Microstructure-Sensitive Crystal Viscoplasticity for Ni ... Library/Events/2016/utsr/Tuesday... ·...

Microstructure-Sensitive Crystal Viscoplasticity for Ni-base Superalloys

(with application to long-term creep-fatigue interactions)

Award DE-FE0011722August 2013 August 2017 (including NCE)

Program Manager: Dr. Patcharin Burke

PI: Richard W. NeuThe George W. Woodruff School of Mechanical Engineering

School of Materials Science & EngineeringGeorgia Institute of Technology

Atlanta, GA [email protected]

University Turbine Systems Research WorkshopVirginia Tech

November 1-3, 2016

The George W. Woodruff School of Mechanical Engineering School of Materials Science and Engineering

Hot Section Gas Turbine Materials

Land-based gas turbines drive to increase service

temperature to improve efficiency; increase life; with minimal increase in cost

replace large directionally-solidified Ni-base superalloyswith single crystal superalloys

Power Output: 375 MW


Superalloys airfoils are subject to a continuation of process during service due to sustained hostile environment and loading.Understanding the influence of the microstructure and its evolution on TMF, creep and fatigue and their interactions is key todelivering better turbine designs

t

R = 0 t

R = -

[Epishin et al., 2010]

PressureTemperature

Time

PressureTemperature

Enginestart-up

Engineshut-down

Hot Section Gas Turbine Materials


PSPP Map for Ni-base Superalloy Airfoils

PROCESS STRUCTURE PROPERTIES

Creep Rate

Elastic Modulus

Yield Strength

Resistance to Environment Degradation

FatigueStrength

FractureToughness

Dendritic structure Primary arm spacing Secondary arm spacing Misorientation

Vacancy concentration

and phases Shape Size Volume fraction Mismatch and

Secondary size

Other Features: Crystal structure APB Energy Texture Eutectic pools Casting pores Freckles High angle grain boundaries Carbides

Dislocations Density in and Configurations Distributions

Load profile

Temperature profile

Environment profile

Composition:Ni, Co, W, Ta, Al, Cr, Re, Ti, Mo , Hf, C, B,

Age Treatment

Casting andSolidification

Solution Treatment

Continuation of Process During Service

PERFORMANCE

MaximumCreep-Rupture Life

MaximumThermomechanical

Fatigue Life

CTE

MinimalCost


Influence of Service Conditions on Microstructure, Properties, & Performance


Creep Rate

Elastic Modulus

Yield Strength


FatigueStrength

FractureToughness




Secondary size



Load profile

Temperature profile

Environment profile


Age Treatment


Solution Treatment


PERFORMANCE



Fatigue Life

CTE

MinimalCost


Role of Chemical Composition


Creep Rate

Elastic Modulus

Yield Strength


FatigueStrength

FractureToughness




Secondary size



Load profile

Temperature profile

Environment profile


Age Treatment


Solution Treatment


PERFORMANCE



Fatigue Life

CTE

MinimalCost


Single Crystal Alloy being Investigated for IGT Applications

CMSX-8: 1.5% Re "alternative 2nd gen alloy" replacing 3.0% Re containing alloys (e.g., CMSX-4, PWA1484)

[Wahl and Harris, 2012]

Strength

Creep

Alloy

Cr

Co

Mo

W

Al

Ti

Ta

Re

Hf

C

B

Zr

Ni

Mar-M247LC-DS

8.4

10.0

0.7

10.0

5.5

1.0

3.0

-

1.5

0.07

0.015

0.05

Bal

CM247LC-DS

8.1

9.2

0.5

9.5

5.6

0.7

3.2

-

1.4

0.07

0.015

0.01

Bal

CMSX-4

6.5

9.0

0.6

6.0

5.6

1.0

6.5

3.0

0.1

-

-

-

Bal

SC16

16

0.17

3.0

0.16

3.5

3.5

3.5

-

-

-

-

-

Bal

PWA1484

5.0

10.0

2.0

6.0

5.6

-

9.0

3.0

0.1

-

-

-

Bal

CMSX-8

5.4

10.0

0.6

8.0

5.7

0.7

8.0

1.5

0.2

-

-

-

Bal


Influence of Temperature, Loading Direction and Crystal Orientation on Modulus and Strength

