Fatigue Behavior of Dented 1-42

download Fatigue Behavior of Dented 1-42

of 42

Transcript of Fatigue Behavior of Dented 1-42

  • 7/28/2019 Fatigue Behavior of Dented 1-42

    1/42

    ContractNo.DTRs56-95-C-0003

    TE XA S TRANSPORTATION /NSTITUTE

    Fatigue Behavior ofDented

    Petroleum Pipelines

    -

    Task 4

    in cooperation with theU.S.DepartmentofTransportation

    Research andSpecialProgramsAdministrationOfficeofPipeline Safety

  • 7/28/2019 Fatigue Behavior of Dented 1-42

    2/42

    I RcponNoTask 4

    3 Recipient's Catalog No2. Government AccessionNo.

    16 AbsPact

    Dents ispipelines can seriously reduce their design lives. Dents cause the developmentofstressconcentrations which make dents susceptible to fatigue failures. Acceptance criteria ofdents is currentlybased on dent depth alone. To understand the behavior of dents in pipelines, a research program involvingboth experimental testing and finite element analysis was conducted. Fifteenpipe specimens ranging indiameters from 12 in. to 36 in. withwall thickness ofeither 1/4 in. or 3/8 in.Various types of damage werestudied, including dents due to rocks, dents formed with backhoe teeth, andshort longitudinal dents with

    '

    simulated damage. The effect of dent restraint was also investigated. Two typesof fatigue failures wereobserved: cracks developing fiommachined gouges in the dent troughs and periphery cracks at plain(without machined gouges) dents. Three dimensional shell element finite element models were given elasto-plastic behavior for modeling dents inpipelines. Parameters modeled include dent depth, dent type, dentrestraint, pipe diameter, thickness, grade, longitudinal stress,pipe support, and pressure at indentation. For

    the given parameters, the reboundbehavior and stress behavior of dentswas investigated. The displacementdatawas studied to understand the rebound characteristicsofdents. From the finite element modeling, therebound characteristics canbepredicted for a broad range of values for the various parameters.

    1 Title and SubtitleFATIGUE BEHAVIOR OF DENTED PETROLEUM PIPELINES(TASK4)

    5 . Report Date-May 19976. Performing Organization Code

    7 -2UthOr(S)

    PeterB. Keating and RogerL. HoffmannTexas Transportation Institute

    College Station, Texas 77843-3 135

    COTR:U.S. Department of TransportationResearch and Special Programs Administration

    400 Seventh Street, SW, Room 2335

    Washington, D.C. 20590-0001ATTN: Mr. Gopala Vinjamuri

    9 Performing OrganirauonName and Addms

    The Texas A&M University System12 Sponsoring Agency Name and Address

    Offce ofPipeline Safety (DPS-13)

    8 Performing Organmion Report NoTask 410 WorkUnitNo (TRAIS)

    I 1 ContractorGrant NoContractNo. DTRS56-95-C-0003

    Final:June 1995 to February 1997

    13 Typeof Report and Pen04 Covered

    14 Sponsoring Agency Code

    17. Key Words

    Pipelines, Fatigue, Dents, Rupture, Damage18. DistributionStatcment

    No restrictions.

    19 SecurityClassif.(ofthis report) 20. Security Classif.(of this page) 21. No. ofP a g e sUnclassified Unclassified 309

    22. Pnce

  • 7/28/2019 Fatigue Behavior of Dented 1-42

    3/42

    Fatigue Behavior ofDentedPetroleum Pipelines

    Final Report

    May 1997

    For

    Office ofPipeline Safety

    U.S. Department ofTransportation

    PeterB.Keating, P bD .Associate ResearchEngineer

    RogerL. HofitinannGraduate Research Assistant

    Texas TransportationInstituteTexas A&M University SystemCollege Station,Texas

  • 7/28/2019 Fatigue Behavior of Dented 1-42

    4/42

    DISCLAIMER

    The contents ofthis report reflect the views ofthe authors who are responsible forthe

    facts and the accuracy ofthe datapresented herein. The contents do not necessarily reflect the

    official views orpolicies ofthe Office ofPipeline Safety,U.S. Department ofTransportation.This report does not constitute a standard,specification, or regulation.

    i

  • 7/28/2019 Fatigue Behavior of Dented 1-42

    5/42

    TABLEOF CONTENTS

    -

    ACKNOWLEDGMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .SUMMARY ................................................................. vi

    CHAPTERONE

    -

    A

    -

    I

    .

    INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .PROBLEMSTATEMENT ............ ... .. . ............. .. .. . .. .. . .1CURRENT ACCEPTANCE CRITERIA . ............ .......... . . . .... 21.2.1 ASME B31.4 (1992 Edition) .. .... ... . . . . ... .... .. .... ....... .. . 31.2.2 ASME B31.8 (1992 Edition) . .. .. .. ... ... . . . . . .... . .. .... . . . . . . 31.2.3 Repair Procedures .. .... ... .. ....... . .. . .. .. .. . . ....... . . . ...4DAMAGETYPES.. ... ... .. . .... . .. ... . . .. ....... . . . ... ... . .... . ..4DENTING PROCESS .. . ....... . . . . . . ... ....... .. . . . . .. ..... ... . .. 6OBJECTIVES OFTHE STUDY .. . .. ....... .. ..... . ... . . . . . . .. .. ... . 8LIMITATIONSOF THE STUDY .... ........ . . . . . . ... .. . . . . .... .. . . . 8

    1 . 1

    1.2

    1.3

    1.4

    1.51.6

    CHAPTER TWO

    REVIEW OF EXISTING DATA AND RESEARCH RESULTS ........... .. . .... ... . . 9INTRODUCTION .. .... ........ ...... ... . . ..... .. ..&....... ... .. 9CANADIAN STANDARDS ASSOCIATION TECHNICAL COMMITTEE .. 92.3.1 Expeiimental Program .. .... . . ....... .. . ....... ... . .. ... ... . . . 1 12.3.3 Finite Element Analysis ... . .. . . .... ..... ... . . . ....... . ... ... . 13

    2.3.4 Conclusions . ... .... .. ...... . .... . . .... ... .. ..... .... ... .. . 14REVIEW OF DOCUMENTS PERTAININGTO PIPELINE OPERATIONS . 1 5

    2.12.22.3 PRC-AGA RESEARCH PROGRAM ... ... ... . ... ..... .. . .. ...... ., . .11

    2.42.5 CONCLUSIONS'............ ... .......... ... ..... . .. ....... ... .,.i5CHAPTERTHREE

    EXPERIMENTAL PROGRAM3.13.23.3

    INTRODUCTION .... . . .... .. . .. . . .. .. ...... .. ... ...............17PIPE SPECIMENS .. ... .... .... .... .. . . .. . ... ..... ... ... .... . .. . .17DENTTYPES ... ..... ...... .................. . . .......... ..... . . 183.3.1 Dentsand Machined Damage (Type A) .... .... ...... ... .. ......193.3.2 DentsCaused by Backhoe Teeth(TypeBH-L&BH-T) ........... . .233.3.3 RockDents (Type R) . . . . ... . . . ..... ... ..... .. ...............23RESTRAINT . . ..... .. .... ........ .... .. ........ ..... . ... . . ... .. .30SPECIMEN PRESSURIZATION .. ..... ... . ............ .. . .... . ... . .323.6.1 Hydraulic System ... . ...... ....... .... ......... ..... ....... .32

    3.4 DENTING. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 43.53.6

    ..11

  • 7/28/2019 Fatigue Behavior of Dented 1-42

    6/42

    . .

