STAAD Response Spectrum Questions

Hi,

I am curious about a couple of things on the way STAAD handles response spectrum analysis, and below are my questions. I would be very glad if you can provide me some answers . I'm using build 20.07.05.15, by the way.

Due to the client's own design manual, I had to use the NEHRP provisions to come up with a response spectrum which is to be used in seismic analysis of a bridge. Just as a side note, the functions used in NEHRP's response spectrum are slightly different than AASHTO LRFD Specs. Since STAAD.Pro does not have NEHRP among the built-in codes, I entered the spectrum by hand using the editor (in time vs. spectral acceleration format).

1. Once I entered the spectrum using the editor, I switched to the GUI to check the shape of the curve by using the Edit option. Here, I was surprised to see that the numbers listed in the edit dialog are somewhat different than what I initially entered through the editor. Below are two screenshots comparing the numbers in the editor and in the GUI. No scale factor is applied, everything is in terms of "g" for now:

Why would the spectral acceleration values be different ? Is there some kind of conversion going on behind ? Most important, which set of values would STAAD use in calculations ?

2. This one is about damping. The spectral acceleration values from the maps in AASHTO LRFD (not sure about direct values from USGS though) include damping, if I do not misinterpret the code. For example, the title of AASHTO LRFD Figure 3.10.2.1-5 says:

"Horizontal Response Spectral Acceleration Coefficient for Region 1 at Period of 0.2 Seconds with 7 Percent Probability of Exceedance in 75 Years (Approx. 1000-Year Return Period) and 5 Percent Critical Damping."

Again, if I am not wrong, the design response spectrum I construct using the AASHTO LRFD maps would already be 5% damped. So, wouldn't the default 0.05 in the "Damping" box reduce the amplitudes further ? As far as I know, If I enter "0" to the "Damping" box, it would revert back to the default 0.05. Is there any way to tell STAAD not to use damping ? In other words, do I have to make sure that my spectral acceleration values are undamped before entering them ?

3. What does the "Save" checkbox under the "Combination Method" combobox do exactly ? I checked the help file, but could not find any information on this. It looks like the help file for the GUI belongs to a previous version (I assume 2006), since the screenshot in the help file is different than what I see.

4. This is out of curiosity: What I am planning to do for load combinations was plugging-in spectrums seperately into two different primary load cases for X and Z directions. For example the command for X-direction spectrum should look like this where participation factors for Y and Z are 0:

SPECTRUM CQC X "some number depending on the R factor" ACC SCALE 1 DAMP 0.05 LIN

After that, I should be able to create the required number of 100% +- 30% combinations via "Define Combinations".

My real question is, what would be the purpose of declaring X, Y, Z participations at the same time for the same spectrum ? The only thing that came to my mind is that maybe this flexibility would allow application of seismic forces from an arbitrary angle (by using the participation factors as the amplitudes of the unit vectoral components). If there is an other reason to have this option, could you please explain ?

Thanks a lot for your time !

  • Hi Ekin,

    I can help you with a few of these.

    1.  The RS values in the editor differ from those in the GUI by a factor of 12.  That's just a units thing.  From an earlier version of your model, I could see that your first Period-Acceleration data pair was 0  0.05696 where the units were based on inches.  If you multiply 0.05696 by 12 it equals 0.68352, which is exactly the value that is shown in the GUI screenshot that you included.  So it look like the UNIT INCH command has since been removed, so the program is interpreting your Period-Acceleration data pairs in units of feet, because 0.05696 feet per second^2 equals 0.68352 inches per second^2.

    2.  I'm going to have to let the dynamics experts weigh in on the issue of damping.  What I can say is, based on the earlier version of your model, as I decreased the damping from 0.05 to 0.005 and then to 0.0005, I saw slight decreases in the dynamic base shears according to the CQC combination method.  As soon as I set the damping to 0.00005, the CQC method base shears jumped right back to the values I saw when damping was set to 0.05.  So when I get to 5E-5, that must be so small that the program considers it to be zero, which, as you say, is interpreted as a command to use the default value, hence the results that coordinate with the damping value of 0.05.

    3.  From the Technical Reference, 5.32.10.1.1 Response Spectrum Specification - Generic Method, the save option results in the creation of a file (file name with "Acc" extension) containing the joint accelerations in g's and radians/sec2.

    4.  I would like to hear more discussion on this one, too.  So far, my interpretation has been the same as yours...create two different RS Load Cases, one for each of the two plan directions for seismic evaluation, and calculate the scale factors for each separately.  Then combine the results using REPEAT LOADs (or LOAD COMB, if no non-linearities exist). 

    Cheers,

    Chris



  • Hi Chris,

    You have clarified more than a few points for me, thanks a lot !

    1. I initially thought STAAD treats period/spectral acceleration as unitless values, then adds the units after the application of the scale factor. Now I see it's the other way around.

