The reason is that the members under consideration are under selfweight load and are calculated not only as compressed or tensile elements, but as eccentrically compressed and eccentrically tensioned ones. In case of eccentric compression and eccentric tension of unsymmetrical sections the coefficients FIB and ETA are not defined in Russian codes. This curious case is related to the imperfection of SNiP methods and computer realization of them.
For simplified engineering calculation the design codes specify additionally the cases of pure bending or axial tension/compression (what is impossible in the real life). But the phase shifts from one case to another are not described in design codes. Here lies the problem of computer realization of the methods. E.g. we have a beam and assume it as flexural element; but this element will have tension or compression due to work with other elements of a structure. STAAD.Pro will catch this effect and switches to the other algorithm of calculation. And here it gets clear that the FIB and ETA coefficients are not described by the same design codes.
This is the nature of an effect which might leaded to the errors in your model.
Here are few workarounds which could help:
1. Use TRUSS specification for bracings and lattices – this will exclude their flexure.
2. Model the desired stress-strain behavior for the double check – try to remove SELFWEIGHT option.
3. Isolate the element and load it separately and perform an additional check.
Might be that the combination of these steps should be applied to get the correct picture.
If you need some more assistance on this, please create a Service Ticket for Bentley Technical Support Team and attach the model for the investigation.
Thanks,
Modestas
