an experimental study of the fracture behaviour of ... · overheight compact tension (oct) test...

An Experimental Study of the Fracture Behaviour of Stitched RFI Composites –

Evaluation of the Damage Process Zone to Calibrate a Strain-Softening Model

Jason Mitchell and Anoush Poursartip

Department of Materials Engineering The University of British Columbia,

Vancouver, British Columbia, Canada

Corresponding author:Anoush PoursartipE-mail: [email protected]: +1(604)822-3665Fax: +1(604)822-3619

AcknowledgementsAcknowledgements

Natural Sciences and Engineering Research Council of CanadaThe Boeing Company

Mr. Roger Bennett, UBCDr. Reza Vaziri, UBCDr. Göran Fernlund, UBCDr. William Avery, The Boeing Company

*copyright Tourism Vancouver

OutlineOutline

IntroductionMaterialsOCT TestingPost OCT TestsDamagePreview of Numerical PredictionsConclusionsFuture Work

Stitched Resin Film Infused MaterialsStitched Resin Film Infused Materials

S/RFI is one approach to compensate for poor out-of-plane properties of composite laminatesThe aim is to improve damage tolerance to a level competitive with metallic materials, as well as other benefits

Flow

Tool Resin Preform Dam

Flow

Tool Resin Preform Dam

Stitched Resin Film Infused MaterialsStitched Resin Film Infused Materials

Produced by stitching a preformResin is infused via vacuum and heat from a partially cured film placed under the preformS/RFI material used here developed by The Boeing Company under the NASA ACT Program

Stitched Resin Film Infused MaterialStitched Resin Film Infused MaterialAdvantages of stitching– Shown to improve delamination resistance under low velocity

impact loading– Large stitched preforms easier to handle, cost effective– Removes the need for traditional mechanical fasteners

Disadvantage of stitching– Imparts fibre damage; spreading and breakage

Elongated Stitch 2.0 mmElongated Stitch 2.0 mm

Top of Stitch

1.48 mm1.48 mm

Side of Stitch

Aims of StudyAims of Study

There appears to have been no research into the effect of stitching on fracture response of notched composites. Other research has been mainly involved with inhibiting the initiation of damage or reducing amount of delamination caused by impact loading. Thus:

1. To gain a fundamental understanding of how damage initiates and propagates in notched RFI materials.

– allowing for more accurate physically based inputs to a continuum damage mechanics strain softening model

2. To investigate what effect the introduction of stitching may have on in-plane fracture response of notched RFI materials.

Materials Examined in this StudyMaterials Examined in this Study

Stitched and Unstitched Resin Film Infused (RFI) Laminate Systems

Carbon Fibre (std. and intermediate modulus fibres) – Epoxy Resin (3501-6)

though-thickness reinforcement (S) compared to none (U)Lay-up [+45/-45/02/90/-45/+45]s (where s=4, 5 or 6)

Stitched/RFI Panel Panel and Stiffener – joined by Kevlar® Stitching

Experimental MethodOverheight Compact Tension (OCT) Tests

Experimental MethodOverheight Compact Tension (OCT) Tests

Overheight Compact Tension (OCT) test methodology based on work of Kongshavn and Poursartip1.Loaded from pins on notch flanksSaw-cut slightly blunt notchLines inscribed to measure local surface displacements (line analysis)Extensometer records crack mouth opening displacement (CMOD) of notch

Applied Displacement

Notch

Inscribed LinesCMOD

GaugeGuide

Loading Pins

y

x1 Kongshavn and Poursartip, 1999

Why use the OCT?Why use the OCT?

Geometry has been shown to promote stable damage growth in tape compositesSmall specimen size– 207× 104 mm– Material waste kept to a

minimumDamage includes fibre fracture in a small specimen– usually only seen in large

specimensFracture Process Zone (FPZ) is large enough for post-test analysis.

FPZ

Test ApparatusTest Apparatus

Specimen loaded in InstronUniversal Testing Machine

y

x

Typical Stitched 5-stack Material Behaviour(Test # SS67-s0)

Load – Crack Mouth Opening Displacement Curve

05

101520253035404550

0 1 2 3 4 5 6 7CMOD (mm)

Load

(kN

)

Photographs

Tests ConductedTests Conducted

Specimens were tested in both the primary principal (0°) and the transverse principal (90°) directionsFewer unstitched experiments were performed as results compared favourably with data from previous research

90°90°

32 102

0°0°

134

Stitched (4/5/6)Unstitched (4/6)

17 Tests Performed

Test OrientationsTest Orientations

Applied Displacement, δ

Direction of 0ºplies



0°





90°

Specimens were tested in both the primary principal (0°) and the transverse principal (90°) directions.

