UPLIFT EVALUATION OF STRUCTURES UNDER SEISMIC

LOADING: ASSESSMENT OF DIFFERENT CALCULATION

METHODS AND DESIGN GUIDELINES
Charisis T. CHATZIGOGOS1, Pierre-Alain NAZE2, Jean-Mathieu RAMBACH3, Pauline
BILLION4, Matthieu CAUDRON5 and Sylvain GUISARD6

ABSTRACT

It has been widely recognized that in the case of strong shaking and significant foundation uplifting,
the designer can no longer accept the assumption of linear behaviour and is obliged to adopt a
nonlinear analysis approach. This observation motivates asking to what extent linearity assumption
can be retained in presence of foundation uplifting.
The problem is confronted via a parametric study of an industrial building inspired from the
nuclear industry founded on a mat foundation. The soil-structure system is parameterized on the basis
of four dimensionless parameters reflecting aspect and embedment ratio of the building, rigidity
contrast between soil and superstructure and foundation bearing capacity with respect to building
mass. Dimensionless parameters are varied to give raise to different soil-structure configurations. Each
configuration is studied for increasing levels of seismic intensity using five analysis methods, ranging
from the simplest one (linear pseudo-static analysis, denoted LP) up to a complete nonlinear transient
analysis (denoted NLT).
Three aspects of the response are studied: a) maximum uplift ratio, b) floor response spectra in
positions of interest within the building and c) resultant force and moment at specific sections of the
structural system. The results highlight the accuracy of each method of analysis with respect to NLT
(reference result), and allow quantifying aspects such as the effect of kinematic interaction, the effect
of soil plasticity in dissipating uplift potential, the gain in passing from pseudo-static to transient
analysis, the relocation of seismic surcharge within the structure when uplift is initiated etc.
The results are compiled in a set of guidelines focusing on the range of validity of the examined
analysis methods with respect to seismic intensity and uplift potential.

INTRODUCTION

Foundation uplift has long been recognized as a major factor for determining seismic demand on
structures founded on shallow foundations. Increasing inevitably with seismic intensity, uplift can alter
dynamic characteristics of interacting soil-structure systems (typically shifting fundamental vibrating
modes to smaller frequencies) to the extent that modal superposition (linear) analysis methods, as

1
Project engineer, Géodynamique & Structure, Bagneux, charisis.chatzigogos@geodynamique.com
2
Managing Director, Géodynamique & Structure, Bagneux, pierre-alain.naze@geodynamique.com
3
Chief engineer, Géodynamique & Structure, Bagneux, jean-mathieu.rambach@geodynamique.com
4
Project engineer, EDF/SEPTEN, Lyon, pauline.billion@edf.fr
5
Project engineer, EDF/SEPTEN, Lyon, matthieu-m.caudron@edf.fr
6
Team Manager engineer, EDF/SEPTEN, Lyon, sylvain.guisard@edf.fr

1
prescribed in industrial and civil norms (Eurocode 8, ETC-C/RCC-CW for the nuclear industry), may
cease to be applicable.
For pseudo-static methods in particular, and in presence of uplift, overall stability evaluation
with respect to overturning is known to be problematic for high seismic intensity levels and elongated
structures as these ones are very quickly shown to overturn. From that point on, it is necessary for the
designer to upgrade to some sort of nonlinear and/or transient analysis method, with all subsequent
modeling difficulties. This observation motivates asking to what extent linearity assumption and the
pseudo-static approach can be retained in a rigorous seismic design methodology for structures
allowed to uplift. In addition, it sets the challenge of probing the relevance of different analysis
methods (linear/nonlinear, transient/pseudo-static) for increasing seismic demand in the presence of
foundation uplifting.

PARAMETRIC STUDY

The present paper aims at shedding light to the above questions through a comprehensive parametric
study of a typical industrial building founded on a mat foundation which is allowed to uplift. The
building is inspired from the nuclear industry but entails a structural system simple enough to allow
extrapolating conclusions for civil structures in general. The examined configuration is presented in
Figure 1.
The examined structure is a symmetric embedded reinforced concrete building composed of
peripheral external and internal vertical walls, horizontal slabs and an internal structure modeled as a
single-degree-of-freedom oscillator, attached to the basemat. Soil is modeled as a purely cohesive
homogeneous half-space characterized by uniform undrained shear strength together with curves
for ratio / and corresponding damping ratio versus shear distortion.

