Bernardini Et Al - ANIDIS - Aggregati

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 12

ANIDIS 2017

PISTOIA

Sensitivity analysis of seismic behaviour of masonry aggregates


Chiara Bernardinia, Sonia Boschia, Martin Empelmannb, Andrea Vignolia
a
Dipartimento di Ingegneria Civile e Ambientale, Via di Santa Marta 3, 50139 Firenze.
b
Institut für Baustoffe, Massivbau und Brandschutz, Beethovenstraße 52, 38106 Braunschweig.

Keywords: seismic vulnerability, masonry aggregates, seismic interactions, sensitivity analysis, linear and non-
linear analysis

ABSTRACT
Masonry load-bearing buildings in aggregate are the most widespread structural typology in the Italian city centers
and they are characterized by a wide structural variety as a result of their historical evolution; indeed they may
differ in geometry, materials, construction techniques and state of preservation. For this reason, their seismic
vulnerability is different compared to that of isolated existing masonry buildings and, additionally, the ordinary
lack of information about the entire aggregate makes their assessment more difficult.
Seismic response of buildings enclosed in aggregate is influenced by the presence of adjacent ones, considering
that they frequently share the same boundary walls. Therefore reciprocal interactions should be taken into account
in seismic analysis in order to consider the “aggregate effect”. Even if several studies have been conducted in
recent years, the main problems still concern the definition and the modality of the reciprocal interactions between
different portions of the same aggregate and the identification of the optimal portion of the aggregate to be
considered for a unitary modelling.
Hence, the aim of this paper is to better understand the interaction effect among adjacent buildings by assessing the
influence of several structural parameters on the seismic global behaviour of an ideal case study, based on the
typical masonry aggregates in the North area of Florence. For this purpose, the results of a sensitivity analysis
performed through linear and non-linear analyses, by considering different boundary configurations, are presented
in terms of shear forces distribution, displacements and capacity curves and extensively discussed.

SG03-403
Sensitivity analysis of seismic behaviour of masonry aggregates
Chiara Bernardinia, Sonia Boschia, Martin Empelmannb, Andrea Vignolia
a
Dipartimento di Ingegneria Civile e Ambientale, Via di Santa Marta 3, 50139 Firenze.
b
Institut für Baustoffe, Massivbau und Brandschutz, Beethovenstraße 52, 38106 Braunschweig.

Keywords: seismic vulnerability, masonry aggregates, seismic interactions, sensitivity analysis, linear and non-
linear analysis

ABSTRACT
Masonry buildings in aggregate are the most widespread structural typology in the Italian city centers and they are
characterized by a wide structural variety; indeed, they may differ in geometry, materials, construction techniques
and state of preservation.
Seismic response of buildings in aggregate is influenced by the presence of adjacent ones, since they frequently
share the same boundary walls. Therefore, the reciprocal interactions should be taken into account, in order to
consider the “aggregate effect”. However, the ordinary lack of information about the entire aggregate makes their
analysis more difficult than the case of isolated buildings.
Even if several studies have been conducted, the main problems still concern the assessment of the reciprocal
interactions between different parts of the same aggregate and the choice of the optimal portion of the aggregate for
a unitary modelling. Hence, the aim of this paper is to better understand the “aggregate effect” by assessing the
influence of the floors stiffness on the seismic global behavior of ideal case studies, based on the typical masonry
aggregates in the North area of Florence. For this purpose, the results of a sensitivity analysis, performed through
linear and nonlinear analyses by considering different boundary configurations, are presented.

During a seismic event, reciprocal interactions


may occur among adjacent buildings within a
1 INTRODUCTION masonry aggregate, that frequently share the same
Italian historical city centres are mainly boundary walls, creating an “aggregate effect”
composed by masonry load-bearing buildings that influence the overall seismic response.
which, over the centuries, were built alongside
the pre-existing ones, leading to the complete
saturation of the free spaces and to the creation of
complex structural systems, identified as masonry
aggregates (Figure 1).
A masonry aggregate can be defined as a not
homogeneous group of structural units, which are
connected by links more or less structurally
effective and that naturally interact under seismic
action. A structural unit (SU) is characterized by
structural continuity along the height or by a Figure 1. Diachronic construction process and building
common manufacturing process and it has a interaction (adapted from Giuffrè, 1989).
unique dynamic behaviour (Circ. Min. 617/2009). In recent years, various methodologies have
Recent seismic events have shown again how been developed in order to reduce the seismic risk
much Italian historical city centres are prone to be of historical structures. Just think, for example, to
damaged; this is due to several factors, as poor the vulnerability assessment methods, among
quality of materials, complex structural which we can mention the vulnerability index
organization, heterogeneities, etc. (Borghini et al. methods, used for a large scale assessment, or the
2011, Menna et al. 2016). capacity spectrum based methods, suitable for
detailed analyses (D’Ayala and Novelli 2010).