Test Directions (Stitching)Test Directions (Stitching)Kevlar® stitching is oriented in direction of 0° plies

90°

Stitching oriented

parallel to load

Stitching oriented

perpendicular to load

0°

Effect of Thickness and Stitching on 90ºSpecimens


6-stack 90° (SS81-u90)

6-stack 90° (SS78-s90)

4-stack 90° (SS76-s90)

4-stack 90° (SS83-u90)

5-stack 90° (SS77-s90)

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

2

0 1 2 3 4 5CMOD (mm)

Load

/Thi

ckne

ss (k

N/m

m)

6-stack 90° (SS81-u90)

6-stack 90° (SS78-s90)

4-stack 90° (SS76-s90)

4-stack 90° (SS83-u90)

5-stack 90° (SS77-s90)

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

2

0 1 2 3 4 5CMOD (mm)

Load

/Thi

ckne

ss (k

N/m

m)

a/W = 0.40



0

1

2

3

4

5

6

0 1 2 3 4 5 6CMOD (mm)

Load

/Thi

ckne

ss (k

N/m

m)

5-stack 0° (SS70-s0)

4-stack 0° (SS82-u0)







0

1

2

3

4

5

6

0 1 2 3 4 5 6CMOD (mm)

Load

/Thi

ckne

ss (k

N/m

m)









a/W ~ 0.40-0.43

Summary of P-δ Response TypesSummary of P-δ Response Types

BABA6--BB5BABA4

90°0°90°0°UnstitchedStitched

BABA6--BB5BABA4

90°0°90°0°UnstitchedStitched

Bδ

P

A

P

δB

δ

P

A

P

δ

Post Test AnalysisPost Test Analysis

Used to evaluate the fracture response– Damage initiation and propagation

To investigate the damage evolved in Fracture Process Zone:– Sectioning -- (fibre damage, delamination, matrix

damage)– Deplying -- (fibre damage, fracture path)

To investigate progression of damage across specimen width– Surface Line Displacement Analysis

measures material displacement in front of notch tip

SectioningSectioning

Sectioned Area

Process ZoneNotch

Loading Pins Sectioned Area

Process ZoneNotch

Loading Pins

Sectioning (Example)Sectioning (Example)

0

7.50

-7.50

NotchM.P

-5.63

-1.88

-3.75

5.63

3.75

1.88

7.00 mm

90 90 90 90 90

TopF B TopF B

0

7.50

-7.50

NotchM.P

-5.63

-1.88

-3.75

5.63

3.75

1.88

7.00 mm

90 90 90 90 90

TopF B TopF B

0

7.50

-7.50

NotchM.P

-5.63

-1.88

-3.75

5.63

3.75

1.88

7.00 mm

90 90 90 90 90

TopF B TopF B

Cross-sectioning technique performed on a stitched 5-stack 0°specimen (SS70-s0-B) at 2.00 mm in front of the notch tip.

DeplyDeplyDeplied Area

Process ZoneNotch

Loading Pins Deplied Area

Process ZoneNotch

Loading Pins

Each ply is removed

individually

Section heated at ~400°C for 4

hours to oxidise off resin

Deplied Area

Process ZoneNotch

Loading Pins Deplied Area

Process ZoneNotch

Loading Pins

Each ply is removed

individually

Section heated at ~400°C for 4

hours to oxidise off resin

Deply (Example)Deply (Example)

