Sensitivity of structural responses to the processing of near-fault ground motion records containing fling-step

Abstract

Fling-step is one of the important characteristics of near-fault ground motions that is caused by a permanent static offset of the ground and appears as a one-sided velocity pulse and a large residual displacement at the end of the motion. Due to baseline errors in many ground motion records, they are processed (processing is a combination of baseline correction and filtering) before adding accelerograms to some earthquake databases whereby the residual displacements (i.e., the permanent static offsets) in the fling records are typically eliminated. It is not clear to what extent the record processing can affect the nonlinear responses of structures subjected to fling-step records. To address this issue, the current study is conducted to investigate the sensitivity of the nonlinear structural responses to record processing. To achieve this goal, the seismic responses of three code-conforming reinforced concrete (RC) special moment-resisting frames with different heights subjected to two versions of fling-step ground motions including a processed version that static offsets are removed, and a baseline-corrected version with static offsets preserved are evaluated. The findings of this assessment indicate that even though the processing procedure does not preserve the static offsets of the fling-step ground motions, the seismic responses of the structures to the processed records may resemble those obtained under only the baseline-corrected ground motions (with static offsets preserved).

Keywords

Near-fault records pulse-like ground motions fling-step record processing reinforced-concrete (RC) buildings nonlinear dynamic analysis

Introduction

Ground motions recorded near a ruptured fault can be significantly different from the usual earthquake ground motions recorded further away from the seismic source. In the proximity of an active fault system, ground motions may be characterized by a high amplitude and long-period distinctive pulse, which is concentrated in a narrow frequency band and is evident at the beginning of their velocity and displacement time histories (Bray and Rodriguez-Marek, 2004; Somerville et al., 1997). Near-fault ground motions with such pulses impose high input energy on structures at the onset of an earthquake and bring about severe structural damage (Bray and Rodriguez-Marek, 2004; Fang et al., 2018, 2021). Forward-directivity is one of the most prominent characteristics of near-fault ground motions that can be observed when the rupture and slip direction relative to a site coincide, and a significant portion of the fault rupture propagates towards the site with a velocity close to the shear-wave velocity (Somerville et al., 1997). Forward-directivity effects generate short-duration ground motions with large amplitude and long-period two-sided velocity pulses (Bray and Rodriguez-Marek, 2004). The other important feature of near-fault pulse-like ground motions is fling-step. Fling-step is an engineering term for the effects of the permanent tectonic offset, caused by a rupturing fault, in the near-fault region (Kamai et al., 2014). This phenomenon was mostly obscure until 1999, but the 1999 Chi-Chi and 1999 Kocaeli earthquakes and their detrimental effects on structures have brought this important phenomenon into focus. While forward-directivity is a dynamic phenomenon with two-sided velocity pulses and no permanent ground displacement, fling-step manifests as a one-sided velocity pulse and a large permanent displacement at the end of the shaking (Stewart et al., 2001), which occurs in the strike-parallel direction for strike-slip faults and the dip direction in the case of dip-slip faults (Bray and Rodriguez-Marek, 2004). However, residual displacements (i.e., permanent static offsets) caused by fling are typically removed from ground motion records via processing (i.e., a combination of baseline correction and filtering) before records are added to the library of some ground motion databases, such as the Pacific Earthquake Engineering Research (PEER) center. This procedure is performed to eliminate the low-frequency noise and errors due to baseline offsets (Burks and Baker, 2015; Kamai et al., 2014; Kamai and Abrahamson, 2015) resulting from tilting and transducer response to sharp pulses in ground motions (Burks and Baker, 2015). Filtering removes the static offsets from records by applying low- and high-cut filters in the frequency domain and baseline correction eliminates baseline offsets by fitting and then subtracting a function from the velocity pulse whereby a long-period signal is removed. Neglecting the static offset of fling may influence the nonlinear dynamic responses of structures subjected to fling-step ground motions. Therefore, it might be better to preserve static offsets of fling-step ground motions for engineering applications. Although the processing of raw accelerograms (i.e., unprocessed records) removes the static offset from fling-step records, some special forms of baseline correction can preserve the residual displacement at the end of the shaking (Boore and Akkar, 2001; Burks and Baker, 2015). Moreover, a baseline correction method has been recently proposed by Lin et al. (Lin et al., 2018) wherein an additional offset displacement is introduced to comprehend an expected target final displacement. It should be noted that the amplitude of the residual displacement is highly sensitive to the choice of baseline correction (Boore and Akkar, 2001).

