HELP Autodesk,I have detected errors in the spectral analysis. RSA 2022 y 2023

[bartolome.tavera]

Hello, I am an instructor of a "Robot Structural Analysis" course, we are ATC here in Barcelona Spain.

I have done comparative tests against other analysis and design programs and Robot (which used to do it well) is giving incorrect results very different from those expected in the spectral analysis.

The error is serious, the result is far from correct, and here is a proof: the same building model in Robot 2022 and Robot 2023 gives different results, the dynamics of the structure don't match.

The results for self-weight gravitational loads, live loads and dead loads are correct; but the error in the spectral analysis is disappointing and very sad.

I attach the model, and attach comparative images:
Different results between RSA 22 and RSA 23, (and also both are a bit far from those obtained in other programs)

Model: https://drive.google.com/file/d/1ByWkWF0lScdSIiGfdYmr4HvfQfDOKH38/view?usp=sharing

I would like someone to review this, check that the results are wrong, and contact us, we trust Robot and we think they need to improve. Robot did this very well until version 2020 but apparently since 2022 it was damaged.

[okapawal · Accepted solution]

Hi @bartolome.tavera 

This is an issue in 2022 release.  It is corrected in version 2023 which gives correct results.

It was reported on another thread  https://forums.autodesk.com/t5/robot-structural-analysis-forum/issues-with-mass-and-stories/m-p/1089...

Modal analysis gave wrong results, because masses converted from loads are doubled.  It happens when the option is checked on "Create the main node for floor diaphragm at the Gravity Center of the story projected onto the floor" in stories definition dialog.

You can switch this option off to get proper results in 2022

[bartolome.tavera]

Thank you very much Okapawal, I have seen how you mentioned that the option is no longer default in version 2023, thanks for correcting the error.

There is another thing pending and that is the calculation of the mass of each level. I have already found the reason why Robot gives higher mass results than ETABS, CYPECAD, Midas Gen and Staad; for Robot, the mass is calculated taking into account the entire level, while the other programs do it at half height in each level; and they neglect half the height of the columns at the base level since that mass would be assigned to the ground.

Robot does not do this, and therefore the result of the seismic weight of the building will be higher than in other programs if the model is identical.

Thanks for your time, regards.

Image describing the theme of the mass:

[zeeblake]

Things like these are what is missing in the documentation, in my honest opinion. Most of the times, the main limitation I find in Robot is that I don't know what assumptions are really used by the software for some specific calculation, and there's no place for finding them. The only way is by experimentation, like you did. And that's a big limitation. By the way, I believe the correct assumption here is what Etabs, CYPECAD, Midas, Staad, and the rest do: the mass of each floor should be mid-floor to mid-floor.

[bartolome.tavera]

I agree with you, the mass of each floor should be mid-floor to mid-floor.

[sergey.churilov]

I believe that you can obtain such mass distribution with a little bit of additional work.

You can divide the floor columns in two parts and then define stories from mid-floor to mid-floor.

The approach is not optimal, but it can be achieved.