To: "INTERNET:seaint(--nospam--at)seaint.org" <seaint(--nospam--at)seaint.org>
Subject: Re: Overturning check
From: Mark Gilligan <MarkKGilligan(--nospam--at)compuserve.com>
Date: Sat, 12 Feb 2000 15:06:23 -0500
Continually engineers tend to assume that the design methodologies they
have learned reflect reality. The truth is that all of what we calculate
is a fantasy, albeit in most cases a useful fantasy.
It was stated by Jake:
>>>Why are we being penalized by overturing? Were there a lot of
overturning failures in Northridge?<<<
You are not being penalized by overturning. In the case of seismic loads
you only need consider a factor of safety or load factor of 1.0 since the
loads are not real. The actual seismic loads can be considerably more but
we have found that if you design to this criteria, overturning does not
seem to be a problem. For those structures that are limited by uplift
resistance this approach allows us to estimate the stability of a dynamic
non-linear system using relatively simple static analyses.
The example you showed used should have assumed a load factor of 1.0 (not
1.4) for earthquake induced forces when checking overturning. Overturning
under earthquake induced forces, like soil design should be checked using
working stress load combinations.
As Charles Greenlaw said it "... the static-force seismic code approach is
a mere contrivance, a convenient simplification that when originated,
catered to engineers' familiarity with wind design, whose loads are more
"As a contrived simplification, there are anomalies and sinkholes in the
array of concepts and sub-provisions the static seismic code bears. There
is no integrated, harmonized, seamless conceptual whole that is free of
While I typically prefer to use WSD I believe that there are some
situations where LRFD provides a better conceptual framework. As such I
would suggest that even if you use WSD you also consider how the answer
would be different when LRFD is used.
The LRFD approach explicitly recognizes that the code specified loads can
be considerably smaller or larger. As a result it will often do a better
job of identifying those situations where the design is particularly
sensitive to variations in loads. For example a WSD approach may infer
that you can stop your reinforcing at a certain point while LRFD suggests
that they should extend further. If the overload situation happened and
you had stopped the bars as suggested by WSD then your system could
possibly fail in an unanticipated mode and not have the factor of safety
I believe that one area that creates difficulties when using LRFD is the
checking of soil pressures. First ignore the temptation to require that
the resultant occurs in the middle third. This criteria was intended as a
way to insure that you had a reasonable factor of safety against
overturning. Since LRFD directly deals with this issue by means of load
factors you only need to assure that you will not have a soil failure.
This need to check "ultimate" soil capacities creates a problem in that the
Structural Engineer generally doesn't have a way to calculate the
"ultimate" soil capacities. This would require some agreement on stress
distributions that can be used and also would require the Geotechnical
Engineer to provide ultimate capacities and strength reduction factors for
each soil failure mode. You will likely find that the Geotechnical
Engineers will resist this change.
Given the variation in soil properties we might find that a LRFD approach
to checking soil capacities is more work than is appropriate. Because of
the variations in soil properties and the fact the Geotechnical Engineer
sets the criteria, I would suggest that the Geotechnical is the one who
ultimately determines whether there will be a failure in the soil. If the
calculations say I have a soil failure and the Geotechnical says it will
not fail my typical approach would be to document the discussion and to
assume I will not have a soil failure. Remember if the Geotechnical does
not like the final design he can change the criteria by modifying his
report. In this context the allowable soil stresses are only a way for the
Geotechnical to communicate with the Structural Engineer and thus changing
to an LRFD approach, for checking soil bearing, would result in little if
any change in the actual design.
If you accept the fact that there is more than one route to enlightenment
and try to understand the differences in the several approaches you will
probably be more successful in dealing with the use of LRFD in the
structure and WSD when checking soil capacities. It also helps to develop
an ability to believe in two mutually contradictory concepts at the same
Message text written by INTERNET:seaint(--nospam--at)seaint.org
> Let me throw another problem in here. If you use the LRFD loading
seismic, you end up with larger overturning. In many cases an ASD
approach would work out with no net overturning (stable) and with the
LRFD the system is unstable.
Case in point: say you have a 10 ft wide by 10 ft tall X-brace.
10 kips service shear horizontally at the top with 25 kips dead load
centered on the brace. The service O.T. moment is 100*kip*ft and the
dead load R.M. is 125*kip*ft, so (0.90)(125*kip*ft)-(1.0)(100*kip*ft) =
12.5*kip*ft (stable). In LRFD: (0.90)*(125*kip*ft)-1.4(100*kip*ft) =
-27.5*kip*ft (unstable). And this doesn't even include the vertical
acceleration reductions for LRFD that will generally place 0.90 DL
factor under 0.80.
So my point is, how do you design footing soil pressure with ASD
the footing with LRFD when the footing is no longer stable? Note this
only applies to seismic conditions, wind is a whole other topic. Also,
for those of you looking to move to LRFD wood, you will end up with many
more holdowns. When holdowns are used they will probably be smaller
(ultimate values from the anchors). In know LRFD is a four letter word,
but in this case it appears much more conservative than the ASD
approach. Why are we being penalized by overturing? Were there a lot
of overturning failures in Northridge?
One last thought. The material that would best be served by LRFD
approach is probably soil, yet I haven't heard a rumor on for it. Why?
Jake Watson, E.I.T.
Salt Lake City, UT<