Real case of a C-Hook and considerations about Mecway

Andrea

I work with Mecway from 3/4 months.
I had a good impression on Mecway. I think that is well structured and now I can create models for CCX in a quick way.
For example I verified a 20 metric tons C-hook in half hour with two models: isolated hook and considering contact between the hook and the plate. (the plate is very rigid)

For the first analysis I used some functions not supported by Mecway. For example the supposed contact surface is restrained in the same direction of the lifting force which is tilted of 60° from the horizontal. To do this is necessary to transform coordinate system and apply this transformation to the relative nodeset. But I know that the interface to CCX is quite young and will be developed.

I appreciate your job

Victor

Thanks for the comments.

I don't intend to ever include node coordinate transformation as a visible feature in Mecway. It's not a physically real concept and there should always be an more conceptually simple way of achieving the same result. Mecway generates node transformations automatically where they're needed on constraints and corrects for their side effects, such as changing the directions of loads.

For constraints, you can specify arbitrary directions at constraints themselves. I'm not sure how you did it here (contact?), but if there wasn't a way to specify the direction without changing the node coordinate system, I'd like to find out so I can put something in.

Andrea

I used transform in the isolated hook. So I have the node reactions in the same directions of the force.

For example how you can apply a slider restraint to a tilted surface?
I do that transforming the local coordinates.
There is a simple way?

Victor

Ah, reaction forces are something that needs a way to see the results in an appropriate coordinate system. That would be a postprocessing option though.

For a sliding surface in any orientation, you can use frictionless support. If it needs to only slide along one of the in-plane directions, you can put another frictionless support on another surface that's also parallel to that sliding direction. Much like how a real slider can be made.

If you need to constrain displacement in a direction that's not normal to a surface, you can use a displacement constraint and specify the direction as 3 vector components in global coordinates.

In both cases, it generates whatever *TRANSFORM cards are needed to satisfy the constraints.

Andrea

I tried this kind of support (frictionless).
It tranforms every nodes of a node set into a rectangular coordinates, also cylindrical surfaces.
In a general case of non regular surface is usefull
But for planar or cylindrical surface I don't like this, but is only my opinion.

Victor

I see. Yes, it can create ugly transformations like rectangular when cylindrical feels more natural but hopefully you don't look at them :P My intention with the design of it is that you should never have to think about what coordinate system the nodes are using - or even know that there's such a thing as rotated node coordinates. With the internal solver, those details are completely hidden so it's a non issue. Doesn't help getting reaction forces transformed though.

Sergio

And what about using TRANSFORM card to apply radial deformations on cilindrical node sets? I found usefull that card in that situation, don´t know how to apply such a condition without TRANSFORM to circular coordinate system the node set

Regards!

Victor

Frictionless support now has a displacement value you can specify, which will be the radial displacement if you use it on a cylindrical surface.

Sergio

Great! Have tested and works very well, for doing it by means of CCX cards I have to write and think and check the references manual as is not everyday used, but now is so easy. The only thing now is the name, I meant is not doing a support but applying a displacement, maybe could have a more significative name.

Regards, and thanks.

Andrea

It runs but look the the difference plotting the results on cylindrical coordinates Radial displacements
First image with frioctionless support, second applying *Transform to the inner nodes

Andrea

I think that there is a little error on the midside nodes

Victor

Do you mind uploading the file(s) for that cylinder, Andrea?

Andrea

Here the LIML file.

For CCX user: to manage quickly the CCX input deck I created a spreadsheet linked to a text file.
Once updated the text file I copy and paste the text on "Custom model definition"

Victor

Thanks. This is mainly because the surface isn't very cylindrical - the element faces are flat. I repositioned the nodes to be on a cylinder and it came out with a much more uniform displacement around the edge.

If you need radial displacement on a non-cylindrical surface, you would still have to use node coordinate transformations. In a future version, I expect to have it allow formulas for displacement so you can define the direction as a function of position and get radial displacement on arbitrary shapes by using trig functions.

Andrea

I check the stp file. Nothing is wrong. I think.
But is not important. I prefer a inp file more simple as possible.

Andrea

P:S: Does is possible to move in other place the button "SOLVER"?