytic a l procedure

    specifically

    in terms

    o f t h e s ti n ge r -I a yb a rg e c o ns tr uc t io n method.

    The b as i c theory

    used in

    the an al y s i s i t s

    advantages

    over other

    the orie s an d

    a

    compari

    so n of r e s u l t s

    with results using other

    the orie s

    a re p re se nt ed .

    THE

    MARINF.

    PIPELINE

    SUSPENSION

    PROBLEM

    The

    problem

    o f p r ed ic t in g the

    suspended

    geometl.,Y c f

    marine

    pipelines

    in

    deep water i s

    one o f nonJi::. 3ar, large-angle bending.

    This

    problem

    may be

    s o lv e d n u m er i ca ll y through

    th e

    us e

    of a f iro.i.te-beam

    element i n i t i a l - v a l u e

    approach

    L l l ~ t r e a t s

    th e

    to ta l

    beam

    as

    a

    se rie s

    o f sm -ll

    hemna,

    each

    o f

    which

    is

    analyzed with

    l .

    .inear

    theor.i. Combining

    a l l

    th e

    small-angle

    bending so]U: ;iOl.S produces

    the

    large-angle

    bending marj :Qc p i pe li n e s o lu t io n . In the

    in itial-valtl -;--a.pproacn.;imknown

    boundary

    condi

    t ions the

    ocean

    fl o o r are assumed

    an d

    then

    s Y f l t e m a t i \ a l l ~ _

    changed

    unt i l

    th e

    p r o f i l e which

    has been bui l t up

    segment

    by segment s a t i s f i e s

    boundary conditions

    s p eci fi ed a t th e

    u pp er e nd .

    The appendiX to th is paper

    contains

    th e b as ic

    e 9. ua ti on s a nd ~ h e i r

    derivation.

    The t h e o l ~

    accounts

    fo r applied tensions

    ex t ern al fluid pressures va ria tions

    i n

    water

    currents

    With

    depth pipe s t i f f n e s s va ria tions

    due

    to

    weakness o f

    the weight coating

    a t

    th e

    f ield j oi nt s

    an d

    support buoys [ i f used]. The

    approach

    is l uite general an d is applicable to

  • 8/10/2019 OTC-1071-MS Offshore Pipelines

    2/14

    I I lO

    STRESS ANALYSIS

    OF

    OFFSHORE

    PIPELINES

    DURING INSTALLATION

    OTC 10

    both two-and t h re e -d i m en s io n a l f o r ce s an d de

    flections;

    crosscurrents an d

    la tera l barge

    movement ca n

    be considered.

    Although th e

    develoIJed

    procedure

    can>

    be

    used

    to analyze

    the stresses in th e

    suspended

    portion of a p i p e l i n e u r J ~ e r an y

    support

    con

    d i t i o n i t is discussed>here in

    terms

    o f using

    t h e s t in g er -l ay b a rg e

    construction

    method [Fig.

    1]. The stresses calculated

    account

    for

    s ta t ic

    loads

    only

    considering

    ocean

    currents as

    s tatic

    loads. The e f f e c t s

    o f

    dynamic loads an d ex

    t e r n a l hydrostatic

    pressure

    on th e actual long

    i td u in a l s tr es s

    must

    be considered

    in th e selec

    t io n of allowable s tr es s levels used in con

    junction with this procedure.

    USE OF

    FINITE-ELEMENT

    CALCULATION PROCEDURE

    The

    laborious

    an d r ep etit iv e

    nature

    of th e

    finite-elem ent procedure

    makes

    i t necessary to

    us e a computer. The

    p r oc ed u re h a s

    been

    p r o ~

    grammed fo r the

    IBM 360.

    In a dd it io n

    to

    the

    boundary-condition assumptions

    the

    procedure

    requires th at w a te r d e pt h p i pe properties

    applied tension height

    to upper

    support point

    an d eith er moment or slope

    a t

    th e

    upper support

    point be given. When

    analyzing

    th e stinger

    laybarge

    method the

    upper support

    point w i l l

    be

    the

    l i f t o f f point on th e stinger [ i .e . th e

    las t p o i n t

    t ha t t he

    pipE l is in cgnt.?-ct with.j:;he

    s tin g er ] . I f

    buoys

    or

    currents

    are being

    included in the

    analysis

    th e p oin ts of

    appli

    cat ion

    a nd m a gn it ud e

    o f th ese forces

    i s

    al80

    re luired.

    To

    adequately analyze th e stresses imposed

    on a

    pipeline

    when using the s t i n g ~ r l a y b a r g e

    ~ o n s t r u c t i o n

    method

    some means

    must

    be

    found

    to

    simultaneously

    describe

    the

    s t i n g e r

    and the

    suspended pipeline geometry as an y change in

    stinger

    position

    w i l l

    r e s u l t in a change in

    the

    geometry of

    the

    suspended

    l i n e .

    This

    ca n

    be

    accomplished

    by

    u t i l i z i n g th e p l o t s

    shown

    in

    Figs. 2 an d 3.

    Fig.

    2

    is

    a p lo t

    of

    the maximum

    s t r e s s in

    the

    suspended portion under various

    a pp l ie d t en s io n s as

    a

    function of the >l:i.ft-off

    angle. Fig. 3 is the corresponding p lo t of

    height

    to

    l i f t o f f point

    fo r variQus

    tens tons

    as

    a function of th e l i f t o f f angle. The

    a p p l i e d t e n s i o n s a ~ ~ ~ ~ ~ a x i a l

    tensions

    applied a t th e barge an d do not allow

    fo r

    fr ic ti on i n

    the

    r o ller s

    o f

    the

    stinger. How-

    ever the calculation procedure considers the

    tension reqUired

    to

    hold

    the

    wei ght

    o f

    th e pipe

    on the stinger between th e barge an d the l i f t -

    o f f point. The

    l i f t o f f angle is

    the

    angle

    be

    tween th e

    horizontal

    an d a tangent

    to

    th e p ip e

    l in e a t th e point where leaves the

    s t i n g e r .

