Seismic isolation of buildings for power stations considering soil-structure interaction effects

Author’s Accepted Manuscript Seismic isolation of buildings for power stations considering soil-structure interaction EFFECTS Arturo Tena-Colunga, Luis Eduardo Pérez-Rocha, Javier Avilés, Cuauhtémoc Cordero-Macías www.elsevier.com/locate/jobe

PII: DOI: Reference:

S2352-7102(15)30015-2 http://dx.doi.org/10.1016/j.jobe.2015.08.001 JOBE39

To appear in: Journal of Building Engineering Received date: 20 January 2015 Revised date: 29 June 2015 Accepted date: 6 August 2015 Cite this article as: Arturo Tena-Colunga, Luis Eduardo Pérez-Rocha, Javier Avilés and Cuauhtémoc Cordero-Macías, Seismic isolation of buildings for power stations considering soil-structure interaction EFFECTS, Journal of Building Engineering, http://dx.doi.org/10.1016/j.jobe.2015.08.001 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

SEISMIC ISOLATION OF BUILDINGS FOR POWER STATIONS CONSIDERING SOIL-STRUCTURE INTERACTION EFFECTS Arturo Tena-Colunga1, Luis Eduardo Pérez-Rocha2, Javier Avilés 3 and Cuauhtémoc Cordero-Macías 4 1

Departamento de Materiales, Universidad Autónoma Metropolitana Azcapotzalco Edificio 4P, Av. San Pablo 180, 02200 México, DF, MEXICO [email protected] 2

Instituto de Investigaciones Eléctricas Calle Reforma 113, Col. Palmira, 62490 Cuernavaca, MEXICO [email protected] 3

Instituto Mexicano de Tecnología del Agua Paseo Cuauhnahuac 8532, 62550 Jiutepec, MEXICO [email protected] 4

Comisión Federal de Electricidad Río Mississippi 71, 06500 México, D.F., MEXICO e-mail: [email protected] ABSTRACT Nowadays, the power industry in Mexico is building encapsulated power stations, because they are a more efficient, safer and cheaper technology for power transformation, and also requires of smaller spaces. Encapsulated power stations have been already built in cities of Mexico where the seismic risk and hazard is negligible. Given all the advantages described above, Comisión Federal de Electricidad (CFE), the Mexican power company, is studying the possibility to build this type of stations in the Valley of Mexico, where the earthquake risk and hazard are high. Encapsulated power stations are composed of a series of high diameter pressured steel pipes that contains a toxic gas as an insulation media for the electric cables that host inside. Therefore, it is required that for severe earthquake shaking, encapsulated power station remain fully operational and that there will be no risk of a gas leaks, particularly in the soft soil sites of the Valley of Mexico. The use of base isolation in soft soils as those found in the lakebed region of Mexico City is atypical and must be seriously evaluated. In soft soils, there is a higher risk of dynamic instability of base isolated structures for severe earthquake shaking due to the likeliness of resonant response with the ground, global rocking effects or sudden differential settlements of the ground. Research studies devoted to assess the feasibility of employing pendular isolation systems for the benchmark architectural model of SF6 power stations for six soft soil sites in the Valley of Mexico are

summarized in this paper. Soil-structure interaction (SSI) effects were included. The results of the study allowed concluding that pendular isolation systems can be used for an effective seismic isolation of SF6 power stations in relatively soft soils.

Keywords: encapsulated power stations, power industry, seismic isolation, base isolation, pendular isolators, friction pendulum, soft soils, site effects, soil-structure interaction.

INTRODUCTION The Mexican power industry is in a modernization process of its infrastructure. For this reason, new encapsulated power plants are being built, as well as the replacement of obsolete old power plants for new encapsulated ones. A more modern, secure an efficient technology for power transformation is provided in encapsulated power plants; besides, a smaller land space is needed to build them. Encapsulated power plants are composed of high-diameter pressured steel pipes that hold a toxic gas as an insulation media for the power cables hosted inside (Figure 1). Some encapsulated power stations have already being built in Mexico in sites where the earthquake risk and hazard is low or negligible (Figure 1). Given all the advantages such facilities offers, the Mexican Power Industry is considering to replace obsolete power stations for encapsulated power stations in Mexico City and its Metropolitan Area, where the earthquake risk and hazard are high. Besides, important site effects are triggered in soft soils for the old lakebed region. In order to build encapsulated power stations in such sites, it is required that: a) the building will remain operational and, b) there would not be a serious risk of toxic gas leaking. Such design goals pose an important challenge for its structural design, particularly in the soft soil sites of the Valley of Mexico. Base isolation could be a good structural solution to insure that encapsulated pipes (Figure 1) would not experience relative movements of importance during severe earthquake shaking. However, the use of base isolation in soft soils as those found in the lakebed region of Mexico City is atypical. As a matter of fact, in previous research studies for the region,

1-2

it

was found that the use of conventional elastomeric bearings (lead-rubber bearings, LRB, and high damping rubber bearings, HRB) is not effective for ground acceleration records of very soft soils sites in Mexico City with site periods of 1.25 seconds or greater (Ts≥ 1.25s), due to its potential resonant response with the ground. In fact, the dynamic instability of LRB isolators was obtained when subjected to representative accelerograms recorded in the

lakebed zone during the 1985 Michoacán Earthquake, even when soil-structure interaction effects were not included into the modeling 2. The use of base isolation in soft soils must be seriously assessed because important soil settlements (long-term and sudden) and global uncontrolled rocking effects due to soil-structure interaction could also be developed. On the other hand, a conventional design in soft soil sites for buildings that host encapsulated power stations does not warrant, after a strong earthquake that: a) such installations would remain operational and, b) the pipes would remain undamaged with no risk of gas leaking. Therefore, civil engineers from the Mexican power industry considered that the use of base isolation should be evaluated to assess if it can provide a higher level of protection for encapsulated power plant buildings than a traditional structural design, for six different sites of their interest within the Metropolitan Area of Mexico City, located within the lakebed region. The most relevant aspects of the research study devoted to assess the feasibility of employing pendular isolation systems for the benchmark architectural model of SF6 power stations for six different soil sites in the Valley of Mexico are summarized in following sections. Soilstructure interaction (SSI) effects were evaluated in the dynamic response of the studied baseisolated projects.

PREVIOUS RESEARCH ON SOIL STRUCTURE INTERACTION EFFECTS ON BASE-ISOLATED STRUCTURES Few studies are available on the impact of soil-structure interaction (SSI) effects in baseisolated structures 1,3-11 and most of them are devoted to the seismic isolation of bridges 4-9. Among the pioneering works is the parametric study presented by Constantinou and Kneifati 3

, where a simplified one-story superstructure mounted on a generalized elastomeric bearing

base-isolation system supported in a rigid foundation which rests at the surface of a homogeneous and viscoelastic half-space was studied. The excitation consisted of vertically propagating S-waves, specified by the free-field acceleration of the ground surface. From their parametric study they concluded that soil-structure interaction in base-isolated structures does not have as important effects as it does in conventional structures. The importance of interaction decreases with increasing values of the ratio o/b, this is, the ratio between the fundamental frequency of the fixed-base superstructure (o) and the effective frequency of the base-isolated structure supported on a rigid foundation (b). In other words, the importance of interaction decreases as the base isolation system is effective in uncoupling the

effective base-isolated period for the structure (TI) from the fixed-base period for the superstructure (Tfb), this is, a large TI/Tfb ratio. González and Noguez

1

presented an analytical study for the hypothetical project of an

existing 9-story reinforced concrete building founded in soft soils of Mexico City, with a foundation composed by a rigid slab supported in point-bearing piles. Base-isolation projects with LRB were designed and studied for three records, among them the following two that are well-known worldwide: a) 1940 El Centro NS (firm soils in California) and, b) 1985 SCT-EW record (soft soils in Mexico City). Soil-structure interaction (SSI) effects were also considered in the modeling, assuming the properties of a soft site within Mexico City. It was observed from the results obtained for soft soils under the SCT EW record that: a) absolute acceleration time histories (i.e., roof) for the base-isolated were amplified with respect to the non-isolated project; when including SSI into the modeling, peak accelerations were even higher, b) nonlinear demands on the framed building (story ductility demands and beam yielding) were higher for the base-isolated project than for the fixed-based project. Story ductility demands and beam yielding were more pronounced when including SSI effects. Therefore, it was found that inelastic demands on the frame elements were much higher for the base-isolated project than for a traditional non-isolated project, considering or neglecting SSI effects. The study concluded that: a) base-isolation is efficient for ground motion records typical of firm soils (i.e., 1940 El Centro NS), even when SSI are included and, b) baseisolation is not efficient for ground motion records typical of soft soils (i.e., 1985 SCT EW), where resonant responses are located in a medium-long period range, including or ignoring SSI effects. Thakkar and Maheshwari

4

studied a base isolated bridge to determine soil-structure

interaction effects on its seismic response when elastomeric bearings are provided between superstructure and the substructure. Modal analyses were conducted using the following response spectra: a) Indian code, b) an alluvium site and, c) a rock site. From their limited study they concluded that the effectiveness of elastomeric bearings clearly decrease in alluvium soils when compared to hard soils or rock sites. Chaudhary et al. 5 conducted a very comprehensive study where they identified soil-structure interaction effects on four base-isolated bridges with rubber bearings in Japan from using acceleration data recorded during 18 earthquakes. Two bridges (Matsunohama viaduct bridges) were founded in alluvial/dilluvial soils with average site periods Ts of 0.48s (bridge

