Below - Grade Diaphragm Seismic Forces

[cal91]

We frequently have buildings with multiple levels below grade. The seismic forces to apply to the below grade elements is a common topic of debate.

Do you use Sds / (R / Ie) ?

Or do you use Sds / (Ie) ?

To me it is rational to not include R; the seismic forces are delivered from the earth, and these levels will directly experience the ground accelerations. Any ductility / energy dissipation of the structure above will not reduce the accelerations experienced by the basement levels.

Others argue that it is overly conservative because ASCE 7 does not provide provisions stating to do as I suggest, and that there will be zero amplification of the accelerations, only attenuation. They argue to use R of the building, or at least 0.5*R.

What are your thoughts / experiences? Does anyone have pertinent literature or code references?

Thanks!

 

[Deker]

Traditionally it is assumed that subterranean diaphragms do not experience differential displacement from the earth and therefore do not generate their own inertial forces. However, depending on soil conditions, this may not be appropriate. One common approach is to ignore the retained soil all around the building, set the seismic base to be at the lowest basement level, and model the entire building including the substructure. Below-grade diaphragms would be designed for 1) inertial forces consistent with the R factor used in the substructure and, perhaps more importantly, 2) any transfer forces induced by the backstay effect. I recommend amplifying the transfer forces by the overstrength factor as required by ASCE 7-16. Below are a few documents that should cover most of what you need to know:

Link #1
Link #2 (Section 6.2.8)
Link #3 (search "basement")

 

[cal91]

Deker,

Thank you for the response and the links.

I've heard that subterranean diaphragm are commonly assumed to not generate their own inertial forces. I don't understand how one can assume that being below grade prevents elements from accelerations due to earthquakes. Inertial forces are not generated from differential displacement from the earth. When the soil accelerates, the basement accelerates with it.

 

[WARose]

This kind of becomes a soil-structure interaction problem.

Also, if something is buried (deep), you may be wise (in some high seismic areas) to consider the forces developed form the various types of seismic waves hitting the widget (and the differential forces produced). I'm not sure if ASCE includes this yet (especially at what point it is to be considered).....but I checked a buried tunnel on a project some years ago and the forces produced were greater than any from just considering the soil-structure interaction.

A good resource on this is:

'Seismic Analysis and Design of Retaining Walls, Buried Structures, Slopes, and Embankments' (2008)

 

[DaveAtkins]

I don't understand how one can assume that being below grade prevents elements from accelerations due to earthquakes

Because if the structure does not move relative to the adjacent soil, no force is imparted to the structure due to differential sway.

DaveAtkins

 

[cal91]

Forces might not be imparted from the soil accelerating against the structure... but inertial forces are imparted through the structure accelerating with the soil. I.E. if the structure accelerates at all, there will be inertial forces. No?

 

[BridgeSmith]

The forces are imparted by the structure above ground trying to 'catch up' with the ground and the parts of the structure embedded in the ground that have all moved. The force is applied above ground and resisted by the soil against the walls below ground.

 

[WARose]

Forces might not be imparted from the soil accelerating against the structure... but inertial forces are imparted through the structure accelerating with the soil. I.E. if the structure accelerates at all, there will be inertial forces. No?

I think that is right. IIRC, some methods of calculating lateral forces on retaining walls include wall inertia effects.

 

[Deker]

I rationalize it this way: a large concrete box confined by competent soil all around is as close as we can practically get to rigid body motion, hence the acceleration experienced by the substructure will more or less match peak ground acceleration (around 0.4Sds, much less than the superstructure will experience). In addition, the inertial forces from the diaphragm have a direct axial load path to be resisted by the passive pressure of the retained soil. This will further reduce the effective inertial forces that get transferred to the in-plane basement walls. Of course if you were inclined to quantify how much gets resisted by passive pressure and how much gets transferred through diaphragm action your model would need to include soil structure interaction of the retained soil.