Kinematic constraints

[GabrieleNovembri1027]

Hi,
I'm doing some tests on the kinematic constraints. They look like the Columbus Egg. It seems that I can define a mesh type for different parts of a structure without worring about the inconsistencies in the points of contact between different elements.

I have some doubts and questions:
1) is an approach that can be used on a large scale? (eg for an entire structure)
2) which is the accuracy of this approach?
3) What is the difference between using the system in the point of incoherence of meshes or for the entire structure?
 

I did some tests.

I can not understand if variations of the values are caused by greater accuracy of mesh or by the introduction of kinematic constraints.

Thanks

[Artur.Kosakowski · Accepted solution]

I have some doubts and questions:
1) is an approach that can be used on a large scale? (eg for an entire structure) Yes, it can but mind that as any simplified/approximate approach this is not one of the 'fire and forget' options and you should carefully investigate the results before using them further.

2) which is the accuracy of this approach?

I'm very sorry but this is the kind of question I will not be able to give you the exact answer. It all depends on where you want to use them and what kinds of incoherence of meshes you want to compensate. As you can see on your test model the results are very close and by far more accurate comparing to the model with incoherent meshes of the walls and the slab and no kinematic constrains.

3) What is the difference between using the system in the point of incoherence of meshes or for the entire structure?
 

In the first approach Robot will create meshes trying to generate them in such a way that they are coherent along their edges and kinematic constrains are used only in the locations where the meshes are not coherent. The other approach assumes that the user is fully aware how to control the model as the meshes will be generated exactly as they are declared (based on their parameters only) with no attempt for coherence along the edges.

I did some tests. 

I can not understand if variations of the values are caused by greater accuracy of mesh or by the introduction of kinematic constraints.

Both. You can compare results between the model with denser mesh on the walls (no kinematic constrains) and similar model (also with no kinematic constrains) but with the mesh for the slab that corresponds to the meshes of the walls and the model with the kinematics constraints and the same model (denser mesh on the slab) with no kinematic constrains.

If you find your post answered press the Accept as Solution button please. This will help other users to find solutions much faster. Thank you.

[Lennox1973]

I am having real problems with Kinematic restraints, in that nearly every frame I run it fails with some stupid error message about kinematic circle loops.

Whats really annoying is that if I run the model from foundation upto 1st floor or 1st to second floor it will run and I'll get results, however if I try to run a model with the 1st and second floor in it, it will mesh but gets to the calculation running part and then comes up with the kinematic loop, try remeshing or turn kinematic off.

When I do both it still wont work 

Has anybody got any ideas on what is causing this sort of failure?

[Artur.Kosakowski]

The warning suggests that the correct generation of kinematic constrains is not possible. My assumption is that the meshes that are actually generated are very different for neighboring panels and the difference is too 'large' to be solved by the use of the kinematic constrains. The suggestion of remeshing is just to make these meshes more coherent and use kinematic constrains as 'additional help' in selected locations of the model rather than relay on them as the main solution for getting the 'right' results. The suggestion of switching them off just means use iterative meshing and get the coherent mesh (in such case no kinematic constrains are needed any more). Hope this helps.

[Lennox1973]

Thanks for the reply, but unfortunately it doesn't make any sense as it doesn't explain why indivually the models work but combin them and they dont. Especially when I have made sure that all columns are in line and in some cases redrawn core walls just to be sure that they are inline.

Even when I turn off the kinematic restaints it won't run.

I'll attach the model file when I get in to work on monday.

Hopefully then you may be able to help as our help desk is worthless.

[Artur.Kosakowski]

Assuming that there is a junction of 3 or 4 panels (top wall - bottom wall - floor defined as one or two panels at each sides of the walls) which have different meshes this may be the situation where kinematic constrains cannot be generated in the way that secures correctness of the results. I'll have a look when you attach the model.

[Lennox1973]

Thanks Artur