NASA Langley Damage Tolerance Experiences
by
Ivatury S. RajuNASA Engineering and Safety Center
NASA Langley Research CenterHampton, Virginia
presented at the
FAA Symposium on Composite Damage Tolerance
And MaintenanceChicago
July 19-21, 2006
NASA Langley’s Composites Programs
SOA Analysis
Emerging Continuum Methods
Future Directions
Outline
Allowable Damage Limit
(ADL)
Increasing Damage Severity
Ultimate
~ Maximum load per fleet lifetime
Design LoadLevel
Continued safe flight
Limit
Critical Damage Threshold (CDT)
1.5 Factor of Safety
Structural durability affects the frequency and cost of inspection, replacement,
repair, or other maintenance
Structural damage tolerance ensures damage will be found by maintenance
practices before becoming a safety threat
Discrete source events (e.g., engine burst, birdstrike) can cause severe damage but it is known to pilot
ResidualStrength(notional)
The structure must always be able to sustain design ultimate loads in the presence of nondectabledamage.
(FAR 25.571 & Mil-17)
Durability and Damage Tolerance Requirements
Textile Materials
Details of Stitched Plates
Load Direction
StitchingSpacedAt 3.2mm
48 ply stitched laminate[+45/0/-45/90]6s
Compression After Impact Strength
0
100
200
300
400
500
600
700
0 20 40 60 80 100 120
Unstitched AS4/3501-6
Stitched AS4/3501-6
Toughened MatrixCompositesComp.
Stress,MPa
Impact Energy, J
Stitching Improves Damage Tolerance
No damage or permanent deformation at DLLTest Article with repair of simulated damage failed at 97% of DUL
AS4/3501-6 and IM7/3501-6 in textile preform
NASA ACT Program – Full Scale Wing Box Test (2000)
Dislocations
Twinning
Stacking Fault[1 1 0]
[0 0 1]
x
y
[1 1 0]
Void Coalescence
10 nm
Damage ScienceDamage Science
20 20 μμmm
Emerging Continuum MethodsEmerging Continuum Methods
Com
plex
ity, C
ompu
tatio
nal E
xpen
se
Time
1/41/23/4
0
1
G /(G + G )ΙΙ ΙΙ ΙΙΙ
GT
GIc + GIIc − GIc( ) GII +GIII
GT
⎛
⎝ ⎜
⎞
⎠ ⎟
η
+ GIIIc − GIIc( ) GIII
GII +GIII
GII +GIII
GT
⎛
⎝ ⎜
⎞
⎠ ⎟
η ≥ 1
t≈0
LaRC Decohesion Element(Technology adopted by ABAQUS)
Mixed-Mode Fracture
Bilinear Traction-Displacement Law
Perfectly plastic (pp)
Linear softening (lin)
Progressive softening (pro)
Regressive softening (reg)
G
σ
pro lin reg
c
σ
Needleman (Ne)
Nepp0
c
CGdF
=∫ δδσδ
0 )(
QuickTime™ and a decompressor
are needed to see this picture.
Wide range of element sizes
Narrow range
Classical
Enhanced
Enhanced formulation allows the useof elements up to three times larger thanwith the classical damage model.
MMB Specimen
Delamination growth
FY05 Advances
Current StateCurrent State--ofof--thethe--ArtArt
0
0.5
1
1.5
2
2.5
3
0 50 100 150 200
G I
in-lb/in 2
Pressure, P (psi)
b=0.750in.b=0.625in.b=0.500in.
Evolution of Damage Tolerance at NASA Langley (1999-)
The liquid hydrogen composite tank failed during the protoflight ground test.
Hypersonic Experimental Vehicle (X-33 Program)
The lobes are sandwich construction:The inner face sheet is [45/903/-45/03/-45/903/45]TThe outer face sheet is [65/0/-65/90/-65/0/65]T
The face sheets are IM7/977-2 laminates.The core is a honeycomb Korex 3/16 - 3.0 (1.5 in. thick).The adhesive is AF-191.