A) and 0.57s (bridge B) respectively, whereas the other two bridges were founded on firmer soils, with site periods Ts=0.29s (Onneto bridge) and Ts=0.19s (Yammage bridge). Soilstructure interaction effects were examined by comparing the identified and physical stiffness of the substructure components. They found a strong SSI effect in Onneto bridge and Matsunohama viaduct bridge B, as the identified substructure stiffness was almost half the fixed-base stiffness and the share of horizontal pile group was significant. Then, they concluded that soil-structure interaction was relatively pronounced in bridges founded in weaker soils and it was more strongly related to the ratio of the pier flexural stiffness and horizontal foundation stiffness rather than the soil shear modulus (G) alone. It is worth noting that these soils, although weak, are not as weak as those found in Mexico City lakebed region. Vlassis and Spyrakos

6

conducted a parametric study devoted to assess the effects of soil-

structure interaction on the response of seismically isolated bridges. Rubber bearings were located between the rigid viaduct and the bridge piers that were founded on a shallow soil stratum overlying a rigid bedrock. Isolator bearings were characterized using an equivalent linearization model assuming an equivalent secant stiffness and an equivalent viscous damping. Soil-structure interaction was modeled including equivalent linear translational and rotational springs and translational and rotational dashpots. Harmonic excitations were considered, in order to be able to get solutions in the frequency domain. Among the conclusions of this parametric study where the following: a) the fundamental period of the bridge-soil system was significantly increased when SSI was considered and, b) SSI did not appear to play a major role as far as increasing the effective damping for the system. Tongaonkar and Jagid 7 assessed SSI effects on peak responses of three-span continuous deck bridge seismically isolated by elastomeric bearings. Beam elements were used to model bridge piers. Springs and dashpots were used to model laminated rubber bearings. Abutments were assumed to be rigid. The soil surrounding the foundation of pier was modeled by frequency independent coefficients and the complete dynamic analysis was carried out in time domain using the complex modal analysis method. In order to quantify the effects of SSI, peak responses of isolated and non-isolated bridges were compared with the corresponding bridge ignoring these effects. Three different types of soils were considered: a) a hard soil with shear wave velocity Cs=394 m/s, b) a medium soil with Cs=134 m/s and, c) a soft soil with Cs=100 m/s. Three ground motions typical of hard soils were used: a) 1940 El Centro, b) 1994 Northridge and, c) 1995 Kobe. Among their conclusion were the following:

a) SSI must be included in the analysis of isolated bridges when the stiffness of the supporting soil medium is less than 10 times the stiffness of isolation bearings, b) SSI affects the bearing displacements at the abutment and ignoring these effects will underestimate the design displacement at abutment, c) the effects of SSI were found to be more pronounced for stiff bridges in comparison to flexible bridges and, d) the effects of SSI are more pronounced for bridges with stiff isolation system in comparison to flexible system. Soneji and Jangid

8

assessed the influence of dynamic SSI on the behavior of seismically

isolated cable-stayed bridge supported on a rigidly capped vertical pile groups, which pass through moderately deep, layered soil overlying rigid bedrock. The Quincy Bay View Bridge crossing the Mississippi River at Quincy, Illinois was used as the model. High-damping rubber bearings were considered for the isolation system. Piles closely grouped together beneath the towers were modeled as a single equivalent upright beam. Soil–pile interaction was idealized as a beam on nonlinear Winkler foundation using continuously distributed hysteretic springs and viscous dashpots placed in parallel. The hysteretic behavior of soil springs was idealized using Bouc–Wen model. Three types of layered soil strata were considered: soft, medium and firm. Ground accelerations recorded in firm soils during following strong earthquakes were used: (1) The 1940 Imperial Valley, (2) 1995 Kobe, (3) 1989 Loma Prieta and, (4) 1994 Northridge. Bidirectional seismic excitations were considered. Among the conclusions, it was found that SSI modeling is essential for effective design of seismically isolated cable-stayed bridge, specifically when the towers are very much rigid and the soil condition is soft to medium. For soft soil conditions, the bearing displacement may be underestimated if SSI is ignored. Significant influence of soil–pile interaction is observed on tower base shear response in the transverse direction. Higher base shear forces were obtained when SSI effects were modeled compared with a fixed-base model. Stehmeyer and Rizos

9

studied the effects of SSI on both an un-retrofitted and seismically

isolated typical bridge structure in its transverse direction. Eight different soils were considered, with shear wave velocities ranging from 56 m/s to 190 m/s. The 1940 E-W El Centro record was used for the analyses. From the physical model considered, overall SSI effects were identified as: a) elongation of damped period of vibration, b) increased relative pier displacements, c) increased composite damping ratio for the unisolated soil-foundationstructure systems and, d) decreased composite damping ratio for the isolated soil-foundationstructure systems. All of these identified effects were amplified for softer soils such as silty

sands, and less pronounced for moderate to very stiff soils. The effects of considering SSI in the analysis of seismically isolated structures were shown to reduce the effectiveness of the isolation system of reducing the demands on the isolated structure for moderate to soft soils. Cho et al. 10 analyzed the seismic behavior of a base-isolated liquid storage tank modeling the soil–structure–liquid interaction with a hybrid formulation, which combines finite shell elements for the structure and boundary elements for the liquid and soil. An earthquake response analysis for the liquid storage tank was assessed using the N–S component of the El Centro earthquake as the input ground motion with a peak ground acceleration of 0.348 g. The input motion in the time domain was transformed to the frequency domain motion using a Fast Fourier Transform and, in order to represent the response of the structure, the inversion of the FFT was used. A hard soil characterized by a shear-wave velocity Cs=508 m/s and a softer soil with Cs=150 m/s were considered. Among the conclusions of this very interesting study, the authors found that maximum radial displacements and resultant forces were reduced and responses were generally delayed as soil stiffness decreases, but liquid sloshing height and motion were not changed. Although liquid storage tanks with the base isolation system were less affected by soil stiffness compared with fixed-base tanks, the effect of the soil–structure interaction may increase the response. Spyrakos et al.

11

studied SSI effects on the response of base-isolated buildings. The

equations of motion were formulated in the frequency domain, assuming frequencyindependent soil stiffness and damping constants. An equivalent fixed-base system was developed that accounted for soil compliance and damping characteristics of the base-isolated building. Closed-form expressions were derived, followed by a thorough parametric study involving the pertinent system parameters. The described study concluded that the effects of SSI are more pronounced on the modal properties of the system, especially for the case of squat and stiff base-isolated structures. It is clear from the literature review that very little attention has been paid to the impact of SSI effects in base-isolated buildings, particularly those found in very weak soils. Previous parametric studies conducted for buildings were mostly solved in the frequency domain only 3, 11

. Previous studies that included SSI effects have considered ground accelerations recorded

in firm soils only 5, 7-10, and only one in the very soft soils as those found in Mexico City 1. In fact, taking aside the study by González and Noguez 1, the softer soil considered had a shearwave velocity of 56 m/s 9, which is still higher than those found in the lakebed region of

Mexico City. Nevertheless, most previous studies have warned that soil-structure interaction effects reduce the effectiveness of the isolation system as soils became softer. Therefore, the assessment of the impact of SSI in tentative base-isolated building projects for the very soft soils found in the lakebed region of Mexico City is relevant, particularly when in previous research studies for that region

1-2

dynamic instability of LRB isolators was

obtained when subjected to ground accelerations recorded in the lakebed zone during the 1985 Michoacán Earthquake, even when SSI effects were not included into the modeling 2. SCOPE OF THE RESEARCH STUDY Civil engineers from the Mexican power company requested to study the feasibility of using suitable base isolation systems for an architectural project for the main building for encapsulated power stations. Such buildings are planned to be build in the following six sites within the Valley of Mexico: Narvarte, Verónica, Culhuacán, Chimalpa, Los Reyes and El Rosal. The first step proposed in the research project was to evaluate if the structural system related to the architectural project for the encapsulated power station was adequate for the design goals of the project from an earthquake-resistant design viewpoint. The use of elastomeric bearings was not found to be a good solution for soft soil sites before 1-2

. In fact, lead-rubber bearings (LRB) would have trouble to isolate buildings directly (i.e.,

with their initial elastic stiffness) to a target period equal to three times the dominant period for the site (Ts) to “jump” the resonant period range for the site. If LRB are designed using an effective secant stiffness strategy, then these systems would irremediably transit first within the period range for resonant responses and, as shown in previous studies, these structures most likely would fail 2. Therefore, pendular systems were considered as a more viable option for this project. As a first approximation, base-isolated projects were studied considering site effects in the characteristics for the generated artificial ground motions, but ignoring soil-isolationstructure interaction effects, in order to distinguish the most plausible projects. Later, soilisolation-structure interaction effects were evaluated in 2-D models (under unidirectional seismic input) and 3-D models (under bidirectional seismic input).

