Available online at www.sciencedirect.

com

ScienceDirect
Procedia Engineering 125 (2015) 1050 – 1056

The 5th International Conference of Euro Asia Civil Engineering Forum (EACEF-5)

Performance of force based design versus direct displacement based

design in predicting seismic demands of regular concrete special
moment resisting frames
Ima Muljatia,*, Fransiscus Asisia, Kevin Willyantoa
a
Civil Engineering Department, Petra Christian University, Siwalankerto 121-131 Surabaya 60236, Indonesia

Abstract

As an approximate method in seismic design, direct displacement based design (DDBD) has been widely accepted in conjuction
with the former method, force based design (FBD). In its procedure FBD uses building displacement as the final check to
determine the structural performance while DDBD use it as the designed target performance. If the final displacement in FBD
larger than the value specified by the standard then the design procedure should be recalculated. On the other hand, under some
common practices the procedure of DDBD is more straight forward compared to FBD. Unfortunately, the prospective use of
DDBD is not applied well, especially in Indonesia. Therefore, the objective of this study is to evaluate the performance of DDBD
on a regular concrete special moment resisting frame compared to two varians of FBD, equivalent lateral force procedure and
response spectrum analysis. All methods are designed using the latest Indonesian seismic code and verified using the exact
method nonlinear time history analysis. All design method are run in a single cycle of design without any effort to improve the
performance level to experience the effectiveness of the each metod in predicting seismic demand. The parameters used for
evaluating structural performance are story drift, damage indices, and structural failure mechanism. As the results, DDBD
performed better than FBD in predicting story drift. All methods experience excellent damage indices level. Although all
methods show tendency poor mechanism, DDBD needs no improvement because the structures fulfill the targeted peformance
while FBD needs extra effort to improve the structural performance due to the design procedure should be repeated from the
beginning. In this point of view, DDBD procedure is more effective compared to FBD.
© 2015
© 2015TheTheAuthors.
Authors.Published
Publishedby by Elsevier
Elsevier Ltd.Ltd.
This is an open access article under the CC BY-NC-ND license
Peer-review under responsibility of organizing committee of The 5th International Conference of Euro Asia Civil Engineering
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of organizing committee of The 5th International Conference of Euro Asia Civil Engineering
Forum (EACEF-5).
Forum (EACEF-5)

* Corresponding author. Tel.: +62-816 504 496; fax: +62-31-298-3392.

E-mail address: imuljati@petra.ac.id

1877-7058 © 2015 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of organizing committee of The 5th International Conference of Euro Asia Civil Engineering Forum (EACEF-5)
doi:10.1016/j.proeng.2015.11.161
 Ima Muljati et al. / Procedia Engineering 125 (2015) 1050 – 1056 1051

Keywords: direct displacement based design; force based design; equivalent lateral load procedure; response spectrum analysis; performance
based design.

1. Introduction

Indonesian seismic code SNI 03-1726-2012 [1] which is adopted from ASCE 7-10 [2] permits three options of
analytical procedure so called equivalent lateral force analysis (ELF), modal response spectrum analysis (MRS), and
seismic response time history analysis (THA) for designing regular concrete moment resisting frame systems
(MRF). ELF and MRS are based primarily on forces i.e. force based design (FBD). Forces are induced by the
earthquake resulting displacement in the structures. During elastic stage, these forces are related to the elastic
stiffness of the system, but in inelastic condition the relationship becomes more complex, depending on the
displacement history throughout the excitation. Therefore force considerations are important in FBD. Structural
strength should exceeds the design applied loads to avoid structural collapse.
Ahead of the 80s, seismic design recognised that strength was less important compared with ductility. A ductile
structure able to deform inelastically responding to the earthquake without loss of strength, although it has been
designed using a lower design strength. Thus, it is common to use the reduced design force level in the FBD
procedure. In 90’s, some problems are found in the application of FBD, mainly due to interdependency between
strength and stiffness. Member sizes should be determined in the early stage of design, then forces are distributed
among members in proportion to their assumed stiffness. If member sizes are modified, then the calculated design
forces will no longer be valid, and the design process should be recalculated.
Priestley et al [3] founds at least three main weakness of FBD. Firstly, FBD relies on assumption of initial
stiffness to determine the structural period and the distribution of design forces among different structural elements.
Since the stiffness is dependent on the strength of elements, this cannot be known until the design process is
complete. Secondly, allocation seismic force among elements based on initial stiffness is illogical for many stuctures
because it incorrectly assumes that the different elements can be forced to yield simultaneously. Thirdly, there is no
unique force-reduction factors (based on ductility capacity) for a given structural type and material.
Some effort has been developed to overcome the deficiencies of FBD. Initially the approaches were designed to
fit within, and improve the existing FBD until finally leading to the more realistic approach which is based on
displacement consideration. Among several of them, direct displacement based design (DDBD) method proposed by
Priestley et al [3] seems to be the most promising and has been codified [4] refering to some clauses in Eurocode 8
[5]. The aim of DDBD is to design a structure which would achieve, rather than be bounded by a given performance
limit state under a given seismic intensity.
DDBD determines the strength required at designated plastic hinge locations to achieve the targeted design in
terms of defined displacement objectives. It must then be combined with capacity design procedures to ensure that
plastic hinges occur only where intended, and that non-ductile modes of inelastic deformation do not develop. These
capacity design procedures must be calibrated to the DDBD approach based on [3].
The comparison of both methods has been done in several cases [6‒8] and mostly reported that DDBD has
succesfully overcome the drawback of FBD including in the Indonesian case [9]. However, Indonesian seismic
design code has been revised, then the effectiveness of DDBD should be rechecked altogether with FBD. Therefore,
the objective of this study is to evaluate the performance of DDBD compared with two varians of FBD i.e.
equivalent lateral force procedure and response spectrum analysis in predicting seismic demands of a concrete
special moment resisting frame (SMRF).

