13th World Conference on Earthquake Engineering

Vancouver, B.C., Canada

August 1-6, 2004
Paper No. 266

PERFORMANCE-BASED SEISMIC DESIGN OF STEEL MOMENT

FRAMES USING TARGET DRIFT AND YIELD MECHANISM

Soon-Sik LEE1, Subhash C. GOEL2, and Shih-Ho CHAO3

SUMMARY

A performance-based seismic design procedure for steel moment frames based on pre-selected yield
mechanism and target drift is proposed in this paper. The design base shear is derived from a modified
energy balance equation incorporating the concept of seismic force reduction factor and the displacement
amplification factor. A new seismic lateral force distribution based on nonlinear dynamic analyses is also
presented. Three steel moment frames were designed using the proposed lateral force distribution and the
performance plastic design methodology with specified pre-selected target drifts. The inelastic seismic
behaviors of the three frames were studied through nonlinear static and dynamic analyses. The results
showed that all three frames performed very well under design level ground motions; in particular, all
frames exhibited expected yield mechanism with story drifts within the target limit. That is, the structural
performance can be well predicted and “controlled” by employing the proposed performance-based design
procedure.

INTRODUCTION

Many studies have shown that building structures designed by modern seismic code procedures are
expected to undergo large cyclic deformations in the inelastic range when subjected severe earthquake
ground motions. Nevertheless, most seismic design codes are still based on elastic methods using
equivalent static lateral design forces. This procedure can result in unpredictable and poor response during
severe ground motions with inelastic activity unevenly distributed among structural members. To solve
this problem, Leelataviwat [1] developed a new performance-based plastic design procedure for steel
moment frames using the concept of energy balance applied to a pre-selected yield mechanism, with
adequate strength and ductility. This study is an extension and modification of the previous study by
Leelataviwat [1]. It is well known that force reduction and displacement amplification factors, intended to
account for damping, energy dissipation capacity as well as overstrength, play important roles in seismic
design. However, since the previous proposed design method (Leelataviwat, 1998) did not consider the
above factors as influenced by the structural periods, the method could result in conservative design for

1
Senior Bridge Engineer, URS Corporation, Roseville, USA. Email: [email protected]
2
Professor, University of Michigan, Ann Arbor, USA. Email: [email protected]
3
Ph.D. Candidate, University of Michigan, Ann Arbor, USA. Email: [email protected]
long period structures and unconservative design for low-rise, short period structures. For simplicity, a
linear distribution of lateral design forces has been generally used in the codes. However, many studies
have shown that this distribution may not be applicable in the inelastic stage and may underestimate the
story shears. It can also be too conservative for the design of columns in the performance based plastic
design procedure. Moreover, this distribution does not satisfactorily recognize the higher mode effects for
high-rise building structures. In this study, a new distribution of the lateral forces derived from nonlinear
dynamic analysis is presented and applied to the proposed performance based plastic design procedure.
The results of nonlinear static and dynamic analyses of three example steel moment frames (3, 9, and 20
story) designed by the proposed method are also presented and discussed.

MODIFIED ENERGY BALANCE EQUATION

Design base shears in current seismic codes are calculated by reducing the elastic strength demands to the
inelastic strength demands using the response modification factors. These inelastic strength demands are
further increased according to the importance of specific structures using occupancy importance factor.
Generally, the design base shear is determined from the code prescribed design acceleration spectrum and
expressed in the following form:

⎛I⎞
V = C e ⎜ ⎟W (1)
⎝R⎠

where Ce is the normalized design pseudo-acceleration; I is the occupancy importance factor; R is the
response modification factor; and W is the total seismic weight. After selecting the member sizes for
required strengths (which is generally done by elastic analysis) the calculated drift using elastic analysis is
multiplied by deflection amplification factor, such as Cd given in the codes, and kept within specified drift
limits (in the order of 2%).