DESIGN EARTHQUAKE Recent advances in seismology have significantly contributed to the knowledge of seismic hazard in Mexico, particularly in the improved assumed geometry for the Cocos plate in its portion subducted beneath the North American continental plate. In Mexican model seismic codes

12-13

the Mexican Republic has been divided in dozens of

sources of earthquakes. There are three possible origins for earthquakes in Mexico: a) subduction earthquakes, b) normal faulting of intermediate depth earthquakes and, c) shallow earthquakes of the continental crust. These sources obey the regional tectonic process and are reflected by the instrumental history of recorded earthquakes. Each one of these sources generates earthquakes with a constant rate per unit area. The activity of the i-th seismic source is specified in terms of the exceedance rate of magnitudes they generated, i(M). The exceedance rate of magnitudes measures how frequent earthquakes whit magnitude greater than a given one are generated. After the activity rate of each one of the seismic sources is determined, it is necessary to evaluate the effects (in terms of seismic intensity) that each source produces in a specific site (supposedly located at firm ground). For this purpose, it is required to know what intensity would take place in the site if an earthquake with a given magnitude occurs in the i-th source. Equations that relate magnitude, site-source distance and intensity are known as attenuation laws. It is considered that the seismic intensities of interest are the ordinates of the response spectrum Sa (pseudo accelerations, 5% of critical damping), quantities that are approximately proportional to the lateral inertial forces generated in the structures during earthquakes and depend on the natural period of vibration. Three attenuation laws dependent of the path from the source to the site are used in this study in a spectral scheme to account for the fact that the attenuation is different for waves of different frequency, so there are parameters for each vibration period. These laws are: costal earthquakes 14, intermediate depth earthquakes 15 and superficial earthquakes 16. Given magnitude and epicentral distance, the seismic intensity cannot be considered deterministic, because it is not free of uncertainties. It is often assumed that, for a given magnitude and distance, the intensity Sa is a random variable lognormally distributed, with median Am(M, R), given by the attenuation law and typical deviation of the natural logarithm equal to σlnA. For instance, if an earthquake with magnitude M and distance R has occurred, the probability that the spectral acceleration SA would be greater than a given value, Sa is;

 1 Sa  PrS A  Sa M , R   1    ln   ln A Am M , R 

(1)

Once the seismicity of the sources and the attenuation patterns of the generated waves in each one of them are known, the seismic hazard can be computed regarding the sum of the effects of all the sources and the distances between each source and the referred site. In this study, seismic sources are areas, where a spatial integration process to account for all possible focal locations is performed. Generally, it is assumed that, in a seismic source, all point have the same probability to be an epicenter (constant seismicity per unit area). In this case, the rates of exceedance of acceleration due to one seismic source (the i-th) are computed with the following equation 17:  d M   vi Sa    j wij     Pr S A  Sa M , Rij dM dM  M o



Mu



(2)

where j is the index for each one of the sub-elements in what the source has been divided, M0 and Mu are the minimum and maximum magnitudes considered in the analysis, Pr (SA>SaM, Rij) is the probability that the acceleration exceed the value Sa in the site, given that an earthquake with magnitude M is generated at the distance Rij, and Rij are the distances between the site and sub-element j of the source i. A weight wij for each sub-element is assigned, which is proportional to its size. Finally, the contributions of all sources -N- are summed to the seismic hazard of the site. This analysis is performed for several structural periods. This approach allows providing a uniform hazard spectrum (UHS) at rock as a seismic excitation. For 5% damping and a return period of 475 years, the basis of regulatory design spectra in Mexico´s Federal District Code

18

was obtained by Ordaz

19

and it is used in this

study. It will be named as “design earthquake” (Figure 2a). Once the target response spectrum is known, synthetic acceleration records can be obtained (Figure 2b). Such records are representative for the ground motion at the basement level. From these acceleration records, it is possible to compute the free field response at the surface. Specifically, a simulation method that fulfills the prescribed spectral amplitudes is used

20-21

. The principle of the method is to build transient signals whose response spectra

iteratively fit the target response spectrum. The synthetic acceleration records are shown in Figure 2b, as well as the compatibility matched with the objective response spectra (Figure

2a). These movements are artificial and conservative, since the calculated spectral ordinates are adjusted in all vibration periods, not only in the characteristic periods for the ground motion. These spectra exhibit two peaks in vibration periods near 0.25 and 1 s. The first is associated with near-field earthquakes (local and intraplate normal faulting) and the second far-field earthquakes (subduction).

SITE EFFECTS For the specific sites of interest in Mexico City, subsurface data are scarce. To account for the uncertainty in the data, plausible representations of the stratigraphic profile were generated randomly. For this purpose, the dominant site period ( Ts ) and the depth to the basement ( H s ) specified in the seismic zoning maps of Mexico´s Federal District Code

18

were taken as

benchmark values. For these sites, the depth of the basement is between 15 m and 40 m. The effective velocity for the site was obtained as Vs  4Hs Ts . Estimated basement shear waves velocity Vr was such that the resonant response for the sites, obtained with analytical transfer functions (stratigraphic model), was similar to that obtained with empirical transfer function (spectral ratios). For each simulation of the stratigraphic profile, free field response was calculated assuming vertical propagation of shear waves. During intense earthquakes, soils may respond in the nonlinear range of behavior. To account for the stiffness degradation and increment of damping that occurs during great deformations for the soil, the equivalent linear method

22

is applied. An iterative linear analysis is

conducted to correct dynamic parameters, which stops until the stiffness modulus and damping are compatible with the strain level reached by the soil under the corresponding extreme shaking. The degradation curves of the stiffness modulus and increment of internal damping are the accepted curves for Mexico City Valley 23. The shape of response spectra is strongly dependent on the site period. Free-field spectra provide information of amplification effects due to local soil conditions exclusively. Specific site response spectra, corresponding to the design earthquake and for 5% equivalent viscous damping, are shown in Figure 3 for all sites of interest. It can be observed from Figure 3 that, fortunately, none of the sites of interest has a long site period (in all cases T s≤1.1s); this is, they do not possess some characteristics of very soft soils. In fact, site effects are almost negligible in El Rosal site.

Ten synthetic acceleration records were generated for each site of interest. For space constraints, only the records generated for the Narvarte site are shown in Figure 4, as this is one of the studied sites where softer soils were found.

PRELIMINARY DESIGN FOR THE PROTOTYPE SF6 BUILDING The original proposed architectural project for the main building for encapsulated power stations is depicted in Figure 5. The following requirements were established by the Mexican power company for the seismic design of such buildings: a) The structural system must be made of structural steel. b) Composite steel decks must be used as floor systems. c) The seismic design must comply with all the requirements set in Mexico´s Federal District Code 24 and its complementary technical norms 18, 25-27. d) A fully compensated mat foundation should be design according to RCDF-04 and the power company specific guidelines, including specific in-depth geotechnical studies for the site. e) Dynamic soil-structure interaction effects must be included.

Preliminary structural design As it can be observed in Figure 1, steel moment frames have been used before as the structural system for the main building of encapsulated power plants already built in Mexico in non-seismic regions. Therefore, that is the reason why in the preliminary architectural project, moment frames were also proposed (Figure 5). Then, as a first approximation, the encapsulated SF6 building was solved with special moment-resisting steel frames (SMRSFs) and composite steel decks as floor systems. The ETABS model is shown in Figure 6a. A chessboard distribution of the supporting beams and steel decks was also done to balance the lateral stiffness and strength of the floor system in the main orthogonal directions (Figure 6b). Steel decks were designed according to AISC and ASCE recommendations for composite construction. For space constraints, details for the preliminary design according to the Mexican seismic codes guidelines using the inelastic design spectrum according to NTCS-2004 (Figure 7) including site effects for (Narvarte site, Ts=1.1s) are not presented. However, it is worth noting that it was demonstrated to the power company that a low‐rise moment frame solution was not a good structural solution for

encapsulated SF6 power station main buildings, because of their very large lateral flexibility with large fundamental periods (Table 1) and their strong first soft-story potential (Table 2). Both conditions are inappropriate for an effective base isolation project, even in favorable soil conditions (firm soils or rock). For that reason, three alternate structural projects were also studied, which ETABS models are depicted in Figure 8. In Model SF6-Contra (Figure 8a), the perimeter of the first story was stiffened adding steel chevron braces, trying to solve both the soft first story potential and the lateral flexibility. In Model SF6-TContra (Figure 8b), all stories are stiffened in the perimeter adding steel chevron braces to reduce the lateral flexibility significantly. Model SF6ContraCom (Figure 8c) is an alternate model for Model SF6-Contra, but using composite steel reinforced concrete (SRC) columns instead of steel columns of square box sections. Best results were obtained for the design of model SF6-Contra (Figure 8a). The final design for SF6-Contra model consisted of: a) columns of square box cross sections 50x50 cm (19.7 x 19.7 in) with a plate thickness t=7/8” at the first story and t=5/8” at the remaining stories, b) chevron braces of square box cross sections 20x20 cm (7.9x7.9 in) with a plate thickness t=1/4” at the first story, c) beam sections for perimeter frames are W16x77 for the first two stories and W16x57 for the remaining stories and, c) long-span beam sections for interior frames are also W16x77 for all stories and their design was ruled by the serviceability limit state. The main dynamic characteristics for SF6-Contra model are summarized in Table 3. It can be observed that the first modes are almost uncoupled (close to 100% pure). The lateral flexibility was considerably reduced, as the fundamental translational period of vibration is reduced 1.91 times the one for the original proposal (Table 1). Therefore, adding chevron bracing at the first story increased the global lateral stiffness for the structure around 3.64 times. Design story drifts (i), shear forces (Vi), lateral shear stiffness (ki) and lateral stiffness ratios (ki+1/ki) for the SF6-Contra model are shown in Table 4. It can be observed that story drifts are well below the Mexican Code drift limit =0.015 for special moment-resisting steel braced frames

13

. However, although the soft first story potential is still important in the Y

direction (k2/k1 =1.84<2.0), perhaps an effective isolation project would lessen the effects of such structural irregularity, as the structure above the isolation system should experience reduced lateral displacements. In general, the global dynamic characteristics for model SF6Contra are much more favorable for a successful seismic isolation project, as the effective

base-isolated period (TI) to the fixed-based period (Tfb) ratio would allow an important dynamic uncoupling.

