Simplified Design of Building Frames using

First-Order Analysis and K = 1.0

G.H. Kuchenbecker1, D.W. White2, and
A.E. Surovek-Maleck3
INTRODUCTION
Valid assessment of the strength and stability of steel framing systems
requires that second-order effects be considered. This may be done
either using general purpose second-order algorithms, or by judicious
use of approximate approaches such as the NT-LT method provided in
AISC (1999), in which internal first-order forces and moments are
amplified to establish second-order values.

Recommendations for the 2005 AISC Specification include two

methods for assessing frame stability: the Critical Load and the Direct
Analysis approaches (Deierlein et al. 2002, Maleck & White 2003,
Surovek-Maleck & White 2004). The essential attributes of the Critical
Load approach are the same as the (AISC 1999) column buckling or
effective length based procedures. Conversely, the Direct Analysis
approach represents a departure from previous methods in the AISC
Specifications in that it takes advantage of more direct modeling of the
aspects that affect the system and member strength. The Direct
Analysis method gives more accurate estimates of the internal forces
and moments, and in as such, eliminates the need for calculation of
column effective lengths (Surovek-Maleck & White 2004). The Direct
Analysis approach may be used either with general-purpose second-
order algorithms, or with approximate methods based on first-order
analysis, such as the NT-LT approach.

This paper illustrates the use of the Direct Analysis approach in the
context of an alternative first-order elastic analysis method for

1
Graduate Asst., South Dakota School of Mines and Technology Rapid City, SD 57701
2
Associate Professor, Georgia Institute of Technology, Atlanta, GA 30332-0355
3
Asst. Professor, South Dakota School of Mines and Technology Rapid City, SD 57701
calculation of second-order forces and moments in rectangular framing
systems. In the proposed underlying analysis method, the second-order
sidesway effects are determined by applying equal and opposite P∆
shear forces at the top and bottom of each story. These forces are
calculated using the amplified sway displacements from a first-order
analysis. The combination of the broader Direct Analysis approach and
the proposed underlying second-order analysis method is particularly
powerful in that:
• No K factor calculations are required,
• No free-body diagrams are necessary to determine accurate
second-order beam and connection moments,
• The relative significance of second-order effects is established in a
straightforward manner from preliminary or intermediate
calculations,
• The design of all types of braced, moment and combined framing
systems is handled in a unified and consistent fashion
• The Engineer’s attention is focused on the important attribute of
providing adequate overall sidesway stiffness.
The first part of the paper gives an overview of the broader Direct
Analysis approach for design of steel frames, as well as the proposed
underlying amplified first-order elastic analysis procedure for
determining second-order forces and moments. The step-by-step
application of these combined methods is then discussed. The paper
concludes by presenting two design examples.

DIRECT ANALYSIS APPROACH

The development of the Direct Analysis approach for general framing
systems is discussed in (Deierlein 2003), (Maleck & White 2003) and
(Surovek-Maleck & White 2004). In this approach, two modifications
are made to a conventional elastic analysis:
1. For any member in which Pr > 0.5Py, the flexural rigidity is taken
as 0.8τEI, where
Pr ⎛ Pr ⎞
τ=4 ⎜1 − ⎟ (1)
Py ⎜ Py ⎟
⎝ ⎠
 All the contributions to the elastic stiffness are reduced by a base
value of 20 percent relative to their nominal values, i.e., they are
multiplied by a factor of 0.8.
2. A notional load of Ni = 0.002Yi is added at each level, where Yi is
the total factored gravity load acting on the ith level. This
modification is equivalent to a uniform nominal out-of-plumbness
of L/500, based on erection tolerances specified in (AISC 2000).
The above reductions in the stiffness are used only for checking
strength. Serviceability limits are checked using nominal (unreduced)
stiffnesses. It should be noted that the uniform reduction of 0.8
influences only the second-order effects within the system. That is, for
structures in which the second-order effects are small, the 0.8 stiffness
reduction has a negligible effect on the magnitude and/or distribution of
the system internal forces. The rationale for the above modifications is
discussed in detail in (Surovek-Maleck & White 2004.)

