full text
TRANSCRIPT
RESEARCH ARTICLE
Wai Fah CHEN
Advanced analysis for structural steel building design
E Higher Education Press and Springer-Verlag 2008
Abstract The 2005 AISC LRFD Specifications for
Structural Steel Buildings are making it possible for
designers to recognize explicitly the structural resistance
provided within the elastic and inelastic ranges of beha-
vior and up to the maximum load limit state. There is an
increasing awareness of the need for practical second-
order analysis approaches for a direct determination of
overall structural system response. This paper attempts
to present a simple, concise and reasonably comprehens-
ive introduction to some of the theoretical and practical
approaches which have been used in the traditional and
modern processes of design of steel building structures.
Keywords advanced analysis, structural steel building
design, introduction
1 Introduction
The purpose of structural design is to produce a physical
structure capable of withstanding the environmental con-
ditions to which it may be subjected. Many factors affect
the design process, from loading to foundation to dimen-
sion lay out to risk and cost, but basically the ultimate
design is a reflection of the properties of the structural
material and the geometrical imperfection of its structural
members, and in particular its mechanical properties and
the residual stresses induced in the structural members
during manufacturing and fabrication, which define the
characteristic response of the material and the member to
the environmental forces.
At present, in engineering design practice, there is a
fundamental two-stage process in the design operation:
firstly, the forces acting on each structural member in
the structure must be calculated; secondly, the load car-
rying capacity of each of these structural members to
those forces acting on it must be determined. The first
stage involves an analysis of the distribution of forces
and moments acting on each of these structural members;
the second stage involves knowledge of the load carrying
capacity of these members to resist these forces and
moments acting on them. The more comprehensive this
knowledge, the more exact will be the design and the more
reliable will be the structure.
Since the load carrying capacity of structural membersdepends on the type of loads acting on the member, geo-
metrical imperfections, properties of material and residual
stresses, the knowledge of load-carrying capacity of these
structural members has been determined mostly on the
basis of full scale tests in the form of pin-ended column
strength curves for axially loaded members, simply sup-
ported beam strength curves for bending dominatedmem-
bers, and beam-column interaction curves for membersunder combined axial force and bending moment. These
member strength curves are formally coded as the member
strength curves or equations for design practice.
Having divided structural members in a framed struc-
ture into three classes, namely columns, beams, and beam-
columns, and determined their respective strengths by full
scale tests with ideal end or boundary conditions, the next
stage must be to drastically simplify the material behavior
under stress in such a way as to readily assist the engineerin analyzing the stress distribution in the structure in order
to size up the structural members in a framed structure. At
present, the practicing engineer bases design primarily on
the simple model of linear elasticity of the material for
those early designs based on the allowable stress method.
In this process, time-dependent effects of material are
assumed insignificant. This is indeed a drastic simplifica-
tion of material properties over long term behavior. Sowith this time independent simplification, the design pro-
cess is now focused and concentrated on reducing the
stress level or the structure is loaded only in the working
load level. The large safety factor is therefore used to
adjust the design to take into account inelastic features
in the material to avoid failure. Most structural analyses
in engineering practice have been based on linear elastic
analysis. First-order linear elastic analysis has been thehallmark of structural engineering in early years, while
the second-order linear elastic analysis for a structural
Received December 13, 2007; accepted March 10, 2008
Wai Fah CHEN (*)Department of Civil Engineering, University of Hawaii at Manoa,Honolulu, HI 96822, USAE-mail: [email protected]
Front. Archit. Civ. Eng. China 2008, 2(3): 189–196DOI 10.1007/s11709-008-0024-8
system has been developed and is increasingly being uti-
lized in recent years.
2 First-order elastic structural analysis withK-factors
The boundary conditions of a framed member in a frame
structure are quite different from that of an isolated mem-
ber that is used as the basis for the development of column
strength curves (pin-pin end conditions) or beam strength
curves (simply supported end conditions). In order to size
up the framed members, the framed member boundary
conditions must be adjusted to the equivalent pin-pin
end conditions for the case of column design, for example,
so that the column strength curves can be properly used in
determining the required size of the framed member under
consideration (see Fig. 1).