It is noted that the response modification factors, R, specified in design codes for various structural
systems are determined primarily based on engineering judgment. Moreover, as stated earlier, the
conventional design procedures in the codes are based on elastic force-based analysis methods rather than
displacement-based methods, thus the inelastic response beyond the elastic limit for a structure cannot be
predicted with good precision. A more rational design approach to overcome the shortcomings in the
conventional approach was proposed by Leelataviwat [1] and modified by Lee and Goel [2], which uses
energy balance equation as the design basis with the structure pushed monotonically up to a target drift
after the formation of a selected yield mechanism. The amount of external work needed to do that is
assumed as a factor γ times the elastic input energy E (= 12 MSv2 ) . The modification factor γ is dependent
on the natural period of the structure which has significant influence on the earthquake input energy, as
observed by many investigators [3]. Thus, the modified energy balance equation can be written as:

γE = ( E e + E p ) (2)

where E e and Ep are, respectively, the elastic and plastic components of the energy needed as the structure
is pushed up to the target drift; Sv is the design pseudo-velocity; M is the total mass of the system. The
modification factor, γ , depends on the structural ductility factor ( µ s ) and the ductility reduction factor
( Rµ ), which is related to the structure’s period. Fig. 1 shows the relationship between the base shear (CW)
and the corresponding drift ( ∆ ) of an elastic system and an elastic-plastic system and Eqn. 2 can be
written as:

1 ⎛1 ⎞
γ C euW∆ e = ⎜ C yW (2∆ max − ∆ y ) ⎟ (3)
2 ⎝ 2 ⎠

Using the expression for drifts ( ∆ ), Eqn. 3 can be rewritten as:

∆ e (2∆ max − ∆ y )
γ = (4)
∆y ∆e

where ∆ e and ∆ max from Fig. 1 are equal to Rµ ∆ y and µ s ∆ y , respectively. Substituting these terms in
Eqn. 4, the energy modification factor γ can be determined as:

2µ s − 1
γ = (5)
Ru2

According to Eqn. 5, the modification factor is a function of the ductility reduction factor and the
structural ductility factor. Using different approaches, many investigators have studied the relationship
between ductility reduction factor ( Rµ ) and structural ductility factor ( µ s ) [4]. In this study, the
relationship suggested by Newmark and Hall [5] is used. Fig. 2 shows the relationship between µ s , Rµ ,
and the structural period, and Fig. 3 shows the resulting values of γ .

The design input energy can be determined from the elastic design pseudo-acceleration spectra as given in
the building codes. In this study, the design is based on the UBC [6] design spectrum which, for elastic
systems, is specified as:

A = Ce g (6)

where A is the design pseudo-acceleration, g is the acceleration due to gravity, and Ce is the normalized
design pseudo-acceleration as defined in Eqn. 1. Note that no occupancy importance factor is included in
the design pseudo-acceleration for the approach proposed. The occupancy importance factor, I, raises the
design force level in an attempt to decrease the drift and ductility demand for the structure for a given
level of ground shaking. However, that cannot be considered as a direct method to achieve the intended
purpose such as damage control. The reduction of potential damage should better be handled by using
appropriate drift limitations. In this regard, the method of calculating the design base shear proposed in
this study uses the target drift as an important parameter. It is assumed that the selected target drift will
have the occupancy importance factor built into it.

The energy balance equation can be rewritten as:

 2
⎛1 ⎞ 1 ⎛ T ⎞
( E e + E p ) = γ ⎜ MS v2 ⎟ = γM ⎜ Ce g ⎟ (7)
⎝2 ⎠ 2 ⎝ 2π ⎠

Akiyama [7] showed that the elastic vibrational energy can be calculated by assuming that the entire
structure is reduced into a single-degree-of-freedom system, i.e.,

2
1 ⎛ T V ⎞
Ee = M⎜ ⋅ ⋅ g⎟ (8)
2 ⎝ 2π W ⎠

where V is the design yield base shear and W is the total seismic weight of the structure (W=Mg).
Substituting Eqn. 8 into Eqn. 7 and rearranging the terms gives:

WT 2 g ⎛⎜ 2 ⎛ V ⎞ ⎞⎟
2
Ep = γC − ⎜ ⎟ (9)
8π 2 ⎜⎝ ⎝ W ⎠ ⎟⎠
e

Figure 1. Structural idealized response and energy balance concept.

 Figure 2. Ductility reduction factors Figure 3. Modification factors for energy
proposed by Newmark and Hall (1982). equation versus period.