X-33 Composite Liquid Hydrogen Tank Failure
X-33 Composite Liquid Hydrogen Tank Failure
Teflon Tape in CoreInner Skin Microcracking
Weak Core to Face Sheet Bond Strength/Toughness
Causes of the X-33 Composite Tank Failure
0
0.5
1
1.5
2
2.5
3
0 50 100 150 200
GI
in-lb/in2
Pressure, P (psi)
b=0.750in.b=0.625in.b=0.500in.
mean-Gcr
3σ-Gcr
Lowest measured-Gcr
b
2.625”
R=0.5”
Hig
hest
mea
sure
d -P
3σ -
P
Strain Energy Release Rates for an F.O.D. Debond
Local FE Model
Global FE Model
Lug and Pin
Rear Fuselage and Tail Configuration
Attributes25,931 nodes21,519 elementsContact modeled200 plies in lugGlobal-local coupled analysisDamage monitored as load incremented
x
yz
3D-Shell Finite Element Model
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
1.1
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2
Load Factor
Fres
(MN
)
FresMx
11.0
Onset of Damage
Peak Mx
Peak Fres
0.0
7.0
1.0
3.0
2.0
4.0
5.0
6.0
8.0
9.0
10.0
Mx
(kN
-m)
Damage Propagation from PFA
Test Failure Load 907 kNPredicted Failure Load896 kN
Failed Test LugFailed Test Lug
(b) View through the outer thickness of the lug
Unsymmetric Damage Due to Configuration and
Applied Moments
Elements with Predicted Damage
(a) View normal to the hole
(b) View through the outer thickness of the lug
Unsymmetric Damage Due to Configuration and
Applied Moments
Elements with Predicted Damage
(a) View normal to the hole
(b) View through the outer thickness of the lug
Unsymmetric Damage Due to Configuration and
Applied Moments
Elements with Predicted Damage
(a) View normal to the hole
(b) View through the outer thickness of the lug
Unsymmetric Damage Due to Configuration and
Applied Moments
Elements with Predicted Damage
(a) View normal to the hole
(c) View through the outer thickness of the lug
(b) View through the outer thickness of the lug
Unsymmetric Damage Due to Configuration and
Applied Moments
Elements with Predicted Damage
(a) View normal to the hole
(b) View through the outer thickness of the lug
Unsymmetric Damage Due to Configuration and
Applied Moments
Elements with Predicted Damage
(a) View normal to the hole
(b) View through the outer thickness of the lug
Unsymmetric Damage Due to Configuration and
Applied Moments
Elements with Predicted Damage
(a) View normal to the hole
(b) View through the outer thickness of the lug
Unsymmetric Damage Due to Configuration and
Applied Moments
Elements with Predicted Damage
(a) View normal to the hole
(b) View through the outer thickness of the lug
Unsymmetric Damage Due to Configuration and
Applied Moments
Elements with Predicted Damage
(a) View normal to the hole
(b) View through the outer thickness of the lug
Unsymmetric Damage Due to Configuration and
Applied Moments
Elements with Predicted Damage
(a) View normal to the hole
(b) View through the outer thickness of the lug
Unsymmetric Damage Due to Configuration and
Applied Moments
Elements with Predicted Damage
(a) View normal to the hole
(b) View through the outer thickness of the lug
Unsymmetric Damage Due to Configuration and
Applied Moments
Elements with Predicted Damage
(a) View normal to the hole
(c) View through the outer thickness of the lug
Elements with predicted damage
Unsymmetricdamage
due to loading
Predicted Failures
Comparison of Predicted and Test Results
(b) View through the outer thickness of the lug
Unsymmetric Damage Due to Configuration and
Applied Moments
Elements with Predicted Damage
(a) View normal to the hole
(b) View through the outer thickness of the lug
Unsymmetric Damage Due to Configuration and
Applied Moments
Elements with Predicted Damage
(a) View normal to the hole
(b) View through the outer thickness of the lug
Unsymmetric Damage Due to Configuration and
Applied Moments
Elements with Predicted Damage
(a) View normal to the hole
(b) View through the outer thickness of the lug
Unsymmetric Damage Due to Configuration and
Applied Moments
Elements with Predicted Damage
(a) View normal to the hole