AMPLIFIED FIRST-ORDER ANALYSIS EQUATIONS FOR

BRACED FRAMING
The idealized model shown in Figure 1 represents the essential
attributes within a story of a rectangular structure composed of pinned
gravity framing combined with a braced frame system. The vertical
load carrying members are represented by a single column. The
bracing system is represented by a spring at the top of the column. An
axial load of ΣPr is applied to the model, where Σ Pr is the required total
vertical load supported by the story (i.e. factored gravity load). A
horizontal load ΣH is also applied to the model, where ΣH represents
the required story shear due to the applied loads on the physical
structure. The minimal bracing stiffness required to prevent sidesway
buckling under the load ΣPr (with ∆o and ΣH equal to zero) is defined
as the ideal stiffness (Galambos 1998)
ΣP (2)
βi = r
L
whereas the sidesway stiffness provided by the bracing system is
denoted by:
 ∆ o ∆1 ∆ 2

ΣPr
ΣH β (∆1 H + ∆ 2 ) = F1 + F2

∆ o = initial out-of-plumb

L ∆1 = 1st-order drift due to applied loads = ∆1 H + ∆1 P

∆1 H = drift due to lateral loads = ΣH / β

∆1 P = drift due to vertical loads (caused by nonsymmetry

of vertical loads and/or frame geometry)
A
∆2 = additional drift due to 2nd-order (P∆) effects
β ∆2
ΣPr

Figure 1. Idealized model of a braced story within a tiered structure

ΣH
β= (3)
∆1H
In Eq. (3) and throughout the subsequent developments, the overbars
indicate parameters that are affected by the stiffness reduction used in
the Direct Analysis approach; the fundamental equations, however, are
valid also when used with nominal stiffness values. By application of
basic structural analysis principles (LeMessurier 1976; White et al.
2004a), the total drift of the lateral load resisting system is obtained as:

( ) (
∆ tot = ∆ o + ∆1 + ∆ 2 = ∆ o + ∆1 + ) (∆β0 + ∆1 ) = Blt (∆ o + ∆1 ) (4)
−1
βi
where
β
βi 1
Blt = = (5)
β βi
−1 1−
βi β
The term Blt is the sidesway displacement amplification factor. The
subscript “lt” is used in this work rather than “2” since the subscripts 0,
1 and 2 are reserved to denote initial, first-order and second-order
quantities. For a complete derivation of the above equations, the reader
is referred to (White et al. 2004a).

The total drift of the story ∆ tot is equal to the total first order
displacement (∆ o + ∆1 ) multiplied by Blt . The total lateral force
developed within the lateral load resisting system is in turn

F1 + F2 = ΣH + ΣPr Blt
(∆ o + ∆1 ) (6)
L
where F2 represents the second order component of the internal forces.
The total force within the lateral load resisting system is equal to the
first-order internal shear force ΣH plus the shear force associated with
the gravity loads acting through the total inter-story displacement ∆ tot .

AMPLIFIED FIRST-ORDER ANALYSIS EQUATIONS FOR

STRUCTURES CONTAINING MOMENT FRAMES
In this section, a moment- rather than a braced frame system is
employed to provide the lateral-load resistance and stability. For a
more detailed discussion of the formulation and complete derivations of
the following equations, see (White et al. 2004b).

Under the hypothetical case that the moment frame supports zero
gravity load, the solution presented in the previous section still applies.
In the more typical case in which the moment frame system also
supports gravity load, the free-body diagram shown in Figure 2 applies.
LeMessurier (1977) provides a general approximate method for the
second-order analysis of this type of structure. In the limit that the
columns are relatively rigid and the flexibility of the moment frame is
predominantly due to deformations within the beams and connections,
the lateral-load resisting columns remain essentially straight under the
drift of the story. Therefore, the solution presented in the previous
section also applies to this problem. However, if the flexural rigidity of
the beams and connections in the moment frame system are relatively
large, additional Pδ moments exist within the lateral-load resisting
columns as illustrated by the diagram on the right-hand side of Figure
2. These Pδ moments cause additional sidesway displacements for a
given horizontal shear ΣH. That is, they cause an additional
amplification of the first-order lateral displacements (∆ o + ∆1 ) .