To achieve this equivalency, the effective length fac-
tor or the K-factor has been widely used in the past to
relate the pin-ended column strength curves to the
framed member design in a structural system. The effec-
tive length method generally provides a good method
for the design of framed structures. This method has
been widely used for the development of modern steel
design codes including the allowable stress design and
the plastic design in early years; and the load and res-
istance factor design in more recent years [1]. However,
despite its popularity, the approach has several major
limitations and shortcomings.
The first of these is that it does not give an accurate
indication of the factor against failure, because it does not
consider the interaction of strength and stability between
the member and the structural system in a direct manner.
It is a well recognized fact that the actual failure mode of
the structural system often does not have any resemblance
whatsoever to the elastic buckling mode of the structural
system that is the basis for the determination of the effec-
tive length factor K.
The second, and perhaps the most serious limitation, is
probably the rationale of the current two-stage process in
design: elastic analysis is used for the determination of
distribution of forces acting on each member of a struc-
tural system, whereas the member’s ultimate strength
curves are developed for design either on the basis of full
scale tests or by inelastic analysis with each member
treated as an isolated component [2,3]. There is no veri-
fication of the compatibility between the isolated member
and the member as part of a frame. The individual mem-
ber strength equations as specified in specifications are
not concerned with system compatibility. As a result,
there is no explicit guarantee that all members will sustain
their design loads under the geometric configuration
imposed by the framework.
The other limitations of the effective length method
include the difficulty of computing a K factor, which is
not user-friendly for computer-based design, and the
inability of the method to predict the actual strength of
a framed member, among others. To this end, there is an
increasing tendency of the need for practical analysis/
design methods that can account for the compatibility
between the member and system. With the rapid develop-
ment of computing power and the availability of desk-top
computers and user-friendly software, the development of
an alternative method to a direct design of structural sys-
tem without the use of K-factors becomes more attractive
and realistic.
3 Second-order elastic structural analysiswith K-factors
The adoption of elastic structural analysis with K-fac-
tors for steel design may be divided into two stages of
progress. The simplest first stage of progress with K-
factors in the design process is the first-order elastic
analysis with amplification factors to include the sec-
ond-order effects as generally provided by the specifi-
cations [4]. This is described in the preceding section.
Logically, the next stage of progress is a direct second-
order elastic analysis without the use of amplification
factors for second-order effects [4]. Both methods are
based on the formation of first plastic hinge defined as
the failure of the system (see Fig. 2).
As mentioned previously, the effective length factor will
generally yield good designs for framed structures, but it
does have the following drawbacks:
Fig. 1 Interaction between structural system and its com-ponent members
190 Wai Fah CHEN
1) It cannot capture accurately the interaction between
the structural members and the structural system behavior
and strength;
2) It cannot reflect the proper inelastic redistributions
of internal forces in a structural system;
3) It cannot predict the failure modes of a structural
system;
4) It is not easy to implement in an integrated computer
design application with the use of alignment charts in the
K-factor calculation process;
5) It is generally a time-consuming process requiring
separate member capacity checks with different K-factors
for different framed members.
Even for the most recent AISC-LRFD procedures, sim-
ilar difficulties are encountered when performing a seismic
design because the same strength interaction equation
checks must be performed [4]. Some of the difficulties
are even more so on the seismic designs since additional
questions are raised such as:
1) How is the structure going to behave during an earth-
quake?
2) Which part of the structure is the most critical area?
3) What will happen if part of the structure yields or
fails?
4) What might happen if forces greater than those spe-
cified in the code occur?
None of these questions can be answered by the con-
ventional LRFD method with K- factors. In contrast, the
advanced analysis to be described in the following can
provide this information (see Fig. 3).
