- Email: [email protected]
Influence of the annex on seismic behavior of historic churches
Engineering Failure Analysis 45 (2014) 300–313
Contents lists available at ScienceDirect
Engineering Failure Analysis journal homepage: www.elsevier.com/locate/engfailanal
Influence of the annex on seismic behavior of historic churches A. Dal Cin ⇑, S. Russo IUAV University of Venice, Dept. of Design and Planning in Complex Environments, Dorsoduro 2206, 30123 Venice, Italy
a r t i c l e
i n f o
Article history: Received 21 January 2014 Received in revised form 3 July 2014 Accepted 6 July 2014 Available online 16 July 2014 Keywords: Historical church Annex Interface Damage identification Modal analysis
a b s t r a c t In May 2012, two major earthquakes occurred in Emilia Romagna region in Northern Italy, causing widespread damage. The hypocentre of the second one, strokes Mirandola where is located the Gesù Church investigated in this research. The church has a long and important annex to the south built during the same period of the church. This paper addresses how the important annex influenced the seismic response of this historical church and how, more generally, this kind of asymmetric mass can influence the behavior of historic churches. The final considerations are based on the comparison between the structural damage pattern survey and modal and seismic FE analysis. A FE model was constructed considering four different configurations: (i) isolated church, (ii) the church with the presence of the real annex with a perfect connection, (iii) the church with the presence of the same annex but with an interface between the church and the annex and (iv) this last configuration with the stiffness degradation of the interface. Firstly the dynamic modal analysis and subsequently the seismic spectral analysis were performed. The results indicate that the annex’s presences play a significant role in the dynamic response of the church and affect the distribution of damages for the whole building. The results of the seismic simulation are in agreement with the observed damage. Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction The recent earthquakes that hit Italy in the last century proved the high vulnerability of cultural heritage, with particular reference to churches. These particular monumental buildings cannot be reduced to any standard structural scheme and this makes it difficult to evaluate their seismic reliability. To overcome this problem, the macro-element approach has been proposed a few years ago and since then repeatedly used to recognizing the collapse mechanisms in the different macroelements of the church [1,2]. The common collapsing configurations are shown in PCM-DPC-MiBAC M.-A.-D. [3]. The historical centers often occur as the result of an uncontrolled constructive evolution, whose complex configurations lead the structures to strongly interact with each other when are subjected to seismic action [4]. As a matter of fact, several historic churches are not isolated from the urban context but are often characterized by the presence of adjacent buildings, usually named annex (convents, sacristy, tower, minor constructions, etc.) at the same time of the church or subsequently constructed as we have seen in other case studies (Fig. 1). Besides the monitoring of historical constructions in seismic areas is a predominant issue also in Europe and especially in Italy because of the richness of its inestimable architectural heritage [5,6]. In the last century the seismic events stroked severely the cultural historic heritage and in particular the 49% of the damaged structures are churches highlighting their
⇑ Corresponding author. E-mail addresses: [email protected] (A. Dal Cin), [email protected] (S. Russo). http://dx.doi.org/10.1016/j.engfailanal.2014.07.004 1350-6307/Ó 2014 Elsevier Ltd. All rights reserved.
A. Dal Cin, S. Russo / Engineering Failure Analysis 45 (2014) 300–313
301
Fig. 1. Some examples of churches characterized by the presence of the annex and damaged by earthquakes.
intrinsic vulnerability [7–17] due first to the absence of any kind of diaphragm except the covering. Tuned with this issue it becomes necessary to take a census of the weakness of these monuments through the structural identification procedures and the evaluation of the related ground motion characteristics [18,19]. By the way the paper presents the case study of Gesù church in Mirandola, which was severely damaged by the strong earthquake that occurred in Emilia Romagna region in Northern Italy, in May 2012. The question of how the important annex had engraved on the dynamic behavior during the earthquake arise from the observation of an unsymmetrical damage survey. The construction of the undamaged FE model, firstly, of the isolated church (WA) and then with the presence of its annex (A) was fundamental to simulate the dynamic and seismic behavior. As previously said, the annex could have been built later than the church; this hypothesis is considered in an additional FE model (Ai) with the inclusion of an interface between the church and the annex. One of the main problems of the historic masonry buildings is certainly the materials deterioration, particularly of the mortars. The reason of the deterioration of this part is often due to the low quality of conservation and other exceptional events or new additional loads [20]. It is chosen therefore to simulate also this problem in the FE model with a deteriorated interface Ai_d. The dynamic modal analysis and the response spectrum analysis have been performed to determine the possible collapse mechanisms and to study the stresses distribution on each macro-element of the whole church analyzed. This case study highlights the importance of a deeper study on the incidence of the presence of an important annex on the dynamic response of the churches. The analysis of this case study achieves some results, which may be useful also for correctly driving future strengthening interventions on similar context.
302
A. Dal Cin, S. Russo / Engineering Failure Analysis 45 (2014) 300–313
2. Description of the case study The seventeenth-century Gesù Church (Fig. 2) was built on the orders of Alessandro I Pico on the occasion of his investiture as Duke of Mirandola. The church has a typical basilica plan, indeed it is characterize by a wide single nave joined on either side by side chapels separated by Corinthian fluted pillars and on the right side to border with the transept there is an entirely wooden pulpit. This is also an important transept characterized by two extraordinary wooden altars in the walls of back and majestic frames; the church ends with an apse regular where the main altar lies. The transept and the apse are connected on both sides by two volumes of two levels, the sacristy, probably dating on the same period of the church building (Figs. 2c and 3). All structural entities of the church are covered with a cross and barrel vaulting that are undecorated. As it can be seen in Fig. 3a the church has a long and important annex to the south, the College of Gesuiti. This building was opened in 1690 a ‘year later of the completion of the church, has two floors and a classic style. The exterior looks rather poorer in particular the façade remained unfinished (Fig. 2b). All external walls are brick surface that makes legible all the renovations that took place over time. The roof of the building was restored after years of work in 1998. It is possible to see the intervention with steel brackets focused to the node strut-chain in the support of the truss to the wall. The bracket of the beam has been applied to steel tie which acts in cooperation with the chain itself. The church can be divided into different macro-element: the nave, the side chapels, the transept, the apse and the façade (Fig. 3c). It presents a large nave joined on both sides by the chapels of modest size, the façade has a very regular plant and prospects, the transepts and the apse have significant dimensions both in plan and in height. The plant shows a longitudinal length along the main axis x that is equal to 32 m and a width along the y axis of 18 m. The height of the Church reaches 22 m.
3. Damage survey The Emilia-Romagna has been subject to an important seismic sequence in 2012 that hit the region lead to severe damage to the historical and cultural heritage. The first serious earthquake, with a magnitude of 5.9 ML was recorded May 20, 2012 with its epicenter in Finale Emilia, at a depth of 6.3 km. Another powerful quake of magnitude 5.8 ML and depth of hypocenter 9.6 km was recorded on 29 May 2012. In Ferrara area, the main damage is caused by the earthquake of 20 May, instead in the most western part of the Modena (between Mirandola, Carpi, Novi di Modena, Cavezzo and Concordia) the most serious effects are certainly due to the earthquake of 29 May, mainly due by the proximity to the epicenter. The earthquake has severely affected the structure, in particular the most visible damage is the collapse of the vaults of the church along the main axis x (Fig. 4). The nave was characterized by cross vaults and the apse by barrel vaults. The parties collapsed give the possibility to identify the type of texture characterized by bricks placed in horizontal position that identifies indeed a very thin structure element. The only cover parts that have survived are the arches that marked the vaults where the bricks placed in vertical position and where we observed the presence of metal chains (Fig. 4). The geometry of the thin vaults and their static behavior predominantly in compression have proved to be vulnerable also in response to the severe amplitude of the vertical component. The roof trusses have well reacted to the seismic activity due to the consolidation work carried out recently while the wooden roof suffered a collapse in the attached part of the façade (Fig. 5). The inner part of the façade is seriously damaged (Fig. 5b and d) especially the level above the trabeation for the loss of structural stiffness and probably due to a reduced sectional dimension than the first level. The crack most serious propagates from the center of the top of the architrave of the door affecting the central opening that is also a point of vulnerability and from the trabeation and the arms at the center. The top of the façade is a very weak point of the macro-element caused by the poor connection with the cover.
Fig. 2. Church of Gesù, view of north later part (a), the façade (b) and the apse (c).