Figure 1. Soil-structure configuration examined in the parametric study

The entire configuration is parameterized on the basis of four dimensionless parameters reflecting
aspect and embedment ratio of the building, rigidity contrast between soil and superstructure and
foundation bearing capacity with respect to building mass. The definition of dimensionless parameters
employed in the parametric study is presented in Table 1.
Dimensionless parameters are varied appropriately to give raise to different instances of soil-
structure systems to be analyzed in the parametric study. This is performed by modifying the
geometric and mechanical parameters presented in Figure 1. The total number of examined
configurations is 17. Starting from a base configuration we obtain four families with five
configurations each (the base configuration belongs to all four families) by varying one dimensionless
parameter while the other three are kept constant (to the extent that this is possible since there is

2
 C. T. Chatzigogos et al. 3

dependence between the parameters). The values of dimensionless parameters for the 17 examined
configurations are summarized in Table 2. The configurations are grouped as follows:

- Study of the aspect ratio: configurations 1 to 5

- Study of the embedment ratio: configurations 1 and 6 to 9
- Study of rigidity contrast ratio: configurations 1 and 10 to 13
- Study of bearing pressure ratio: configurations 1 and 14 to 17

Table 1: Definition of dimensionless parameters used in the parametric study

Aspect Embedment Rigidity Bearing pressure

ratio ratio contrast ratio

= = = =

: fundamental circular frequency of building ; fixed-base conditions

: shear wave velocity of soil medium
: total mass of the building

Table 2: Summary of examined soil-structure configurations

BASE Variation of Variation of Variation of Variation of
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
2.0 1.0 1.5 2.5 3.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0

0.2 0.2 0.2 0.2 0.2 0.0 0.1 0.3 0.4 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2

2.5 3.1 2.8 2.4 2.4 2.5 2.5 2.5 2.5 4.0 3.0 1.7 1.3 2.8 2.7 2.5 2.5

4.0 4.2 4.0 4.5 4.6 4.0 4.0 4.0 4.0 3.5 4.1 6.0 7.2 2.0 3.0 5.0 6.0

Each configuration is studied for five levels of increasing seismic intensity (PGA ranging from
0.15g to 0.55g, step 0.1g) using five analysis methods. These range from a conventional linear pseudo-
static analysis method (denoted LP) to a complete nonlinear transient analysis with explicit modeling
of the superstructure and the surrounding soil (denoted NLT). The characteristics of each analysis
method are summarized in Table 3.

Table 3: Synthetic presentation of analysis methods used in the parametric study

Method Loading SSI Nonlinearity Seismic Input Analysis method
Sub-structuring Modal analysis
H: FF*
LP Response spectrum Dynamic impedance None Spectral combination
V: FF
Effective soil properties Linear pseudo-static analysis
Sub-structuring Modal analysis
H: FF
NLP Response spectrum Dynamic impedance Uplift Spectral combination
V: FF
Effective soil properties Nonlinear pseudo-static analysis
Sub-structuring
H: Kinematic int.
Three acceleration Dynamic impedance Linear transient
LT None V: Kinematic int.
time history scenarios Effective soil properties analysis
θ: Kinematic int.
Kinematic interaction
Sub-structuring
H: Kinematic int.
Three acceleration Dynamic impedance Nonlinear transient
NLTP Uplift V: Kinematic int.
time history scenarios Effective soil properties analysis
θ: Kinematic int.
Kinematic interaction
Three acceleration Uplift H: FF deconvolution Nonlinear transient
NLT Direct method
time history scenarios Soil plasticity V: FF ground surface analysis
*FF free field
 For nonlinear methods (NLP, NLTP and NLT) soil-structure interfaces (along the mat foundation
and the embedded vertical walls) are modelled with tensionless node-to-node contact elements
allowing for detachment between the structure and the soil. The basemat-soil interface exhibits an
infinite friction resistance (rough interface) whereas the vertical walls-soil interface exhibits zero
friction resistance (smooth interface).
The modelling principle for the methods that use the substructuring technique (LP, NLP, LT and
NLTP) is presented in Figure 2. The soil domain is replaced by four groups of spring-dashpot elements
(denoted T1 to T4) which aim at reproducing the dynamic impedance of the mat foundation. The
values attributed to the vertical and horizontal stiffness of the springs in each group are given in Table
4. Similar expressions apply for the damping terms.