It is of significance to note that the acceleration and velocity time histories of the processed and baseline-corrected versions of a fling-step record are greatly similar to each other, but there are significant differences in their displacement time histories. Hence, the PGA and PGV values are not significantly affected by processing, while the effects on the PGD are considerable (Kamai and Abrahamson, 2015). In addition, recent studies (Akkar and Boore, 2009; Boore and Akkar, 2003; Kamai et al., 2014; Wu and Wu, 2007) have shown that filtering and baseline correction typically have a negligible effect on the elastic response spectra of a fling-step ground motion record at periods shorter than the filter corner period. All the mentioned observations are evident in Figure 1, which shows the acceleration, velocity, and displacement time histories and the corresponding elastic response spectra of the processed (without the permanent static offsets of fling) and only baseline-corrected (with the permanent static offset of fling) versions of the 1999 Chi-Chi earthquake. Nevertheless, it is important to note that in contrast to the elastic response of single-degree-of-freedom (SDOF) systems (see Figure 1(b)), standard record processing may have more effect on the nonlinear responses of complex multi-degree-of-freedom (MDOF) structures when they respond beyond the elastic range (Akkar and Boore, 2009; Burks and Baker, 2015), which requires to be investigated.

Figure 1.

Processed (without permanent static offsets) and baseline-corrected (with permanent static offset) versions of the (a) time histories and (b) elastic response spectra in the North-South direction from the TCU052 station in the 1999 Chi-Chi earthquake.

Furthermore, recent studies on the effect of fling-step ground motions on structures revealed that there is no clear consensus on the type of fling-step records (i.e., processed or unprocessed) selected for seismic evaluation. That is, some studies (e.g., (Kalkan and Kunnath, 2006b)) utilized original fling records without any processing (with static offsets preserved), while some others (e.g., (Daei et al., 2022)) used processed records selected from the PEER database (without static offsets) for seismic assessment. This inconsistency creates a scientific challenge about the extent of the impact of standard record processing of fling-step ground motions on the nonlinear seismic behavior of structures.

Recently, some studies were conducted to investigate the consequence of the elimination of some fling-step effects in engineering applications. Some of these investigations (Kamai et al., 2014; Kamai and Abrahamson, 2015) focused on the parameterization of the fling-step effects and examining the influences of processing on the time histories and elastic response spectra (linear SDOF systems). However, to the authors’ best knowledge, a few analytical investigations (Akkar and Boore, 2009; Burks and Baker, 2015) have been undertaken to study the difference between the nonlinear structural responses obtained from the processed ground motions with fling removed and the baseline-corrected records with fling preserved. (Burks and Baker, 2015) compared the collapse capacity of two SDOF systems with different periods of 1.32 and 3 s using multiple versions of three ground motion recordings, including a version with permanent displacements preserved via baseline correction, a version obtained from the NGA database that static offsets are removed by processing and an artificial version with an idealized fling pulse superimposed to the processed records. It was demonstrated that for both of the systems considered and for all records, the baseline-corrected and processed versions of the ground motions give rise to similar collapse capacities. They also concluded that, in most of the cases, adding an artificial fling pulse to the processed record results in a conservative estimate of collapse capacity. As mentioned previously, this study focused on SDOF systems, while the effect of the removal fling permanent displacement due to the record processing on the nonlinear responses of MDOF systems is still questionable. In another study, (Ventura et al., 2011) evaluated the effects of fling removal due to processing on the nonlinear seismic behavior of SDOF systems and tall buildings. It was concluded that the elimination of fling-step effects from ground motion records has a great impact on the seismic responses of structures. Their results showed that the ground motions with fling retained can produce higher seismic demands in comparison to the records without fling-step. Although this investigation was on both SDOF and MDOF systems, the results in this study were obtained on the condition that the fling-step effects were completely removed from the records. To be precise, (Ventura et al., 2011) removed both the velocity pulse and residual displacement of fling for the ground motions designated as “without fling effects”, while in the current study, the velocity pulse of fling-step records is kept intact during the processing process, and only the permanent displacement is removed.

In general, the main thrust of this research is to investigate the sensitivity of the nonlinear seismic responses of the MDOF systems to record processing by comparing their nonlinear seismic responses subjected to two versions of fling-step ground motions, including processed versions with static offsets removed and baseline-corrected versions with static offsets preserved. To this end, three code-conforming RC special moment-resisting frame structures with different heights are evaluated under the two mentioned versions of fling-step ground motions at two earthquake intensity levels. Peak interstory drift ratios and story shears are two engineering demand parameters selected for this purpose.