A. Dal Cin, S. Russo / Engineering Failure Analysis 45 (2014) 300–313
303
Fig. 3. View of the church and the annex (a), detail of the annex (b) and macro-elements plan (c).
Fig. 4. View of the vaults before (a) and after the earthquake (b).
Fig. 5. View of the external and internal façade before (a and c) and after the earthquake (b and d).
Observing Fig. 6, the masonry façade undergoes a detachment from the side walls, in particular in the second level in height of the church; analyzing carefully the corner south is evident that the lesion in corner, characterized by a slipping of the bricks, affects in particularly the top part and propagates downwards up to damage the entire trabeation and the capital of the pillar corner (Fig. 6a, detail 1). In proximity on the corner south is also detected a deeper crack that propagates from the upper part of a window to the frame of the trabeation below (Fig. 6a, detail 2). It is probably due to the presence of the adjacent system support bell, also the trabeation of the north corner are cracked at the same point (Fig. 6b, detail 3). Fig. 7 shows the south side chapels before (Fig. 7a) and after the earthquake. In the general view of the damage, the side chapels in the south of the nave (Fig. 7b) present any serious damage for the presence of annexes on the south side of the church that helped to have a greater stiffness in the first level in height. However, the side chapels to the north (Fig. 7c) have a severe crack pattern and the presence of deep diagonal lesions along the transverse walls of the nave and in the barrel vaults which also present partial detachment of material. The activation of a detachment of the side chapels of the
304
A. Dal Cin, S. Russo / Engineering Failure Analysis 45 (2014) 300–313
Fig. 6. View of the damage in the south corner (a) and the north corner (b) between the sidewalls and the façade. Cracks 1, 2 in corner left and 3 in corner right.
Fig. 7. View of the south chapels before the earthquake (a) and view of south (b) and north chapels (c) after earthquake
longitudinal wall of the nave it is visible, as a matter of fact there are deep vertical cracks that propagate along the perimeter of the connection. About the transept, observing the outer wall of the north side (Fig. 8a) some lesions are found, whose are propagated by diatones stone (more stiffness of bricks) present in the masonry and indicate a vertical detachment of the bottom wall from the side. In the outer wall side of the south transept (Fig. 8b) is evident a lesion deep diagonal and the loss of material. The apse has two deep lesions on the trabeation below the opening of the back wall and the lesion of the right continues so severe for almost half of the total height (Fig. 8c). On the top of this wall there is a clear detachment of material caused by
Fig. 8. View of damage of the external wall of the north (a) and south transept (b) and view of damage of the back wall (c) and south sidewall (d) of the apse.
A. Dal Cin, S. Russo / Engineering Failure Analysis 45 (2014) 300–313
305
the collapse of the vault. The side walls of the apse (Fig. 8d) have two diagonals deep lesions. In this case the not continuous presence of the attached side, has not prevented the activation of an overturning mechanism. 4. Finite element modal analysis A Finite Element Model (FEM) of the church and of its coeval annex was elaborated with the DIANA computer program [21], the detailed geometric model is based on the geometrical survey. The construction of the undamaged FE model, firstly, of the isolated church (WA) and then with the presence of its annex (A) was fundamental to simulate the dynamic and seismic behavior. The hypothesis of the later annex construction was considered in an additional FE model (Ai) with the inclusion of an interface between the church and the annex. Finally in the Ai_d FE model the interface of this last FE model configuration was not degraded with the change of the mechanical property of the mortar material but removing the continuity of the interface connections (Fig. 9). For the realization of the four models of the church and of the annex four-node quadrilateral shell elements according to the Mindlin–Reissner theory have been adopted including shear deformation already used in similar analysis [22,23]. These elements of the FE model are characterized at the base by perfect constraints and the beam element constitute the structure of roof trusses. In the optimized mesh of all the surface, an average size element of 0.35 m was considered. In the Ai and Ai_d FE model an interface element was included between the south walls of church and the annex (see the first column of the Fig. 10), assuming a minimum thickness of 0.02 m. Defined in this way the geometry, the model was first subjected to a dynamic modal analysis using the mechanical characteristics of masonry showed in Table 2, a ‘‘masonry bricks and lime mortar’’ [24] was chosen as the material that identifies the whole church. In Table 2 are also showed all the mechanical parameters adopted; the highest value of the elastic modulus, equal to 1800 N/mm2 proposed by rules has been adopted since the masonry appears in good condition and also for the shear elastic modulus has been assumed the highest value characterizing the orthotropic behavior of the masonry. The other values are the Poisson’s ratio of 0.15 and the density of 18 kN/m3. The mechanical characteristics of the interface element between the different walls was assumed equal to those of a good mortar, the stiffness moduli equal to 170 N/mm2 and considering the tensile strength equal to zero. Through dynamic modal analysis, performing a free vibration eigenvalue analysis with DIANA solver, the main modes of vibration of the four configuration church-annex have been identified and compared (Fig. 10). A relevant remark is the dispersion of the mode shapes and the highest participating mass calculated is equal to 42.40% in the X axis and 40.00% in Y axis for the WA and even lower in the FE model of the church with the presence of the annex. The results of model Ai presents the minor participating masses. Observing the mode shapes of the A, Ai and Ai_d (Fig. 10) it is clear that the shapes mainly involve the macro-element free from the presence of the annex. In particular in the mode shape 4 the side walls nave presents the main shape in the free side unlike the uniform behavior of these walls of the isolated church. The mode shapes in the translational X (TX) direction suffer the greatest changes due to the presence of the annex in the same axis while the modes characterize by the displacements in the longitudinal direction seem not to undergo remarkable variations. For this reason Fig. 11 focuses the attention on the mode shapes with the greatest participation mass in transversal direction. The first mode shape in TX direction regards in particular the transept and the apse, due to the presence of the steel ties and the presence of the lateral chapels (see Section 2) that make the wall’s nave stiffer. The macro-elements in the WA model shift in a symmetrical way differently to the behavior of the other configurations. It is visible that the south transept of the A model shows the lowest displacement and that the back wall of the apse presents an out of the plan mechanism unlike the behavior of this part in the other models. The frequency of A in this specific mode presents an increase of 27% compared to WA and of 10% compared to Ai and Ai_d. Furthermore, the increase of Meff in WA model is almost 50% compared to other models. The most evident difference between the two models with the inclusion of the interface (Ai and Ai_d) is the behavior of the south side nave that shows biggest displacement in the Ai_d model (Fig. 11).
Fig. 9. Detail of the connection in the A, Ai and Ai_d FE model configurations.
306
A. Dal Cin, S. Russo / Engineering Failure Analysis 45 (2014) 300–313
Fig. 10. Mode shapes obtained by the FE dynamic modal analysis.
Table 2 Mechanical characteristics used in the FE model. Elements
Material
E (N/mm2)
G (N/mm2)
v
q (KN/m3)
Macroelements Annex Interface Trusses
Masonry Masonry Mortar Wood
1800 1800 170 15000
600 600 – –
0.15 0.15 0.15 0
18 18 20 7.5
A. Dal Cin, S. Russo / Engineering Failure Analysis 45 (2014) 300–313
307
Fig. 11. Principal mode shapes in TX direction.
Observing the façade (Fig. 10), it can be seen that the mode shapes are very similar to those of the WA model, presenting however lower displacements in real condition simulated in the model A. From the analysis of the frequencies of the first principal modes, that shift the façade, it is clear that the A model presents the highest result with a slow difference compared to the mode shapes in TY direction of the Ai and Ai_d model. Nevertheless, even in TY direction, observing the frequencies of A with respect to WA emerges an increase of 4.5% that reveals once again the increase of stiffness of the configuration with additional mass. In the model A the structure of the annex (particularly floors and roof) directly affects the structure of the church, thus leading to greater interaction between the two buildings. In the mode shapes of the TX direction (Fig. 11) there is a moderate difference apparently because the structure of the annex in Ai and more in the Ai_d behave independently and therefore more similar to the behavior of the model WA. The comparison between the overall frequencies of the WA and A model presents differences equal to 15%. 5. Response spectrum analysis The accelerations recorded on 29 May 2012 by the accelerometric station of Mirandola (MRN station) extracted from the ITalian ACcelerometric Archive (ITACA) [25], was used to calculate the time history of each seismic component. A modal dynamic analysis was carried out by using the response spectrum of the two horizontal seismic component (NS and EW).