∆ o ∆1 ∆ 2 ΣPr = ΣgPr + ΣmPr

ΣgPr Σm P r Moment Diagram

ΣH (Right-Side Column)

x Σ Hx
x
ΣPr (∆ o + ∆1 + ∆ 2 )
L
L
ΣmPr δ
δ

Σ gPr Σ gPr
A
(∆ o + ∆1 + ∆ 2 ) ΣH + (∆ o + ∆1 + ∆ 2 )
L L

Σ HL + Σ Pr (∆ o + ∆1 + ∆ 2 )
ΣgPr ΣmPr

Figure 2. Idealized model of a story within a tiered structure

composed of gravity framing combined with a moment frame system

Based on the solutions presented by LeMessurier (1977), a reasonable

approximation of the above Pδ effects is obtained simply by modifying
the term Blt in Eq. (4) from the form given by Eq. (5) to
1 (7)
Blt =
β
1− i
βRm
where
Σ m Pr (8)
R m = 1 − 0.15
ΣPr
(White et al. 2004b). The term Rm is equal to one in the limit that the
moment frames support zero gravity load (i.e., ΣmPr = 0), or the
structural system does not contain any moment frames, and is equal to
0.85 for a framing system in which all of the gravity load is supported
by moment frame columns (i.e., ΣmPr = ΣPr).

The proposed amplified first-order elastic analysis method does not

apply amplification factors to the first-order internal stress-resultants
due to sidesway. Rather, the sidesway displacements ∆1 are calculated
and then the corresponding P∆ shears
⎛ ∆ + ∆1 ⎞ (9)
H P∆ = ΣPr Blt ⎜⎜ o ⎟⎟
⎝ L ⎠
are applied to the structure as equal and opposite forces at the top and
bottom of each story to determine the second-order component of the
forces and moments throughout the lateral load resisting systems. This
leads to a more general, straightforward, accurate and intuitive
calculation of the P∆ effects on the internal forces and moments than
the AISC (1999) NT-LT approach (White et al. 2004a & b):
• The proposed approach is uniformly applicable to braced, moment
and combined framing systems.
• Second-order internal forces and moments in beams, connections
and bracing members are determined automatically.
• Approximations associated with multiplication of the first-order
internal moments by the (AISC 1999) B2 amplifier comparable to
Eq. (5) are avoided.
• The Engineer does not need to subdivide the analyses into artificial
NT and LT parts.
Furthermore, if H P∆ from Eq. (9) is smaller than a certain fraction of
the story shear due to the applied lateral loads ΣH (e.g., five percent),
the Engineer can exercise his or her judgment and assume that the
second-order sidesway effects are negligible. Proper application of the
story shears in multistory buildings is addressed in (White et al. 2004a).
Considerations associated with the Pδ amplification of non-sway
moments are addressed in (White et al. 2004b). The authors suggest
that if
 Cm
Bnt = ≤ 1.05 (10)
1 − Pr / Pe.nt
where all the terms in this equation are the same as the comparable
terms in the AISC (1999) B1 amplifier except that Pe.nt is calculated
using the reduced elastic stiffness, non-sway second-order effects do
not need to be considered. In many types of structures, all the framing
members will satisfy Eq. (10). The authors point out that in the limited
cases where Eq. (10) is violated, beam and connection moments often
are affected significantly in addition to the beam-column moments.
Typically, a general-purpose second-order analysis is required to
determine the second-order beam and connection moments properly in
these types of structures.

STEP-BY-STEP APPLICATION
For each load combination, the basic steps of the proposed procedure
are as follows:
1. Perform a first-order elastic analysis of the structure
2. Obtain the total first-order story lateral displacements ∆1
3. Calculate the sidesway displacement amplification factor, Blt
4. Calculate the P∆ shears
5. Apply the P∆ shears and determine the second-order forces and
moments directly from the analysis.
The above P∆ shears include both the effect of the reduction of
structure stiffness and the effect of the nominal out-of-plumbness ∆0
associated with the Direct Analysis method.

