Von mises for plastics

[pawanhv]

Hi I am currently working on plastic part design (PA6) reinforced by GF 40% . I would like to analyse its structural rigidity in ANSYS . So I would like to know is it appropriate to look at the vonmises stresses. I dont know if i have to consider it as completely isotropic to look at von mises. Or is there any other criterion i am missing for plastic composites?

Meanwhile I am currently lookinf at tensile and compressive stresses alone as a criterion to look for rupture. Will it make sense?

Thanks .

 

[rb1957]

is this chopped fiber ? ie not lamina.

von mises doesn't sounds very appropriate, maybe the randomness of the orientation of the chopped fibers and the (poresumably) uniform distribution through the matrix is sufficient to allow you to treat it as homogenous.

personally, i'd test the part, or at least a coupon.

 

[pawanhv]

hi rb1957
Thanks for your reply. Yes I am currently working on chopped short glass fiber reinforcements.

I completely agree, I was hesitant to use vonmises in the first place.
but what sort of test do you recommend (to check the homogeneity ? )

From what i know chopping of a samlpe injected part in differnt directions wrt its moulding direction and testing it should give an idea is its homogenous ? or do you have anything else in mind ?

I got a lot of questions on this. The more I think the more I am interested

Thanks again

 

[rb1957]

to test homogenity ... mix up a batch, make a panel, representative thickness, cut coupons along both x- and y-axes, pull ...

i'd probably want to test a finished part, depending on the margin and the criticality of the component.

 

[pawanhv]

thanks again, I always wondered, how would you differentiate if the material is anisotropic and not isotropic? I mean what is the difference in tensile curve I am looking at say +/-20% along the curve (or average of curve!)? or is there a specific method for that?

 

[rb1957]

isotropic is an ideal.

20% sounds enough to say it is anisotropic, maybe 10%.
whatever, if i'd taken the time to determine the values, i think i'd use them. mind you, that requires controlling grain direction, so i might model using conservative values. and i might correct the most critical stress to include the effects (if i could, but i haven't thought about it enough).

 

[GregLocock]

I'd cut coupons from a similar part, unless you are very lucky the fibres will orient themselves during manufacture .

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[rb1957]

i thought of that too.

the obvious advantage is your coupon geometry is as close as you're going to get the real manufactured part. the disadvantage (if you want to get analytically anal about it) is you don't know the direction.

i thought manufacturing a plate would give a similar distribution as the manufactured part, and would have the advantage that you could cut orthogonal coupons.

 

[pawanhv]

well thanks for your responses, in case if I recognize it as anisotropic , which criterion would you use to dimension the plastic part ? to see if there is a rupture . Anyone has experience in such sort of analysis? thanks

 

[irq]

von Mises to evaluate fibre reinforced plastics? I think this is not a good idea, since this criterion is mostly applicable to ductile materials and your plastic is more brittle. Max. normal stress seems in this case to be more appropriate. A good approach is to use higher stress of both criterion.

Do you have complex shape of your design? If the answer is yes, you need access to injection molding simulation to see fiber orientation are good positioned in critical regions of your parts or check the weld-lines etc. Without this information the whole game is a guess especially if you want to have weight optimized part! If you have no such information use high MS. Otherwise you can experience a surprise during the testing phase.

The cut from the part is not so easy to perform. Very often the wall has non-constant thickness while many standard test methods require it. I have similar problem right now, see the picture. Do we get correct values if the specimen has different thickness on a whole gauge length (approx. 2:1)? Stress is simply F/A, but I can not imagine that the slope has no influence on the results. What difference would you expect comparing to the correct tension test specimen with constant thickness? Can we anyhow consider this deviation in the test value? Is the test with a slope more conservative? The grinding of the specimen is not an option, as surface of the wall has the best properties and in my opinion it would reduce the test value.

There is also another issue. You get different values if you cut the specimen (surface quality) from the part and in case of casting. Unfortunately I can not remember the recommended reduction factors.

 

[pawanhv]

Thanks irq for your reply. I will try and respond to it when I will be on my PC.
Normally for a plastic part what is the recommended security factor? Like for ex if we are testing for rupture and the limit is 120MPa , if for example we consider a loading case of 1600N and the max stress is 115MPa; what would you do ? look to optimize the part or accept it? I cannot chant plastic material? how much should I try to optimize the part so that the stresses is reduced (20% less stress for ex ? ) I see that a lot of people try to overdo this hence loosing time with no results.