308
A. Dal Cin, S. Russo / Engineering Failure Analysis 45 (2014) 300–313
Fig. 12. Plan of the church with the earthquake’s acceleration directions (a), recorded accelerograms (b) and pseudo-acceleration response spectra (c).
Table 1 Principal values of the two main shock earthquakes. Station
Seismic event
Magnitude Ml
Depth (km)
Distance epicenter (km)
Channel
PGA (cm/s2)
Mirandola MRN
May 29, 2012
5.8
9.6
4.0
NS EW UP
290 220 900
Observing the horizontal components of motion (Fig. 12 and Table 1), the most interesting value is the N-S component that is the most severe (PGA = 290 cm/s2). However the highest value of PGA equal to 900 cm/s2 is recorded by the vertical component caused by the important manifestation of the primary wave P, due by the small distance from the epicenter. The horizontal components N-S and E-W are not perfectly aligned at the coordinate axes of the plan of the church, however it can be affirmed that, the serious horizontal N-S component, invests the church orthogonally to its longer side that lies along the main axis x (Fig. 12). The FE model, previously used to perform the modal dynamic analysis, is subsequently subjected to a base acceleration übase = 1 m/s2 in the horizontal X direction and then in Y direction. To perform the response spectrum analysis the earthquake spectrum was added to the model. The frequencies and the corresponding load amplification factors for the base excitation load have been used. The spectral accelerations were combined using the Square Root of the Sum of the Squares (SRSS) method. Fig. 13 presents the results of all different FE model configurations of the absolute maximum displacement obtained by the sum of the all component of displacement through the modal dynamic analyses carried out by using the response spectrum of the two horizontal seismic component (NS and EW). It is clear that the isolated church presents symmetrical displacements compared to the configurations of the church with the presence of annex. The A configuration presents again the lowest overall displacements. From the observation of the mode shapes is rather difficult to understand the difference in the mode shapes between the two configurations Ai and Ai_d due for the reason that the two configurations have the same mass and the only variation is the quantity of interfaces which material degradation of the connection is showed in Fig. 9. The differences are indeed also present between these two models and for this reason are presented subsequently in terms of displacements and base shear stresses. Observing the outcomes of the mode shapes resulting from the analysis with the NS seismic component is interesting to observe the back wall of the apse and façade: in the model WA these walls have a completely different behavior from the configurations with the presence of annex which show the displacement out of the plane and also a shift in the plane due to the additional lateral mass (Fig. 13). As already shown in the damage survey (see Section 3) the major cracks involved the connection between the façade and the side walls nave. For this reason the global displacements of these corners were investigated (Fig. 14a). Appears immediately clear the difference of the response’s results between the analysis with the spectral component with the NS and EW. In the second one (Fig. 14c and e) the presence of the annex on the south side of the church allows the façade to move independently and presents some difference between the four model configuration. The perfect connection of the A model leads to present the lowest displacements in each corner proving the benefit of the presence of the annex. The perfect interface connection Ai shows a moderate difference (an average equal to 4%) compare with the Ai_d model. Comparing these graphs it is interesting the increase of the displacements in the corner 2 in Ai and Ai_d model suggesting a rotational mechanism of the façade. On the other hand, looking at the results of NS component analysis (Fig. 14b and d) there is a homogenization of the displacements in the A, Ai, Ai_d configuration, in particular in the corner 1 and an important increase of the values in the WA model. The annex that lies in the same axis of the NS component greatly restricts the displacements. It is also visible the variation, the increase of the displacements after the annex’s quota.
A. Dal Cin, S. Russo / Engineering Failure Analysis 45 (2014) 300–313
309
Fig. 13. Displacements obtained by the spectrum response analysis with NS and EW seismic components.
It was also necessary to investigate the seismic base shear. As a matter of fact, were analysed the values of the base shear stress first of the global model and then of the distribution of these stresses in the different macro-elements (Fig. 15). The results of the base shear of the global model and of each macro-element (Vtot and VME) are normalized to the total weight (Wtot and WME). From the data analysis it is possible to provide some first observations. The difference between WA and the global model with the presence of the annex A and Ai is equal to 24% in the seismic transversal direction and to 11% in the longitudinal direction. From a first quick analysis of the macro-elements graphs is evident that the values of the longitudinal direction analysis showing lowest differences between the all configurations compared to the results obtained from the transversal direction analysis. It is interesting to note that the Ai_d presents a global results more similar to the WA model suggesting that the degradation of the mortar leads to a more independent behavior of the church.
310
A. Dal Cin, S. Russo / Engineering Failure Analysis 45 (2014) 300–313
Fig. 14. Corners façade-nave 1 and 2 of the church (a), displacements of these points obtained in the NS (b and d) and in the EW (c and e) response spectrum analysis.
Fig. 15. Base shear of the global model and of each macro element in transversal (a) and longitudinal (b) directions.
A. Dal Cin, S. Russo / Engineering Failure Analysis 45 (2014) 300–313
311
The macro-elements analysis of the first spectra combination (Fig. 15a) shows the highest differences between WA and A– Ai, in particular for the north side nave (NS_N) and for the north transept (TR_N). Also in the global analysis with the second spectra combination (Fig. 15b) the macro-elements analysis presents high differences between the same models. These results concerning the north side of the church in the transverse direction analysis indicate that the presence of the annex in the opposite side is certainly a negative factor leading to an average increase of the base shear stress equal to 40%. Also the apse that has the presence of the annex in the all side wall shows a high difference compared with the isolated church (WA) in particular in the transversal direction. On the other hand the façade (F), the south transept (TR_S) and the south side nave (NS_S) present in the transversal direction analysis similar values for all global model configuration. Observing these macroelements in the longitudinal analysis, the WA model presents a modest increment of the values compared to the A and Ai model. In this case the presence of the annex seems to give a positive contribution. The variation of the results between the A and Ai model configurations is moderate. The macro-element NS_S in the transversal direction analysis and the macro-element TR_S in the longitudinal direction analysis present the main differences.
6. Comparison between FEA and effective damage The damage survey carried out after the earthquake allows us to demonstrate the reliability of the FE analysis results previously obtained. The façade presents similar results for all four model configurations. In the isolated configuration (WA) the shear base stresses results, both in the analysis with the prevailing longitudinal component that with the transversal one, are higher (10%) compared to those of the church with the annex presence. In addition, the mode shapes of the façade in all the model configurations (Fig. 10) present an out of plane mechanism that confirms an independent behavior. The presence of the annex involves a slight benefit to the base shear stress of the façade but does not prevent its prevailing out of plane structural behavior. The overturning of this rigid wall is the most likely local collapse mechanism in masonry structures [26]. These results are compatible with the damage observation (Fig. 6) that show deeper cracks on the both corner connections with the side nave walls. The seismic damage survey of the north side nave and of the chapels presents a severe crack pattern (Fig. 7b). Also in this case the modal analysis results support the effective damage scenario. As a matter of fact, in particular in the A model configurations, that simulate the real condition of the church-annex connection, both the mode shapes that the base shear stress of the NS_N are significantly greater than the NS_S. This difference is more evident compared with the symmetric behavior of same macro-element in the WA model. Finally, the damage survey of the apse suggest an overturning of all the macro-element. Both the modal analysis that the NS seismic combination show a different behavior of the apse in the A model compared with the other model configurations. As a matter of fact, the mode shapes of the back wall presents a mechanism out of the plane.
7. Discussion and generalization of the structural role of annex All that was explained above showed a strong relevance of the annexes’ presence to historic buildings subjected to earthquake compared to those isolated. Although the first part of the paper highlights the different types of annex and the impossibility address all of them in a single work due to the complex of the matter, the case of the Gesù Church allows to make some initial considerations on the structural role of the annex during the seismic action. The response spectrum analysis have evidenced the strong vulnerability of all the configurations presented in the transversal direction, in particular in the church-annex model. With reference to the lateral annex investigated and the orientation of the seismic vector that generated the damage, it is believed that the macro-elements contiguous to the annex are safe. However, the wall of the nave and chapels at the opposite side appear much more damaged than they could be without the presence of the annex, as confirmed by the numerical analysis. Nonetheless, it is possible to state that, wherever the annex is on one side of the church and is slightly lower in height, its influence is mainly beneficial. It is possible to say that the small size of the annex and good connection to church’s walls have also prevented the typical pounding phenomena. Besides highlighted by the present research, more favorable outcomes were obtained from the model with rigid connections between the church and annex. From the comparison to another church damaged by the earthquake of L’Aquila – i.e. Santa Maria del Suffragio [14,15], very similar to the object of this research, in term of annex – it was noticed that the presence of a lower and smaller annex has not led significant differences on the damage pattern among all the macro-elements of the church. Moreover, also in this case the annex has not prevented the overturning of the façade – which is very high and massive than the church here investigated. As a matter of fact, the most common collapse mechanism is the overturning of this masonry wall; the presence of annex, which in most cases is located at the sides, does not prevent the activation of this mechanism. This mechanism regards also the back wall of the apse that in the Gesù Church was prevented due to the contiguity of annex for the whole height of the macro-element (Fig. 3). Differently, the damage survey of the Santa Maria del Suffragio church presents a serious detachment from the lateral walls confirming the negligible presence of the small annex.