    .

    ...

    I

    .

    .

    .

    -93.6.2 Pressure Histones . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .333.7 FATIGUETEST PROCEDURES .................................... 37

    . 3.8 FATIGUE CRACK REPAIR PROCEDURE . . . . . . . . . . . . . . . . . . . . . . . . . . .393.9 SUMMARYOF RESULTS .................................... i . . .42

    3.9. I Denting Forces ............................................. 423.9.2 Restraint Forces ..................1 ......................... 463.9.3 Dent Rebound Behavior...................................... 50

    Unrestrained TypeA Dents ................................... 51TypeBHD- ............................................ 543.94 Fatigue BehaviorofDentswithMechanical Damage (TypeA) . . . . . . .56Unre-d TypeA De@ .................................. 1 56d TvDeA DQ& .....................................583.............................. 59

    ofTvbeA Dent Behawor ............................603.9.5 Fatigue BehaviorofDents fiom Backhoe Teeth (Type BH-L andBH-T)603.9.6 Fatigue BehaviorofRock Dents (Type R) ........................663.9.7 Yield Stress Determination ................................... 693.9.8 Effect ofStraps on Hoop Stress ................................ 71

    CHAPTERFOUR

    FMITE ELEMENT ANALYSIS4.1 INTRODUCTION ................................................ 74

    4.2.1 Dent Parameters ............................................ 744.2.2 Pipe Parameters ............................................ 78FMITE ELEMENT MODEL ....................................... 834.3.1 Mesh Design and Behavior ................................... 834.3.2 Model History andOutput ...................................... 894.3.3 Problems EncounteredwithModels ............................ 91REBOUNDBEHAVIOROFUNRESTRAZNEDLONGITUDINALDENTS .944.4.1 Introduction ............................................... 944.4.2 Dent Rebound Behavior...................................... 964.4.3 . Dent Type A Rebound Behavior ':............................. 1064.4.4 DentTypeBH Rebound Behavior.................. .......... 117

    Dent Type G Rebound Behavior .............................. 124Dent Type H Rebound Behavior .............................. 132

    STRESSBEHAVIOROF UNREST"ED LONGITUDINALDENTS ... 138Dent TypeA StressBehavior ................................ 140Dent Type BH Stress Behavior ............................... 153DentTypeG StressBehavior ................................ 163

    4.2 PARAMETERS .................................................. 74

    4.3

    4.4

    4.4.54.4.6

    4.5.1 Introduction 1384.5.24.5.34.5.4

    4.5

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

    4.5.5 Dent Type H Stress Behavior ................................ 169

    ...

    iii

  • 7/28/2019 Fatigue Behavior of Dented 1-42

    7/42

    4.6 BEHAVIOROF U N R E S " E D TRANSVERSE. SPHERICAL. ANDPLATEDENTS ................................................. 1714.6.1 TransverseDents .......................................... 1714.6.2 Spherical Dents ........................................... 1844.6.3 Plate Dents ............................................... 188

    4.7 RESTRAINED DENTS ........................................... 1944.7.1 Introduction .... ......................................... 1944.7.2 PeriphedFailures ......................................... 1954.7.3 Inside Surface Cracking Under Restrained Rocks ................. 199

    4.8 MISCELLANEOUS PARAMETERS ................................ 2084.8.1 Longitudinal Stress ........................................ 2084.8.2 Pipe Grade and Pressure History .............................. 2094.8.3 Support ConditionsDuring Indentation ......................... 2144.8.4 PressureDuring Indentation ................................. 219COMPARISONS OFMODELSWITHEXPERDENTAL TESTSPECIMENS ................................................... 2224.9.1 Introduction ............................................... 2224.9.2 Rebound Behavior ............................. 1 ...........2234.9.3 Failure Modes ............................................ 227

    4.9

    CHAPTERFIVE

    ACCEPTANCE CRITERIA OF DENTS ......................................... 231

    CHAPTERSIX

    CONCLUSIONSAND RECOMMENDATIONS ..................................242

    CONCLUSIONS ...................................... .........2426.1.1 Rebound Behavior .............................. :..........2426.1.2 Stress Behavior ........................................... 2426.1.3 Restrained Dent Behavior ................................... 2436.1-4 Acceptance Criteria ........................................ 243

    6.2 RECOMMENDATIONSFORADDITIONAL RESEARCH ........ ....243

    6.1

    REFERENCES ...............................................................246

    APPENDIX ................................................................ 247

    iv

    ..

  • 7/28/2019 Fatigue Behavior of Dented 1-42

    8/42

    ACKNOWLEDGMENTS

    The fatigue test program andits results reported are the results ofa research project sponsoredby

    the Office of Pipeline Safety, U.S. Department ofTransportation ( SolicitationNo.DTRS56-95-R-0003) The research wasperformed by the Texas Transportation Institute, TexasA&M University System in College Station, Texas. The research project was coordinated forthe Office ofPipeline Safety by William Gute; Gopala (Krishna)Vinjamuri, Materials Engineer,provided technical direction as the Contracting Officers Technical Representative.

    I

    V

  • 7/28/2019 Fatigue Behavior of Dented 1-42

    9/42

    SUMMARY

    The fatigue behavior of damaged on-shore petroleum pipelines was investigated both

    experimentally and analytically. Various types of damage were studied, including dents due to

    rocks, dents formed with backhoe teeth, and short longitudinal dents with simulated gouges. A

    dent causes a localized concentration of the pipe wall stresses that result from internal pressures.

    Fluctuation of the pressure can result in fatigue crack initiation and propagation. The effect of

    dent restraint was also investigated.

    A total of fifteen specimens were fatigue tested. These specimens had a range of nominal

    outside diameters from 12 in. to 36 in. with wall thickness of either114 in. or3/8 in. Thisresulted in a range of diameter-to-thickness ratios from 34 to 96. The number of dents per

    specimen varied from 10for the smaller diameter pipes to 3 for the larger diameter pipes. Dent

    depths rangedfrom5 to 18.75percent of the pipe diameter. Each specimenwas subjected to a

    hydraulically varied pressure distribution to simulate actual pipeline pressure conditions. Each

    specimen was pressure-cycled until all the dents in the specimen failed or 100,000 cycles hadaccumulated. Failed dents, asdefinedby leakage, were either repaired or removedfromthe

    specimen by sectioning.

    -

    Three types of dents classifiedastypeA, BH, andRwere tested. Type A consists of a

    four-inch longitudinal dent. TypeBH consists of longitudinal and transverse dents caused by a

    backhoe tooth. TypeRconsists of rocks. TypeA andBH dents were tested with andyithoutrestraint. Type Rdents were tested with restraint. Mechanical damage including machinedgouges and scratches were tested with alltypeA dents.

