Академический Документы
Профессиональный Документы
Культура Документы
INTRODUCTION
The principles of design of steel structures for strength are well known, in particular
as they are presented in design specifications such as the AISC Specification for allowable
stress design or the recently issued Load and Resistance Factor Design Specification (1,2).
The criteria that are given by these specifications are well-defined formulations that are
based on accurate or simplified strength models, and the underlying limit states are in most
cases clearly definable characteristics of member or structural behavior. Some typical ex-
amples are tension member gross cross section yielding, laterally supported beam yielding
in the extreme fibers (yield moment) or reaching the fully plastic moment, or column buck-
ling, to mention a few prominent cases. Common to all of these is the fact that given the
loading or internal moment condition, the corresponding response of the member or struct-
ure can be determined and understood; the results are relatively unambiguous.
The design criteria are considerably less well defined when it comes time to deter-
mine how the structure or any of its parts will respond under service loads, and even more
so if it is required to give an assessment of how the occupants of the structure will per-
ceive the response of the building to a variety of loading situations. However, it is clear
that such knowledge should be a part of the designer's package of criteria that are to be
used in evaluating the design for future performance. Now commonly referred to as service-
ability criteria or parameters that express how the structure will behave under service or
working loads, it appears that there are no unified conditions that can be or are used by
"all" designers, basically because the understanding of how structures behave and how to de-
termine this behavior vary from engineer to engineer. In other words, the loads and the
structural models that are formulated may differ from firm to f i r m and from individual en-
gineer to individual engineer.
Among the serviceability criteria, the deflection limitations for gravity loads seem
to be generally understood and accepted, although their backgrounds are not always obvious.
Thus, the well-known " L / 3 6 0 " limitation for the live load deflection of a bending member
in an office or similar type building is widely used. However, there is widespread confusion
about how it was developed, prompting some designers and textbooks to apply the limit to
all gravity load deflections. On the other hand, designers of a variety of structures are cog-
nizant of the fact that it is quite reasonable to place less restrictions on the expected de-
flections in a structure that will be used only rarely for human occupancy, not to mention
what can be tolerated by the structure if no human usage is intended. Thus, designers of
industrial structures and components routinely utilize much more liberal gravity deflection
criteria.
Responding to a universally felt need to establish rational procedures and the factors
that influence the response of steel structures to lateral loads, as well as to determine if
7-2
© 2003 by American Institute of Steel Construction, Inc. All rights reserved.
This publication or any part thereof must not be reproduced in any form without permission of the publisher.
it is possible or even advisable to specify response criteria for structures under lateral loads,
members of the American Society of Civil Engineers Committee on Design of Steel Build-
ing Structures, chaired by the author, set out to examine current procedures and techniques
for lateral load evaluation of steel-framed building structures. The members also intended
to provide their own evaluation of currently used versus needed criteria, with the aim of
giving the profession a cohesive reference document on lateral load serviceability of steel
structures.
It was decided at an early stage that it would be necessary to separate the consider-
ations that applied to wind loads and seismic loads. In the following only wind load service-
ability is dealt with; seismic criteria are not addressed.
A Task Committee was selected from the main committee to carry out the investi-
gation. The effort was chaired by James M. Fisher of Computerized Structural Design, Mil-
waukee, with the following as members: Horatio Allison, Allison, McCormac & Nickolaus,
Rockville, Maryland; William F. Baker, Jr., Skidmore, Owings & Merrill, Chicago; Reidar
Bjorhovde, University of Arizona; Robert O. Disque, AISC, Chicago; Lawrence G. Griffis,
Walter P. Moore & Associates, Houston; Carl W. Hatmaker, TVA, Knoxville; Enrique Mar-
tinez Romero, EMRSA, Mexico City; Donald R. Strand, Brandow & Johnston, Los Angeles;
and Kenneth B. Wiesner, LeMessurier Consultants, Cambridge.
Briefly, the survey aimed at finding the approaches that are in current use, and the
questions that were asked can be grouped into the following primary categories:
(1) Should wind drift criteria be codified, and in what form?
(2) Height of buildings where drift control becomes a consideration.
(3) Methods of drift analysis and the input parameters that are used.
(4) Overall versus interstory drift values.
(5) Influence of structural system, exposure, wind speeds, etc.
(6) Influence of building code and corresponding wind loads.
(7) Dynamic versus static wind drift criteria; use of wind tunnel tests.