+452

-45

02

02

-45+45

-45 02

90 02 -45

1 2 3

7

8 10 11

13 14 16

+4518

0221

905 5

+4517 -4519

9022-4520

Ply Drop

+4

+6

+9 +12

+15

+452

-45

02

02

-45+45

-45 02

90 02 -45

1 2 3

7

8 10 11

13 14 16

+4518 +4518

0221

905 905 5

+4517 -4519 -4519

9022-4520

Ply Drop

+4

+6

+9 +12

+15

9002

-4502

-45

23 24 25

29

30 31 32

34 35 36

38

27 28

37 39

+452+26

-45

02 +452+33

-45 02 90

02 -45 +45

61 mm

9002

-4502

-45

23 24 25

29

30 31 32

34 35 36

38

27 28

37 39

+452+26

-45

02 +452+33

-45 02 90

02 -45 +45

9002

-4502

-45

23 24 25

29

30 31 32

34 35 36

38

27 28

37 39

+452+26

-45

02 +452+33

-45 02 90

02 -45 +45

61 mm61 mm

Documentation of SS69-s05-stack S/RFI 0°

Significant Fibre BreakageSignificant Fibre Breakage

Notch mid-plane

Fibre Damage

11.75 mm

Notch mid-planeNotch mid-plane

Fibre Damage

11.75 mm

Notch mid-plane

Damage at 45°

11.75 mm

Notch mid-plane

Damage at 45°

11.75 mm

This example: 5-stack 0°

damage growth in 0° plies damage growth in -45° plies

Surface Line AnalysisSurface Line Analysis

0

5

10

15

20

25

30

35

40

45

50

0 1 2 3 4 5 6 7CMOD (mm)

Load

(kN

)

Photographs (See Line Analysis SS69-s0)

3

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

-10 0 10 20 30 40 50Position in front of notch tip (mm)

Dis

plac

emen

t of l

ine

#5(1

2.5

mm

from

NM

P) (m

m)

Photo 3

Back End of specimen

NOTCH

Relating P-CMOD curves to Surface Line

Analysis Plots

Load- CMOD

Line Analysis

Specimen Surface

0

5

10

15

20

25

30

35

40

45

50

0 1 2 3 4 5 6 7CMOD (mm)

Load

(kN

)


3

6

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4


Dis

plac

emen

t of l

ine

#5(1

2.5

mm

from

NM

P) (m

m)

Photo 3

Photo 6


NOTCH


Analysis Plots

Photo 6

0

5

10

15

20

25

30

35

40

45

50

0 1 2 3 4 5 6 7CMOD (mm)

Load

(kN

)


3

67

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4


Dis

plac

emen

t of l

ine

#5(1

2.5

mm

from

NM

P) (m

m)

Photo 3

Photo 6

Photo 7


NOTCH


Analysis Plots

Photo 7

0

5

10

15

20

25

30

35

40

45

50

0 1 2 3 4 5 6 7CMOD (mm)

Load

(kN

)


3

8

67

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4


Dis

plac

emen

t of l

ine

#5(1

2.5

mm

from

NM

P) (m

m)

Photo 3

Photo 6

Photo 7

Photo 8


NOTCH


Analysis Plots

Photo 8

0

5

10

15

20

25

30

35

40

45

50

0 1 2 3 4 5 6 7CMOD (mm)

Load

(kN

)


9

3

8

67

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4


Dis

plac

emen

t of l

ine

#5(1

2.5

mm

from

NM

P) (m

m)

Photo 3

Photo 6

Photo 7

Photo 8

Photo 9


NOTCH


Analysis Plots

Photo 9

0

5

10

15

20

25

30

35

40

45

50

0 1 2 3 4 5 6 7CMOD (mm)

Load

(kN

)


9

3

12

8

67

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4


Dis

plac

emen

t of l

ine

#5(1

2.5

mm

from

NM

P) (m

m)

Photo 3

Photo 6

Photo 7

Photo 8

Photo 9

Photo 12


NOTCH


Analysis Plots

Photo 12

0

5

10

15

20

25

30

35

40

45

50

0 1 2 3 4 5 6 7CMOD (mm)

Load

(kN

)


9

3

1412

8

67

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4


Dis

plac

emen

t of l

ine

#5(1

2.5

mm

from

NM

P) (m

m)

Photo 3Photo 6Photo 7Photo 8Photo 9Photo 12Photo 14


NOTCH


Analysis Plots

Photo 14

0

5

10

15

20

25

30

35

40

45

50

0 1 2 3 4 5 6 7CMOD (mm)

Load

(kN

)


9

3

221412

8

67

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4


Dis

plac

emen

t of l

ine

#5(1

2.5

mm

from

NM

P) (m

m)

Photo 3Photo 6Photo 7Photo 8Photo 9Photo 12Photo 14Photo 22


NOTCH


Analysis Plots

Photo 22

0

5

10

15

20

25

30

35

40

45

50

0 1 2 3 4 5 6 7CMOD (mm)