2. Case Study

A six story concrete MRF building on site class E with typical plan, and typical story height of 3.15 m shown in
Fig.1(a) are chosen as case study. The design response spectra are shown in Fig.1(b), represents low and high
seismicitiy level in Indonesia (EQ1 and EQ2). Each structure is designed as a special moment resisting frame using
two varians of FBD which are equivalent lateral force analysis (ELF) and modal response spectrum analysis (MRS)
as well as direct displacement-based design (DDBD) method, regarding to [1] and [3] for damage control condition.
1052 Ima Muljati et al. / Procedia Engineering 125 (2015) 1050 – 1056

Concrete and steel strengths are 25 MPa and 400 MPa respectively. The internal forces are generated by ETABS
[10]. The story drift design target is set 2% for all methods.
All methods are designed using the latest Indonesian seismic code [1] and verified using the exact method
nonlinear time history analysis. All design methods are run in a single cycle of design without any effort to improve
the performance level to experience the effectiveness of the method. The parameters used in evaluating structural
performance are story drift, damage indices, and structural failure mechanism.
y

4200
4200
19,100
2300
4200
4200

3900 2800 3700 3300 3700 2800 3900

24,100

(a) (b)

Fig. 1. (a) building plan; (b) design response spectrum.

Nonlinear time history analysis is conducted by SeismoStruct V.6.5 [11] with the consistent ground acceleration
spectrum (EQ1 and EQ2), modified from N-S component of El-Centro 1940. The modification is achieved using
Seismomatch [12]. Moment-rotation relationship is modeled as biliner using Cumbia [13]. The acceptance criteria
for evaluating structural performance are based on damage control condition (consistent with the targeted design
performance) involving some parameters including story drift, damage indices, and failure mechanism of the
structure.

3. Results and Discussion

3.1. Beams and Columns Dimension

The detailed structural dimension resulting from these three methods are shown in Table 1. There are significant
differences among beams and columns dimension resulting from each method, especially between FBD (ELF and
MRS) and DDBD. Following the design procedures of FBD and DDBD, member dimensions are predicted in the
beginning of the design process. However, in FBD the dimension adequacy can only be checked in elastic condition
due to ultimate loading. While in DDBD, the dimension adequacy can be determined accurately in the begining of
the design process based on the ductility demand of each member under the targeted displacement in inlastic
condition , i.e. 2% of story drift.

3.2. Story Drift

Story drift is the ratio of relative displacement between two consecutive story and story height. The results of
these three methods in both directions are shown in Fig.2. DDBD succesfully results story drift less than 2% in all
stories in both direction because the structure is deliberately designed to achieve 2% story drift. Conversely, FBD
(i.e. ELF and MRS) failed to fulfill the limitation of 2% story drift especially in lower stories. Thus, FBD unable to
 Ima Muljati et al. / Procedia Engineering 125 (2015) 1050 – 1056 1053

predict the story drift effectively. At this stage, the design procedure of FBD should be recalculated in order to
improve the story drift performance. In this case, enlarge the columns dimension would be preferable option.
However, the study did not look further on the matter because the main objectives of the study is only to investigate
the effectiveness of each design procedure. It also answer why the structural dimension of DDBD larger than FBD
in section 3.1.

