Conference PaperPDF Available

Fast generative tool for masonry structures geometries

Authors:

Abstract and Figures

Modelling masonry bond pattern is still challenging for the scientific community. Though advanced Laser Scanning methods are available and allow to extract blocks sizes and shapes of actual masonry structures, they are up to now very time-consuming and complex to set up. Therefore, modelling masonry as an ideal and regular assemblage of regular units is still very common in the scientific field. This paper presents a generative algorithm for masonry specimens built with a single-leaf cond pattern. It is based on C# programming under the environment offered by Rhinoceros (+ Grasshopper). Five components have been constructed (wall, corner, T and cross-connections, and opening). They can be assembled, up to infinity, to build complex masonry specimens. Moreover, they are all parametrised to account for every wish of the modeller. The global methodology is found highly time-efficient, with the creation of an initial geometry composed of 5-10 components requiring around 10 minutes and, while the update due to a parameter variation is done in less than one second. The paper finally discusses the next developments of the promising generative algorithm.
Content may be subject to copyright.
14
TH
C
ANADIAN
M
ASONRY
S
YMPOSIUM
M
ONTREAL
,
C
ANADA
M
AY
16
TH
M
AY
19
TH
,
2021
F
AST GENERATIVE TOOL FOR MASONRY STRUCTURES GEOMETRIES
Savalle, Nathanaël
1
; Mousavian, Elham
2
; Colombo, Carla
3
and Lourenço, Paulo B.
4
ABSTRACT
Modelling masonry bond pattern is still challenging for the scientific community. Though
advanced Laser Scanning methods are available and allow to extract blocks sizes and shapes of
actual masonry structures, they are up to now very time-consuming and complex to set up.
Therefore, modelling masonry as an ideal and regular assemblage of regular units is still very
common in the scientific field. This paper presents a generative algorithm for masonry specimens
built with a single-leaf cond pattern. It is based on C# programming under the environment offered
by Rhinoceros (+ Grasshopper). Five components have been constructed (wall, corner, T and
cross-connections, and opening). They can be assembled, up to infinity, to build complex masonry
specimens. Moreover, they are all parametrised to account for every wish of the modeller. The
global methodology is found highly time-efficient, with the creation of an initial geometry
composed of 5 10 components requiring around 10 minutes and, while the update due to a
parameter variation is done in less than one second. The paper finally discusses the next
developments of the promising generative algorithm.
KEYWORDS: Masonry, Parametric modelling, C# coding, single-leaf, Grasshopper
1
PhD, University of Minho, ISISE, Department of Civil Engineering, Guimarães, Portugal, n.savalle@civil.uminho.pt
2
PhD, Department of Structures for Engineering and Architecture, University of Naples Federico II, Italy,
elham.mousavian@unina.it
3
PhD Student, University of Minho, ISISE, Department of Civil Engineering, Guimarães, Portugal,
carla.colombo95@gmail.com
4
Full Professor, University of Minho, ISISE, Department of Civil Engineering, Guimarães, Portugal,
pbl@civil.uminho.pt
INTRODUCTION
Masonry is one of the most used building material in the world. Masonry buildings also represent
a significant part of our cultural heritage, being the physical memory of our civilisation’s history.
Because they are often old, these historical structures are often highly damaged (due to, e.g.
foundation settlements [1], past earthquakes [2-4], insufficient initial design [5-6], or weathering).
They need today maintenance operations. Their assessment and the identification of the
maintenance or strengthening actions are carried out using either limit analysis [7-10] or numerical
[1, 10-16] tools. Both categories can then be divided into micro-modelling [8-9, 11-12] or macro-
modelling approaches [1, 7, 10, 13-16]. In micro-modelling approaches, each block is modelled
separately and interact one with another at an interface. In macro-modelling approaches, masonry
is represented by an homogeneous continuum with mechanical properties that depend on the one
of the the masonry units and of the mortar. Both of them need to define a geometrical parameter:
the size of the masonry units. This directly dictates the size of the blocks in micro-modelling [8-9,
11-12], the failure criteria in limit analysis macro approaches [7, 10] or lastly, the parameter of an
equivalent homogenised masonry material [1, 13-16]. Indeed, blocks size has been found to
influence the final load capacity of a masonry structure subjected to a foundation settlement [8].
