Meso Scale Modelling of geometry and mechanical properties of an - - PowerPoint PPT Presentation

Guillaume PERIE Stepan LOMOV Ignaas VERPOEST (MTM KULeuven) David MARSAL (SNECMA)

Meso Scale Modelling of geometry and mechanical properties of an interlock reinforced composite

Geometrical and mechanical modelling of 3D Interlock Fabrics

Introduction

07/04/2009 2nd World conference on 3D fabrics 2

• Snecma Fan blade Project
• Carbon / epoxy composite, 3D reinforcement, RTM process
• 3D = Better resistance to impact and fatigue
• The fan blade
• 1m long, complex shape
• Thickness, Vf and Weave pattern change along the length of the blade

Geometrical and mechanical modelling of 3D Interlock Fabrics

Snecma 3D interlock Fabric

07/04/2009 2nd World conference on 3D fabrics 3

• 3D angle Interlock :
• Several weft layers linkes by warp yarns
• Shifted weft layers
• Thick Fabrics
• Parameters of the fabric:
• Weave pattern
• Spacing between yarns
• Type of yarns
• Fiber volume fraction
• Shear angle

Impossible to perform mechanical tests on all configurations Need of a modeling tool WiseTex + TexComp

Geometrical and mechanical modelling of 3D Interlock Fabrics

Objectives :

07/04/2009 2nd World conference on 3D fabrics 4

• Geometrical modelling of Interlock fabrics :
• Modification of WiseTex
• Input data : compression behavior of carbon yarns
• Validation of the models with image analysis on samples
• Mechanical modelling of Interlock fabrics :
• Calculation of static mechanical properties
• producing database of mechanical properties
• Damage modelling ?

Geometrical and mechanical modelling of 3D Interlock Fabrics

WiseTex : Modeling of internal geometry of relaxed & deformed textiles

07/04/2009 2nd World conference on 3D fabrics 5

• Model a RVE (Representative Volume Element) using the minimum energy principle,

calculating the equilibrium of yarn interactions

• Covers a wide range of textile structures (2D, 3D, Braided, Knitted, Laminates, non

crimp)

Geometrical and mechanical modelling of 3D Interlock Fabrics

WiseTex Modifications

1. « Missing Wefts »

• Possibility to remove weft yarns In order to model shifted weft layers

07/04/2009 2nd World conference on 3D fabrics 6

• Modification of the mathematical coding of the weave pattern with

negative values

4 1 2 3 1 2 3 4 layer 1 layer 2 level 0 level 1 level 2

            1 2 1 1 2 1 1 1 2 2 1 1

warp 1 warp 2 warp 3 warp 4

WiseTex WiseTex modified

Geometrical and mechanical modelling of 3D Interlock Fabrics

WiseTex Modifications

• 2. Modification of interaction algorithm between warp and weft

07/04/2009 2nd World conference on 3D fabrics 7

• Different definitions of bending intervals for calculation of bending energy

Interval n°2 Interval n°7 Interval n°3 Interval n°5 Interval n°2 Interval n°7 Interval n°3 Interval n°5

N°1 N°2 N°3

• Interpenetrations between

yarns

• No crimp on weft yarns
• Low Vf
• Interpenetrations between

yarns

• Bad contact zones between

warp and weft

• Good modelisation of undulation
• f yarns for warp and weft
• reduced interpenetration

Geometrical and mechanical modelling of 3D Interlock Fabrics

Input Data : Compression behaviour of carbon yarns

07/04/2009 2nd World conference on 3D fabrics 8

• Critical data for prediction of deformation of the fabric
• Standard tests (Kawabata) not designed for heavy yarns Fmax= 10N and no

measurment of yarn width

• Compression set up developed in LPMT

(Laboratoire de Physique et Mecanique des Textiles, Mulhouse, France)

• First designed for polymer monofilaments
• Fmax= 1.5 kN
• Compression between glass plates = width

measurement

• G. Stamoulis , Ch. Wagner-Kocher and M. Renner

An experimental technique to study the transverse mechanical behaviour of polymer monofilaments,vExperimental Techniques, vol.29, issue 4, 2005

Geometrical and mechanical modelling of 3D Interlock Fabrics

Input Data : Compression behaviour of carbon yarns

07/04/2009 2nd World conference on 3D fabrics 9

• Matlab routine developed to analyse automaticaly pictures and

measurements

• Greyscale to binary image
• Correction of angle of the yarn axis
• Measurment of width on each pixel row and averaging
• Synchronization of the image with force and thickness