    The water

    depth

    pipe

    size

    an d over-all pipe

    s p ec i fi c g ra v it y

    for this s e t

    o f curves

    are

    fixed.

    Two p o t e n t i a l problems

    ex is t

    in

    the

    sus

    pended portion

    of

    th e pipeline. The

    pipeline

    may bend excessively an d

    buckle

    in th e

    suspende

    p or ti on o r

    i t may

    f a i l

    to l i f t o ff th e s t i n g e r

    an d

    b uc kle o ve r th e

    en d

    of th e s t i n g e r . Both

    p ro bl em s c an be

    lessened

    by

    increasing

    tension

    decreasing the specific

    gravity or

    lengthening

    th e s t i n g e r . Fo r

    a

    given

    tension an d

    specific

    gravity

    th e maximum s tr es s in th e suspended

    portion

    is reduced

    when the

    l i f t o f f

    angle

    is

    reduced

    [Fig.

    2 ] .

    However th e submerged en d

    o f the

    stinger m us t h ave

    s u f f i c i e n t

    slope an d

    depth

    to

    permit the pipe

    to l i f t

    o f f t he s ti ng en

    Fo r a given s t i n g e r length the maximum pipe

    l in e s tr es s is m i ~ i m i z e d when the l i f t o f f angl

    is the minimum obtainable.

    To

    achieve l i f t o f f

    when laying in deep water i t is desirable t h a t

    the

    stinger be curv ed to lim it s t i n g e r length.

    Just how stinger curvature

    is

    obtained is un

    important to the following

    discussion. optimum

    stinger

    deployment is

    defined

    as

    th at

    which

    yields the

    minimum

    obtainable l i f t o f f angle

    an d minimum flooding-schedule s tr es s or curva

    ture

    with

    a

    given

    amount

    of tension.

    optimum

    deployment a ls o re qu ire s th e pipe

    to l i f t o ff

    near the en d

    o f

    th e

    s t i n g e r . This

    analysis

    .assumes optimum l i f t o f f

    to be

    20

    f t

    from the

    end.

    Obviously the stinger shape or geometry

    must

    be

    known. The

    stinger geometry curves are

    obtained

    by

    eith er

    choosing varying

    amounts

    of

    bui l t in curvature or by making a s tr es s anal

    y s is o f

    a

    p a r t i c u l a r

    s t i n g e r ~ u si ng f lo o di ng

    sc he dul e s c orre spondi ng to different

    s tr es s

    l e v e l s .

    These

    curves

    are

    then

    superimposed on

    on

    the

    height

    to l i f t o f f point vs l i f t o f f

    angle

    p l o t as shown on Fig. 3.

    I f

    a s tr aig h t

    s t i n g e r

    is flooded

    to

    obtain

    a

    given

    curvature

    th e flooding schedule

    is

    usually

    selected

    b y

    t r i a l

    an d e r r o r .

    Since

    s ti ng e rs c ur re n tl y being

    used are hinged

    o r

    pinned

    a t

    th e barge each

    t r i a l

    flooding

    arrangement

    must

    be

    chosen

    so

    the stinger w i l l have j u st enough buoyance

    to

    support the pipe r e s t ~ n g on the stinger plUS a

    p or ti on o f

    th e

    suspended

    pipe weight. The

    suspended pipe weight

    to

    be

    supported

    can be

    determined

    from t h e f in i te - el e m en t calculation

    procedure.

    To

    o bt ai n t he stinger geometry

    curves

    in

    Fig. 3 a 550-ft

    s t r a i g h t

    stinger was theore

    t ica l ly

    flooded

    to

    cause

    maximum

    s t r e s s

    [due

    to

    flooding schedule]

    o f

    0 lO l5 20 an d 25 k s i

    in the stinger members. These same curves

    could have

    been

    generated

    by

    assuming 0

    0

    4

    0

    9

    10

    0

    an d

    14.5

    0

    b u i l t - i n

    curvature.

    The

    s t i n g e r

    is assumed

    to

    have

    an

    in i t ia l slope of 13

    0

    a t

    th e water surface. The stinger geometry curves

    are formed

    by

    p l o t t i n g th e slope of th e s t i n g e r

    as a function of height above th e ocean floor

    for the four stinger s tr e ss l ev e ls . A

    curve i s

    then drawn connecting a p o i n t 20 f t from

    the

  • 8/10/2019 OTC-1071-MS Offshore Pipelines

    3/14

    GTC 1071

    J . T. POWERS and L. D FINN

    I I 11

    en d

    of the

    s tin ge r f or eaCh s t i n g e r s tr es s

    lev el. This

    i s

    r f r ~

    to

    as

    th e

    l i f t - o f f

    l i n e . The

    intersection

    o f each s tin g er geometr

    curve w ith th e a pp lied tension curves i s

    then

    transferred

    to

    F ig . 2. Also transferred

    to Fig.

    2 i s the i nt er s ec t io n o f

    the

    applied tension

    curves with

    th e

    l i f t - o f f l i n e s. This

    l i f t - o f f

    l i n e on Fig. 2

    i s

    a l i m i t of

    stinger

    p l o y m n ~

    in

    other words for

    t h i s

    p a r t i c u l a r

    s tin g er

    there

    i s

    no

    solution

    to

    the

    l e f t

    o f

    th e l i f t

    o ff lin e

    Fig. 2. I f the s t i n g e r geometry

    s tr es s i s limited

    to

    25 000

    p s i

    then

    there

    i s

    no solution to

    the r ig h t

    of

    t he 2 5-k si l i n e .

    The p lo tted output shown in Figs. 2

    an d

    3

    ca n

    be used to determine [1] th e

    tension

    re

    quired

    t o p re ve nt overstressing the

    pipeline

    with a given s t i n g e r

    curvature

    [2 ]

    the

    tension

    required to l i f t o f f the s tin g er

    an d

    [3] the

    e f f e c t of changing the curvature of the stinger.

