perform a spectrum analysis with spectral values on 3 directions

[marcomer]

Hi everyone,
i'm trying to perform a spectrum analysis with ansys apdl 13; i've seen several tutorials about performing that analysis with a single response spectrum, but i have to combine spectrum values on three directions (x,y,z); in this special case, does anyone knows if the combination of modal shapes has to be done on post-processor fase or i can simply define the three spectrum and combine everything with a single SRSS solution command?

thank you all,
regards

 

[JoshPlumSE]

I know very little about Ansys. However, response spectrum for the three directions are usually analyzed independently. You then combine results from each response spectra analysis together using some combination method.

The 1997 UBC and older codes allowed an SRSS combination of the results. The newer ASCE code, for some reason, insist that you use a 100%+30% combination method for the two horizontal directions. I've seen this extended out to 100%+30%+30% for spectra applied in all three directions. However, I believe the nuclear codes have now moved towards a 100%+40%+40% combination method.... though I am not positive about that because it has been a long time since I worked on a nucler job.

 

[marcomer]

Thank you joshplum,
i think i will use the SRSS combination; however, i don't understand how i can do this procedure in ansys: i found some tutorial about spectrum analysis, but i can't find anything about combination of results for each sysmic input. Any ideas?

 

[JoshPlumSE]

Ansys doesn't have a concept of load combinations do they? That's the tough thing about using a mechanical FEM program for a Structural project.

If they have any post-processor type Macros that you can build then you could run three separate analysis and build some Macros to export the results to Excel for combination... or something along those lines.

 

[marcomer]

thanks for your reply, in this way exporting results to Excel I obtain just numerical results. Instead I want to obtain graphical results which Ansys allows to visualize for single response spectrum analysis. So I have to combine results within Ansys...

 

[DST148]

I had worked on nuclear projects in Asia in the early nineties. We used STAAD-III for the dynamic analysis. The combination of the three spatial components was done using SRSS / 100 - 40 - 40 method. For the version we had, these were the limitations of STAAD those days: 1)It did not account for correction required for missing mass. 2) It did not have load combinations per ACI-349: Code requirements for nuclear concrete structures.
We analyzed the structure for acc. in one direction, reanalyzed for missing mass correction, and wrote a post processor to combine the two results to get forces in one direction. This was repeated for other two directions. After getting the forces in three directions, they were combined using SRSS 100 - 40 - 40 method. The seismic response was then combined with gravity to get the final design forces.
We also wrote a post-processor in 'C' language for design of concrete elements per ACI - 349. We had to get the approval for the post-processor from the review board.
With SRSS the signs are lost. So all the six forces either could be positive or negative for load combination purposes.

 

[JoshPlumSE]

Marcomer -

I assume that you can get plotted results in Ansys for the response spectrum analysis in one direction, right? What do you think of the results that you're plotting?

The reason why I ask is because response spectrum results involve a statistical combination (SRSS or similar) of individual modal results. While the results for each mode make sense. You get positive and negative deflections and forces, you look the positive / negative sign convention when you do a statistical combination of the individual modes.... So, each numerical value is a valid maximum design value. However, you don't know whether the sign is positive or negative. That can cause problems for plotted contours and such. So, you frequently have to turn to the spreadsheet results anyway. At least for the final data.

Just something extra for you to think about....

 

[TERIO]

If it is of any help, ANSYS does have load combinations. Look up commands like LCASE and LCOPER as a start.

 

[marcomer]

first of all, thank you all for your help..anyway i've solved the problem by inserting spectrum values on the three directions with coefficient (1 or 0.3), doing the three analyses separatly and just at the end of the three analyses, reviewing results with command read from file (reviewing .mcom file)

 

[Elite7]

Just my $ 0.02:

Combination of modal participations is done using SRSS/ CQC/ 10% or other methods.

Combination of multi-direction responses is done using 100-30-30 or 100-40-40 methods.

Best to be clear about the difference between modal combination and directional combination.