    The dent rebound behaviorwas studied for the unrestrained dents. Dent reboundincreases with an increase in the diameter-to-thickness ratio. TypeA dents experienced more

    rebound than Type BH dents. The rebound ofTypeA dents allowed the center of the dent to

    bulge to put the area in reversecurvature. Thisbehavior is relatedto the length of the dent. Type

    BH dents did not have the behavior ofTypeA dents due to the short dent length andsharpdent

    Vi

  • 7/28/2019 Fatigue Behavior of Dented 1-42

    10/42

    e

    geometry. The dent Type A rebound after pressurization of dentswithdifferent depths yielded

    similar final depths which makes it impossible to know the extent ofdamage if only the final

    dent depth is known.

    Three types of fatigue failures were observed experimentally: cracks developing in the

    troughs ofTypeA unrestrained dents, periphery cracks in all dent types, and internal cracks in

    the trough of restrained rocks.As expected, dents with deeper initial dent depths exhibited

    shorter fatigue lives thanshallower dents. However, due to the rebound after denting and during

    pressurization, it is difficult to determine the initial dent depth. The unrestrained Type A dents

    had relatively low fatigue lives due to the dent rebound behaviorin addition to the mechanical

    damage. The fatigue lives ofdents with periphery cracks is longer. Restraint ofTypeA dents

    did not allow the rebound, causing peripheral cracking insteadofcracking in the dent trough.

    Failures of Type A dents occurred at all dent depth-to-diameter ratios tested. Failures of Type

    BH only occurred for dent depth-to-diameter ratios of ten and above. The failures from intemal -cracks under rocks hadsimilar fatigue lives asType BH failures. Only a few of these cracktypesdeveloped.

    A finite element parametric study was performed to understaud the behavior of dents in

    pipelines. Three dimensional shell element models were given elasto-plastic behavior formodeling dents in pipelines.Variousdent and pipe parameters were modeled. The dent

    parameters include dent depth, denttype,and dent restraint. Dent type includes geometry

    considerations such as dent length and dent orientation. Pipe parameters include diameter,

    thickness, grade, longitudinalstress,pipe support, and pressure at indentation. For the given

    parameters, the rebound behavior andstressbehavior of dents was investigated. The

    displacementdatawas studiedto understand the rebound characteristics of dents. Stress data

    was studied to determine failure modes of different combinations of dents and pipes.

    Restrained dents were modeled to understand the failure modes found in the experimental

    program. Spherical restrained dents were used to simulate dents caused by rocks. Restrained

    longitudinal dents were alsomodeled.

    vii

  • 7/28/2019 Fatigue Behavior of Dented 1-42

    11/42

    Models were compared with data from the experimental program to verify the results ofthe models. Rebound data and failure modes were compared between models and

    experimentally tested dents.

    viii

  • 7/28/2019 Fatigue Behavior of Dented 1-42

    12/42

    Chapter One

    INTRODUCTION

    1.1 PROBLEMSTATEMENT

    Pipeline failures are often associatedwith fatigue crack development. Fluctuation in

    pressure fiom pumping, scheduled pressure testing, pressure surges, and product density changes

    causes stress cycles to accumulate over the life of pipes. In the design and evaluation of

    pipelines for fatigue, both the longitudinal seam and the full penetration groove weld used to

    joint pipe sections are critical. Current fatigue design provisions, coupledwithcompetent

    construction and inspection, yield satisfactory performance over the intended service life of the

    pipeline. However, accidental damage to the pipe can result in a defect, suchk a gouge or dent,with a stress condition more severe than the designed welds. Gouges, nicks, or scratches

    typically result in initial flaws that are largerthanthose tolerated in the pipeline weldments.

    Stress concentrations caused by dents increase the range ofthe stress cycles. The increase in

    flaw sizes and stress concentrations accelerate the growthoffatigue cracks which can leadto

    leakage rupture of pipelines.

    Two types of pipeline dents are typically categorized: excavation dents and settlement-

    induced dents. Excavation dents occurfiom contactk t h earth-moving equipment (backhoebucket teeth, for example) during construction or Serviceofthe pipeline. The excavation dents

    are usually accompanied by gouges ordamage that contain initiation sites for fatigue crack

    development. These dents aremost often found in the upperhalf of the pipe due to the locationand directionofexcavation. Settlement-induced or rock dents occur when a pipeline is

    inadvertently laidonorsettles onto an unyielding rock. The weight ofthe pipe, the soil

    overburden, hydrostatic test water, orproduct can cause thepipe to deform around the rock.

    Unlike excavation dents, settlement-induced dents are unable to rebound with pressurization,

    thus changing its fatigue behavior when compared to unrestrained excavation dents. Settlement-

    induced dents are found on the bottomhal fof the pipe.

    -1-

  • 7/28/2019 Fatigue Behavior of Dented 1-42

    13/42

    Acceptability of dents in pipelines is currently based on the dent depth-to-diameter ratio

    (&D)at the time of detection without regard to diameter, thickness, or any aspect of the dentgeometry. This method does not account for differences in dent shape or pipe variables.

    Acceptance based strictly on dent depth may leadto overly restrictive estimates of damage so as

    to include the entire spectrum of different dent types for all pipe sizes. This, in turn,may result

    in unnecessary repair of damaged pipelines that may otherwise have satisfactory performance.

    Conversely, certain combinations of dent types and pipe sizes may result in fatigue failure from

    dents acceptable under the current criteria. Without a clear understanding of the denting.

    mechanismand what types of dents can form under various load conditions, a single measure ofdent depth may not always provide information to estimate dent criticality with a reasonable

    degree of certainty. No knowledge of the dent history (initial dent depth, rebound characteristics,

    etc.), which influences dent fatigue behavior, may lead to unreliable acceptance or rejection of

    dents. Thus,an improvement in the acceptance criteria can improve the operational reliability of

    the pipeline, aswell as reducing costs by providing a rational method of dent evaluation. This

    will require an understanding of the possible failure modes forvarious combinations of dent

    shapes and pipe sizes.

    1.2 CURRENT ACCEPTANCE CRITERIA

    The design, fabrication, and operation of pipelines for the transportation or transmission

    of both liquid and gas petroleum product is governed by the pressure piping codes issued by the

    American Society of Mechanical Engineers (ASME). Separate codes areprovided forboth

    liquid and gas petroleum pipelines. Gaspipelines typically operateat higher pressures than

    liquid pipelines. Due to the compressibility of gas, thepossibility ofan explosive rupture is

    much g r a t a thanin a liquid pipeline, and the dent acceptancecriteria is somewhat morerestrictive. The following is abriefsummary of the current acceptance criteria for both gas and

    liquid petroleum pipelines, aswell as the requi redrepairprocedures.

    -2-

  • 7/28/2019 Fatigue Behavior of Dented 1-42

    14/42

    -

    -

    1.2.1 ASME B31.4 (1992Edition)

    The ASME B31.4pressure piping code provides for the design, inspection, and operation

    of liquid transportation systems for hydrocarbons, liquid petroleum gas, anhydrous ammonia,

    and alcohols. This code defines a dent as a gross disturbance in the curvature of the pipe. A

    gouge, scratch, or groove is any local disturbance that results in a reduction of wall thickness.