4 Various advanced analysis methods
Advanced analysis refers to any method that captures the
strength and stability of a structural system and its indi-
vidual members in such a way that separate member capa-city checks are not required. Thus, there is no need for K-
factor calculation. Usually these analyses are also referred
more formally as the second-order inelastic analysis for
frame design (see Fig. 2). With the present computing
power, it is a rather straight forward process to combine
the theory of stability [5] with the theory of plasticity [6]
for structural system analysis [7]. The real challenge is
therefore to make this type of new approach to designwork and competitive with the current methods in engin-
eering practice [8].
4.1 Elastic-plastic hinge method
The elastic-plastic hinge method is the simplest approxi-
mation of inelastic behavior of material by assuming all itsinelastic effect is concentrated at the plastic hinge loca-
tions. In this idealization, it assumes that the element
remains elastic except at the ends where zero-length plas-
tic hinges can form. This method accounts for inelasticity
but not the spread of yielding or plasticity at sections, or
the influence of residual stresses.
Here, as in previous sections for elastic analysis,
depending on the geometry used to form the equilibriumequations, the elastic-plastic hinge method may be divided
into first-order and second-order plastic analysis. For the
Fig. 2 General analysis types for steel building structure design
Advanced analysis for structural steel building design 191
first-order elastic-plastic hinge analysis, the undeformed
geometry is used, and nonlinear geometric effects are
neglected. As a result, the predicted ultimate load is the
same as conventional rigid-plastic analysis. In the second-
order elastic plastic analysis, the deformed shape is con-
sidered and geometric nonlinearities can be included using
stability functions which enable use of only one beam-
column element per member to capture the second-order
effects. A comprehensive presentation of the plastic design
and second-order analysis method can be found in the
book by Chen and Sohal [9].
The second-order elastic-plastic hinge analysis is only
an approximate method [10,11]. For slender members
whose dominant failure mode is elastic instability, the
method provides a good approximation; but for stocky
members as well as for beam-column elements subjected
to combined axial load and bending moment, this method
overestimates the actual strength and stiffness in the inel-
astic range due to spread of yielding effects.
This method is therefore a good first approximation of
the second-order inelastic analysis for frame design within
the applicable range described above. It requires further
refinement before it can be recommended for analysis of a
wide range of framed structures [12].
4.2 Refined plastic-hinge method
The refined plastic-hinge methods are based on some sim-
ple modifications of the elastic-plastic hinge method
described above. The notional-load concept is first intro-
duced to the conventional elastic-plastic hinge method by
applying additional fictitious equivalent lateral loads to
account for the influence of residual stresses, member
imperfections, and distributed plasticity that are not
included in the conventional procedures. With certain
modifications, this refined approach is accepted in the
European Convention for Construction Steelwork [13],
the Canadian Standard, and the Australian Standard.
However, Liew’s research [10] shows that this method
under-predicts the strength in the various leaning column
frames by more than 20% and over-predicts the strength
by up to 10% in the isolated beam-columns subjected to
axial forces and bending moments.
In the following, we shall make further modifications
and simplifications of the elastic-plastic hinge method to
improve its performance and at the same time to make it
practical and work in engineering practice. These modifi-
cations are grouped into three categories: geometry,
material, and connection. Details of these modifications
can be found in the two doctoral theses [10,14,15] and
their subsequent papers (for example, Refs. [16,17]).
This refined plastic-hinge method will now be called the
practical advanced method for frame design in what fol-
lows.
4.3 Plastic-zone method
The plastic-zone method is considered to be ‘‘exact’’ since
it is based on the most refined finite element analysis for a
structural system. This plastic-zone analysis has been well
developed and documented by a research team at Cornell
University over the last two decades under the leadership
of Professor McGuire [18]. The team members include
Ziemian [19], White [11], Attala and Deierlein [20], among
others. As a comparison, the elastic-plastic hinge model is
considered to be the simplest; while the elastic-plastic-
zone model exhibits the greatest refinement.