    The output ca n

    also

    be used

    to

    d e J ~ e r w i n e the

    e f f e c t of

    using

    d i f f e r e n t equipment

    p p

    s i z e s

    specific g r av ities e tc .

    Examples of Finite-Element

    Calculations

    The following examples i l l u s t r a t e the uses

    o f

    th is an aly tical procedure.

    Determination

    of

    T ension Re qui re d to

    Prevent Overstressing

    th e

    Pipeline

    for

    a Given Maximum Stinger Curvature

    Using Figs. 2 an d ~ p r e p r e s p ecif ically

    20-in. pipe i n 200 water l e t

    us

    assume t h a t th e following

    conditions

    have been

    se t :

    Maximum

    allowable

    s tin g er

    s tr es s l5 OOO

    p si

    Maximum allowable s tr es s

    in

    pipeline

    36 000 p si

    Enter

    Fig. 2 with a 36-ksimaximum combined

    s t r e s s

    proceed

    h oriz on ta lly to th e 1 5-k si

    s t i n g e r

    st r e ss

    l in e an d observe t h a t 40 OOO lb

    tension i s

    required.

    In

    th is case

    th e

    pipe

    would l i f t o ff

    the s t i n g e r

    more than 20 f t from

    the

    end.

    T e nsion Re qui re d to Li f t

    Of f

    Stinger

    Assuming t h a t i t

    i s

    desired

    to l i f t

    o ff

    20

    f t from the en d o f the

    s tin g er

    an d s t i n g e r

    s tr es s i s limited to 10 000 ps Fig. 3 in d i

    cates t h a t 60 000

    Ib

    of tension i s required.

    This

    i s

    determined b y f ol lo wi ng

    the lO OOO psi

    s t i n g e r

    st r e ss

    l in e

    to

    i t s intersection

    with

    the l i f t - o f f

    l in e [ignore the

    dashed

    l in es

    which are used

    in

    Ex. 3 ]. This intersection

    occurs a t

    the 60 OOO lb applied tension

    curve.

    [An

    applied te ns io n o f les s

    than

    60 000

    Ib

    would

    allow

    the

    pipe

    to droop

    over

    the en d o f

    th e s tin g er . ] The l i f t - o f f angle would be

    about 17

    0

    occurring

    a t a point about 64 f t

    above

    th e

    ocean

    floor. Fig.

    2

    indicates

    t h a t

    th e maximum combined s tr es s

    in

    the suspended

    portion

    o f the

    pipeline

    for t h i s condition

    would

    be

    27 000 p s i ..

    Effect o f Changing the

    Curvature

    o f the

    Stinger

    F i r s t

    assuming

    40 OOO lb

    applied

    tension

    an d a stinger c u r va t ur e p r ov i de d

    by

    15 000 p s i

    flooding

    schedule

    s t r e s s

    Fig.

    3 shows

    [see

    dashed l in es ] t h a t the p p w i l l l i f t

    o ff

    68

    f t

    above th e ocean floor

    a t

    a l i f t - o f f

    angle

    o f

    about 19.8

    0

    Referring

    to F ig.

    2

    the s tr es s

    in th e

    pipeline i s found

    to be 36 000 p s i fo r

    the

    19.8

    0

    l i f t - o f f

    angle

    an d

    applied

    tension

    o f

    40 000 I b s . However

    the

    f l oo d in g s ch ed ul e

    ca n b e r ed uce d to a bo ut 1 2 5 00

    p s i

    an d

    s t i l l

    allow

    the

    pipe to l i f t o f f

    20

    f t

    from the

    en d o f

    the s tin g er . This

    can be determined

    in

    Fig. 3

    by

    moving

    down

    along th e 40 000-10

    tension curve from th e 15 OOO psi s t i n g e r s tr es s

    lin e

    to the

    l i f t - o f f

    l i n e

    an d interpolating

    be

    tween the 10 OOO and 15 OOO psi s t i n g e r

    st r e ss

    l i n e s. The

    pipeline

    w i l l

    now

    l i f t

    o f f

    59

    f t

    0

    above the ocean f l o o r a t an

    angle

    of about 18

    Fig.

    2 indicates t h a t th e maximum pipeline

    s tr es s i s reduced to 34 000 p s i .

    Other Uses

    of

    Finite-Element

    Calculations

    New Stinger Design

    Fo r a specific

    i n st a l l a t i o n

    where the water

    depth pipe s i z e

    e t c .

    are known curves simi

    l a r to those in

    F igs.

    2

    an d

    3

    ca n be

    used to

    s elect the

    proper

    s te e l grade and s t i n g e r

    leng th. S tinger geometry curves

    could

    be drawn

    corresponding

    to

    available

    grades

    o f

    s te e l.

    Kilowing the maximum value o f te ns io n which

    c ou ld b e

    applied in

    the

    p a r ti c u la r s i tu a t io n

    th e

    s tin g er

    length

    could be

    determined from th e

    i nt e rs e ct io n o f the stinger s tr es s curve

    with

    th e maximum available

    applied

    tension

    curve.

    S el ec ti on o f Pipe

    Wall Thickness

    Grade an d

    s p ecif ic

    Gravity

    In a s itu atio n where

    the

    p p size has

    been

    selected an d the lay barge i s equipped

    with a s p ec if ic s t in g er an d a

    tensioning device

    o f known

    capacity the proper

    selection of

    p p

    wall thickness an d grade ca n be made by pre

    paring

    a

    se t

    o f

    curves

    similar

    to

    F igs.

    2

    an d

    3

    fo r each wall thickness and each specific grav

    i t y .

    COMP RISON WITH

    OTHER

    THEORIES

    Results obtained

    with

    th e

    finite-element

    approach were compared with r es u lts from sever

    a l existing

    deflection

    t h e o r i e s . The f i r s t

    comparison was made

    with

    the simple beam theory

    t h a t

    does not consider ax ial tension. Thus

    for

    a f a i r comparison

    i t

    was

    necessary to

    apply

  • 8/10/2019 OTC-1071-MS Offshore Pipelines

    4/14

    1112

    TRESS ANALYSIS OF OFFSHORE PIPELINESDURING INW .MXATION

    cTc lo~

    approximately14 kips tension at the barge to

    have zero effectivetension at the ocean floor.