    The acceptance criteria for dents and gouges is given as:

    0

    0

    0

    0

    Dents which affect the curvature of the pipe at a seam or girth weld mustberemoved by cutting out the damaged portion of the pipe as a cylinder.

    Dents that exceed6percent of the nominal pipe diameter (greater thanNPS4 insize) are not permittedinpipelines that operate at a hoop stress greater than 20

    percent of the specifiedminimumyield strength of the pipe.

    Dents containing a stress concentrator, such as a scratch, gouge, groove, or arcburn mustbe removed.

    Gouges and grooves having a depth greaterthan12.5percent of the nominal wallthickness mustbe removed or repaired.

    1.2.2 ASME B31.8 (1992Edition)

    The ASME B31.8pressurepiping code provides for the design, inspection, and operationof gas transportation and distribution piping systems. A dent is defined by this code as a

    depression which produces a gross disturbance in the curvature of the pipe wall. The dpeth of

    the dent is measuredas the gap between the lowest point of the dent and a prolongation of the

    original contourofthepipe in any direction. A scratch or gouge is definedas a depression that

    reduces the thicknessofthe pipe wall. The acceptance criteria for dents and gouges is given as:

    0 Dents which affectthe curvature of the pipe at a seam orgirthweld mustberemoved.

    Dents that exceed2percent of the nominal pipe diameter (greater thanNPS 12 in

    -3-

  • 7/28/2019 Fatigue Behavior of Dented 1-42

    15/42

    size) are not permitted in pipelines that operate at a hoop stress greater than 40percent of the specified minimum yield strengthofthe pipe.

    0 Dents containing a stress concentrator, such asa scratch, gouge, groove, orarcburnmustbe removed by cutting out the damaged portion ofthe pipe as acylinder.

    Gouges and grooves having a depth greaterthan 10percent of the nominal wallthickness mustbe removed or repaired.

    1.2.3 RepairProcedures

    The repair procedures for both codes are similar. Scratches, gouges, and grooves may be

    removed by grinding, provided the wall thickness is not reduced below 90percent ofthe nominal

    wall thickness required for the operating design pressure. When too large a reduction in the wall

    thickness results, the damaged section of pipe mustbe cut out and replaced, though the ASME

    B31.4 (liquid) code allows a complete encirclement welded patch. Insert patching is prohibited

    in both codes. ASME B31.4 (liquid) allows a gouge tobe repaired by the use of an approved

    welding procedure while ASME B31.8 (gas) does not.Any dent which contains a scratch,

    gouge, or groove must be removed by cutting out the damaged portion of the pipe asa cylinder.

    Dents without gouges that exceed the specified depths mustbe removed by cutting out the

    damaged portion of the pipe as a cylinder. The ASME B31.4 (liquid) code allows for the use of

    a full encirclement welded or mechanically applied split sleeve.

    -

    1.3 DAMAGETYPES

    Indirectly,the ASME dent acceptance criteria classifies dents into two categories: dents

    with damage and plain dents. The damage that occurswithdenting is usually associated with the

    contact force casing thisdamage. Different dent formation processes can lead to different dent

    damage, which in turnwill alterthe fatigue behavior. Consequently,the typeof damageassociatedwitha dentbecomesan important criterion for dent acceptance. It is important to

    -4-

  • 7/28/2019 Fatigue Behavior of Dented 1-42

    16/42

    note, however, that damage residing in the pipeline that can lead to fatigue failure may not be

    easily associated with dents. Dent rebound behavior often playsan important role in determining

    fatigue behavior.

    Damage, in the form of dents and gouges, can be introduced into the pipeline at various

    stages of construction and operation. The following discussion summarizes the possible damage

    associated with each stage.

    Shipment Damage - Damage that occurs during shipment of the pipe from themanufacturer/supplierto the installation site. If pipe segments are improperly secured duringtransit by either rail or truck, dents and/or fatigue damage may occur. Inadequate support along

    the pipe length can results in excessive high-bearing forces which in turn cause denting or

    localized bending. The repeated flexing ofthe pipe wall at a support point can initiate and

    propagate a fatigue crack. Once the pipe is offloaded, the pipe wall may return to its original

    shape making detection of the fatigue damage difficult. It wouldbe expected that any dent

    formed during shipping wouldbe cause for rejection ofthepipe segment unless adequately

    repaired. Shipping damage occurs most often at the ends of the pipe where, as aminimum, theyare supported.

    -

    Construction Damage - Dents that occur during construction of the pipeline. These dentscan occurfrom chains orstrapsusedto lift and move thepipe and by earthmoving equipment.

    Also,they canbe caused by preexisting dents prior to construction which hadnot been detectedor removed. Constructiondents can exist anywhere around the circumference of the pipe, but aremost prevalent in the upperhalfofthe pipe cross section.

    ServiceDamage

    -Dents

    thatoccur during the service life ofthe pipeline. The dents are

    typically from earthmoving equipment and can vary widely in t e r n ofsize andseverity. Thesetypes of dents are always located on the upperportion ofthe pipe since they are associatedwith

    excavation above.

    -5-

  • 7/28/2019 Fatigue Behavior of Dented 1-42

    17/42

    RockDamage - Dents that form due to thepipebeing laid on the rock during construction(improper bedding) or settlementofthe pipe onto the rock during service. Both local and global

    bending stresses can result due to the settlement of the pipe (Fig. 1-1). Flexible pipe design

    requires the consideration of the bending stresses that developwithdifferential settlement along

    the pipe. A uniformsettlementwill not resultinbending stresses unless the pipe settles on arigid object such asa rock. Inthis case, only localized bending stresses occurin the dent region

    ifthe pipe wall is relatively flexible. Differential settlement of the pipe onto a rock oran

    increase in wall stifkess may result in a superposition of global bending stresses in the dentedregion. This can lead to altered fatigue behavior.

    \ !W

    Figure 1-1: Schema of local and global bending due topipe settlement on rock.

    1.4 DENTING PROCESS

    As an object is forcedagainst theouter surface of a pipe, a resistance to this force is

    developed by in- forces (both plate bending and membrane) in thepipe wall aswell as theintemal pressure. If the denting force is sufficient in magnitude, the plate bending moments will

    cause inelastic behavior due to the yielding ofthe extreme, outerfibersofthepipe wall. Large

    bending strains can result in the development ofa fullplastic hingeat the time ofmaximumindentation.

    -6-

  • 7/28/2019 Fatigue Behavior of Dented 1-42

    18/42

    The depth of the dent is typically measured in terns of percent pipe diameter or the dlDratio. This ratio for a given dent will assume several values depending on the stage of the dent

    forming process. The initial dent depth (dQiniMis the maximum depth of the dent (in terms ofpercent). For construction and service dents, this occurs at themaximumapplication of the

    denting force. For rock dents, this ratio increases with time as the pipeline continues to settle

    onto the rock.

    The initial dent depth is a function of the internal pressure at the time ofthe application of

    the denting force. Higher internal pressures will result in shallower dents for the same denting

    force magnitude. Construction dents can form in the absence of internal pressure.