(c) View through the outer thickness of the lug
(b) View through the outer thickness of the lug
Unsymmetric Damage Due to Configuration and
Applied Moments
Elements with Predicted Damage
(a) View normal to the hole
(b) View through the outer thickness of the lug
Unsymmetric Damage Due to Configuration and
Applied Moments
Elements with Predicted Damage
(a) View normal to the hole
(b) View through the outer thickness of the lug
Unsymmetric Damage Due to Configuration and
Applied Moments
Elements with Predicted Damage
(a) View normal to the hole
(b) View through the outer thickness of the lug
Unsymmetric Damage Due to Configuration and
Applied Moments
Elements with Predicted Damage
(a) View normal to the hole
(b) View through the outer thickness of the lug
Unsymmetric Damage Due to Configuration and
Applied Moments
Elements with Predicted Damage
(a) View normal to the hole
(b) View through the outer thickness of the lug
Unsymmetric Damage Due to Configuration and
Applied Moments
Elements with Predicted Damage
(a) View normal to the hole
(b) View through the outer thickness of the lug
Unsymmetric Damage Due to Configuration and
Applied Moments
Elements with Predicted Damage
(a) View normal to the hole
(b) View through the outer thickness of the lug
Unsymmetric Damage Due to Configuration and
Applied Moments
Elements with Predicted Damage
(a) View normal to the hole
(c) View through the outer thickness of the lug
Elements with Predicted Damage
FWDFWDUnsymmetricDamage Due to
Configuration and Applied Moments
LoadLoad
12
W375 Accident Conditions – Damage
Normalized Failure Load for 1985-Certification Test, 2003-Subcomponent Test and W375 Accident Condition
0.0
0.5
1.0
1.5
2.0
2.5
3.0
1985 Test SC Test W375Accident Case
Nor
mal
ized
Fai
lure
Loa
d,kN
PFA Analysis Failure LoadPFA Analysis Load at Maximum Moment MXTest Failure Load
0.0
0.5
1.0
1.5
2.0
2.5
3.0
1985 Test SC Test W375Accident Case
Nor
mal
ized
Fai
lure
Loa
d,kN
PFA Analysis Failure LoadPFA Analysis Load at Maximum Moment MXTest Failure Load
Dislocations
Twinning
Stacking Fault[1 1 0]
[0 0 1]
x
y
[1 1 0]
Void Coalescence
10 nm
Damage ScienceDamage Science
20 20 μμmm
Emerging Continuum MethodsEmerging Continuum Methods
Com
plex
ity, C
ompu
tatio
nal E
xpen
se
Time
1/41/23/4
0
1
G /(G + G )ΙΙ ΙΙ ΙΙΙ
GT
GIc + GIIc − GIc( ) GII +GIII
GT
⎛
⎝ ⎜
⎞
⎠ ⎟
η
+ GIIIc − GIIc( ) GIII
GII +GIII
GII +GIII
GT
⎛
⎝ ⎜
⎞
⎠ ⎟
η ≥ 1
t≈0
LaRC Decohesion Element(Technology adopted by ABAQUS)
Mixed-Mode Fracture
Bilinear Traction-Displacement Law
Perfectly plastic (pp)
Linear softening (lin)
Progressive softening (pro)
Regressive softening (reg)
G
σ
pro lin reg
c
σ
Needleman (Ne)
Nepp0
c
CGdF
=∫ δδσδ
0 )(
QuickTime™ and a decompressor
are needed to see this picture.
Wide range of element sizes
Narrow range
Classical
Enhanced
Enhanced formulation allows the useof elements up to three times larger thanwith the classical damage model.
MMB Specimen
Delamination growth
FY05 Advances
Current StateCurrent State--ofof--thethe--ArtArt
0
0.5
1
1.5
2
2.5
3
0 50 100 150 200
G I
in-lb/in 2
Pressure, P (psi)
b=0.750in.b=0.625in.b=0.500in.
Evolution of Damage Tolerance at NASA Langley (1999-)
Building Block Integration.
Verification of Design Data and Methodology
Development of Design Data
Number of Specimens
Chr
onol
ogic
al S
eque
nce
Spec
imen
Com
plex
ity
Certification Methodology (Mil-Hbk.-17)
Structural Levels of Testing & Analysis
Static/Fatigue
Material Selection and Qualifications Coupons
Design Allowables Coupons
Structural Elements
Sub-components
Components
Full Scale Article
Ana
lysi
sA
naly
sis Reduced Testing
More accurate design tools reduced recurring costs
Reduced reliance on testingFaster design process reduced non-recurring costs
High-Fidelity Progressive Failure Analysis
Building Block Approach
Conceptual Preliminary Detailed
DesignKnowledge,DesignFreedom
Time into Design Process
DesignFreedom
Goal
GoalKnowledge about Design
Design Freedom vs. Knowledge
Building Block Integration.