Analytical modeling of the test buildings

Buildings’ configuration and design

Three residential RC buildings with special moment frame lateral load resisting systems are generated for this work. The case studies have 3-, 9-, and 18-story heights. A typical height of 3.2 m was equally selected for each story. Each building consists of three bays with a bay span of 5 m. The buildings are located in Van Nuys, Southern California (latitude/longitude = 34.218/-118.451), on a site with very dense soil or soft rock (site class C). the buildings are designed in compliance with the provisions of ASCE 7–10 (American Society of Civil Engineers (ASCE), 2010) and ACI 318–14 (American Concrete Institute (ACI), 2014). The floor dead and live loads are equal to 5.3 kN/m² and 1.96 kN/m², respectively. The interior wall load is equal to 1.1 kN/m² and is distributed throughout the floors. The loading width of the frames was assumed to be 5 m and the seismic mass at each floor level was assumed to comprise the dead load plus 20% of the live load. The design process is conducted using the ETABS software (Computers and Structures, 2017) and all the checks including the strong-column/weak-beam concept, the joint shear panel provision, and the drift criterion are carried out to ensure that the test buildings satisfy the provisions of ASCE 7–10 and ACI 318–14. The plan and elevation view of the case studies along with the beams and columns dimensions and the amounts of longitudinal reinforcements can be found in Appendix A.

Numerical models

Nonlinear time-history analyses are conducted within the Open System for Earthquake Engineering Simulation (OpenSees, 2016) platform. The fiber-based modeling strategy is utilized to develop numerical models. In this regard, each member cross-section is divided into three parts: the unconfined concrete section, the confined concrete section, and the reinforcement layer. All the structural elements are modeled using the “dispBeamColumn” element in the OpenSees platform. To improve the accuracy of the solution because of the use of displacement-based beam-column elements, each beam and column is divided into 10 and 9 sub-elements, respectively. Five integration points based on the Gauss-Legendre quadrature rule are used along each sub-element.

Steel material model

To define the stress-strain relationship of the reinforcing steel bars, the “Steel02” uniaxial steel material based on the Giuffré–Menegotto-Pinto (Menegotto and Pinto, 1973) model is adopted. A 1% strain hardening ratio is assumed for this material model. Figure 2(a) shows the stress-strain relationship of the Steel02 model. The yield strength of the longitudinal and transverse reinforcements is considered 400 MPa and 300 MPa, respectively.

Figure 2.

Stress-strain relationship used in the OpenSees for: (a) Steel02, (b) Concrete07 material model (Daei et al., 2022).

Concrete material model

In this study, the compressive strength of the unconfined concrete is considered 30 MPa. To simulate the unconfined and confined concrete in members, the “Concrete07” material model developed by (Waugh, 2009) is employed. This material model is based on (Chang and Mander, 1994) hysteretic concrete model with simplified unloading and reloading curves. Chang and Mander’s concrete model is an advanced rule-based model that takes advantage of different formulas to define the parameters of the monotonic envelope curve corresponding to the unconfined and confined concrete. This material model is successful in capturing important behavioral characteristics, namely strength deterioration, progressive stiffness degradation associated with smooth unloading and reloading curves at increasing strain values, pinching effect, and gradual crack closure effect and incorporates the effects of both complete and partial reloading to monotonic envelope curve (Chang and Mander, 1994; Waugh, 2009). The backbone curve of the Concrete07 material model is illustrated in Figure 2(b). Several parameters are incorporated in developing this envelope curve including the initial elastic modulus ( $E_{c}$ ), peak strain and stress at compression ( $ε_{c}^{'}$ , $f_{c}^{'}$ ), critical strain at compression ( $ε_{c r}^{-}$ ) where the compression part of the envelope curve turns to a straight line and continues until the spalling strain ( $ε_{s p}$ ) with zero compressive stress is reached, peak strain and stress at tension ( $ε_{t}$ , $f_{t}$ ), critical strain at tension ( $ε_{c r}^{+}$ ) where a straight line is followed by the tension envelope curve until the concrete reaches the cracking strain ( $ε_{c r k}$ ), shape parameter (r) that controls the descending branch of the envelope curve, and $x_{p}$ and $x_{n}$ that are non-dimensional terms defining the strain at which the straight line starts descending in tension and compression, respectively.

As an example, the values of the above-mentioned modeling parameters of the Concrete07 model for the unconfined and confined concrete of typical beam and column sections shown in Figure 3 are listed in Table 1.

Figure 3.

Typical column and beam sections.

Table 1.

Input parameters of the Concrete07 material model for typical beam and column sections.

Section	$E_{c}$ (MPa)	$f_{c}^{'}$ (MPa)	$ε_{c}^{'}$ (MPa)	$f_{t}$ (MPa)	$ε_{t}$	$x_{p}$	$x_{n}$	r
Unconfined	29359	30	0.00204	3.39	0.00023	2	2.3	3.87
Confined (column)	29359	39.94	0.00541	3.39	0.00023	2	30	1.34
Confined (beam)	29359	34.84	0.00368	3.39	0.00023	2	30	1.48

Note: Based on the Concerete07 formulation, the monotonic envelope for the tension side of the confined concrete follows the same curve that is used for unconfined concrete.