SG03-404
As regards the masonry aggregates, the main  Difference among opening areas, that may
difficulties in seismic risk assessment concern influence the distribution of horizontal
how to evaluate reciprocal interactions between actions between the façades.
different part of the same aggregate and the  Position of the SU within the aggregate:
choice of its optimal portion to be considered for an internal SU is constrained by the others
a unitary modelling. In order to achieve these respect to a corner or header building.
goals, understanding which are the features that  Difference in height: for example, a
most influence the masonry aggregates is an building in adjacency with taller buildings
essential task. may suffer minor damage compared to
In this paper, the influence of the floors one in adjacency with buildings
stiffness on the seismic global behaviour of ideal significantly more low, even if the lower
masonry aggregates is investigated. The case construction may be subject to a
study is based on the typical structures in the vulnerability induced by possible collapse
North area of Florence. Different configurations of the higher ones.
have been considered and a sensitivity analysis  Presence of staggered floors, that can
has been performed by applying linear and generate local thrusting forces on
nonlinear analyses. common walls as well as effects of
pounding.
These aspects are representative of the
2 INTERACTION AMONG BUILDINGS IN interaction among buildings and they are
AGGREGATE considered in the vulnerability index method
As a consequence of their historical evolution, forms recently implemented for the evaluation of
masonry aggregates are often characterized by a the vulnerability assessment of historical city
wide structural variety and frequently they centres at large scale (Formisano et al. 2015,
coincide with the urban blocks. Their Vicente 2008, Ferreira et al. 2012).
configuration is also influenced by the Therefore, the knowledge of the surrounding
characteristics of the site and they may be conditions is essential for the assessment of the
classified in simple or complex (Carocci 2006). seismic response of a building within an
As already seen in Caniggia et al. (1979), one of aggregate. However, it is a common practice to
the most common typologies in Italian historical analyse the single SU of interest not considering
centres is the row aggregate, composed by several the adjacent buildings and so not taking into
structural units (SUs) organized in line. account the “aggregate effect”; this is mainly due
The analysis of these structures could be very to the lack of information on the aggregate and to
difficult, due to uncertainties related to the the impossibility of accessing the other buildings.
knowledge of construction techniques, the On the other side, a detailed analysis of a large
mechanical characteristics of materials and the part of the aggregate should be performed, but
effective structural organization. As suggested by this would require a large amount of data. As a
Carocci et al. (2007), the “critical reading” consequence, some approximations should be
procedure should be applied. This is a systematic introduced. As suggested in ReLUIS (2010), in
analysis of the constructive technique and of the case of particularly complex and extensive
evolutionary phases, in order to investigate the aggregate, a subdivision in Minimum
mechanical characteristics and the vulnerability Intervention Unit (MIU) can be applied. The MIU
factors that may influence the seismic response of is a portion of the aggregate, composed by one or
masonry aggregates. more homogeneous SUs, whose repair or
Several aspects could influence the seismic reconstruction must be unitedly planned. The
behaviour of a SU in different ways; for example, optimal choice of the MIU has to minimize the
the following features should be considered in the reciprocal seismic interactions and it can be
analysis . defined by finding structural discontinuities. In
 Heterogeneities between adjacent the analysis, the Minimum Unit of Analysis
buildings: in presence of a more rigid (MUA) should be considered, in order to assess
building in the centre of the aggregate, the possible interaction effects; this is defined as the
side buildings can withstand higher level portion of the aggregate, generally larger than the
of excitation, probably because the whole MIU in question, including some adjacent SUs.
aggregate is reinforced (Ulrich et al. As deepened in Amadio et al. (2011), the
2012). seismic response of extended and regular row
aggregates, composed by SUs characterized by
constant stiffness, is uniform and translational for

