An investigation into the face sheet (skins) debonding of glass balsa sandwich composites

An investigation into the face sheet (skins) debonding of glass balsa sandwich

composites

Comptest Lausanne 02/2011

Dr M. Colin de Verdiere ([email protected])

Professor J.M Dulieu - Barton, Professor R.A Shenoi

and Dr J.I.R Blake

Content• Introduction.

• Manufacture of specimens.

• Testing for debonding characterisation.

• Material and crack characteristics.

• Experimental results.

• Digital image correlation for added material

information.

• Parameters estimation.

• Conclusion.

Introduction

• Advantages

• Requirement

Mine blast

Mine countermeasure vessels using glass-balsa sandwiched structures

Weather (hail)

Tool falling on deck

Rough seas

Variability of balsa wood

Defects (gaps in between blocks)

Each blocks has different mechanical properties

End grain balsa sheets are made of many different blocks

Balsa core coated in resin and drying in oven to avoid excessive resin

absorption

Specimen manufacture

Large panel

Reduced panel

Crack film Crack film

Sp

eci

men

1

Sp

eci

men

2

Sp

eci

men

3

Sp

eci

men

4

Sp

eci

men

5

Sp

eci

men

6

Sp

eci

men

7

Sp

eci

men

8

Sp

eci

men

9

Sp

eci

men

1

0

Sp

eci

men

1

1

Sp

eci

men

12

Sp

eci

men

13

Sp

eci

men

1

Sp

eci

men

2

Sp

eci

men

3

Sp

eci

men

4

Sp

eci

men

5

Sp

eci

men

6

Sp

eci

men

7

Sp

eci

men

8

Sp

eci

men

9

Sp

eci

men

1

0

Sp

eci

men

1

1

Sp

eci

men

14

Sp

eci

men

15

Sp

eci

men

1

6

Sp

eci

men

1

7

Sp

eci

men

1

8

Sp

eci

men

19

Sp

eci

men

20

Sp

eci

men

21

Sp

eci

men

22

Sp

eci

men

23

Sp

eci

men

2

4

Sp

eci

men

25

Sp

eci

men

26

Specimen manufacture

Vacuum bagFlow mediaPeel plyDBL 800CSM Mat

•CSM Mat•DBL 800•Peel ply•Flow media•Mould

Baslsa coreCrack film

Specimen characteristics

Face sheet (skin)

Core

Face sheet (skin)

Pre crack area

Width: 35 mmCrack length: 50 mmLength: 200mmCore thickness: 13-40mmFace sheet (skins) thickness: 4 mm

Specimen characteristicsDuring debonding of the face sheet (skins) the following effects are looked at:

•Core thickness

•Crack film thickness

•CSM mat or no CSM mat

•Epoxy or vinylester resin

13 mm

40 mm

14 μm 60 μm 120 μm

CSM Mat No CSM Mat

Epoxy resin Vinylester resin

Debond testingMode I

Debonding of epoxy

specimenMode I (14μm crack film, core thickness 13 mm, Mat CSM)

Debonding of epoxy

specimenMode I (14μm crack film, core thickness 13 mm, Mat CSM)

Debonding of epoxy specimenEffect of core thickness (14μm crack film, core thickness 40 mm, Mat CSM)

Two different modes of failure depending on the specimen position in the panel: interface or wood crack propagation (just

below interface)

Interface crack propagation

Balsa cracks propagation

Debonding of epoxy specimenEffect of crack film thickness (60-120 μm crack films, core thickness 13 mm, Mat CSM)Crack film thickness 60

μmCrack film thickness 120 μm

Mode IIDebond testing

Debonding of epoxy

specimenMode II (14μm crack film, core thickness 13 mm)

No mat CSM

Mat CSM

Mixed Mode

Debond testing

Debonding of epoxy

specimenMixed Mode (14μm crack film, core thickness 13 mm,

Mat CSM)

Vinylester specimenMixed Mode (14 μm crack film, core thickness 13 mm, Mat CSM)

Epoxy – vinylester

Digital image correlation

Strain versus crack path through time0 20 40 60 80 100

Strain extraction along the crack path.

The strain is extracted at different time during

crack loading and propagation:

Pixels

10%

5%

0%

-5%

-10%

Strain εyy

Parameter estimation-Face sheet and core stiffness

-Stress at which the crack propagate in Mode I and II

- G strain release energy rate in Mode I and II

F (N)

d (mm)

σ : Stress to propagate the crack

G

Conclusion•Variability of balsa wood is important

•Thicker core specimens provided less reproducible results and lower GIC.

•Thick crack film led to unsteady crack tip initiation and propagation and should be avoided.

•The presence of mat layer is beneficial.

•Vinylester resin was weaker than epoxy resin in mix mode loading.

•GIC computed by recording of the crack tip location during loading.

•GMMB calculation scatter due to the difficulty to locate the crack.

•Mode II crack location was not detectable to the naked eyes.

• To improve this reading usage of digital image correlation may help to refine the crack location.

Future work

• Validation of the optical crack location method in Mode I and usage in Mode II for GIIC calculation.

• Refinement of materials parameters

• Numerical validation in Mode I, II and mix mode and comparison to experimental results.

Thank you