No convergence of nonlinear problem.

[sak_n]

I got calculation message "No convergence of nonlinear problem. Case no. = XXX" when I use the tension members for sag rod and roof bracing. I read the older post and try to change value of tolerance for residual force and displacement in the nonlinear analysis parameters. the message still has existed. If I ignore the message whether it has problem on the member design or not. How can I find solution for these problems. the model is attached for checking.

Thanks for your help

[Rafacascudo · Accepted solution]

Only problem on your model is that there is no horizontal support at all . Only vertical springs. Assign some to you support nodes and all combinations will converge very fast

Other than that , nonlinear analysis for simple cases, other than self weight/Dead load comb load cases ,are pointless as in a nonlinear analysis each combination is solved as it was a single load case. No individual results of simple load cases are summed.

So you can set all your simple load cases as auxiliary and save lots of analysis time.

Also , very often is faster and easier to the model converge with the setting below

[Artur.Kosakowski]

Hi @sak_n

I just wanted to ask if you managed to calculate your model following the suggestions posted by @Rafacascudo (thank you)? If yes, would you mind to mark his post as the solution so that it can be found faster by others? Thank you.

[Anonymous]

Try changing the tolerance in 'job preferences'.

If the model is non-linear due to tension only members then the solver will pick the tolerance from job preferences rather than the analysis parameters.

[Rafacascudo]

Sorry , but  I think That´s not the solution for this model . Analysis does not converge because the model is unstable due to having no horizontal restraints.

Even if it converged (Impossible for load cases and combinations with horizontal loads like wind) ,results would be totally wrong!!!

"If the model is non-linear due to tension only members then the solver will pick the tolerance from job preferences rather than the analysis parameters"

That´s new for me. I tried that but it didn´t work . Paramaters set on analysis options remained active.

[sak_n]

Hi Rafael,

I do following steps as your instructions and the analysis is converge. 

1. I assign the horizontal spring support at the end of pile.

2. The analysis type of simple load cases is changed to  auxiliary only except combination loads.

3. Set parameters for all load cases (both simple load cases and combination loads) as non-linear analysis parameter and select matrix update after each iteration as addition.

As the above steps, is it correct? Does the results is correct? The model is attached for checking.

[Rafacascudo · Accepted solution]

perfect!!!