-1.5 -1 -0.5 0 0.5 1 1.5

strain [%]

-1000

-500

0

500

1000

stre

ss [M

Pa]

Virgin CMSX-8 monotonic response in the dir

20 C

750 C

950 C

1025 C

1100 C

1100 C

1025 C

950 C

750 C

20 C

Increasing

Temperature

Increasing

Temperature

-1.5 -1 -0.5 0 0.5 1 1.5

strain [%]

-1000

-500

0

500

1000

stre

ss [M

Pa]

Virgin CMSX-8 monotonic response in the dir

Increasing

Temperature

750 C

750 C

1100 C

20 C

20 C

1100 C

Increasing

Temperature


Thermomechanical Fatigue (TMF)

Linear In-Phase (IP) Linear Out-of-Phase (OP)

Time, t

Stra

in,

(%)

Thermal StrainMechanical StrainTotal Strain

Cold

Tension

Compression

Hot

Time, t

Stra

in,

(%)

Thermal StrainMechanical StrainTotal Strain

Cold

Tension

Compression

Hot

Hardening Process

Plastic Deformation

Plastic Deformation

Creep

Oxidation

Cyclic Ageing and Coarsening effects

Hardening Process

Plastic Deformation

Cyclic Ageing and Coarsening effects

Plastic Deformation

Creep

Oxidation

=0 =180


Effect of Tmin on OP TMF of CMSX-4

[Arrell et al., 2004]

3x

2x drop

[Kirka, 2014]

CMSX-4 [001]

CM247LC DS in Longitudinal Dir.


Life Modeling Approach

Damage Mechanism Modules

Fatigue

Creep-Fatigue

Environment-Fatigue

CreepRatchetting

Dominant Damage

Deformation Response

Load

ing

Par

amet

er

Life

Accurrate representations of the deformation responsehighly critical for predicting crack formation


Primary Objectives of UTSR Project

Creep-fatigue interaction experiments on CMSX-8

Influence of aging on microstructure and creep-fatigue interactions

Microstructure-sensitive, temperature-dependent crystal viscoplasticity to capture the creep and cyclic deformation response


Major Activities and Accomplishments

Characterize creep-fatigue interactions on CMSX-8 Creep-fatigue Thermomechanical fatigue Creep (either tension or compression) followed by fatigue Fatigue followed by creep

Characterize the influence of aging on microstructure and creep-fatigue interactions

Experimentally establish the creep-fatigue interactions in a single-crystal Ni-base superalloy that is being targeted for use in industrial gas turbines (CMSX-8)

TMF life: In-Phase (R=0) vs Out-of-Phase (R=-inf)


Conventional Creep-Fatigue (baseline)

0.30

0.50

0.70

0.90

1.10

1.30

1.50

100 1000 10000

Stra

in ra

nge

[%]

Cycles to Failure

Effect of hold on LCF life: R = -

(T = 950 [C], R = -1)(T = 1100 [C], R=-1)(T = 950 [C], R = -, hold = 3 min.)(T = 1025 [C], R = -, hold = 3 min.)(T = 1100 [C], R = -, hold = 3 min.)

1100 [C]

950 [C]

950 [C]

1000 [C]1025 [C]

0.30

0.50

0.70

0.90

1.10

1.30

1.50

100 1000 10000

Stra

in ra

nge

[%]

Cycles to Failure

Effect of hold on LCF life: R = 0

(T = 950 [C], R = -1)(T = 1100 [C], R = -1)(T = 950 [C], R = 0, hold = 3 min.)(T = 1025 [C], R = 0, hold = 3 min.)(T = 1100 [C], R = 0, hold = 3 min.)