    Once the denting force is removed, an elastic bound of the dent will occur. .The amount

    of rebound will be a function of the degree of restraint.Most construction and service dents are

    unrestrained and can therefore rebound through their fully elastic limit. The rebound dent depth .

    ratio is given as (dD)*& The depth will almost never reboundtozeropercent since somedegree of plasticity will occur at ( L I / D ) ~ ~ ~ .Rock dents do not rebound unless the pipe settlementis reversed.

    The important consequence of the denting process is the residual stress distribution thatremains in the pipe wall. The localized stresses, particularly those that are at the outer portions

    of the wall thickness, will govern the fatigue behavior of the dent since the location of the initial

    defect for fatigue crack development will coincide. Compressive residual stresses resulting fiom

    the dentingprocess canwillhelp suppresscrack formation. Conversely, tensile residual stresses

    can accelerate crack growth rates. The dent history then becomes an important consideration for

    determining dent acceptance. Estimating the fatigue life of a dent using only the final detected

    depth cannot always provide reasonable predictions.

    -7-

  • 7/28/2019 Fatigue Behavior of Dented 1-42

    19/42

    1.5 OBJECTIVESOFTHE STUDY

    The primary objective of the current study is to develop a fundamental understanding of

    the fatigue behavior of dents inpipelines. Thiswill form the basis for the refinement of dent

    acceptance criteria. Specifically, the following objectives:

    0 Investigate the fatigue behavior of various dent types to determine if dentparameters in addition to the dent depth need to be considered.

    0 Investigate the effect of restrain on the fatigue behavior of dents. Restrain limitsorprevent rebound under cyclic pressure loading, thus limiting the tensile bendingstresses.

    0 Quantify the fatigue behavior of rock dents. Thistypeof dent has beeninadequately studiedin the past with regard to cyclic pressure loading.

    Development acceptance criteria based on the finding of the current study in

    concurrence with past research andstandards.

    The above objectives willbe accomplishedbyboth experimental and analytical studies.c

    1.6 LIMITATIONSOFTHE STUDY

    It shouldbe noted that the current research effort was performed with the following

    limitations:

    0 Considers the fatigue behavior of dents residing in straight sections of pipe. Theadditional stresses due to circularbends inpipes are not considered.

    The effect of residual stresses due to manufacturingand welding other than thosealreadywithin thepipe test specimens.

    0 Location of dents in relation to welds orseams.

    0 The mechanical damage process, As discussed, significant localizedstresses willoccuratthecontact region.

    0 Corrosioneffects.-8-

  • 7/28/2019 Fatigue Behavior of Dented 1-42

    20/42

    Chapter Two

    REVIEWOF EXISTING DATA AND RESEARCH RESULTS

    2.1 INTRODUCTION

    Prior to the start ofthe experimental program and the finite element analysis, a review of

    previous research program resultswasperformed. Thisprovided a basis for the current research

    and reduced duplicate efforts. Documents reviewed are:

    a Efects ofDents on Failures ofGas Transmission Pipelines, Urednicek, M.,prepared for the CanadianStandardsAssociation Technical Committee, ReportNo. 14032,September, 1986.

    a Cyclic Pressure Fatigue Life ofPipelines with Plain dents, Dents with Gouges,and Dents with Welds,Fowler, J.R., Alexander, C.R., Kovach, P.J., and Connelly,L.M., prepared for the Offshore and Onshore Design Applications SupervisoryCommitteeofthe Pipeline Research Committee at the AmericanGas Association,Final Report for PR-210-927and PR-2 10-9324, June 1994.

    -

    The followingdiscussion summarizes each document as itpertains to the current study.

    2.2 CANADIAN STANDARDSASSOCIATIONTECHNICAL COMMITTEE

    A review of the acceptance criteria of the CanadianGasPipelhe Standard (CSA 2184-M1983) was conducted for the repair of gouges, dents, and gouged dents on new gastransmission pipelines under construction. New experimental and fielddatawere analyzed to

    determine if changes in the acceptance criteria were warranted. The experimental data reviewed

    included:

    BritishGasCorporation - Both plain dents and gouged dents were staticallypressure tested to failure in30 and36in. diameter pipe. One plain dent, at d/D=43%, failed at 60percent ofthe yieldstress. The remaining plain dents failed atstress levels above nominal yield. The faiiure stress ofthe mechanically notcheddents were generally below the yield stress in all cases. {Jones 19823.

    -9-

  • 7/28/2019 Fatigue Behavior of Dented 1-42

    21/42

    0 Battelle Columbus Laboratories - Hydrostatic bust tests were conducted on pipeswith both plain and gouged dents. Pipe diameters rangedfrom 16 in. to 42 in.

    . The location of the dent relative to longitudinal seam welds was examined. Plaindents failed only when located near the weld due to crack formation in the weldduring denting. Dents with machined notches generally failed at stress levels

    below nominal yield. [McClure 1962,Kiefher 1969,Eiber1981, Mayfield 19791.0 ColumbiaGasCompany - Five 20 in. diameter pipes with 3percent dents (d i l l )

    were statically pressure tested to failure. In all cases, failure occuned after thenominal yield stress was reached. [Belonos 19583

    0 Alberta NaturalGas Plain dents detected by inspection of in-service pipelineswere examined. A total of46 of the dents was confirmedas rock dents. Ten ofthe rock dents exceeded the acceptance criteria of2percent and were cutout. Twoof these were plain dents and were hydrostatically tested to above nominal yieldstress levels. Two dents containing gouges at the rock contact region were

    hydrostatically tested to above nominal yield stress. Four dents contained fatiguecracks that initiated on the inside surface at the dent apex. prednicek 19861

    0 NOVA - Both plain and gouged dents were hydrostatically tested to failure in 36in. diameter pipe. No failures occurred at stress levels below the nominal yieldstress.

    The conclusionsdrawn rom this investigation were basedon the review of experimentaldata producedby statically pressurizing pipe section to failure. No cyclic pressure tests were

    conducted. The conclusionsare as follows:

    0 Plain dents, up to 10percent of the pipe diameter (O.D.), could remain in servicewithout affecting the pipeline integrity.

    0 Rock dents behave similar to plain dents. However, crack formation on the insidepipe wall at the dent apex is possible. Inspection to confirm the absence ofcracking is required for acceptance.

    Dents with gouges reduce the pipeline integrity, regardless ofthe dent depth, andshouldbe removed. (No changes to the current criteria were recommended.)

    -10-

  • 7/28/2019 Fatigue Behavior of Dented 1-42

    22/42

    2.3 PRC-AGARESEARCHPROGRAM

    A research program was sponsored by the Pipeline Research Committee (PRC) of the

    American Gas Association (AGA) to investigate the fatigue behavior of dents in gas transmission

    pipelines. The research program involved both experimental and finite element analysis of pipes

    with dents and gouges. The primary objectives of the research program were:

    0

    0

    Determine important variables that govern the fatigue behavior of dents.Classification of dents with gouges.Investigate influence of weldseain on dent fatigue behavior.

    The following briefly summarizes the'methods and conclusions of the research program.

    23.1 Experimental Program

    The experimental program was divided into the testing of two dent types: plain dents and

    dents with gouges.

