DAMAGE INITIATION AND DEVELOPMENT IN CARBON-EPOXY TRIAXIAL BRAIDED - - PowerPoint PPT Presentation

DAMAGE INITIATION AND DEVELOPMENT IN CARBON-EPOXY TRIAXIAL BRAIDED COMPOSITE AND FINE STRUCTURE OF DAMAGE

• D. Ivanov, S.V. Lomov, I. Verpoest, F.Baudry, H.Xie

Department MTM, Katholieke Universiteit Leuven, Belgium

Cont ent

Geometry of the composite Tensile test program Strain mapping measurements Damage stages Crack geometry Conclusions

Geom et ry of braided t ex t ile c om posit e: in-plane

MD BD BD CD Experiments on the composite:

Tensile test in 3 directions (MD, BD, CD): accompanied by: examined posterior by: Strain mapping measurements X-ray Acoustic emission measurements Cross-sectioning

Carbon-epoxy composite, 4 layers

14.4 mm

Int ernal geom et ry of t he c om posit e

20 mm Inlay yarns Layer borders Local fibre distribution at the yarn edges;

Int ernal geom et ry of t he c om posit e

• 50 -40 -30 -20 -10 0 10 20 30 40 50

1.05 1.00 0.95 0.90 0.85

I II III IV V VI VII

71 63 55 47

Four inlay yarns in different layers

%

Position across the yarn, % Location of maximum fibre volume fraction in inlay

Normalised and symmetrised fibre volume fraction across inlay yarn (obtained with V.Koissin) Fibre volume fraction in the inlay yarns along their path Position of the measurements of fibre volume fraction along the path;

Int ernal geom et ry of t he c om posit e

Cross section of the 4-layer triaxial braided composite in the machine direction.

Maximum thickness of the layers up to 44% of the composite thickness. Minimum thickness of the layers - 18% of the composite thickness).

1. Surface fields are influenced by inner layers 2. Strong nesting and “interpenetration” of the layers may blur the structural features of the surface field

Cont ent

Geometry of the composite Tensile test program Strain mapping measurements Damage stages Crack geometry Conclusions

Ex perim ent

Specimens: Width: 40 mm (2 unit cell) Thickness: 3 mm Total length: 250 mm Strain mapping (Aramis 4.7): Facet size: 1 mm Facet step: 1 mm Inspection area: 40 mm (width direction), 55 mm (length direction) Load step size: 0.033%

• f applied strain

Acoustic emission: Distance between the sensors: 110 mm Loading: Speed: 1 mm/min

Loading in different m at erial direc t ion

Tensile diagrams Acoustic emission diagrams

MD BD BD CD

1.11 0.03 1.25 0.18 1.45 0.15

0.33 0.07 0.43 0.36 0.05 0.45 0.29 0.09 0.55 1 ,% 2 ,% 15.90.7 0.390.03 36.81.8 0.070.02 32.61.1 0.730.06 E, Gpa

• CD

BD MD

1 - damage initiation; 2 – AE intense grow;

critical 1 ,%

Cont ent

Geometry of the composite Tensile test program Strain mapping measurements Damage stages Crack geometry Conclusions

St rain m apping m easurem ent s

Proportionality of strain fields on each load step to the applied strain. Periodicity: expected at least in the direction of loading. Smoothness: A thin layer of matrix covers all the surface of the composite => no inner interface boundary on the interface. Natural requirements to the results:

< MD >, % MD, %

• < MD > =0.063%

Strain history of a point: Strain field MD:

St rain fields at different load st eps

0.189 0.063 0.122 ,%

• Local strain fields along a line in the direction of loading

(MD).

Load steps: 1 - at 0.063%; 2- at 0.122%; 3- at 0.189%;

1 2 3

• ,%

F iltered fields along a line

Initial deformation field (resolution 3mm);

Filt ered fields along a line

Period of the structure

MD, % MD, % MD, % x, mm x, mm x, mm

Probable position of the line Filtered (least square regression, resolution 3mm); Filtered (least square regression, resolution 5 mm);

Init ial and proc essed fields

Normalised strain fields MD in the direction of loading (MD) Initial field at applied strain 0.1%; Filtered field (resolution 3 mm); Filtered field (resolution 5 mm);

0.524 0.162 Finite element analysis 0.491 0.142 Filtering and area smoothing 0.509 0.052 Filtering by load steps 1.836