    It was found that the simple beam and finite-

    element solutions compare very well for lift-

    off angles as large as 55 [Fig. 4]. Because

    the simplebeam solutioncannot account for

    axial tension, it has limited application.

    Abeam column solutionwas developedby

    solving the small-mgle bending, linear differ-

    ential equation for a singlebeem element in

    tension. In water depths to at least 250 ft

    with a relativelystiff pipe the beam column

    and finite-elementsolutionsagree very well

    for lift-off mgles as large as 30 [Fig. 4].

    It is evident from Fig. 5 that the sus-

    pended length is considerablylonger and flatte

    for 60 kips applied tension thaa.for.20kips,

    so the small-auglebending theory accuracy

    improves for higher tensions.

    The beam columu

    solutiongenerallypredicts stresseswhich are

    higher than those predictedby the finite-

    element solution [Fig.6]. IU deeperwater

    [1,000ft] with a relativelyflexiblepipe the

    bean-column solutionand the finite-element

    solution do not comparenearly as well [Figs.

    7and8]. Althoughthe.bendingstress diagrsm

    [Fig.81 is significantlydifferent,especially

    near the upper end, the maximum stress differs

    by only 13 percent.

    With the natural catenary,it was found

    that for a constantapplied tension,the pro-

    file describedby the suspendedportton was

    independentOf the lift-offheight.

    Thi.s same

    trend was observedwith three of the other

    theories for deep water with relativelylarge

    tensions am.dflexiblepipelines. Although the

    profile describedby the natural catenaryin

    deep water differs from the others [Fig. 7],

    the slope at the up$er end comparesvery well

    with the slopes of the finite-elementand stiff

    ened catenarysolutions.

    As the natural cate-

    nary solutionignorespipe stiffness,the pipe

    can bend more freely at the ocean floor, so the

    horizontal distanceis shorter than obtained

    with solutionswhich considerstiffness. The

    maximum bending stresspredicted by the natural

    catenarystress diagram follows the finite-

    element and stiffenedcatenarydiagrsms through

    the middle part of the suspendedpipeline,but

    the boundary conditionsat the end are not

    satisfied.

    In shallowerwater with relativelysmall.

    tensions and stiff pipes, the natural catenaxy

    solutionis grossly in error [Figs.4, 5 and 6]

    Fig. 5 illustratesthat the error in the profil

    decreasesas the tension increases. The maxtiw

    stress in the suspendedpipelineprgdicted by

    the finite-elementmethod asymptotically

    approachesthe maxim~ stresspredictedby the

    natural catenaryas the lift-offangle in-

    -~

    manner that the results couldb= cmpa:redto

    those from tl:efi.nite-e].ementme~hod.

    In deeg

    water with re?&%i.ve.Qflexiblepipes and Ioug

    suspendedlengths,

    the resu.1.tsbtr,inedwere

    identicalwith ~,hosefrcm ttief%.i+;e-element

    method [Figs.7 ant 8].

    The stiffei~.e

  • 8/10/2019 OTC-1071-MS Offshore Pipelines

    5/14

    OTC 1071

    ~_.--J...PQWjlRSand L. D. FINN

    11 13

    the finite-elementsolutionsagree in deep

    forces at n along X, Y and

    water, the former can be used in deep water to

    Z axes, respectively

    obtain a two-dimensionalsoultion.

    However,

    Fq Fq Fq

    XiI)Yn) Zn

    = summationbetween poin-t1

    ocesn currentsand nonzero moments at the lift-

    and.n of current drag

    off point cannotbe considered,and the resul*-

    forces along X, Y and Z

    ing suspendedgeometrycannotbe made compatible

    axes

    with a curved stinger.

    F

    xopFYopFzop

    = componentsof the outside

    fluid pressure acting

    CONCLUSIONS

    normal to the pipe wall

    F

    Xip> Yip) Zip

    = componentsof the inside

    The calculationtechniquepresented is

    fluid pressure acting

    quite general and is applicableand accurate

    normal to the pipe wall

    for a wide range of OffshO.repipelayingprob-

    g

    = gravitationalconstant

    lerns.

    Becausethetechniqueis very flexible,

    H=

    total water depth

    it can accurately considervarious boundary

    moment of inertiaof pipe

    conditionssmd, as is shown in the appendix,

    ~= length of beam segnent

    currentprofiles,weight coatings>awiliary

    %> Yn~ %

    = moments requiredat point u

    supportbuoys, crosscurrents>and lateral barge

    for equilibriwnabout X,

    drift.

    Y and Z axes

    % ~ zn

    moments at point n about X,

    The calculationprocedure can be used to

    y and z axes

    evaluateadequacy of constructionequipment,

    % ~n %

    = moments at point n resultin~

    design new equipment,and.selectPiPe character-

    from currentdrag on sec-

    istics for a.proposed.pi~ellne Szn

    = forces at point n along x,

    ACKNOWLEDGMENT y and z axes

    T

    The authors wish to acknowledgethe contri- en

    = effectivetension at n

    .

    horizontal currentvelocity

    butions of D. W. Dareing, U. of Arkansas, to the

    J

    on jth

    segment

    calculationprocedure describedin this paper.

    v

    They also wish to thank the management of Esso

    yj zj

    = current velocityalong y

    and z axes

    productionResearch Co. for their permissionto

    Wy? Wz

    =

    uniformly distributedload

    publish this paper.

    per unit length on beam i~

    REFERENCE

    y and z directions

    wb

    .

    = buoyancyper unit length

    .

    e

    = effectiveweight per unit..

    1.