    Plain Dents -A surveywas sent to 43 oil and gas companies to classify the types of plain

    dents typically encountered with the operation ofa pipeline. Whilethreedent types weredetermined to be common, only one typewas usedin the experimental program.. A finiteelement analysis determined that both the denttypeahd dent length were not important whenconsidering fatigue behavior. The dents were formed by forcing a 24"x24"x2"flat steel plateoriented flat-wise into the si& of the pipe specimens. The dent length, definedas the distance

    between two Undeformed cross sections adjacent to the-dent, was approximately three times thenominal 12 in. diameterofmost of the specimens. The dent depths used, in terms ofdlD, were5,10,and20percent. These depths were measured after the indenter was removed and the dentrebounded. Therefore, the initial dent depth hadto be over-estimatedby a factor oftwo.Denting was performed without the pipe specimens under pressure.

    -1 1-

  • 7/28/2019 Fatigue Behavior of Dented 1-42

    23/42

    The experimental analysis of plain dents was conducted in twophases. The first phase

    involved eight 12-in. nominal diameter pipe specimens, each 20 ft. long. Varying the wallthickness of the specimens gave a range ofDlt from 18 to 50. The dent depths examined were 5,10, and 20 percent ofthe nominal pipe diameters. The second phase of testing involvedtwopipe

    specimens with largerDlt ratios: 64(D=12.75 in., andt =.0.2 in.) and94 (D=24 in., andt =0.26 in.). The dent depths examined were 5 , 10, 15 percent for the 12-in. pipe and5 and 10

    percent for the 24 in. pipe.

    Each specimenwas first statically pressurized to amaximum. of 1200 psi. for thespecimens in the firstphase and400 and900psi. for the 12 in. and24 in. pipe specimens,

    respectively, in the second phase. Dents in the thinner-walled pipe specimenspopped-out orremoved but remained unchanged in the thicker pipe specimens.

    Cyclic pressure testing involved a two-stage pressure range spectrum. Initially, a low

    pressure range was for a specific numberofcycles. If failure did not occur, a higher pressure

    range was used until failure occurred or100,000total pressure cycles hadaccumulated. The

    actualpressure range and number of cycles for each variedfrom specimen to specimen.

    .-

    Dents wth Geum - Since earlier research indicatedthat dents with gouges were morecritical thanplain dents, a second series of testing was performed. Five specimens with nominal

    diameters of 12 in. were used. Wall thicknesses rangedfkom 0.22 to 0.51 for a range ofDlt of25to 58. An easily reproduced machined groove with a radiusof0.002 in. wasusedto simulate thegouge. A 16 in. 1ongitudinaUyoriented gouge waskmachined into the pipe specimens atvarious depths. The specimenwas then dented using the Same process aswith the plain dents.

    As with the plain dents,a two-stage pressure cycle spectrumwas used. The influence oftheseam weld location relative to the dents was

    alsoinvestigated.

    -12-

  • 7/28/2019 Fatigue Behavior of Dented 1-42

    24/42

    2.3.3 Finite Element Analysis

    Both elastic and elasto-plastic finite element analyses were performed with the primary

    objective ofdetermining stress concentrations for dents. These stress concentration factors could

    then be used in analytical fatigue analyses.

    The elastic analysis involved the use ofthree-dimensional models using elastic shell

    elements. Pipe sizes with D/t ratios of18 to 50 were modeled. The models were generated foraspecific dent shape. Indentationwas not modeled elastically. Zero residual stress was given to

    the model with the dented shape. The models were analyzed with internal pressures of500psi

    and1200psi. The stresses forthe models with D/t =50with large dent depths were beyond theyield strength of the pipe. Elastic modeling was found to overestimate the stress distribution.

    Elasto-plastic modeling is needed since the material yields. Residual stresses also should .

    be accounted for. The modeling needs to include the denting process to account forthis. The

    limitations of the elastic model were studied using the elasto-plastic two dimensional models.

    Elasto-plastic three dimensional models were not studied.IElasto-plastic analysis was performed on two dimensional rings comprisedofshell

    elements. The objectives of the elasto-plastic analysis were to determine stress vs. pressure data

    for fatigue calculation and to examine the dent removal caused by pressurization. Support

    conditions were provided to simulate the denting process for the test specimens. A saddle was

    used to support the model. The indenter plate was simulated to form the dents. This model

    actually representsan infinitely long dent in a pipe not subjected to longitudinal stress.

    Parametersstudied includedD/t, dent depth to diameter ratiodD,yieldstress, and pressure.Modeling was performedin threesteps - denting, removal of indenter, and pressure cycling. The

    stress data revealed possible locationsoffatigue crack development. The locations were found

    based on stress levels aswell as stress range. The locations found were similarto the failure

    locations found for the test specimens. The dent removal caused by pressurization was significant

    for the models. Residual dent depth ratios above 1%were not maintained for values ofD/t =50

    -

    -13-

  • 7/28/2019 Fatigue Behavior of Dented 1-42

    25/42

    for any initial dent depth. The final dent depths were smaller than depths from the testspecimens.

    Stress range data was recorded to perform fatigue calculations for the parameters

    modeled. Data was tabulated for the ratio of change in stress to change in pressure. This datacan directly be used to determine fatigue life. Miner's rule fatigue analysiswas used todetermine the fatigue life.

    The results of the tabulated finite element stress range data show that fatigue life is

    reduced by increasing dent depth and increasing pipe slenderness. Comparisons between the

    models and test specimens were made. The predicted cycles until failure for the models were

    mostly lower than the actual test specimens. A few test specimens failed before the predicted

    values were reached.

    23.4 Conclusions

    From both the experimental (plain dents and dents with gouges) and finite analysis, the

    following conclusions were drawn:

    in Dents

    0 Plain dents have longer fatigue lives thandents with gouges.

    0 Fatigue design life is infinite for plain dents in pipeswithD /fless than30.0 Fatigue failures inplain dents can occur whenDit is greaterthan30and the initial

    dent depth,dD is greatert&in 5percent:0 Three-dimensional elastic finite element analysis indicatedthat the dent depth

    ratio, d D is an importantvariable in determumg fatiguebehaviorwhile the dentlength is not.

    0 Gouges that are 10percent or more of the wall thicknesscan result in zero fatiguelife.

  • 7/28/2019 Fatigue Behavior of Dented 1-42

    26/42

    0 Gouges that are 5percent ofthe wall thickness may reduce the fatigue life of adent to 2.5percent of that ofa comparableplain dent.

    0 Gouges in the absence of dents have longer fatigue lives.

    Dents with Welds0 The location of a dent on a longitudinal seam weld did not significantly affect the

    fatigue behavior of the dent.

    Dents located on girth welds have lower fatigue lives than those locatedonlongitudinal welds.

    2.4 REVIEW OF DOCUMENTS PERTAINING TO PIPELINE OPERATIONS

    In additional to the review of results fromboth experimental and analytical researchprograms, various documents pertaining to the operational reliability of pipelines and specific

    fatigue failures in pipelines were reviewed. From this review, it became evident that fatigue life

    estimates for pipeline dents were being based on insufficient information, specifically, the dent

    history. As discussedin Chapter One, the residual stress distribution that results from thedenting process cannot be estimatedh m the final observed dent depth. Without a fundamentalunderstanding of the denting process, basic assumptions that aremade with regard to fatigue life

    estimates are often suspect. Thisoften resulted in conflicting predictions offatigue life.