SG03-405
transverse earthquake (Figure 2a); it is therefore analysis of the single SU and that depends on the
possible to study separately each SU, but stiffness ratio between adjacent SUs.
considering the portion of mass deriving from the Senaldi et al. (2010) have investigated the
adjacent units, that weighs on the common walls. effects of the length of the row aggregates and of
On the contrary, in presence of a longitudinal the presence of flexible floors, by applying
earthquake, each SU is influenced by the nonlinear dynamic analyses on idealised
presence of the adjacent ones; indeed, a resistant configurations. In relation to the transversal
arch mechanism could be generated, producing a response, the external units show a higher
significant capacity increase of all SUs, with the deformation compared to the internal ones.
exception of the external ones (Figure 2b). For A simplified approach to assess the seismic
these, the study of the single SU is therefore capacity of a building in aggregate is to analyse it
generally in favour of safety. When the aggregate both as isolated and inserted in aggregate.
is characterized by plan irregularity or by SUs Pujades et al. (2010) have analysed the seismic
with different stiffness, its seismic response is behaviour of typical structures in Barcelona,
more complex, because it could be influenced by evaluated as individual buildings and as an
translational and torsional effects. aggregate, through a capacity spectrum based
method. The authors have noticed that, in
longitudinal direction (direction of compound),
the aggregate does not improve significantly the
seismic behaviour of the isolated buildings; rather
the expected damage in the aggregate is a sort of
average of the damage expected for the individual
Earthquake in transverse direction
buildings. Instead, in transverse direction, the
expected damage of the aggregate is similar to the
lower value obtained for the individual buildings,
maybe because the confinement may produce a
small improvement of their individual behaviour.
Applying the same approach, Ferrito et al.
(2016) have assessed the seismic vulnerability of
Earthquake in longitudinal direction
old mixed masonry–RC buildings in Lisbon using
Figure 2. Seismic response of a row aggregate in presence linear dynamic analysis and nonlinear static
of transversal (a) and longitudinal (b) earthquake (adapted
from Amadio et al., 2011)
analysis. The authors have observed that, when
the building is considered within the compound,
In ReLUIS (2010), a mechanical model for the its seismic vulnerability is reduced.
global seismic analysis of buildings in aggregate In Ulrich et al. (2015), the effects of
is introduced, based on the linear static analysis. aggregation on the seismic response are assessed
It allows a quantitative assessment of the collapse by subjecting isolated and aggregated ideal
acceleration of the unit of analysis, assuming that buildings to both in-plane and transverse loads.
all vertical elements simultaneously achieve their The buildings at the extremities of the aggregate
maximum resistance. In presence of flexible are subjected to asymmetric loads because of the
floors, each wall can be analysed as an gravitational loads transferred by the adjacent
independent structure subject to vertical loads of structures; moreover, as already seen in Amadio
competence and to the seismic action in parallel et al (2011), they are more prone to damage,
direction. In presence of rigid floors, the SU is because, contrary to the buildings inside the
analysed by considering the simultaneous aggregate, they are not efficiently braced by the
contribution of the different resistant walls in the neighbouring structures.
examined. From the abovementioned contributions, it
Similarly, Amadio et al. (2011) have proposed appears clear that the hypothesis of isolated
a simplified method for linear static analysis of building is not always in favour of safety. On the
masonry buildings in aggregate, based on the contrary, taking into account the group
subdivision into cells, that identifies two limit mechanism is important both for correctly
situations: the overall response of the aggregate estimate seismic actions and for optimizing
and the response of the single cell; stiffness is interventions.
calculated by considering the stiffness per unit of In order to better understand the interaction
surface of each cell. The authors have noticed that effect among adjacent buildings, ideal aggregates
there is a redistribution of forces by considering have been analysed by assessing the influence of
the group mechanism, which is lost in the the floors stiffness on their seismic global

SG03-406
behaviour. A sensitivity analysis has been  1 SU-aggregate,
performed through linear and nonlinear analyses,  3 SUs-aggregate,
by considering different boundary configurations.  5 SUs-aggregate.
Floors have been modelled as rigid or flexible. Each SU has plan dimensions of 4.80 x 9.60 m
Rigid floors have been considered as slabs that and 2 storeys of height of 3 m. It has 2 rooms per
allow the distribution of both vertical and storey and there are 3 alignments in X direction
horizontal loads to the vertical structures, (x1, x2, x3) and 2 alignments in Y direction (y1,
guaranteeing a global response of the building. y2). The openings in X direction are regular and
Flexible floors are characterized by in-plane aligned; in the Y direction, walls have no
flexible elements that only share vertical loads to openings.
the walls; in this case, these are mainly stressed The characteristics of the basic SU are shown
out of their plain. in Figure 4, while the plans of the 3 SUs and 5
The characteristics of the considered ideal SUs aggregates are shown in Figure 5.
cases study are described in the following (b)
paragraph. (a)
y1 y2 120
300
x3
60 120 120 120 60
120
3 IDEAL CASES STUDY
480 120
300
240
3.1 Row aggregates in Florence 120

The ideal aggregates analysed in this paper are x2 960 60 120 120
480
120 60

based on the properties of the typical aggregates 60 120 300

of the North area of Florence (Figure 3),


considered because of their recurring 480
180
300

characteristics and their spread in Italian territory.


x1
y 60 120 120 120 60 300
480 240

x
60 120 300
480
(c)
Figure 4. 1 SU plan (a), front and back façade (b) and
internal wall (c).
(a)
y1 y2 y3 y4
x3

480

x2 960

Figure 3. Typical row aggregates of the North area of 480

Florence.
x1
These are row aggregates built at the 480 480
1440
480
(b)
beginning of the XX century and composed by 2- x3
y1 y2 y3 y4 y5 y6

3 storeys buildings. Each building was built


adjacent to the neighbour one and, consequently, 480

they share the same boundary walls.