    Dixon, D. A. and Rutledge,D. R.: Stiffene

    d

    length of pipe in fluid

    CatenaryCalculationsin Pipellne Laying

    P

    = weight per unit length of

    problem Traus. ASME, Petroleum Division, coated pipe in air

    paper 67-=[Sept., I-9671.

    x> Y> z

    = local coordinates

    Yn? n

    = deflectionof point n rela-

    NOMENCLATURE tive to n-1 along y and z

    [forAppendix.]

    x Yn, Zn =

    )

    global coordinatesat point

    n

    Ai =

    cross-sectionalsrea inside pipe

    X:g, Zcg =

    coordinatesof center of

    Ao

    = total area enclosedwithin the

    n

    gravity of portion sus-

    outermostdiameter

    CD

    = drag coefficient

    pended between point 1 an(

    n

    Do =

    outermostdiameter of piPe line

    en

    = angle in X-Z pleae between

    E=

    modulus of elasticity

    Z and z axes

    Fn)Fyn)FZn

    = forces required at point n for

    Yn

    ~ angle in x-y plane between

    equilibriumalong X, Y and Z

    Y and y axes

    ~b Fb ~b

    axes, respectively

    Y@ ezn

    change in angle between

    Xn Yn) Zn

    = componentsof capped end fluid

    point n-l and n in y-x

  • 8/10/2019 OTC-1071-MS Offshore Pipelines

    6/14

    STRESS ANALYSIS OF OFFSHORE PIPELINESDURING INSTALLATION

    OTC 10?1

    1

    .i

    and z-x planes, respectively

    Yf~~ Yfi

    = specificweight of fluid outside

    and insidepipe, respectively

    P

    = mass density of fluid outside

    pipe

    @ = angle horizontal currentmakes

    with X axis

    APPENDIX

    Equations for Finite-Be&mElement,

    Initial-ValueApproach

    ;eneralApproach

    A pipeline suspendedbetween the ocean ___

    floor and stingeror laybargebehaves as a long

    beam smd deflectsaccording to the forces actine

    m it.

    This nonlinear,large-anglebending

    problemmay be solved-throughthe use of a

    finite-beam-element,initial-valueapproach

    whichtreats the suspendedpipe as a series of

    small beams each of which is solved by linear

    theory. In the initial value approach, the

    unknownboundary conditionsat the ocean floor

    are initiallyassumed and then systematically

    changeciuntil the profile which has been built

    up segmentby segment satisfiesthe boundary

    conditionsat the upper end.

    The origin of the coordinateaxes is tsken

    at the point where the pipe first touches the

    ocean floor.

    For two-dimensionalproblems, the

    ~mentj tension,position, and slope are how

    ~t the origin.

    The trial solution is beg& by

    assumingthe unknown shear force at this point.

    Ihe shear, moment md tension at the other end

    [point2] of a short-segment are calculated

    N

    Cos 43XCos Q

    Y

    -Cos 9

    x

    sin 0

    Y

    Y= sin @

    Y

    Cos 9

    Y

    z

    sin (2Xcos Q

    Y

    -sin 6X sin Q

    L

    Y

    from equilibriumequations. These internal

    forces along with the external forces are then

    used to determinethe incrementaldisplacement

    and slope at tiat point.

    The internal forces

    at Point 2 could be used to obtain the internal

    forces at Point 3, but for computational-error

    reasons the forces are obtained from equilibria

    of the segnent suspendedbetween Point 3 and

    the origin. This procedure is repeateduntil

    the pipe reaches a designatedheight above the

    ocean floor. If the specifiedboundary condi-

    tions at that point are not satisfied,a new

    and improved shear force is assumed until they

    are satisfied.

    For three-dimensionalproblems,

    the same approach is taken. In this case more

    than one boundary conditionmust be assumed at

    the origin. These unknown boundary conditions

    are convergedupon separatelyuntil the tbree-

    dimensionalboundary conditionsspecifiedat the

    upper end are satisfied.

    CoordinateSYstem

    The global coordinatesystem [X, Y, Z] will

    be defined with the X axis the direction of lay,

    the Y axis vertical, and the Z axis the lateral

    direction forming a right-hand system. The

    local coordinatesystem [x, y, z] will be de-

    fined with the x axis along the pipe axis, the

    z axis perpendicularto x and in the X - Z

    plane, snd the y axis normal to the pipe and

    upward forming an orthogonalright-hand system

    [Fig.9].

    The transformation of displacement, force,

    or moment vectors from local coordinatesto

    global coordinatescan be accomplishedby the

    transformationmatrix equation:

    -sin Q.

    x

    K]

    o

    Y

    [1]

    Cos Q

    x

    z

    and the reverse transformationby the transpositionof the above,

    [H

    cos Q

    x

    Cos Q

    Y

    sin Q

    Y

    sin

    fax Cos

    Qy

    10

    Y=

    -Cos Q

    x

    sin 9

    Y

    eos Q

    Y

    -sin QX sin 0

    Y

    Y . . . . .+. . . . . .

    [2]

    z

    -sin Q

    x

    o

    cos Q

    x

    z

    where @y is the angle between the Y s.udy sxes, and ~ is the angle between the X axis and the

    projection of the x ~es.oq the X-Z plane_____

    Forces on Pipeline

    The effect of the fluid acting around the outside of the pipe segment can be derived by

    consideringa section of pipe immersedin fluid with both ends capped [Fig. I-O]. Archimedes

    principle states thgt the total ~orce or buoyancy acts in a vertical directionand is equal to

    the weight of displacetlwater, ~yfo~, where & is the area of the end plate, yfo is the

    specificweight of the fluid and A is the segment length.

    This total buoyancy can be divided

  • 8/10/2019 OTC-1071-MS Offshore Pipelines

    7/14

    JTf -071

    J. T. POWERS and L. D. FINN

    11 15

    into three parts, the two end forces Pn.I.and Pn, and the net fluid force per unit length qw,

    where qw is obtainedby integratingthe fluid pressure aro~d the circumference.

    The summation

    of forces in the vertical direction equals the total buoys.ncy:

    n

    Aoyfo .$ . Pn ~ sin Qti-l - Pn sinQyn +

    J

    ~cos Qyds . . . . . . . . . . . . . .[3]

    n-1

    and in the horizontaldirectionsequals zero,

    n

    O = Pn ~ cos Qm ~ cos Qyn ~ - Pn cos Qh cos Qyc -

    [

    ~ cos Q

    x

    SiZ10 ds. . . . . . .