Verification of numerical modeling

The finite element modeling procedure used in the current study is verified with an experimental test conducted by (Lehman and Moehle, 2000) on an axially loaded RC column specimen with a circular cross-section under reversed cyclic loading. The result of this validation is displayed in Figure 4. From this figure, an excellent agreement can be seen between the force-displacement hysteretic response of the column obtained from the experimental test and numerical analysis. For more details regarding this validation, the reader is referred to Ref. (Daei et al., 2022).

Figure 4.

Verification of the modeling procedure used in the current study with the experimental test conducted by (Lehman and Moehle, 2000).

Dynamic analysis assumptions and modal analysis

In the dynamic analyses, the Newmark scheme is employed for numerical integration. The damping matrix is simulated using the Rayleigh damping formulation. The number of modes selected for defining the Rayleigh damping is selected such that the sum of their effective modal participating mass ratios (α_n) is at least 90%. The sum of α_n for the first two modes of the 3-story frame, the first three modes of the 9-story frame, and the first four modes of the 18-story frame is 0.9520, 0.9190, and 0.9187, respectively. So, the damping is defined assuming a critical damping ratio of 5% in the first two modes of vibration in the case of the 3-story building, the first and third modes in the 9-story building, and the first and fourth modes in the 18-story building. In all the analyses, P-delta (i.e., the second-order) effects are taken into account using the corotational geometric transformation command available in OpenSees. The most important merit of the corotational theory is the decoupling of the formulation of the structural element from the geometric transformation of its response quantities (Uriz et al., 2008). Moreover, it is assumed that the columns are fixed to the base and the diaphragms are rigid.

The first three natural vibration periods of the frames based on the uncracked and cracked concrete properties obtained by OpenSees and ETABS, respectively, are provided in Table 2. For more detailed information about the numerical models, refer to (Daei et al., 2022; Daei and Poursha, 2021).

Table 2.

Natural periods of the first three vibration modes.

Frame	Vibration periods (s)
	Uncracked section			Cracked section
	$T_{1}$	$T_{2}$	$T_{3}$	$T_{1}$	$T_{2}$	$T_{3}$
3-Story	0.472	0.148	0.082	0.617	0.180	0.091
9-Story	1.160	0.416	0.240	1.626	0.559	0.312
18-Story	1.608	0.590	0.347	2.265	0.828	0.483

Selected ground motions

To perform the nonlinear dynamic analyses, a total of 12 near-fault ground motion records containing the fling-step effect are selected. The selected ground motions were recorded from the Kocaeli, Turkey (M_w = 7.4) and Chi-Chi, Taiwan (M_w = 7.6) earthquakes in 1999 within 15 km of the rupturing fault. They were recorded on very dense soil and soft rock (class C) or stiff soil (class D) sites. Table 3 provides the main properties of the selected ground motions. In this table, pertinent information on the selected ground motions including the magnitude (

M_{w}

), the closest distance of the station from the rupturing fault (

R_{r u p}

), peak ground acceleration (PGA), peak ground velocity (PGV), peak ground displacement (PGD), and pulse period (

T_{p}

) of the records are provided. To achieve the objective of this study, two types of the selected records are compiled: the first type consists of processed records wherein the residual displacement caused by fling is removed, and the second type contains only baseline-corrected (without filtering) records in which the residual displacement at the end of the displacement time-histories of the records is preserved. The processed records (the first type) are adopted using the PEER NGA-West2 database tool (https://ngawest2.berkeley.edu), while the baseline-corrected records (the second type records) are obtained from the fling-step ground motion database compiled by (Kalkan, 2006). The permanent displacement of the baseline-corrected fling-step records were determined based on the GPS measurements for the Chi-Chi earthquake ground motions (Kalkan, 2006; Kalkan and Kunnath, 2006b). The information for this purpose is derived from Ref. (Boore and Akkar, 2001). In the case of the Kocaeli earthquake, baseline corrections were only conducted considering the requirement of the zero-velocity crossing at the end of the time history. More information about the baseline-corrected ground motions can be found in Refs. (Kalkan, 2006; Kalkan and Kunnath, 2006b).

Table 3.

The selected fling-step ground motion records.

