P eculiarities of damage behaviour of N C F carbon/epoxy - - PowerPoint PPT Presentation

downloaded from: http://www.mtm.kuleuven.ac.be/Research/C2/poly/index.htm 1

P eculiarities of damage behaviour

• f N

C F carbon/epoxy laminates under tension

Stepan V. Lomov, Dmitry S. Ivanov, Katleen Vallons, Ignaas Verpoest, Thanh Chi Truong Department MTM, Katholieke Universiteit Leuven

downloaded from: http://www.mtm.kuleuven.ac.be/Research/C2/poly/index.htm 2

C

• ntents

1.

• Introduction. Features of the damage in NCF composites

2. Premature fibre damage, as discovered by Mattsson et al – “Swedish effect” 3. “Is it a bug or is it a feature?” – new experimental observations 4. Challenge for the modelling work

downloaded from: http://www.mtm.kuleuven.ac.be/Research/C2/poly/index.htm 3

1. Introduction

• Non-crimp fabrics (NCF)
• Internal structure on NCF composites
• Damage in NCF composites

2. Premature fibre damage, as discovered by Mattsson et al – “Swedish effect” 3. “Is it a bug or is it a feature?” – new experimental observations 4. Challenge for the modelling work

downloaded from: http://www.mtm.kuleuven.ac.be/Research/C2/poly/index.htm 4

N

• n-crimp fabrics = Multi-axial multi-ply fabrics

Source: www.liba.de

Weft insertion Warp knitting unit

downloaded from: http://www.mtm.kuleuven.ac.be/Research/C2/poly/index.htm 5

A pplications

Floor pan TECABS project Rear pressure bulkhead, AIRBUS

Sources: www.tecabs.org; AIRBUS

downloaded from: http://www.mtm.kuleuven.ac.be/Research/C2/poly/index.htm 6

E xamples of N C F

B1 B2 Q

downloaded from: http://www.mtm.kuleuven.ac.be/Research/C2/poly/index.htm 7

Internal structure of N C F composite 90 90

Cross-ply laminate NCF Stitching yarn Gap Inner opening Surface channel

downloaded from: http://www.mtm.kuleuven.ac.be/Research/C2/poly/index.htm 8

S tudying the damage initiation and development in tension

Acoustic emission sensor Transverse extensometer Longitudinal extensometer Longitudinal strain gauge Q 2

• M

D 2 1 2 3 4 5 6 7 8 .0 0 .2 0 .4 0 .6 0 .8 1 .0 1 .2 1 .4 1 .6 1 .8 2 .0 S tra in , % Event 1 2 3 4 5 6 Strees, MPa e v e n t s tre s s

• s

tra in

Transition point 1: strain at which initial damage is expected Impregnation

• f specimens

before tensile loading Tensile loading on specimens up to 1 2 3 X-ray inspection

• n specimens

before tensile loading X-ray inspection

• n specimens

after tensile loading Comparison of X-ray images for damage evaluation 1 2 3

downloaded from: http://www.mtm.kuleuven.ac.be/Research/C2/poly/index.htm 9

D amage development –loading in fibre direction

downloaded from: http://www.mtm.kuleuven.ac.be/Research/C2/poly/index.htm 10

G eneral features of damage in N C F composites –tension

1. Mechanical properties of carbon/epoxy NCF composites are close to the mechanical properties of UD laminates. 2. Damage in carbon/epoxy NCF composites is initiated very early: 0.3% strain for fiber direction loading, 1.3% for bias loading. These values are twice as low as the damage initiation threshold for UD laminates. Damage sites coincide with the stitching positions. 3. For loading in fibre direction the damage develops by multiplication of transversal cracks in 90°plies. 0°plies stay intack and the stiffness of the sample does not change. 4. Breackage of the longitudinal fibres happens immediately before the sample failure

