Composites analysis and modelling

mike262

Hi all im looking for a good guide and some info on the best way to model composite laminates using Mecway. i'm analysing a composite wing for a light aircraft that uses fairly simple layups with unidirectional spar caps and bidirectional +- 45 degree layups for wing ribs and shear webs. im also keen to find out the best way to model cored panels made of nomex honey comb and and carbon face sheets
Cheers
Michael

prop_design

composite support in mecway are pretty basic at the moment. you have to define your own material properties and failure criteria. you can't really view stress too well for laminates. the models get very large as well, due to the ccx expanded shell method. i did mess around with laminates some but ended up using a different method. so i can't give you too much help. i'd suggest just playing around with something simple first and see if you really want to go down that road. there are a lot of products out there specifically geared towards composite analysis, that would make it a lot easier on you.

mike262

i have done a fair bit of spread sheet analysis and have a fair bit of data on material properties for the composites i intend to use....and one text book i have on composite light aircraft suggests using the tsai wu failure criterion in spread sheet analysis. I have done fea in the past professionally on metallic components using nastran. (i dont have the budget for a Nastran licence!)......there must be some way to get some info on cored panels and such. i guess many of the composite light aircraft designed in the 80 and 90's before widespread availability of pc based fea were done fairly crudely using a "black metal" approach. I prefer FEA to spread sheet analysis for work flow reasons and because you produce reports that are easier to read than with just spread sheets alone.

prop_design

hi mike,

yeah i know what you mean. i'm in the same boat as well. i used algor way back in the day. nastran then ansys. the licensing fees are nuts. mecway is great. you can do composites. it's just not it's strong suit. compared to a spreadsheet it would be way better. since you have experience with them, i think you have a good chance at success. i don't have that much experience with them and i still managed to get what i wanted. it just takes awhile at first. i would still suggest a small model or two before you try a whole wing. i'm not sure if you have used mecway yet but you might have a bit of learning to do in general and then there is the composite aspect. there are a few ways you can go about it. you can use multiple parts and contact. you can use the laminate feature. lastly, you might be able to do something like what i did and got around having to use the laminate feature. i went back and watched the fea videos i did yesterday. unfortunately, there is lots of things spread out over all the videos. so it's hard to say any particular one. in the advanced composites i forgot some things and had to go back with the following two videos to cover those aspects.

you can define tsai-wu or any failure criteria you want. that is one of the great things about mecway. the built in tsai-wu is only for the mecway solver. unfortunately the mecway shell element is very limited so you probably won't use that at all. if you use the ccx shell or laminate, they turn into solids. you have to define your own criteria with those. with either solver you have to define your own material properties.

prop_design

mostly the challenges i had was figuring out how to do the laminate orientation, model composite fillets, and learning mecway/ccx. luckily i only work with single parts most of the time. so i didn't also have to setup contact etc. i have a laptop with 8gb of memory and the ccx solver that comes with mecway gets limited fast. in the propeller hub thread we discovered you can use intel mkl pardiso solver. it makes the models run way faster and with less memory. so that will help you out a lot.

mike262

Here is a composite mesh of a aircraft structural wing box i have made up. It has unidirectional fibers of varying thickness along the spar caps and BI directional carbon fibre for the skins and ribs. I have run it and it compares well with other analysis i have done and data i have. I would like to see what the effect of using carbon fibre/Honeycomb sandwich panels in the wing skins and ribs would be. I would like to reduce the number ribs too. How can i model sandwich panels in Mecway? . Im mainly interested in capturing the stiffness increase they provide. I have done some spread sheet analysis to work out the basic properties of simple sandwich panels but i am unsure how to put them in to MECWAY. I am not so much interested in accurately modelling failure modes in sandwich panels....there are other ways to work that out i just need a reasonable idea of the stiffness they provide to the wing.
thanks
michael

Victor

You can define a sandwich panel in the same way as a laminate, but with one thick layer for the core. The internal solver might not be very accurate though because the shell elements have:

• No elasticity in the thickness direction.
  
• A single shear angle for all layers. I remember finding poor accuracy for sandwich materials in shear that I attributed to this, but I'm not sure of the details. It might be worth checking with a simple beam first to find its limitations. Take extra care with rotational boundary conditions because of the low core shear stiffness.
  

mike262

what typical values would you use for a core material? below is the laminate entry i have tried unsuccessfully ....The face sheets are Kevlar with thickness 0f 0.03in E of 3200000 psi, poissons ratio of 0.1 and an in - plane shear modulus of 120000psi.
the core material is .25 in thick honeycomb and only its shear modulus of 5800psi and its shear strength is given in the material spec sheets....The values i entered for E and Poissons ratio for the core material were just guesses to see if it would work. is this the correct way to enter in laminate details?



0.03, 0, 3200000, 3200000, 0.1, 120000, 0, 0
0.25, 90, 100, 100, 0.1, 5800, 5800, 0
0.03, 0, 3200000, 3200000, 0.1, 120000, 0, 0

mike262

I have been getting pretty good results so far just need to stiffen up the the wing skins!

mike262

Is there some way to turn the skins off in the results so i can see whats happening internally to the wing?

Sergio

Nice model @mike262 ! In the results tree there is a branch called Components where you can hide/unhide it.

VMH

mike262, you can only hide individual if they are individual. If the outer layer and framing are the same component then you can hide the outer and see the framing.