No.	Earthquake	Year	$M_{w}$	Station	Component	Site	$R_{r u p}$ (km)	PGA (g)	PGV (cm/s)	PGD (cm)	$T_{p} (s)$
1	Kocaeli	1999	7.4	Yarimca (YPT)	60	D	4.8	0.23	69.71	62.32	4.9
2	Chi-chi	1999	7.6	TCU049	EW	C	3.8	0.28	53.54	74.26	10.2
3	Chi-chi	1999	7.6	TCU052	NS	C	0.7	0.44	172.33	226.58	11.9
4	Chi-chi	1999	7.6	TCU065	EW	D	0.6	0.79	125.35	108.73	5.7
5	Chi-chi	1999	7.6	TCU067	EW	C	0.6	0.50	92.00	101.38	2.3
6	Chi-chi	1999	7.6	TCU068	EW	C	0.3	0.51	249.60	297.15	12.3
7	Chi-chi	1999	7.6	TCU071	NS	C	5.8	0.65	67.57	66.90	5.4
8	Chi-chi	1999	7.6	TCU074	EW	C	13.5	0.59	70.37	21.31	1.5
9	Chi-chi	1999	7.6	TCU076	EW	C	2.7	0.34	51.84	33.27	4.7
10	Chi-chi	1999	7.6	TCU082	EW	C	5.2	0.22	54.92	95.08	8.1
11	Chi-chi	1999	7.6	TCU128	EW	C	13.1	0.14	63.75	81.84	9.0
12	Chi-chi	1999	7.6	TCU129	EW	C	1.8	1.00	62.81	60.89	7.5

Figure 5 shows the acceleration, velocity, and displacement time histories of both the processed and baseline-corrected versions of the selected ground motion records. The elastic acceleration, velocity, and displacement spectra of the selected records and the mean spectra of the processed and baseline-corrected versions are also shown in Figure 6.

Figure 5.

Time histories of the processed and baseline-corrected versions of the selected ground motion records.

Figure 6.

Elastic pseudo-acceleration, pseudo-velocity, and displacement response spectra of the selected ground motion records along with the mean response spectra of the processed and baseline-corrected record sets.

All the selected ground motions are scaled such that their spectral acceleration at the fundamental period of the structure, S_a(T₁), equals that of the design basis earthquake (DBE) target spectrum of ASCE 7–10. This scaling approach is better than the PGA method, which does not take into account the dynamic behavior of the structure and is not efficient for soft soil and near-fault sites (Shakeri et al., 2018). In order to evaluate the results in a higher intensity level (i.e., the risk-targeted maximum considered earthquake ( ${M C E}_{R}$ ) level), the records are scaled again to 1.5S_a(T₁). These two intensity levels are the usual earthquake levels used for the design and evaluation of buildings based on seismic codes.

Preliminary overview of the structural behavior

Pushover analysis as a widely used analysis procedure is able to provide valuable insight into the nonlinear behavior of a structure like the lateral loading capacity of the structure subjected to seismic loads and give the damage pattern of the structure in the nonlinear range. Also, it can predict the seismic deformation demands (Fajfar, 2000) and force demands (Amini and Poursha, 2018; Goel and Chopra, 2005) of the structure by using the design spectrum of a code. In view of the simplicity and less computationally nature of the pushover procedure, it has gained growing popularity among practitioners and scholars in recent years as a tool for seismic design and evaluation of structures. Extensive research efforts (Antoniou and Pinho, 2004; Chopra and Goel, 2002; Daei and Zarrin, 2021; Kalkan and Kunnath, 2006a; Kreslin and Fajfar, 2011; Poursha et al., 2009; Zarrin et al., 2021) have been devoted to the development of advanced multi-modal pushover methods to overcome the shortcomings of the primary versions of this procedure. To gain a preliminary sense of the inelastic response characteristics of the case study frames, the pushover analysis is conducted on the frames prior to presenting the results of the nonlinear time history analyses. For this purpose, the frames are pushed by lateral forces proportional to the first vibration mode. The global pushover curves (base shear versus roof drift angle) of the frames obtained by the analysis are displayed in Figure 7. The dashed vertical lines in this plot indicate the target roof displacements of the buildings at the DBE and MCE levels, which are the mean value of the roof displacements resulting from the nonlinear dynamic analysis subjected to the selected processed records. According to the pushover curves, several key points can be identified in the response of the structures. The figure illustrates that the 3-story frame deforms well beyond the elastic limit at both hazard levels, while this holds true for the 9-story and 18-story frames at the MCE level, and these structures step only slightly into the inelastic range at the DBE level. All the frames exhibit a small strength plateau after the yielding point, followed by a strength loss and consequently a negative post-yield stiffness. This negative post-yield stiffness, which has serious ramifications on the dynamic response of a structure (Gupta and Krawinkler, 1999), is because of mainly the P-delta effect, as well as the strength deterioration characteristic considered by the concrete material model used in this study (refer to the section of concrete material model). It is also worth mentioning that a trivial decrease in the lateral load strength of the structures (<5% for the 3-story frame and <1% for the 9- and 18-story frames) can be observed from the pushover curves at the DBE level. This strength deterioration increases as the frames experience a high level of nonlinearity (i.e., the MCE level) – that is, the strength deterioration for the 3- and 9-story frames is about 13% and for the 18-story frame is roughly 20% at the MCE level.