downloaded from: http://www.mtm.kuleuven.ac.be/Research/C2/poly/index.htm 11

1. Introduction 2. Premature fibre damage, as discovered by Mattsson et al – “Swedish effect” 3. “Is it a bug or is it a feature?” – new experimental observations 4. Challenge for the modelling work David Mattsson, Roberts Joffe and Janis Varna, Damage in NCF composites under tension, Engineering Fracture Mechanics, 2007, in print

downloaded from: http://www.mtm.kuleuven.ac.be/Research/C2/poly/index.htm 12

T wo stacking sequences

NCF NCF NCF NCF

strain, % stress 0.8 1.6 B A

Vf=60%

downloaded from: http://www.mtm.kuleuven.ac.be/Research/C2/poly/index.htm 13

D amage in lay-up T ype B

stress B 1.6 strain, %

0.91%

downloaded from: http://www.mtm.kuleuven.ac.be/Research/C2/poly/index.htm 14

D amage in lay-up type A

strain, % stress 0.8 1.6 A 0.66% 0.87%

downloaded from: http://www.mtm.kuleuven.ac.be/Research/C2/poly/index.htm 15

E xplanation by Mattssonet al

“The main reason is the larger number of inside 0°

• bundle breaks in laminates

where imperfect bundles of this orientation are supported by 90°

• bundles

compared to cases where two 0°

• bundles from different layers are together. The

mechanics is not clear at present but it may be related to larger bending resistance of two imperfect bundles as compared with one. The difference is larger if these bundles would be delaminated which may be initiated by transverse cracks in 90°

• bundles.

Another possible reason for larger stiffness reduction in [0/90/0/90]S composite is, according to FE calculations, directly related to the larger effect of each surface 0°

• bundle break on the composite stiffness: due to less constraint from

the surrounding material the opening of surface bundle breaks is much larger”

<

downloaded from: http://www.mtm.kuleuven.ac.be/Research/C2/poly/index.htm 16

1. Introduction 2. Premature fibre damage, as discovered by Mattsson et al – “Swedish effect” 3. “Is it a bug or is it a feature?” – new experimental observations 4. Possible explanations 5. Challenge for the modelling work

downloaded from: http://www.mtm.kuleuven.ac.be/Research/C2/poly/index.htm 17

O ld results: 0° /90° , Vf = 45%

Truong Chi, T., M. Vettori, S.V. Lomov and I. Verpoest Carbon composites based on multiaxial multiply stitched preforms. Part 4: Mechanical properties of composites and damage observation Composites part A 36 2005 1207-1221

16 Plies in the laminate 45 Vf, % (0/90,90/0,0/90,90/0)s (90/0,0/90,90/0,0/90)s Stacking sequence MD CD 8 Fabric layers in the laminate 329 Fabric areal density, g/sq.m

downloaded from: http://www.mtm.kuleuven.ac.be/Research/C2/poly/index.htm 18

N

• “S

wedish effect”…

40.80.9 0.050.01 1.590.04 65432 43.61.9 CD 39.51.3 0.060.01 1.550.05 65936 45.51.1 MD B2 0°/90° Vf (%) Poisson ratio Ult.strain (%) Strength (MPa) Modulus (GPa) Test direction Material

CD = A ? MD = B ?

downloaded from: http://www.mtm.kuleuven.ac.be/Research/C2/poly/index.htm 19

N ew results. Vf = 56%

Toughtened Matrix 8 Plies in the laminate 56 Vf, % [+45/-45,+45/-45,-45/+45;-45/+45] Stacking sequence MD 4 Fabric layers in the laminate 540 Fabric areal density, g/sq.m Tests in BD+ and BD- directions BD- BD+ BD + BD- = B = A Stacking sequence the same as with Mattsson et al

downloaded from: http://www.mtm.kuleuven.ac.be/Research/C2/poly/index.htm 20

F irst series –some effect is present

0.36 0.03 72 2 1.57 0.05 1115 55

• 45 (BD-)