In addition, the survey respondents were asked to provide wind drift data for a number of
buildings that they had been responsible f o r , as well as to give their drift limits and asso-
ciated wind load criteria for a given type of structure.
7-3
The following descriptions are limited to paraphrasing the comments and other an-
swers that were offered by the survey respondents. More detailed discussions and evaluations
are provided in the forthcoming ASCE committee report (3). The topics are those that were
used in the survey questionnaire, as indicated by the Appendix.
Although not representative to the survey of current practices of wind drift design,
the committee was interested in establishing the opinions of the structural engineering com-
munity insofar as drift criteria codification were concerned. Many pros and cons have been
voiced by various designers; the survey responses indicated that although a majority of the
engineers would like to see some form of codified drift criteria, almost all voiced a need
for detailed qualifications of same. Thus, it would appear that although it is considered an
advantage to be given the limits of drift, at the same time most engineers recognize that
the problem is considerably more complicated than that. In other words, it is not sufficient
to prescribe a drift index without at the same time also giving a number of conditions that
apply for the particular limit. For example, several respondents indicated that it is necess-
ary to tie drift values to the type of human occupancy of the building, type of cladding, and
so on. A large number of engineers also noted that they regarded drift limitations as the
purview of the designer, since only he/she is fully conversant with the performance require-
ments of the individual structure. In other words, these designers felt that codification has
the potential of being too restrictive or limiting, especially towards new analysis approaches
or methods of construction.
It is of some interest to note that several designers appeared to "want to have their
cake and eat it, too" : while at the same time requesting rigorous drift limitations within
the codes, presumably for the protection of the public, these engineers also preferred to be
left free to use their own approaches to determine drift values. Clearly, such opinions are
contradictory.
Somewhat surprising, the responses of the survey engineers were almost equally di-
vided between "often", "sometimes", and "never". It was only surprising because a two-level
drift approach implies a level of design consideration that is quite advanced, and reflects a
clear understanding of what governs strength and serviceability. These designers clearly in-
dicate the proper understanding that the wind loads that are used to size the members of
the structure require a different wind load treatment than when the sway is computed. The
philosophy is in agreement with current thinking among load study experts (4).
This portion of the questionnaire specifically asked the designers to indicate when
© 2003 by American Institute of Steel Construction, Inc. All rights reserved.
they would start considering
This publicationchecking drift,
or any part thereof asbeareproduced
must not function in any of
form the
withoutheight
permissionof the
of the building. It
publisher.
was interesting to note that many of the respondents indicated that they felt that drift
should be computed regardless of the height of the structure. This is in agreement with the
opinions of the committee, reflecting the need to clearly establish the lateral load resistance
capacity of the structure through a positive system.
It was deemed very important to establish the basic parameters for the analysis of
the bare steel frame, recognizing that this was the frame that formed the basic one for the
7-4
evaluation of the frame response (i.e. not influenced by cladding type and similar elements).
The final results are clearly influenced by a number of geometric and other parameters, and
the following items were addressed:
(a) Influence of Shear Deformations of Beams and Columns: About one half of the re-
spondents indicated that they included shear deformations in their drift analyses. It is felt
that the response reflects a misunderstanding of what shear deformations are, and that the
answers rather indicated that the computations incorporated panel zone deformations.
(b) Panel Zone Deformations: About two thirds of the answers were "no", and of the
remainder, some of the responses indicated that the panel zone deformations were only tak-
en into account when certain sizes of the zones were encountered, or when certain systems
were used for the structure (such as large panels, tube-type structure, etc.).
(c) Axial Deformations of the Columns: As expected, a large percentage of the engi-
neers indicated that they always incorporated the effects of column axial deformations. Some
reservations were expressed about using this for very low-rise frames, for example, but the
respondents generally exhibited good appreciation of this effect.
(d) Width of the Beam-to-Column Joints: Most answers were in the affirmative, al-
though some were in contradiction to the answers given under item (g). Clearly, one cannot
at the same time base these computations on both center-to-center dimensions and include
the effect of the width of the joints.
(e) Second Order Analysis: More than two thirds indicated that they run second order
analyses as a matter of course, although a number of qualifying statements were made. On
the other hand, about 40 percent said that they would not do such evaluations for 1 to 3
story buildings, making no reservations in this regard.