    Y

    [4]

    n-1

    The two end forces can be defined as

    Pn

    =Aoyfo( H- Yn), . ..o o..... . . . . . . . . . . . . ..- . . . .. [61

    P

    n-1

    =Aoyfo( H- Yn-l ), . . . . . . . . . . . . . . . . . . . . . . . . . . . .. [7]

    where H is the total flutd depth and Yn sad Yn-~ are the heights from the bottom to point n and

    n-l, respectively.

    In an actual, suspendedpipeline the ends are not capped. The vertical and horizontal sums

    of the q forces are t~e terms of interest.

    Since the deflected shape of the segment is not

    known, q cannotbe defined in terms which can be integrated.

    Since the above equationsare not

    restricted to short segments,

    they can be applied to the section suspended between Point 1 and

    Point n by substituting1 for n-1 and s for ~, where s is the suspendedlength from the origin.

    Expressionsfor the vertical snd horizontal sums [Fxop,Fyop sad Fzop] are derived from Eqs. 3

    through 7:

    n

    Xop =

    J

    - COS Qx Sin Qy ds = A y

    [

    -H cos Qn

    o fo

    1

    COS Q~+(H-Yn) COS ~fi COS Qyn . . . [8]

    T

    L

    n.

    F

    Yop =

    J

    [

    ~cos Qyds =Ao~fo

    S- HSin QU+(H-Yn)Si.n Om

    . . . . . . . . .. [9]

    1

    -1

    -L

    n

    F

    J

    op = -

    [

    si nxfi Oy

    ds = A. Yfo -H s~ Qu cos Qn+(H-yn) sfi @wcos fi . . . [IO]

    1

    These equations show that the total effect of the fluid pressure along the pipe can be repre-

    sentedby the weight of water displaced acting verticallythrough the centroidof the suspended

    portion minus the capped end forces acting at each end.

    If the pipeline is

    fille_d_wit_hfluid__Q~_inghe lay operation,the effect of this fluid

    acting inside the pipe that extends from the ocean floor [Point1] to Point n can be similarly

    derived.

    F

    Xip

    = Ai Yfi

    [

    cos(3n cosQ~-(H-Yn) COS@mCOS~h .. o.OOO. .o*. [8A1

  • 8/10/2019 OTC-1071-MS Offshore Pipelines

    8/14

    11 16

    STRESS ANALYSISF OFF~ORE PIPFLINESDURINGINSTALLATION

    OTC 10

    F

    Yip

    = Ai yfi

    c

    -s+ Hsin Q

    -( H-Yn)sin Oyn -[9A]

    F

    Zip

    [

    Aiyfi Hsfi Q~cos Qu - (H-yn)sin Qfi.cos@yn . . . . . . . . . . . .[1OA]

    where Ai is the inside cross-sectionalarea and yfi is the specificweight of the inside fluid.

    me weight of a pipe line suspendedbetween Point 1 aud n is Wps, where Wp iS the weight

    per unit length of the coated pipe in air.

    ___Thebuoyancy per unit length cm be defined by

    %AoYfo

    -Alyfi, . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

    [11]

    and the effectiveunit weight of pipe in water by

    we

    W

    -Wb. . . . . .. . , . . . . . . . . . . . . . . . . . . . . . . . . . .

    [12]

    P

    The forces acting on the pipe suspendedbetween the origin and Point n are shown in Fig. 11.

    The effectivetotal suspendedweight of the segment.oflength s is WeS. me componentsof the

    capped end forces w~H and ~[H-Yn] that must be subtractedat Point 1 SUd n are defined as:

    b

    W =

    Wb (H - Yn) COS Qfi

    COS ~h

    b

    Yn =

    Wb (H - Yn) SillQM

    1

    .

    b

    Zzl =

    Wb (H - Yn) sin Qfi cos Qvn

    b

    /

    . . . . . . . . . . . . . . . . . . . .

    [131

    i l =

    Wb H COS Q

    xl

    cos Q

    Xl

    ~b

    n=

    Wb H sin Q

    n

    I

    b

    21 =

    Wb H Sin QD COS Qn

    J

    Other forces acting on the pipe section 1 to n could includebuoy forces

    Q . The= forces

    would always act verticallyand could be randomly spaced, but for ease of formulationthey will

    be assumed to act only at node points.

    The current forces per unit length acting on the jth segmentwill be designatedqj. This

    force is the drag which acts on a body innnersedin flowing fluid and is expressedby

    =1/2PAC V?

    j

    D

    J ~. . . . .. . . . . . . . .. . . . . . . . .

    .

    [14

    where the mass density P is 7fo/g2 A is the area exposed per ~it length [outsidedismeter

    DO in

    this case], CD is the dimensionlessdrag coefficient~d Vj is the comPonentof velocitYno~al

    to the pipe on the jth segment. The assumed horizontal current acts at a arbitrary angle 0 wit

    X axes, as shown in Fig. 12.

    The assumed constantcurrent velocity acting on the jth segment

    [Vj] is first resolved into componentsalong the X and Z axes [Vjcose, Vjsin o], and then

    resolved into local coordinatecomponentsnormal to the pipe using Eq. 2.

  • 8/10/2019 OTC-1071-MS Offshore Pipelines

    9/14

    -r-r. n

    ,-i 1 n-l

    .T W

    .lL -L{

    XHJE R.~ and T.. T). 17TNN

    I -Lu J-

    . ~. .,,

    u,w_ - .,., . --- . . . .. .

    v

    = Vj Cos @ -COS 0

    sin QYj)+ Vj sin @ (-sinQ

    sin 0

    )

    yj

    Xj

    Xj

    Yj

    )

    . . . . . . . . . .