Figure 7.

Global pushover curves of the case study frames.

Results of the seismic evaluation

In this section, the sensitivity of the nonlinear structural responses to record processing is scrutinized by comparing the deformation and force demands of the three test buildings subjected to the two versions of fling-step ground motions described in the section of selected ground motions. To this end, the height-wise distribution of the floor displacements, interstory drift ratios, and story shears of the buildings are examined, but because the conclusions for floor displacements and interstory drift ratios are similar, the floor displacements are not presented herein for conciseness. The evaluation is carried out at two earthquake intensity levels including the DBE and MCE levels according to the ASCE 7 standard. The analysis is set to terminate when the simulation encounters convergence problems or if the maximum story drift value of 10% is reached.

Figure 8 to 10 show the interstory drift ratios and story shears of the buildings resulting from the processed and baseline-corrected versions of the selected ground motions having fling-step pulses (see Figure 5). The results for the two aforementioned intensity levels (the DBE and MCE levels) are presented in separate plots. The figures illustrate that the structural responses under the processed version (in which static offsets are removed) are, to a great extent, similar to those of the baseline-corrected version (with static offsets preserved). This demonstrates that even though the standard processing does not preserve the static offsets (i.e., permanent displacement in displacement time histories) in fling-step ground motions, it may result in dynamic effects that are similar to those produced by the baseline-corrected version with residual displacement. These dynamic effects of fling-step containing high seismic energy lie in the high-amplitude and long-duration velocity pulse.

Figure 8.

The seismic responses of the 3-story frame subjected to the processed and baseline-corrected versions of the fling-step ground motions (a) DBE level (b) MCE level.

Figure 9.

The seismic responses of the 9-story frame subjected to the processed and baseline-corrected versions of the fling-step ground motions.

Figure 10.

The seismic responses of the 18-story frame subjected to the processed and baseline-corrected versions of the fling-step ground motions (a) DBE level (b) MCE level.

Despite the similarities between the results obtained from the two versions of fling records, there are negligible differences between them. Interestingly, the processed records induce even slightly higher seismic demands in comparison with the baseline-corrected counterparts in most cases. These insignificant differences are more evident in the interstory drift demands, while the processing has a more negligible effect on the story shear demands.

For better understanding, the maximum differences between the results of the processed records and those of the baseline-corrected records are shown in Figure 11. The maximum differences are defined as the absolute maximum of the ratio of the structural responses obtained from the processed records to those from the baseline-corrected ground motions. As can be seen in the figure, for the DBE intensity level, the maximum differences in the interstory drift ratios between the processed and the baseline-corrected versions are 6.84, 7.20, and 2.96% for the 3-, 9-, and 18-story buildings, respectively. These values decrease to 0.37, 0.95, and 1.73% for the story shear demands in the case of the 3-, 9-, and 18-story buildings, respectively. For the MCE level, the maximum differences between the interstory drift demands obtained from the two types of fling-step records are 4.75, 3.40, and 7.22% for the 3-, 9-, and 18-story buildings, respectively. Finally, in the case of the story shear demands at the MCE intensity level, the maximum differences reach 0.12, 1.67, and 3.76% for the 3-story, 9-story, and 18-story buildings, respectively. The results reveal that the seismic responses of the structures are not sensitive to record processing and either type of records can be used for seismic assessment.

Figure 11.

The maximum ratios of the structural responses obtained from the processed records to those from the baseline-corrected records.

The ratios of the fundamental period of the case study frames to the pulse period of the selected fling-step records (T/T_p) is provided in Table 4. This table indicates that almost all the T/T_p ratios (except one record in the 18-story frame) are much lower than unity, implying that the pulse period of the selected records are greater than the first-mode period of the structures. Therefore, all the conclusions in this study are limited to the structures studied in this paper and it is true as long as the fundamental period of the structure is shorter than the pulse period of the ground motions. Future studies are required to investigate the effect of record processing on tall buildings with fundamental periods longer than the pulse period, other structural typologies (e.g., steel and masonry buildings), and also other structures like bridges and dames.

Table 4.

Ratios of the fundamental period of the structures to the pulse period of the selected records (T/T_p).