(a) (b)
Figure 2. Modelling principle: (a) for linear (LP and LT) methods and (b) for nonlinear
substructuring methods (NLP and NLTP)

Table 4: Stiffness terms for the definition of foundation impedance springs

Group (T1) Group (T2) Group (T3) Group (T4)

1
= # = % − 2!" ( ) =0 + =0
!" !

, - ."
, , =0 ,# =0 ,) = ,+ = / − 2!" 0
! ! . -"

! : number of nodes in the mat foundation

!" : number of nodes of vertical wall in contact with the soil
: embedment depth
- : area of the mat foundation
-" : contact area between vertical wall and soil
. : moment of area of mat foundation
." : moment of area of contact area between vertical wall and soil
: horizontal impedance of the mat foundation (real part)
, : vertical impedance of the mat foundation (real part)
: rotational impedance of the mat foundation around axis 1 (real part)
: coupling term (horizontal translation-rotation, real part)

4
 C. T. Chatzigogos et al. 5

Method NLT is a direct method for soil-structure interaction (no substructuring) with explicit
modelling of the superstructure and the surrounding soil. The modelling principle for NLT is presented
in Figure 3.

Figure 3. Modelling principle for method NLT and related mesh

A detailed description of the modelling and calculation hypotheses adopted for each analysis
method can be found in the internal reports of Géodynamique & Structure (2012) and (2013a).
The earthquake excitation in the horizontal direction is represented by the EUR design response
spectrum for medium soils (cf. references – document EUR). In all cases, vertical seismic component
is defined as 2/3 of the horizontal one. In addition, a set of three spectrum-compatible acceleration
time histories has been defined in order to establish the seismic input motion samples to be used in
transient analyses. For convenience, a planar seismic excitation is considered, parallel to symmetry
plane xz as in Figure 1 (i.e. horizontal component parallel to axis y is zero).
All nonlinear seismic analyses are performed after gravity initialization in which permanent and
variable loads in the building are considered. In method NLT, gravity initialization in building and soil
is performed simultaneously.
The implementation of analyses for each configuration (all analysis methods included) is
organized in ten strictly predefined steps that allow controlling calculation procedure with
programming scripts. A summary of the adopted procedure with calculation times required for each
step is presented in Table 5. Calculations have been performed using an Intel-2Quad-CPU@2.83GHz.
Calculation times in Table 5 correspond to sequential execution of analyses (no parallelism).
Parallelism has only been introduced in the treatment of different configurations.

Table 4: Calculation procedure for each configuration – Required calculation times

Calculation time Number of analyses per
Step
per configuration configuration
1. Mesh generation for configuration <1min 1
2. Calculation of effective soil parameters 1h 5 (intensity levels)
3. Deconvolution of seismic motion (for NLT) <1min 5 (intensity levels)
4. Calculation of basemat impedance functions 1h to 3h 5 (intensity levels)
5. Implementation of method LP ~40min 5 (intensity levels)
6. Implementation of method NLP 4h to 8h 5 (intensity levels)
7. Effect of kinematic interaction (for transient methods) ~3h 5 (intensity levels) x 3 (records)
8. Implementation of method LT 4h to 8h 5 (intensity levels) x 3 (records)
9. Implementation of method NLTP 20h to 30h 5 (intensity levels) x 3 (records)
10. Implementation of method NLT 60h to 120h 5 (intensity levels) x 3 (records)
RESULTS

Emphasis has been given on three specific aspects of dynamic response, namely: a) maximum uplift
ratio developed during loading (uplift ratio is defined as the ratio of area of uplifted zone over total
area of mat foundation), b) floor response spectra in positions of interest within the building and c)
resultant internal forces and moments at specific locations of the structural system.
The selected positions for calculation of these quantities are presented in Figure 4. Floor
response spectra are calculated in five positions either in the horizontal direction, the vertical direction
or in both. Resultant vertical internal forces are calculated in three positions lying along external and
internal walls normal to horizontal seismic input motion. Similarly, resultant bending moments are
calculated in three positions lying along external and internal walls parallel to horizontal seismic input
motion. Moreover, obtained results for transferred response spectra are more conveniently presented in
dimensionless form as spectral ratios; to do so, all calculated response spectra are normalized by
division with the corresponding spectrum obtained with method NLT.

