Altair RISE

A program to recognize and reward our most engaged community members

Nominate Yourself Now!

3-point-bending non-linear transient

Hello everyone,

i have to create a timedependent 3-point-bending simulation. I created a table with a constant displacement for 10s, but the simulation doens't start..

There are fixed spcs for the fixed cylinders and one spc (with dof 1 in x direction), which are combined with a spcdadd.

I also created a spcd for the pressing cylinder with C1 and a given displacement.

The table and the spcd were connected with the dload-card.

I get error 186.

Thanks in advance!

Find more posts tagged with

OptiStruct

Auto Tagged

Accepted answers

All comments

Adriano_Koga

Hi,

as per the error message, you're missing a SPC.

Using SPCD requires that you also apply a SPC at the same node and DO as well.

I.e: you have a SPCD in node 16917, DOF1. You will need to apply a SPC at the same node 16917, DOF1, constraining it to 0.0.

A fatal error has been detected during input processing:

*** ERROR # 186 ***
An SPCD component must be constrained by an SPC or SPC1 data.
GRID ID = 16917
component = 1
LOAD set ID = 10
SPC set ID = 13
subcase ID = 1

Alt

Hey,

thanks for the fast answer. The other DOF's will be free, right?

Adriano_Koga

correct. You will need to add SPC at the same nodes/dofs where you apply your SPCDs.

Other than that, only the fixing SPCs, as usual…

Make sure the SPCs are in the right load collector, and not mixed with the SPCD collector.

Alt

i have started the Simulation, but it doesnt convergate.. There is also a distance between the contacted surfaces. And als the spcs are combined in a spcadd.

Maybe i have wrong parameters for nlparm, nladapt or nlout?

The displacement is velocity-based (constant velocity with table for 10s)

Could you check the file please?

Thanks in advance! Really helpful tips!

Biegeversuch_originale_Matrix-Matrix_Cont.hm

Adriano_Koga

i won't be able to investigate in details, but a few important things:

you need to constraint the moving cylinder in the other directions to avoid it to fly away in Y and Z (and rotations)
you will probably need to run using LGDISP, so instead of regular NLPARM, you should use NLPARM(LGDISP) insted
you have used a CNTSTB in your model, with S0 only 0.01..This might be ok, but maybe you would need something larger at the beginning.
NLPARM could probably start with smaller increment, so probably starting NINC with 100 would be interesting, making initial convergency better
for debuggin purposes, you could use PARAM,NLMON,DISP, that will write an additional H3D file with displacements. With these you could check the intermediate iterations, including non-convergent ones, to debug your model, and check possible issues, like parts flying away..

Adriano_Koga

a few tips…

https://community.altair.com/discussion/33277/tips-tricks-optistruct-nonlinear-analysis

Alt

Thank you very much! :)