TECHNICAL NOTE

An Anomaly in the Equivalent

Linearization Approach for the Estimation
of Inelastic Response
Vijay Namdev Khosea) E.Aff.M.EERI, and
Yogendra Singhb) E.Aff.M.EERI

The estimation of inelastic displacement is an essential step in displacement-

based design. Estimation of inelastic response using response spectra is a simpler
and attractive alternative to nonlinear dynamic analysis. A number of methods
based on different approaches are now available to estimate nonlinear res-
ponse from design spectra. Equivalent linearization is one of the widely used
approaches for this purpose. The approach is known to yield reasonably accurate
results. However, an anomaly in the approach is observed, which results in sig-
nificantly non-conservative estimates of inelastic displacement, particularly in
displacement-controlled spectral range. This paper illustrates the anomaly
using displacement modification factors and numerical examples of oscillators
of different periods and ductility. The results from the equivalent linearization
approach have been compared with the results of nonlinear dynamic analysis
and yield spectrum approach. [DOI: 10.1193/031112EQS088T]

INTRODUCTION
Displacement-based design (DBD) methods for seismic actions are slowly gaining popu-
larity over the conventional force-based design. The estimation of the peak displacement of
an equivalent inelastic single-degree-of-freedom (SDOF) system is an integral part of DBD.
Although nonlinear dynamic (time history) analysis is the most accurate method of estimat-
ing inelastic displacement, it is time-consuming and complex. Further, it requires the selec-
tion of appropriate earthquake records, whereas most of the national seismic design codes
specify only design response spectra without adequate guidelines for the selection of design
ground motions. Under such conditions, the estimation of inelastic response (peak response
of the inelastic system) using elastic response spectrum is a simpler and attractive alternative.
Different approaches are now available to estimate peak inelastic displacement using the
response spectrum. The equivalent linearization (EL) approach (Gülkan and Sozen 1974,
Iwan 1980, Kowalsky 1994, Kwan and Billington 2003, Grant et al. 2005, Dwairi et al.
2007, Priestley et al. 2007, Pennucci et al. 2011) is one of the popular approaches for esti-
mating inelastic displacement. In this approach, the inelastic response is obtained from the
response of an equivalent elastic SDOF system having an appropriate period and damping.
The other approaches to estimate inelastic response include the yield spectrum (YS), or

a)
Senior Engineer, Thornton Tomasetti, Inc., Mumbai-400013, India
b)
Professor, Department of Earthquake Engineering, IIT Roorkee, Roorkee-247667, India

965
Earthquake Spectra, Volume 30, No. 2, pages 965–972, May 2014; © 2014, Earthquake Engineering Research Institute
966 V. N. KHOSE AND Y. SINGH

constant ductility, approach (Newmark and Hall 1982, Riddell et al. 1989, Krawinkler and
Nassar 1992, Miranda 1993, Vidic et al. 1994, Ordaz and Pérez-Rocha 1998, Miranda 2000,
Riddell et al. 2002, Cuesta et al. 2003, Chopra and Chintanapakdee 2004, Ruiz-García and
Miranda 2004) and the displacement modification approach (ASCE 41-06 2007). It can be
shown that with the recently developed models, the inelastic displacements obtained using
different approaches tend to converge, except an anomaly in the EL approach in long periods
(displacement-controlled) spectral region. This paper illustrates the anomaly in the EL
approach by comparing displacement modifications factors (ratios of peak inelastic to
peak elastic displacement) with those obtained from the YS approach and an analytical
study using nonlinear time history analysis. Numerical examples are also presented to
demonstrate the effect of this anomaly on estimated peak inelastic displacement.