BASE ISOLATION PROJECTS IGNORING SOIL-STRUCTURE INTERACTION Base-isolated projects for the SF6-Psuave and SF6-Contra models were first studied considering site effects in the characteristics for the generated artificial ground motions, but ignoring soil-isolation-structure interaction effects. An effective and efficient isolation project would allow the structure above the isolation system to move almost as a perfect rigid body. In order to achieve this goal, it is convenient to have a TI/Tfb ratio as large as possible 28. In recent studies, it was found that when TI/Tfb=8, even torsional effects due to stiffness eccentricities in the superstructure are considerably reduced

29

. It is also important that the target effective base-isolated period (TI) would be

much larger than the dominant period for the site (Ts). For soft soils in the Valley of Mexico, where site periods could be large, it seems convenient to try to achieve the following relation: TI/Ts≥3. From previous research experiences, lead-rubber bearings were not effective to isolate an apartment building

1

as well as school building structures

2

when subjected to

acceleration records typical of soft soils in Mexico City 2. In theory, pendular systems or sliding systems could be more effective in “jumping” the resonant period range in soft soils. Then, pendular systems were preferred in this exploratory study. Base-isolated projects were studied for the following three different idealized pendular systems: a) a pendular system with no additional damping, b) a pendular system with an addition equivalent viscous damping of 15% the critical and, c) a friction pendulum

30

. The

first two options are representative of commercial systems such as the GT-BIS system 31. In fact, the cheapest version for the GT-BIS system does not provide additional damping ( very close to 0% for practical purposes). Different commercial options are now available for the friction pendulum system. The target effective base-isolated period for the models was TI=3.5s. It was assumed that base isolators were placed below each column line (Figure 6b), this is, 13 isolators were considered. For the friction pendulum system, the following properties were initially considered: a) maximum coefficient of sliding friction fs=0.095, b) a) minimum coefficient of sliding friction ss=0.05, c) constant which controls the transition of coefficient of sliding friction from maximum to minimum value of 0.8 and, d) yield displacement y=1.27 cm (0.5 in).

Nonlinear dynamic analyses of different base-isolated projects for three different idealized pendular systems were conducted in 3D-Basis 32. Bidirectional effects were considered using two horizontal orthogonal components for the simulated ground motions for each site (for example, Figure 4), which were selected randomly. Peak dynamic responses for some response parameters of interest for the different base-isolated projects of SF6-Psuave and SF6-Contra models are reported in Tables 5 to 8 for some of the softer soils of interest. The parameters of interest to monitor were: a) peak displacement for the isolation system (Di-max), b) peak base shear for the structure above the isolation system (Ve-max), c) peak roof displacement for the structure above the isolation system, measured from the isolation level (De-max). The following general observations can be done from the results reported in Tables 5 to 8, where “unacceptable” results are highlighted in orange and “less acceptable” results are highlighted in yellow: 1. As it was assumed a-priori, the original architectural proposal (SF6-Psuave model) is not adequate for an effective base isolation project in the soft soil sites of interest. The structure above the isolation system is very flexible, therefore, the ratio TI/Tfb<3.5 in the critical direction, so modal coupling occurs. As the structural project also has a strong stiffness soft first story potential (Table 1), using a pendular isolation system for a target base isolated period TI=3.5s does not help the structure to avoid an unstable response, even with an effective equivalent viscous damping =15%. From the studied options, the only possibility for this structure to have a somewhat effective isolation project is using friction pendula, but the lateral displacements for the structure above the isolation system are not small enough for a building of that height. Then, it was confirmed that SF6-Psuave model should not be considered in the following detailed analyses, as there is no warranty that it would promote an effective and safe enough isolation project when SSI effects would be considered. 2. In contrast, without including SSI effects, it is confirmed that, in theory, isolation projects for the SF6-Contra model are more promising and potentially effective. The reasons behind it are that such project is both reasonably uncoupled with the structure above the isolation system (TI/Tfb≈7) and with the site period (TI/Ts>3). However, it is also observed that using a pendular system alone with no additional damping (=0%) is not good enough, as very large displacements (around 65 cm or 25.6 in) could be developed by the isolation system. Therefore, a bigger gap between the base-isolated structure and the surrounding ground is required. Providing additional viscous

damping (=15%) for the pendular system is an option to reduce the displacements for the isolation system to 32% or less and the base shear to 43% or less of those for the undamped pendular system. Studied friction pendula were found very effective to reduce the displacements for the isolation system to 26% or even less with respect to the undamped pendular system. However, peak base shears and top lateral displacements are increased with respect to a pendular system with an additional viscous damping (=15%). It is worth noting that the maximum coefficient of sliding friction considered in these simulations is relatively high. Therefore, most detailed studies were planned to consider also a smaller coefficient of sliding friction.

BASE ISOLATION PROJECTS INCLUDING SOIL-STRUCTURE INTERACTION SSI effects were evaluated in 2-D models (unidirectional seismic input) and 3-D models (bidirectional seismic input), as described in following sections.

2D modeling approximation The idealized model for the analysis of SSI in buildings with seismic isolation is shown in Figure 9a. For each direction of analysis, the structure is modeled as a shear beam with four degrees of freedom in horizontal translation and the foundation as a rigid block with two degrees of freedom, one in horizontal translation and the other in rotation. Additionally, the deformation of the isolator is considered, which is modeled as a spring and dashpot in the interface between the superstructure and the substructure support. The soil is modeled with spring and dashpots that depend on the excitation frequency; although in Figure 9b only the springs are shown, parallel dashpots also exist. Because damping for the soil and for the isolator are significantly higher than the one for the structure, and therefore no proportional to their mass and stiffness (non-proportional damping), the system lacks classical modes of vibration, not allowing performing a conventional modal superposition analysis. The analytical solution for the system depicted in Figure 9 was obtained and it is reported in detail elsewhere

33

. It is worth noting that, in the

solution, the following definitions are relevant for the understanding of some of the results reported in following sections. To characterize the isolation system, it was necessary to introduce the following two parameters:

To 

2

o

with o 

ko 4

Mo   Mn

(3)

n1

o 

Co 4 2 M o   M n o n 1  

(4)

where To (basically TI) is interpreted as the natural period for the base-isolated structure neglecting SSI effects and  o is the effective damping ratio for the natural period of a conventional base-isolated structure. In order to determine the response for the system, sometimes it is convenient to use the complex frequency response method

34

. Thus, natural frequencies for the system and

amplitudes of vibration of the considered degrees of freedom can be simultaneously computed. It can be proved that the frequencies for the fixed-base structure can be identified by the transfer function of the roof as 33: H e ( ) 

xb  xc  h4 (b  c )  xo  x4 xg

(5)

where all terms are defined in Figure 9. For the computation of floor spectra, it is necessary to assess the transfer function for the base: H o ( ) 

xb  xc  ho ( b   c )  xo xg

(6)

These complex functions relate the roof or base-structure responses with the excitation on the ground surface. For SSI analyses considering 2D response and unidirectional seismic input only, the fixedbase SF6-Contra model was used (Figure 8a), as previous results lead one to believe it was the “best” structural option from the limited set that was studied. Weights and heights for each floor for the model, as well as interstory stiffnesses in two orthogonal directions are given in Table 9. The weight of the isolated structure is Wo  833.3 t (1,835.5 kips). The stiffness of the system for a target isolated period To  3.5s is Ko  4 2Wo gTo2  2.74 t/cm (15.33 kips/in). The mat foundation is of a rectangular section

with length of 20.1 m (65.9 ft), 13.80 m (45.3 ft) in width and 3.5 m (11.5 ft) in depth.

Transfer functions The use of transfer functions allows one to simultaneously identify natural frequencies for the system and amplitudes of vibration for the degrees of freedom. Roof transfer functions for the fixed-base structure ( K o   and Vs   ), the structure with flexible base (including SSI effects) and for the base-isolated structure without and with SSI effects are shown in Figures 10 to 13. Computations were done for the two directions of analysis, considering effective viscous damping  o  5 and 15% for the isolation system. Higher amplifications for the transfer functions are obtained for the fixed-base structure, whereas the smallest amplifications are those corresponding to the isolated structure with SSI. SSI effects are noticeable for the non-isolated and base-isolated structures for Narvarte (Figure 10), Verónica (Figure 11) and Culhuacán (Figure 12) sites, and negligible for Los Reyes (Figure 13), Chimalpa (not shown) and El Rosal (not shown) sites. The most pronounced SSI effects are observed for Culhuacán site (Figure 12). It is also observed from these figures that characteristic high frequency peaks for the fixed‐base structure that are eliminated in the base isolated structure without SSI effects are restored (though shifted) in the base isolated structure when SSI effects are considered, but their amplitude is considerably smaller with respect to the non-isolated structures. Therefore, one of the advantages of base‐isolation, which is to reliably reduce floor accelerations to protect acceleration sensitive nonstructural components and equipment is maintained, as it is shown later using floor spectra. It is worth noting that the obtained results are a consequence that, for all sites under study, kinematic SSI effects (observed in the amplitude reduction with respect to the fixed-base structure) were more important than inertial SSI effects (observed in the frequency shift) which were almost negligible.