Load

(kN

)


9

3

2214

24

12

8

67

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4


Dis

plac

emen

t of l

ine

#5(1

2.5

mm

from

NM

P) (m

m)

Photo 3Photo 6Photo 7Photo 8Photo 9Photo 12Photo 14Photo 22Photo 24


NOTCH


Analysis Plots

Photo 24

0

5

10

15

20

25

30

35

40

45

50

0 1 2 3 4 5 6 7CMOD (mm)

Load

(kN

)


9

3

2214

24

28

12

8

67

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4


Dis

plac

emen

t of l

ine

#5(1

2.5

mm

from

NM

P) (m

m)

Photo 3Photo 6Photo 7Photo 8Photo 9Photo 12Photo 14Photo 22Photo 24Photo 28


NOTCH


Analysis Plots

Photo 28

0

5

10

15

20

25

30

35

40

45

50

0 1 2 3 4 5 6 7CMOD (mm)

Load

(kN

)


9

3

2214

24

28

29

12

8

67

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4


Dis

plac

emen

t of l

ine

#5(1

2.5

mm

from

NM

P) (m

m)

Photo 3Photo 6Photo 7Photo 8Photo 9Photo 12Photo 14Photo 22Photo 24Photo 28Photo 29


NOTCH


Analysis Plots

Photo 29

Correlation of P-CMOD Curves and Line Analysis


Self-similar damage growth is associated with:

– load-softening behaviour in P-CMOD– steady moving damage front in Line Analysis

plot

05

101520253035404550

0 1 2 3 4 5 6 7CMOD (mm)

Load

(kN

)

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

-10 0 10 20 30 40 50

Position in front of notch tip (mm)

Dis

plac

emen

t of l

ine

#5(1

2.5

mm

from

NM

P) (m

m)

Photo 3Photo 6Photo 7Photo 8Photo 9Photo 12Photo 14Photo 22Photo 24Photo 28Photo 29

SS69-s0 Stitched 5-stack

Techniques are consistentTechniques are consistent

Deply

Compliance

Line Analysis

Heavy Fibre Damage

Light Fibre Damage

27 mm

26.6 mm

34 mm

4mm

Notch Tip

Notch

Notch

Notch

Deply

Compliance

Line Analysis

Heavy Fibre Damage

Light Fibre Damage

27 mm

26.6 mm

34 mm

4mm

Notch Tip

Notch

Notch

Notch

Self-similar Crack Growth (A)Self-similar Crack Growth (A)

Notch Crack Process Zone

Virgin MaterialNotch Crack Process

ZoneVirgin

MaterialNotch Crack Process Zone

Virgin Material

All 90º and 5-stack 0º materials

0º Behaviour: 4 and 6-stack vs. 5-stack0º Behaviour: 4 and 6-stack vs. 5-stack

Only the stitched 4 and 6-stack 0°did not show significant amount of fibre damage

– delamination damage, blunting few millimeters from notch tip

4 and 6-stack 0º

BehaviourStitched and Unstitched

5- stack (and 4, 5 and 6-stack 90º

Behaviour)

Applied displacement,δ

Stitching (Perpendicular to Load/applied displacement)

ii)

Applied displacement,δ

Stitching (Perpendicular to Load/applied displacement)

ii)



Blunting (Delamination) fracture response is associated with:

– absence of load-softening in P-CMOD– characteristic kinks in the Line Analysis plot

0.0

0.2

0.4

0.6

0.8

1.0

1.2

-10 0 10 20 30 40 50

Position in front of notch tip (mm)

Dis

plac

emen

t of l

ine

#5(1

2.5

mm

from

NM

P) (m

m)

0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

8.0

9.0

10.0

Equi

vale

nt S

trai

n (%

)

Photo 15

Photo 23

Photo 28

Photo 34

Photo 36

SS73-s0 Stitched 6-stack

0

5

10

15

20

25

30

35

40

45

50

0 1 2 3 4 5 6 7CMOD (mm)

Load

(kN

)


15

2328

3436

0

5

10

15

20

25

30

35

40

45

50

0 1 2 3 4 5 6 7CMOD (mm)

Load

(kN

)