EQUIVALENT LINEARIZATION APPROACH

In the EL approach, the behavior of the inelastic system is represented by an equivalent
linear system having equivalent damping, ξeq , and an effective or equivalent period, T eq .
A number of EL models are available, which provide expressions for equivalent damping
and effective periods, which either corresponds to secant stiffness at peak displacement, or an
optimal stiffness between initial and secant stiffness. Once the equivalent damping is esti-
mated, damping reduction factor, Rξ , can be obtained using available relationships (Newmark
and Hall 1982, CEN 1998, CEN 2004). Recently, Pennucci et al. (2011) have observed that
the damping reduction factor has substantial random variability and that it is more appropriate
to use displacement reduction factors (DRF) obtained directly from the ductility and hyster-
esis type, without estimation of equivalent damping as an intermediate step. They have also
noted that the slope of the elastic demand spectrum between initial and effective periods has a
significant effect on the values of DRF in long-period ranges. However, the model proposed
by them does not incorporate this effect.
Figure 1 shows inelastic displacement response spectra (IDRS) obtained from a few
selected EL models for ductility values of 2%, 4%, 6%, and 8%, and 5% elastic damping.
ASCE 7-10 (2010) design spectrum for a PGA value of 0.4 g on site class B has been con-
sidered as the basis elastic spectrum, assuming the long-period transition period (also known
as the corner period between velocity-controlled and displacement-controlled spectral
ranges), T L , as 6 sec. In the case of all the models, elasto-plastic hysteresis type has
been considered, except in the case of the Pennucci et al. (2011) model, where the
Takeda-Thin hysteresis type has been considered, as the model for elasto-plastic hysteresis
type is not available. The spectra are normalized by maximum elastic spectral displacement
and plotted against the effective period. It is observed that the Priestley et al. (2007) and
Pennucci et al. (2011) models estimate quite close IDRS for all ductility values. Further,
the inelastic displacement estimated by Pennucci et al. model is generally conservative com-
pared to other models, except for a ductility value of 2, wherein the Grant et al. (2005) model
predicts slightly higher inelastic displacement.
As noted earlier, different EL models consider different definitions of effective periods,
and hence a more rational comparison can be made by transforming the IDRS to the initial
period using the relationships between initial and effective periods. Figure 2 shows the trans-
formed displacement spectra against the initial period for ductility values of 2, 4, 6, and 8.
ANOMALY IN THE EL APPROACH FOR THE ESTIMATION OF INELASTIC RESPONSE 967

Figure 1. Normalized inelastic displacement spectra obtained from various equivalent lineariza-
tion models, for 0.4 g PGA, 5% elastic damping and varying ductility, μ. The spectra are normal-
ized by the peak elastic spectral displacement.

An interesting observation about the transformed spectra is the dependence of corner periods
on ductility and the definition of effective period. As the effective period in the EL models is
a function of ductility, it increases (shifts) with ductility, but the corner period in the elastic
spectra remains constant for all the values of damping. This results in the observed shift in the
corner period when the spectra are transformed to the initial period. Further, the magnitude of
shift depends on the definition of the effective period and two different classes of EL models
based on the two definitions of the effective period are distinctly visible in Figure 2. This shift
in the corner period leads to the anomaly in long-period spectral ranges, resulting in much
lower estimates of inelastic displacement.

THE ANOMALY
As observed earlier, the effective period of the inelastic systems shifts with ductility, but
there is no shift in the corner period in the elastic response spectra for different damping. This
results in an anomaly, causing the reduction of inelastic displacement in the long-period
(displacement-controlled) range. This anomaly can be better illustrated by plotting displace-
ment modification factors (DMFs) obtained from the EL approach, YS approach, and ana-
lytical study, plotted against the initial period. As noted earlier, the Pennucci et al. (2011) and
Priestley et al. (2007) models yield generally conservative results compared to other EL
968 V. N. KHOSE AND Y. SINGH

Figure 2. Normalized inelastic displacement spectra transformed to initial period, obtained

from various equivalent linearization models, for 0.4 g PGA, 5% elastic damping, and varying
ductility, μ. The spectra are normalized by the peak elastic spectral displacement.