Verification of Design Data and Methodology
Development of Design Data
Number of Specimens
Chr
onol
ogic
al S
eque
nce
Spec
imen
Com
plex
ity
Certification Methodology (Mil-Hbk.-17)
Structural Levels of Testing & Analysis
Static/Fatigue
Material Selection and Qualifications Coupons
Design Allowables Coupons
Structural Elements
Sub-components
Components
Full Scale Article
Ana
lysi
sA
naly
sis Reduced Testing
More accurate design tools reduced recurring costs
Reduced reliance on testingFaster design process reduced non-recurring costs
High-Fidelity Progressive Failure Analysis
Building Block Approach
• Failure Criteria are used for predicting damage initiation and final failure
• Composites have multiple damage modes; each requires a different criterion
• Failure Criteria are used for predicting damage initiation and final failure
• Composites have multiple damage modes; each requires a different criterion
σ11σ11
σ22σ22
m
m
-XC-XC
ϕ
2a0
σ22, YT
tτ12, S
Lis is
α=53º
τL
τTσn
σ22
τ12α
Fiber orientation
Matrix Tension & Shear
Fiber Compression
Matrix Compression and Shear
LaRC04 Criteria• In-situ matrix strength
prediction• Advanced fiber kinking
criterion• Prediction of angle of
fracture (mat. compression)• Criteria used as activation
functions within framework of damage mechanics
LaRC04 Criteria• In-situ matrix strength
prediction• Advanced fiber kinking
criterion• Prediction of angle of
fracture (mat. compression)• Criteria used as activation
functions within framework of damage mechanics
Failure Criteria for Laminated Composites
Gibbs Free Energy
Strains:
Lamina Secant Relation
Rate of Damage Growth
fi: LaRC04 failure criteria as activation functions
Softening
Compression Tension
CDM ensures consistent material degradation and mesh-independent solution
LaRC04 in Continuum Damage Model
DefinitionPultruded graphite rods positioned through-thickness (z-direction) of a composite laminatePins are 0.2-0.5mm diameter rodsTypical range of areal density between 0.5% and 4%Inserted into uncured laminate using ultrasonic hammer
Purposes / DrawbacksImprove composite laminate transverse strengthProhibit delaminationDegrade laminate (in-plane) properties, see micrograph
ApplicationsAreas with significant out-of-plane loads such as bonded stiffener terminationAreas exposed to impact damage threat
Z-Pin preform: Insertion side**Z-Pin preform: Upper side**
*James Ratcliffe, NIA. **Pierre Minguet, Boeing. ***Jeffery Schaff, Sikorsky Aircraft.
Z-Pins protruding from laminate*
Fiber misalignment from z-pins***
z-pins
Z-Pin Technology
GT
GIc + GIIc − GIc( ) GII
GT
⎛
⎝ ⎜
⎞
⎠ ⎟
η ≥ 1B-K Criterion
2D Fracture Criterion(only Mode I and Mode II)
3D Problems Contain Mode III
sublaminate buckling problem
Pure Mode III Testing
edge crack torsion test
ECT produces pure mode III data
GIIIc normally higher than GIIc
No mixed-mode test with mode III component
circulardelamination
contours show out-of-plane
displacement
New Delamination Criterion Needed
1/41/23/4
0
1
G /(G + G )ΙΙ ΙΙ ΙΙΙ
GT
GIc + GIIc − GIc( ) GII +GIII
GT
⎛
⎝ ⎜
⎞
⎠ ⎟
η
+ GIIIc − GIIc( ) GIII
GII +GIII
GII +GIII
GT
⎛
⎝ ⎜
⎞
⎠ ⎟
η ≥ 1
Mode I-III interaction similar to the measured mode I-II interactionLinear interpolation between mode III and mode II
Proposed 3D Mixed-Mode Criterion
Dislocations
Twinning
Stacking Fault[1 1 0]
[0 0 1]
x
y
[1 1 0]
Void Coalescence
10 nm
Damage ScienceDamage Science
20 20 μμmm
Emerging Continuum MethodsEmerging Continuum Methods
Com
plex
ity, C
ompu
tatio
nal E
xpen
se
Time
1/41/23/4
0
1
G /(G + G )ΙΙ ΙΙ ΙΙΙ
GT
GIc + GIIc − GIc( ) GII +GIII
GT
⎛
⎝ ⎜
⎞
⎠ ⎟
η
+ GIIIc − GIIc( ) GIII
GII +GIII
GII +GIII
GT
⎛
⎝ ⎜
⎞
⎠ ⎟
η ≥ 1
t≈0
LaRC Decohesion Element(Technology adopted by ABAQUS)
Mixed-Mode Fracture
Bilinear Traction-Displacement Law
Perfectly plastic (pp)
Linear softening (lin)
Progressive softening (pro)
Regressive softening (reg)
G
σ
pro lin reg
c
σ
Needleman (Ne)
Nepp0
c
CGdF
=∫ δδσδ
0 )(
QuickTime™ and a decompressor
are needed to see this picture.