• 0.877

Initial experimental field max, % min, %

Max im um and m inim um values of st rain field

Fourier analysis of filt ered and sm oot hed st rain

Frequency of the signal Amplitude, mm Period of the structure Misalignment of textile orientation with loading direction causes an increase of apparent unit cell size

20 mm 20 mm

Cont ent

Geometry of the composite Tensile test program Strain mapping measurements Damage stages Crack geometry Conclusions

St at ist ic al feat ures of t he noise

< >, %

• Strain history MD of a point:

1

• Relation of error

standard deviation to average strain

Relative error for 16 points in square grid: The distribution of noise is normal and has the same standard deviation for all the load steps Noise local strain deviation from its linear regression Few locations in structure are characterised with “abnormal behaviour”

Dam age init iat ion and developm ent in t ex t ile c om posit es

1 2 Loading direction

BD CD MD 0.3 %

1

• Cumulative energy
• f acoustic emission

Relative error for 16 points in square grid Cumulative event curve

• f acoustic emission

Stress

• Relation of error

standard deviation to average strain

1

Dam age m ec hanism in t riax ial braided c om posit e

1 - damage initiation: transverse cracks (inter-fibre failure) 2 - damage propagation: inter-yarn delamination critical - damage propagation: fibre failure

Microscopy X-Ray

• btained with A.Salehi

Cont ent

• Geometry of the composite
• Tensile test program
• Strain mapping measurements
• Damage stages
• Crack geometry
• Conclusions

Crac k dist ribut ion: BD

Sample loaded in the bias (BD) direction

Line of the cross-section Cross-section of the composite X-ray image Cracks: in the yarns of upper and middle layers

Inlay yarns Braiding yarns

2

Crac k dist ribut ion: MD

X-ray image Cross-section of the composite Transverse cracks

Sample loaded in the machine (MD) direction

2

Crac k dist ribut ion: CD

X-ray images Cross-section of the composite Transverse cracks

Inlay yarns Braiding yarns

Sample loaded in the cross (CD) direction

2

Crac k densit y at t he different load st ages

1 2

0.29 0.33 0.37 0.41 0.45 0.49 0.53 0.57

%

Crack density, mm/dm2

1600 1200 800 400

CD MD BD

Conc lusions

AE initiation corresponds to occurrence of first cracks High energy of AE signal = multiple cracks or/and delamination Transverse cracks cross all the thickness of yarn with no exceptions A difference of damage accumulation braided composite with the process in laminates or non-crimp composite - finite length of transverse cracks At a first stage of damage accumulation: periodic increase of a crack pattern without crack length grow Maximum crack density is observed in a specimen subjected to macroscopic shear (braiding direction)

Conc lusions

No extensive micro cracking is observed Transverse cracks are always co-oriented to fibre direction An angle of the crack is 90 15

Requirem ent s for dam age m odelling

Geometry of the composite: yarn architecture yarn curvature fibre volume fraction in yarns Criterion for damage initiation Criterion for damage stoppage Criterion for inter-yarn delamination Criterion for delamination stoppage Criterion of fibre failure Engineering application Main criterion for the composite failure

FE c alc ulat ions

Loading in MD direction up to the critical strain MD =0.3%, <CD>= -0.02%. Stress index: (a) longitudinal L; (b) transverse T; (c) shear LT. Hoffman criterion predicts crack occurrence at MD =0.2%

1 Unit cell 2 Half of the UC: 3 Quarter of the UC Rotation around the x1 axis by Rotation around the x3 axis by

Puc k c rit erion for braided c om posit e

1 - damage initiation: transverse cracks (inter-fibre failure) crack

• /2

1

• failure angle

Front view Back view Stress exposure factor

Prediction of transverse failure of composite loaded in machine direction (MD) up to 0.35% of applied strain

Conc lusions

• Geometry of the braided composite is characterised: in-

plane symmetry, small nesting, big variation of layer thickness and local fibre volume fraction orientation are

• bserved.
• Tensile test is held accompanied by surface strain and

acoustic emission measurements.

• Noise filtering and area smoothing have been proposed to

process strain mapping field.

• Fourier analysis allows to identify periodicity of the strain

fields

• Statistics analysis of noise allowed identifying the damage

initiation threshold, which confirmed acoustic emission measurements and point out the location of damage.

• The obtained data serves a good benchmarking for FE

modelling of this textile