Geometrical and mechanical modelling of 3D Interlock Fabrics

Input Data : Compression behaviour of carbon yarns

07/04/2009 2nd World conference on 3D fabrics 10

Force kN 0.01 0.025 0.05 0.1 0.15 0.2 0.3 0.4 0.5 0.6 0.7 0.9 1.1 1.3 1.48 Mean

0.68 0.70 0.72 0.74 0.75 0.76 0.78 0.79 0.80 0.81 0.81 0.82 0.83 0.84 0.85

Std

0.53 0.52 0.50 0.49 0.48 0.47 0.46 0.46 0.46 0.45 0.45 0.45 0.45 0.45 0.45

CV %

15.10 14.39 13.60 12.88 12.47 12.10 11.69 11.42 11.20 11.06 10.92 10.77 10.58 10.47 10.39

Normalized Width measurements made on 20 samples

Force kN 0.01 0.025 0.05 0.1 0.15 0.2 0.3 0.4 0.5 0.6 0.7 0.9 1.1 1.3 1.48 Mean

0.82 0.70 0.62 0.54 0.50 0.47 0.44 0.42 0.40 0.39 0.38 0.36 0.35 0.34 0.33

Std

0.10 0.07 0.06 0.05 0.04 0.04 0.04 0.03 0.03 0.03 0.03 0.03 0.03 0.03 0.03

CV %

10.70 9.48 8.76 8.08 7.83 7.59 7.39 7.27 7.18 7.19 7.11 7.04 7.04 6.99 6.94

Normalized Thickness measurements made on 20 samples

Geometrical and mechanical modelling of 3D Interlock Fabrics

Input Data : Compression behaviour of carbon yarns

07/04/2009 2nd World conference on 3D fabrics 11

• Deviation between samples in the same serie of tests but also

between different series of test :

Slope of the curves is similar for every test but curves look shifted up

• r down which shows a strong dependency on the initial dimensions
• f the yarns
• Initial thickness d10 and width d20 :
• It is not possible to measure accurately d10 and d20, as the yarn begins to

flatten before the measurments starts (when the load cell mesure 0.01 kN

• D10 depends a lot on how the yarn is fixed on the glass plate by the operator

(different tension and torsion of the yarn)

• A proposed solution would be to add tension to the yarn during the

compression test A modification of the set up is needed to introduce tension

Geometrical and mechanical modelling of 3D Interlock Fabrics

20 25 30 35 40 45 50 55 60 65 70

Vf % Pressure

Specimen WiseTex

Input Data : Compression behaviour of carbon yarns

07/04/2009 2nd World conference on 3D fabrics 12

d1 d2

• Average behaviour = WiseTex Input:
• Improvment in the prediction of the fabric

compression

• Overestimation due to interpenetration of

yarns in WiseTex In the fabric widening of one yarns is constrained by the other yarns

• Need to make tests with side boundaries

Yarns compresion Fabric Compresion

Geometrical and mechanical modelling of 3D Interlock Fabrics

Validation of models with image analysis

07/04/2009 2nd World conference on 3D fabrics 13

1. Transverse Yarns

• Filtering on grey scale levels to isolate the yarns and clean the image
• Manual separation of contours in Photoshop
• Measurments made with image analysis softwares

Measurements :

• Spacing between yarns
• Vf inside yarns
• Orientation of cross sections

Geometrical and mechanical modelling of 3D Interlock Fabrics

Validation of models with image analysis

07/04/2009 2nd World conference on 3D fabrics 14

• 2. Longitudinal yarns
• Filtering on grey scale levels to isolate the yarns and clean the image
• Manual separation of contours in Photoshop
• Matlab routines to calculate crimp and orientation of the yarn

In Matlab :

• Isolation of contours and center line for each yarn
• Smoothing of lines using cubic smoothing splines, in order to avoid pixel effect
• Measurement of the length of the spline, and orientation of spline at different

points

Geometrical and mechanical modelling of 3D Interlock Fabrics

Image analysis results

07/04/2009 2nd World conference on 3D fabrics 15

Warp Weft

Sample WiseTex Model Warp d1 1 0.89 Warp d2 1 1.12 Warp Vf % 77.1 70 Warp Crimp % 1.17 1.5 Weft d1 1 0.87 Weft d2 1 1.08 Weft Vf % 69.6 63.9 Weft crimp % 1.01 1.2 FabricThickness 1 1.02

Conclusion:

• Yarns are flatenning too much
• Crimp is overestimated due to local

changes of curvature on the models

• Crimp is not enough accurate :

Need to compare histograms of yarns

• rientation for better characterization
• f the fabric.