NEW LATERAL FORCE DISTRIBUTION FOR PROPOSED DESIGN PROCEDURE

Inelastic dynamic analyses were conducted to study the distribution of maximum story shears. The
nonlinear analysis program SNAP-2DX [8] was used to perform the analyses. The study frames were
subjected to a set of four earthquake records [9]. Fig. 4 shows an example 5-bay, 9 story steel frame. The
relative distributions of maximum story shears due to four selected earthquake records and the UBC static
story shears are compared in Fig. 5 and a considerable difference can be noticed between the UBC lateral
force distribution and that obtained from the analyses. The ratio of the earthquake induced story shear at
level i to that at the top level, n, is assumed to be of the form:

b
⎛V ⎞
β i = ⎜⎜ i ⎟⎟ (10)
⎝ Vn ⎠

where Vi and Vn , respectively, are the static story shears at level i and at the top level as computed from
the design forces given by the UBC lateral force equations. The static story shears, Vi and Vn, at level i and
the top story were assumed as:

⎛ n ⎞
⎜ ∑ wi hi ⎟
⎜ ⎟
Vi = ⎜ ni ⎟ ⋅V (11)
⎜⎜ ∑ w j h j ⎟⎟
⎝ j =1 ⎠

⎛ ⎞
⎜ ⎟
⎜ w h ⎟
Vn = ⎜ n n n ⎟ ⋅V (12)
⎜⎜ ∑ w j h j ⎟⎟
⎝ j =1 ⎠

Substituting Eqs. 11 and 12 into Eqn 10 gives:

 b
⎛ n ⎞
⎜ ∑ wi hi ⎟
β i = ⎜⎜ i ⎟
⎟
(13)
w h
⎜ n n ⎟
⎝ ⎠

where V is the base shear, wi (or wj) is the weight of the structure at level i (or j), hi (or hj) is the height of
the structure at level i (or j); wn is seismic weight of the structure at the top story and hn is the height of top
story. An equation of 0.5T -0.2 for the exponent b in Eqn. 13 was found to give a good fit with the analysis
results as shown in Fig. 5 [9]. Thus, the proposed lateral force applied at the top story, Fn, is given by:

0.5T −0.2
⎛ ⎞
⎜ ⎟
⎜ w h ⎟
Fn = V ⋅ ⎜ n n n ⎟ (14)
⎜⎜ ∑ w j h j ⎟⎟
⎝ j =1 ⎠

The force applied at level, Fi , can be written as:

Fi = ( β i − β i +1 ) Fn (15)

where β i and β i +1 , respectively, are the shear proportioning factors at level i and level i + 1 . Note when
i = n, β i +1 = 0 .

9 El Centro

W24x68 W24x68 W24x68 W24x68 W24x68 Newhall

8
W27x84 W27x84 W27x84 W27x84 W27x84
Sylmar

7 Synthetic
W30x99 W30x99 W30x99 W30x99 W30x99

6
W36x135 W36x135 W36x135 W36x135 W36x135 (Vi/Vn )
8 stories @ 13'-0"

-UBC
Story Level

W36x135 W36x135 W36x135 W36x135 W36x135 5

0.5T^(-0.2)
(Vi/Vn )
W36x135 W36x135 W36x135 W36x135 W36x135 4

W36x135 W36x135 W36x135 W36x135 W36x135

3

W36x160 W36x160 W36x160 W36x160 W36x160

2
W36x160 W36x160 W36x160 W36x160 W36x160
18'-0"

0
0 1 2 3 4
5 equal bays @ 30'-0"
Vi/Vn
Original Frame

Figure 4. Example 5-bay, 9 story frame. Figure 5. Distribution of maximum story shears
due to selected ground motions.
 DESIGN BASE SHEAR BASED ON TARGET DRIFT

By using a pre-selected yield mechanism as shown in Fig. 6 and equating the plastic energy term Ep to the
external work done by the design lateral forces gives:

n
E p = ∑ Fi hiθ p (16)
i =1

where θ p is the global inelastic drift of the structure, which is the difference between the target drift ( θ u )
and yield drift ( θ y ). Substituting Eqs. 9 and 15 into Eqn. 16, and solving for V/W gives:

V − α + α + 4γC e
2 2
= (17)
W 2

where V is the design base shear and α is a dimensionless parameter, which depends on the stiffness of the
structure, the modal properties and the intended drift level, and is given by:

0.5T −0.2
⎛ ⎞
⎜ ⎟
⎛ n
⎞ ⎜ wn h n ⎟ ⎛ θ p 8π 2 ⎞
α = ⎜⎜ ∑ ( β i − β i +1 )hi ⎟⎟ ⋅ ⎜ n ⎟ ⋅⎜ 2 ⎟ (18)
⎝ i =1 ⎠ ⎜ w h ⎜ T g ⎟
⎜∑ j j ⎟⎟ ⎝ ⎠
⎝ j =1 ⎠

Eqn.17 gives the required design base shear corresponding to an intended inelastic drift level, θp,
and pre-selected yield mechanism.