0.40 0.05 68 6 1.5 0.3 840 77 +45 (BD+) Damage initiation strain (%) E modulus (GPa) Failure strain (%) Tensile strength (MPa) Test direction Six tests each direction BD + BD-

downloaded from: http://www.mtm.kuleuven.ac.be/Research/C2/poly/index.htm 21

S plitting of the outer plies in BD + samples

BD + BD- Images taken just before final failure of the samples

downloaded from: http://www.mtm.kuleuven.ac.be/Research/C2/poly/index.htm 22

D amage development

BD + + BD

• 0.3%

0.5% 0.9% 1.3% BD+ BD- No difference in damage development; No major delaminations / fibre breakage

downloaded from: http://www.mtm.kuleuven.ac.be/Research/C2/poly/index.htm 23

A coustic emission: cumulative event energy

50000 100000 150000 200000 250000 300000 0.2 0.4 0.6 0.8 1 1.2 strain, % cumulative AE event en

BD + + BD

• AE sensors

removed

downloaded from: http://www.mtm.kuleuven.ac.be/Research/C2/poly/index.htm 24

S ummary first series

1. Difference in the composite strength for two stacking sequences: same as in the “Swedish effect”, albeit less pronounced 2. Difference in the damage initiation strain: opposite 3. No evidence of the premature fibre damage 4. Splitting of the plies on the sample surface (not observed by Mattsson et al). Probably, we deal here with different phenomena…

downloaded from: http://www.mtm.kuleuven.ac.be/Research/C2/poly/index.htm 25

S econd series: almost no effect… BD+

Stiffness (GPa) Strength (MPa) Strain to failure 1 70 1037 0,013 2 64 969 / 3 65 994 0,015 4 66 1013 0,016 5 68 985 0,015 6 65 998 0,016 7 73 958 0,013 8 69 979 0,015 average 68 992 0,015 stdev 3 25 0,001

BD-

Stiffness (GPa) Strength (MPa) Strain to failure 1 / 1010 / 2 65 970 0,015 3 67 1032 0,015 4 67 1029 0,015 5 67 1058 0,016 6 65 1091 0,017 average 66 1036 0,016 stdev 1 41 0,001

ST R EN G T H

300 600 900 1200 Strength (MPa BD+ BD-

ST R AIN T O FAILU R E

0.005 0.01 0.015 0.02 Strength (MPa BD+ BD-

BD + + BD

downloaded from: http://www.mtm.kuleuven.ac.be/Research/C2/poly/index.htm 26

… also some disturbances for BD + at the latter stage

200 400 600 800 1000 1200 0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80

BD

• BD

+ +

downloaded from: http://www.mtm.kuleuven.ac.be/Research/C2/poly/index.htm 27

1. Introduction 2. Premature fibre damage, as discovered by Mattsson et al – “Swedish effect” 3. “Is it a bug or is it a feature?” – new experimental observations 4. Challenge for the modelling work

downloaded from: http://www.mtm.kuleuven.ac.be/Research/C2/poly/index.htm 28

S ummary

1. The effect observed by Mattsson et al in carbon NCF/epoxy 0/90 composites constitutes the premature fibre damage for a stacking sequence:

• with 90°plies concentrated in the centre of the laminate
• with 0°plies lying on the surface of the laminate

(two different interpretations of the same lay-ups) 2. Using same material system “carbon NCF/epoxy”, but different fabrics and different resin, we did not observe the influence of the stacking sequence for thicker plates with more plies, having more double 0°plies. Also fibre volume fraction for these tests is less (about 45%) 3. Using same material system “carbon NCF/epoxy”, but different fabrics and different resin, and the same fibre volume fraction and stacking sequences, we have observed the difference in strength for different types of lay-ups, less pronounced then in Sweedish tests, but with the same sign. However, the mechanism of the change of strength is different and is associated with the splitting of the fibres in outer layers.

downloaded from: http://www.mtm.kuleuven.ac.be/Research/C2/poly/index.htm 29

C hallenge

The effects of premature damage, both “Swedish” and “Belgian”, could be qualitatively understood, but are not predicted by the current damage models of NCF composites. Such a prediction would be of great value, as it would help to establish safer design boundaries for these materials, which are used in load-carrying automotive and aircraft parts.