Displacement and shear force profiles To compute the time response under seismic excitation, convolutions for the system transfer functions with design simulated earthquakes were done for each site of interest. Mean values of peak shear forces and displacements for the fixed-base structure and the isolated structure without and with SSI were computed. Shear force and lateral displacement profiles from the base of the foundation to the roof, for the fixed-base structure ( K o   and Vs   ) Vs   ), the structure with flexible base

(including SSI effects) and for the base-isolated structure without and with SSI are shown in Figures 14 and 15 for the two sites with softest soils. Calculations were done for the two directions of analysis, considering equivalent viscous damping  o  5 and 15% for the isolation system. SSI effects for shear forces are relatively more important than for displacements, especially in the upper floors. The effectiveness of the isolation when SSI is accounted for is preserved even for small damping. In fact, for the sites under study, peak responses for the base-isolated structure without (blue lines) and with SSI (red lines) are identical. As expected, the superstructure experienced an almost perfect rigid-body motion. Therefore, relative displacements become insignificant when compared with the deformations of the isolators, which are smaller than 30 cm for 5% damping. It is worth noting that the computation of seismic forces and lateral displacements for the design earthquake was done considering an equivalent linearization model for the base isolators. In contrast, the non-linear behavior of the supporting soil was accounted for as explained in detail elsewhere 12. Floor spectra Floor spectra are a relatively simple tool to assess quickly the impact of SSI effects in the effectiveness of a seismic isolation. Floor spectra for the isolated structure without and with SSI effects are shown in Figures 16 (Narvarte site) and 17 (Verónica site), where they are also compared with the free-field spectra. Computations were done for both directions of analysis, considering equivalent viscous damping  o  5 and 15% for the isolation system. Changes on response spectra in short periods are due to SSI effects, being kinematic effects more important than inertial effects. It can be observed from these results that for baseisolated structures and within the short period range, slightly smaller seismic forces are developed when SSI are neglected than when SSI are considered. However, this increment is too small to reduce the effectiveness of base isolation to protect acceleration sensitive nonstructural components and equipment. Additionally, it is shown that the effectiveness of the isolation system is notably increased for a 15% equivalent viscous damping for the isolator.

3D modeling Complete 3D ETABS models for the SF6-Contra project were built to model both base isolators and soil structure interaction, as shown in Figure 18. The box mat foundation (3.5 m or 11.5 ft in depth) was composed of 25 cm (10 in) thick reinforced concrete (RC) perimeter walls and a 25 cm (10 in) thick RC slab. The perimeter gap between the mat foundation and the strong floor was assumed to be 60 cm (23.6 in) at all four sides. Soil-structure interaction modeling included translational and rotational springs and dashpots in contact with the outer face of the perimeter walls and the bottom face of the slab of the foundation. The values for the spring and dashpot constants were obtained including the geotechnical information for each site, according to the procedure outlined in Mexico´s Federal District Code 18. Isolators were modeled as 3-D nonlinear spring as suggested in ETABS manuals

35-36

. For

pendular systems, an isolator was located below each column line (13 in total). Each isolator was connected to the strong floor at the ground level and to the bottom slab of the mat foundation (Figure 18a). Pendular systems with no additional damping (=0%) and with an additional equivalent viscous damping of 15% the critical were studied. For the friction pendulum system, additional basement columns were modeled to connect the bottom slab of the mat foundation with each isolator, and then each isolator to the strong floor at the ground level. Two different options were studied. In the first one, the following properties were considered: a) maximum coefficient of sliding friction fs=0.095 and, b) minimum coefficient of sliding friction ss=0.05. In the second one, the following properties were considered: a) maximum coefficient of sliding friction fs=0.04 and, b) minimum coefficient of sliding friction ss=0.03. In both cases, the constant which controls the transition of coefficient of sliding friction from maximum to minimum value (variation of the coefficient of friction with velocity) was set to 20, the radius of curvature of the spherical contact surface was R=3.04 m (10 ft) and the total effective lateral stiffness of the pendular system in each horizontal orthogonal direction was KI=2.4 t/cm (13.4 kips/in). Ten mode shapes were required to obtain at least 90% of the modal mass participating in each orthogonal direction (Table 10). The first three mode shapes are coupled, as a consequence that the plan is irregular and then, the centers of mass for the floor systems do not coincide

with the centers of stiffness of both the isolators and the superstructure. Nevertheless, it is known that friction pendula are effective to reduce torsional responses 30, 37-41. It is worth noting that although the same properties for the isolation systems were considered with respect to the 3D models without SSI interaction (3D-Basis models). The objective effective period TI=3.5s was slightly underestimated (less than 0.78% for the fundamental mode, Table 11) because: a) the presence of a stiff mat foundation was more important than the added mass of the same mat foundation, b) the stiffness constants to model SSI effects were high, so inertial SSI effects are reduced. In addition, it is worth noting that because the spring constants to model SSI effects vary for each site, then, the period of vibration for the mode shapes change for each site of interest, as reported in Table 11. However, it can be observed in that table that the period variation is negligible for practical purposes (in fact, to three digits there is no difference for Los Reyes and El Rosal sites). Once the main dynamic properties for the 3D models which includes SSI effects were reasonably calibrated, nonlinear dynamic analyses were conducted under the action of two horizontal orthogonal components for the simulated ground motions for each site (for example, Figure 4), which were selected randomly. Peak dynamic responses for some response parameters of interest for the different base-isolated projects of SF6-Contra model are reported in Tables 12 to 17 for each site of interest. The parameters of interest to monitor were: a) peak displacement for the isolation system (Di-max), b) peak base shear for the structure above the isolation system (Ve-max), c) peak roof displacement for the structure above the isolation system, measured from the isolation level (De-max), d) peak story drift (max), that always occurred at the third story (N3), e) peak horizontal displacement for the mat foundation (Dmax-h-mf) and, f) peak vertical displacement for the mat foundation (Dmax-vmf).

The following general observations can be made from the results reported in Tables 12 to

17 where again, “unacceptable” results are highlighted in orange and “less acceptable” results are highlighted in yellow: 1. Base isolation is potentially effective for the SF6-Contra structural solution for encapsulated SF6 power stations for the sites of interest, even considering SSI effects. The reasons behind it is that the project is reasonable uncoupled with the structure above the isolation system (TI/Tfb≈7) and with the considered site periods (TI/Ts>3). 2. It is confirmed that a simple pendular system with no additional damping (=0%) is not good enough for this project, as large displacements are required for the isolators.

A perimeter gap of 60 cm (23.6 in) between the mat foundation walls and the strong base floor was considered in the 3D models. From the results reported in Tables 12 to 17, one can conclude that pounding effects (which were not modeled) should be included, as peak isolator displacements surpass 60 cm for all sites, except Culhuacán (Table 14). Of course, a better structural solution would be to use a much larger gap. In addition, peak roof displacements (De-max) and story drifts (max) are always the largest ones, compared to other isolation systems, despite the fact that the mat foundation displacements (Dmax-h-mf and Dmax-v-mf) are not necessarily the largest ones. 3. From the structural behavior viewpoint, it seems that overall, the “best” option from the ones studied is the pendular system with an additional equivalent viscous damping =15%. Better balances of all relevant structural parameters under study are obtained, for example: a) reasonable peak displacements for the isolators, Di-max (less than 25 cm or 10 in), b) peak base shears for the superstructure (Ve-max) are the smallest for all sites, c) for most sites, De-max and max are also the smallest ones and, d) except for the Culhuacán site (Table 14), the smallest displacements for the mat foundation (Dmax-hmf and

Dmax-v-mf) are obtained.

4. It is confirmed that friction pendula is also an attractive solution. The advantage that they offer is that peak displacements for the isolation systems could be further reduced with respect to pendular system with additional damping. However, peak base shears for the superstructure are higher. It is also observed that these systems seem to be more sensitive to soil-structure interaction effects, as larger peak displacements for the mat foundation (Dmax-h-mf and Dmax-v-mf) were obtained. 5. Regarding the coefficients of sliding friction, for most of the sites of interest, a better overall performance was obtained using the maximum coefficient of sliding friction fs=0.04 and ss=0.03 than fs=0.095 and ss=0.05. Using smaller coefficients of sliding friction (fs=0.04) reduced the following peak parameters with respect the larger ones (fs=0.095): Ve-max, De-max, max, Dmax-h-mf and Dmax-v-mf. The only advantage of using a larger coefficient of sliding friction (fs=0.095) is to further reduce peak displacements for the isolators, Di-max. 6. It is worth noting that all the reported peak drifts for the structure above the isolation system (max) are below the allowable drift limit u=0.007 proposed for ductile steel

frames according to the guidelines for the seismic design of base-isolated structures 42 of the Manual of Civil Structures, a model code in Mexico

12-13

. However, the

allowable drift limit u=0.004 proposed for braced steel frames in the same guidelines is only satisfied entirely for friction pendula with coefficients of sliding friction fs=0.04 and ss=0.03 and for pendular systems with =15% for all sites except Culhuacán (Table 14). 7. The rocking effect of the foundation was more important when friction pendula was used. As the efficiency of this system depends on compressive axial forces transmitted to an isolation system that it is layered in a perfect horizontal plane, it is worth noting that rocking effects due to SSI could be very important for much softer soils than the ones considered in this study. Fortunately, for the sites of interest, peak vertical displacements for the foundation (at its corners) were smaller than 2.5 cm (1 in), so the dynamic stability of friction pendula was not jeopardized. 8. Comparing the results where SSI are neglected (Tables 5 to 8) and included (Tables 12 to 15), it is found that including SSI effects: a) peak top displacements for the structure (De-max) noticeable increase, b) peak displacements for the isolators (Di-max) increase for pendular isolators with added damping, but not necessarily for friction pendula, c) peak base shears for the structure (Ve-max) decrease for friction pendula, but not necessarily for pendular isolators with added damping. Finally, it was concluded that effective base-isolation projects for encapsulated SF6 power station in the sites of interest (with relatively soft soils and site effects) using pendular systems are possible, even when SSI effects are accounted for.