Each building is usually characterized by a x2 960

width of 4 - 6 m and a depth of 8 - 12 m, for a


480
plan dimension ratio of ½. The geometry of the
openings is quite regular and they are aligned on x1

the main façade.


y 480 480 480 480 480
2400
x

Figure 5. 3SUs-aggregate plan (a) and 5SUs-aggregate plan


3.2 Ideal models (b).

Three main models have been considered, by The walls have a thickness of 40 cm and they
varying the number of the structural units along are characterized by irregular stone masonry,
the longitudinal axis (X direction): corresponding to the 1st category of the

SG03-407
classification in Tab. C8A.2.1 of Circ. Min. 2009. 4 RESULTS
The considered mechanical characteristics (Table
1) are the average values. In the hypothesis of not
4.1 Linear analysis
elastic conditions, the values of E and G are
halved. In order to perform linear analyses, the
Table 1. Mechanical characteristics of masonry.
aggregates have been modelled by using a FEM
software. Shell elements have been used to model
w E G fm W0 the masonry walls. Rigid floors (R) are modelled
[kN/m3] [N/mm2] [N/mm2] [N/cm2] [N/cm2]
with shell–thick elements to which the
19 435 145 140 2.60
mechanical characteristics of the concrete C25/30
are assigned (Table 5). Flexible floors (F) are not
The floors are organized in parallel to the main modelled but there are holes in their
front. As previously said, they are modelled as correspondence, in order to not considering their
rigid (R) or flexible (F); the applied loads are contribution of stiffness and connection. The
constant (Table 2). foundations are modelled as fully restrained.
Table 2. Loads on floors.
Table 4. Mechanical characteristics of concrete C25/30.
Value
Loads E [N/mm2] G [N/mm2] fck [N/cm2]
[kN/m2]
31476 13115 25
Permanent G1 2.50
Non-permanent G2 2.50 4.1.1 Modal analysis
Variable Q 2.00
The changes in periods and modal shapes have
been evaluated for all the models.
In the sensitivity analysis, each SU is
In presence of rigid floors, the aggregates
characterized by the same floor type at the 2
show a global behaviour: the 1st mode is
storeys. Therefore, the basic model (1SU-
translational in X direction for all the three
aggregate) is made of one SU with rigid (R) or
aggregates. In presence of flexible floors, walls
flexible (F) floors. Different boundary
are subjected to local mechanisms (out of their
configurations have been considered, by
plane), with a consequent reduction in
aggregating the basic model (R or F) in different
participating mass.
layouts. In Table 3 are summarized all the cases
In Figure 6, the periods of the first 3 modes are
that have been considered in the modelling.
compared. As we can observed, in case of rigid
Table 3: Aggregate models. floors (shades of red), the periods are the same
1 SU-aggregate for the three aggregates. Considering flexible
Floor type floors (shades of violet), in general the periods
Model
1° SU are higher; the 1 SU-aggregate has lower periods,
R Rigid while the 5SUs-aggregate shows the highest ones.
F Flexible In case of models with mixed floors, the 3SUs-
aggregate is more rigid compared to the 5SUs-
3 SUs-aggregate
aggregate.
Floor type
Model
1° SU 2° SU 3° SU 0.75 R_1SU

R Rigid Rigid Rigid R_3SU


F Flexible Flexible Flexible R_5SU
0.60
RFF Rigid Flexible Flexible F_1SU
FRF Flexible Rigid Flexible F_3SU
Period [s]

FRR Flexible Rigid Rigid 0.45 F_5SU


RFF
5 SUs-aggregate RRFFF
0.30
Floor type FRF
Model
1° SU 2° SU 3° SU 4° SU 5° SU FFRFF
R Rigid Rigid Rigid Rigid Rigid 0.15 FRR
1 2 3
F Flexible Flexible Flexible Flexible Flexible Number of modes FRRRF
RRFFF Rigid Rigid Flexible Flexible Flexible Figure 6. Comparison of the modal periods.
FFRFF Flexible Flexible Rigid Flexible Flexible
FRRRF Flexible Rigid Rigid Rigid Flexible
4.1.2 Dynamic linear analysis
In order to analyse the aggregate effect as a
function of the floors stiffness, the results of a
dynamic linear analysis have been evaluated in

SG03-408
term of shear percentage distribution in the + Ex
analysis direction on the walls and of shear trend 100%
on each SU. 90%
The response spectrum has been defined

Shear Force Percentage


80%
Sum 81% Sum 19%
according to the NTC 2008, considering an 70%
Sum 66% Sum 34%
60%
ordinary building and II use class. The site is in 50%
Florence (Tuscany), the foundation soil type is A 40%
and the topographic category is T1. For this 30%
study, an elastic response spectrum has been 20%
considered. 10%
0%
x1 x2 x3 y1 y2
The next figures (Figure 7, Figure 8, Figure 9) Rigid Flexible
summarize the percentage distribution of the
shear force in the analysis direction on each
+ Ey
alignment for all the models, considering 100%
separately the two directions of analysis. 90%
For all the aggregates, referring to the analysis

Shear Force Percentage


80%
Sum 4% Sum 96%
in X direction, the X-alignments are the most 70%
Sum 8% Sum 92%
60%
stressed; they absorb more than 80% of the total
50%
shear in presence of rigid floors and more than 40%
60% in presence of flexible floors. As a 30%
consequence, we can observe that a reduction of 20%
the floors stiffness leads to an increase of shear 10%
on the Y-alignments, being more stressed out-of- 0%
x1 x2 x3 y1 y2
their plain, with a consequent reduction on the X-
Rigid Flexible
alignments. Moreover, the internal wall x2 is
generally more stressed than the external ones, Figure 7. Shear percentage distribution on walls for the
1SU-aggregate for both directions of linear analysis.
since it is more rigid than the others, given the
inferior number of openings.
Considering the analysis in Y direction, more 100%
+ Ex
than 90% of the shear is on the Y-alignments, 90% Sum 86% Sum 14%
regardless of the type of floors.
Shear Force Percentage

