Return to index: [Subject] [Thread] [Date] [Author]

RE: Foundation design for seismic

[Subject Prev][Subject Next][Thread Prev][Thread Next]
When you ask if the "foundation" should be designed for the capacity of the
superstructure, I presume that you are just talking about the concrete

The soil is a different issue in that soil is still designed for allowable
stress with the safety factors varying markedly depending on the consultant.

In regard to applying the ductility requirements on the concrete elements, I
agree that the concrete structural elements (including concrete foundations)
need to be designed using the special seismic provisions.

There are other areas that have somewhat ambiguous wording in the UBC and in
the ACI.  It is best to look at the ACI 318 comparable sections and look at
the commentary.  In this case you would look at the ACI 318 Section R21.2.2.
Basically it says that you need to use the special seismic provisions unless
you use a nonlinear response model.  Given that choice, I use the special
seismic provisions.

It is always good to keep in mind that the structure will go into the
nonlinear range during the design event.  Using that logic, I have been able
to convince engineers looking for loop holes that it is best to design with
the seismic detailing.

Harold Sprague
The Neenan Company

-----Original Message-----
From: T. Eric Gillham PE [mailto:gk2(--nospam--at)]
Sent: Tuesday, May 18, 1999 3:28 PM
To: seaoc list
Subject: Foundation design for seismic

Here is a question that has been on my mind for quite some time.  I have
asked other engineers knowledgeable in seismic design, and gotten differing

Here it is:

Does the UBC have in it language that either explicitly or implicitly
requires the foundation of a structure (say a mat foundation for example) be
designed for the CAPACITY of the superstructure?

Looking at UBC94, I have found some sections, notably 1921.2.1.4 and
1921.2.2.3, that together seem to indicate this is the case.

For the record, I feel strongly that the foundation SHOULD be designed to
either yield the superstructure, or as an less favorable option be designed
as a ductile foundation.  This only makes sense to me, as we are assuming
(usually) that the superstructure will be yielding during a major EQ, so the
use of reduced seismic forces (Rw=12 for a SMRF) will NOT give the expected
demands on the foundation.

For example, a SMRF supported by a mat foundation.  It seems logical that
the punching shear capacity of the mat should exceed the maximum expected
axial load in any of the columns, NOT just the code level axial loads
obtained from an elastic analysis.  Otherwise, the mat foundation is in
danger of failing in punching shear before the superstructure does all of
its work, and the design assumptions which were made would likely become
much less appropriate.

In the end, I really don't mind if the UBC doesn't explicitly require this,
as I do it anyway.  But, in some cases, where I am for example checking
another company's design, having code support becomes much more important.

Anyone have any thoughts on this one?

T. Eric Gillham PE
GK2 Inc.
PO Box 3207  Agana, Guam  96932
Email - gk2(--nospam--at)
Ph:  (671) 477-9224
Fax: (671) 477-3456