CONCLUDING REMARKS It is a common belief of many researchers and practicing engineers worldwide that by not considering the soft soil site effects for a building structure, the resulting seismic design remains on the safe side. Although that may happen in many cases of practical interest, particularly in soils that are not soft enough, the fact is that in very soft soil sites, such as those found in the Valley of Mexico, SSI should be included for the following reasons: (1) impressive collapses of foundations and complete buildings related to SSI rocking effects have been observed during strong earthquakes because of the long strong shaking durations, starting with the July 28, 1957 Earthquake (M=7.5)

43

and the September 19, 1985

Michoacán Earthquake (Ms=8.1)

44-45

that severely affected Mexico City and, (2) previous

research studies have confirmed that SSI effects within the Valley are not always beneficial, as amplified responses can also be obtained 1, 46. Therefore, because the aim of this study was to evaluate the feasibility of using seismic isolation in buildings for encapsulated power stations in the Valley of Mexico, then the impact of SSI effects (kinematic and inertial) should be assessed. Seismic excitation was given in terms of a uniform hazard spectrum (UHS) at rock for 475 year return period. The studied sites of interest have soft soils deposits with site periods below 1.1 seconds. The original architectural proposal for the main encapsulated SF6 building consisted of steel moment frames, as previous projects of this kind in non-seismic regions of Mexico were solved with such structures. This structural system was found to be inadequate for seismic isolation projects in the sites of interest. Reasons were the high lateral flexibility of the structural system and its strong soft first story potential. Unstable responses were obtained for pendular isolation projects, even when SSI effects were not considered. To solve the lateral flexibility and the soft first story potential, three additional structural solutions were evaluated. The “best” solution consisted of adding chevron braces to the perimeter frames at the first story (double height), solving the first story potential and providing the system with enough lateral stiffness to make the structure candidate for a suitable base isolation project. Isolation projects based on simple pendular systems with added damping and friction pendula were found effective for this structural solution, as a consequence that reasonable uncoupled responses were obtained considering or neglecting SSI effects, because in the design process the following period ratios were targeted: TI/Tfb≈7 and TI/Ts>3. When assessing SSI effects, the following observations can be done:  Using the results from transfer functions (unidirectional seismic input), it was found that dynamic uncoupling between the fundamental base-isolation period (TI) and the fundamental fixed-base period (Tfb) for the structure is maintained when SSI is considered, such that the seismic isolation continues to be effective.  For linear isolators, SSI effects are relatively more important for shear forces than for displacements, especially in the upper floors.  From floor spectra results of 2D models under unidirectional seismic input, and considering pendular isolators with added damping only, it was observed that smaller shear forces are developed in the superstructure when SSI are neglected than when

SSI are included, although SSI effects are relevant only for short periods. It was also observed that characteristic high frequency peaks for the fixed‐base structure that are eliminated in the base isolated structure without SSI effects are restored (though shifted) in the base isolated structure when SSI effects are considered, but their amplitude is considerably smaller with respect to the non-isolated structures. Therefore, one of the advantages of base isolation, which is to reliably reduce floor accelerations to protect acceleration sensitive nonstructural components and equipment is maintained.  From detailed 3D models under bidirectional seismic input, and comparing the results where SSI are neglected or included, it was found that, including SSI effects: a) peak top displacements for the structure noticeable increase, b) peak displacements for the isolators increase for pendular isolators with added damping, but not necessarily for friction pendula, c) peak base shears for the structure decrease for friction pendula, but not necessarily for pendular isolators with added damping.  From detailed 3D models under bidirectional seismic input, it was found that rocking effects of the mat foundation were more important when friction pendula were used.  From the structural behavior viewpoint, it seems that overall, the best option for an effective isolation project from the ones studied is the pendular system with an additional equivalent viscous damping =15%. Better balances of all relevant structural parameters under study were obtained. In particular, the following peak response parameters for the structure, which are those of the main interest for this project, were reduced: peak base shears, peak top displacements, peak story drifts and peak vertical and horizontal displacements for the mat foundation. However, friction pendula is also an attractive solution. The advantage that they offer is that peak displacements for the isolation systems could be further reduced with respect to pendular system with additional damping, but structural responses increases. Based upon the results obtained in this research, it was concluded that effective base-isolation projects using pendular systems for encapsulated SF6 power station in the sites of interest in the Valley of Mexico (with relatively soft soils and site effects) are possible, even when SSI effects are accounted for. However, additional requirements were recommended to the Mexican power company, in order that engineering design firms would be able to achieve successful design projects, primarily:

 The proposed structural system should be designed to avoid soft stories and be perfectly documented in the structural design files.  The structural system should possess enough lateral stiffness so the fundamental fixed-base period should be around or below 0.5 seconds. 

Dynamic soil-structure interaction effects should be included.

 The effective period for the base-isolated structure (TI) at the maximum displacement for the design earthquake scenario must be greater than: a) three times the site period (TI/Ts≥3), b) seven times the fixed-base period for the structure (TI/Tfb≥7) or, c) 3.5 seconds (TI=3.5s).  Pendular systems with added damping or friction pendula should be used. Other pendular systems available in the market could be used, granted that it is demonstrated with detailed studies that they provide a similar level of protection. Finally, it is also worth noting that although pendular isolation systems were found promising for base-isolation projects in relatively soft soils of the Valley of Mexico, the studied sites do not correspond to the softer soils found within the Valley, which site periods range from 1.25s to 4s approximately (1.25s ≤ Ts ≤ 4s)

18, 47

. Previous studies have reported very

important rocking effects due to SSI in conventional buildings located in soft soils with site periods around 1.25 s

44-45, 48-49

. The efficiency of many isolation systems depend on that the

isolation system is layered in a perfect horizontal plane. Rocking effects due to SSI could be very important for much softer soils than the ones considered in this study, as well as sudden or long-term differential soil settlements, particularly important in non-consolidated soft soils. These aspects should be carefully assessed in future research projects.

ACKNOWLEDGEMENTS The financial support of Comisión Federal de Electricidad (CFE) through a research grant to Instituto de Ingeniería, UNAM, is gratefully acknowledged. The comments and suggestions of anonymous reviewers were important to improve this paper and are gratefully acknowledged.

REFERENCES 1. González, R. and Noguez, F. J. (1994), “Respuesta teórica de un edificio con aislamiento de base”, Revista de Ingeniería Sísmica, 46 23-52 (in Spanish). 2. Tena-Colunga, A. (1996), “Some retrofit options for the seismic upgrading of old low-rise school buildings in Mexico”, Earthquake Spectra, 12 (4) 883-902. 3. Constantinou, M. C. and Kneifati, M. C. (1988), “Dynamics of soil-base-isolatedstructure systems”, ASCE Journal of Structural Engineering, 114 (1) 211-221. 4. Thakkar, S. K. and Maheshwari, R. (1995), “Study of seismic base isolation of bridge considering soil structure interaction”, Proceedings: Third International Conference on Recent Advances in Geotechnical Earthquake Engineering and Soil Dynamics, St. Louis, Missouri, Vol. I, Paper No. 5.36, pp.397-400. 5. Chaudhary, M. T. A., Abé, M. and Fujino, Y. (2001), “Identification of soil-structure interaction effect in base-isolated bridges from earthquake records”, Soil Dynamics and Earthquake Engineering, 21 713-725. 6. Vlassis, A.G. and C. C. Spyrakos (2001), “Seismically isolated bridge piers on shallow soil stratum with soil structure interaction”, Computers and Structures, 79 2847-2861. 7. Tongaonkar, N. P. and R.S. Jangid (2003), “Seismic response of isolated bridges with soil–structure interaction”, Soil Dynamics and Earthquake Engineering, 23 287–302. 8. Soneji, B. B. and R.S. Jangid (2008), “Influence of soil–structure interaction on the response of seismically isolated cable-stayed bridge”, Soil Dynamics and Earthquake Engineering, 28 245–257. 9. Stehmeyer III, E. H. and D. C. Rizos (2008), “Considering dynamic soil structure interaction (SSI) effects on seismic isolation retrofit efficiency and the importance of natural frequency ratio”, Soil Dynamics and Earthquake Engineering, 28 468–479. 10. Cho, K. H., M. K. Kim, Y. M. Lim and S. Y. Cho (2004), “Seismic response of baseisolated liquid storage tanks considering fluid–structure–soil interaction in time domain”, Soil Dynamics and Earthquake Engineering, 24 839–852. 11. Spyrakos, C.C., I. A. Koutromanos and Ch. A. Maniatakis (2009), “Seismic response of base-isolated buildings including soil–structure interaction”, Soil Dynamics and Earthquake Engineering, 29 658–668. 12. MOC-2008 (2009), Manual de diseño de obras civiles. Diseño por sismo. Recomendaciones y Comentarios, Instituto de Investigaciones Eléctricas, Comisión Federal de Electricidad, December (in Spanish). 13. Tena-Colunga, A., Mena-Hernández, U., Pérez-Rocha, L. E., Avilés, J., Ordaz, M. and Vilar, J. I. (2009), “Updated seismic design guidelines for buildings of a model code of Mexico”, Earthquake Spectra, 25 (4) 869-898, doi: 10.1193/1.3240413. 14. Ordaz, M., Jara, J. M. and Singh, S. K. (1989), “Riesgo sísmico y espectros de diseño en el estado de Guerrero”, Informe conjunto del II-UNAM y el Centro de Investigación Sísmica AC de la Fundación Javier Barros Sierra al Gobierno del estado de Guerrero, Instituto de Ingeniería, UNAM, Proyectos 8782 y 9745, México (in Spanish).