The following is a more detailed description of the steps:

(1) Perform a first-order analysis for each of the load types under
consideration. If the analyses are run for each load type (D, L, W,
Lr, E, etc.) using nominal (unfactored) load levels, the results can
be scaled and combined using superposition for each considered
load combination. Establish the first-order internal member forces.
(2) Obtain ∆1 , the total first-order story lateral displacements for a
given load combination, from the analyses in step (1).
(3) Calculate the sidesway amplification factor Blt associated with a
given strength load combination from Eq. (5), with βi and β
calculated using Eqs. (2) and (3) respectively. If ΣH = 0 (and
therefore ∆1H = 0) determine β by applying any fraction of ΣPr as
an equal and opposite set of shear forces at the top and bottom of
each story).
(4) For each of the load combinations, calculate the P∆ story shear
forces H P∆ from Eq. (9) and apply to each story of the frame
within a separate first-order analysis to determine the second-order
component of the internal forces.
(5) Add the second-order forces from step (4) to the first-order forces
associated with the strength load combination from step (1) to
obtain the required strengths for the entire structure.

EXAMPLE 1 – BRACED FRAME

Figure 3 shows a long-span roof structure initially presented by
LeMessurier (1976). The lateral load resistance for this structure is
provided by the brace ab. Minor modifications have been made to the
original design to satisfy the (AISC 1999) requirements.

The following strength load combinations are considered:

1.2D + 1.6Lr + 0.8W, 0.9D + 1.6W and 1.2D + 0.5Lr + 1.6W
The size of column bc and the tension force requirement in brace ab are
governed by the strength load combination 1.2D + 1.6Lr + 0.8W with
the wind acting to the right, whereas the size of the diagonal ab is
governed by its compressive strength under the load combination 0.9D
+ 1.6W with the wind acting to the left. Separate first-order analyses of
the above structure under the nominal (unfactored) gravity and wind
loads are presented first. A discussion of results from the base analyses
and one of the strength load combinations follows.
 2 kips/ft
2.7 kips b d

Brace
HSS4.5x4.5x3/16 W8x48
Aab = 2.93 in2 Ac = 14.1 in2 L =18 ft
Lab/ry = 125 L/ry = 104
Fy = 46 ksi Fy = 50 ksi
a c e

B = 3 ft 165 ft

Wind = 15 psf, Dead = 25 psf, Live = 75 psf

Frames 20 ft. o.c.

Figure 3. LeMessurier’s Example Frame (1976) designed by LRFD

Base First-Order Analysis of Nominal Loads. The derivation of the

base first-order elastic analysis equations for the structure is presented
in LeMessurier (1976). The nominal column loads are P = 165 kips
and the total nominal story vertical load is ΣP = 330 kips. The axial
compression in column bc causes it to shorten, and compatibility
between the diagonal ab and the top of the column at b requires a
corresponding story drift of
L
P
∆1P 1
= B = = 0.00242
L A c E 413
where L is the story height, B is the horizontal distance between the
base of column bc and the base of the diagonal ab, the column area, Ac
= 14.1 in2 and E = 29000 ksi. The nominal elastic sidesway stiffness of
the structure is given by the equation
AcE kips
β= = 8.743
⎡ ⎛L⎞
2 ⎤ in
L ⎢(1 + λ )⎜ ⎟ + λ ⎥
⎢⎣ ⎝ B⎠ ⎥⎦
where:
 L ab A c
λ=
A DL
The length of the diagonal is 18.25 ft. The story drift due to the
nominal horizontal load of ΣH = 2.7 kips is
∆1H ΣH 1
= = = 0.00143
L βL 699
It is interesting and important to note that the story drift under the
nominal gravity load alone is significantly larger than H/500.
Furthermore, the drift due to the nominal gravity load is 1.7 times that
due to the nominal wind load. This attribute of the response and the
magnitude of the total vertical load make this example a severe test of
the validity of the various stability analysis and design procedures. It
also illustrates the importance of considering second-order effects in
general for braced systems; second-order effects are often neglected in
these types of systems.

Strength Analysis (1.2D + 1.6Lr + 0.8W), Wind Acting to the Right.