Table 1. Dimensions.
Low Seismicity (EQ1) High Seismicity (EQ2)
Design Method
ELF (mm) MRS (mm) DDBD (mm) ELF (mm) MRS (mm) DDBD (mm)
Main Beam 250×500 250×500 500×700 400×800 400×800 500×700
Secondary Beam 250×300 250×300 250×300 250×300 250×300 250×300
Column 500×500 500×500 600×600 600×600 600×600 800×800
Note: ELF: equivalent lateral force design; MRS: modal response spectrum design; DDBD: direct displacement based design.

Story Drift: X-dir. Story Drift: Y-dir

6 6

5 5
ELF-EQ1 ELF-EQ1
4 4
MRS-EQ1 Story MRS-EQ1
Story

3 DDBD-EQ1 3 DDBD-EQ1
ELF-EQ2 ELF-EQ2
2 2
MRS-EQ2 MRS-EQ2
1 DDBD-EQ2 1 DDBD-EQ2
Limit/Target Limit/Target
0 0
0 0.5 1 1.5 2 2.5 3 0 0.5 1 1.5 2 2.5 3
Story Drift (%) Story Drift (%)

(a) (b)

Fig. 2. (a) story drift in x-direction; (b) story drift in y-direction.

3.3. Damage Index

Damage index is a parameter to measure the level of damage of structural member. It is defined as the rotation
demand divided by the available rotation capacity which can be calculated using:

Tn  T y
DI (1)
Tu  T y

where Tn is the rotation occured at the member, Ty and Tu are the yield- and ultimate-rotation of the member. Table 2
and 3 show the damage indices of beams and columns of the observed structure resulting from the three methods.
All methods able to meet the targeted performance i.e. damage control condition, which is limited by the damage
index ranged from 0.25 to 0.40. Although the member dimension of ELF and MRS lower than the required
dimension to fulfill 2% story drift, both methods still produce acceptable damage. Only beams in high seismicity
region (EQ2) capable of reaching damage control limit state, while all columns and beams in low seismicity region
(EQ1) are still in serviceability limit state. It seems that strong-column-weak-beam condition is well maintained.
The pastic hinge locations are discussed further in the subsequent section.
1054 Ima Muljati et al. / Procedia Engineering 125 (2015) 1050 – 1056

Table 2. Damage indices for beams, both in x- and y- direction.

Performance level
EQ level Design Method
First Yield Serviceability Damage Control Safety Unacceptable
EQ1 ELF 9
MRS 9
DDBD 9
EQ2 ELF 9
MRS 9
DDBD 9 (x-dir) 9 (y-dir)
Damage Index Limitation < 0.10 0.10 – 0.25 0.25 – 0.40 0.40 – 1.00 > 1.00
Note: design target is damage control limit state

Table 3. Damage indices for columns, both in x- and y-direction.

Performance level
EQ level Design Method
First Yield Serviceability Damage Control Safety Unacceptable
EQ1 ELF 9
MRS 9
DDBD 9
EQ2 ELF 9
MRS 9
DDBD 9
Damage Index Limitation < 0.10 0.10 – 0.25 0.25 – 0.40 0.40 – 1.00 > 1.00
Note: design target is damage control limit state

3.4. Failure Mechanism

Table 4 shows plastic hinges formation at the worst condition in interior/exterior frames resulted by nonlinear
time history analysis. The others configuration can be found in detail in [14]. All structures fail to meet the beam-
side-sway-mechanism as expected which is only allow plastic hinges formation at three locations, i.e. all beams end,
top columns at top story, and base columns. For low seismicity zone (EQ1), FBD results soft story mechanism at the
first story in both direction, leading to unsafe failure mechanism. The forbidden plastic hinge locations are denoted
by the arrow in Table 4. In contrast, DDBD performed well and show a safe mechanism in both direction.
Unfortunately, in high seismicity zone (EQ2), all methods result poor mechanism leadings to soft story mechanism.
It is well noted that the application of capacity design procedure indeed significantly reduces the risk of failure
for both method. All structures show that the condition of strong-column-weak-beam are well maintained. Yield
occurence starting from the beams followed by the columns. Some beams in highly seismic zone reach damage
control limit state but columns only reach serviceability limit state. Although structures experience soft story
mechanism, they do not collapse during the excitation. However, soft story mechanism at the first story of FBD
structures cause large floor displacement in the first- and second-stories resulting large story drift, exceedings the
limitation of 2% as discussed in section 3.2. While in DDBD structures which use larger member dimension, soft
story mechanism does not significantly affect the structural performance, structures successfully meet the story drift
provision without any significant damage at their structural member.
 Ima Muljati et al. / Procedia Engineering 125 (2015) 1050 – 1056 1055

Table 4. Plactic hinges configuration at the most severe frame.

