To: "INTERNET:seaint(--nospam--at)seaint.org" <seaint(--nospam--at)seaint.org>
Subject: Re: Dynamic Analysis Question
From: Mark Gilligan <MarkKGilligan(--nospam--at)compuserve.com>
Date: Fri, 14 Jan 2000 01:35:44 -0500
Assuming that you have used reasonable modeling practices in all of your
models. The combined model should give more realistic answers. In fact
given programs such as Etabs and Sap I see little advantage in modeling
them separately given the added work in dealing with two models and the
less realistic results from separate models.
All computer models are approximations and require the analysts to
understand the way the program works as well as understand dynamics. In
addition you need software that you can trust. It seems that many of the
warnings provided relate to concerns about questionable software and poor
modeling practices. Modeling the structure as two separate parts will not
automatically improve the reliability of the software nor will it
automatically guarantee that your modeling assumptions are any better.
It was stated:
>>As one might expect when the building is analyzed as whole, a substantial
portion of the forces induced by the base is getting dumped into the tower
and penalizing the tower frames.<<
When trying to analyze a dynamic analysis it is a good idea to not think in
terms of forces induced at the base being dumped at certain locations. To
use those terms is to try to explain a dynamic analysis in terms of a
static analysis. I find it useful to think of the structure vibrating in
its natural modes and as a result of the induced deformations in the modes,
forces are induced in the members.
We also have a tendency to use loaded terms like "penalizing the tower
frames" that implies that the results are un-necessarily conservative.
Could it be that the other analysis is unconservative? Just because we
have designed many buildings using the results of static analyses doesn't
make them correct nor on the other hand am I implying that the buildings
designed using static analyses are unsafe.
What aspect of the loading or properties of the frame are causing these
members in the tower to have higher forces? Look at the mode shapes that
are producing these higher forces. One of the benefits of a dynamic
analysis is that it can give you a better understanding of the structure,
if you will study it. In addition as a result of studying the modes you
will often find errors in the model that are much harder to find in a
>> Now let me provide you some additional relevant information, or
maybe the second part of the question. The dynamic analysis had revealed
that the dominant modes for upper tower and lower podium were dramatically
different. Assume the decision is made to use the distribution of forces
the model of building as a whole. When it comes time to scale the forces
back to the code level forces what do you scale it to? And would you scale
the forces for each section separately? According to which period?<<
Find the period of the first mode for the total structure, using your
combined model, and then calculate the code level base shear. You would
then perform a response spectrum analysis using the combined model. The
results from the combined model can then be scaled so that the base shear
from the response spectrum analysis matches the code level based shear.
When you do this you will likely find that there are smaller forces and
story drifts in the upper floors.
I would like to understand what was meant by the statement " In summary,
doing it in parts is better considering most of today's software
capabilities." My experience is directly the opposite.