    [15]

    v C o.g@ -sin Qxj) + Vj sin @ ( Cos Qxj)

    V=j

    zj

    J

    The

    drag forcesper unit length are obtained from Eq. 14:

    %j =

    I/2p D CD~j (sign Vyj)

    o

    1

    [16]

    . .. *O. .. .

    . . .

    q

    zj =

    1/2 p Do CD ~j (Si.@ Vzj)

    These unit forces are then resolvedinto componentsalong the global coordinateaxis using

    Eq. 1:

    qxj = qyj

    (-COS Gxj .cJ~Q

    ~j) + qzj (-sin Oxj)

    q~j

    = S ( Cos YJ )

    1

    .

    . . . . . . . . . . . . . . .

    .

    [17]

    q~j

    = ~j (-sin X,j sin Qyj) + qzj ( Cos Xj )

    The global coordinateforces at n are obtainedby applying equilibriumto the free body

    diagrsm in Fig. 11.

    n-1

    ml

    =F~+F~-F&-

    2

    1.qxj

    j=l

    n

    n-1

    Yn

    = Fm+Fb~-Fb +Wes-

    Xn

    z

    J -

    z

    1

    ~qyj . . . . . . . . . . . . .

    .[I-81

    j = j a

    n-1

    h

    = Fm+F:l-F:n -

    Z

    g q~j

    j =

    These force componentsare transformedinto local coordinatecomponents

    [Sm, SW>

    Szn] using

    Eq. 2.

    The moment at n resultingfrom the current is obtainedby knowing the current forces and

    moments at n-1 [Fig. 12] Wd then applying equilibriumto the n-1 segment:

    % = .%-1+ n-l 4 Os Yn-l h Xn-l - %-l 2 b yn-~

    4 yn-l

    ,4/2cos Qyn , sin QX_ ~ +4 qti ~ Q/2sin gti-l

    --

    ?

    = %-1 -1-;n-l ~ COS @fi-l SiilQm-l - F~-l ~ s~ gyn-l

    1

  • 8/10/2019 OTC-1071-MS Offshore Pipelines

    10/14

    n 18

    ST@EE-ANKLYSTS OF OFFSHOREPTPELTNT SllTTRTNGNSTALLATION OTc 10;

    .

    .-.

    -.. ..- --.-

    ),

    .$2/2(~j sfi ~-l - zj

    Os m-l in %-l

    &q

    --1

    -F&4 cos Qmlsin Qh1 +F&l Aces Qyn&os Qh-l

    [191

    = ~n-~ -

    - 22/2 (qzj

    Cos Yn-l

    ),

    %

    = ~-~ + ~-~ L in Yn-~ - ~n-~ 2 Os fi-~ Os yn-~

    + .42/2(qzj sin ~-l

    h Xn-l + %d Os fi-l)

    where the force componentsat n-l resulting from current forces are:

    n-2

    1 .

    I

    1

    A Xj

    j=l

    n-2

    ~~1 .

    z

    4qYj + . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .[20]

    j=l

    n-2

    ~q.l .

    Z

    A q~j

    j=l

    The total moment componentsin global coordinatesare obtained from applyingmoment equilibrium

    about n [Fig. 11]:

    % = ~-yn(Fzl

    b,

    FL) + Zn (F~ + F~

    n-1

    c~)+~Qj zn

    +Wes zn-zn

    -zJ)+M ,

    j=l

    l =

    ~+xn ~zl+~~)-zn Fn+F~)+qn ,

    =

    ~1 - Xn ~n + FL) + Yn Fn + FL)

    }

    . . . . . . . . . . . . .

    [21]

    .

    n-1

    x,) +%

    g)+ ~Qj(Xn-J

    we s (Xn - ~

    j=l

    These moment componentsare transformedinto local coordinatecomponents [Mn,

    Y

    , Mzn] using

    Eq. 2.

    The coordinatesof.the ceriter-ofravity for point n are obtainedfrom t e center of

    gravity of n-l:

  • 8/10/2019 OTC-1071-MS Offshore Pipelines

    11/14

    Tfl1071

    -J. T..PQWERS anl_L. D. FINN __

    II 19

    x;~ =

    [

    &

    ~-~ -A)+ Xn+x

    J

    4/2]

    s

    )

    ....................

    [

    .[22

    ~:g . ~::1

    1

    s-2)+ n+znJ4/2 /s

    In Eq.s.

    13 through 22 the coordinatesXn, Yn and ~ and the sloPe ~ and @yn are not known

    until the forces and moments have been calculated. If the segment length is short, the change in

    slope and deflectionof n-relative.ton-l is small. Very little error in the csJ-culatedforces

    and moments will result if it is assumed that the slope

    }

    Yn = Yn-l , snd

    . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

    .

    [23

    in = Xl-1

    and the coordinates

    }

    .

    .[24

    z = Znl

    n

    4 tiom-l Cos Yn-l

    J

    After the forces have been obtained,the deflectionand change in slope at Point n can be calcu-

    lated and an improved slope and coordinatesof Point n obtained.

    This cycle can be repeated

    se~eral times for each segment, if necessary.

    The effectivetension at Point n is defined as the axial force at n which has an effect on

    the moment, shear and deflectionsof the suspendedportion,

    en = xn

    tWb[H-yn] . . . . . . .:. . . . . . . . . . . . . . . . . . . . . . . .

    .[25

    1

    1

    1

    1

    It is apparentfrom Fig.

    11 that the term wb[H-Yn] could have been left out of the force summa-

    tions F~, Fyn and FZn, and we

    still would

    have obtained the value for the shears ~ and Szn,

    since this term is first broken into vertical and horizontal componentssad then transformedback

    to a single axial component.

    The axial componentthus obtainedwould equal the effective tension

    defined aboye. Although the pipe is not actually capped at Point n, the capped end axial force

    is present at that point. The sxial stress should include this force and hence should be based

    on s=

    as previously defined.

    The uniformly distributedloads normal to the pipe acting on segment n-1 are:

    Y =

    p-l -

    e

    Os Yn-l>

    )

    ..............................