Figure 6. Steel moment frame in the target drift response state with the pre-selected yield mechanism.
 PLASTIC DESIGN OF MOMENT FRAMES

It is desirable that the required distribution of beam strength along the height closely follow the
distribution of story shears induced by design lateral forces. Therefore, the required beam strength at each
level can be determined as follows:

n n n

∑ 2β M
i =1
i pbr + 2M pc = ∑ Fi hi = ∑ ( β i − β i +1 )hi Fn
i =1 i =1
(19)

where Mpbr is the reference plastic moment of beams and the only unknown variable in the above
equation. The required plastic moment capacity of the first-story columns to prevent the story mechanism
in the first story can be taken as [1]:

1.1Vh1
M pc = (20)
4

where V is the total base shear, h1 is the height of the first story, and the factor 1.1 is the overstrength
factor to account for possible overloading due to strain hardening [1]. In order to ensure the selected
strong column-weak beam plastic mechanism at the ultimate drift level, it is essential that columns are
designed by assuming that all beam plastic hinges are fully yielded and strain-hardened when the drift is at
the target ultimate level. The detailed procedure for designing the columns can be found in the previous
study [1].

SEISMIC EVALUATION OF THE STUDY FRAME

The three example frames were designed using the proposed design procedure. Fig. 7 shows the resulting
member sizes of these frames. In this study, the frames were designed by selecting an assumed 2% target
drift, and all members to have specified yield strength of 50 ksi. The members were designed by using
AISC-LRFD specification. Design parameters for the example frames are shown in Table 1.

A series of non-linear analyses including inelastic static analysis and inelastic time history analysis was
conducted to investigate the response of these frames. All analyses were carried out by using the computer
program SNAP-2DX [8]. Strain-hardening and viscous damping values of 2 % were used for all members.
Fig. 8 shows the base shear versus roof displacement plot for the frames obtained from the static pushover
analysis. As can be seen, the yield drift and the design base shear of the frame are very close to the values
assumed in the design stage as shown in Table 1. Fig. 9 shows the envelopes of maximum story drifts of
the three frames due to four selected ground motions. It can be noticed that the story drifts are generally
within the target design limit of 2%, as expected.

The locations of inelastic activity and the rotational ductility demands at plastic hinges are shown in Fig.
10. As expected, the plastic hinge in the frames formed only in the beams and at the column bases without
developing any undesirable mechanism, such as a soft story mechanism. It was observed that the plastic
hinges at the column bases formed later than those in the beams in most cases, and the rotational ductility
demands at the column bases are much smaller than those in the beams. This suggests that the chances of
premature failure at the column base can be significantly reduced by designing the structure according to
the proposed methodology.

Table 1. Design Parameters (2% Target Drift Limit) of the Three Structures.

Number of Period Ce Assumed θp γ α V/W

Stories (sec.) Yield Drift
3 0.546 1.100 0.01 0.01 0.820 2.775 0.325
9 1.285 0.635 0.01 0.01 0.750 1.505 0.179
20 2.299 0.431 0.0075 0.0125 0.609 1.272 0.083

W30x90 W30x90 W30x90 W30x90

3 stories @ 13'-0"

W30x108 W30x108 W30x108 W30x108

W14x211
W14x211

W14x370

W14x370

W14x370
W30x116 W30x116 W30x116 W30x116

4 equal bays @ 30'-0"

(a) 3-Story Frame

W30x90 W30x90 W30x90 W30x90 W30x90

W14x257

W14x426

W14x426

W14x426

W14x426

W14x257
W30x116 W30x116 W30x116 W30x116 W30x116

W33x130 W33x130 W33x130 W33x130 W33x130

W14x342

W14x550

W14x550

W14x550

W14x550

W14x342

W36x135 W36x135 W36x135 W36x135 W36x135

8 stories @ 13'-0"

W40x149 W40x149 W40x149 W40x149 W40x149

W14x426

W14x665

W14x665

W14x665

W14x665

W14x426

W40x149 W40x149 W40x149 W40x149 W40x149

W40x149 W40x149 W40x149 W40x149 W40x149

W14x455

W14x665

W14x665

W14x665

W14x665

W14x455

W40x149 W40x149 W40x149 W40x149 W40x149

W40x149 W40x149 W40x149 W40x149 W40x149

W14x500

W14x665

W14x665

W14x665

W14x665

W14x500
18'-0"

5 equal bays @ 30'-0"

(b) 9-Story Frame

 19 stories @ 13'-0"
18'-0"

5 equal bays @ 20'-0"

(c) 20-Story Frame

Figure 7. Member sizes of the 3, 9, and 20-Story frames designed for 2% target drift.
 0.5

Base Shear Coefficient (V/W)

0.4 3-Story
9-Story
0.3 20-Story

0.2

0.1

0
0 1 2 3 4
Roof Drift (%)

Figure 8. Base shear versus roof drift responses from nonlinear pushover analysis.