950 [C]

950 [C]

1000 [C]

1025 [C]

1100 [C]

t

R = 0

t

R = -


Conventional Creep-Fatigue (baseline)

t

R = 0

t

R = -

0.3

0.5

0.7

0.9

1.1

1.3

1.5

100 1000 10000

Stra

in ra

nge

[%]

Cycles to Failure

Life for low cycle creep-fatigue

(T = 950 [C], R = 0, hold = 3 min.)(T = 950 [C], R = -, hold = 3 min.)(T = 1025 [C], R = 0, hold = 3 min.)(T = 1025 [C], R = -, hold = 3 min.)(T = 1100 [C], R = 0, hold = 3 min.)(T = 1100 [C], R = -, hold = 3 min.)


Cycle Evolution

150

200

250

300

350

400

450

500

0 0.5 1 1.5 2 2.5 3

Stre

ss [M

Pa]

Hold [min.]

Cycle 1 Cycle 2

Cycle 3 Cycle 4

Stress relaxation

1st 10 cycles

Stress evolution

T = 1100 C, R = 0, = 1.0 %


Crack Characteristics

R = 0, T = 1100C, = 0.8%Nf = 1420

R = -, T = 1100C, = 0.8%Nf = 980

t

R = 0

t

R = -


Half-life at 1025 C

= 1.0 [%]

= 0.8 [%]


Half-life at 1100 C

= 1.0 [%]

= 0.8 [%]

Plastic strain range alone does not

explain life data at high temperatures.A Coffin-Manson

relations would not be sufficient


Major Activities and Accomplishments

Stress-free and stress-assisted (rafting) aging experiments under tensile and compressive stresses

Establishing process-structure linkages using physical models, 2-point statistics and PCA

Microstructure-sensitive Crystal Viscoplasticity (CVP) Model to Determine Service Process-Structure-Property Linkages

t0 tf

virgin evolved

Experiments & Models to Predict Current State of Microstructure (Service Process-Structure Linkages)

Sensitivity of Local Composition on Diffusivity for Input in Aging and Viscoplasticity Models

Thermo-CalcDICTRA

Databases: TCNi5 / MOBNi2

Composition segregation in and phase

Determination of composition sensitive effective diffusivity to characterize aging activation energy and diffusivity parameter in viscoplasiticity models


Microstructure Evolution in Blades

2 cm

Dis

tanc

e fro

m R

oot


Rafting and Coarsening of

[Epishin et al., 2008]

N-raft

P-raft

CMSX-4


High-Throughput Stress-assisted Aging


Aged Microstructure under Compressive Stress

Compression Creep Frame

5um

P-Raft

Ceramic Compression Creep Extensometer

Bottom Holder

Top Holder


Physical Models to Describe Aging Behavior

Physical aging models

Rafting Coarsening

, = . .

900 C

950 C1100 C

Experimental data pointsLSW Model

3- 0 3 =

= 0

LSW model:

8 = 269.4 /

= ( 0)

A (m/h) Q (kJ/mol)

UT(J/mol.MPa.Kn) n

CMSX-8 2.0105 206 0.033 1.525CMSX-4 9.31104 222 0.19 1.294

Saturated level of rafting


2-point Correlation Method

2-point correlation A statistical representation of the microstructure that provides a magnitude measure in addition to the associated spatial correlation.

2-point correlations (,) capture the probability density associated with finding an ordered pair of specific local state at the head and tail of a randomly placed vector into the microstructure.

[Niezgoda, Fullwood and Kalidindi, 2008]

Cross-correlationDefined by (, ). The function gives the probability that a random vector of length x start in one phase and end in the other.

(a) An example of two phase microstructure (b) Cross-correlation for the black and white phases at the head and tail of a vector [Fullwood, Niezgoda, Adams, Kalidindi, 2010].

(a) (b)


2-point Correlation Application to CMSX-8

precipitate and channel sizes are determined from the first peak and valley of the probabilistic curve of the corresponding micrograph.


