Compression only material

mathiasf

Hi Victor,

I use Mecway to study truss structures (static analysis mainly), with a particular interest on the joints. I post-process the results given by Mecway (N, V, M of beams) and could not do the same with ccx results.

There is a tension-only material for bar elements in mecway, but no compression-only material. Do you plan to add it to Mecway (internal solver) in a foreseeable future ? Or is there a reason why it is not implemented ? Or a "trick" to transform tension-only truss elements to compression-only elements ?

Victor

There's no technical reason why it couldn't be there.

You can arrange beams to connect to the opposite ends of the tension-only member though it's a bit fiddly. Here's an example that also uses an artificial beam with extreme section properties to couple rotations while releasing longitudinal displacement.

I think another way would be to couple your structure to the opposite ends of the tension-only element using constraint equations.

mathiasf

Hi Victor,

So could compression only elements be added in a future version of Mecway ?

I cannot think of a way to make your present solution to work for me (compression only via tension only elements), but maybe I did not understood it fully. Here is an example of a simple structure.



My goal (in this example) is to have the diagonal elements to behave in compression only. I use short bar elements to control the stiffness of the connections, and also to retrieve the forces at the connections (for postprocessing). On the left there is the full structure. On the right there is the structure in which I removed the diagonal element in tension.

If I could double the connection bar elements, one with elastic behaviour and A~0, and the other with compression only behaviour, I would get more or less what I am aiming for. Am I missing something ? Is there a way to make it work with tension only elements ?

Mishal

Hi Guys

Not sure if this possible as I have not used contact that much on Mecway.
Would it be possible to apply contact at the joints of the truss members, this will allow for only compression to transfer. I have done a similar thing on Strand7 when checking how the loads will redistribute on conveyors when a leg lifts of the ground during high wind loads.

mathiasf

Hi Mishal,

Thank you for your answer, but it does not apply to this case use.

Victor

Hello @mathiasf. I've put my compression-only structure into your model. Adding Flexible joint on beam to both ends makes the whole little assembly behave like a single compression-only truss element.

The spacing between the two middle nodes would lead to error in the compression stiffness, so move them as close together as needed.

The deformed view shows it opening up like a Z-shape but I'm not sure if that's something wrong or an artifact of the scaling. The displacement of everything else isn't quite the same as the structure on the right and I'm not sure why.


I'm not sure about having compression-only truss elements in the next version.

mathiasf

Hi Victor,
Thank you for your example, it is what I was hoping for.

cwharpe

Just a follow-up question whether it is yet plausible to develop a two-pt. Compression_Only spring as a simpler element vs. the "zig-zag" work around shown herein using tension-only slider element, etc?

Since I often use shells in my modeling, I've been using the zig-zags to compensate for the shell element not being able to model the material compression normal to plane. Setups are with zig-zags between spiders (in parallel with pre-tension bolts), and are used when modeling joint prying forces. Also used when examining multiple pre-loaded bolted joints over a large surface for probable pressure leaks (when material compression reaches zero even though bolt retains some tension). I have found the method a bit "fiddly" as was mentioned, and I'm always wary of spurious displacements through the zig-zag lateral to bolt axis.

I'm naive to the coding requirements, but I'm wishing whatever process "turns off" a Tension-Only spring when it goes into compression might be employed in reverse to turn-off a Compression-Only spring when it goes into tension??

Victor

I agree and intend to include compression-only 2-node elements in a future version. Probably with a spring stiffness (k) rather than E and A the way tension-only currently is.

cwharpe

That would be very

disla

Hi,

Isn't tension and only compression materials nonlinear behaviors?.
Is there a reason why the problem is not approach with ccx solver?
Is it because there is not such a variety of beam sections.?

disla

This is the truss system with ccx and only compresion beams in the diagonals.
It is nonlinear (I think it must be nonlinear for this kind of material)

Final results are:

(+ = tension)
Side in Tension. Final Tension 8.999927E+01 N
Side in Compresion :Final Compression -2.305519E+05 N

The maximum tensile stress the user is willing to allow is controled by the second parameter in the Custom card :

*MATERIAL,NAME=COMPRESSION_ONLY_1
*USER MATERIAL,CONSTANTS=2
110E9,1E3 <------

A minimum value is needed to help convergence.
This kind of material behaviour has an unspected issue. As it can only sustain compresion, it colapse under self weight.

cwharpe

@disla : Thanks for the example. I'll still hope for the Comp.Only spring to be added to the element arsenal -- for internal solver work as well.