However, given the variety of stones and methods to assemble them, no universal and unique
description of the bond pattern is possible. In practice, it leads to very irregular masonry patterns,
even when using rectangular blocks (Figure 1). Finally, when using regular and identical bricks as
masonry units, one can still observe various bond patterns worldwide (Figure 2).
Figure 1: Masonry pattern using regular blocks for the historic perimeter wall of
Guimarães, Portugal
Figure 2: Different typical bond patterns
Some in-field studies have been developed to acquire the actual sizes and shapes of masonry units
[16-19]. They are mainly based on Terrestrial or Mobile Laser Scanning technologies that create
a 3D data Point Cloud of the monitored object. Though they all showed the high potential of Laser
Scanning approaches, they also demonstrated that the extraction of the block sizes and shapes out
of the created Point Cloud is not an easy task [17-18]. For instance, Valero et al. developed an
automatic plugin for this extraction, but that only handles “straight” walls (i.e. without significant
curvatures).
For this reason, masonry structures have been (and still today) often modelled as an assemblage
of regular blocks, either from a computational point of view [8-9, 11] or from an experimental
point of view [11-12, 20-22]. In this respect, it is evident that a micro-modelling strategy [8-9, 11]
may become significantly time-consuming during the preparation of the model geometry because
of the number of units and the definition of all possible interacting interfaces [8-9]. Computer-
Aided-Design (CAD) can be of great help in this laborious task [23]. One of the noteworthiest
efforts in this regard is BIM-M (Building Information Modeling for Masonry) [24]. That initiative
more specifically focuses on developing the construction-oriented data structure. Yet, providing
the database collecting the structural aspects of the masonry assemblage is missing in the literature.
To tackle these challenges and improve the CAD already existing solutions, this work aims to
present an innovative tool that generates masonry geometries in a very time-efficient way. Its main
advantage lies in the parametrisation of all geometrical data, allowing geometrical updates very
quickly.
GENERAL DESCRIPTION AND PURPOSES OF THE TOOL
The generative algorithm is implemented in the environment offered by Rhinoceros3D-
Grasshopper to use the programming language C#. Its purposes can be gathered as follows:
1. Allowing the geometrical modelling of masonry structures, from simple shapes (e.g. U-
shaped, Corners [21]) to more complex ones. The time required to build these models
should be proportional to the complexity of the structure and comparable to other CAD
software [23].
2. Covering all possible configurations and let the user choose his/her design. Hence, several
geometrical parameters must be defined.
3. Allowing a fast update of the geometry when some parameters vary.
4. Enabling the easy exportation of the generated geometries to structural software for
analysis (e.g., DIANA, Abaqus, 3DEC, LiABlock_3D, etc.).
The present tool allows modelling geometries with single-leaf masonry bond patterns. The first
three objectives are already fulfilled for single-leaf structures. More specifically, the time
requested to build a classical U-shaped masonry specimen is approximately ten minutes, while it
updates the model in less than one second for each parameter variation. The following sections
describe, in a more detailed way, the present generative algorithm.
PARAMETRISATION CHOSEN TO CREATE MASONRY GEOMETRIES
For simplicity, the generative algorithm uses five components corresponding to five different
masonry elements. The first one corresponds to a masonry wall. The second to fourth ones
correspond to masonry connections (corner, T-connection and cross-connection, respectively) that
assemble two (or more) walls. The last component has been developed to create an opening in an
already existing wall.
One should note that using these different typologies allows creating almost all possible masonry
configurations. However, curved walls are not yet included, though the same methodology can be
easily applied without significant work.
Wall component
The wall component is the main component of the generative algorithm. Figure 3 gives the
perspective and plane views of it. One can note that the single-leaf bond pattern only considers
two distinct courses (even and odd) that are repeated all along with the component’s height. This
characteristic is shared by all features described in the following.
Figure 3: Wall component. a) Perspective, b) plane, and c) front views.
In total, a wall component has 19 parameters (Figure 4):
- Three geometrical parameters describe the masonry units: length l_b, width w_b and height
h_b (Figures 3 & 5).
- Two geometrical parameters define the wall: its length L and height H (Figure 3).