    2. This damping issue is still bugging me. I would be very glad to hear from others who deal with seismic analysis, as well. How do you enter an already 5% damped response spectrum into STAAD?

    Thank you for checking the lower bound before the default 5% kicks in though, I was planning to do something similar. On the other hand, I would initially guess the dynamic base shear would go up with decreasing amounts of damping. Shouldn't the demands increase when the spectrum gets less damped, as shown in the figure below ?

    3. Thanks for the clarification !

    4. As I said, this was just out of curiosity. Although I am not planning to use this feature, it's nice to have such flexibility. My plan for introducing the seismic loads is exactly as you explained, I just have to make sure I cover all possible combinations.

    Thank you,

    Ekin

  • I scratched my head a little on that issue of the decreasing base shears with decreasing damping, too.  Had a couple discussions with our Senior Advisory Software Developer, and here is a synopsis:

    1. The spectrum input is presumed to be created using a specified damping. STAAD does not know this damping value. The value, in effect, is used for all of the modes.
    2. Alternatively, you can enter the spectrum data via a file. On this file you can enter several spectrum curves, one for each damping level. STAAD will use double interpolation to obtain the spectral acceleration value for each mode based on that mode's damping value (not the damping value on the spectrum command) and the period of the mode.
    3. The damping value on the spectrum command is for constructing the modal interaction matrix for the CQC and ASCE4 methods. This value plays no role in the results for the SRSS, CSM, ABS, & Ten PCT methods.
    4. The DAMP on the spectrum command is the damping value to use in the CQC mode interaction equation. This calculates how much greater the CQC result should be as compared to the SSRS result (due to close mode response interaction).  Since the spectral acceleration data was generated assuming a particular damping, it only makes sense to use that damping value in the interaction calculation.  Since STAAD does not what that assumed damping value was, the DAMP input option is made available. Ideally the CDAMP option would be used along with several spectral acceleration versus period curves for various damping values that are entered on a file. Reiterating, to really change damping in a response spectrum case, you need to change the spectrum acceleration versus period curve.

    Hope this helps.

    Regards,

    Chris



  • Chris,

    This indeed helps.To summarize what I understood:

    • STAAD needs the damping factor corresponding to a user defined set of Sa/g values as a starting point, so that Sa/g values corresponding to other modes with specific modal damping factors can be calculated, if necessary. In other words, the DAMP command defines the input response spectrum itself, and it has nothing to do with "further damping" it, as I thought initially.
    • The damping factor is actually used in constructing the modal damping matrix. In default state (if no mode specific modal damping factors are introduced), this would assign the same damping factor to all modes.
    • If known, STAAD allows the user to introduce mode specific damping factors. In case not all of the modal damping factors are known, I assume Rayleigh's approximation or some other numerical method is used to construct the modal damping matrix for n number of modes.

    Well, thank you one more time for the clarifications, I do appreciate it.

    Ekin

  • Hi Ekin,

    I did a little more digging and learning.  Here are responses to your summary items:

    _________________________________

    You said: STAAD needs the damping factor corresponding to a user defined set of Sa/g values as a starting point, so that Sa/g values corresponding to other modes with specific modal damping factors can be calculated, if necessary.

    My response: I don't think STAAD uses the DAMP value to calculate values corresponding to other modes with specific modal damping factors. It is my understanding that the DAMP value is only used for constructing the modal interaction matrix for the CQC and ASCE4 methods.
    _________________________________

    You said: In other words, the DAMP command defines the input response spectrum itself, and it has nothing to do with "further damping" it, as I thought initially.

    My response: It is my understanding that the value specified by DAMP is only used by the CQC and ASCE4 combination methods, which incorporate the damping value in the combination process. The DAMP value is not for the purpose of providing further damping. It seems logical that the value specified for DAMP should be equal to the damping value that is inherent in the spectrum data that is being used.
    _________________________________

    You said: The damping factor is actually used in constructing the modal damping matrix.

    My response: It is my understanding that the DAMP factor is actually used in constructing the modal INTERACTION matrix. If there is a desire to incorporate individual damping ratios for individual modes, that is achieved through the use of the MDAMP option, as explained under Section 5.26.4 of the STAAD Technical Reference manual.
    _________________________________

    You said: In default state (if no mode specific modal damping factors are introduced), this would assign the same damping factor to all modes.

    My response: Yes, the default value for DAMP is 0.05 (regardless of the inherent damping value in the spectrum data), and it will be applied to all modes by default unless MDAMP or CDAMP are specified.
    _________________________________

    You said: If known, STAAD allows the user to introduce mode specific damping factors. In case not all of the modal damping factors are known, I assume Rayleigh's approximation or some other numerical method is used to construct the modal damping matrix for n number of modes.

    My response: Take a look at Technical Reference section 5.26.4 Modal Damping Information for many different options to define mode-specific damping values.
    _________________________________

    Regards,

    Chris