In the plastic-zone method, each member is discretized
into many sections along the length and each segment is
subdivided into many finite elements. The material
Fig. 3 Analysis and design methods: conventional vs advanced analysis/design
192 Wai Fah CHEN
properties of each finite element are specified for specific
properties and residual stresses. The behavior of the mem-
ber is obtained by numerical integration of these finite ele-
ments. The second-order geometric effects are captured by
updating the geometry at each incremental load step. The
exact solutions were verified with some benchmark beam-
column tests. This method is ideal for verification of various
simplified methods developed for engineering practice.
5 Practical advanced method
The practical advanced method was developed at Purdue
University over the last 15 years beginning with the work
of Liew [10], verified with the well-documented cal-
ibration frames [7,21], and completed with the adoption
of the advanced analysis method in the 2005 AISC Load
and Resistance Factor Design Specification for Structural
Steel Buildings. The details of this development were sum-
marized in the 1997 book ‘‘LRFD Steel Design Using
Advanced Analysis’’ by Chen and Kim [22]. Herein, as
in our previous experiences we learned from engineering
practice, the practical advanced analysis method must be
reduced to a sequence of equivalent linear elastic analysis
as was the case for the widely popular allowable stress
design in early years as well as the elastic-plastic hinge
design in later years in order to gain acceptance in engin-
eering practice for general use.
In the following, I shall briefly summarize the highlights
of this development in terms of idealizations and simpli-
fications of geometry, material and connection to achieve
this goal while meeting the current LRFD requirements.
In this process of development, it initially sets out to use
the familiar but simplified stability functions to capture
the second-order effects, going on to show how the spread
of plasticity due to residual stresses can be accounted for
with a reduction of material tangent modulus, and finally
outlining the influence of geometric imperfections that
may be simulated by a further reduction of tangent modu-
lus in the ultimate processes of design of structural system
with an equivalency of simple elastic analysis.
5.1 Second-order effects
To capture the second-order effects, the simplified
stability functions reported by Chen and Lui are adopted
[23].
5.2 Cross-section plastic strength
The LRFD cross-section plastic strength curves are
adopted for both strong and weak-axis bending (see
Fig. 4) [5]. The same reduction factors used in the
LRFD specification are selected as 0.85 for axial strength
and 0.9 for flexural strength.
5.3 Residual stresses
The CRC tangent modulus is employed to account for the
gradual yielding effect due to residual stresses along the
length of members under axial loads. In this approach, the
elastic modulus E instead of the moment of inertia I is
reduced to account for the reduction of the elastic portion
of the cross-section, because the reduction of elastic
modulus is easier to implement than that of the moment
of inertia for different sections. The reduction rate in stiff-
ness for both strong and weak axis is taken to be the same
as reflected by the CRC tangent modulus curve Et (see
Fig. 5).
5.4 Geometric imperfections
The degradation of member stiffness due to geometric
imperfections may be simulated by a further reduction
of member stiffness. This may be achieved by a further
Fig. 4 Smooth stiffness degradation for work-hardeningplastic hinge based on LRFD sectional strength curve
Fig. 5 CRC and reduced tangent modulus, Et, for memberswith residual stresses and geometric imperfections
Advanced analysis for structural steel building design 193
reduction factor of the tangent stiffness E0t~0:85Et. This
further reduction modulus is applicable for both braced
and unbraced members and frames (see Fig. 5).
5.5 Semi-rigid connections
The actual behavior of a semi-rigid connection in a struc-
tural steel building falls between the two extremes of
pinned-joint and rigid-joint. These two extreme idealiza-
tions have been widely used in the past in conventional
structural analysis and design. In recent years, attention
has been focused toward a more accurate modeling of
such connections (see Fig. 6). To this end, the 1994
AISC-LRFD Specifications designated two types of con-
struction in their provisions: Type FR (fully restrained)
and Type PR (partially restrained). The Type FR is the
traditional rigid connection; while the PR Type is known
as the semi-rigid connections in the conventional termino-
logy of structural steel analysis and design. If Type PR
construction is used, the effect of connection flexibility
must be taken into account in the analysis and design
procedures (see Fig. 7).