- Two Boolean parameters ask if the length and height of the blocks need to be recomputed
based on the wall length and height, respectively to avoid very small or thin blocks. If
activated, they ensure that L = n × l_b and H = n × h_b with n to be an integer.
- A Boolean parameter asks whether the first course of the wall is an odd or an even type
(Figure 3).
- For each side, three parameters handle the connection properties. Specifically, the first one
is a Boolean asking if the side should be straight or indentated (Figure 6). If the side is
indentated (Figure 6b), the second parameter determines the length of the indent l_1 and
l_2, respectively (Figure 5).The last parameter is also a Boolean specifying if the
indentation is done on the even or odd courses.
- The horizontal distance between heading joints of even and odd layer is given by e (Figure
5).
- The minimum length that blocks have to be indeed created (Figure 6c).
- Three parameters that determine the coordinates of the local origin O of the wall, the
orientation of its longitudinal dimension (V
1
) and the orientation of its elevation dimension
(V
2
), see Figure 6a.
Figure 4: An overview of the GH component to model a wall with its inputs and outputs.
As for the output, all automatically generated blocks of a wall are stored in a DataTree
(Rhinoceros3D data structure), where each branch of the tree corresponds to a layer. It allows
a much easier utilisation of the output later on.
The position O and orientation vectors of both side of the wall component is also outputted.
Since all the other features use the same methodology to define their positions, it allows
directly connecting the output position parameters of a given component to the input position
parameters of another component to be connected with (Figure 7).
Figure 5: Parametrisation used for the wall component
Figure 6: a) Straight and b) indentated connections. c) Effect of the minimum allowable
block length
Figure 7: Connection of components to create the global geometry. a) Visual programming
and b) preview
Corner component
A masonry corner component binds two different walls together connected orthogonally. It has 13
inputs (Figure 7), and most of them are similar to the inputs of a wall component. One of its
particularities lies in the fact that the width of the two connected walls can be different (Figure 8).
Another particularity is the possibility to connect either the 1
st
branch or the 2
nd
one, resulting in a
“clockwise” or “counter-clockwise” connection (Figure 8).
Figure 8: Corner component: a) and b) plane view and parametrisation; c) and d) previews
with clockwise and counter-clockwise connections
T-connection and Cross-connection components
T-connection components connect three walls and have a total of 17 inputs. Most of them are
shared with the previous features (wall and corner connection). Again, the width of the three
branches to be connected can be different (Figure 9).
Figure 9: Parametrisation for a T-connection: a) even and b) odd layers; c) larger block
length; d) misalignment of wall’s façade and e) deleting of small blocks
Two specific features of T-connections are also presented on Figure 9. First, blocks can be merged
if long enough blocks are available (Figure 9b-c). Moreover, in case the widths of the façade’s
walls are different, the user can determine the misalignment m (Figure 9d) so that the walls can
either be aligned along their external (m = 0%) or internal faces (m = 100%). Finally, the last input
parameter corresponds to the minimum length of blocks and avoid creating too small blocks
(Figure 9d-e).
The cross-connection component has 19 inputs and works similarly to the T-connection features.
The refinements discussed in Figure 9 apply in both x and y-directions, thus leading to more input
parameters. Examples of T-wall and Cross-wall are shown in Figure 10.
Figure 10: Final model geometries created using the five essential components.
Opening
The last feature regards the creation of an opening in an existing wall. The component has 23 input
parameters, of which three corresponds to the wall DataTree (and eventually of two connections
components) to cut. Then, six parameters define the relative and absolute position of the opening,
while three other dimensions (l
op
and h
op
) control its size (Figure 11). Six parameters corresponding
to the block and wall geometries are also needed. Finally, one Boolean parameter determines if a
lintel should be created above the opening, and if yes, the lintel height (h
l
) and width (w
l
) as well
the length of the supporting columns (b
l
) must be assigned.
Figure 11: Parametrisation of the opening and the lintel above it
DISCUSSION AND CONCLUSIONS
The five parametrised components presented above are enough to create diverse masonry
geometries (Figure 10). A simple structure like those shown in Figure 10c does not require more
than 10 minutes to be built. Besides, as Figures 4 & 7 demonstrate, the time to update the geometry
is less than one second. The proposed generative algorithm, therefore, reaches high performance
in terms of real-time modeling. On the one hand, it can find application as an input of structural
analysis software that aims at doing fast calculations (e.g. micro-modelling Limit Analysis tools
[8-9, 23]). Indeed, the construction of complex geometrical models of masonry structures can be
a laborious task that can largely overpass the structural analysis time, reducing the time-efficiency
of these practical tools. Moreover, this tool is particularly suitable for the parameterically shape
exploration to find e.g., the optimal geometry manually..
