Seismosoft | Earthquake Engineering Solutions

Dear Huffte,
thanks for your response, i have already encountered the above two scenarios that you have hilighted i have assign the strain values of Life safety to Immidate occupancy level but still that problem happen. Analysis get terminated after achieving the maximum base shear value and the failure at that point was just spalling of concrete from beam and columns.if anything else you can suggest.
Dear huffte i would like to have your comments on couple of points.
1) In Pushover analysis=>Nodal load=> incremental load=>type =force and displacement.what is the difference in selecting the force and displacement option bcz when i select displacement opt, program gives this pop-up(Zero diagonal encountered at the following nodes)before running the analysis, while when i select force opt porgram does not give any such pop-up before running the analysis.So whats d/f B/w these 2 and which one should use?
2) I first restrained the control nodes of each floor in x and Y dir only, the above warning message was apears so when i restrained the control nodes in all X+Y+Z+RX+RY+RZ dir my analysis run completely,no warning message appears but the T.P of my structre reduces to the half. if firstly it was 0.78 sec after doing this it comes out to 0.34 sec.why is it so?my orignal T.P of str 0.78 which matches to the Etabs value is now changed?why?Now which T.P i should use in order to find Target displacement?
Regards.

1) The message regarding the zero diagonal items in the stiffness matrix is a warning message of the program to notify users about this potential problem. With zero diagonal elements in [K], it is relatively easy to get divergence problems. If you have calibrated your model and it runs without problems, you can safely ignore it.
2) It does not seem reasonable to apply constraints to the control node in all DOFs X+Y+Z+RX+RY+RZ. It seems more logical to apply the diaphragmatic constraints in its level X+Y+RZ. This is probably the reason why you get such a reduced fundamental period.
3) Regarding the sudden drop in the capacity curve, it also seems to me that it is a structural failure, rather than numerical problems. Check the following: (1) fracture strain of your reinforcement material (apply a very large value, e.g. 1000 or even 1e8). (2) performance criteria or code-based checks that have residual strength. Once these criteria are reached the strength of the relevant element will drop to a much smaller value (e.g. the default is 20% of the previous strength).
Seismosoft Support

1) As per your directions i have apply X+Y+RZ rigid link constrains, alter the fracture strain of reinforcement and in the performance criteria i have increased the strain value limits. unfortunatelyt got the same response capacity curve break after attaning the maximum value of base shear. The thing is the failures i am getting in the post process is spalling of concrete from beams and col. The concrete and rebar strain values reaches the IO level only.
2) The only way in which my analysis got complete is by providing Rigid link restrain in X+Y+Z+RX+RY+RZ direction to control node, which is obviously seems strange bcz T.P of my structure get reduce to half of orignal.
3) i did static POA and NLTHA of a practical building before which was weaker than this current building in which i got my all results without such issues. This current building is a Theoritical building which i have designed on ETAB, how is it possible that this building with full fledge section and material properties fails very earlier without completing analysis.i have gone through all the things that was possible (convergance setting,material,performance criteria,loading phase,constrains etc). i don't know what else.To me its seems to be software issue this time bcz i did all these things before and no such thing happen.
Regards

The fact that you cannot achieve convergence does not mean that you should change the rigid diaphragms of the model to X+Y+Z+RX+RY+RZ. This will completely alter the structural behavior.
If you believe that you do not have structural failure in your building, then the convergence problem seems to be numerical. There are hundreds of posts on numerical stability and convergence in the forum, you will find many tips on how to solve your problem. Furthermore, make sure that you have read the 'List of SeismoStruct Convergence and Divergence Flags' and 'Tips to Solve Convergence Problems' in Appendix A of the manual. Finally, you could also try making the convergence criteria more stringent, rather then more loose (e.g. 1e-5 instead of 1e-4).
Last but not least, from our previous experience it seems very unlikely that this is a software issue, as you claimed. The program has been used by thousands of users for the past 20 years (or more likely tens of thousands) and all bugs are spotted and fixed very quickly. Because one user faces certain modelling problems, does not necessarily mean that the software is problematic.
Seismosoft Support