(f) Connection Flexibility: Approximately two thirds of the answers indicated that no
account were made for connection flexibility, as expected. For the number that answered in
affirmative, it is believed that they did not understand the nature of the question. Clearly,
it is extremely rare to see that such effects are taken into account in the design of struct-
ures today, even by the most advanced design firms.
(g) Center-to-Center Dimensions: A majority said that they considered only the center-
to-center dimensions of the frames. This is in contradiction to the responses given to item
(d), as discussed above.
In a broad question, the designers were asked for their opinions as regards the use of
allowable drift limitations, and how the latter could be determined. Specifically, the respon-
dents were asked for their thoughts on the influence of the structural system for the build-
ing, the building use and height, the cladding type, the design wind speed, the terrain expo-
sure, the method of drift calculation, and the use of probable versus maximum expected
winds. Briefly, most answered "no" to the question regarding the use of a drift that was a
function of the type of structural system, and a majority also answered in the negative in
response to the influence of terrain exposure. The answers were about evenly divided be-
tween "yes" and "no" on the© 2003 question regarding
by American the
Institute of Steel method
Construction, ofrights
Inc. All calculating
reserved. the drift, presum-
This publication or any part thereof must not be reproduced in any form without permission of the publisher.
ably considering different methods and levels of accuracy in the frame analyses. The majo-
rity answered "yes" to all of the other questions, including, as expected, the effect of build-
ing height, cladding, building use, and design wind. The positive responses also incorporated
the answers to item (h) on the questionnaire, dealing with a two-level design approach. This
would appear to be in agreement with the responses to the item dealing with design wind.
This part of the questionnaire focused on the calculations of story drifts, including
the special criteria that are sometimes used for first stories of buildings. There appeared
to be no general agreement with respect to either of the two parts of the question, and the
impressions was left that story drift values essentially were chosen as close to the overall
structural drifts. A few of the answers indicated that they recognized that the first stories
7-5
of some buildings exhibit large drift values; special considerations are used in such cases.
However, there seemed to be little recognition of the fact that story drifts in some types
of framing systems tend to be much larger than the average (overall) building drift, which
is what most designers use in their evaluations.
An important part of the survey addressed the level of wind loads that was pre-
scribed by the various model codes that are used in the United States today. As illustrated
by Question 8 in the survey questionnaire (see Appendix), the respondent was asked to indi-
cate a variety of preferred drift ratios for different types of structures, occupancies, loca-
tions, commercial versus industrial, and so on.
The responses to this questions varied a great deal, as expected, although it is inte-
resting to note that the most frequently occurring drift ratio for all types of structures and
model codes was 0.0025 (= building height/400), and the minimum and maximum values were
0.0015 and 0.003, respectively. Other than that, the data are not very conclusive, other than
to show that there is a distinct trend - as expected - to use lower drift values for residen-
tial buildings than commercial, for obvious reasons. There are no clearcut differences be-
tween the different model codes.
Two thirds of the engineers indicated that they felt that dynamic criteria should be
used for establishing lateral vibration characteristics. On the other hand, a large percentage
mentioned that they had never utilized such criteria in practice, but rather felt that the
dynamic response would be a better measure of the structures' characteristics. For those of
the respondents that had used the dynamic solutions, accelerations of 10 to 30 m-g for a ten
year return period seemed to be the norm.
It is clear that the state-of-the-art of dynamic lateral vibration criteria for building
structures is not very advanced, and much research needs to be conducted.
This portion of the questionnaire addressed the possible uses of wind tunnel tests to
establish wind pressure distributions on structures, as well as vibration characteristics for
the structures (aeroelastic studies) and the overall wind forces. Fifty percent of the engi-
neers indicated that they had utilized wind tunnel tests of some form at one time or other,
although mostly to establish the wind loads for complex structures in heavily built urban
areas. It appears that more tests have been run to determine wind pressure distributions (ri-
gid models) than vibration characteristics (dynamic models). Most respondents indicated that
they would routinely specify wind tunnel testing if there were anything unusual about the
building and its location, or if it were of great height. Other considerations that were men-
tioned were large aspect ratios for the building (height to width greater than 4 to 1 or 6 to
1), unusual structural geometries and exposure conditions.
7-6
© 2003 by American Institute of Steel Construction, Inc. All rights reserved.
This publication or any part thereof must not be reproduced in any form without permission of the publisher.