Figure 4. Positions for calculation of transferred floor spectra and internal resultant forces

The results obtained through the parametric study are presented in a synthetic form in order to
allow for a quantitative characterization of each analysis method with respect to method NLT, which is
viewed as the reference result. The obtained characterization is performed in terms of various aspects
of the overall response, such as the effect of kinematic interaction, the effect of soil plasticity in
dissipating uplift potential, the gain in passing from pseudo-static to transient analysis, the relocation
of seismic surcharge within the structure when uplift is initiated etc. A synthetic presentation of results
is achieved by constructing multi-axial radar diagrams. In these diagrams, each axis corresponds to a
performance criterion 56 . Selected performance criteria form vector 7 = {56 }. This vector is
represented in each radar diagram as function of analysis method, seismic intensity or some
dimensionless parameter of the study.
In the following, we will restrain presentation of obtained results to the base configuration. This
already allows anticipating the main conclusions of the parametric study. An extensive presentation of
the entire corpus of results can be found in the internal report of Géodynamique & Structure (2013b).

Maximal Uplift ratio

Figure 5 presents the obtained radar diagram for calculated uplift ratio in base configuration.
The diagram possesses five axes, which correspond to the calculated uplift ratio for each seismic
intensity level. Five curves are presented in the diagram, one for each analysis method. Curve color
and type correspond to a particular feature of the analysis methods: blue curves correspond to pseudo-

6
 C. T. Chatzigogos et al. 7

static methods, red curves to transient methods. Continuous lines correspond to linear methods, dotted
lines to nonlinear methods. Green continuous curve finally corresponds to method NLT, the reference
method. Axes represent maximal uplift ratio, which may vary from 0 (no uplift) to 1 (total basemat
detachment).

Figure 5. Radar diagram of uplift ratio in the base configuration

The results in Figure 5 reveal the pronounced conservatism of method NLP, which leads to the
highest uplift ratio for all seismic intensity levels. Moreover, the values increase rapidly to more than
70% for intensities of 0.25g or larger anticipating loss of overall stability from overturning and/or
bearing pressure insufficiency.
Results also confirm the envelope character of pseudo-static methods (linear LP and nonlinear
NLP) with respect to corresponding transient methods (linear LT and partial nonlinear NLTP). For linear
methods in particular, passage from LP to method LT may lead to a reduction in the calculated uplift
ratio of the order of 15%. In what regards transient methods, passage from LT to NLTP has negligible
effect on results for intensities smaller than 0.25g. For larger intensities however the two methods
diverge because in any case, uplift ratio for linear methods cannot exceed 50% (area of springs in
tension is equal to area of springs in compression).
Finally, reference method NLT produces by far the smallest uplift ratios compared to the other
methods. This is due to several factors, such as:

(a) transient character of the method

(b) consideration of soil plasticity in the analyses (soil plasticity tends to flatten the basemat-
soil contact area)
(c) seismic motion attenuation due to soil modelled as elastoplastic medium
(d) reduction of incident motion at the basemat level with respect to ground surface

It should be noted however, that for certain configurations not herein presented, method NLT
may produce uplift ratios larger than linear methods: this may be the case when system characteristics
are such that a resonance may be generated between soil column and the superstructure or when there
is significant permanent rotation accumulation in the basemat-soil interface which tends to accentuate
uplift. Both of these features can only be captured with method NLT.

Floor spectrum ratios

Figures 6 and 7 present radar diagrams related to floor spectrum ratios in the base configuration.
Spectral ratios are calculated in two characteristic frequencies: a) at f = 35Hz (cf. Figure 6) and b) at
the frequency which yields the maximal spectral ratio (cf. Figure 7). The first result highlights
accuracy of different methods in evaluating maximum seismic accelerations at different parts of the
structure. The second reveals safety margins and performance of each method related to the
dimensioning of secondary equipment.
Each diagram possesses seven axes: the first one (axis parallel to +x) represents the calculated
uplift ratio. The following six axes provide calculated spectral ratios in the retained positions within
the building (these positions have been presented in Figure 2). Each radar diagram contains five curves
as in Figure 5. Spectral ratios for method NLT are by definition equal to 1. A separate radar diagram is
provided for each seismic intensity level. It should be noted that no distinction is made between LP and
NLP (pseudo-static methods). Transfer response spectrum calculation in these cases is performed using
code FSG (Igusa & Der Kiureghian, 1995) directly based on modal characteristics of the examined
configuration.