312
A. Dal Cin, S. Russo / Engineering Failure Analysis 45 (2014) 300–313
8. Conclusion The case study of the Gesù Church hit by the Mirandola earthquake in 2012 was useful to comment the influence of the annexes in the structural dynamic behavior of churches. The observation of the damage scenario suggested this research that, with the support of a consciously restricted FE model, allows to present some results about a very complex item. The dynamic modal analysis and the use of the two horizontal seismic combinations of the MRN spectra were used. The following final consideration can thus be proposed: – the presence of the annex on only one side of the church brings a benefit to the base shear stress at the connected side, but also an increase at the opposite side; as a matter of fact, the different structural responses of the two side walls of the nave highlight the highest vulnerability at the free side; – the seismic combination analysis shows a more significant weakness along the transversal direction, concerning in particular the nave’s and transept’s unconstrained sides; – the annex has increased the total seismic mass and this addition during the seismic action has affected particularly the northern side of the church; the prevailing seismic component of the earthquake, that was almost aligned with the transverse axis of the church, turns out to be detrimental for the macro-elements with a free side; – the differences in the mode shapes between the two configurations – i.e. with and without the inclusion of the interface – is visually of little relevance. However, the analysis of the base shear stresses shows that the results of the model with the deterioration of the interface get closer to those of the isolated church; – concerning the mode shapes that have modal participating mass ratio (Meff) in longitudinal direction and the mode shapes of the response spectrum analysis EW, the façade presents a similar response for all the FE model configuration proposed. This macro-element presents the most significant detachment from the side walls of the nave. Consequently, it tends to become structurally independent from the global behavior of the church. The analysis shows that the annex has not substantially prevented the activation of the façade’s predominant out-of-plane behavior; – the obtained results support, as expected, the hypothesis that the particular height and position of the annex, in this case, respectively moderate and asymmetrical and the seismic vector play an important role in the seismic risk evaluation for the historic churches.
Acknowledgment Financial support by the research Project AQ DPC/ReLUIS 2014-2018_UR IUAV-DPPAC_RS 4_ ‘‘Seismic Observatory of Structures & Monitoring’’ is gratefully acknowledged by the Authors. References [1] Doglioni F, Moretti A, Petrini V. Le Chiese e il terremoto. Trieste: Edizioni Lint; 1994 [in italian]. [2] Lagomarsino S, Podestà S. Damage and vulnerability assessment of the churches after the 2002 Molise, Italy. Earthquake, Earthquake Spectra 20, Earthquake Reconnaissance Report 2004. In: Bazzurro P, Maffei J, editors [Special Issue 1, 2002 Molise Italy]. [3] PCM-DPC-MiBAC M.-A.-D. Form for the survey of damage in cultural heritage –Churches. Scheda per il relievo del danno ai beni culturali – Chiese; 2011 [in Italian]. [4] Vicente RS, Parodi S, Lagomarsino S, Varum H, Mendes da Silva JAR. Seismic vulnerability and risk assessment: case study of the historic city centre of Coimbra, Portugal. Bull Earthq Eng 2011;9(4):1067–96. [5] Russo S. Testing and modelling of dynamic out-of-plane behaviour of the historic masonry façade of PalazzoDucale in Venice, Italy. Eng Struct. http:// dx.doi.org/10.1016/j.engstruct.2012.07.032. ISSN: 0141–0296, 2012. [6] Pau A, Vestroni F. Vibration assessment and structural monitoring of the Basilica of Maxentius in Rome. Mechanical Systems and Signal Processing 2013;41(1–2):454–66. [7] Brandonisio G, Lucibiello G, Mele E, De Luca A. De Luca, Damage and performance evaluation of masonry churches in the 2009 L’Aquila earthquake. Eng Fail Anal 2013. http://dx.doi.org/10.1016/j.engfailanal.2013.01.021. [8] Betti M, Vignoli A. Numerical assessment of the static and seismic behaviour of the Basilica of Santa Maria all’Impruneta (Italy). Constr Build Mater 2011;25(12):4308–24. [9] Lagomarsino S, Podestà S. Seismic vulnerability of ancient churches: I. Damage assessment and emergency planning. Earthq Spectra 2004;20:377–94. [10] Boscato G, Dal Cin A, Russo S, Sciarretta F. SHM of historic damaged churches. Adv Mater Res 2014;838–841:2071–8. [11] Boscato G, Pizzolato M, Russo S, Tralli A. Seismic behavior of a complex historical church in L’Aquila. Int J Architec Heritage 2014;8(5):718–57. [12] Milani G, Russo S, Pizzolato M, Tralli A. Seismic Behavior of the San Pietro di Coppito Church Bell Tower in L’Aquila, Italy. Open Civil Eng J 2012;6(SPEC.ISS.1):131–47. [13] Russo, S. Performance of a PFRP structure covering a historic building struck by an earthquake. In: Advanced Composites in Construction, ACIC proceedings of the 5th international conference, 2011. p. 51–57. [14] Russo S. On the monitoring of historic Anime Sante church damaged by earthquake in L’Aquila. Structural Control and Health Monitoring. ISSN: 1545– 2263, 2012http://dx.doi.org/10.1002/stc.1531. [15] Boscato G, Rocchi D, Russo S. Anime Sante Church’s Dome After 2009 L’Aquila Earthquake. Monitoring and Strengthening Approaches, Adv Mater Res 2012;446–449:3467–85, ISSN: 1022–6680, doi: 10.4028/www.scientific.net/AMR.446-449.3467. [16] Gattulli V, Antonacci E, Vestroni F. Field Observations and Failure Analysis of the Basilica S. Maria di Collemaggio after the 2009 L’Aquila Earthquake. Eng Fail Anal 2013. http://dx.doi.org/10.1016/j.engfailanal.2013.01.020. [17] Mele E, De Luca A, Giordano A. Modelling and analysis of a Basilica under earthquake loading. J Cult Heritage 2003;4(4):355–67. [18] Di Giulio G, Azzara RM, Cultrera G, Giammarinaro MS, Vallone P, Rovelli A, Effect of local geology on ground motion in the city of Palermo, Italy, as inferred from aftershocks of the 6 September 2002 Mw 5.9 earthquake. Bull Seism Soc Am 2005;95:2328–41 [19] D’Ayala DF, Paganoni S. Assessment and analysis of damage in L’Aquila historic city centre after 6th April 2009. Bull Earthquake Eng 2011;9:81–104.
A. Dal Cin, S. Russo / Engineering Failure Analysis 45 (2014) 300–313
313
[20] Casalegno C, Cecchi A, Reccia E, Russo S. Heterogeneous and continuous models: comparative analysis to investigate masonry wall subjected to differential settlements. Composites: mechanics, computations, applications. Int J 2013;4(3):187–207. [21] DIANA User’s Manual Release 9.4.2. 2010. Delft, TNO. [22] Lourenço PB, Trujillo A, Mendes N, Ramos LF. Seismic performance of the St. George of the Latins church: Lessons learned from studying masonry ruins. Eng Struct 2012;40:501–18. [23] Roca Pere, Cervera M, Gariup G, Pela’ L. Structural Analysis of Masonry Historical Constructions. Arch Comput Methods Eng 2010;17:299–325. http:// dx.doi.org/10.1007/s11831-010-9046-1. Classical and Advanced Approaches. [24] NTC 08. Decreto Ministeriale 14/01/08. ‘‘Norme tecniche per le costruzioni’’. Gazzetta Ufficiale No. 29, Ministero delle Infrastrutture; 2008 [in Italian]. [25] ITACA (the ITalian ACcelerometric Archive at http://itaca.mi.ingv.it/ItacaNet/). [26] Giuffrè A. Safety and conservation of historical centers: the case of Ortigia. Bari: Laterza; 1993 [in Italian].