Currently, the most commonly used approach to
describe the moment-rotation curve of a semi-rigid con-
nection is to curve fit the experimental data with simple
expressions based on large data base [24–26]. Several
curve fitted semi-rigid connection models have been pro-
posed but the most notable is the three-parameter power
model proposed originally by Richard [27], modified by
Kishi and Chen [28] and Kishi et al. [29], and refined by
Wu [30].
The influence of partially restrained connections on
structural analysis is not only in changing the moment
distribution among beams and columns, but also in
increase frame drift and accordingly the increase of the
P-delta effect in the frame analysis. Extensive work has
been made in recent years on connection modeling [31],practical procedures for implementation [32], and guide-
lines on proper selection of model parameters for PR con-
struction [33]. Details of these developments can be found
in the comprehensive books by Chen [34], and by Chen,
Goto and Liew [32], among others.
A one-day workshop on frames with partially
restrained connections was organized and conducted by
the Structural Stability Research Council [35] in Atlanta,Georgia on September 23, 1998. The subsequent work-
shop proceedings covered various aspects of the state-
of-the-art progress on the practical analysis for partially
restrained frame design. This included codes, database,
modeling, classifications, analysis/design, and design
tables and aids. A comprehensive design guide with design
examples was also provided in a companion SSRC book-
let entitled ‘‘Practical Analysis for Partially RestrainedFrame Design’’ by Chen and Kim [33].
6 Practical advanced analysis for seismicdesign
Based on the response of structures in the Northridge
earthquake in 1994, significant changes were made to
the 1997 LRFD seismic provisions [36], especially the
beam-to-column connection design in moment frames
[37]. These changes are conceptually trying to achieve
both ‘‘structural fuse concept’’ and ‘‘performance-based
design’’ (see Fig. 8). Structural fuse concept was designedfor controlled failure or excessive deformations at pre-
selected locations. This type of design will help in the
repair of major structures with least cost and quick recov-
ery for business [38]. The performance-design is for differ-
ent levels of performance of structural components. To
achieve the required performance with adequate safety
and economy, advanced analytical procedures are neces-
sary for such an evaluation process.
Advanced analysis can directly address the overall sys-
tem response including second-order effects, gradual
Fig. 6 Rotational deformation of semi-rigid connection
Fig. 7 AISC-LRFD classification of flexibility of beam-to-column connections [4]
194 Wai Fah CHEN
yielding, inelastic force redistribution and semi-rigid con-
nections. Thus, the advanced analysis can address the
seismic design very well by providing answers to the crit-
ical questions. It is the most efficient tool in evaluating the
code besides practical design. A comprehensive report on
the development of practical advanced analysis method
for seismic steel frame design along with an evaluation
of the major impact of the 1997 LRFD steel seismic code
revision in the USA was made in the doctoral thesis by
Chen [39] along with its subsequent publication [40].
A comparative study of the LRFD methods with the
practical advanced analysis for steel building design was
reported in a recent thesis by Hwa [41]. It was found that
similar results were obtained since the advanced analysis
method was calibrated against the LRFD code. This pro-
vides structural engineers with more confidence in adopt-
ing the advanced analysis. As a demonstration of its
capability for a much wider application of the advanced
analysis method, both performance-based fire resistance
design and seismic design were also studied in some depth
by Hwa [41].
7 Summary
This paper describes the state-of-the-art progress of the
second-order inelastic analysis method for steel
building design now known as the ‘‘advanced analysis’’.
Extensive developments have been made in recent years
to make the advanced analysis work and practical in
engineering practice. To this end, a design method based
on the requirements put forward by the AISC-LRFD
specifications was developed and refined to achieve both
simplicity in use and, as much as possible, a realistic
representation of actual behavior. This practical
advanced analysis method for steel building design has
reached such a level of maturity that is now ready for
general use in engineering design.
Research works on advanced analysis are now in full
swing to develop nonlinear procedures and software for
practical use in design office. The theory and approaches
for advanced analysis of plane frames composed of mem-
bers of compact sections, fully braced out-of-plane, have
been well developed, verified and coded by the American
Institute of Steel Construction in the 2005 AISC
Specifications as well as others around the world.