The ideal stiffness of the example structure associated with the load
combination (1.2D + 1.6Lr + 0.8W) is given by
1.5ΣP kips
βi = = 2.292
L in
where the factor 1.5 = (1.2D + 1.6Lr) / (D + Lr). By the Direct Analysis
approach, the sidesway displacement amplifier is given by
1 1
Blt = = = 1.487
βi β
1− 1− i
β 0.8β
and the total story drift is computed as
∆ tot ⎡ ⎛ ∆ ∆ ⎞ ⎛ 1 ⎞⎤ 1
= Blt ⎢0.002 + ⎜1.5 1P + 0.8 1H ⎟ ⎜ ⎟⎥ =
L ⎣ ⎝ L L ⎠ ⎝ 0.8 ⎠⎦ 83
where the term 0.8 in the denominator of these equations is the elastic
stiffness reduction factor. This results in a P∆ story shear of
 ⎛ ∆ total ⎞
H P∆ = 1.5ΣP⎜ ⎟ = 5.87 kips
⎜ L ⎟
⎝ ⎠
a tension force requirement in the diagonal ab of

( )
⎛L ⎞
Fab = 0.8ΣH + H P∆ ⎜ ab ⎟ = 48.8 kips
⎝ B ⎠
and a maximum compressive strength requirement in column bc of

( )
⎛L⎞
Fbc = 1.5P + 0.8ΣH + H P∆ ⎜ ⎟ = 295.7 kips
⎝ B⎠
The Direct Analysis approach is equivalent to the simplified AISC
(1999) approach with φ = 0.8 in the limit that β = 2βi and ∆o + ∆1 =
0.002L (White et al. 2004a). However, the Direct Analysis approach is
more straightforward to understand and apply, since it works directly
with a basic reduced stiffness representation of the nominally imperfect
structure at the maximum strength limit state. Given the nominal
stiffness reduction of 0.8 and the nominal geometric imperfection of ∆o
= 0.002L, the Direct Analysis solution follows “directly” from the
fundamental mechanics. The simplified AISC (1999) approach, while
providing an important practical limit on the minimum lateral stiffness
of the structure, allows the Engineer to conduct a conventional analysis
of the perfect nominally-elastic structure and then imposes an
additional stability bracing force requirement of 0.004ΣPr to account
approximately for the above effects. In many cases, β >> 2βi and the
second-order force component determined by the proposed method, F2
in Eq. (6), is significantly smaller than 0.004ΣPr.

Summary of Overall Results. Table 1 summarizes the overall results

for the LRFD strength load combinations considered in this example.
For a more complete discussion of the example and the listed results,
the reader is referred to (White et al. 2004a). The results using the
AISC (1999) stability bracing provisions and the proposed Direct
Analysis approach are comparable for this frame. Direct Analysis
gives a maximum tension in brace ab of 48.8 kips, which is nine
percent larger than that based on the AISC (1999) second-order elastic
analysis (i.e., with ∆o = 0 and nominal elastic stiffness) and
subsequently including the stability bracing force requirement. It gives
a maximum compression within the diagonal brace of -28.3 kips, which
is two percent smaller than that calculated using the above AISC (1999)
approach. White et al. (2004a) also present plastic zone analysis results
for this structure. The internal forces obtained by plastic zone analysis
are similar to the values listed in Table 1.

Table 1. Summary of Analysis Results, braced frame

Load Combination Analysis Procedure Fab Fbc
1.2D + 1.6Lr +
0.8W AISC (1999) 44.7 291.6
Wind to the right Direct Analysis 48.8 295.7
0.9D + 1.6 W AISC (1999) -28.9
Wind to the left Direct Analysis -28.3
1.2D + 1.6Lr +
0.8W AISC (1999) 37.7
Wind to the right Direct Analysis 37.2