80% Sum 65% Sum 35%


For the 1SU-aggregate, the Y-alignments are 70% Sum 80%
Sum 83%
Sum 20%
Sum 17%
equally stressed, despite the floors type, for both 60%
Sum 85% Sum 15%
50%
directions of analysis, as shown in Figure 7. 40%
With refers to the 3SUs-aggregate (Figure 8), 30%
the Y-walls are equally stressed in presence of 20%
rigid floors; when floors are flexible, the shear 10%
increases on the internal walls, since they have 0%
x1 x2 x3 y1 y2 y3 y4
higher vertical loads and they undergo major
deformations and, consequently, the external ones R F RFF FRF FRR

are discharged. So, the presence of different type


of floors among adjacent SUs leads to a different 100%
+ Ey
distribution of the shear on the walls in Y 90% Sum 6% Sum 94%
direction; in particular, with refers to the other
Shear Force Percentage

80% Sum 10% Sum 90%


models, we can observe that the external walls are 70% Sum 7% Sum 93%
Sum 9% Sum 91%
more stressed in presence of rigid floors, while 60%
Sum 6% Sum 94%
50%
the internal ones are more stressed in presence of
40%
flexible floors. 30%
Considering the 5SUs-aggregate (Figure 9), 20%
we have to observe that, for the Y analysis in 10%
presence of rigid floors, the Y-walls are not 0%
x1 x2 x3 y1 y2 y3 y4
equally stressed as previously seen, but the latter
ones are the most stressed (12% on y1, 18% on R F RFF FRF FRR
y6). Anyway, the shear percentage distribution on Figure 8. Shear percentage distribution on walls for the
the walls follows the same trend noticed for the 3SUs-aggregate for both directions of linear analysis.
3SUs-aggregate.

SG03-409
+ Ex
+ Ex 600
100% 1°SU 2°SU 3°SU 4°SU 5°SU
90% Sum 87% Sum 13% 500
Shear Force Percentage

80% Sum 67% Sum 33%


70% Sum 80% Sum 20% 400

60% Sum 78% Sum 22%

V [kN]
Sum 84% Sum 16% 300
50%
40% 200
30%
20% 100

10%
0
0% 0.0 1.6 3.2 4.8 6.4 8.0 9.6 11.2 12.8 14.4 16.0 17.6 19.2 20.8 22.4 24.0
x1 x2 x3 y1 y2 y3 y4 y5 y6 x [m]
R_5SU R_1SU F_5SU F_1SU
R F RRFFF FFRFF FFRRR RRFFF FFRFF FRRRF

+ Ey
+ Ey 600
100% 1°SU 2°SU 3°SU 4°SU 5°SU
90% Sum 6% Sum 94% 500
Shear Force Percentage

80% Sum 10% Sum 90%


70% Sum 7% Sum 93% 400

60% Sum 9% Sum 91%

V [kN]
Sum 7% Sum 93% 300
50%
40% 200
30%
20% 100

10%
0
0% 0.0 1.6 3.2 4.8 6.4 8.0 9.6 11.2 12.8 14.4 16.0 17.6 19.2 20.8 22.4 24.0
x1 x2 x3 y1 y2 y3 y4 y5 y6 x [m]
R_5SU R_1SU F_5SU F_1SU
R F RRFFF FFRFF FFRRR RRFFF FFRFF FRRRF
Figure 9. Shear percentage distribution on walls for the Figure 11. Shear force on each SU for the 5SUs-aggregate
5SUs-aggregate for both directions of linear analysis. for both directions of linear analysis.

+ Ex
Additionally, the trend of the shear force in
600
1°SU 2°SU 3°SU
each SU is shown both for the 3SUs-aggregate
500
(Figure 10) and for the 5SUs-aggregate (Figure
11) as a function of the floors stiffness.
400
It’s easy to note that analysing a SU as an
isolated building leads to an overestimation of the
V [kN]

300
shear stresses, regardless of the type of floors;
200
this may be unfavourable in order to optimize the
100 interventions. Moreover, the shear on each SU is
0
higher in presence of rigid floors.
0.0 1.2 2.4 3.6 4.8 6.0 7.2 8.4 9.6 10.8 12.0 13.2 14.4 With refers to the 3SUs-aggregate, in X
x [m]
direction (Figure 10), both for R and F model,
R_3SU R_1SU F_3SU F_1SU
RFF FRF FRR there is a slightly upward shear trend from the
first to the last SU. Considering the other models,
+ Ey
600
1°SU 2°SU 3°SU
we can observe that the shear is higher in
correspondence of the SUs with rigid floors, for
500
which stresses are almost twice. For example, the
400 biggest difference among the SUs can be noticed
in the models FRR (3rd SU the most stressed) and
V [kN]

