UNDERGROUND STRUCTURE ANALYSIS

 

I am performing a dynamic analysis (Response Spectrum) for a Structure with 2 basement levels. I have some queries regarding the use of STAAD commands to model the soil - structure interaction.

 

Results of static and dynamic analysis are combined using LOAD COMB command.

A finite element model of the basement levels was developed with STAAD. The basement walls were modeled as PLATE elements connected to beams and columns.

In order to model the soil as Winkler spring calculated on the basis of the modulus of subgrade reaction of soil, each joint of beam and column members have to be attached to a soil spring calculated on the basis of the influence area of that joint. Similarly, each joint of the plate elements have to be attached also to a soil spring representing the influence area of that plate element............. How to do this?

I was looking the example problem No. 27, an example to illustrate the usage of commands necessary to apply the compression only attribute to spring supports for a slab on grade, where the command used to instruct STAAD to generate support springs which are effective in the global Y direction was the STAAD facility for automatic generation of spring supports to model footings and foundation mats ..... But for the basement levels I don't think this is OK....?

All the appropriate loads including hydrostatic pressure on the basement walls were input in STAAD.

I was thinking to use the Spring Tension/Compression Specification, but based on the Technical Reference Manual, Section 5.27.5,........

"The SPRING TENSION command should not be used if the following load cases are present : Response Spectrum load case, Time History Load case, Moving Load case. If used, the SPRING TENSION /COMPRESSION will be ignored in all load cases."

"Do not use Load Combination to combine these cases. Tension/Compression cases are non-linear and should not be linearly combined as in Load Combination. Use a primary load case with the Repeat Load command."


What can I do.......?

Any help you can give me to model the soil-structure interaction would be appreciated.


Best Regards,


VETP

  • The nature of the Response Spectrum analysis makes it difficult to incorporate compression-only spring effects, because the analysis inherently describes vibrations that would need to alternately activate and deactivate supports. If the shear wall effect is well-behaved in your model, you might consider doing a test run that completely eliminates the lateral restraint from the backfill during the RS analysis, and see what effect it has on nodal displacements at the ground floor level and basement wall shears. It could turn out that the building resists the seismic load primarily by shear in the parallel walls, rather than leaning against the soil on the passive perpendicular wall. Another consideration would be to evaluate the amount of resistance provided by skin friction of soil on the parallel walls during a seismic analysis. Maybe between these two alternate load paths, you can get comfortable that the "error" is minor if the springs associated with the walls are allowed to produced tensile reactions. Another potential option might be to compare the results of the response spectra analysis to the results of an equivalent lateral force procedure analysis. If you can be comfortable that the ELFP analysis produces more severe effects than the RS analysis, then you could implement REPEAT LOAD combinations and consider compression-only springs using the ELFP results. Hope this is of some help. Cheers, Chris



  • Dear Mr. Chris,

    Thank you very much by your time and by your valuable opinion about this matter. Very usefull to improve my knowledge and to learn more about the capabilities of the STAAD to perform this type of analysis.

     Best Regards,

    VETP

  • Our pleasure, as always. Good luck with the analysis. Cheers, Chris