It is important to recognize that in LeMessurier’s example frame, the

tension force requirement in ab would be dangerously underestimated
as 13.1 kips if the Engineer conducted only a first-order elastic
analysis. If a first-order elastic analysis were employed and the
0.004ΣPr requirement was added to these results, the tension force
requirement in ab still would be only 25.2 kips (not shown in Table 1).
This underestimates the force of 48.8 kips obtained from Direct
Analysis by 48 percent, and is based on a false assumption that ∆o + ∆1
is less than or equal to 0.002L. The governing compression force
requirement in ab from the load combination (0.9D + 1.6W), and the
force in ab for the combination (1.2D + 0.5Lr + 1.6W) are less sensitive
to the type of analysis. This is largely because the total gravity loads as
well as the drift of the frame are significantly smaller in these load
cases.
The authors emphasize that the brace tension force ab in the example
frame is highly sensitive to deviations in the story drift due to: (1) the
small angle of the brace relative to the vertical position, (2) the large
gravity load supported by the structure, (3) the relatively small wind
load requirements and (4) the correspondingly small lateral stiffness of
the bracing system. If a symmetrical configuration of the bracing were
introduced, the second-order P∆ effects in this frame are dramatically
reduced. Also, the second-order internal story shears approach
0.002ΣPr when the bracing system has substantial stiffness relative to
the ideal stiffness βi (White et al. 2004a). In many cases, particularly
if ΣPr is relatively small, these internal shear forces are only a small
fraction of the lateral load resistance of the bracing system. However,
in sensitive stability critical frames such as this example, the
application of either Direct Analysis or the AISC (1999) stability
bracing provisions is considered prudent.

EXAMPLE 2 – MOMENT FRAME

The following is an example moment-gravity frame that demonstrates
the advantages and ease of the proposed approach compared to
conventional methods. Figure 4 shows an idealized model of an 11-bay
single-story frame developed in (Maleck 2001) and further studied in
(Maleck & White 2001), (Surovek-Maleck & White 2004) and
(Deierlein 2003). Five gravity columns on each side of the interior
moment frame are represented by single leaning columns with load 4P,
where P is approximately the axial load in each of the interior gravity
columns. To illustrate the robustness of the proposed procedure, large
gravity loads are applied which produce significant P∆ effects; one of
the primary attributes of this frame is that it is sensitive to initial
imperfection effects, and yet it meets drift limits based on a service
wind load of 0.7W (Surovek-Maleck & White 2004).

The strength of the example frame is assessed using: (1) a conventional

analysis of the geometrically-perfect nominally-elastic structure (the
AISC (2000) NT-LT approach is employed) and with the column axial
load resistances within the design checks determined based on buckling
loads or effective length factors, and (2) Direct Analysis of the
structure, with the proposed amplified first-order method used to
determine the internal forces and moments. These solutions are
compared to the results from a refined plastic zone analysis from
(Surovek-Maleck & White 2004). The plastic zone analysis models the
spread of plasticity explicitly, including the effect of nominal residual
stresses and a nominal out-of-plumbness of ∆o = 0.002L. The yield
strength and the stiffness are both factored by 0.9 for the plastic zone
analysis, as discussed by Surovek-Maleck & White (2004) and White et
al. (2004a & b).
4P 4P
Frame spacing = 35'
W
W27 x 84

L = 18'
W10 x 49

3 @ 35' =105'

DL = 80 psf Load Combinations:

LL = 40 psf 1.2D + 1.6L
Wind = 20 psf

Figure 4. Idealized model of 11-bay single story frame

In the conventional analysis-design approach, the effective length for

in-plane buckling is determined per Eq. (C-C2-6) of (AISC 1999),
including the effect of the leaning columns. A theoretical value of
G = 0 is used at the column base, rather than the recommended practical
value of one, to be consistent with the assumption of fully-fixed base
conditions within all of the analysis models. Due to the symmetry of
the structure, the sidesway moments Mlt are zero within the
geometrically-perfect structure; also, B1 is calculated to be equal to 1.0
based on Eq. (C1-2) of (AISC 1999).