Figure 6. Radar diagrams of spectral ratios @< = 35Hz in the base configuration

Figure 7. Radar diagrams of maximal spectral ratios in the base configuration

Radar diagrams of spectral ratios reveal that, as regards maximum accelerations (cf. Figure 6),
differences for increasing intensity levels within the same analysis method remain relatively minor. In
terms of comparisons among different methods, FSG methodology yields in all cases envelope results
with respect to NLT. However, the same cannot be said for methods LT and NLTP: these methods
produce results that can be either superior with respect to FSG or inferior with respect to NLT: an
uncertainty is thus persistent in characterizing conservatism of partial transient methods for maximum
acceleration evaluation.
Secondly, concerning maximal spectral ratios (cf. Figure 7), it can be readily observed that
variability in terms of intensity level is more significant than in the previous case. It is also shown that
NLT yields in all cases (positions and intensity levels) the most favorable result. Finally, FSG is
generally more conservative than partial transient methods (LT and NLTP) but not in all cases: in
particular spectral ratios at node 4676 (this node corresponds to the SDOF oscillator which models
internal structure) are found to be much larger with LT and NLTP compared to FSG: this suggests that a
critical appreciation of FSG results should always be required, especially for parts of global structural
models that exhibit some sort of singularity.

Comparison of internal forces

The third examined aspect of structural response concerns internal resultant seismic forces and
seismic moments at selected locations. With reference to Figure 2, we calculate

- the seismic vertical force FZ [MN] along sections C1, C3 and C5

- the seismic bending moment MY [MNm] along sections C2, C4 and C6

8
 C. T. Chatzigogos et al. 9

It is stressed that only the seismic component is presented; forces and moments due to gravity
are not included in the presented results. Related radar diagrams are presented in Figure 8. These
diagrams follow the same presentation approach as in Figures 6 and 7; note that resultant seismic
forces are presented in dimensional form but with no sign: this is due to the alternate character of
seismic excitation.

Figure 8. Radar diagrams of internal forces in the base configuration

Figure 8 reveals that the envelope character of simplified methods with respect to NLT is less
evident in what regards evaluation of forces. Factors related to this observation are:

a) the large variability that characterizes nonlinear transient analyses,

b) differences in structural damping models between transient and modal-spectral approaches
c) consideration of static and dynamic earth pressures in NLT, which modify overall stress state

Figure 8 also reveals that increasing seismic intensity leads to a distinct ovalization in the form
of radar diagram for method NLP: this is explained by the redistribution of internal forces occurring
within the structure as uplift ratio increases up to more than 50%, creating force concentrations in the
zone of contact and distress reduction in the uplifted zone. This phenomenon does not take place in
other methods because uplift ratio is by far smaller than in NLP.
It can additionally be noted that method LP is strictly envelope to method LT, proving the
conservative character of the pseudo-static approach and the CQC modal combination method with
respect to transient methods.
Finally, NLTP leads to a moderate distress reduction with respect to LT; this reduction increases
with intensity level and confirms the observation that uplift can globally be seen as a favorable seismic
isolation mechanism relieving demand in the majority of structural elements.

GUIDELINES FOR DESIGN

Upon completion of the parametric study, there has been a systematic effort to reinterpret the most
fundamental conclusions as a set of design guidelines for evaluation of seismic response of structures
on shallow foundation allowed to uplifting. In order to simplify the presentation, we consider that the
superstructure is designed in the linear range and that the only one source of nonlinearity in the system
is located at the foundation level.
The main outcome of the study has been compiled as a design methodology which is
concretized by performing the following steps:

1. Method LP is in all cases the reference method for seismic design. Linear springs connected to
the basemat are used to model soil-structure interaction. These springs can develop both tension
and compression forces. Given the design seismic excitation prescribed for the structure, the
designer uses LP to calculate uplift ratio at the basemat level. Uplift ratio obtained in this initial
calculation is denoted as κ>? .
2. If κ>? < 10%, it is implied that uplift remains limited for the examined configuration/excitation
and has negligible effect on the overall structural response. Structural design everywhere in the
structure (the basemat included) is performed on the basis of internal forces as calculated with
LP. Floor response spectra can be calculated either by the FSG methodology (modal synthesis
techniques) or by means of linear transient analyses.