Frames	T/T _p
Frames	Yarimca	TCU049	TCU052	TCU065	TCU067	TCU068	TCU071	TCU074	TCU076	TCU082	TCU128	TCU129
3-Story	0.096	0.046	0.040	0.082	0.205	0.038	0.087	0.315	0.100	0.058	0.052	0.063
9-Story	0.237	0.114	0.097	0.203	0.504	0.094	0.215	0.773	0.247	0.143	0.129	0.155
18-Story	0.328	0.158	0.135	0.282	0.699	0.131	0.298	1.072	0.342	0.198	0.179	0.214

Conclusions

Some ground motion databases remove the static offsets from the end of fling-step ground motions via standard record processing to eliminate baseline errors before making them available for engineering applications. This study was conducted to investigate the extent of the impact of permanent displacement removal from fling-step records on the nonlinear responses of RC building structures. To this end, three RC special moment-resisting building frames with different heights were subjected to two types of fling-step ground motions, including a type where the static offset is removed via processing and another type with static offset preserved. Interstory drifts and story shears are two seismic responses selected for this comparative study. The evaluation was carried out at two earthquake intensity levels including the DBE and MCE levels. The results of this investigation lead to the following findings:

• To a great extent, similar nonlinear structural responses were observed under the two mentioned versions of fling-step ground motions (with and without permanent displacements), implying that even though the processing does not retain the residual displacement of fling ground motions, it may result in dynamic effects that are similar to those produced by the records with residual displacement preserved.

• The maximum differences in structural results between the processed and baseline-corrected records are greater for deformation demands than the force demands. That is, for all the test buildings, the maximum difference between the results of the processed and baseline-corrected records is just above 7% for the interstory drift ratios, while this value is lower than 4% for the story shear demands.

• While no specific trend can be observed in differences between the results of the processed and baseline-corrected versions in the case of the drift demands, a consistent trend is observed for story shears where the differences raise by increasing the structural height. To be precise, the greatest differences between the fling records with and without static offsets in the case of interstory drift ratios belong to the 9-story and 18-story frames at the DBE and MCE levels, respectively, while the 18-story frame has the largest discrepancy between the processed and baseline-corrected versions for the story shears at both intensity levels.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Data availability

If necessary, it can be requested from the corresponding author

Ethics approval

The authors state ethics approval.

Consent of participate

On behalf of all authors, the corresponding author states the authors consent to participate.

Consent of publication

On behalf of all authors, the corresponding author states the authors consent for publication.

ORCID iDs

Mehdi Poursha

Mohamad Zarrin

Plan and elevation view of the case studies

The plan and elevation view of the case studies along with the beams and columns dimensions and the amounts of longitudinal reinforcements are presented below.

Figure A1.

Plan view of the archetype buildings (Daei et al., 2022).

Figure A2.

The elevation view of the case study buildings (Daei et al., 2022).

References

American Concrete Institute (ACI) . (2014) Building Code Requirements for Structural Concrete and Commentary (ACI 318–14). Farmington, MI.

Akkar

Boore

(2009) On baseline corrections and uncertainty in response spectrafor baseline variations commonly encounteredin digital accelerograph records. Bulletin of the Seismological Society of America. 99(3). 1671–1690. DOI: 10.1785/0120080206.

American Society of Civil Engineers (ASCE) . (2010) Minimum Design Loads for Buildings and Other Structures. Reston, VA: ASCE. pp. 7–10.

Amini

Poursha

(2018) Prediction of the force demands of tall buildings through the enhanced pushover procedures. Structural Design of Tall and Special Buildings 27(17): 1–12. DOI: 10.1002/tal.1540.

Antoniou

Pinho

(2004) Development and verification of a displacement-based adaptive pushover procedure. Journal of Earthquake Engineering 8(5): 643–661. DOI: 10.1080/13632460409350504.

Boore

Akkar

(2001) Effect of baseline corrections on displacements and response spectra for several recordings of the 1999 Chi-Chi, Taiwan, earthquake. Seismological Society of America, 91(5): 1199–1211. DOI: 10.1002/eqe.299.

Boore

Akkar

(2003) Effect of causal and acausal filters on elastic and inelastic response spectra. Earthquake Engineering and Structural Dynamics 32(11): 1729–1748. DOI: 10.1002/eqe.299.

Bray

Rodriguez-Marek

(2004) Characterization of forward-directivity ground motions in the near-fault region. Soil Dynamics and Earthquake Engineering 24(11): 815–828. DOI: 10.1016/j.soildyn.2004.05.001.

Burks

Baker

(2015) Ground Motion Simulations: Validation and Application for Civil Engineering Problems. Report 191, the John A. Blume Earthquake Engineering Center. Stanford CA: Stanford University.

10.

Chang

Mander

(1994) Seismic energy based fatigue damage analysis of bridge columns: Part 1 - evaluation of seismic capacity. NCEER Technical Report No. NCEER-94-0006. Technical Report NCEER-94-0006.

11.

Chopra

Goel

(2002) A modal pushover analysis procedure for estimating seismic demands for buildings. Earthquake Engineering and Structural Dynamics 31(3): 561–582. DOI: 10.1002/eqe.144.

12.

Computers Structures (2017) Structural Software for Building Analysis and Design. ETABS. Available at: https://www.csiamerica.com/products/etabs (accessed 12 February 2018).