Contents lists available at ScienceDirect
Engineering Failure Analysis journal homepage: www.elsevier.com/locate/engfailanal
Influence of the annex on seismic behavior of historic churches A. Dal Cin ⇑, S. Russo IUAV University of Venice, Dept. of Design and Planning in Complex Environments, Dorsoduro 2206, 30123 Venice, Italy
a r t i c l e
i n f o
Article history: Received 21 January 2014 Received in revised form 3 July 2014 Accepted 6 July 2014 Available online 16 July 2014 Keywords: Historical church Annex Interface Damage identification Modal analysis
a b s t r a c t In May 2012, two major earthquakes occurred in Emilia Romagna region in Northern Italy, causing widespread damage. The hypocentre of the second one, strokes Mirandola where is located the Gesù Church investigated in this research. The church has a long and important annex to the south built during the same period of the church. This paper addresses how the important annex influenced the seismic response of this historical church and how, more generally, this kind of asymmetric mass can influence the behavior of historic churches. The final considerations are based on the comparison between the structural damage pattern survey and modal and seismic FE analysis. A FE model was constructed considering four different configurations: (i) isolated church, (ii) the church with the presence of the real annex with a perfect connection, (iii) the church with the presence of the same annex but with an interface between the church and the annex and (iv) this last configuration with the stiffness degradation of the interface. Firstly the dynamic modal analysis and subsequently the seismic spectral analysis were performed. The results indicate that the annex’s presences play a significant role in the dynamic response of the church and affect the distribution of damages for the whole building. The results of the seismic simulation are in agreement with the observed damage. Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction The recent earthquakes that hit Italy in the last century proved the high vulnerability of cultural heritage, with particular reference to churches. These particular monumental buildings cannot be reduced to any standard structural scheme and this makes it difficult to evaluate their seismic reliability. To overcome this problem, the macro-element approach has been proposed a few years ago and since then repeatedly used to recognizing the collapse mechanisms in the different macroelements of the church [1,2]. The common collapsing configurations are shown in PCM-DPC-MiBAC M.-A.-D. [3]. The historical centers often occur as the result of an uncontrolled constructive evolution, whose complex configurations lead the structures to strongly interact with each other when are subjected to seismic action [4]. As a matter of fact, several historic churches are not isolated from the urban context but are often characterized by the presence of adjacent buildings, usually named annex (convents, sacristy, tower, minor constructions, etc.) at the same time of the church or subsequently constructed as we have seen in other case studies (Fig. 1). Besides the monitoring of historical constructions in seismic areas is a predominant issue also in Europe and especially in Italy because of the richness of its inestimable architectural heritage [5,6]. In the last century the seismic events stroked severely the cultural historic heritage and in particular the 49% of the damaged structures are churches highlighting their
⇑ Corresponding author. E-mail addresses: [email protected] (A. Dal Cin), [email protected] (S. Russo). http://dx.doi.org/10.1016/j.engfailanal.2014.07.004 1350-6307/Ó 2014 Elsevier Ltd. All rights reserved.
A. Dal Cin, S. Russo / Engineering Failure Analysis 45 (2014) 300–313
301
Fig. 1. Some examples of churches characterized by the presence of the annex and damaged by earthquakes.
intrinsic vulnerability [7–17] due first to the absence of any kind of diaphragm except the covering. Tuned with this issue it becomes necessary to take a census of the weakness of these monuments through the structural identification procedures and the evaluation of the related ground motion characteristics [18,19]. By the way the paper presents the case study of Gesù church in Mirandola, which was severely damaged by the strong earthquake that occurred in Emilia Romagna region in Northern Italy, in May 2012. The question of how the important annex had engraved on the dynamic behavior during the earthquake arise from the observation of an unsymmetrical damage survey. The construction of the undamaged FE model, firstly, of the isolated church (WA) and then with the presence of its annex (A) was fundamental to simulate the dynamic and seismic behavior. As previously said, the annex could have been built later than the church; this hypothesis is considered in an additional FE model (Ai) with the inclusion of an interface between the church and the annex. One of the main problems of the historic masonry buildings is certainly the materials deterioration, particularly of the mortars. The reason of the deterioration of this part is often due to the low quality of conservation and other exceptional events or new additional loads [20]. It is chosen therefore to simulate also this problem in the FE model with a deteriorated interface Ai_d. The dynamic modal analysis and the response spectrum analysis have been performed to determine the possible collapse mechanisms and to study the stresses distribution on each macro-element of the whole church analyzed. This case study highlights the importance of a deeper study on the incidence of the presence of an important annex on the dynamic response of the churches. The analysis of this case study achieves some results, which may be useful also for correctly driving future strengthening interventions on similar context.
302
A. Dal Cin, S. Russo / Engineering Failure Analysis 45 (2014) 300–313
2. Description of the case study The seventeenth-century Gesù Church (Fig. 2) was built on the orders of Alessandro I Pico on the occasion of his investiture as Duke of Mirandola. The church has a typical basilica plan, indeed it is characterize by a wide single nave joined on either side by side chapels separated by Corinthian fluted pillars and on the right side to border with the transept there is an entirely wooden pulpit. This is also an important transept characterized by two extraordinary wooden altars in the walls of back and majestic frames; the church ends with an apse regular where the main altar lies. The transept and the apse are connected on both sides by two volumes of two levels, the sacristy, probably dating on the same period of the church building (Figs. 2c and 3). All structural entities of the church are covered with a cross and barrel vaulting that are undecorated. As it can be seen in Fig. 3a the church has a long and important annex to the south, the College of Gesuiti. This building was opened in 1690 a ‘year later of the completion of the church, has two floors and a classic style. The exterior looks rather poorer in particular the façade remained unfinished (Fig. 2b). All external walls are brick surface that makes legible all the renovations that took place over time. The roof of the building was restored after years of work in 1998. It is possible to see the intervention with steel brackets focused to the node strut-chain in the support of the truss to the wall. The bracket of the beam has been applied to steel tie which acts in cooperation with the chain itself. The church can be divided into different macro-element: the nave, the side chapels, the transept, the apse and the façade (Fig. 3c). It presents a large nave joined on both sides by the chapels of modest size, the façade has a very regular plant and prospects, the transepts and the apse have significant dimensions both in plan and in height. The plant shows a longitudinal length along the main axis x that is equal to 32 m and a width along the y axis of 18 m. The height of the Church reaches 22 m.
3. Damage survey The Emilia-Romagna has been subject to an important seismic sequence in 2012 that hit the region lead to severe damage to the historical and cultural heritage. The first serious earthquake, with a magnitude of 5.9 ML was recorded May 20, 2012 with its epicenter in Finale Emilia, at a depth of 6.3 km. Another powerful quake of magnitude 5.8 ML and depth of hypocenter 9.6 km was recorded on 29 May 2012. In Ferrara area, the main damage is caused by the earthquake of 20 May, instead in the most western part of the Modena (between Mirandola, Carpi, Novi di Modena, Cavezzo and Concordia) the most serious effects are certainly due to the earthquake of 29 May, mainly due by the proximity to the epicenter. The earthquake has severely affected the structure, in particular the most visible damage is the collapse of the vaults of the church along the main axis x (Fig. 4). The nave was characterized by cross vaults and the apse by barrel vaults. The parties collapsed give the possibility to identify the type of texture characterized by bricks placed in horizontal position that identifies indeed a very thin structure element. The only cover parts that have survived are the arches that marked the vaults where the bricks placed in vertical position and where we observed the presence of metal chains (Fig. 4). The geometry of the thin vaults and their static behavior predominantly in compression have proved to be vulnerable also in response to the severe amplitude of the vertical component. The roof trusses have well reacted to the seismic activity due to the consolidation work carried out recently while the wooden roof suffered a collapse in the attached part of the façade (Fig. 5). The inner part of the façade is seriously damaged (Fig. 5b and d) especially the level above the trabeation for the loss of structural stiffness and probably due to a reduced sectional dimension than the first level. The crack most serious propagates from the center of the top of the architrave of the door affecting the central opening that is also a point of vulnerability and from the trabeation and the arms at the center. The top of the façade is a very weak point of the macro-element caused by the poor connection with the cover.
Fig. 2. Church of Gesù, view of north later part (a), the façade (b) and the apse (c).
A. Dal Cin, S. Russo / Engineering Failure Analysis 45 (2014) 300–313
303
Fig. 3. View of the church and the annex (a), detail of the annex (b) and macro-elements plan (c).