To conclude, the presented tool shows auspicious results and need to be further developed to
account for multi-leaf bond patterns, gable walls, curved and inclined walls as well as non-
homogeneous block size, which are all envisioned as next development steps. Furthermore, the
automatisation of the exportation of the generated masonry geometries to classical structural
analysis software is currently being investigated.
ACKNOWLEDGEMENTS
This study has been funded by the STAND4HERITAGE project (new standards for
seismic assessment of built cultural heritage) that has received funding from the European
Research Council (ERC) under the European Union’s Horizon 2020 research and innovation
programme (Grant agreement No. 833123), as an Advanced Grant. Its support is gratefully
acknowledged. The opinions and conclusions presented in this paper are those of the authors and
do not necessarily reflect the views of the sponsoring organisation.
REFERENCES
[1] Alessandri, C., Garutti, M., Mallardo, V., & Milani, G. (2015). Crack patterns induced by
foundation settlements: Integrated analysis on a renaissance masonry palace in Italy.
International Journal of Architectural Heritage, 9(2), 111-129.
[2] Vlachakis, G., Vlachaki, E., & Lourenço, P. B. (2020). “Learning from failure: Damage and
failure of masonry structures, after the 2017 Lesvos earthquake (Greece).” Engineering
Failure Analysis, 117, 104803.
[3] Bruneau, M. (1994). “Seismic evaluation of unreinforced masonry buildings—A state-of-the-
art report.” Canadian Journal of Civil Engineering, 21(3), 512-539.
[4] Leite, J., Lourenco, P. B., & Ingham, J. M. (2013). Statistical assessment of damage to
churches affected by the 2010–2011 Canterbury (New Zealand) earthquake sequence. Journal
of Earthquake Engineering, 17(1), 73-97.
[5] Lourenço, P. B., Mendes, N., Ramos, L. F., & Oliveira, D. V. (2011). “Analysis of masonry
structures without box behavior. International Journal of Architectural Heritage, 5(4-5),
369-382.
[6] de Felice, G., De Santis, S., Lourenço, P. B., & Mendes, N. (2017). “Methods and challenges
for the seismic assessment of historic masonry structures.” International Journal of
Architectural Heritage, 11(1), 143-160.
[7] D’Ayala, D., & Speranza, E. (2003). “Definition of collapse mechanisms and seismic
vulnerability of historic masonry buildings.” Earthquake Spectra, 19(3), 479-509.
[8] Gagliardo, R., Portioli, F. P. A., Cascini, L., Landolfo, R., & Lourenço, P. B. (2021). A rigid
block model with no-tension elastic contacts for displacement-based assessment of historic
masonry structures subjected to settlements. Engineering Structures, 229, 111609.
[9] Portioli, F., Casapulla, C., Gilbert, M., & Cascini, L. (2014). Limit analysis of 3D masonry
block structures with non-associative frictional joints using cone programming. Computers &
Structures, 143, 108-121.
[10] Funari, M. F., Spadea, S., Lonetti, P., Fabbrocino, F., & Luciano, R. (2020). Visual
programming for structural assessment of out-of-plane mechanisms in historic masonry
structures. Journal of Building Engineering, 31, 101425.
[11] Bui, T. T., Limam, A., Sarhosis, V., & Hjiaj, M. (2017). Discrete element modelling of the
in-plane and out-of-plane behaviour of dry-joint masonry wall constructions. Engineering
Structures, 136, 277-294.
[12] Meriggi, P., de Felice, G., De Santis, S., Gobbin, F., Mordanova, A., & Pantò, B. (2019).
Distinct element modelling of masonry walls under out-of-plane seismic loading.
International Journal of Architectural Heritage, 13(7), 1110-1123.
[13] Valente, M., Milani, G., Grande, E., & Formisano, A. (2019). Historical masonry building
aggregates: advanced numerical insight for an effective seismic assessment on two row
housing compounds. Engineering Structures, 190, 360-379.
