Seismic design of footing

[SHBH]

Hi,
I am designing a steel rack for very high seismic zone in South America. For seismic design of steel structure i am using responce reduction factor R as 3.25 (ASCE-7), means i am designing my structure for a less seismic load taken into account it's ability to entre into post elastic mode with some inelastic deformation. So far so good. Now my question is while designing foundation of this structure why i use the support reaction directly from this analysis as seismic loads have been reduced by a factor "R". To my understanding foundation should be designed for actual seismic loads without any reduction.(FOS in sliding or overturning can be taken as 1 but against the actual seismic loads received by this foundation......without any reduction)
I coudn't find any literature on this topic. So i would like to receive some suggestion or comment on my views.

 

[CROWN06]

Refer to Page 122, ASCE 7-05, note a.
Hence after analysing your steel rack, what ever reaction comes, the E will be taken the value it comes, if you are designing your conc ftg using Strength method and factored loads.
when your are checking soil bearing capacity, divide E values with 1.4. Actually if you are using the
ASCE load comb, you are doing it any way.
R to my knowledge is reduced based on the energy dissipation or dampness ability. So the
structure will throw less load on ftg too as it has lost some energy.

 

[Lion06]

SHBH-
I would ask why you are using an R value of 3.25. If I am correct, using an R>3 means that you have to follow the AISC's seismic provisions. If you use R=3 or less, you do not (unless you are in seismic design catefory D, E, or F. I can't imagine that getting that extra 0.25 is worth all of the detailing that you will have to do to make R>3.

 

[SHBH]

StructuralEIT ,

Actually i have got site specific spectra for this site and the accelerations are huge (peak spectral acc is 1.3g in OBE and 1.81g in SSE cases). That's why i need to consider the actual value of R, and would have been happier if it was even more than that.

 

[SHBH]

Thanks CROWN06,

But even after dividing the reaction from steel analysis(which in fact LRFD load comb as per ASCE7) by 1.4 ( or infact using ASD load comb for foundation in which E load factor is 0.7, as you said)should one allow some relexation in allowable bearing capacity.

For overturning 12.13.4 allows 10% reduction in the FOS, what about the sliding ( i guess it is applicable for sliding as well).

 

[CROWN06]

As per my knowledge you cannot reduce any further if you are using ASCE 7-05 combos. As far as section 12.13.4 is concerned, I am sure 25% reduction is ok when without seismic load the stability is 1.5. One small tip, I have worked overseas and standard of construction are mostly sub-standard. So personally I will not push US codes to limit as the execution format is non - US. As an engineer I think you should consider this also in your design and play safe. Also if the ground acceleration is much higher than highest in ASCE map, than, probably we should not use it's formula's directly and do some research.

 

[civilperson]

ASCE 7-05 Map of "Maximum Considered Earthquake Ground Motion for US 0.2 SEC Spectral Response Acceleration(5% of Critical Damping), Site Class B" has upper values of: 275(California), 175 (Utah/ Wyoming), 300 (Missouri/ Illinois/ Kentucky/ Tennessee), 258 (South Carolina). The values for the oversea project of 130 and 181 are within the scope of the ASCE formula.

 

[OldPaperMaker]

Instead of chapter 12 in ASCE 7-05 shouldn't you be using chapter 15 "Seismic Design requirements for Nonbuilding Structures"?

 

[Qshake]

I know these are code requirements, But I really don't think it useful to say "actual value of R".

Unless there are some specific ductility requirements, the "R" value is empirical and I doubt that the difference from 3 to 3.25 can be quantified as accurately as the code implies.

I digress, the original post isn't about seismic philosophy it's about the code application.

Regards,
Qshake

Eng-Tips Forums:Real Solutions for Real Problems Really Quick.

 

[Zhuge]

As you designed the steel member with ductility = 3.25,the column member size should be much more smaller than if you design it elastically. However, your foundation should be able to have bigger capacity to allow the steel member yield first.
Therefore, You need to use steel member section moment capacity multiply by an overstrength factor roughly about 1.1 - 1.25, to make sure your steel member yield first before the foundation fails.

 

[SHBH]

Zhuge,
you want to say that moment capacity of column divided by 1.4(strength design factor), multiplied by 1.1-1.25 should be used in the overturning requirement of the foundation.
seems to be unrealistic to me, specially when your section sizes are governed by the deflections.

 

[Zhuge]

Hi SHBH,
I am not quite sure whether you're familiar with capacity design or not. However, since you decided to design the steel member using ductility factor = 3.25, You don't have any other choices but to design with capacity design.
That means you will expect to have some yielding at your steel member during extreme earthquake.Therefore your foundation has to be able to carry the forces that causes the steel member to yield.
Steel member moment capacity = S * Fy
But you need to take into account the overstrength factor due to strain hardening and material variation = 1.10 -1.25 ( roughly).
Therefore, your foundation need to be able to cary = 1.25 * S * Fy (rough estimate).

That's why for petrochemical piperack, where the displacement limit is governed we tend to design it with ductility factor of 1 -1.25 ( elastically). The steel member will be bigger but it may save the foundation cost.

 

[SHBH]

Thanks Zhuge, for giving a good suggestion.
Due to my limited knowledge on seismic design, one more thing i really need to know.
even if the pass ratio of the column id in the range of 0.6-0.8, do i need to design the foundation to cary = 1.25 * S * Fy.

 

[Zhuge]

yes. the reason is you need to ensure if it fails, it fails at steel member rather that at foundation. (it would be quite costly to repair).
I think design codes state that you can avoid using overstrength factor when you design the foundation as long as S*fy*overstrenght factor < design moment if you use elastic spectra.