models, and thus these models have been chosen for comparison. Similarly, in the YS
approach, the Chopra and Chintanapakdee (2004) and Miranda (2000) models are the latest
models and yield almost identical results, and these have been considered as representative
YS models for the comparative study.
The DMF, which is also known as inelastic displacement ratio (Miranda 2000, Chopra
and Chintanapakdee 2004), is defined as the ratio of the inelastic spectral displacement to
elastic spectral displacement and can serve as the common basis for the comparison of dif-
ferent approaches. In the case of analytical study and YS approach, the DMF is obtained
directly from the estimated elastic and inelastic displacements, whereas in the case of the
EL approach, the dependence of the effective period on ductility is to be considered, and
the DMF for a given ductility is expressed as:

ðΔinelastic ÞT e
DMF μ ¼ (1)
ðΔelastic ÞT
EQ-TARGET;temp:intralink-;sec3;41;162

where ðΔelastic ÞT is the elastic displacement at initial period, T; and ðΔinelastic ÞT e is the inelastic
displacement at corresponding effective period, T e .
ANOMALY IN THE EL APPROACH FOR THE ESTIMATION OF INELASTIC RESPONSE 969

Figure 3. Target elastic spectrum (5% elastic damping) considered for study.

The analytical study has been conducted using an ensemble of five recorded ground
motions made compatible with ASCE 7-10 (2010) spectrum (Figure 3) for PGA value of
0.4 g on site class B and the long-period transition period, T L , equal to 6 sec. The time his-
tories have been selected from the PEER Ground Motion Database (PGMD 2011) to best
match with the target spectrum. In the selection of ground motions, the magnitude is chosen
in the range 6.5–7, consistent with long-period transition period, T L , equal to 6 sec (FEMA
2003); and shear wave velocity, V S30 is chosen as 760–1;500 m∕sec (corresponding to site
class B). The selected time histories have been further modified using a wavelet-based tool
developed by Mukherjee and Gupta (2002) to make them compatible with the target spectra.
The nonlinear response analysis software BISPEC (2010) has been used to develop inelastic
spectra for elasto-plastic hysteretic model. Average inelastic displacement spectra obtained
from the five spectrum-compatible ground motions have been used for obtaining the DMF.
Figure 4 shows the comparison of the DMFs obtained from the EL and YS approaches
and the analytical study. It can be observed that in short- and intermediate-period ranges
pffiffiffi
(T < T L ∕ μ), the DMFs obtained from the three approaches are close, and the EL approach
yields mostly conservative estimates for ductility ratio 4 and higher. Further, the conserva-
tism in the EL approach increases with the increase in ductility. However, for longer periods
pffiffiffi
(T > T L ∕ μ), the DMFs from the EL approach start decreasing and become constant beyond
T L . In this range of periods, the EL approach underestimates the DMFs compared to the YS
approach and the analytical results.

NUMERICAL EXAMPLES
The effect of the anomaly in the EL approach on the estimated peak inelastic displace-
ments has been illustrated using examples of oscillators having different initial periods (3, 4,
5, and 6 sec) and ductility (2, 4, 6, and 8). The same EL and YS Models considered in pre-
vious section are also considered for the numerical study. The analytical inelastic displace-
ment is obtained from the average response to the five spectrum-compatible ground motions,
as discussed earlier.
Table 1 shows the peak inelastic displacements of the oscillators, obtained using different
approaches. The analytically obtained displacement is shown in mm, whereas the displace-
ment obtained using other approaches is shown as percentage of that obtained analytically.
It can be observed that for ductility values of 2–6, the EL approach estimates inelastic
970 V. N. KHOSE AND Y. SINGH

Figure 4. Comparison of displacement modification factors obtained by the EL approach, the YS

approach, and analytical study for 0.4 g PGA, 5% elastic damping, and varying ductility, μ.

Table 1. Comparison of inelastic displacement estimated by different approaches. (The

analytically obtained displacement is shown in mm, whereas the displacement obtained using
other approaches is shown as percentage of the analytically obtained displacement.)