8 Concluding remarks
Unlike the older second-order analysis used in seismic
and static design which is used for supplemental check-
ing only, the concept of advanced analysis can now be
used for primary design and member size framing.
Recent collapses of several shallow roof domes in
Russia, Germany and Poland clearly show the deficiency
of the current practice. The advanced analysis can be
easily applied to the design of these special or other
conventional structures.
For advanced analysis to achieve its full potential as a
tool for practical design for steel building structures,
future work is needed to integrate the effects of local buck-
ling of cross sections, lateral-torsional buckling of mem-
bers, flexibility of connections, and 3-D member behavior
into the advanced analysis. Furthermore, other structural
elements such as concrete floor slab, composite joints and
walls could also be included in the analysis for a more
realistic representation of the actual behavior of a steel
building structure during its life cycle performance and
service. Good progress has been made on some aspects
of this advancement, but much more remains to be done
(see for example, Ref. [42]).
References
1. Salmon C G, Johnson J E. Steel Structure Design and
Behavior. 3rd ed. New York: Harper Coliins, 1990
Fig. 8 Special truss moment frame ----- an example of struc-tural fuse concept in which plastic hinges serving as structuralfuses are developed at pre-determined locations
Advanced analysis for structural steel building design 195
2. Chen W F, Atsuta T. Theory of beam-columns. In: PlaneBehavior and Design. Vol 1. New York: McGraw–Hill,1976, 513
3. Chen W F, Atsuta T. Theory of beam–columns. In: SpaceBehavior and Design. Vol 2. New York: McGraw–Hill,1977, 732
4. AISC-LRFD. The Load and Resistance Factor DesignSpecification for Structural Steel Buildings. Chicago:American Institute of Steel Construction, 1986, 1994 and 2005
5. Chen W F, Lui E M. Structural Stability–Theory andImplementation. New York: Elsevier, 1986, 490
6. Chen W F, Han D J. Plasticity for Structural Engineers. NewYork: Springer-Verlag, 1988, 606
7. Chen W F, Toma S. Advanced Analysis of Steel Frames. BocaRaton: CRC Press, 1994, 384
8. Chen W F, Lui E M, eds. Chapter 5: Steel frame design usingadvanced analysis. In: Handbook of Structural Engineering.2nd ed. Boca Raton: CRC Press, 2005
9. Chen W F, Sohal I. Plastic Design and Second-Order Analysisof Steel Frames. New York: Springer-Verlag, 1995, 509
10. Liew J Y R. Advanced analysis for frame design. Dissertationfor the Doctoral Degree. West Lafayette: Purdue University,1992, 393
11. White D W. Plastic hinge methods for advanced analysis ofsteel frames. Journal of Constructional Steel Research, 1993,24(2): 121–152
12. White D W, Chen W F. Plastic Hinge Based Methods forAdvanced Analysis and Design of Steel Frames: AnAssessment of the State-of-the-Art. Bethlehem, Pennsylvania:Structural Stability Research Council, 1993, 299
13. ECCS. Essentials of Eurocode 3: Design Manual for SteelStructures in Buildings. ECCS-Advisory Committee 5, 1991,(65): 60
14. Liew J Y R, White D W, Chen W F. Second-order refinedplastic hinge analysis for frame design: Part I. Journal ofStructural Engineering, ASCE, 1993, 119: 3196–3216
15. Kim S E. Practical advanced analysis for steel frame design.Dissertation for the Doctoral Degree. West Lafayette: PurdueUniversity, 1996, 271
16. Kim S E, Chen W F. Practical advanced analysis for bracedsteel frame design. Journal of Structural Engineering, ASCE,1996, 122(11): 1266–1274