Quantifying the Current Aged State

Stress decreasing

Reduced order representation of aged microstructures using PCA of 2-point spatial correlations


Influence of Variation in Composition

Thermo-CalcDICTRA

Databases: TCNi5 / MOBNi2

Composition sensitive diffusivity parameter

Aging behavior Temperature-dependent constitutive models

=642

9

= 0, (

)

Coarsening effective diffusivity

LSW model

Diffusivity parameter

Composition segregationThermo-Calc


Influence of Composition on Diffusivity

Interdiffusion coefficients of Ni-x binary systemsDICTRA

= 0, (

) = =1

Composition variation (Re) effect on effective diffusivityof the complex Superalloy system

DICTRA

Molar fraction of element x

Effective activation energy/Coarsening activation energy = 275.9 KJ/mol


[Ma, Dye, and Reed, 2008]

Distinct deformation in the and phasesIn : 12 octahedral slip systems

moving as dislocation ribbons

deformationIn : 12 octahedral slip

systems active

deformation

Deformation predictions sensitive to the and phase attributes

Alloy structure

[Shenoy, Tjipowidjojo, and McDowell, 2008]

0 100 200 300 400 500

time [hrs]

0

10

20

30

40

cree

p st

rain

[%]

Tertiary creep: 950 C, Stress = 400 MPa

0.09 m

0.10 m

0.11 m

Channel thickness:

0 500 1000 1500

time [hrs]

0

1

2

3

4

5

6

cree

p st

rain

[%]

Primary creep: 750 C, Stress = 680 MPa

0.13 m

0.14 m

0.15 m

Channel thickness:

Microstructure-sensitive Crystal Viscoplasticityfor Single-Crystal Ni-base Superalloys

0 0.2 0.4 0.6 0.8 1

strain [%]

-800

-600

-400

-200

0

200

400

600

800

stres

s [MP

a]

Creep-fatigue test at T =1025 C 801-15B

First 10 cycles CMSX-8[B/C]UMAT


Crystal Viscoplasticity Kinematic Relations

Kinematic relations includingtemperature dependence

Macroscopic plastic velocity gradient

Deformation gradient

Velocity gradient

In : 12 octahedral slip systems active

In : 12 octahedral slip systems moving as dislocation ribbons


Creep Deformation Mechanisms of Superalloys

600

700

800

900

1000

1100

1200

1300

0 200 400 600 800 1000

Tem

pera

ture

[C]

Stress [MPa]

CMSX-4 PrimaryCMSX-8 PrimaryCMSX-4 TertiaryCMSX-8 TertiaryCMSX-4 RaftingCMSX-8 Rafting?

Tertiary Creep

Rafting Primary Creep

Tertiary dislocation activity restricted toa/2 form operating on {111} slip planes in the channels

Primary particles are sheared by dislocation ribbons of overall Burgers vector a dissociated into superlattice partial dislocations separated by a stacking fault; shear stress must above threshold stress (about 550 MPa)

Rafting transport of matter constituting the phase out of the vertical channels and into the horizontal ones (tensile creep case)

[Reed, 2006; Ma, Dye, and Reed, 2008, CMSX-8 Data]

Influence of stress and temperature on modes of creep deformation




Inelastic Shear Strain Rate

Crystal Viscoplasticity (CVP) Rate Eqn

Evolution Equations

Inelastic Velocity Gradient


Orientation Sensitivity


Primary creep performancebased on experiments

[MacKay and Meier, 1982]

CMSX-4

Creep strain in tertiary regime900C/400MPa (111 hr)

Creep strain in primary regime750C/600MPa (111 hr)

CMSX-4 Model Predictions


Creep in different orientations

0 200 400 600 800 1000 1200 1400 1600

time [hrs]

0

5

10

15

20

25

30

35

40

cree

p st

rain

[%]

Temp = 950[C], Stress = 250[MPa]

Data

CVP

CVP

CVP

CVP

CVP

CVP


Primary and Tertiary Creep

0

5

10

15

20

25

30

0 500 1000 1500

Cree

p St

rain

[%]

Time [hrs]

871 C, 448 MPaCMSX-8 Exp. 1CMSX-8 Exp. 2CMSX-4 Ma et al. ModelCMSX-8 Model

0

5

10

15

20

25

30

0 200 400 600

Cree

p St

rain

[%]

Time [hrs]

