No convergence of nonlinear analysis for portal frame

[Gabriela.Bercot]

Hi,

I've been working on the model of a portal frame, and I had set up my bracings as tension-compression members, so it's being calculated as a nonlinear analysis. I am coming across 'no convergence issues'. I tried a few things as outlined on this page but without success ("No convergence of nonlinear problem" during analysis in Robot Structural Analysis (autodesk.com)

Could someone help me to sort this issue please?

[Krzysztof_Wasik]

HI @Gabriela.Bercot 

I have verified your model and have the following remarks.

- you have used compressed only members for bracings - used sectiopn types (angles) suggest that your intention was using tension only members for bracings (correct me if I have missed your idea). I would change bracings to tension only members.

-you are using default setting/parameters for nonlinear load cases. I would suggest activation Matrix update after each iteration setting as described in Non convergence.......-article

For bracings changed to tension only members and modified parameters of nonlinear cases  model converges for all load cases.

[Gabriela.Bercot]

Hi @Krzysztof_Wasik 

Thanks for your reply. My mistake, I thought I had set them up as a tension bars.

Related to the parameters for the analysis, I added automatic calculations (manual calculations - generated) and I am getting the same convergence error. My question is, can I set up the parameters for nonlinear combinations as well? It would this interfere with the results?

[Krzysztof_Wasik]

Hi @Gabriela.Bercot 

You can set calculation parameters for arbitrary load case/combination number. Type all for all load cases in Case list

[Gabriela.Bercot]

@Krzysztof_Wasik I had this settings as advised. I am having really high values for the bending moments in most of the combinations, not too sure why.

[Krzysztof_Wasik]

HI @Gabriela.Bercot 

In my opinion there are too many releases defined in the model. 

Verify displacements - especially for wind loads - they are unrealistic

• remove not necessary releases
• add supports if necessary
• if necessary add additional members for stucture stabilization

for instance when supports for UY are defined in nodes 380 and 381 moments drops down

.

[Simau]

@Gabriela.Bercot 

Try pinned supports in the right side in place of rollers

https://forums.autodesk.com/t5/robot-structural-analysis-forum/portal-frame-wrong-moments-values/td-...

[Gabriela.Bercot]

Hi @Krzysztof_Wasik 

 Yes, definitely there's some issues with the releases on these members. The problem is, I can't add supports where you suggested, this structure will be on top of rails to move north-south and will be lifted to ships when needed. I divided the column on the door frame in 3 parts, the beams below (marked in yellow) need to work as a cantilevers.

I have set up the releases like this and I still have higher displacements.

I attached the model, it is a bit hard to see on the snip.

[Krzysztof_Wasik · Accepted solution]

Hi @Gabriela.Bercot 

Set members  31 32 45 47  as truss members (tension +compression)

This will reduce displacements.

[Gabriela.Bercot]

Hi @Krzysztof_Wasik 

Thanks for that, it worked!