300
RFF (1st SU the most stressed). So it’s possible to
200
observe that, when two buildings of different
100 stiffness are in adjacency, the rigid one is more
stressed, while the other is discharged.
0
0.0 1.2 2.4 3.6 4.8 6.0 7.2 8.4 9.6 10.8 12.0 13.2 14.4 Referring to the Y direction (Figure 10), the
x [m] shear is higher for the external SUs; for equal
R_3SU R_1SU F_3SU F_1SU
RFF FRF FRR
stiffness of the walls in this direction, this is due
Figure 10. Shear force on each SU for the 3SUs-aggregate to their higher vertical loads of competence. This
for both directions of linear analysis. difference is grater in the R model, while it is
almost nothing in the F model; for the other

SG03-410
models, it is highlighted as the number of SUs for them are similar, in the following only the
with rigid floors increases. uniform one is considered.
Also for the 5SUs-aggregate the shear trend in
X direction is increasing from the first to the last Similarly to the case of linear analysis, in the
SU (Figure 11); the increase of the shear on the following figures (Figure 13, Figure 14, Figure
last SUs is higher in the F model. With regards to 15) the percentage distribution of the shear force
the other models, as previously noticed, the shear in the analysis direction on each alignment is
is higher in correspondence of the SUs with rigid shown for all the models, considering separately
floors. the two directions of analysis. Since local
In Y direction (Figure 11) the shear mechanisms of collapse are restricted, the walls
distribution is more influenced by the vertical are stressed in their plain and the seismic action is
loads of the SUs than the floors stiffness; indeed, in the analysis direction.
also in this case, the highest stresses are in As previously seen, in X direction the wall x2
correspondence of the external SUs. is generally more stressed than the external ones;
but in this case, the difference among internal and
Besides, in order to evaluate how the floors external X-walls is greater in presence of flexible
stiffness influences the response of the walls of floors and it decreases for the other models.
each SU, in Figure 12 the shear trend on the Indeed, the response for this modelling is less
alignment x1 is shown as an example. This aspect affected by the stiffness variation of the walls in
will be deepened in the future works. their plane, but it is more affected by their
The shear trend is influenced by the influence area.
distribution of the openings. Both for R and F
model, the trend is stable among the SUs and the + Ex
100%
shear reaches higher values in presence of rigid 90%
floors. For the combined models, the stresses
Shear Force Percentage

80%
increase in the SU with rigid floors. This is in 70% Sum 100% Sum 0%
Sum 100% Sum 0%
accordance with the previous results. 60%
50%
+ Ex 40%
120
1°SU 2°SU 3°SU 30%
100 20%
10%
80
0%
Vx [kN]

x1 x2 x3 y1 y2
60
R F
40

20 + Ey
100%
0
0.0 1.2 2.4 3.6 4.8 6.0 7.2 8.4 9.6 10.8 12.0 13.2 14.4 90%
Shear Force Percentage

x1 [m] 80%
R R_1SU F F_1SU 70% Sum 0% Sum 100%
RFF FRF FRR Sum 0% Sum 100%
60%
Figure 12. In-plane shear trend on alignment x1. 50%
40%
30%
4.2 Nonlinear analysis 20%
10%
The same models previously introduced have 0%
been analysed by performing pushover analysis, x1 x2 x3 y1 y2
by using an equivalent frame model. The floors R F
have been modelled as orthotropic membrane Figure 13. Shear percentage distribution on walls for the
finite elements of equivalent thickness; the rigid 1SU-aggregate for both directions of nonlinear analysis.
floors are characterized by an infinite stiffness,
while the flexible floors have been modelled as In Y direction, similarly to the linear case, the
timber elements characterized by an equivalent external walls are more stressed in presence of
modulus of elasticity E1,eq= 17187 MPa in the rigid floors, while the internal ones are more
warping direction. stressed in presence of flexible floors. In Figure
Two distributions of lateral loads have been 14 we can observe that for the 3SUs-aggregate
considered, uniform (proportional to the mass) the shear distribution is symmetrical, while for
and pseudo-triangular (proportional to the 1st the 5SUs-aggregate (Figure 15) the y5 wall is the
mode), without eccentricities. Since the results

SG03-411
most stressed (22%) and y1 is the less stressed 600
+ Ex_mass

(13%), as seen in the linear case. 1°SU 2°SU 3°SU

500

+ Ex 400
100%
90%

V [kN]
Sum 100% Sum 0% 300
Shear Force Percentage

80% Sum 100% Sum 0%


70% Sum 100% Sum 0% 200
Sum 100% Sum 0%
60%
Sum 100% Sum 0%
50% 100

40%
0
30% 0.0 1.2 2.4 3.6 4.8 6.0 7.2 8.4 9.6 10.8 12.0 13.2 14.4
20% x [m]
10% 3SU_R 1SU_R 3SU_F 1SU_F
0% RFF FRF FRR
x1 x2 x3 y1 y2 y3 y4
+ Ey_mass
R F RFF FRF FRR 600
1°SU 2°SU 3°SU
550