Inelastic displacement

μ¼2 μ¼4 μ¼6 μ¼8

EL EL EL EL

Ti AS* PCK† PSC‡ YS§ AS* PCK† PSC‡ YS§ AS* PCK† PSC‡ YS§ AS* PCK† PSC‡ YS§
(sec) (mm) (%) (%) (%) (mm) (%) (%) (%) (mm) (%) (%) (%) (mm) (%) (%) (%)
3 290 92 97 103 297 111 118 101 261 121 130 115 255 122 131 118
4 409 87 91 97 274 120 128 146 276 115 123 145 272 115 122 147
5 390 96 102 128 325 101 108 153 374 85 91 133 362 86 92 138
6 567 66 70 105 440 75 80 136 423 75 80 141 443 70 75 135

*Analytical study.
†
Priestley et al. (2007).
‡
Pennucci et al. (2011).
§
Chopra and Chintanapakdee (2004).
ANOMALY IN THE EL APPROACH FOR THE ESTIMATION OF INELASTIC RESPONSE 971

displacement either close to or more conservative than the analytical results (the Pennucci
et al. model yielding slightly higher estimates compared to the Priestley et al. model) for
periods up to 5 sec, whereas for periods equal to 6 sec, the inelastic displacements estimated
using the EL approach are significantly underestimated. In the case of ductility equal to 8, the
inelastic displacement from the EL approach is underestimated even for a 5-sec period. In the
case of the YS approach, the estimated inelastic displacements are either conservative or
close to the analytical results for the whole range of period and ductility, considered in
the study. It can be concluded that the EL approach underestimates inelastic displacement
in the whole of the displacement-controlled and in some part of the velocity-controlled spec-
tral ranges. Further, the period beyond which the EL approach underestimates inelastic dis-
placement reduces with an increase in ductility.

CONCLUSION
Equivalent linearization is one of the popular approaches to estimate peak inelastic
displacements using response spectrum. An anomaly in the EL approach has been illustrated,
which results in underestimated inelastic displacement in the whole of the displacement-
controlled and in some portion of the velocity-controlled spectral ranges. The anomaly
has been demonstrated by plotting the displacement modification factors and estimates of
inelastic displacement for oscillators of varying period and ductility using the EL and
YS approaches and an analytical study. Further, it has been noted that the period beyond
which the EL approach underestimates inelastic displacement decreases with an increase
in ductility.

REFERENCES
American Society of Civil Engineers (ASCE 7-10), 2010. Minimum Design Loads for Buildings
and Other Structures, ASCE 7-10, Reston, VA.
American Society of Civil Engineers (ASCE 41-06), 2007. Seismic Rehabilitation of Existing
Buildings, ASCE 41-06, Reston, VA.
Bispec Advanced 2.03.0000, Nonlinear Spectral Analysis Software, Earthquake Solutions.
Chopra, A. K., and Chintanapakdee, C., 2004. Inelastic deformation ratios for design and evalua-
tion of structures: Single-degree-of-freedom bilinear systems, Journal of Structural Engineer-
ing, ASCE 130, 1309–1319.
Cuesta, I., Aschheim, M. A., and Fajfar, P., 2003. Simplified R-factor relationships for strong
ground motions, Earthquake Spectra 19, 25–45.
Dwairi, H. M., Kowalsky, M. J., and Nau, J. M., 2007. Equivalent damping in support of direct
displacement-based design, Journal of Earthquake Engineering 11, 512–530.
European Committee for Standardization (CEN), 2004. Eurocode 8: Design of Structures for
Earthquake Resistance, Part 1: General Rules, Seismic Actions and Rules for Buildings,
EN1998-1, Brussels, Belgium.
European Committee for Standardization (CEN), 1998. Eurocode 8: Design of Structures for
Earthquake Resistance, Part 1: General Rules, Seismic Actions and Rules for Buildings,
prEN1998-1, European Committee for Standardization (CEN), Brussels, Belgium.
Federal Emergency Management Agency (FEMA), 2003. Recommended Provisons for Seismic
Regulations for New Buildings and Other Structures, Part 2: Commentary, FEMA 450 Part 2,
Washington, D.C.
972 V. N. KHOSE AND Y. SINGH