15. García, D., Singh, S. K., Herráiz, M., Ordaz, M. and Pacheco, J. (2005), “Inslab earthquakes of Central Mexico: peak ground-motion parameters and response spectra”, Bulletin of the Seismological Society of America, 95 2272-2282. 16. Abrahamson, N. A. and Silva, W. J. (1997), “Empirical response spectral attenuation relations for shallow crustal earthquakes”, Seismological Research Letters, 68 94-127. 17. Esteva, L. (1967), “Criterios para la construcción de espectros para diseño sísmico”, Proceedings: 3er Simposio Panamericano de Estructuras, Caracas, Venezuela (in Spanish). 18. NTCS-2004 (2004), “Normas técnicas complementarias para diseño por sismo”, Reglamento de Construcciones para el Distrito Federal, Gaceta Oficial del Distrito Federal, October (in Spanish). 19. Ordaz, M. (2014). Personal communication. 20. Tena-Colunga, A., Godínez-Domínguez, E. A. and Pérez-Rocha, L. E. (2007), “Vulnerability maps for reinforced concrete structures for Mexico City’s Metropolitan Area under a design earthquake scenario”, Earthquake Spectra, 23 (4) 809-840. 21. Godínez, E. A., Pérez, L. E. and Tena, A. (2013), “Metodología para la generación de acelerogramas sintéticos”, Pakbal, 12 (27) 21-29. 22. Seed, H. B. and Idriss, I. M. (1969), “Influence of soil conditions on ground motions during earthquakes”, ASCE Journal of the Soil Mechanics and Foundation Division, 95. 23. Jaime, A. (1987), “Características dinámicas de las arcillas del Valle de México”, Ph.D. Thesis, Facultad de Ingeniería, UNAM (in Spanish). 24. RCDF-2004 (2004), “Reglamento de Construcciones para el Distrito Federal”, Gaceta Oficial del Distrito Federal, January (in Spanish). 25. NTCC-2004 (2004), “Normas técnicas complementarias para diseño y construcción de estructuras de concreto”, Reglamento de Construcciones para el Distrito Federal, Gaceta Oficial del Distrito Federal, October (in Spanish). 26. NTCDC-2004 (2004), “Normas técnicas complementarias para diseño y construcción de cimentaciones”, Reglamento de Construcciones para el Distrito Federal, Gaceta Oficial del Distrito Federal, October (in Spanish). 27. NTCEM-2004 (2004), “Normas técnicas complementarias para diseño y construcción de estructuras metálicas”, Reglamento de Construcciones para el Distrito Federal, Gaceta Oficial del Distrito Federal, October (in Spanish). 28. Skinner, R. I., Robinson, W. H. and Mc Verry, G. H. (1993), An introduction to seismic isolation, John Wiley and Sons, London. 29. Tena-Colunga, A. and Escamilla-Cruz, J. L. (2007), “Torsional amplifications in asymmetric base-isolated structures”, Engineering Structures, 29 (2) 237-247. 30. Zayas, V. A., Low, S. S. and Mahin, S. A. (1990), “A simple pendulum technique for achieving seismic isolation”, Earthquake Spectra, 6 (2) 317-333. 31. Foutch, D. A., Gambill, J. B. and Garza-Tamez, F. (1993), “Investigation of a seismic base isolation system based on pendular action”, Structural Research Series No 578, Department of Civil Engineering, University of Illinois at Urbana-Champaign.

32. Nagarajaiah, S., Reinhorn, A. M. and Constantinou, M. C. (1991), “3D-Basis: Nonlinear dynamic analysis of three-dimensional base isolated structures: Part II”, Technical Report NCEER-91-0005, National Center for Earthquake Engineering, State University of New York at Buffalo. 33. Pérez-Rocha, L. E., Avilés, J., Tena-Colunga, A. and Cordero, C. (2014), “Influence of site effects and soil-structure interaction on seismic isolation of buildings on soft soils”, Proceedings, 10th National Conference in Earthquake Engineering: Frontiers in Earthquake Engineering, Anchorage, Alaska, Paper No. 1437, July. 34. Clough, R. W. and Penzien, J. (1993), Dynamics of Structures, McGraw-Hill. 35. CSI-2005a (2005), “CSI analysis reference manual for SAP2000, ETABS, and SAFE”, Computers and Structures, Inc., Berkeley, California, USA, October. 36. CSI-2005b (2005), “ ETABS, integrated software for structural analysis and design. Software verification examples”, Computers and Structures, Inc., Berkeley, California, USA, November. 37. Nagarajaiah, S., Reinhorn, A. M. and Constantinou, M. C. (1993), “Torsional coupling in sliding base-isolated structures,” ASCE Journal of Structural Engineering, 119 (1) 130-149. 38. Ryan, K. L. and Chopra, A. K. (2004), “Estimating the seismic displacement of friction pendulum isolators based on non‐linear response history analysis”, Earthquake Engineering & Structural Dynamics, 33 (3), 359-373. 39. De la Llera, J. C., Almazán, J. L. and Vial, I. J. (2005), “Torsional balance of planasymmetric structures with frictional dampers: Analytical results”, Earthquake Engineering & Structural Dynamics, 34 (9) 1089-1108. 40. Vial, I. J., De la Llera, J. C., Almazán, J. L., and Ceballos, V. (2006), “Torsional balance of plan-asymmetric structures with frictional dampers: Experimental results”, Earthquake Engineering & Structural Dynamics, 35 1875-1898. 41. Seguín, C. E., Almazán J. L and De la Llera, J. C. (2013), “Torsional balance of seismically isolated asymmetric structures”, Engineering Structures, 46 703-717. 42. Tena-Colunga, A. (2013), “Seismic design of base-isolated buildings in Mexico. Part 1: Guidelines of a model code”, The Open Civil Engineering Journal, 7 17-31. 43. Tena, A. (2010), “Irregularidad estructural y su efecto en la respuesta sísmica de edificios”, Proceedings: V Congreso Iberoamericano de Ingeniería Civil en honor del Dr. José Grases, Mérida, Venezuela, CD-ROM pp 1-57 (in Spanish). 44. Esteva, L. (1988), “The Mexico Earthquake of September 19, 1985—Consequences, lessons, and impact on research and practice”, Earthquake Spectra, 4 (3) 413-426. 45. Mendoza, M. J. and Auvinet, G. (1988), “The Mexico Earthquake of September 19, 1985—Behavior of building foundations in Mexico City”, Earthquake Spectra, 4 (4) 835-853. 46. Avilés, J. and Pérez-Rocha, L. E. (2005), “Influence of foundation flexibility on Rμ and Cμ factors”, ASCE Journal of Structural Engineering, 131 (2), 221–230. 47. Tena-Colunga, A. (1999), “International seismic zone tabulation proposed by the 1997 UBC code: Observations for Mexico”, Earthquake Spectra, 15 (2) 331-360.

48. Del Valle, E., Foutch, D. A., Hjelmstad, K. D., Figueroa-Gutiérrez, E. and TenaColunga, A. (1988), “Seismic retrofit of a RC building: a case study”, Proceedings, 9th World Conference on Earthquake Engineering, Tokyo-Kyoto, Japan, VII 451456. 49. Foutch, D. A., Hjelmstad, K. D., Del Valle, E., Figueroa, E. and Downs, R. (1989), “The Mexico Earthquake of September 19, 1985 – Case studies of seismic strengthening for two buildings in Mexico City”, Earthquake Spectra, 5 (1) 153-174.

Figure 1. Typical encapsulated power station built in Mexico

a) Target uniform hazard spectrum (thick dashed line), response spectra simulations (thin lines) and average response spectrum (continuous thick line)

b) Synthetic acceleration records

Figure 2. Response spectra for rock for a return period of 475 years and their associated synthetic acceleration records

Figure 3. Response spectra of synthetic acceleration records (thin black lines) and average response spectrum (thick red line), for the design earthquake (475 years return period) at the sites of interest.

Figure 4. Synthetic acceleration records for Narvarte site for the design earthquake (475 years return period)

a)

c)

Main facade

Elevation, section A-A

b) Ground level plan

d) Elevation, section B-B

Figure 5. Architectural project for the main building of encapsulated SF6 power stations

b) Typical plan for steel decks

a) 3D ETABS model

Figure 6. 3D ETABS model SF6-Psuave for preliminary design

Figure 7. Elastic e inelastic design spectra for Narvarte site (Ts=1.1s)

a)

SF6-Contra

b)

SF6-TContra

c)

SF6-ContraCom

Figure 8. 3D ETABS models for alternate system for encapsulated SF6 power station main building

a) Reference model

b) Deformed configuration

Figure 9. Reference model and deformed configuration for soil-structure systems with seismic isolation

Figure 10. Roof transfer functions for the fixed-base structure (black), structure with SSI effects (green) and for the base-isolated structure without (blue) and with (red) SSI effects, Narvarte site

Figure 11. Roof transfer functions for the fixed-base structure (black) ), structure with SSI effects (green) and for the base-isolated structure without (blue) and with (red) SSI effects, Verónica site

Figure 12. Roof transfer functions for the fixed-base structure (black) ), structure with SSI effects (green) and for the base-isolated structure without (blue) and with (red) SSI effects, Culhuacán site

Figure 13. Roof transfer functions for the fixed-base structure (black) ), structure with SSI effects (green) and for the base-isolated structure without (blue) and with (red) SSI effects, Los Reyes site

Figure 14. Shear force and displacement profiles for the fixed-base structure (black), structure with SSI effects (green) and for the base-isolated structure without (blue) and with (red) SSI. Narvarte site

Figure 15. Shear force and displacement profiles for the fixed-base structure (black), structure with SSI effects (green) and for the base-isolated structure without (blue) and with (red) SSI. Verónica site

Figure 16. Floor spectra for the fixed-base structure (black), structure with SSI effects (green) and for the base-isolated structure without (blue) and with (red) SSI. Narvarte site

Figure 17. Floor spectra for the fixed-base structure (black), structure with SSI effects (green) and for the base-isolated structure without (blue) and with (red) SSI. Verónica site

a) Model

b) First mode of vibration

Figure 18. 3D ETABS model for the SSI-base isolation SF6-Contra project for encapsulated power station. It includes the box of the mat foundation, springs and dashpots for SSI and springs and dashpots (if necessary) for the isolation system