Comparison of Internal Forces. Table 2 shows the axial force and

maximum internal second-order moments in the lateral load resisting
columns obtained by each of the above approaches. The internal
moment in the conventional NT-LT analysis approach is simply the
first-order NT moment, since B1 = 1 and Mlt = 0. The internal moment
within the Direct Analysis approach is simply equal to this moment
plus the effect of the factored story gravity load ΣPr acting through the
nominal initial out-of-plumbness, ∆o= L/500, amplified due to the
second-order effects

The sidesway amplifier used in the proposed Direct Analysis procedure

is calculated from Eq. (5) as
1 1
Blt = = = 3.65
ΣPr / L 2,156 kips / 216 in
1− 1−
(0.8ΣH / ∆1H )R m (14.17 kips / in )(0.970)

where β = 0.8ΣH/∆1H may be calculated for this frame by applying

horizontal load ΣH = 1 kip at the roof level of the nominally elastic
structure, determining the first-order sidesway deflection ∆1H due to
this load, and then multiplying by 0.8 to account for the required
stiffness reduction. The column axial loads Pr are smaller than 0.5Py
within the example frame, and therefore the column inelastic stiffness
reduction factor τ is equal to one. The term Rm = 0.970, which
accounts for the Pδ effects in the moment frame columns on the
sidesway deflections, is calculated using Eq. (8).

The P∆ shears are calculated as

⎛ ∆ + ∆1 ⎞
H P∆ = ΣPr Blt ⎜⎜ o ⎟⎟ = (2,156 kips)(3.65)(0.002) = 15.73 kips
⎝ L ⎠
v
Since the first-order lateral displacement ∆1 is zero, the P∆ shear force
is based solely on the axial load acting through the initial imperfection
of ∆o = L/500. The internal moments generated by these P∆ shears are
then added to the previous internal moments using superposition. The
results are summarized in Table 2.

Whereas, the axial forces are similar in each method, the column
internal moments differ substantially. The percent difference in the
moments relative to that determined by plastic zone analysis is
expressed as ε = 100(M - Mpz)/Mpz, where Mpz is the moment
determined by plastic zone analysis. One can observe that the
conventional NT-LT analysis of the geometrically-perfect nominally-
elastic structure substantially underestimates the internal moments
determined from the plastic zone analysis. The internal moments in
this frame are highly sensitive to any potential out-of-plumbness of the
structure, in spite of the fact that it satisfies typical drift limits under the
service wind load. The design gravity loads are near the frame’s elastic
sidesway buckling load, as indicated by the large Blt .

Table 2. Analysis results, moment frame example

Design Method Pr [kips] Mr [in-kips] ε

NT-LT Analysis 216 397 -61 %
Direct Analysis 218 1276 24 %
Plastic Zone Analysis 215 1030

The Direct Analysis approach is conservative in calculating the internal

moments within this stability critical structure. The column end
moments are overestimated somewhat in this frame predominantly
because the columns are still essentially elastic at the design load level.
Therefore, the stiffness reduction of 0.9 used in the plastic zone
analysis is sufficient for this problem. Because of the frame’s large
Blt , the effects of the reduced elastic stiffness are accentuated. The 0.8
elastic stiffness reduction factor within the Direct Analysis approach
accounts generally for a reduction of 0.9 in the stiffness due to yielding,
etc. as the strength limit of the most critical component is approached,
as well as an additional reduction of φb = φc = 0.9 applied to the
stiffness, as in the plastic zone analysis solution. The product of these
factors is rounded to 0.8.

Design Check. Since Pr/φPn ≥ 0.2 in each design approach, the beam-
columns are checked by AISC (1999) Eq. H1-1a for in-plane and out-
of-plane strength. The beam columns are presumed to be braced at
their ends in the out-of-plane direction. The results are presented in
Table 3.

Table 3. Strength checks, moment frame example

Interaction
Design Strengths Check
Out-
φPn(x) φPn(y) φMn In- of-
Design Approach Kx [kips] [kips] [in-k] plane Plane
NT-LT Analysis 2.3 236 361 2718 1.05 0.73
Direct Analysis n/a 511 361 2718 0.78 0.95
Plastic Zone n/a n/a 361 2718 n/a 0.93

In the AISC (1999) assessment, the in-plane interaction check is

violated due primarily to the large effective length factor (relative to
1.0). As discussed previously, because of the lack of inclusion of any
initial imperfections in the conventional AISC (1999) approach, the
moments are substantially underestimated. Therefore the out-of-plane
interaction check is in effect unconservative.