+ Ey 500
100%
450
90% Sum 0% Sum 100%

V [kN]
Shear Force Percentage

80% Sum 0% Sum 100% 400


70% Sum 0% Sum 100%
350
Sum 0% Sum 100%
60%
Sum 0% Sum 100% 300
50%
40% 250

30% 200
20% 0.0 1.2 2.4 3.6 4.8 6.0 7.2 8.4 9.6 10.8 12.0 13.2 14.4
x [m]
10%
0% 3SU_R 1SU_R 3SU_F 1SU_F
RFF FRF FRR
x1 x2 x3 y1 y2 y3 y4
Figure 16. Shear force of each SU for the 3SUs-aggregate
R F RFF FRF FRR for both directions of nonlinear analysis.
Figure 14. Shear percentage distribution on walls for the
3SUs-aggregate for both directions of nonlinear analysis. + Ex_mass
600
1°SU 2°SU 3°SU 4°SU 5°SU

+ Ex 500
100%
90% 400
Sum 100% Sum 0%
Shear Force Percentage

80% Sum 100% Sum 0%


V [kN]

300
70% Sum 100% Sum 0%
60% Sum 100% Sum 0%
Sum 100% 200
Sum 0%
50%
40% 100
30%
20% 0
0.0 1.6 3.2 4.8 6.4 8.0 9.6 11.2 12.8 14.4 16.0 17.6 19.2 20.8 22.4 24.0
10% x [m]
0% 5SU_R 1SU_R 5SU_F 1SU_F
x1 x2 x3 y1 y2 y3 y4 y5 y6 RRFFF FFRFF FRRRF

R F RRFFF FFRFF FRRRF + Ey_mass


600
1°SU 2°SU 3°SU 4°SU 5°SU
+ Ey 500
100%
90% Sum 0% Sum 100% 400
Shear Force Percentage

80% Sum 0% Sum 100%


V [kN]

70% Sum 0% Sum 100% 300


60% Sum 0% Sum 100%
Sum 0% Sum 100% 200
50%
40%
100
30%
20% 0
10% 0.0 1.6 3.2 4.8 6.4 8.0 9.6 11.2 12.8 14.4 16.0 17.6 19.2 20.8 22.4 24.0
x [m]
0%
x1 x2 x3 y1 y2 y3 y4 y5 y6 5SU_R 1SU_R 5SU_F 1SU_F
RRFFF FFRFF FRRRF
R F RRFFF FFRFF FRRRF Figure 17. Shear force of each SU for the 5SUs-aggregate
Figure 15. Shear percentage distribution on walls for the for both directions of nonlinear analysis.
3SUs-aggregate for both directions of nonlinear analysis.