[14] Gams, M., Anžlin, A., & Kramar, M. (2017). Simulation of shake table tests on out-of-plane
masonry buildings. Part (III): two-step FEM approach. International Journal of Architectural
Heritage, 11(1), 94-102.
[15] Landolfo, R., Gagliardo, R., Cascini, L., Portioli, F., Malena, M., Tomaselli, G., & de Felice,
G. (2020). Rigid block and finite element analysis of settlement-induced failure mechanisms
in historic masonry walls. Frattura ed Integrita Strutturale, 14(51), 517-533.
[16] Milani, G., Esquivel, Y. W., Lourenço, P. B., Riveiro, B., & Oliveira, D. V. (2013).
Characterisation of the response of quasi-periodic masonry: Geometrical investigation,
homogenisation and application to the Guimarães castle, Portugal. Engineering structures,
56, 621-641.
[17] Riveiro, B., Lourenço, P. B., Oliveira, D. V., González‐Jorge, H., & Arias, P. (2016).
Automatic morphologic analysis of quasi‐periodic masonry walls from LiDAR. Computer
Aided Civil and Infrastructure Engineering, 31(4), 305-319.
[18] Valero, E., Bosché, F., Forster, A., & Hyslop, E. (2019). Historic digital survey: reality
capture and automatic data processing for the interpretation and analysis of historic
architectural rubble masonry. In Structural Analysis of Historical Constructions (pp. 388-
396). Springer, Cham.
[19] Park, H. S., Lee, H. M., Adeli, H., & Lee, I. (2007). A new approach for health monitoring of
structures: terrestrial laser scanning. ComputerAided Civil and Infrastructure Engineering,
22(1), 19-30.
[20] Savalle, N., Vincens, E., & Hans, S. (2018). Pseudo-static scaled-down experiments on dry
stone retaining walls: Preliminary implications for the seismic design. Engineering
Structures, 171, 336-347.
[21] Restrepo Vélez, L. F., Magenes, G., & Griffith, M. C. (2014). Dry stone masonry walls in
bending—Part I: Static tests. International Journal of Architectural Heritage, 8(1), 1-28.
[22] Graziotti, F., Tomassetti, U., Penna, A., & Magenes, G. (2016). Out-of-plane shaking table
tests on URM single leaf and cavity walls. Engineering Structures, 125, 455-470.
[23] Cascini, L., Gagliardo, R., & Portioli, F. (2018). LiABlock_3D: a software tool for collapse
mechanism analysis of historic masonry structures. International Journal of Architectural
Heritage, 14(1), 75-94.
[24] Gentry, T. R., Eastman, C., Haymaker, J., Lee B., Development and Documentation of Case
Studies of Masonry Construction Projects, technical report, 2016, project-ii-benchmark-
project-final-report-from-georgia-tech-january-16-2016.pdf (bimformasonry.org)
... These are the models which have a similar load bearing mechanism as simple models, but the bearing walls may be divided into two parts. The first part being the masonry columns at the corners, and the other is the bearing walls joining the corner columns [6]. Both, the corner columns and the bearing walls bear the combination of the loading from the roof (the dead and live loading and the self-weight of the roof structure), the only difference is the interaction of the bearing wall and the masonry column. ...
Conference Paper
Full-text available
Masonry structures may be of different kinds, including stone, brick and other masonry blocks assembled in different ways using mortar. For study of such structures, modelling them is necessary which can be done in two ways, either by modelling large classes of masonry as single units (Simple Model) or by regular arrangement of blocks and the interaction between blocks and mortar (Refined Model). Simple Model approach takes crude material assumptions and produces results affected by large approximations , while the second approach is sophisticated and close to reality. The present research deals with the creation of a method for modelling of Finite Element Models of masonry buildings through their fast regenerative modelling in an environment offered by Rhinoceros 3D and Grasshopper. Since exact modelling of such structures is complicated and time-consuming, macro modelling of such structures is carried out by simplistic assumptions of the structural system as load bearing walls. Simple architectural plans of each storey may be imported from AutoCAD to Rhinoceros 3D, scaled according to the dimensions. These can then be assigned as geometry in Grasshopper and extruded in Z-direction (according to storey height) to get the basic volume of the curves. The plan of each storey can then be placed on top of one another by maintaining a distance equal to the storey height and the whole 3D model may be created. These models are user friendly and consume very less time, if definitions for extrusion and openings etc. are created beforehand in Grasshopper, and can be used for multiple structures. Such macro models are then used for simple static analysis by using a similar script in Grasshopper or may be imported to other Finite Element Analysis (FEA) softwares.
