8/3/2019 Efek Ketaksempurnaan Geometrik Pada Renspon Kompresif Dari Panel Komposit_Transportasi Maritim_de_Verdiere_

1/1

Failure criteria such as Tsai-Hill and Tsai-Wu can be applied in Fortran

user subroutine Usermatps.F and compiled in Ansys. Such failure

criteria could not reproduce material non linearties and failure load

predictions well.

An existing damage model in an explicit code was preferred for enhanced load

buckling and damage predictions and ran in PAM-CRASH.

A small amount of defect depth (0.7mm) reduces considerably strength and

stiffness on unidirectional or [0/90] lay ups (more than 20%), but has little

influence on bias direction lay up (less than 8%).

The defect influence is greater on stiff and brittle composites (carbon)

than on more compliant ones (glass).

Results are sensible to the compressive stiffness, strain and strength used which

are difficult to acquire experimentally in the axial direction in the first instance.

An Ansys input file was generated that created various numerical plates withdefects rapidly.

To predict better the compressive response and the defect influence; a damagemodel was used on CFRP and GFRP with the PAM-CRASH code. Implementingsuchmodel in Ansys was rather problematical and was avoided.

Defect influence was considerable on [0/90] lay up and small on [+-45] lay up. It

was also stronger on carbon than glass lay-up.

Small mould defect could reduce strength by 10 to 25% on axial lay up and deeperone up to 40% . It is an issueon long structure subjected to buckling such hasboatmast, wind turbine blades or glider wings. Significant safety factor should be used.

Panels made via resin infusion.

Panel testing:

o The apparatus should allow for regular enlightenment of the panel to permit DIC

algorithm to perform.

o The apparatus should allow for compressive loading of the panel and be

representative of the in service loading as such as the one of the blade spar and skin.

o The panels are compressed by an Instron test machine.

o The apparatus uses anti buckling guides on its side and full clamping at its bottom.

o Defects: bumps in the form of a sinus wave are used (Figure 4).

Monitoring of imperfections

o DIC allows for the capture of the defect changes in 3D through out loading.

Ultimate objective: to asses the effect of

geometrical imperfections on the compressive

strength of carbon and glass composite panels

(CFRP and GFRP).

Currently compressed structures such as wind

turbine blades are produced by resin infusion. As

the blade have important aspect ratio the moulds

can be rather long (Figure 3) and the blade can

then be sensit ive to buckling as compressive

properties of composites are known to be

significantly lower than in tension (15 to 40 %). To

make it worse it is possible to get some bashes on

the mould surface after the production of several

blades due to human error (a tool falling on the

mould surface) or the mould deformation under

heat for example. Such defects on the mould will

be transferred to the blade geometry that would

lead to the reduction of the buckling resistance of

the structure.

This research considers the simulation of

composite panels (GFRP and CRFP) with various

lay-up and geometrical imperfections (size and

depth) to simulate and investigate the reduction in

buckling load and ultimate compressive strength.

Secondly a novel compressive apparatus is being

manufactured to allow for the buckl ing of the

panels and the 3 D monitoring of the deformation

via digital image correlation (DIC).

Effect of geometric imperfections on the compressiveresponse of composite panels

Mathieu Colin de Verdiere mcdv1u07@soton.ac.uk - School of Engineering SciencesCEC Marstruct funding

Supervisors

Dr Steve Boyd and Professor Ajit Shenoi

FSI Away Day 2010

Fluid Structure Interactions

Research Group

Motivation & Aim

Figure 1: Bending of the windturbine blade Source [1]

Generic Panels

Research has designed generic panels to be representative of wind turbine blade.

[0-90] lay-up 210 x 210 mm, 4 mm thick.

Bump depth vary from 0 mm to 10 mm.

Bump diameter vary from 50 to 100mm.

Parametric simulations are ran under Ansys with basic failure criteria andbuckling analysis on glass and carbon fibres.

A few simulations are run under PAM-CRASH with a damage models. The modelis calibrated in thefirst instance and then ran on panels with and without defects.

References:

[1] : Full scale testing of wind turbine blade to failure - flapwise loadingErik R. Jrgensen, Kaj. K. Borum, Malcolm McGugan, Christian L. Thomsen, FindM. Jensen, Christian P. Debel og Bent F. SrensenRis National Laboratory, Roskilde

June 2004[2]: http://www.bayviewedisonindustries.com

Testing

Results

Concept

Figure 2: typical wind turbinecross section [1]

Figure 3: 60 foot wind turbinemould [2]

Simulations

a) b)

c)

Figure 4: geometrical defect

Future work:

Simulation with various thicknesses

Strain field monitoring

Comparison of damage for different defects

Comparison of numerical and experimental results

Conclusion

Figure 5: Tensile, compressive and shear specimens

Figure 6: axial stresses prior to buckling on small tolarge defects (left to right)

Figure 6: a) complete apparatus, b) GFRP specimen, c) GFRP panel infused

Specimens testing

Compressive apparatus

Figure 7: Ansys input model