Grant, D. N., Blandon, C. A., and Priestley, M. J. N., 2005. Modelling Inelastic Response in
Direct Displacement-Based Design, Report 2005/03, IUSS Press, Pavia, Italy, 104 pp.
Gülkan, P., and Sozen, M. A., 1974. Inelastic responses of reinforced concrete structure to earth-
quake motions, ACI Journal Proceedings 71, 604–610.
Iwan, W. D., 1980. Estimating inelastic response spectra from elastic spectra, Earthquake Engi-
neering & Structural Dynamics 8, 375–388.
Kowalsky, M. J., 1994. Displacement-Based Design-a Methodology for Seismic Design Applied
to RC Bridge Columns, M.S. Thesis, University of California, San Diego.
Krawinkler, H., and Nassar, A. A., 1992. Seismic design based on ductility and cumulative
damage demands and capacities, in Nonlinear Seismic Analysis and Design of Reinforced
Concrete Buildings (P. Fajfar and H. Krawinkler, Eds.), Elsevier, New York, 316 pp.
Kwan, W.-P., and Billington, S. L., 2003. Influence of hysteretic behavior on equivalent period
and damping of structural systems, Journal of Structural Engineering, ASCE 129, 576–585.
Miranda, E., 1993. Site dependent strength reduction factors, Journal of Structural Engineering,
ASCE 119, 3503–3519.
Miranda, E., 2000. Inelastic displacement ratios for structures on firm sites, Journal of Structural
Engineering, ASCE 126, 1150–1159.
Mukherjee, S., and Gupta, V. K., 2002. Wavelet-based generation of spectrum-compatible time-
histories, Soil Dynamics and Earthquake Engineering 22, 799–804.
Newmark, N. M., and Hall, W. J., 1982. Earthqukae Spectra and Design, Earthquake Engineer-
ing Research Institute, Oakland, CA.
Ordaz, M., and Pérez-Rocha, L. E., 1998. Estimation of strength-reduction factors for elastoplas-
tic systems: A new approach, Earthquake Engineering & Structural Dynamics 27, 889–901.
Pennucci, D., Sullivan, T. J., and Calvi, G. M., 2011. Displacement reduction factors for the
design of medium and long period structures, Journal of Earthquake Engineering 15, 1–29.
PEER Ground Motion Database (PGMD), 2011. PEER Ground Motion Database (Beta Version
as on 10 October 2011), University of California, Berkeley, available at http://Peer.Berkeley
.Edu/Peer_Ground_Motion_Database.
Priestley, M. J. N., Calvi, G. M., and Kowalsky, M. J., 2007. Displacement-Based Seismic Design
of Structures, IUSS Press, Pavia, Italy.
Riddell, R., Garcia, J. E., and Garces, E., 2002. Inelastic deformation response of SDOF systems
subjected to earthquakes, Earthquake Engineering & Structural Dynamics 31, 515–538.
Riddell, R., Hidalgo, P., and Cruz, E., 1989. Response Modification Factors for Earthquake
Resistant Design of Short Period Buildings. Vol. 5, Earthquake Engineering Research Insti-
tute, Oakland, CA.
Ruiz-García, J., and Miranda, E., 2004. Inelastic Displacement Ratios for Design of Structures on
Soft Soils Sites. Vol. 130, American Society of Civil Engineers, Reston, VA.
Vidic, T., Fajfar, P., and Fischinger, M., 1994. consistent inelastic design spectra: Strength and
displacement, Earthquake Engineering & Structural Dynamics 23, 507–521.
(Received 11 March 2012; accepted 4 December 2012)