Conference Paper
Full-text available
Zagreb Cathedral was severely damaged by the 5.4 magnitude earthquake on March 22, 2020. The church had already been hit by a severe earthquake in 1880 and was restored in the neo-Gothic style with cross vaults and arches supported by slender columns. In the restoration, the spatial connection between the load-bearing elements was poorly executed and the structural system for horizontal loads has not been significantly improved, which resulted in serious damage from the new earthquake. For the purposes of a new reconstruction, a structural analysis of the Cathedral is currently being done using the distinct element method (DEM). This paper focuses on the geometry of blocks that form a cross vault. The available solutions for the automatic generation of masonry geometries developed in Grasshopper are mostly focused on the generation of walls [4] or barrel vaults. The vaults in Zagreb Cathedral are non-standard cross vaults that are not square in plan, but strongly rectangular. The crowns of orthogonal barrel vaults that form a cross vault are not at the same height and some are not horizontal. Since the block arrangement in vaults (stereotomy) is of great influence on the behaviour of cross vaults (less depending on mechanical properties [5]), this paper proposes a different definition of groins of untypical cross vaults. Finally, results (failure mechanism) of a numerical model for the cross vault after time history analysis obtained with 3DEC software are shown.
Article
Full-text available
A two-step strategy for the mechanical analysis of unreinforced masonry (URM) structures, either subjected to in- and out-of-plane loading, is presented. At a first step, a semi-automatic digital tool allows the parametric modeling of the structure that, together with an Upper bound limit analysis tool and a heuristic optimization solver, enables tracking the most prone failure mechanism. At a second step, a coupled concurrent FE model with micro- and macro-scales is assumed. A micro-modeling description of the masonry is allocated to regions within the failure mechanism found in the former step. In converse, the other domain regions are modeled via a macro-approach, whose constitutive response is elastic and orthotropic and formulated through closed-form homogenized-based solutions. The application of the framework is based on non-linear static (pushover) analysis and conducted on three benchmarks: (i) an in-plane loaded URM shear wall; (ii) a U-shaped URM structure; and (iii) a URM church. Results are given in terms of load capacity curves, total displacement fields, and computational running time; and compared against those found with a FE microscopic model and with a limit analysis tool. Lastly, conclusions on the potential of the framework and future research streams are addressed. Keywords: Masonry, Micro-modeling, Macro-modeling, two-step approach, Homogenization, URM Applications, Concurrent FE model
Article
Full-text available
This paper deals with the vulnerability assessment of historic masonry structures subjected to settlements using rigid block modelling. A 2-D rigid block model with unilateral elastic contacts and finite friction is developed for the evaluation of the displacement capacity in the large displacement regime by push-down analysis. A variational formulation of the rigid block model is adopted, which relies on associative behavior for displacement rates. Under this assumption, the equation systems governing the behavior of the rigid block model can be uncoupled into two equivalent force and displacement-based problems, thus reducing computational costs. The numerical model was validated against the results of an ad-hoc experimental campaign on small scale tuff panels and against the tests on a small-scale masonry façade subjected to moving supports, from the literature. Applications of the proposed modelling approach are presented to the assessment of a full scale, monumental masonry façade. Following classic force-displacement methods that are used in the case of seismic actions, capacity curves are proposed for the damage assessment induced by settlements. Those state the relation between the base reaction at the moving supports and the displacement of a control point, which is obtained from the push-down analysis. Finally, comparisons with empirical assessment methods from the literature are presented.
Article
Full-text available
The paper is related to the assessment of collapse mechanisms of historic masonry structures suffering settlements induced by ground movements. Two numerical strategies are adopted in order to study the influence of the settled zone on the cracking of masonry buildings: a discrete rigid block model and a continuous homogenized model. The first approach provides an estimate of the collapse load and failure pattern of masonry based on the lower bound theorem of limit analysis. The second approach is formulated in the framework of multi-surface plasticity and implemented in a FE code for the path-following non-linear analysis of masonry wall described as continuous anisotropic plate. Several settlement configurations, of masonry walls under moving ground support are investigated and the corresponding failure patterns resulting from the analysis are obtained resulting in local or global failure modes. The results of the two modeling formulations are compared and discussed in order to highlight the features of the two different approaches in the prediction of settlement-induced damage.
