Disproportionate Collapse and Catenary Action

[PaulGrimes2282]

Hi,

I am trying to do a disproportionate collapse notional element removal analysis on one of my structures which will rely on catenary action in the catastrophic failure case.

Does Robot model catenary action, or is there any way I can approximate to it? Currently as soon as I remove elements I get huge deflections.

Thanks!

[Pawel.Pulak]

To obtain catenary action (reduced deflection on the cost of huge tension forces in members) you should activate 3rd order geometric nonlinear effects for appropriate load cases or combinations:

The name "P-delta analysis" is somehow misleading here:(

---------------------------------------------
If this post answers your question please click "Accept as Solution". It will help everyone to find answer more quickly!

Regards,

[PaulGrimes2282]

Thanks for your reply.

I have tried this, but I am getting non-convergence for all of my load cases. Does this mean that the structure is failing, or does it just mean that it is indeterminate?

[Pawel.Pulak]

Can you attach the model (without results and compressed to ZIP to reduce the size)?

The reason of no convergence may be both  that the structure is failing (buckling of some parts of the structure) and parameters of analysis (like not using full Newton-Raphson method with stiffness update after each iteration).

Regards,

[PaulGrimes2282]

Thanks! here is the file http://wtrns.fr/kdZreogd4TFyGLA

[Pawel.Pulak · Accepted solution]

Attached your model modified to obtain convergence of load case 1.

It was necessary to set incremental method with full Newton-Raphson algorithm ("Matrix update after each iteration" active):

It resulted in reducing the maximum displacement in load case 1 from about 10 meters for linear analysis to about 1 meter when considering the 2nd and 3rd order geometrical non-linear effects. So considering geometrical non-linear effects the displacements are strongly reduced but still very significant. Moreover they are related to rotations exceeding 7 degrees.

When trying to run this analysis for other load cased or combinations you can very easily exceed the assumption of small rotations used in Robot. It may be the additional reason of convergence problems during analysis.

---------------------------------------------
If this post answers your question please click "Accept as Solution". It will help everyone to find answer more quickly!

Regards,

[PaulGrimes2282]

thanks!

I am now getting convergence for LC1 - selfweight, but loadcase 2 (SDL), which is a significantly smaller total load is still giving me the error message 'matrix not positive definite'. Surely if the heavier loadcase works the lighter one should also?

[Rhonnie]

You need to set unnecessary loads to "Auxiliary" when doing the analysis. I do not think windloads and such are necessary for this analysis. 

Replying for future "searchers" that might pass by here.