    2.5 CONCLUSIONS

    From the reviewofvarious documents, the following conclusions were reached withregard to thepresent researcheffort forthe fatigue behaviorofdented petroleum pipelines and

    the need foradditional research

    0 The fatigue behavioroflong, plain dents hasbeen adequately examinedexperimentally by the PRC-AGA sponsoredprogram.

    0 Additional dent types need to be investigated both experimentally and

    -15-

  • 7/28/2019 Fatigue Behavior of Dented 1-42

    27/42

    analytically. This includes short dents and dents restrained against elastic

    rebound.

    0 Denting and dent rebound is an elastic-plastic process. The dent residual stressesas a result ofthis process are greatly influenced by thisprocess.

    0 Dent stiffness, which influences the denting process and reboundbehavior, is athree-dimensional phenomenon. A two-dimensional analysis may not yieldacceptable results.

    -16-

  • 7/28/2019 Fatigue Behavior of Dented 1-42

    28/42

    ChapterThree

    EXPERIMENTAL PROGRAM

    3.1 INTRODUCTION

    The objective of the OPS test program was to duplicate the test parameters of the AGA

    program but with the dents restrained. Therefore, pipe sizes,dent configurations, load pressure

    histories, and mechanical damage similar to those usedin the AGA program (see Sec. 2.3) wereused in the OPS test program. A total of fifteen pipe specimens were tested.

    3.2 PIPE SPECIMENS

    The fifteen water-filled pipe specimens were fatigue testedas summarizedin Table 3-1.

    Pipe diameters ranged fiom 12 in. to 36 in. Both 1/4 in. and3/8in. wall thickness were used

    resulting in a range ofdiameter-to-thickness (Oh)ratios of32 to 96. The length ofthe pipespecimens were 20 t. for the 12,16,and18 in. diameter pipes. To reduce the change in volumedue to the internal pressure changes and, hence, the cyclic time requiredfor100,000 cycles, the

    lengthsofthe larger diameter pipe specimens were reducedto 13 ft. for the 24 in. diameter pipeand10 t. for the 30 in. and36 in. pipe. The reduced pipe length facilitated handling of the largerdiameter pipe, especially when filled with water. A more detailed description ofeach pipe

    specimen and its test parameters is given in the Appendix.

    *

    Elliptical (2:l ratio) end caps were welded on the ends ofeach pipe using the GMAWprocess. Thepipesectionswere ordered fiom the supplier with a beveled cut finish. When

    necessary, the edge ofthe endcapwas beveledso that any difference in wall thickness between

    the pipe and end caps wasproperly transitioned. The welding ofthe end caps wasperformed atthe Testing, Repair, and Machining Facility at the Riverside Campus of Texas A&M University,aswell as a local welding company.

    -17-

  • 7/28/2019 Fatigue Behavior of Dented 1-42

    29/42

    Table3-1: Summary ofpipe specimen properties.

    ~

    Pipe Pipe WallSpecimen Diameter Thickness D/t Length API L2Number D t, in. Grade

    ~ ~

    1 12 0.375 32 20' - 0" X 602 12 0.375 32 20' - 0" X 603 12 0.250 48 20' -0" X 424 16 0.250 64 20' - 0" X 605 16 0.250 64 20' - 0" X 426 18 0.250 72 20' - 0'' X 427 18 0.250 72 20' - 0" X 608 18 0.250 72 20' - 0" X 609 24 0.250 96 13' - 4" X 6010 24 0.250 96 13' - 4" X 6011 24 0.250 96 13' - 4" X 6012 30 0.375 80 10' - 0" Gr. B13 30 0.375 80 lo' - 0" Gr. B14 36 0.375 96 10' - 0" Gr. B15 36 0.375 96 10' - 0" Gr.B

    3.3 DENTTYPES

    Three different dent typeswere investigatedin the experimental test program as

    summarizedin Table 3-2. They include a 6 in. long dentwithsimulated damage (TypeA),

    backhoe dents orientedinboth the longitudinal(TypeBH-L)and transverse(TypeBH-T)

    directions, androckdents(TypeR). These dent types were selected for theprimary objective of

    investigating short dents and restrained dents. The following describes each denttype in detail.

    -18-

  • 7/28/2019 Fatigue Behavior of Dented 1-42

    30/42

    Table3-2: Summary of dent types for each pipe specimen.

    Specimen D/t TotalNo. Dent Type per Specimen Type AofDents Gouge

    A BH-L BH-T RNo.

    8

    9

    10

    11

    12

    13

    14

    32

    32

    48

    64

    64

    72

    72

    72

    96

    96

    96

    80

    80

    96

    6

    10

    9

    8

    9

    8

    10

    8

    4

    5

    4

    3

    3

    3

    6

    10

    8

    2

    8

    10

    4

    2

    3

    3

    3

    0.05"

    N.A.

    5 5 N.A.

    0.024"

    3 4 N.A. .

    scribed

    N.A.

    4 4 N.A.

    scribed3 N.A.

    4 N.A.

    scribed

    N.A.

    scribed

    3.3.1 Dents with Machined Damage (Type A)

    The first type,designatedasTypeA, wascausedby a 6.0 in. long indenter with round

    edges. Figure3-1provides the dimension of theTypeA indenter. Thistypeofdent contained a

    5-inch long machined notch in the dent trough, as shown in Fig. 3-2, to simulate the damage

    typically foundin a gouge. Some of the dents were restrained on various pipes to investigate the

    effect on fatigue behavior.

    -19-

  • 7/28/2019 Fatigue Behavior of Dented 1-42

    31/42

    y __ \_ __-I...Figure3-1: Indenterdhensions forTypeA dent.

    c

    Figure3-2: Gouge location for Dent TypeA.

    Variousmethods were used to simulate the gouge. In the dents of several pipes, the

    gouge was founded bymachining a square-shapednotch using a rotarygrinderas shown in Fig.

    3-3. Thisproduced aunif'om notch to a depth and allowed for reproducibledata not influencedby initial defect parameter. In addition,thisgouge is similarto that usedin other studies and,

    thus,allowed for the comparison ofdata. Two gouge depths were used: 0.05 in. and0.024 in. A

    portion ofa specimen cross section with a machined notch is shown in Fig. 3-4.

    It shouldbe

    notedthat

    this simulation ofa gouge will tendto overestimate the resultingfatigue damage since the compressive residual stresses normally associated with the denting and

    gouging formationare notpresent. Compressive residual stresses reduce the effective stressrange and stress intensity range associated with fatigue crack propagation.

    -20-

  • 7/28/2019 Fatigue Behavior of Dented 1-42

    32/42

    Figure3-3:

    View of machining operation form e c h c a l damage,Dent H, Specimen4.

    Figure 3-4: Cross sectionofDent C, Specimen 1, showingmachined notch (dD 5%).