Article
Full-text available
A rigid block model is proposed for collapse mechanism analysis of three-dimensional historic masonry structures subjected to point live loads, seismic-induced lateral loads and settlements. The model is made of polyhedral rigid blocks interacting at no-tension, frictional contact interfaces and can be used to represent complex assemblages and bond patterns. The formulation and the solution procedure of the underlying limit equilibrium analysis problem were implemented in LiABlock_3D, a MATLAB based tool with Graphical User Interface (GUI). The software was designed to import the geometric model from commercial Computer Aided Design (CAD) tools, thus allowing high flexibility of structural configurations and masonry patterns. The graphical interface is also used to define material properties as well as boundary and loading conditions. Numerical and experimental case studies from the literature were analyzed to show the ability of the model developed in predicting the collapse behavior of a variety of structural typologies. Those include arches, vaults, and domes under vertical and horizontal live loads and spreading supports. A two-story masonry building with a barrel vault at first level is also analyzed under variable lateral loads and support movement. Potentialities and limitations of the proposed formulation and tool are discussed on the basis of the results obtained and also in terms of computational efficiency.
Article
Full-text available
Dry stone retaining walls (DSRWs) are vernacular structures which can be found all over the world. Most of them have been built in the 19th century but they can be as old as two hundred years. Because of decades of neglect, many of these walls are highly damaged; however, in the absence of national rules for this peculiar heritage, any intervention on these constructions is made difficult. A number of former studies in France have tried to settle the bases for a standard aimed at designing slope DSRWs. This paper complement those works in order to give clues for a seismic design of slope DSRWs following the simplified approach proposed by the Eurocode 8 and denoted pseudo-static approach. Firstly, scaled-down experiments have been carried out using a mock-up composed of a wall made of clay bricks retaining a sandy backfill. This mock-up was then tilted and the characteristics of the system at failure were reported and analysed. A particularity of failure in DSRWs is that the failure surface crosses the wall leaving a part of it intact. Secondly, an analytic method based on the limit equilibrium using the Coulomb wedge theory has been designed to predict the tilting angle and validated on the basis of these experimental results. Finally, some preliminary implications for the seismic design of DSRWs have been proposed for walls built in moderate seismicity areas which is typical of mainland France. In low seismicity zones, the extra width required to bear the seismic motion does not exceed 25%% of the width identified through the static design. Results are also given for more critical cases associated to zones of higher seismicity as well as for different wall configurations. Keywords : Masonry; Seismic; Scaled-down tests; Slope
Article
On the 12th of June 2017 an earthquake of Mw=6.3 struck SSE of Lesvos Island, causing one human fatality and severe damage to the built environment. The traditional settlement of Vrissa was the most affected area, having masonry structures as the majority of its building stock. The objective of the present study is two-fold: to present the structural damage and failure patterns induced by the Lesvos earthquake to masonry structures; to highlight the causes and weaknesses that led to damage, or the factors that prevented it. Particular attention is paid to traditional construction techniques and architectural features that contributed to the seismic response of the structures, either having beneficial or detrimental effect. To this end, a field reconnaissance has been conducted and meaningful technical conclusions are drawn by the observations. Structural systems of both unreinforced and timber-reinforced masonry are inspected. Besides the identification of frequent cases of local, out-of-plane and in-plane mechanisms, combined global mechanisms are also pointed out. Finally, insight of the performance of past interventions is also given, assisting the challenging task of engineering practice.
Article
This work aims at proposing a novel procedure for the seismic assessment of historic masonry structures which is computationally efficient and does not rely on destructive material tests. Digital datasets describing the geometric configuration of historic masonry structures are employed to automatically generate a non-linear Finite Element (FE) model and investigate on possible collapse modes. A configuration of failure surfaces is therefore detected through the Control Surface Method (CSM), which is here proposed for the first time. In a following step of the analysis, structural macroblocks are identified, whereas an upper bound limit analysis approach is employed to estimate the structural capacity of the structure. Genetic Algorithms are also employed to detect the actual failure mode for the structure. The procedure is implemented into a visual coding environment, which allows one to parametrically explore all possible failure surfaces and immediately visualize the effects of the user assumptions. This is particularly suited to support a decisions-making process which strongly relay on engineering judgement. The procedure is validated by the analysis of several benchmark cases, whose results are presented and discussed.
