Seismosoft | Earthquake Engineering Solutions

I have a concrete structure with two floors.I split beams and pillars on each of 4 elements and applied vertical loads. The floor it's modeled with rigid diaphragm .When run the static analysis gives me problems of convergence, the solution does not converges.If remove the rigid diaphragm the solution converges.It's correct neglect the rigid diaphragm?Otherwise, what could I do to converge the solution.

Hi,

The use or not of the rigid diaphragm constraints should be decided based on the actual in-plan stiffness of the slab you are modelling (if you have a in-plan flexible slab its use is not recommended).
The use of this solution will constraint only the displacements of the nodes in the plan you select (typically x-y), and don't have the capacity to carry vertical loads. Thus, i don't think the diaphragm itself is the cause of the lack of convergence for the static analysis.
I suggest you to verify if you don't have any load applied in an unsupported node... if not, you might try to use different constraints types (settings/constraints). Read carefully the help system in this topic.

Hope this help you.
Cheers

Can it be the case that you are restraining all 4 nodes of the beams with the diaphragm constraints? If so, try restraining only the end-nodes, so that to let the beams deform.
This will improve convergence and it will not affect the results, due to the axial rigidity of the beams

No because I have bound only end nodes. My structures consists of 20 fields of the floor. Until a 10 the solution converges. By placing more than 10 fields of the floor is not the solution converges.

If you search this Forum for "convergence problems", you will find plenty of hints&tips on how to overcome these challenges.

The same happens in the Help System, you can also search it for "convergence", and you will find plenty of suggestions, especially in the 'Notes' located at the bottom of many of the Help pages.

In the specific case of your model, the first thing we would try would be to relax the displ/rot convergence criterion from 10E-5 to 10E-4 (you will see it works).

Seismosoft Support

Many thanks Seismosoft, it solved it.