    In an effort to reduce the severity ofthe machined notch, the notch in dents ofseveralspecimens were manuallyhammerpeenedas shown in Fig.3 4 6 . Preliminary analysis ofthedata indicates that thepeening did not significantlyaffect the fatiguebehaviorofthe TypeAdents. Forlater specimens, a scratch was formed by manually scribing a shallownotch with a

    hardened steel scribe, as shown in Fig. 3-6, Threepasses were made along a straight-edge for

    each scratch in an effort to imposesomeofthe compressive residual stresses into the gouge that

    would be expectedifa sharp foreign object isdrawnacross the surface ofthepipe. This dent

    -21-

  • 7/28/2019 Fatigue Behavior of Dented 1-42

    33/42

    type was testedinboth the restrained and unrestrained configurations to investigate the effect of

    dent reboundon fatiguebehavior.

    Figure 3-5: Peening ofmechanical damagepriorto denting, Specimen4.

    Figure 3-6: Scribing of scratch along straight-edge to simulate damage from moving indenter.

    -22-

  • 7/28/2019 Fatigue Behavior of Dented 1-42

    34/42

    3.3.2 Dents Caused By Backhoe Teeth (Type BH-L & BH-T)

    ~nothertype ofdent investigated experimentally was formedby forcing a typicalbackhoe tooth into the wall ofthepipe specimens. This resultedin a dent relatively short in

    length since the backhoe toe was not forced horizontally along the surface ofthe pipe. Dents

    with the backhoe tooth oriented (With respect to the longitudinal axis) inboth the longitudinal(TypeBH-L)and transverse(TypeBH-T)direction were used.An example ofa backhoe tooth

    used for this dent type is shown inFig. 3-7.

    Figure 3-7: View ofbackhoe tooth restraint.

    3.3.3 Rock Dents (TypeR)

    The thirdtypeofdent investigated was the rock dent (TypeR). This typeofdent was

    formed by forcing the rock into the pipe wall and then restraining the rock in the same positionprior to cyclic loading. Thisallowed for the investigationofdents fonnedby the pipeline beinglaidon to a rock during initialconstruction or by the settlement of the pipeline during service.Figure 3-8provides a view of a Type Rdent showing the rock restrained in a dent.

    -23-

  • 7/28/2019 Fatigue Behavior of Dented 1-42

    35/42

  • 7/28/2019 Fatigue Behavior of Dented 1-42

    36/42

    Figure 3-9: View ofdenting process in 500 kip testing apparatus, Specimen 7.

    Figure 3-10: View ofdenting process with portable loadM e , Specimen 3.

    -25-

  • 7/28/2019 Fatigue Behavior of Dented 1-42

    37/42

    Foreach dent type used on a particular specimen, different initial dent depths were used.

    as summarizedin Table 3-3. The initial dent depths, given in terms of the percentage ofnominaldiameter. ranged from 5 to 18.75percent. The actuator forces required to attain the initial depth

    were measured for each dent andare summarized inAppendixA. It should be noted that withthe removal ofthe indenter, elastic rebound of the dent occurred. Consequently, the final depth

    was always less than the initial depth. The degree ofelastic reboundis a function ofthe

    diameter. thickness, and the dent type.

    The indenter used to force the dents varies with the dent type. The TypeA indenter. as

    shown in Fig. 3-11,was centered over the dent location so that the machined gouge wouldresidein the dent trough. Figure 3-12shows the same dent as formed in Fig. 3-11. Note that majority

    ofthe plastic deformation in the pipe wall fromthe indenter contact is at eitherendofthe dent.

    Consequently, crack initiation wouldbe expected to be more towards the centerofthe dent,

    removed fiom regions of compressive residual stress.

    Figure 3-11: Indenter Type A atmaximumdepth for Dent B, Specimen 1.

    -26-

  • 7/28/2019 Fatigue Behavior of Dented 1-42

    38/42

    Table 3-3: Summary ofdent depths (dD)by specimenanddent type.

    Type A5.0

    7.5

    1012.5

    15

    17.5

    18.75

    Type BH-L

    2 2 2 2 1 12 : I 2 2 I 2 I 1 1

    2 2 1I I

    2 I 2 2 l 12

    5.0I 1 1 1

    7.5 I 2 2 1 ,10 I 2 1 2 1 1 ' 1

    I12.5 I 2 2 I

    7.5

    5 O I 1 ' I l l iI II ' 1 1' 2 1 I7.5

    I

    I 1 i l l ;I / ' l i,

    15

    -27-

  • 7/28/2019 Fatigue Behavior of Dented 1-42

    39/42

    Figure 3-12: View of Dent B, Specimen 1, after removal of indenter and priorto testing.

    Figure 3-13provides an example ofa backhoe tooth used for the Type BH (L &T) dents. -Figure 3-14shows the sametype dent after removal of the backhoe tooth. Note the impressionofthe backhoe tooth left in thepipe wall indicativeofplastic flow and deformation. High

    residual stresses wouldbe expected in this region. Consequently, fatigue crack initiation would

    be suppressed.

    Figure 3-13: Indenter for dent TypeBH-L at maximum depth for DentA, Specimen 7.

    -28-

  • 7/28/2019 Fatigue Behavior of Dented 1-42

    40/42

    Figure 3-14: View of Dent Type BH-L after removal of indenter and prior to testing(Dent I. Specimen 1).

    The Type Rdents were formedby forcing the rocks into the wall of the pipe specimens,

    as shown in Fig. 3-15. The same rockusedas the indenterwas also used during fatigue testing

    so the same contact forces developed at the same locations in the dent. Severalofthe rocks failedby tensile splitting during the denting process.

    -

    Figure 3-15: Rock indentation, Specimen 3.-29-

  • 7/28/2019 Fatigue Behavior of Dented 1-42

    41/42

    3.5 R E S T W N T

    In order to investigate the effects of restraint on the fatigue behavior of dents, the indenterdevices were held against selected dents of each pipe specimen whiIe the pipe was cyclicallytested. h l e n a dent resulting fiom a backhoe tooth was restrained, the same tooth was used forboth denting andrestraining. Likewise, the same rock was used for both denting and testing.This insured that the dent profile remained unchanged. The 6 in. long indenter used with the

    TypeA dent is shown attached to a smallsection of pipe in Fig. 3-16.

    The typical assembly shown in Fig. 3-16was then placed in the dent andheld in place

    using several industrial tie-down straps. These straps are typical of those used to secure cargo on

    flatbed tractor trailers and canbe tightened to variousdegrees through the use ofa.ratchetsystem. Prior to thestartof the fatigue test program, these straps were proof tested to

    approximately 11,000 lbs. Since each strap looped around the pipe in a continuous fashion, a

    resisting force of approximately 20,000 lbs for each strap waspossible. In order to attenuate the

    strap-bearing forces around the pipe, l"x2" wood strappingwas usedbetween the straps andpipe. Figure 3-17shows Specimen4 withDents E throughH restrained using the indenter and

    strap system. Figure 3-18 showsa restrainedTypeA indenter in contactwith the dented pipe.

    Figure 3-16: View of4" dent restraint(TypeA).-30-

  • 7/28/2019 Fatigue Behavior of Dented 1-42

    42/42