Article
The out-of-plane vulnerability of masonry walls plays a crucial role in the seismic response of existing structures. Depending upon mechanical properties and section morphology, collapse may occur by the onset of a mechanism or, as historic constructions often exhibit, leaf separation, disaggregation or sliding. In these latter cases, structural analyses based on rigid-body mechanics may overestimate the seismic capacity, thus resulting unconservative. The distinct element method (DEM), which represents masonry as an assembly of discrete blocks and nonlinear interfaces, could instead be used. Nevertheless, it is more complex and requires more input parameters, so it is still barely applied in engineering practice. In this paper, the seismic out-ofplane response of masonry walls was modelled with DEM. A shake table test on a two-leaf rubble stone masonry wall and a single-leaf wall in tuff blocks was simulated through nonlinear dynamic analyses. The mechanical properties of joints were calibrated on the basis of dynamic identification under low-intensity white noise input, leading to a good prediction of the seismic response. Then, they were further tuned based on the surveyed crack pattern for an improved matching between experimental results and numerical postdictions. Finally, the results provided by the limit analysis were discussed in the light of DEM simulations.
Article
Historical masonry aggregates represent a large portion of the cultural heritage in Italy and are highly vulnerable to seismic actions, as shown by past seismic events. Typically, they are large and complex structures for which there is a lack of knowledge and information concerning the structural behavior, in particular as far as the response to seismic actions is concerned. This paper investigates the seismic response of two complex historical masonry aggregates located in Sora (Lazio region, Central Italy), through advanced 3D FE numerical simulations. For each aggregate, a detailed 3D FE model is developed and analyzed in the non-linear dynamic range, assuming that masonry behaves as a damaging-plastic material with almost vanishing tensile strength. The seismic performance of the two aggregates is evaluated in terms of damage distribution, energy density dissipated by tensile damage and maximum normalized displacements. The numerical analyses show the high vulnerability of the perimeter walls. In particular, the units at the extremities of the aggregate are subjected to large displacements, being not efficiently braced by the adjacent units and being subjected to the torsional effects induced by the seismic action. The presence of several openings is a fundamental feature that significantly decreases the strength of the perimeter walls, influencing the damage distribution in the aggregate mainly due to out-of-plane actions. The most damaged elements are generally the walls of the tall units without lateral support and the adjacent slabs covering large spans. Numerical results also show that the structural response of a single building unit is affected by the interactions with adjacent structural parts. Moreover, it can be stated that a preliminary structural assessment through kinematic limit analysis on partial failure mechanisms may be reliable only after a proper estimation of the different structural elements playing a role in the horizontal behavior (e.g. interlocking between walls, typology of masonry, distribution of horizontal loads, constraints and dead loads distribution, etc.). The obtained results will be also used in an accompanying paper to benchmark simplified approaches that can be employed by engineers in common design practice to quickly predict the seismic vulnerability of masonry aggregates and define the most suitable strengthening interventions.
Article
This paper aims to improve knowledge on the suitability of the discrete element method (DEM) to simulate the in-plane and out-of-plane behaviour of different in-configuration structural masonry walls constructed with dry joints. The study compares the results obtained from laboratory tests against those predicted using the three-dimensional distinct element 3DEC software. Significant features of the structural behaviour shown by the walls are discussed and conclusions on their ultimate capacity and failure mechanisms are addressed. A key feature of the DEM is the important role that brick discontinuities, i.e. joints, play in the mechanics of masonry. Within DEM, the bricks were modelled as continuum rigid elements while the joints were modelled by line interface elements represented by the Mohr-Coulomb law. The analysis of the results showed that the model developed is capable of representing the crack development and load carrying capacity of masonry structures constructed with dry joints with sufficient accuracy. Moreover, a collection of experimentally verified material parameters is provided to be used by other researchers and engineers and to develop a reliable model to solve engineering challenges worldwide.