-
Finite Element Modeling of aThermoplastic Seal at
HighTemperature and Pressure
Jorgen BergstrJorgen Bergstroomm11, Ph.D., Ph.D.
Brun HilbertBrun Hilbert22, Ph.D., P.E., Ph.D., P.E.
22Exponent Inc.Exponent Inc.
Natick, MANatick, MA
11Veryst Engineering, LLCVeryst Engineering, LLC
Needham, MANeedham, MA
Email: [email protected]
-
Outline of PresentationOutline of Presentation
•• Description of the ProblemDescription of the Problem
•• Mechanical behavior of TeflonMechanical behavior of
Teflon
•• New UMAT for TeflonNew UMAT for Teflon
•• Calibration and validation of theCalibration and validation
of theUMATUMAT
•• FE simulation of Teflon gasketFE simulation of Teflon
gasket
•• ConclusionsConclusions
-
Threaded Connection SimulationThreaded Connection Simulation
Steel
PipeTeflon
Seal
Steel
Coupling
-
Threaded Connection SimulationThreaded Connection Simulation
•• The two steel pipes are threaded togetherThe two steel pipes
are threaded together•• The assembled pipe transports gas at
highThe assembled pipe transports gas at high
temperature and pressuretemperature and pressure
•• The Teflon seal acts as a secondary sealThe Teflon seal acts
as a secondary seal•• How much pressure can the Teflon seal takeHow
much pressure can the Teflon seal take
before leaking?before leaking?
-
Threaded Connection SimulationThreaded Connection Simulation
•• What is the pressure between the Teflon sealWhat is the
pressure between the Teflon sealand the steel pipes at different
temperatureand the steel pipes at different temperatureand
times?and times?
-
Finite Element ModelingFinite Element Modeling
Finite Element
Modeling
Geometry and
BC
Material
RepresentationLoading
Specification
• The results from FEA are only as accurate as the input
values
• The most difficult part is typically the material
representation
-
Geometry and Boundary ConditionsGeometry and Boundary
Conditions
•• AxisymmetricAxisymmetricrepresentationrepresentation
•• The model contains 3The model contains 3parts:parts:
–– Lower steel pipeLower steel pipe
–– Upper steel pipeUpper steel pipe
–– Teflon sealTeflon seal
-
Geometry and Boundary ConditionsGeometry and Boundary
Conditions
•• The analysis isThe analysis isperformed usingperformed
usingABAQUS/ExplicitABAQUS/Explicit
•• The 3 parts areThe 3 parts areinitially overlappinginitially
overlapping
•• No contact activatedNo contact activated
-
Geometry and Boundary ConditionsGeometry and Boundary
Conditions*Expansion, *Expansion, type=orthotype=ortho, zero=,,
zero=,
dependencies=1dependencies=1
** alpha11, alpha22, alpha33, T, field** alpha11, alpha22,
alpha33, T, field
6.000e-4, 0.000e-5, 6.000e-4, , 1 6.000e-4, 0.000e-5, 6.000e-4,
, 1
1.242e-5, 1.242e-5, 1.242e-5, , 2 1.242e-5, 1.242e-5, 1.242e-5,
, 2
*Initial conditions, type=temperature*Initial conditions,
type=temperature
allNallN, ,
*Initial conditions, type=field,*Initial conditions, type=field,
var=1var=1
allN allN, 1, 1
FIRST STEP:FIRST STEP:
*Temperature*Temperature
intPipeintPipe..allNallN,
*Field, op=mod,*Field, op=mod, var=1var=1
allN allN, 1, 1
SECOND STEP:SECOND STEP:
*Temperature*Temperature
intPipeintPipe..allNallN, ,
Apply DIFFERENT TEMPERATURE AND PRESSUREApply DIFFERENT
TEMPERATURE AND PRESSURE
HISTORIESHISTORIES
-
Experimental Data for TeflonExperimental Data for Teflon
T=20°C
Uniaxial Tension
-
Experimental Data for TeflonExperimental Data for Teflon
Stress Relaxation
Triaxial Compression
-
Mechanical Behavior of TeflonMechanical Behavior of Teflon
• Creep
• Stress relaxation
• Temperature
dependence
• Yielding
• Large deformations
The Response is Characterized by:
How can the Teflon material be
modeled using ABAQUS?
-
Constitutive Model DescriptionConstitutive Model Description
*Material, name=Teflon
*User material, constants=17
100, 1, 0, 1.11e-4, , 6.0, 3.5, 600
100, 200, 3.5, 600, 0.0, 1.35, 3.0, 165.0
0.01
*Depvar
18
*Density
2200e-12
-
Constitutive ModelingConstitutive Modeling
8-chain model
8-chain model
Viscoelastic Flow:
• Modeled with a
reptation based
energy activation
representation
Equilibrium
response
Time-dependent
responseViscoplastic Flow
Chain slippage driven by a
stress driven representation
-
Response of the EquilibriumResponse of the
EquilibriumNetworkNetwork
•• LL-1-1((xx) is the inverse Langevin) is the inverse
Langevinfunctionfunction
•• Hyperelastic representationHyperelastic representation••
Micromechanism inspiredMicromechanism inspired•• Accurately
predicts large strainAccurately predicts large strain
multiaxial deformationsmultiaxial deformations
8-chain model
( ) ( )( )
1 *0
*
1*
/
dev 11/
ve lock
A ve ve
A lockve
L
JLJ
! !µ "#
!!
$
$% & % &= + $' ( ' (T B 1
detve veJ ! "= # $F
( ) ( )2/3
*T
ve ve ve veJ
!