El Centro Newhall
Sylmar Synthetic
4
3-Story
3
Story

2 20
1 20-Story

0
0 1 2 3 4
15
10
9 9-Story
Story

8
7 10
6
Story

5
4 5
3
2
1
0 0
0 1 2 3 4 0 1 2 3 4
Story Drift (%) Story Drift (%)

Figure 9. Maximum story drifts of the frames due to selected earthquake records.
El Centro Newhall

Sylmar Synthetic

(a) 3-Story

El Centro Newhall

Sylmar Synthetic

(b) 9-Story

Rotational Ductility Demands: 1.0-1.5, 1.5-2.0, 2.0-2.5, 2.5-3.0

 El Centro Newhall

Sylmar Synthetic
(c) 20-Story

Rotational Ductility Demands: 1.0-1.5, 1.5-2.0, 2.0-2.5, 2.5-3.0

Figure 10. Location of inelastic activity in 3, 9, and 20-Story frames under the four selected earthquakes.
 CONCLUSION

A new performance-based seismic design procedure using modified energy balance concept and plastic
design methodology is proposed. The modified energy balance equation accounts for the structural
ductility factor ( µ s ) and the ductility reduction factor ( Rµ ), which is related to a structure’s period. A
new design lateral force distribution based on nonlinear inelastic dynamic analysis results was also
developed. It is noted that, in the proposed design method, the target story drift is specified as a key design
parameter and, therefore, no further check for drift is required.

In order to validate the proposed method, three steel moment frames with 3, 9, and 20 stories were
designed. Nonlinear static and dynamic analyses were conducted and the results show that the proposed
method can produce structures that meet preselected performance objectives in terms of yield mechanism
and target drift.

REFERENCES

1. Leelataviwat, S. “Drift and Yield Mechanism based Seismic Design and Upgrading of Steel
Moment Frames.” Ph.D. Thesis, Department of Civ. & Env. Engrg., University of Michigan, Ann
Arbor, MI, USA, 1998.
2. Lee, Soon-Sik and Goel, S. C. “Performance-Based Design of Steel Moment Frames Using Target
Drift and Yield Mechanism.” Report No. UMCEE 01-17, Department of Civ. & Env. Engrg.,
University of Michigan, Ann Arbor, MI, USA, 2001.
3. Uang, C.-M. and Bertero, V.V. “Use of Energy as a Design Criterion in Earthquake-Reistant
Design.” Report No. UCB/EERC-88/18, Earthquake Engrg. Res. Ctr., University of California,
Berkeley, CA, USA, 1988.
4. Miranda, E. and Bertero, V.V. “Evaluation of Strength Reduction Factors for Earthquake-
Resistant Design.” Earthquake Spectra, EERI, 1994; 10(2): 357-380.
5. Newmark, N.M. and Hall, W.J. “Earthquake Spectra and Design.” Earthquake Engrg. Res. Inst.,
El Cerrito, CA, 1982.
6. Uniform Building Code (UBC), Int. Conf. of Bldg. Officials, Whittier, Calif, 1997.
7. Akiyama, H. “Earthquake-Resistant Limit-State Design of Buildings.” University of Tokyo Press,
Japan, 1985.
8. Rai, D.C., Goel, S.C., and Firmansjah, J. “SNAP-2DX: A General Purpose Computer Program for
Nonlinear Structural Analysis.” Report No. UMCEE 96-21, Dept. of Civ. & Env. Engrg.,
University of Michigan, Ann Arbor, MI, USA, 1996.
9. Lee, Soon-Sik and Goel, S.C. “A New Lateral Force Distribution for Seismic Design of Steel
Structure.” Proceedings of U.S.-Japan Workshop on Seismic Fracture Issues in Steel Structures,
San Francisco, CA, February 28-March 1, 2000.