Fig. 4. View of the vaults before (a) and after the earthquake (b).
Fig. 5. View of the external and internal façade before (a and c) and after the earthquake (b and d).
Observing Fig. 6, the masonry façade undergoes a detachment from the side walls, in particular in the second level in height of the church; analyzing carefully the corner south is evident that the lesion in corner, characterized by a slipping of the bricks, affects in particularly the top part and propagates downwards up to damage the entire trabeation and the capital of the pillar corner (Fig. 6a, detail 1). In proximity on the corner south is also detected a deeper crack that propagates from the upper part of a window to the frame of the trabeation below (Fig. 6a, detail 2). It is probably due to the presence of the adjacent system support bell, also the trabeation of the north corner are cracked at the same point (Fig. 6b, detail 3). Fig. 7 shows the south side chapels before (Fig. 7a) and after the earthquake. In the general view of the damage, the side chapels in the south of the nave (Fig. 7b) present any serious damage for the presence of annexes on the south side of the church that helped to have a greater stiffness in the first level in height. However, the side chapels to the north (Fig. 7c) have a severe crack pattern and the presence of deep diagonal lesions along the transverse walls of the nave and in the barrel vaults which also present partial detachment of material. The activation of a detachment of the side chapels of the
304
A. Dal Cin, S. Russo / Engineering Failure Analysis 45 (2014) 300–313
Fig. 6. View of the damage in the south corner (a) and the north corner (b) between the sidewalls and the façade. Cracks 1, 2 in corner left and 3 in corner right.
Fig. 7. View of the south chapels before the earthquake (a) and view of south (b) and north chapels (c) after earthquake
longitudinal wall of the nave it is visible, as a matter of fact there are deep vertical cracks that propagate along the perimeter of the connection. About the transept, observing the outer wall of the north side (Fig. 8a) some lesions are found, whose are propagated by diatones stone (more stiffness of bricks) present in the masonry and indicate a vertical detachment of the bottom wall from the side. In the outer wall side of the south transept (Fig. 8b) is evident a lesion deep diagonal and the loss of material. The apse has two deep lesions on the trabeation below the opening of the back wall and the lesion of the right continues so severe for almost half of the total height (Fig. 8c). On the top of this wall there is a clear detachment of material caused by
Fig. 8. View of damage of the external wall of the north (a) and south transept (b) and view of damage of the back wall (c) and south sidewall (d) of the apse.
A. Dal Cin, S. Russo / Engineering Failure Analysis 45 (2014) 300–313
305
the collapse of the vault. The side walls of the apse (Fig. 8d) have two diagonals deep lesions. In this case the not continuous presence of the attached side, has not prevented the activation of an overturning mechanism. 4. Finite element modal analysis A Finite Element Model (FEM) of the church and of its coeval annex was elaborated with the DIANA computer program [21], the detailed geometric model is based on the geometrical survey. The construction of the undamaged FE model, firstly, of the isolated church (WA) and then with the presence of its annex (A) was fundamental to simulate the dynamic and seismic behavior. The hypothesis of the later annex construction was considered in an additional FE model (Ai) with the inclusion of an interface between the church and the annex. Finally in the Ai_d FE model the interface of this last FE model configuration was not degraded with the change of the mechanical property of the mortar material but removing the continuity of the interface connections (Fig. 9). For the realization of the four models of the church and of the annex four-node quadrilateral shell elements according to the Mindlin–Reissner theory have been adopted including shear deformation already used in similar analysis [22,23]. These elements of the FE model are characterized at the base by perfect constraints and the beam element constitute the structure of roof trusses. In the optimized mesh of all the surface, an average size element of 0.35 m was considered. In the Ai and Ai_d FE model an interface element was included between the south walls of church and the annex (see the first column of the Fig. 10), assuming a minimum thickness of 0.02 m. Defined in this way the geometry, the model was first subjected to a dynamic modal analysis using the mechanical characteristics of masonry showed in Table 2, a ‘‘masonry bricks and lime mortar’’ [24] was chosen as the material that identifies the whole church. In Table 2 are also showed all the mechanical parameters adopted; the highest value of the elastic modulus, equal to 1800 N/mm2 proposed by rules has been adopted since the masonry appears in good condition and also for the shear elastic modulus has been assumed the highest value characterizing the orthotropic behavior of the masonry. The other values are the Poisson’s ratio of 0.15 and the density of 18 kN/m3. The mechanical characteristics of the interface element between the different walls was assumed equal to those of a good mortar, the stiffness moduli equal to 170 N/mm2 and considering the tensile strength equal to zero. Through dynamic modal analysis, performing a free vibration eigenvalue analysis with DIANA solver, the main modes of vibration of the four configuration church-annex have been identified and compared (Fig. 10). A relevant remark is the dispersion of the mode shapes and the highest participating mass calculated is equal to 42.40% in the X axis and 40.00% in Y axis for the WA and even lower in the FE model of the church with the presence of the annex. The results of model Ai presents the minor participating masses. Observing the mode shapes of the A, Ai and Ai_d (Fig. 10) it is clear that the shapes mainly involve the macro-element free from the presence of the annex. In particular in the mode shape 4 the side walls nave presents the main shape in the free side unlike the uniform behavior of these walls of the isolated church. The mode shapes in the translational X (TX) direction suffer the greatest changes due to the presence of the annex in the same axis while the modes characterize by the displacements in the longitudinal direction seem not to undergo remarkable variations. For this reason Fig. 11 focuses the attention on the mode shapes with the greatest participation mass in transversal direction. The first mode shape in TX direction regards in particular the transept and the apse, due to the presence of the steel ties and the presence of the lateral chapels (see Section 2) that make the wall’s nave stiffer. The macro-elements in the WA model shift in a symmetrical way differently to the behavior of the other configurations. It is visible that the south transept of the A model shows the lowest displacement and that the back wall of the apse presents an out of the plan mechanism unlike the behavior of this part in the other models. The frequency of A in this specific mode presents an increase of 27% compared to WA and of 10% compared to Ai and Ai_d. Furthermore, the increase of Meff in WA model is almost 50% compared to other models. The most evident difference between the two models with the inclusion of the interface (Ai and Ai_d) is the behavior of the south side nave that shows biggest displacement in the Ai_d model (Fig. 11).
Fig. 9. Detail of the connection in the A, Ai and Ai_d FE model configurations.
306
A. Dal Cin, S. Russo / Engineering Failure Analysis 45 (2014) 300–313
Fig. 10. Mode shapes obtained by the FE dynamic modal analysis.
Table 2 Mechanical characteristics used in the FE model. Elements
Material
E (N/mm2)
G (N/mm2)
v
q (KN/m3)
Macroelements Annex Interface Trusses
Masonry Masonry Mortar Wood
1800 1800 170 15000
600 600 – –
0.15 0.15 0.15 0
18 18 20 7.5
A. Dal Cin, S. Russo / Engineering Failure Analysis 45 (2014) 300–313
307
Fig. 11. Principal mode shapes in TX direction.
Observing the façade (Fig. 10), it can be seen that the mode shapes are very similar to those of the WA model, presenting however lower displacements in real condition simulated in the model A. From the analysis of the frequencies of the first principal modes, that shift the façade, it is clear that the A model presents the highest result with a slow difference compared to the mode shapes in TY direction of the Ai and Ai_d model. Nevertheless, even in TY direction, observing the frequencies of A with respect to WA emerges an increase of 4.5% that reveals once again the increase of stiffness of the configuration with additional mass. In the model A the structure of the annex (particularly floors and roof) directly affects the structure of the church, thus leading to greater interaction between the two buildings. In the mode shapes of the TX direction (Fig. 11) there is a moderate difference apparently because the structure of the annex in Ai and more in the Ai_d behave independently and therefore more similar to the behavior of the model WA. The comparison between the overall frequencies of the WA and A model presents differences equal to 15%. 5. Response spectrum analysis The accelerations recorded on 29 May 2012 by the accelerometric station of Mirandola (MRN station) extracted from the ITalian ACcelerometric Archive (ITACA) [25], was used to calculate the time history of each seismic component. A modal dynamic analysis was carried out by using the response spectrum of the two horizontal seismic component (NS and EW).
308
A. Dal Cin, S. Russo / Engineering Failure Analysis 45 (2014) 300–313
Fig. 12. Plan of the church with the earthquake’s acceleration directions (a), recorded accelerograms (b) and pseudo-acceleration response spectra (c).