Since the Direct Analysis approach uses the actual column unsupported
length in calculating the axial strength, the weak axis always governs
the design if the unbraced lengths are the same in the strong and weak
axis directions (assuming that the same beam-column interaction
equations are used to check both in-plane and out-of-plane beam-
column strength). This is not the case in the LRFD procedure, as
evident in Table 3, where the design is controlled by the in-plane
checks. The plastic zone analysis shows that the frame has more than
adequate in-plane strength (Surovek-Maleck & White 2004).

For the sake of clarity, the examples provided above show the proposed
method in its most basic form for single story frames. For an extension
of the method to multi-story structures as well as further examples, the
reader is referred to (White et al. 2004a & b.)
CONCLUSIONS

The influence of second-order effects on the strength and stability of

steel framing systems must be considered in design. This paper
presents a design approach in which the previously reported attributes
of the Direct Analysis approach are combined with a simplified form of
the fundamental equations from LeMessurier (1976 & 1977) for
calculation of second-order displacements and internal forces from
first-order analysis. The method offers a number of benefits including:
(1) intuitive, straightforward, general and accurate calculation of
second-order internal forces and moments, (2) elimination of the need
to calculate effective length factors, (3) emphasis on providing
adequate sidesway stiffness.

Two practical but stability critical examples, one in which the lateral
load resisting system is provided by bracing and the other where the
lateral load resistance is provided by a moment frame, are presented to
illustrate the proposed approach. The examples emphasize the
applicability of the approach to general rectangular framing systems.

REFERENCES

AISC (2000), Code of Standard Practice for Steel Buildings and Bridges,
American Institute of Steel Construction, Inc., Chicago, IL.

AISC (1999), Load and Resistance Factor Design Specification for Steel
Buildings, American Institute of Steel Construction, Inc., Chicago, IL.

Deierlein, G., Hajjar, J.F., Yura, J.A., White, D.W., and Baker, W.F. (2002),
“Proposed New Provisions for Frame Stability Using Second-Order Analysis,”
Proceedings, Annual Technical Session, Structural Stability Research Council,
Seattle, April.

Deierlein, G. (2003). “Background and Illustrative Examples on Proposed

Direct Analysis Method for Stability Design of Moment Frames,” Background
Materials, AISC Committee on Specifications, Ballot 2003-4-360-2, August
20, 17 pp.
Galambos, T.V. (1998), Guide to Stability Design Criteria for Metal
Structures, 5th Edition, T.V. Galambos (ed.), Structural Stability Research
Council, Wiley.

LeMessurier, W.J. (1976), “A Practical Method for Second Order Analysis.

Part 1 – Pin Jointed Systems,” Engineering Journal, AISC, 12(4), 89-96.

LeMessurier, W.J. (1977), “A Practical Method for Second Order Analysis.

Part 2 – Rigid Frames,” Engineering Journal, AISC, 14(2), 49-67.

Maleck, A.E. (2001), Second-Order Inelastic and Modified Elastic Analysis

and Design Evaluation of Planar Steel Frames, Ph.D. Dissertation, Georgia
Institute of Technology, 579 pp.

Maleck, A.E. and White, D.W. (2003), “Direct Analysis Approach for the
Assessment of Frame Stability: Verification Studies,” North American Steel
Construction Conference, AISC, Baltimore, April.

Maleck, A. E. and White, D. W., “A Modified Elastic Approach for Design of

Steel Frames,” Annual Technical Session, Structural Stability Research
Council, Ft. Lauderdale, May 2001.

Surovek-Maleck, A.E. and White, D.W. (2004), “Alternative Approaches for

Elastic Analysis and Design of Steel Frames. I: Overview,” Journal of
Structural Engineering, ASCE, to be published.

White, D.W., Surovek-Maleck, A.E. and Kim, S.C. (2004a). “Direct Analysis
and Design by Amplification of First-Order Elastic Results. Part 1 – Combined
Braced and Gravity Framing Systems,” Engineering Journal, AISC, submitted
for review.

White, D.W., Surovek-Maleck, A.E. and Chang, C.J. (2004b). “Direct

Analysis and Design by Amplification of First-Order Elastic Results. Part 2 –
Moment Frames and General Structural Systems Including Moment Frames,”
Engineering Journal, AISC, submitted for review.