17. Kim S E, Chen W F. Practical advanced analysis for unbracedsteel frame design. Journal of Structural Engineering, ASCE,1996, 122(11): 1259–1265
18. McGuire W. Computer-aided analysis. In: Dowling P J,Harding J E, Bjorhovde R, eds. Constructional Steel Design-an International Guide. New York: Elsevier, 1992, 915–932
19. Ziemian R D, McGuire W. A method for incorporating liveload reduction provisions in frame analysis. EngineeringJournal, AISC, 1992, 29: 1–3
20. Attala NM, Deierlein G G, McGuire W. Spread of plasticity :quasi-plastic-hinge approach. Journal of StructuralEngineering, 1994, 120(8): 2451–2473
21. Vogel U. Calibrating Frames. Berlin: Stahlbau, 1985, 1–7, 10
22. Chen W F, Kim S E. LRFD Steel Design Using AdvancedAnalysis. Boca Raron: CRC Press, 1997, 441
23. Chen W F, Lui E M. Stability Design of Steel Frames. BocaRaton: CRC Press, 1992
24. Kishi N, Chen W F. Steel Connection Databank Program.Structural Engineering Report No. CE-STR-86-18. 1986
25. Kishi N, Chen W F. Database of Steel Beam-to-ColumnConnections, Vol. 1 & 2. Structural Engineering Report No.CE-STR-86-26, 1986
26. Nethercot D A. Steel beam to column connections – a reviewof test data and their applicability to the evaluation of the jointbehavior of the performance of steel frames. CIRIA ProjectRecord 338 338, 1985, 77–77
27. Richard R M. A study of structural system having nonlinearelements. Dissertation for the Doctoral Degree. WestLafayette: Purdue University, 1961
28. Kishi N, Chen W F. Moment- rotation relations of semi-rigidconnections with angles. Journal of Structural Engineering,ASCE, 1990, 116(ST7): 1813–1834
29. Kishi N, Chen W F, Goto Y, et al. Design aid for semi-rigidconnections for frame design. AISC Engineering Journal,1993: 90–107
30. Wu F H. Semi-rigid connections in steel frames. Dissertationfor the Doctoral Degree. West Lafayette: Purdue University,1988
31. Chen W F, ed. Semi-Rigid Connections in Steel Frames,Council on Tall Buildings and Urban Habitat. New York:McGraw-Hill, 1993, 318
32. Chen W F, Goto Y, Liew J Y R. Stability Design of Semi-Rigid Frames. New York: John Wiley & Sons, 1996, 468
33. Chen W F, Kim Y S. Practical Analysis for PartiallyRestrained Frame Design. Bethlehem, Pennsylvania:Structural Stability Research Council, Lehigh University,1998, 82
34. Chen W F, ed. Practical Analysis for Semi-Rigid FrameDesign. Singapore: World Scientific Publishing Co., 2000, 465
35. Ricles J, Bjorhovde R, Iwankiw N, eds. SSRC, StructuralStability Research Council Workshop Proceedings: Frameswith Partially Restrained Connections, Atlanta, Georgia.September 23, 1998
36. AISC-LRFD. Seismic Provisions for Structural Steel Buildings.Chicago: American Institute of Steel Construction, 1997
37. Chen W F, Yamaguchi E. Spotlight on steel moment frames.Civil Engineering, ASCE, 1996, 44–46
38. NEHRP. NEHRP Guidelines for the Seismic Rehabilitationof Buildings. Washington DC: Building Seismic SafetyCouncil, FEMA, 1996
39. Chen I H. Practical advanced analysis for seismic design ofsteel building frames. Dissertation for the Doctoral Degree.West Lafayette: Purdue University, 1999
40. Chen I H, Chen W F. Practical advanced analysis for seismicframe design. Advances in Structural Engineering, 1999, 2(4):237–363
41. Hwa K. Toward Advanced Analysis in Steel Frame Design.Dissertation for Doctoral Degree. Honolulu: University ofHawaii, 2003
42. Wongkaew K, Chen W F. Consideration of out-of-planebuckling in advanced analysis for planar steel frame design.Journal of Constructional Steel Research, 2002, 58: 943–965
196 Wai Fah CHEN