Table 1 Principal values of the two main shock earthquakes. Station
Seismic event
Magnitude Ml
Depth (km)
Distance epicenter (km)
Channel
PGA (cm/s2)
Mirandola MRN
May 29, 2012
5.8
9.6
4.0
NS EW UP
290 220 900
Observing the horizontal components of motion (Fig. 12 and Table 1), the most interesting value is the N-S component that is the most severe (PGA = 290 cm/s2). However the highest value of PGA equal to 900 cm/s2 is recorded by the vertical component caused by the important manifestation of the primary wave P, due by the small distance from the epicenter. The horizontal components N-S and E-W are not perfectly aligned at the coordinate axes of the plan of the church, however it can be affirmed that, the serious horizontal N-S component, invests the church orthogonally to its longer side that lies along the main axis x (Fig. 12). The FE model, previously used to perform the modal dynamic analysis, is subsequently subjected to a base acceleration übase = 1 m/s2 in the horizontal X direction and then in Y direction. To perform the response spectrum analysis the earthquake spectrum was added to the model. The frequencies and the corresponding load amplification factors for the base excitation load have been used. The spectral accelerations were combined using the Square Root of the Sum of the Squares (SRSS) method. Fig. 13 presents the results of all different FE model configurations of the absolute maximum displacement obtained by the sum of the all component of displacement through the modal dynamic analyses carried out by using the response spectrum of the two horizontal seismic component (NS and EW). It is clear that the isolated church presents symmetrical displacements compared to the configurations of the church with the presence of annex. The A configuration presents again the lowest overall displacements. From the observation of the mode shapes is rather difficult to understand the difference in the mode shapes between the two configurations Ai and Ai_d due for the reason that the two configurations have the same mass and the only variation is the quantity of interfaces which material degradation of the connection is showed in Fig. 9. The differences are indeed also present between these two models and for this reason are presented subsequently in terms of displacements and base shear stresses. Observing the outcomes of the mode shapes resulting from the analysis with the NS seismic component is interesting to observe the back wall of the apse and façade: in the model WA these walls have a completely different behavior from the configurations with the presence of annex which show the displacement out of the plane and also a shift in the plane due to the additional lateral mass (Fig. 13). As already shown in the damage survey (see Section 3) the major cracks involved the connection between the façade and the side walls nave. For this reason the global displacements of these corners were investigated (Fig. 14a). Appears immediately clear the difference of the response’s results between the analysis with the spectral component with the NS and EW. In the second one (Fig. 14c and e) the presence of the annex on the south side of the church allows the façade to move independently and presents some difference between the four model configuration. The perfect connection of the A model leads to present the lowest displacements in each corner proving the benefit of the presence of the annex. The perfect interface connection Ai shows a moderate difference (an average equal to 4%) compare with the Ai_d model. Comparing these graphs it is interesting the increase of the displacements in the corner 2 in Ai and Ai_d model suggesting a rotational mechanism of the façade. On the other hand, looking at the results of NS component analysis (Fig. 14b and d) there is a homogenization of the displacements in the A, Ai, Ai_d configuration, in particular in the corner 1 and an important increase of the values in the WA model. The annex that lies in the same axis of the NS component greatly restricts the displacements. It is also visible the variation, the increase of the displacements after the annex’s quota.
A. Dal Cin, S. Russo / Engineering Failure Analysis 45 (2014) 300–313
309
Fig. 13. Displacements obtained by the spectrum response analysis with NS and EW seismic components.
It was also necessary to investigate the seismic base shear. As a matter of fact, were analysed the values of the base shear stress first of the global model and then of the distribution of these stresses in the different macro-elements (Fig. 15). The results of the base shear of the global model and of each macro-element (Vtot and VME) are normalized to the total weight (Wtot and WME). From the data analysis it is possible to provide some first observations. The difference between WA and the global model with the presence of the annex A and Ai is equal to 24% in the seismic transversal direction and to 11% in the longitudinal direction. From a first quick analysis of the macro-elements graphs is evident that the values of the longitudinal direction analysis showing lowest differences between the all configurations compared to the results obtained from the transversal direction analysis. It is interesting to note that the Ai_d presents a global results more similar to the WA model suggesting that the degradation of the mortar leads to a more independent behavior of the church.
310
A. Dal Cin, S. Russo / Engineering Failure Analysis 45 (2014) 300–313
Fig. 14. Corners façade-nave 1 and 2 of the church (a), displacements of these points obtained in the NS (b and d) and in the EW (c and e) response spectrum analysis.
Fig. 15. Base shear of the global model and of each macro element in transversal (a) and longitudinal (b) directions.
A. Dal Cin, S. Russo / Engineering Failure Analysis 45 (2014) 300–313
311
The macro-elements analysis of the first spectra combination (Fig. 15a) shows the highest differences between WA and A– Ai, in particular for the north side nave (NS_N) and for the north transept (TR_N). Also in the global analysis with the second spectra combination (Fig. 15b) the macro-elements analysis presents high differences between the same models. These results concerning the north side of the church in the transverse direction analysis indicate that the presence of the annex in the opposite side is certainly a negative factor leading to an average increase of the base shear stress equal to 40%. Also the apse that has the presence of the annex in the all side wall shows a high difference compared with the isolated church (WA) in particular in the transversal direction. On the other hand the façade (F), the south transept (TR_S) and the south side nave (NS_S) present in the transversal direction analysis similar values for all global model configuration. Observing these macroelements in the longitudinal analysis, the WA model presents a modest increment of the values compared to the A and Ai model. In this case the presence of the annex seems to give a positive contribution. The variation of the results between the A and Ai model configurations is moderate. The macro-element NS_S in the transversal direction analysis and the macro-element TR_S in the longitudinal direction analysis present the main differences.
6. Comparison between FEA and effective damage The damage survey carried out after the earthquake allows us to demonstrate the reliability of the FE analysis results previously obtained. The façade presents similar results for all four model configurations. In the isolated configuration (WA) the shear base stresses results, both in the analysis with the prevailing longitudinal component that with the transversal one, are higher (10%) compared to those of the church with the annex presence. In addition, the mode shapes of the façade in all the model configurations (Fig. 10) present an out of plane mechanism that confirms an independent behavior. The presence of the annex involves a slight benefit to the base shear stress of the façade but does not prevent its prevailing out of plane structural behavior. The overturning of this rigid wall is the most likely local collapse mechanism in masonry structures [26]. These results are compatible with the damage observation (Fig. 6) that show deeper cracks on the both corner connections with the side nave walls. The seismic damage survey of the north side nave and of the chapels presents a severe crack pattern (Fig. 7b). Also in this case the modal analysis results support the effective damage scenario. As a matter of fact, in particular in the A model configurations, that simulate the real condition of the church-annex connection, both the mode shapes that the base shear stress of the NS_N are significantly greater than the NS_S. This difference is more evident compared with the symmetric behavior of same macro-element in the WA model. Finally, the damage survey of the apse suggest an overturning of all the macro-element. Both the modal analysis that the NS seismic combination show a different behavior of the apse in the A model compared with the other model configurations. As a matter of fact, the mode shapes of the back wall presents a mechanism out of the plane.
7. Discussion and generalization of the structural role of annex All that was explained above showed a strong relevance of the annexes’ presence to historic buildings subjected to earthquake compared to those isolated. Although the first part of the paper highlights the different types of annex and the impossibility address all of them in a single work due to the complex of the matter, the case of the Gesù Church allows to make some initial considerations on the structural role of the annex during the seismic action. The response spectrum analysis have evidenced the strong vulnerability of all the configurations presented in the transversal direction, in particular in the church-annex model. With reference to the lateral annex investigated and the orientation of the seismic vector that generated the damage, it is believed that the macro-elements contiguous to the annex are safe. However, the wall of the nave and chapels at the opposite side appear much more damaged than they could be without the presence of the annex, as confirmed by the numerical analysis. Nonetheless, it is possible to state that, wherever the annex is on one side of the church and is slightly lower in height, its influence is mainly beneficial. It is possible to say that the small size of the annex and good connection to church’s walls have also prevented the typical pounding phenomena. Besides highlighted by the present research, more favorable outcomes were obtained from the model with rigid connections between the church and annex. From the comparison to another church damaged by the earthquake of L’Aquila – i.e. Santa Maria del Suffragio [14,15], very similar to the object of this research, in term of annex – it was noticed that the presence of a lower and smaller annex has not led significant differences on the damage pattern among all the macro-elements of the church. Moreover, also in this case the annex has not prevented the overturning of the façade – which is very high and massive than the church here investigated. As a matter of fact, the most common collapse mechanism is the overturning of this masonry wall; the presence of annex, which in most cases is located at the sides, does not prevent the activation of this mechanism. This mechanism regards also the back wall of the apse that in the Gesù Church was prevented due to the contiguity of annex for the whole height of the macro-element (Fig. 3). Differently, the damage survey of the Santa Maria del Suffragio church presents a serious detachment from the lateral walls confirming the negligible presence of the small annex.