    [26]

    w=

    z

    q

    zn-1

    Deflectionsand Rotations

    The coordinatesof Point n are determinedfrom the deflectionof that point relative to the

    tangents at n-l by imaginingthis sectionas a csatileveredbeam fixed at Point N-1 and loaded

    by shear forces Sp and Sznj moments M

    and &n, effectivetension Ten) ~iformly distributed

    forcesw

    f

    and Wz, snd buoy force Qn.

    ~e local coordinatedeflectioncomponentsfor element n-l

    as calcu ated by small angle bending formulas are:

    r

    >-n = l/EI Mzn ~,2/2+ (S

    + Qn COS 0

    yn

    yn-~ ) 43/3 + T?yIkp

    1

    j

    )

    ...........

    [27]

    z

    .

    l/EI -M

    42/2 + Szn

    ,43/3 +WzA4/8

    n

    w

  • 8/10/2019 OTC-1071-MS Offshore Pipelines

    12/14

    w. -. -- . . . . . . . . . . ..- --

    --- - .. ~_ .__.

    Similarly,the angle changes are:

    m. =

    Yn

    @I

    [

    Mznj + (Syn +

    ~ c Qyn.l)A2/2 + Wy 43/6

    Qzn/EI -M

    ~ + Szn

    P

    &2/2 +

    WZ A

    3/6

    1 f

    .

    . . . . . . . . . . .

    [28]

    The

    new coordinatesfor Pointn are:

    D

    n-1

    ~

    n-1

    z

    n-1

    +

    [

    Cs h- l CsYn-l

    I

    h l

    i n XI I - lsYn-l

    =Os

    xi v- li n h- l

    Cos%-l-l

    sin Xl&l in %-l-l

    -i

    si n L-l

    o

    cos Q

    Xn

    . [29]

    11 20

    -:C ITREW

    ANAT:YSTti m? OliFSHOREPIPELI~S DURING INSTALLATION

    OTC 1071

    and the new angles are

    9

    Yn

    =Q

    ynl+AQ

    Yn

    [ 1)

    in AQ

    Xn

    = %-l-l

    + arc tan

    zn

    .

    . . . . . . . . . . . . . . . . . . . . . .

    [30]

    Cos Q

    Yn-1

  • 8/10/2019 OTC-1071-MS Offshore Pipelines

    13/14

    =

    _ PORTION SUPPORTED

    g

    LIFT-OFF POINT

    %

    e

    *

    n

    %

    SUSPENDED

    =

    s

    z

    g

    .

    Fi g. 1 - Layi ng pi pe by sti ng er- l aybarge method. 3

    0

    \ 19.8

    04812162024 2832364044

    ANGLE AT LIFT-OFF POINT DEG)

    Fig.

    2 - Pipe stress vs lift-off angle in 200 ft.

    of water 2o in. 0.0., .625 in. wall thickness,

    1.3 empty specific gravity pipe).

    200

    180

    160

    140

    120

    100

    80

    60

    40

    20

    00 ~ ~ ,2 ,6 Z. 24 28 ~z 36 ~. 44, -

    A NGL E A T L IF T- OF F PO IN T ( DE G)

    Fig. 3 - Lift-off height vs lift-off ang

    200 ft. of water 20 in. O.D., .625

    i n,

    thickness,

    1.3 empty specific gravity p

    F1g

    100

    [

    FINITE ELEMENT

    ------STIFFENED CATENARY

    80

    - NATURAL CATENARY

    ,4 K

    BEAM COLUMN

    60 -

    40

    ---- ~ >_-

    _-

    ~~oK

    /

    /

    r

    I

    00 816 24

    32

    40

    48

    ANGLE

    AT LIFT-OFFPOINT DEG)

    4 -

    Comparison of maximum stress vs lift-off

    angl e for various tensions in 250 ft. of water

    e in

    as predicted by various theDries 24 in. 0.0.

    wall

    .625 in. wall thickness,

    1.2 empty specific

    pe).

    gravity pipe).

    250

    20K

    1

    TENSIO~

    d-

    20 K 60K

    60K

    200

    , F NITE ELEMENT

    ~ - -- -- - ST lFFE N6~ CAT EN ARY

    /

    /

    150 -

    /

    -

    NAIURA: CATENARY

    ,

  • 8/10/2019 OTC-1071-MS Offshore Pipelines

    14/14

    FINITE ELEMENT

    ~

    5TIFFENED CATENARY

    z

    o

    - NATURAL CATENARY

    g

    BEAM COLUMN

    g 500 -

    ~

    ~

    9

    0

    0

    500 1500

    HORIZONTAL POSITION FT)

    Fig. 7 - Comparison of deflected shapes for 60

    kips tension in 1,000 ft. of water, as predicted

    by various theoriss 16 in. O.D., .50 in- wall

    thickness,

    1.2 smpty specific gravity pipe).

    Y

    K

    I

    20 ,

    I

    I

    I

    1

    I

    I I

    \

    \-

    \

    \

    \

    10

    \

    FINITE ELEMENT

    .-----sTIFFENED CATENARY

    -NATURAL CATENARY

    BEAM COLUMN

    2000

    LOCATION ALONG PIPE FT)

    Fig. 8 - Comparison of bending stress diagrams

    for 60 kips tension in 1,000 ft. of water, as

    predicted by various theories 16 in. 0.0.,

    .50 in. wall thickness, 1.2 empty specif i c

    gravi ty pi pe).

    x

    Fig.

    10 - Fluid pressure on beam element

    /

    z

    Fi g. 9 - Global- and local coordinates.

    nds capped).

    I

    MY.;

    /,

    Fig. 11 - Three-dimensional fores diagram

    Y

    M :n

    1/

    M ;n

    M;n.I

    M ~n

    M;n-l Fq

    Xn-1

    4

    \

    I

    /

    X

    r

    /

    N

    /

    ~q x,

    I

    /

    Yn-1 ~,

    1

    /1

    I ,/

    \

    , 1,

    z

    w

    M :n-1

    Fig. 12 - Current-induced forces.