Response Spectrum Output Problem in SAP2000

[Sannyboy]

thread801-360032

Dear All
In continuation to the above listed thread. I am encountering another problem while performing response spectrum in SAP2000

I am analysing a Mat foundation in SAP
DF is a load case with Dead load and SY is for response spectrum in Orthogonal Y direction. However the Output base pressure for SY shows all regions as positive. My guess is because SRSS method is followed for superposition.
But now the problem is DF+SY load combination isnt giving true picture about the Uplift or Max Base pressure.
Also ideally DF+SY and DF-SY should give either Max or Min Values respectively, however the output base pressure diagram for both the Load combinations is same.

Please suggest some alternative turnaround.

Thanks and Regards
Sunny

 

[Sannyboy]

It is a STG table top foundation in which RSA needs to be performed since it is in very high seismic zone
So basically two different issues i have
1. RSA results to be applied with sign in the combinations : - for which the above query and.
2. To use compression only springs in order to reflect the actual base pressure distribution below the mat, NL load cases (similar to repeat load in STAADPro) is being used

Any guidance would be deeply appreciated.

Sunny

 

[stressed]

Response spectrum analysis is by definition a linear frequency domain analysis. So nonlinear (NL) compression-only soil supports cannot be properly considered in a Response spectrum analysis. Run your seismic load as NL time history and you can considerer compression-only behavior of soil supports. Notice how under Define load cases, the RSA case says "Response spectrum", not 'Nonlinear'. RSA won't work for what you're trying to do.

 

[JoshPlumSE]

I wrote a RISANews blog on this subject a couple of years ago. See link below:
Why Don't Response Spectra Results Satisfy Statics?

Obviously, I'm a RISA guy, so I'm not very familiar with how SAP does things. But, you might check to see if they have a RSA solution option along the lines of RISA's "Use Dominant Mode for signage". This makes response spectra analysis results a whole lot easier to understand whenever you have one mode that dominates the response spectra behavior.

 

[Sannyboy]

@Stressed :: I understand RSA is a linear analysis.. and this is the reason when you try to convert a load combination with seismic to a non linear load case. SAP converts all load cases except the Seismic ones which are later linearly added back in the load comb,.
But i am really confused how to proceed with analysis with RSA even if i am not going for compression only case. The Stresses generated in the Plate Element also do not resemble the correct situation.
Any Suggestion how we can convert RSA to time history problem.

@Joshplum : So i checked the SAP , there is no option as "Use dominant mode "

so my problem still is " USE RSA TO ANALYSE A MAT AND CHECK FOR CONTACT AREA MORE THAN 80% "

Thanks and Regards
Sunny

 

[JoshPlumSE]

Due to a lack of signage, interpreting RSA results can be really, really tricky. It comes down to understanding what the program is giving you and then applying engineering judgment to those results.

In RISA, the Use Dominant Mode for Signage option really helps. Another "trick" I like to do is to apply my ELF (equivalent lateral force) seismic loads to the model. Then I can compare the reactions from RSA to the reactions for ELF. Hopefully, that comparison (with the understanding of the signage issue) helps to make the RSA results more understandable.... though this is probably significantly more helpful for models which have a reasonably dominant mode.

 

[Sannyboy]

@JoshPlum & @Stressed:
I finally took the results of Response Spectrum and applied the loading as a Static Case. Also for the uplift case i applied GAP element concept. However now i have another problem.
I analyzed a small 1m x 1m mat with GAP link. The problem is AFter redistribution SAP should ideally show some area with 0 bearing pressure and redistribute for remaining. however its shows bearing pressure for complete mat.

IN another try.. within the definition of load case for non-linear i checked the option for "store result for each step". there i can see each step of iteration .. and its seems the envelop value is taken in the final reported result without this "store result of each step" option.

 

[Sannyboy]

Attachment as below.

Test 1m x 1m mat with simple load to ensure uplift case.
I try to redistribute bearing pressure

 

[stressed]

Sanny,

Your model shows uplift under static seismic, a load which you applied in simplified fashion as a joint load moment about Y with -Z load. Attached is a screenshot of soil pressures showing one end of the mat with bearing pressure of zero. That's confirmed by deflections which show uplift joint deflections of almost 1/2"

USGS and other agencies often make time history earthquake records available on their sites. This includes earthquake records from around the world. With larger projects, geotechnical engineers may provide a seismic hazard evaluation and site specific seismic design criteria.

For dynamic analysis, time history is the only option for signed results which can consider the possibility of uplift. SAP2000 will take RSA results and apply both + and - whenever RSA is used in a load combination. That + and - in the output can be confusing because the values don't necessarily correspond. You don't know what the values of other responses are doing when that max response occurs. RSA, unlike time history, is a statistical type of analysis of likely system response. I haven't seen any option in SAP2000 for a dominant mode type of output. That option could definitely be useful if you happen to have a model with a large percentage participation in one mode.

If you use one of SAP's code based auto-seismic options for static lateral force, you can take base reactions from that case and use the scale factor in the RSA case to scale the RSA base reactions to match those from the static seismic case in the same direction. The idea is that base reactions should match, but RSA will give a better distribution of load distribution throughout the structure.

 

[Sannyboy]

@ Stressed. : I appreciate you took time out to review the file and the suggestion. Finally i switched to STAAD for static analysis , however i shall explore more about the same in SAP.


Thanks once again.