13.

Daei

Poursha

(2021) On the accuracy of enhanced pushover procedures for seismic performance evaluation of code-conforming RC moment-resisting frame buildings subjected to pulse-like and non-pulse-like excitations. Structures 32(2020): 929–945. DOI: 10.1016/j.istruc.2021.03.035.

14.

Daei

Zarrin

(2021) A multi-mode displacement-based pushover (MDP) procedure for seismic assessment of buildings. Soil Dynamics and Earthquake Engineering 150(2020): 106950. DOI: 10.1016/j.soildyn.2021.106950.

15.

Daei

Poursha

Zarrin

(2022) Seismic performance evaluation of code-compliant rc moment-resisting frame buildings subjected to near-fault pulse-like and non-pulse-like ground motions. Journal of Earthquake Engineering 26(10): 5058–5085.

16.

Fajfar

(2000) A nonlinear analysis method for performance based seismic design. Earthquake Spectra 16(3): 573–592.

17.

Fang

Zhong

Wang

, et al. (2018) Peak and residual responses of steel moment-resisting and braced frames under pulse-like near-fault earthquakes. Engineering Structures 177(September): 579–597. DOI: 10.1016/j.engstruct.2018.10.013.

18.

Fang

Liang

Zheng

, et al. (2021) Seismic performance of bridges with novel SMA cable-restrained high damping rubber bearings against near-fault ground motions. Earthquake Engineering and Structural Dynamics 51(1): 44–65. DOI: 10.1002/eqe.3555.

19.

Goel

Chopra

(2005) Extension of modal pushover analysis to compute member forces. Earthquake Spectra 21(1): 125–139. DOI: 10.1193/1.1851545.

20.

Gupta

Krawinkler

(1999) Seismic demands for the performance evaluation of steel moment resisting frame structures. In: Report No. 132. The John A. Blume Earthquake Engineering Center. Stanford CA: Stanford University.

21.

Kalkan

(2006) Prediction of seismic demands in building structures. In: PhD Thesis. University of California Davis.

22.

Kalkan

Kunnath

(2006a) Adaptive modal combination procedure for nonlinear static analysis of building structures. Journal of Structural Engineering 132132(11): 1721–1731. DOI: 10.1061/(ASCE)0733-9445.

23.

Kalkan

Kunnath

(2006b) Effects of fling step and forward directivity on seismic response of buildings. Earthquake Spectra 22(2): 367–390. DOI: 10.1193/1.2192560.

24.

Kamai

Abrahamson

(2015) Are near-fault fling effects captured in the new NGA west2 ground motion models? Earthquake Spectra 31(3): 1629–1645. DOI: 10.1193/101713EQS270M.

25.

Kamai

Abrahamson

Graves

(2014) Adding fling effects to processed ground-motion time histories. Bulletin of the Seismological Society of America 104(4): 1914–1929. DOI: 10.1785/0120130272.

26.

Kreslin

Fajfar

(2011) The extended N2 method taking into account higher mode effects in elevation. Earthquake Engineering and Structural Dynamics 40(14): 1571–1589. DOI: 10.1002/eqe.1104.

27.

Lehman

Moehle

(2000) Seismic Performance of Well-Confined Concrete Bridge Columns. Pacific Earthquake Engineering Research Center, p. 295.

28.

Lin

Zong

Tian

, et al. (2018) A new baseline correction method for near-fault strong-motion records based on the target final displacement. Soil Dynamics and Earthquake Engineering 114(October): 27–37. DOI: 10.1016/j.soildyn.2018.06.036.

29.

Menegotto

Pinto

(1973) Method of analysis for cyclically loaded R. C. Plane frames including changes in geometry and non-elastic behavior of elements under combined normal force and bending. Proceedings of IABSE Symposium on Resistance and Ultimate Deformability of Structures Acted on by Well Defined Loads: 15–22. DOI: 10.5169/seals-13741.

30.

OpenSees (2016) Open System for Earthquake Engineering Simulation. Available at: http://opensees.berkeley.edu/ (accessed 10 February 2018).

31.

Poursha

Khoshnoudian

Moghadam

(2009) A consecutive modal pushover procedure for estimating the seismic demands of tall buildings. Engineering Structures 31(2): 591–599. DOI: 10.1016/j.engstruct.2008.10.009.

(2018) Ground motion scaling for seismic response analysis by considering inelastic response and contribution of the higher modes. Soil Dynamics and Earthquake Engineering 110(March): 70–85. DOI: 10.1016/j.soildyn.2018.04.007.

, et al. (1997) Modification of empirical strong ground motion attenuation relations to include the amplitude and duration effects of rupture directivity. Seismological Research Letters 68(1): 199–222. DOI: 10.1785/gssrl.68.1.199.

34.

Stewart