=B F F
( )* *tr / 3ve ve! = B
( )0 0expA Abase
! !µ ! µ
!
" #$= % &
' (
-
Constitutive ModelingConstitutive Modeling
•• The details of theThe details of thematerial model
arematerial model areavailable in:available in:
–– ““A Constitutive ModelA Constitutive Modelfor Predicting
thefor Predicting theLarge DeformationLarge
DeformationThermomechanicalThermomechanicalBehavior ofBehavior
ofFluoropolymersFluoropolymers””, J.S., J.S.Bergstrom,
L.B.Bergstrom, L.B.Hilbert, Mechanics ofHilbert, Mechanics
ofMaterials, Materials, vol vol 37, pp.37, pp.899-913,
2005.899-913, 2005.
-
Constitutive ModelingConstitutive Modeling
•• Available for bothAvailable for bothABAQUS standardABAQUS
standardand explicitand explicit
•• Physically motivatedPhysically motivated
•• Incorporates:Incorporates:–– Rate effectsRate effects
–– ViscoelasticityViscoelasticity
–– ViscoplasticityViscoplasticity
–– PermanentPermanentdeformationdeformation
–– Temperature effectsTemperature effects
–– Volumetric creepVolumetric creep
-
Determination of MaterialDetermination of
MaterialParametersParameters
1)1) Calibrate model toCalibrate model toavailable uniaxial
dataavailable uniaxial data(different strain rates,(different
strain rates,temperatures, and straintemperatures, and
strainhistories)histories)
2)2) Simulate multiaxial testsSimulate multiaxial testsusing the
calibratedusing the calibratedmodelmodel
3)3) Evaluate performance ofEvaluate performance ofthe modelthe
model
Calibration
Verification
Evaluation
-
Material Parameters for PTFEMaterial Parameters for PTFE
8.52 MPa
71.2 C
5.0
500 MPa
A
base
lock
µ
!
"
#
=
=
=
=
o
Network A
12.97Bs =
1
9.11
28.9
19.0 MPa
152 GPa
base
vol
C
m
n
!
"
= #
=
=
=
=
0
0.046
1.0
19.0 MPa
a
b
!
=
=
=
Network B
Viscoelastic flow
Plastic flow
-
Glass Fiber Filled PTFEGlass Fiber Filled PTFE
-
PTFEPTFE
-
PTFEPTFE
Triaxial Compression
(T=20°C)
-
Limit of applicationLimit of application
•• DeformationDeformation–– The model works for arbitrary
multiaxialThe model works for arbitrary multiaxial
deformation statesdeformation states
–– The model has been tested forThe model has been tested
fordeformation rates between 10deformation rates between 10-5-5/s
to 1/s/s to 1/s
•• Temperature rangesTemperature ranges–– The model has been
tested forThe model has been tested for
temperatures between 20temperatures between 20°°C and 200C and
200°°CC
•• Software implementationsSoftware implementations––
Implemented and tested for ABAQUSImplemented and tested for
ABAQUS
(both Explicit and Implicit)(both Explicit and Implicit)
-
•• Uniaxial testing onlyUniaxial testing onlyprobes one aspect
of theprobes one aspect of thematerial modelsmaterial models
•• Many models can predictMany models can predictuniaxial
deformation,uniaxial deformation,only a few can predictonly a few
can predictmultiaxial loadingmultiaxial loading
•• In many importantIn many importantapplications the
appliedapplications the appliedload is multiaxialload is
multiaxial
The Need for Multiaxial TestingThe Need for Multiaxial
Testing
-
Verification: Punch testingVerification: Punch testing
Specimen geometry:Specimen geometry:
•• Diameter=6.4 mmDiameter=6.4 mm
•• Thickness=0.5 mmThickness=0.5 mm
-
Experimental DataExperimental Data
-
Model PredictionsModel Predictions
-
Threaded ConnectionThreaded ConnectionSimulationsSimulations
-
Threaded ConnectionThreaded ConnectionSimulationSimulation
-
Threaded Connection SimulationThreaded Connection Simulation
-
Threaded ConnectionThreaded ConnectionSimulationSimulation
-
Threaded Connection SimulationThreaded Connection Simulation
-
ConclusionsConclusions
•• FE analysis generally requires 3FE analysis generally
requires 3parts:parts:
–– Geometry specificationGeometry specification
–– Load and boundary conditionsLoad and boundary conditions
–– Material modelsMaterial models
•• The specification of the materialThe specification of the
materialmodel is often the most difficult partmodel is often the
most difficult part
-
ConclusionsConclusions
•• Accurate FE analysis of polymersAccurate FE analysis of
polymersrequires:requires:
–– Careful experimental testingCareful experimental testing
–– Material model calibrationMaterial model calibration
–– Material model validationMaterial model validation
–– Specialized user material modelsSpecialized user material
models((UMATsUMATs) can provide accurate) can provide
accuratepredictions for many predictions for many ““toughtough””
problems problems
-
Exponent UMAT models forExponent UMAT models for
•• ElastomersElastomers–– Filled or unfilledFilled or
unfilled
•• Semi-crystalline glassy polymersSemi-crystalline glassy
polymers–– PolyethylenePolyethylene
–– FluoropolymersFluoropolymers
•• FoamsFoams–– Silastic foamSilastic foam
-
Special Thanks:Special Thanks:
Shell Exploration and Production CompanyShell Exploration and
Production Company
Sina Ebnesajjad Sina Ebnesajjad at DuPont at DuPont
FluoroproductsFluoroproducts
Pradip Kaladkhar Pradip Kaladkhar at DuPont at DuPont
FluoroproductsFluoroproducts