Normalized measurments

Geometrical and mechanical modelling of 3D Interlock Fabrics

Calculation of mechanical properties

07/04/2009 2nd World conference on 3D fabrics 16

• Tex comp routines are used to calculate homogenized mechanical

properties of the final composites:

• Inputs : WiseTex model for reinforcement and matrix properties
• Calculation of mechanical properties by method of inclusions with different

calculation schemes implemented (Mori-Tanaka, Iso-Strain..)

• Results are compared with measurements made by UTC

(Université de Technologie de Compiegne)

• Implementation of a tool to produce databases of material
• Unique interface linking WiseTex and TexComp
• Automatic Production of several WiseTex models based on a range of

weaving parameters specified by the user and calculation of mechanical properties with export to excel file.

Geometrical and mechanical modelling of 3D Interlock Fabrics

Iso-Strain MoriTanaka Experiment

Results

07/04/2009 2nd World conference on 3D fabrics 17

• E11 and E22 measured on tensile tests

in warp and weft directions

• E33 measured on samples cut in the

thickness direction

0.00 0.20 0.40 0.60 0.80 1.00 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00 Normalized Pick spacing E1 58% 0.00 0.20 0.40 0.60 0.80 1.00 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00 Normalized Pick spacing E2 58% 0.00 0.20 0.40 0.60 0.80 1.00 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00 Normalized Pick spacing E3 58%

Geometrical and mechanical modelling of 3D Interlock Fabrics

Iso-Strain MoriTanaka Experiment

Results

07/04/2009 2nd World conference on 3D fabrics 18

• G12 measured on tensile tests in bias

direction

• Techniques are developed in UTC for

measurment of G13 and G23 based on torsion tests and strain mapping on bending tests

0.00 0.20 0.40 0.60 0.80 1.00 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00 Normalized Pick spacing G12 58% 0.00 0.20 0.40 0.60 0.80 1.00 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00 Normalized Pick spacing G13 58% 0.00 0.20 0.40 0.60 0.80 1.00 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00 Normalized Pick spacing G23 58%

Geometrical and mechanical modelling of 3D Interlock Fabrics

Errors on E1

07/04/2009 2nd World conference on 3D fabrics 19

• There are still some significant errors in E1 results (up to 15%)
• These errors can be explained by the errors on fiber orientation in WiseTex

Models

• For small Pick spacing:

in WiseTex crimp is overestimated and we can

• bserve too important changes of curvature in the

yarn path which can explain the low E1 calculated

• For high Pick spacing:

The higher pressure applied on this material makes the yarn expand between weft rows, leading to a local disorientation of fibers. This is not modeled in WiseTex, and can explain that wiseTex values are higher

Geometrical and mechanical modelling of 3D Interlock Fabrics

Errors on E1

07/04/2009 2nd World conference on 3D fabrics 20

• Proof of the effect of crimp on E1 :

We can artificially reduce crimp by reducing the number of points between contacts, which leads to straight warp yarns (but it introduces interpenetration between warp and weft) You can see on the graph that reducing crimp leads to an increase of E1

0.20 0.40 0.60 0.80 1.00 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00 Normalized Pick spacing E1 58% Crimped Yarns Experiment Straight Yarns

« Straight » yarns « crimped » yarns

A solution to decrease crimp without introducing interpenetration, would be to implement tilt of the weft cross sections to fit the interlock angle.

Geometrical and mechanical modelling of 3D Interlock Fabrics

Conclusion

07/04/2009 2nd World conference on 3D fabrics 21

• Prediction of mechanical properties is quiet good:
• Most of the properties calculated are close to the mechanical tests (in the

standard deviation range)

• The errors can be linked to the errors in modelling the geometry of the

reinforcement

• Further Work:
• Improve compression test to reduce deviation, tests with side boundaries

should also help to get more accurate modeling of the yarn geometry

• Implement tilt of the weft cross section to get a more accurate modelling of

warp path (we can see on pictures that weft yarn tilt to fit the interlock angle)

Geometrical and mechanical modelling of 3D Interlock Fabrics

07/04/2009 2nd World conference on 3D fabrics 22

Thank you for your attention