871 C, 551 MPaCMSX-8 Exp. 3CMSX-8 Exp. 4CMSX-4 Ma et al. ModelCMSX-8 Model

0.0

0.5

1.0

1.5

2.0

0 50 100Cr

eep

Stra

in [%

]

Time [hrs]

Zoom in: 871 C, 551 MPaCMSX-8 Exp. 3CMSX-8 Exp. 4CMSX-4 Ma et al. ModelCMSX-8 Model

Primary, secondary and tertiary creep can be

captured with the model


TMF validation

First 10 cycles: IP-TMF [100-1100-100] C, R = 0, = 2.83[/]

Stabilized hysteresis: OP-TMF [100-850-100] C, R = -, = 2.83[/]

Very good agreement predicting TMF


TMF validation

Stabilized hysteresis: IP-TMF [100-1100-100] C, R = 0, = 2.83[/]

Stabilized hysteresis: OP-TMF [100-850-100] C, R = -, = 2.83[/]

Very good agreement predicting TMF


If we considering activation energy for plastic flow Q0 a function of %Re, the diffusivity parameter could take the form of:

Since Re segregates almost exclusively in the channels, the Activation energy in the phase can be modified to account for Re content as follows:

( ) exp for2

o mQ TT TRT

=

Incorporating the Influence of Re Content

[Miller, 1976; Shenoy et al., 2005]


Remaining Plans for Project

Exercise new CVP codes (1) displacement-controlled algorithm coded in FORTRAN as UMAT for

ABAQUS (suitable for 3D analysis of detailed features, but computationally expensive)

(2) stress-controlled algorithm coded in MATLAB (limited to 1D loadings, but computationally cheap; utility in alloy design; early design iterations)

Characterize aged microstructure and its analysis Establish protocols for reduced-order modeling of the current state of the

microstructure using spatial correlations and principal component analysis Develop PSP linkage models using PC values for describing current state

of the microstructure Creep-fatigue interaction life analysis relationships using all new data Disseminate research

Archival journal articles Intern with Siemens Energy Workshops and Conferences (UTSR, ASME, )

Ernesto Estrada Rodas complete Ph.D. studies in 2017

Microstructure-Sensitive Crystal Viscoplasticity for Ni-base Superalloys(with application to long-term creep-fatigue interactions) Award DE-FE0011722August 2013 August 2017 (including NCE)Program Manager: Dr. Patcharin BurkeHot Section Gas Turbine Materials Slide Number 3Slide Number 4Slide Number 5Slide Number 6Single Crystal Alloy being Investigated for IGT ApplicationsSlide Number 8Thermomechanical Fatigue (TMF)Effect of Tmin on OP TMF of CMSX-4Life Modeling ApproachPrimary Objectives of UTSR ProjectMajor Activities and AccomplishmentsConventional Creep-Fatigue (baseline)Conventional Creep-Fatigue (baseline)Cycle EvolutionCrack CharacteristicsHalf-life at 1025 CHalf-life at 1100 CPrimary Objectives of UTSR ProjectSlide Number 21Microstructure Evolution in BladesRafting and Coarsening of High-Throughput Stress-assisted AgingAged Microstructure under Compressive StressPhysical Models to Describe Aging Behavior2-point Correlation Method2-point Correlation Application to CMSX-8Quantifying the Current Aged StateInfluence of Variation in CompositionInfluence of Composition on DiffusivityPrimary Objectives of UTSR ProjectSlide Number 33Crystal Viscoplasticity Kinematic RelationsCreep Deformation Mechanisms of SuperalloysCrystal Viscoplasticity (CVP) Rate EqnOrientation SensitivityCreep in different orientationsPrimary and Tertiary CreepTMF validationTMF validationIncorporating the Influence of Re ContentRemaining Plans for Project

Microstructure-Sensitive Crystal Viscoplasticity for Ni ... Library/Events/2016/utsr/Tuesday... ·...

Documents

Transcript of Microstructure-Sensitive Crystal Viscoplasticity for Ni ... Library/Events/2016/utsr/Tuesday... ·...