You are on page 1of 23

Toward High-Fidelity Multi-Scale Modeling

of 3D Crack Evolution
Presented at the retirement symposium for
Professor Tony Ingraffea
September 27, 2014

Ashley D. Spear (Ph.D. 2014)


Assistant Professor
Dept. of Mechanical Engineering
University of Utah

Brett R. Davis (Ph.D. 2014)


Associate, Mechanical Engineering Practice
Exponent

Albert R. Cerrone (Ph.D. 2014)


Mechanical Engineer, Lifing Lab
GE Global Research Center
1
Presentation Outline
• PART I: Predicting crack propagation at the component scale
– Toolset to simulate elastic-plastic crack growth in 3D
– Toolset to simulate crack-shape evolution using energy-release-rate
formulation
• PART II: Understanding fatigue-crack formation and early
propagation at the microstructural scale
– Study of crack nucleation in Ni-base superalloy
– Study of crack propagation in Al-Mg-Si alloy
• Lessons learned from our time in the CFG

2
Overview of Work
• Generalized hypotheses:
– Compared to existing approaches for predicting crack evolution,
more accurate predictions can be made:
1) by accounting for three dimensionality of the cracked body and
2) by maintaining high level of fidelity appropriate for given length scale
Micro-scale considerations: Component-scale considerations:
*sensitivity to microstructural *sensitivity to constraint conditions
heterogeneities *tearing/fracture at limit state

LEFM is invalid
at both extremes!

schematic adopted from


Suresh, Fatigue of Materials 3
Toolset I: Generalized 3D Fracture Simulation
• Extended FRANC3D capabilities for EPFM simulations
• Applications:
– Crack-growth predictions when material-state history is important and
LEFM is not valid
• Toolset description:
– Geometrically explicit crack representation
– Adaptive remeshing
– Allows prediction of crack growth direction
– Recent enhancements for EPFM simulations
ABAQUS FRANC3D/NG
*Model definitions *Insert crack(s) or predict
-geometry new crack growth
-materials *Adaptively remesh sub-region
-boundary *Connect sub-region with global mesh
conditions
ABAQUS/ Python script

CTD’s *Map material state to


ABAQUS/Python
ABAQUS script maintain history after
compact tension (C(T)) model
*Analyze model remeshing
*Evaluate CTD
criterion during
nonlinear FE analysis
Collaboration with: Veilleux, Hochhalter 4
Toolset I: Validation Example
• Aluminum-alloy 2024-T3 fracture specimen in Arcan test fixture*
• 30˚ loading angle induces mixed-mode I/II crack growth

2024-T3 fracture specimen Arcan test fixture 3D FE model


P

P
units in mm

*Amstutz, et al., 1995 and 1997 5


Toolset I: Validation Example

plastic strain
*120,000 dof’s
(average per
accumulation
crack step) along crack wake
*25 crack steps

Collaboration with: Veilleux, Hochhalter 6


Toolset I: Validation Example
EPFM framework better predicts crack-extension response compared to LEFM framework

20

15
Applied load (kN)

10

5 Experiment
EPFM Simulation
LEFM Simulation
0
0 5 10 15 20 25
da (mm)
Collaboration with: Veilleux, Hochhalter 7
Presentation Outline
• PART I: Predicting crack propagation at the component scale
– Toolset to simulate elastic-plastic crack growth in 3D
– Toolset to simulate crack-shape evolution using energy-release-rate
formulation
• PART II: Understanding fatigue-crack formation and early
propagation at the microstructural scale
– Study of crack nucleation in Ni-base superalloy
– Study of crack propagation in Al-Mg-Si alloy
• Lessons learned from our time in the CFG

8
Toolset II: Energy-Based Crack Shape Evolution
• Develop simulation capability that permits arbitrary growth with
unknown crack-shape evolution Initial Configuration
– Geometrically explicit cracks
– Re-meshing techniques

Non-self-similar crack growth in


mixed-mode bending specimen Final Configuration

Initial Crack Front

Collaboration with: Wawrzynek, Hwang, Carter 9


Toolset II: Energy-Based Crack Shape Evolution

Energy Release Rate


Expansion:

Target Current Load Extension


Energy Release Rate Energy Release Rate Increment Increment

Local Extension Criterion: SOLVE FOR


LOCAL
EXTENSIONS

Local Extension
Balance Condition:

Collaboration with: Wawrzynek, Hwang, Carter 10


Toolset II: Numerical Example

P
4 cm
4 cm

5 cm

Collaboration with: Wawrzynek, Hwang, Carter 11


Presentation Outline
• PART I: Predicting crack propagation at the component scale
– Toolset to simulate elastic-plastic crack growth in 3D
– Toolset to simulate crack-shape evolution using energy-release-rate
formulation
• PART II: Understanding fatigue-crack formation and early
propagation at the microstructural scale
– Study of crack nucleation in Ni-base superalloy
Micro-scale considerations:
– toStudy
*sensitivity of crack propagation
microstructural 10-2 in Al-Mg-Si alloy
heterogeneities
• Lessons learned from our time in the CFG
regime A regime B
da/dN (mm/cycle)