Wide range of element sizes
Narrow range
Classical
Enhanced
Enhanced formulation allows the useof elements up to three times larger thanwith the classical damage model.
MMB Specimen
Delamination growth
FY05 Advances
Current StateCurrent State--ofof--thethe--ArtArt
0
0.5
1
1.5
2
2.5
3
0 50 100 150 200
G I
in-lb/in 2
Pressure, P (psi)
b=0.750in.b=0.625in.b=0.500in.
Evolution of Damage Tolerance at NASA Langley (1999-)
[0 0 1]
Continuum Representation of Atomistic Behavior
Mechanisms of Nano-crack Propagation
Computational Damage Science
KI = 0.36 MPa-m1/2
stress profileσyy(x)
opening profile δn(x)
δ n[n
m]
σsyy
[GP
a] Brittle Tip
Twinning Tip
Dislocations
Twinning
Stacking Fault[1 1 0]
[0 0 1]
x
y
[1 1 0]
Void Coalescence
10 nm
Calculation of Normal Stress and Crack Opening
Develop a fundamental understanding of the underlying damage processes that contribute to fracture initiation and propagation
20 20 μμmm
SEM micrograph of fatigue crack tip
200 200 μμmm
SEM micrograph of fatigue crack emanating
from EDM notch
Micro-Scale Crack Growth
Experimental Damage Science
Materials characterization and in-situ mechanical testing with
environmental capabilities
In-situ loading frame with heater/cooler and specimen tilt for EBSD analysis
Testspecimen
Develop multi-disciplinary Damage Science approach to:1) Characterize material structure and characteristic damage processes, 2) Develop multi-scale models to predict damage, and 3) Validate models through examination of near-tip damage processes.
Damage Science
Damage tolerance of composite wing boxes and full scale wing structures
Textile compositesStitchingEfficient analysis methods
SOA analysis demonstrated onX-33 LH2 tank failureAA587 composite lug analysis
Emerging continuum methodsNew criteria for interlaminar and intralaminar failureContinuum damage models - Mesh independenceZ-pinning
Damage science to understand failure initiation and growth -Damage Tolerance
Summary
Test Article failed at 83% of DUL under combined bending & torsionUnanticipated shear failure mode at out-of-tolerance gap
AS4/1806 and AS4/974
NASA ACT Program – Center Wing Box Test (1991)
Test article failed at 94% of DUL due to nonvisible impact damageCompression after impact (CAI) strength allowable did not account
for damaged elements (skin/stiffener) interaction
Composite stub box
AS4/3501-6 and IM7/3501-6 in textile preform
NASA ACT Program – Wing Stub Box Test (1996)
Upper Cover Compression Panel
Stringer RunoutPanel
Tension
Skin Ply Drop-off
Bulkhead Shear Clip Specimen
Compression
Upper Cover Splice Specimen
Rib Clip Pull-off SpecimenPostbuckling
Specimen
Hi-Load Panels
Spar Cap Longitudinal Shear Panel
Repair Panel
Substructure Shear Specimen
Building Block Approach – Reliance on Extensive Testing
Modeling ComplexitiesFailure of unidirectional and laminated
composites (in-situ)
Material nonlinearity & material degradation laws
Thermal residual stresses
Effects of stress gradients & notches
Size Effects
Finite Element implementation
Delamination growth: static & fatigue
Damage mode interaction
Stitched composites and textiles
LaRC04 Failure Criteria
LaRC04 Decohesion Elements
FY04 FY05-06
Enhanced Decohesion
Element&
High-Cycle Fatigue Model
Continuum Damage Model
In-Situ Strengths
Progressive Damage Analysis