3. If 10% < κ>? < 30%, it is inferred that moderate to significant uplift takes place but an
estimation using linear methods remains realistic. More specifically, it is suggested to neglect
uplift effects for structural design of all elements except for the basemat. Noting that flexural
curvature is the critical quantity for basemat design, it is understood that consideration of
tension forces in SSI springs is non-conservative for calculation of steel reinforcement in the
basemat. In order to generate a more realistic and conservative force state for basemat design,
the designer is invited to introduce tensionless SSI springs at the basemat level (he thus switches
to method NLP) and to apply the required fraction of the pseudo-static seismic forces in order to
recuperate κ>? as basemat uplift ratio. Note that if the designer applies the entire seismic forces
within NLP he will end up with an unrealistically large uplift ratio. This is why basemat design
in performed using method NLP but only with a reduced fraction of seismic forces. By the way,
the proposed approach is in total agreement with a clear conclusion of the parametric study,
namely that method LP is much more precise than NLP in predicting a realistic uplift ratio with
respect to the reference result, in particular for moderate level of shaking. Except for the
basemat, design of all other structural elements is performed as in step 2.

4. Finally, if κ>? > 30% it is inferred that uplift is significant to the extent that linear methods are
no longer sufficient for quantifying its effects on structural response. The designer is thus
invited to test more elaborate methods for uplift ratio evaluation such as energy methods, LT and
mainly NLTP or NLT (note however that NLP cannot serve at this stage for an improved uplift
ratio evaluation since it leads to exceedingly conservative results). Upon implementation of an
improved uplift ratio calculation, two possibilities can arise:

- Corrected uplift ratio (denoted in this case as κD> ) is shown to be smaller than 30%. This
validates design using linear methods and the designer can implement the procedure as
described in step 3. However, target uplift ratio for implementation of reduced NLP of
step 3 is no longer κ>? but it is now set equal to the corrected value κD> .

- Corrected uplift ratio ED> is shown to remain larger than 30%. In this case it is inferred
that seismic response is dominated by uplift. Consequently, structural design and floor
response spectra calculation must be performed using a nonlinear transient analysis
approach, which correctly accounts for the effects of uplift. Nevertheless, it must be noted
that for the majority of examined configurations in the parametric study, implementation
of method NLT has been sufficient to prove that uplift ratio is actually smaller than 30%,
even in the case of strong and very strong shaking (0.55g). It can thus be expected that for
most cases encountered in practice, linear methods should be sufficient for design as long
as they are eventually enriched with a supplementary (nonlinear/transient) calculation
demonstrating that uplift ratio does not violate the normative limit of 30%.

CONCLUSIONS

The present work has been concerned with a classification of seismic design approaches for structures
that are allowed to uplifting. The problem has been confronted by means of an extensive parametric
study, in which several soil-structure configurations have been examined for increasing seismic
intensities using a variety of analysis methods. Several aspects of the overall response have been
discussed, such as calculated uplift ratio, floor response spectra and most representative internal

10
 C. T. Chatzigogos et al. 11

distress quantities within the structural system. The large corpus of obtained results has been
synthetically presented using radar diagrams that allow for a direct comparison of different analysis
methods with respect to shaking intensities, dimensionless parameters and analysis methods.
Finally, a systematic effort has taken place aiming at a reinterpretation of the obtained
conclusions in the light of establishing a proposal for an integrated seismic design methodology of
buildings in the presence of uplift.
The proposed methodology introduces LP as the basic analysis method and postulates that LP is
sufficient for seismic structural design (with the exception of the basemat) as long as it can be shown
(by any analysis method, even NLT) that maximal uplift ratio is smaller than 30%. In case this
condition is not satisfied, it is inferred that uplift dominates dynamic structural response and thus the
design must be performed with a nonlinear transient method.

REFERENCES

EUR – European Utility Requirements for LWR nuclear power plants.

Géodynamique et Structure (2012). “EDF-SEPTEN – Rédaction RCC-CW – Note
d’hypothèses et de méthodologie pour l’étude des cas supports – Dossier 46-12, report
2, revision A” (in french).
Géodynamique et Structure (2013a). “EDF-SEPTEN – Rédaction RCC-CW – Etude des cas
supports – Présentation des résultats du cas de base – Dossier 46-12, report 6, revision
B” (in french).
Géodynamique et Structure (2013b). “EDF-SEPTEN – Rédaction RCC-CW – Propositions
pour la rédaction de l’annexe sismique – Aspects d’interaction sol-structure – Dossier
46-12, report 8, revision B” (in french).
Igusa, T. and Der Kiureghian, A. (1995). FSG – Floor Spectrum Generator, Last revision
1995.