1 mm/min

crack growth rate


10-4

at 50 Hz
1 mm/hr
regime C
10-6
1 mm/day
1 mm/week
10-8
log ΔK 12
Study I: MSFC Nucleation in Superalloy
• How can we use “big data” to understand highly nonlinear
microstructural phenomena?
• Case Study: Microcrack nucleation in Ni-based superalloy
1. Develop constitutive relations and geometric representations of
superalloy
o calibrate crystal plasticity model
o generate microstructural model for 3D crystal-plastic finite-element
analysis
~mm

~cm

2. Capture relevant physics related0.0045 slip localizing at


grain boundaries
to microcrack nucleation event 0.003 0

3. Correlate grain boundary character


0.002 ~μm
with slip localization
0.001
Collaboration with: Rollett, Stein, Tucker, Pokharel, Hefferan, Lind, Suter 13
Study I: MSFC Nucleation in Superalloy
• Establish correlations between microstructural attributes and
fatigue indicator parameters (FIPs)
− Analyze every grain boundary in Ni-based superalloy
− Quantify correlation between postulated FIPs and grain boundary character
− Determine microstructural characteristics most relevant to nucleation event

Collaboration with:
Rollett, Stein, Tucker,
Pokharel, Hefferan,
Lind, Suter 14
Presentation Outline
• PART I: Predicting crack propagation at the component scale
– Toolset to simulate elastic-plastic crack growth in 3D
– Toolset to simulate crack-shape evolution using energy-release-rate
formulation
• PART II: Understanding fatigue-crack formation and early
propagation at the microstructural scale
– Study of crack nucleation in Ni-base superalloy
– Study of crack propagation in Al-Mg-Si alloy
• Lessons learned from our time in the CFG

15
Study II: MSFC Propagation in Al Alloy
• Application: aluminum alloy used by
NASA in ultrathin pressure-vessel
components
2.5mm  <0.75mm by chemical milling
• Why do ultrathin liners deserve
attention?
– Consequences of crack nucleation and
growth could be catastrophic
– Effect of microstructure potentially
more significant for ultrathin liners 16 cm
250 µm

– Real application for multiscale


materials characterization and 3D
modeling!

41 cm
250 µm
• Little is known about how cracks
propagate in 3D at microstructural
length scale for polycrystalline materials
250 µm
Collaboration with: Suter, Li, Lind, NASA LaRC 16
Study II: MSFC Propagation in Al Alloy
• Broken specimens measured
using synchrotron radiation at
Argonne National Laboratory
– X-ray computed tomography (CT)
• Highly resolved fracture surface z
– High-energy X-ray diffraction y
microscopy (HEDM) *
• Grain geometries and orientations
16 cm

250 µm

Collaboration with: Suter, Li, Lind, NASA LaRC * Lienert, et al., JOM (2011), Suter, et al., RSI (2006); Li, et al., Appl J. Cryst. (2013) 17
Study II: MSFC Propagation in Al Alloy
θz
90
80
60
40
20
x
0
z y
angle (degrees) between local normal and loading direction
y

B 111

001 101

y
crack-plane normal in crystallographic frame 100 µm
Collaboration with: Suter, Li, Lind, NASA LaRC 18
Study II: MSFC Propagation in Al Alloy
observed variability in 3D crack-growth rate

A’

250 µm
x

z
y
y A
3D da/dN
(µm/cycle)
0.010

0.0075

52 µm
0.0050

A’ 277 µm A 0.0025

0.00

Collaboration with: Suter, Li, Lind, NASA LaRC 19


Study II: MSFC Propagation in Al Alloy

Collaboration with: Hochhalter, Cerrone 20


Study II: MSFC Propagation in Al Alloy
σz (MPa)

z x

y y

D3

Lots of data here! Remaining need for quantitative post-processing.


Collaboration with: Hochhalter, Cerrone 21
Presentation Outline
• PART I: Predicting crack propagation at the component scale
– Toolset to simulate elastic-plastic crack growth in 3D
– Toolset to simulate crack-shape evolution using energy-release-rate
formulation
• PART II: Understanding fatigue-crack formation and early
propagation at the microstructural scale
– Study of crack nucleation in Ni-base superalloy
– Study of crack propagation in Al-Mg-Si alloy
• Lessons learned from our time in the CFG

22
Lessons Learned from Tony and the CFG
• Collaborate (it’s required to do the best job possible!)
• Be rigorous (computing time is not an excuse for not doing good work!)
• Don’t forget the two most important Vs in life… (verification and validation)
• “A good leader brings good people together and makes them better.” -Brett
• “The boss was always a big proponent of continuing education, seeking
answers proactively, and never letting ignorance get in the way of scientific
progress.” -Al
• “Oh yeah, and when in doubt, ask Bruce.” -Everyone

23

You might also like