Scan-and-Solve for Rhino

Simulate Early, Simulate Often... In Rhino

# Restraint in 1 direction

This plug in looks very promissing and usefull.

I would like to be able to limit the restraint, not only to x,y,z direction, but also to one side of x,y,z.

If I place a object loosely on a flat surface it is restrained only in the direction of that surface, and not the other way.  Similar on bolted flanges, they are limited in the pull direction to the bolts and in the push direction to the flange surface.

(I am not a FEM expert, so I might be asking for something already possible or something impossible)

Regards,

Arthur

Views: 195

### Replies to This Discussion

Arthur,
If I understand this wish correctly, you are talking about solving a contact problem. Contact problems are not impossible, but they can be very hard to solve, especially in a general purpose setting.

That is you'll have to model these situations manually for the time being.
~Michael
Contact problems, as you call it may be hard to solve, but they are very common, your sculpture video is one of them. If you place a simple square block on a table and put one topside under compression it behaves completely different from if it is glued to the table, the tension side of the block-table contact is completely missing in the non glued situation. Probably it wants to move and I understand that is hard to solve but it is very ussable. Likewise a bolted flange under a bending moment. One half will be under a linear compression one half will be under a linear tension, difficult but very common and perfectly suited to FEM solving.

Regards,
Arthur
The problem is that you have to *assume* what kind of contact occurs and where it occurs in order to solve for displacements and stresses. But then, after you solve, you may find out that your computed displacements are inconsistent with the contact assumptions you made. So basically, you need to model all possibilities for each contact, and then check which of the solutions are consistent ... and then you may have multiple contacts and bolts ... and then you would want to add friction and then ... Can be done, but far from trivial.
You mean a gap might be created on the contact surface and/or the part may start to move? I have no problem if the friction is also considered as a restraint if that makes live easier.
Yes, the gap may be created on the contact surface, OR new contact surfaces may be created where gap was -- you cannot predict until you start the computations, and then you need to iterate -- because new contact or gap may change the loads, which changes the displacement, which changes the contact surfaces, and so on. There also could be a sliding motion at the contact, and friction make the situation even more complicated.
If you have a narrow gap (sawblade cut) in the solid or next to a restrianed surface would't you have the same problem? My concern is not so much with the contact area , but with the compression and tension behaving different within the solid. Having the contact area and displacement artificially fixed and displaying a warning in the report would maybe a usable compromise.
Interesting suggestion. Thank you. How do you envision user interaction with such a model? In other words, what do *you* expect to specify on a Rhino model? ... and of course, right now S&S handles single solids, but these issues are also important as we work towards analysis of assemblies.
Assuming a sample block on a flat horizontal surface and gravity only, I expect to specify:
- Direction of barrier +z
- coefficient of friction x-y directions
Maybe an option to fix x-y like it is fixed in a "full" restraint used in the current version of SnS.

Software checking friction and "gravity" -z direction is ok to keep block in place, if not display a warning displacement is expected and input needs to be redone. (Gravity in this sample can also be a user input force or another restraining surface of course)