Earthquak Design Method
e Level
ELF MRS DDBD

EQ1
Interior
Frame

x-dir x-dir y-dir

EQ1
Exterior
Frame

y-dir x-dir x-dir

EQ2
Interior
Frame

x-dir x-dir y-dir

EQ2
Exterior
Frame

x-dir x-dir x-dir

Limit state condition: first yield serviceability limit state damage control limit state
Arrow (Æ) denotes forbidden plastic hinge location.
1056 Ima Muljati et al. / Procedia Engineering 125 (2015) 1050 – 1056

In order to improve the structural performance, FBD involves extra effort due to the design process should be
repeated from the beginning, using a new and different structural member dimension. Consequently, the structural
performance should be rechecked again using non-linear time history analysis. At the worst case, the design process
could be involving several design cycles until it meets the targeted design performance. Asisi and Willyanto in their
study [14] reported that at least four design cycles needed to improve the performance of FBD structure. This is the
most drawback of FBD. There is no guarantee the design meets the expected performance. On the other hand, no
design repetition process needed by DDBD because structures are deliberately designed to meet the targeted
performance. And this study has proven that DDBD performed well and able to meet the expected performance.

4. Concluding Remarks

The study has explained how to evaluate the performance of force based design (FBD) and direct displacement
based design (DDBD) on a concrete special moment resisting frame. It can be concluded that DDBD performed
better than FBD in predicting seismic demand of the structure i.e. story drift because it deliberately design the
structure to achieve a given performance limit state. In contrast, FBD requires several design process repetitions in
order to achieve acceptable performance specified by the code. Therefore, in this point of view DDBD procedure is
more effective than FBD in predicting seismic demand of a concrete special moment resisting frame. Also well
noted here that the application of capacity design procedure in both method is successfully protecting structure from
unsafe failure mechanism.

References

[1] Departemen Pekerjaan Umum, Tata cara perencanaan ketahanan gempa untuk struktur bangunan gedung dan non-gedung. Badan
Standarisasi Nasional, Jakarta, 2012.
[2] ASCE/SEI 7-10, Minimum design loads for buldings and other structures, American Society of Civil Engineers, Virginia, 2010.
[3] M.J.N. Priestley, G.M. Calvi, and M.J. Kowalsky, Displacement-based seismic design of structures, IUSS Press, Pavia, 2007.
[4] T.J. Sullivan, M.J.N. Priestley, and G.M. Calvi, A model code for the displacement-based seismic design of structures, IUSS Press, Pavia,
2012.
[5] Eurocode 8, Design of structures for earthquake resistance, EN 1998 part 1, Comite Europeen de Normalisation, CEN, Brussels, 2004.
[6] D. Cardone, M. Dolce, and G. Palermo, Force-based vs. direct displacement-based design of buildings with seismic isolation, The 14th
World Conference on Earthquake Engineering, Beijing, 2008.
[7] M.A. Rahman and S. Sritharan, An evaluation of force-based design vs. direct displacement-based design of jointed precast post-tensioned
wall systems, Earthquake Engineering and Engineering Vibration, Vol.5, No.2, pp 285-296, Springer, 2006.
[8] J. Goggins and S. Salawdeh, Comparison of DDBD with FBD procedures for concentrically braced steel frame, The 15th World
Conference on Earthquake Engineering, Lisboa, 2012.
[9] I. Muljati, I.W. Susanto, and P. Rantetana, Direct displacement-based design on reinforced concrete moment resisting frame in Indonesia,
Proceeding the 6th Civil Engineering Conference in Asia Region: Embracing the Future through Sustainability ISBN 978-602-8605-08-3,
paper No.86, pp 14-20, Jakarta, 2013.
[10] Computer and Structures, Inc., ETABS v9.0.7, Extended Three Dimensional Analysis of Building System. Berkeley, California, 2005.
[11] SeismoSoft., SeismoStruct – A Computer Program for Static and Dynamic Nonlinear Analysis of Frame Strcuture v.6.5, IUSS Press, Pavia,
2007.
[12] SeismoSoft. SeismoMatch – User Manual Handbook v.2.1, IUSS Press, Pavia, 2013.
[13] L.A. Montejo, CUMBIA, Department of Civil, Construction, and Environmental Engineering, North Carolina State University, 2007.
[14] F. Asisi and K. Willyanto, Perbandingan kinerja bangunan yang didesain dengan force-based design dan direct displacement-based design
menggunakan SNI Gempa 2012, undergraduate thesis, Petra Christian University, Civil Engineering Department, Surabaya, 2014.