Table 1. Dynamic characterists for model SF6-Psuave Mode 1 2 3

T (s) 1.043 0.991 0.742

X Dir. 94.80 1.22 0.07

Modal mass Y Dir. 1.08 92.39 3.02

Rotation 0.18 3.21 93.46

X Dir. 94.80 96.02 96.08

Modal mass sum Y Dir. Rotation 1.08 0.18 93.47 3.39 96.48 96.85

Table 2. Story drifts, shear forces, lateral shear stiffness and lateral stiffness ratios for SF6Psuave model Story 4 3 2 1

X Direction Vi (t) ki (t/cm) 33.97 40.99 65.52 61.43 91.23 100.86 116.26 30.13

i 0.0131 0.0168 0.0217 0.0291

ki+1/ki 0.667 0.609 3.347

i 0.0124 0.0165 0.0220 0.0300

Y Direction Vi (t) ki (t/cm) 32.69 41.48 63.33 60.71 88.58 96.60 113.23 27.28

ki+1/ki 0.683 0.629 3.541

Table 3. Dynamic characteristics for model SF6-Contra Mode

T (s)

1 2 3 4 5 6

0.547 0.491 0.374 0.198 0.182 0.141

X Dir. 0.06 80.37 0.05 0.01 18.36 0.015

Modal mass Y Dir. 89.56 0.06 0.48 9.44 0.00 0.04

Rotation 0.33 0.05 80.56 0.37 0.04 17.49

X Dir. 0.06 80.43 80.48 80.50 98.85 98.87

Modal mass sum Y Dir. Rotation 89.56 0.33 89.63 0.38 90.11 80.94 99.55 81.31 99.55 81.36 99.58 98.84

Table 4. Story drifts, shear forces, lateral shear stiffness and lateral stiffness ratios for SF6-Contra model Story 4 3 2 1

i 0.0069 0.0079 0.0060 0.0032

X Direction Vi (t) ki (t/cm) 34.70 79.61 59.43 119.20 73.74 298.24 84.84 219.70

ki+1/ki 0.668 0.400 1.358

i 0.0089 0.0104 0.0088 0.0066

Y Direction Vi (t) ki (t/cm) 33.61 60.41 60.69 93.84 79.39 221.61 95.28 120.52

ki+1/ki 0.644 0.423 1.839

Table 5. Results for base isolation projects for SF6 power stations without SSI, Narvarte site Isolation System Pendular, =0% Pendular, =15% Friction pendulum

Di-max (cm) 65.83 20.96 14.06

SF6-Contra Model Ve-max (t) De-max (cm) 97.21 1.33 42.04 0.53 78.15 1.57

Di-max (cm)

27.92

SF6-Psuave Model Ve-max (t) De-max (cm) Unstable Unstable 80.02 5.58

Table 6. Results for base isolation projects for SF6 power stations without SSI, Verónica site Isolation System Pendular, =0% Pendular, =15% Friction pendulum

Di-max (cm) 64.20 19.16 16.62

SF6-Contra Model Ve-max (t) De-max (cm) 94.19 1.29 29.82 0.50 85.71 1.71

Di-max (cm)

22.81

SF6-Psuave Model Ve-max (t) De-max (cm) Unstable Unstable 106.20 5.22

Table 7. Results for base isolation projects for SF6 power stations without SSI, Culhuacán site Isolation System Pendular, =0% Pendular, =15% Friction pendulum

Di-max (cm) 61.79 16.39 3.68

SF6-Contra Model Ve-max (t) De-max (cm) 90.47 1.24 23.84 0.39 76.02 1.36

Di-max (cm)

7.01

SF6-Psuave Model Ve-max (t) De-max (cm) Unstable Unstable 89.18 5.11

Table 8. Results for base isolation projects for SF6 power stations without SSI, Los Reyes site Isolation System Pendular, =0% Pendular, =15% Friction pendulum

Di-max (cm) 61.12 15.88 3.38

SF6-Contra Model Ve-max (t) De-max (cm) 65.63 1.07 26.22 0.36 81.14 1.29

Di-max (cm)

3.00

SF6-Psuave Model Ve-max (t) De-max (cm) Unstable Unstable 84.61 4.16

Table 9. Parameters for isolation project for SF6-Contra model Level 4 3 2 1 0 Base

W (t) 134.8 141.7 142.8 160.3 253.8 331.0

h (m) 20.9 17.1 13.3 10.8 3.5

Kx (t/cm) 79.6 119.2 298.2 219.7 2.7 Variable

Ky (t/cm) 60.4 93.8 221.6 120.5 2.7 Variable

Table 10. Dynamic properties for the SF6-Contra project including SSI and a pendular isolation system. Narvarte site Mode

T (s)

1 2 3 4 5 6 7 8 9 10

3.483 3.298 2.379 0.394 0.353 0.249 0.187 0.185 0.166 0.154

X Dir. 29.18 32.57 3.92 0.00 0.04 0.00 10.48 23.36 0.00 0.36

Modal mass Y Dir. 29.73 33.03 2.87 0.09 0.00 0.00 22.25 10.83 0.07 1.12

Rotation 6.44 0.02 51.02 0.00 0.00 0.02 1.42 0.00 0.01 41.03

X Dir. 29.18 61.75 65.67 65.67 65.71 65.71 76.20 99.55 99.56 99.92

Modal mass sum Y Dir. 29.73 62.76 65.63 65.72 65.72 65.72 87.97 98.80 98.87 99.99

Rotation 6.44 6.46 57.48 57.48 57.48 57.50 58.92 58.93 58.94 99.97

Table 11. First three periods of vibration T(s) for the SF6-Contra Project including SSI and pendular isolation system for all sites of interest Mode 1 2 3

Site Narvarte 3.483 3.298 2.379

Verónica 3.480 3.296 2.378

Culhuacán 3.495 3.311 2.381

Los Reyes 3.473 3.289 2.376

Chimalpa 3.482 3.298 2.379

El Rosal 3.473 3.289 2.376

Table 12. Results for base isolation projects for SF6-Contra model with SSI, Narvarte site Isolation System Pendular, =0% Pendular, =15% Friction pendulum, fs=0.095 Friction pendulum, fs=0.04

Di-max (cm) 69.35 22.57 20.87 21.56

Ve-max (t) 64.76 35.48 59.33 47.06

De-max (cm) 9.48 5.58 7.20 5.45

max (N3) 0.0056 0.0036 0.0046 0.0036

Dmax-h-mf (cm) 0.49 0.26 0.73 0.56

Dmax-v-mf (cm) 1.43 0.97 2.31 2.04

Table 13. Results for base isolation projects for SF6-Contra model with SSI, Verónica site Isolation System Pendular, =0% Pendular, =15% Friction pendulum, fs=0.095 Friction pendulum, fs=0.04

Di-max (cm) 66.97 20.27 15.83 17.14

Ve-max (t) 62.74 33.43 66.76 54.76

De-max (cm) 9.06 5.69 7.03 5.41

max (N3) 0.0053 0.0036 0.0045 0.0035

Dmax-h-mf (cm) 0.38 0.21 0.61 0.48

Dmax-v-mf (cm) 1.29 0.89 2.36 2.09

Table 14. Results for base isolation projects for SF6-Contra model with SSI, Culhuacán site Isolation System Pendular, =0% Pendular, =15% Friction pendulum, fs=0.095 Friction pendulum, fs=0.04

Di-max (cm) 56.18 18.87 10.28 11.89

Ve-max (t) 62.65 51.30 46.73 43.00

De-max (cm) 11.50 8.53 6.19 5.15

max (N3) 0.0067 0.0053 0.0040 0.0033

Dmax-h-mf (cm) 0.74 0.48 0.61 0.51

Dmax-v-mf (cm) 3.11 2.45 2.15 2.00

Table 15. Results for base isolation projects for SF6-Contra model with SSI, Los Reyes site Isolation System Pendular, =0% Pendular, =15% Friction pendulum, fs=0.095 Friction pendulum, fs=0.04

Di-max (cm) 64.53 18.23 9.26 11.14

Ve-max (t) 82.11 79.66 159.26 142.26

De-max (cm) 7.93 4.87 5.46 4.58

max (N3) 0.0046 0.0030 0.0036 0.0029

Dmax-h-mf (cm) 0.12 0.08 0.15 0.13

Dmax-v-mf (cm) 0.88 0.60 1.67 1.56

Table 16. Results for base isolation projects for SF6-Contra model with SSI, Chimalpa site Isolation System Pendular, =0% Pendular, =15% Friction pendulum, fs=0.095 Friction pendulum, fs=0.04

Di-max (cm) 62.95 19.01 11.28 14.59

Ve-max (t) 59.36 29.97 49.14 43.88

De-max (cm) 8.91 5.74 6.36 5.19

max (N3) 0.0053 0.0037 0.0041 0.0034

Dmax-h-mf (cm) 0.42 0.23 0.61 0.50

Dmax-v-mf (cm) 1.36 0.98 2.19 2.01

Table 17. Results for base isolation projects for SF6-Contra model with SSI, El Rosal site Isolation System Pendular, =0% Pendular, =15% Friction pendulum, fs=0.095 Friction pendulum, fs=0.04

Di-max (cm) 61.58 17.78 6.77 7.06

Ve-max (t) 63.96 48.21 99.63 88.88

De-max (cm) 7.84 4.93 5.00 4.41

max (N3) 0.0046 0.0031 0.0034 0.0030

Dmax-h-mf (cm) 0.15 0.09 0.19 0.17

Dmax-v-mf (cm) 0.91 0.63 1.68 1.59

HIGHLIGHTS 

Base isolation projects for encapsulated power plants soft soils.



Site effects are considered in the study.



Soil-structure interaction effects are assessed.



Pendular isolation systems are considered.



Practical general design recommendations are provided.