It is apparent that a concerted effort should be undertaken for industrial buildings as well
as for residential/commercial structures insofar as drift considerations are concerned. It is
somewhat surprising that many designers did not indicate concern for building stiffness
under service conditions that can be very severe, and demand a great deal from the res-
ponse of the structure itself.
To illustrate the approaches of the various respondents to drift analysis and design, they
were asked to indicate their preferred drift limitations for a 30 story, 380 foot high multi-
story frame with lightweight exterior cladding, along with the average wind pressure level
for the projected area of the building (this is given as Question 12 in the questionnaire, as
shown in the Appendix). Analysis of the responses showed that more than one half of the
engineers would use a drift index of 0.0025 (= H/400) for wind pressures of 20 to 35 psf. The
answers ranged between low: and high values of 0.0015 and 0.0050.
SUMMARY
The paper has presented the results of a survey of practicing engineers in the United
States, examining their preferred wind drift analysis approaches and associated considerati-
ons. It is shown that by and large, a drift index of 0.0025 appears to be the most common
value, with extremes of around 0.0015 and 0.005, depending on the type of structure, the
level of the wind loading, structural geometry, and so on. It is evident that no cohesive
approaches are currently in use in the US; it is also clear that much misunderstanding per-
sists as regards the needs/requirements for lateral load analysis.
REFERENCES
1. American Institute of Steel Construction (AISC), "Specification for the Design, Fabri-
cation and Erection of Structural Steel for Buildings", AISC, Chicago, Illinois, Novem-
ber 1, 1978.
2. American Institute of Steel Construction (AISC), "Specification for the Load and Re-
sistance Factor Design, Fabrication and Erection of Structural Steel for Buildings",
AISC, Chicago, Illinois, January, 1986.
3. Task Committee on Drift Control of Steel Building Structures, ASCE, "Wind Drift
Criteria for Steel-Framed Buildings", Committee Report in Preparation, February,
1986.
7-7
© 2003 by American Institute of Steel Construction, Inc. All rights reserved.
This publication or any part thereof must not be reproduced in any form without permission of the publisher.
DRIFT CRITERIA QUESTIONNAIRE
Other
(Specify)
7-8
© 2003 by American Institute of Steel Construction, Inc. All rights reserved.
This publication or any part thereof must not be reproduced in any form without permission of the publisher.
5. In your bare-frame lateral-deflection calculations, do you consider:
c. Axial deformation of
columns?
d. Width of beam-to-column
joints?
e. Second-order drift
analysis?
f. Connection flexibility?
g. Center to Center
dimensions only?
YES NO COMMENTS
b. Building Use?
c. Building Height?
d. Cladding Type?
f. Terrain exposure
g. Method of calculating
drift?
7-9
© 2003 by American Institute of Steel Construction, Inc. All rights reserved.
This publication or any part thereof must not be reproduced in any form without permission of the publisher.
8. Check the drift ratio (total drift/total height) for wind which you
would use for each of the categories below. Indicate on what wind
code you are basing your answers (e.g. ANSI A58.1-1982, UBC, BOCA)
CODE: __________
*CLADDING: "Heavy" would reduce actual drift (e.g. masonry precast concrete,
brick, stone, granite). "Light" represents light curtain walls
(e.g. metal stud facia, marblecrete, artcraft, etc.). __
Do you design for drift for a 100 year recurrence level?
7-10
© 2003 by American Institute of Steel Construction, Inc. All rights reserved.
This publication or any part thereof must not be reproduced in any form without permission of the publisher.
9. Please provide wind design criteria and performance information for actual buildings that you or your firm has designed,
NO. BLDG. BLDG. HEIGHT ASPECT RATIO BASIC WIND RECURRENCE CODE, YEAR LOCATION CLADDING
TYPE FEET STORIES (HEIGHT/WIDTH) SPEED (10 yr, 50 yr) (UBC, BOCA CODE TYPE
(1) (MPH) @30' ETC.) ANSI, ETC.) (2)
YES Explain:
NO
YES Comments:
NO
11. (a) Have you or your firm ever had building designs model-tested in
a wind tunnel?
YES
NO
Rigid Dynamic
(c) Under what circumstances would you specify a wind tunnel study
using a rigid wind tunnel model and an aeroelastic model?
(Building height, cross section, location, aspect ratio, etc.)
7-12
Drift Index
13. Considering all the foregoing questions, state what you believe to be
a drift criteria recommendation that should be presented as guidance
to the profession.
7-13