SG03-412
Furthermore, also for the nonlinear analysis, With refers to the percentage distribution on
the trend of the shear force on each SU is shown each alignment, the shear is mainly distributed
both for the 3SUs-aggregate (Figure 16) and for among the walls arranged longitudinally to the
the 5SUs-aggregate (Figure 17) as a function of analysis direction. A reduction of the floors
the floors stiffness. stiffness leads to an increase of shear on the
For both the aggregates and both analysis transverse alignments, since they are more
directions, considering the R models, when a SU stressed out-of-their plain. In X direction, the
is analysed as an isolated building, the shear internal wall is the most stressed, especially in
stresses are overestimated. When the floors are presence of rigid floors in the linear case and of
flexible, the shear is overestimated in Y direction, flexible ones in the nonlinear case. This
but it is underestimated in X direction. difference between the two approaches is mainly
Moreover, in X direction, the SUs are equally due to the different type of modelling; in
stressed, despite the floors type. It’s possible to particular, the equivalent frame model neglects
note that, the shear is higher for the F model and the out of plane contribution of the walls. In Y
it is lower for the R model. With refers to the direction, for both modelling, the compound of
other models, the results are affected by the floors SUs characterized by different floors type leads to
type of the 1st SU that undergoes to the seismic a different shear distribution on the Y-walls; in
action. Indeed, for the 3SUs-aggregate, the F, particular, when the floors are rigid, the external
FRF and FRR models are equivalent (Figure 16). walls are the most stressed, while the shear
The same trend can be noticed in the 5SUs- increases on the internal ones in presence of
aggregate (Figure 17); indeed, in the FFRFF flexible floors.
model the 1st SU is the most stressed. The shear trend on each SU shows that,
In Y direction it’s possible to observe a considering the longitudinal direction of the
difference in shear among adjacent SUs. aggregate, in presence of two adjacent buildings
Considering the 3SUs-aggregate (Figure 16), the of different stiffness, the more rigid one is the
trend is the same noticed for the linear case: the most stressed, while the other one is discharged;
shear is higher for the external SUs, due to their this trend is more accentuated in the linear case.
higher vertical loads of competence. This In the transverse direction, the shear distribution
difference is accentuated for the R model. is more influenced by the amount of vertical
loads on the walls than by the floors stiffness; in
The shear trend changes a lot for the 5SUs-
particular, it is higher for the external SUs.
aggregate (Figure 17). Unlike the linear case, an
upward trend from the 1st to the last SU may be
noticed and the central SUs are equally stressed.
REFERENCES
5 CONCLUSIONS Amadio, C., Rinaldin, G., Puppin, A., Camillo, M., 2011.
Analisi semplificata della vulnerabilità sismica di un
When assessing the seismic response of aggregato edilizio in muratura: il complesso denominato
masonry buildings in aggregate, reciprocal "Vaticano" (Trieste). XIV Convegno ANIDIS, Bari.
Borghini, A., Del Monte, E., Ortolani, B., Vignoli, A.,
interactions should be evaluated, in order to 2011. Studio degli effetti del sisma del 06/04/2009 sulla
consider the “aggregate effect” and to assess the frazione di Castelnuovo, comune di San Pio delle
effective seismic actions, with the purpose of Camere (Aq). Atti del XIV Convegno ANIDIS, Bari.
optimizing interventions. Caniggia, G., Maffei, G.L., 1979. Lettura dell’edilizia di
In this paper the influence of the floors base. Ed. Marsilio, Padova, Italy.
stiffness on the seismic global behaviour of ideal Carocci, C., 2006. Classificazione degli aggregati. ReLUIS
2005-2008 Research Project No. 1: Assessment and
masonry aggregates is investigated. For this reduction of the vulnerability of masonry buildings.
purpose, the results of a sensitivity analysis Carocci, C., Tocci, C., 2007. Sicurezza sismica degli
performed through linear and nonlinear analyses, aggregati edilizi storici: alcuni casi di studio. XII
by considering different boundary configurations, Convegno ANIDIS, Pisa.
are presented in terms of shear percentage Circolare Ministeriale n.° 617 del 02/02/2009. Istruzioni per
l’applicazione delle “Nuove Norme Tecniche per le
distribution on each alignment and of shear trend Costruzioni” di cui al D.M. 14/01/2008.
on each SU. D’Ayala, D., Novelli, V. (2010). Deliverable 8. Review and
From the results, it is possible to observe that validation of existing vulnerabilty displacement-based
analysing a SU as an isolated building leads in models. Perpetuate Project: Performance-based
most cases to an overestimation of the shear approach to earthquake protection of cultural heritage
stresses, regardless of the floors type. in European and Mediterranean Countries.

SG03-413
Ferreira, T. M., Vicente, R., Varum, H., 2012. Vulnerability
assessment of building aggregates: a macroseismic
approach. 15 WCEE. Lisboa.
Ferrito, T., Milosevic, J., Bento, R., 2016. Seismic
vulnerability assessment of a mixed masonry–RC
building aggregate by linear and nonlinear analyses.
Bulletin of Earthquake Engineering, 14(8), 2299-2327.
Formisano, A., Florio, G., Landolfo, R., & Mazzolani, F.,
2015. Numerical calibration of an easy method for
seismic behaviour assessment on large scale of masonry
building aggregates. Advances in Engineering Software,
80, 116–138.
Giuffrè, A., 1989. Mechanics of historical masonry and
strengthening criteria. XV Regional Seminar on
Earthquake Engineering, Ravello.
Menna, C., Frascadore, R., Moroni, C., Lignola, G.P., De
Martino, G., Salzano, A., Di Ludovico, M., Prota, A.,
Manfredi, G., Cosenza, E. ReLUIS. Rapporto
fotografico relativo ai danni subiti da alcuni edifici a
seguito del sisma del Centro Italia del 2016. NTC 2008.
D.M. del Ministero delle Infrastrutture e dei Trasporti
del 14/01/2008. Nuove Norme Tecniche per le
Costruzioni. G.U. n. 29 del 04.02.2008, S.O. n. 30.
Pujades, L. G., Barbat, A. H., González-Drigo, R., Avila, J.,
Lagomarsino, S., 2012. Seismic performance of a block
of buildings representative of the typical construction in
the Eixample district in Barcelona (Spain). Bulletin of
Earthquake Engineering, 10, 331–349.
ReLUIS, 2010. Linee guida per il rilievo, l'analisi ed il
progetto di interventi di riparazione e consolidamento
sismico di edifici in muratura in aggregato.
Senaldi, I., Mageners, G., Penna, A., 2010. Numerical
Investigations on the Seismic Response of Masonry
Building Aggregates. Advanced Materials Research,
Vols 133-134, 715-720.
Ulrich, T., Gehl, P., Negulescu, C., Foerster, E., 2012.
Seismic vulnerability assessment of masonry building
aggregates. 15 WCEE, Lisboa.
Ulrich, T., Negulescu, C., Ducellier, A., 2015. Using the
discrete element method to assess the seismic
vulnerability of aggregated masonry buildings. Bulletin
of Earthquake Engineering, 13, 3135-3150.
Vicente, R. (2008). Estratégias e metodologias para
intervenções de reabilitação urbana. Ph.D. Thesis.
Universidade de Aveiro.

SG03-414

You might also like