15

2328

3436

Blunting and Specimen Size EffectBlunting and Specimen Size Effect

There was no self-similar crack growth in the OCT specimen for the 4 and 6-stack 0º materialsHowever, this is because the damage height is 60-70 mm as opposed to the ~5 mm seen for the all 90ºspecimens and 5-stack 0ºmaterialTo get self-similar crack growth, specimen size must be large enough

2-5 mm

~60-70 mm

Delamination in 4- and 6-stack Specimens

2-5 mm

~60-70 mm

Delamination in 4- and 6-stack Specimens

Surface Loops in stitched/RFI materialSurface Loops in stitched/RFI material

4-stack4-stack

6-stack6-stack

5-stack5-stack

Tight Stitching vs. Loose StitchingTight Stitching vs. Loose Stitching

For this notched geometry, tight stitching does not inhibit the natural tendency of these materials to delaminate

– Damage height does not change as much

4-stack (SS70-s0)stitched 0°

Blunting ResponseTight Stitching


Brittle Response Loose Stitching


Blunting ResponseTight Stitching

Sections 2 mm ahead of notch at notch mid-plane

1.48 mm1.48 mm 1.82 mm1.82 mm 2.20 mm2.20 mm2.20 mm

Stitching Inhibits Surface DelaminationStitching Inhibits Surface Delamination

4-stack un-stitched 0°specimen(SS82-u0)

4-stack stitched 0°specimen(SS72-s0)

Strain-softening material modelStrain-softening material model

Three parameter input:fracture energy release rate GF

material tensile strength σpeak or strain at σpeak

initial modulus E0

A simple bilinear damage model has been implemented in ABAQUS finite element code.

• implicit solution scheme allows efficient solution of quasi-static tests

• orthotropic → parameters defined in each direction

G (J/m2) = γ x he

σ

εεpeak

σpeak

γ

εf

McClennan M.A.Sc work 2004

Numerical Predictions: Parametric StudyNumerical Predictions: Parametric Study

0

2

4

6

8

10

12

14

16

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5CMOD (mm)

Load

(kN

)

Increasing σ peak

(420, 440, 460, 480 MPa)

Increasing G c

(70, 75, 80, 85 kJ/m2)


Load - CMOD ResultsLoad - CMOD Results

0

2

4

6

8

10

12

14

16

0.0 1.0 2.0 3.0 4.0 5.0

CMOD (mm)

Load

(kN

)

0

2

4

6

8

10

12

14

16

0.0 1.0 2.0 3.0 4.0 5.0

CMOD (mm)

Load

(kN

)ρ = 1 mm

ρ = 4 mm

E0 = 32 GPa

GF = 80 kJ/m2

σpeak = 460 MPa

Input Parameters:


Input Parameters:

0

2

4

6

8

10

12

14

16

18

0.0 1.0 2.0 3.0 4.0 5.0

CMOD (mm)

Load

(kN

)

Load - CMOD ResultsLoad - CMOD Results

ρ = 32 mm

ρ = 38 mm

E0 = 32 GPa

GF = 80 kJ/m2

σpeak = 460 MPa0

2

4

6

8

10

12

14

16

18

0.0 1.0 2.0 3.0 4.0 5.0

CMOD (mm)

Load

(kN

)


ConclusionsConclusions

Self-similar crack growth consisting of a process zone that coalesces into a through crack observed

– For 3 out of 4 cases tested using OCT geometry– Damage height was ~5 mm, consisting of fibre breakage,

matrix cracking and delaminationOne case (0º 4 and 6-stack) did not show self-similar growth, and blunted in this OCT geometry

– Much larger damage height at ~65 mm, and thus unable to grow in self-similar manner in this OCT geometry

In this OCT geometry, tight stitching had no effect on behaviour of 0º specimens

– Large delamination height leading to blunting.– Fibre damage due to stitching present but did not affect

measured behaviour

Conclusions (2)Conclusions (2)

However, loose stitching, as observed in 5-stack panel, had an effect, inhibiting delamination and leading to self-similar crack growth– Poses question regarding stitch tension control and quality

controlCombination of P-δcurves, line analysis, sectioning, and deply provides very coherent and consistent picture of response– Combined with other tests such as tensile and bend tests,

allows for calibration of damage and strain-softening modelsA simple strain-softening model does good job of predicting response– Including transition from stable to unstable response

an experimental study of the fracture behaviour of ... · overheight compact tension (oct) test...

Documents