312
A. Dal Cin, S. Russo / Engineering Failure Analysis 45 (2014) 300–313
8. Conclusion The case study of the Gesù Church hit by the Mirandola earthquake in 2012 was useful to comment the influence of the annexes in the structural dynamic behavior of churches. The observation of the damage scenario suggested this research that, with the support of a consciously restricted FE model, allows to present some results about a very complex item. The dynamic modal analysis and the use of the two horizontal seismic combinations of the MRN spectra were used. The following final consideration can thus be proposed: – the presence of the annex on only one side of the church brings a benefit to the base shear stress at the connected side, but also an increase at the opposite side; as a matter of fact, the different structural responses of the two side walls of the nave highlight the highest vulnerability at the free side; – the seismic combination analysis shows a more significant weakness along the transversal direction, concerning in particular the nave’s and transept’s unconstrained sides; – the annex has increased the total seismic mass and this addition during the seismic action has affected particularly the northern side of the church; the prevailing seismic component of the earthquake, that was almost aligned with the transverse axis of the church, turns out to be detrimental for the macro-elements with a free side; – the differences in the mode shapes between the two configurations – i.e. with and without the inclusion of the interface – is visually of little relevance. However, the analysis of the base shear stresses shows that the results of the model with the deterioration of the interface get closer to those of the isolated church; – concerning the mode shapes that have modal participating mass ratio (Meff) in longitudinal direction and the mode shapes of the response spectrum analysis EW, the façade presents a similar response for all the FE model configuration proposed. This macro-element presents the most significant detachment from the side walls of the nave. Consequently, it tends to become structurally independent from the global behavior of the church. The analysis shows that the annex has not substantially prevented the activation of the façade’s predominant out-of-plane behavior; – the obtained results support, as expected, the hypothesis that the particular height and position of the annex, in this case, respectively moderate and asymmetrical and the seismic vector play an important role in the seismic risk evaluation for the historic churches.
Acknowledgment Financial support by the research Project AQ DPC/ReLUIS 2014-2018_UR IUAV-DPPAC_RS 4_ ‘‘Seismic Observatory of Structures & Monitoring’’ is gratefully acknowledged by the Authors. References [1] Doglioni F, Moretti A, Petrini V. Le Chiese e il terremoto. Trieste: Edizioni Lint; 1994 [in italian]. [2] Lagomarsino S, Podestà S. Damage and vulnerability assessment of the churches after the 2002 Molise, Italy. Earthquake, Earthquake Spectra 20, Earthquake Reconnaissance Report 2004. In: Bazzurro P, Maffei J, editors [Special Issue 1, 2002 Molise Italy]. [3] PCM-DPC-MiBAC M.-A.-D. Form for the survey of damage in cultural heritage –Churches. Scheda per il relievo del danno ai beni culturali – Chiese; 2011 [in Italian]. [4] Vicente RS, Parodi S, Lagomarsino S, Varum H, Mendes da Silva JAR. Seismic vulnerability and risk assessment: case study of the historic city centre of Coimbra, Portugal. Bull Earthq Eng 2011;9(4):1067–96. [5] Russo S. Testing and modelling of dynamic out-of-plane behaviour of the historic masonry façade of PalazzoDucale in Venice, Italy. Eng Struct. http:// dx.doi.org/10.1016/j.engstruct.2012.07.032. ISSN: 0141–0296, 2012. [6] Pau A, Vestroni F. Vibration assessment and structural monitoring of the Basilica of Maxentius in Rome. Mechanical Systems and Signal Processing 2013;41(1–2):454–66. [7] Brandonisio G, Lucibiello G, Mele E, De Luca A. De Luca, Damage and performance evaluation of masonry churches in the 2009 L’Aquila earthquake. Eng Fail Anal 2013. http://dx.doi.org/10.1016/j.engfailanal.2013.01.021. [8] Betti M, Vignoli A. Numerical assessment of the static and seismic behaviour of the Basilica of Santa Maria all’Impruneta (Italy). Constr Build Mater 2011;25(12):4308–24. [9] Lagomarsino S, Podestà S. Seismic vulnerability of ancient churches: I. Damage assessment and emergency planning. Earthq Spectra 2004;20:377–94. [10] Boscato G, Dal Cin A, Russo S, Sciarretta F. SHM of historic damaged churches. Adv Mater Res 2014;838–841:2071–8. [11] Boscato G, Pizzolato M, Russo S, Tralli A. Seismic behavior of a complex historical church in L’Aquila. Int J Architec Heritage 2014;8(5):718–57. [12] Milani G, Russo S, Pizzolato M, Tralli A. Seismic Behavior of the San Pietro di Coppito Church Bell Tower in L’Aquila, Italy. Open Civil Eng J 2012;6(SPEC.ISS.1):131–47. [13] Russo, S. Performance of a PFRP structure covering a historic building struck by an earthquake. In: Advanced Composites in Construction, ACIC proceedings of the 5th international conference, 2011. p. 51–57. [14] Russo S. On the monitoring of historic Anime Sante church damaged by earthquake in L’Aquila. Structural Control and Health Monitoring. ISSN: 1545– 2263, 2012http://dx.doi.org/10.1002/stc.1531. [15] Boscato G, Rocchi D, Russo S. Anime Sante Church’s Dome After 2009 L’Aquila Earthquake. Monitoring and Strengthening Approaches, Adv Mater Res 2012;446–449:3467–85, ISSN: 1022–6680, doi: 10.4028/www.scientific.net/AMR.446-449.3467. [16] Gattulli V, Antonacci E, Vestroni F. Field Observations and Failure Analysis of the Basilica S. Maria di Collemaggio after the 2009 L’Aquila Earthquake. Eng Fail Anal 2013. http://dx.doi.org/10.1016/j.engfailanal.2013.01.020. [17] Mele E, De Luca A, Giordano A. Modelling and analysis of a Basilica under earthquake loading. J Cult Heritage 2003;4(4):355–67. [18] Di Giulio G, Azzara RM, Cultrera G, Giammarinaro MS, Vallone P, Rovelli A, Effect of local geology on ground motion in the city of Palermo, Italy, as inferred from aftershocks of the 6 September 2002 Mw 5.9 earthquake. Bull Seism Soc Am 2005;95:2328–41 [19] D’Ayala DF, Paganoni S. Assessment and analysis of damage in L’Aquila historic city centre after 6th April 2009. Bull Earthquake Eng 2011;9:81–104.
A. Dal Cin, S. Russo / Engineering Failure Analysis 45 (2014) 300–313
313
[20] Casalegno C, Cecchi A, Reccia E, Russo S. Heterogeneous and continuous models: comparative analysis to investigate masonry wall subjected to differential settlements. Composites: mechanics, computations, applications. Int J 2013;4(3):187–207. [21] DIANA User’s Manual Release 9.4.2. 2010. Delft, TNO. [22] Lourenço PB, Trujillo A, Mendes N, Ramos LF. Seismic performance of the St. George of the Latins church: Lessons learned from studying masonry ruins. Eng Struct 2012;40:501–18. [23] Roca Pere, Cervera M, Gariup G, Pela’ L. Structural Analysis of Masonry Historical Constructions. Arch Comput Methods Eng 2010;17:299–325. http:// dx.doi.org/10.1007/s11831-010-9046-1. Classical and Advanced Approaches. [24] NTC 08. Decreto Ministeriale 14/01/08. ‘‘Norme tecniche per le costruzioni’’. Gazzetta Ufficiale No. 29, Ministero delle Infrastrutture; 2008 [in Italian]. [25] ITACA (the ITalian ACcelerometric Archive at http://itaca.mi.ingv.it/ItacaNet/). [26] Giuffrè A. Safety and conservation of historical centers: the case of Ortigia. Bari: Laterza; 1993 [in Italian].