Seismic Assessment of a Multi-storey

Masonry Building

N. Ademovi, M. Hrasnica
Faculty of Civil Engineering in Sarajevo, Bosnia and Herzegovina

ABSTRACT:
Seismic assessment of multi-storey unreinforced masonry (URM) buildings is described throughout the paper. A
typical building for massive construction during the 50s and 60s in Western Balkans and through the region was
chosen for the study.

The vulnerability of this building lies in its height, having load bearing walls mostly in only one direction, no
vertical R.C. confining elements and no seismic rules have been applied. According to EMS-98 it belongs to the
unconfined masonry, younger than approx. 60 years with reinforced concrete floors and for the 7 th degree of
seismic intensity moderate to heavy damages (grade 3-4) could be expected. Numerical modelling has been done
in two different software packages utilizing pushover and time history analyses. Several comparisons have been
made. Results were found to be in a very good correlation. As a result possible ways of strengthening have been
proposed.

Keywords: nonlinear behaviour of masonry, Pushover, Time History Analysis, finite element method, equivalent
frame method

1. INTRODUCTION

The assessment of existing masonry buildings is often a complex issue due to the high degree of
uncertainties regarding geometry, typology and mechanical properties of materials, load history, soil
characteristics and so on. Gathering as much information as possible before starting the study is
required.

The paper shows the assessment of a typical multi-storey masonry building in Bosnia and Herzegovina
built in the 60s designed without any seismic guidelines one of many similar structures spread
throughout Western Balkans. Assessment was done by utilizing the approach proposed by ICOMOS.
In the last decades several strong earthquakes struck this region and underlined the importance of
seismic assessment including evaluation of possible strengthening and retrofit measures. After the
earthquake in Skopje in 1963, first seismic codes were published and vertical confining R.C. elements
were introduced in masonry building.

Nonlinear behavior of masonry was taken into account during the calculations. Pushover Analysis
according to the Eurocode 8 and Time History Analyses were conducted. Several comparisons have
been made. On the basis of the results possible ways of strengthening have been proposed.

2. MULTI-STOREY MASONRY BUILDING

2.1. Structure and its geometry

The examined building is located in Sarajevo; see Figure 1 (a).

 (a) (b)
Figure 1. Multi-storey building, built in the year of 1957: (a) photo of the present building; (b) floor plan.

In plan, the structure is of dimensions 38.0m by 13.0m with 7 levels (basement + ground floor + 5
storeys). The plan view illustrated in Figure 1 (b) shows the layout of the structure with load bearing
walls mainly in one direction with slabs made of semi-prefabricated concrete elements. Faade walls are
made of solid brick masonry walls of 0.25m thick (bearing part) and hollow bricks of 0.125m thick (non
bearing part), while the inner bearing walls are solid brick walls of 0.25m thick. Bricks are of standard
dimensions 25x12x6.5cm, connected by cement mortar. The slabs are made out of "Herbst" concrete
hollow elements. Construction of these blocks was regulated by ex-Yugoslavian standards
B.D1.030-1965 (Peuli, 2002). The basement walls are made out of concrete. Inner perpendicular walls
are 0.38m thick, while the outer (longitudinal) walls are 0.30m thick, and two inner walls are 0.25m
thick.

3. NUMERICAL RESULTS

Modeling of masonry structures is not an easy task. The main difficulties were attributed to the nonlinear
behavior of masonry material and inadequate experimental information regarding mechanical properties
of structural elements. The structure has been modeled resorting two modeling strategies, Finite
Element Method (FEM) utilizing the software package DIANA 9.4 (DIANA 9.4, 2009), and the second
in 3MURI Program, which uses the Frame by Macro Elements (FME) method (S.T.A.-DATA, 2010).

2.1. Finite Element Method

The structure was modeled by curved shell quadrilateral element CQ40S type. Figure 2 (a) shows a final
3D numerical model consisting of 84523 nodes and 28522 CQ40S elements. Due to a very large number
of elements and limited time and the fact that the structure is symmetric, half of the structure was
considered. Adequate boundary conditions were used as shown in Figure 2(b).

Physical non-linear behavior of the masonry walls is defined through the adoption of the total strain
fixed crack model, detailed in DIANA (DIANA 9.4, 2009). For hysteretic behavior of masonry a
parabolic stress-strain relation for compression, based on Hill-type yield criterion, was chosen with no
lateral confinement and no lateral crack reduction. Tension path, based on Rankine-type yield criterion,
was described by an exponential tension softening diagram. Finally, constant shear retention was chosen
see Figure 3. For more details see (Ademovi, 2011, Ademovi, 2012).
 Z

(a) (b)
Figure 2. (a) Numerical model performed in DIANA and (b) Constrains for half of the model

Figure 3. Hysteretic Behavior of Masonry (Mendes, et al., 2009).

2.1. Frame by Macro Elements

The model (input simplified) having the same characteristics as previously, consists of 7 levels, 218 - 3D
nodes, 34 - 2D nodes, and 506 elements. The view of the structure and elements is presented in Figure 4.

Figure 4. View of the structure and its elements in 3MURI.

2.1.1. Comparison of results obtained by two models

The maximum difference for the first three modes is 19.2% while for mass participation is 9.2%. The
mode shapes obtained by both models have the same configuration. Therefore, it can be concluded that
a good correlation between the two calculations has been obtained. For more info see (Ademovi, 2011)
and (Ademovi et. al., 2013).

Here only the "+ Y" direction is shown. The horizontal load was applied in a stepwise manner
proportional to its mass. A control node was chosen in the line of symmetry at the roof level, node no.
44014, as shown in Figure 5 (a).
 (a) (b)
Figure 5. (a) Location of the nodes; wall labeling; and direction of the horizontal force (b) Capacity Curves:
DIANA vs. 3MURI in "+Y" direction

The capacity curves for the "+Y" direction obtained by Finite Element Method and Frame by Macro
Elements are compared in Figure 5(b). Once the maximum strength was obtained the capacity curve
achieved using FEM, stopped due to convergence issues. A horizontal plateau is evident in the capacity
curve obtained from FME after reaching the maximum strength, as it adopts elastic-ideal plastic
constitutive curves for the structural elements. Frame by Macro Elements assumes rigid connection
between spandrel and pier elements so the difference in the stiffness can be attributed to this. The
difference regarding the maximum load coefficient is about 6.9%, which can be declared as acceptable
given the differences between the modeling strategies.

The largest amount of cracks is located at the walls W-Y6 and W-Y5 (see Figure 6 (b) and (a)) when the
structure reached the final stage of loading (=0.518). Shear damage of these walls is more than evident,
which would further on lead to shear failure. Due to bending above the openings, bending damage and
maybe later on even failure is to be expected in these locations. This has a direct implications on the
development of the cracks on the facade wall W-X1 (see Figure 6 (c)), and W-X3, but to a smaller
extent, where evident compression (seen from the principal compressive stresses, not presented here)
damage and even failure at the ground level is noticed, with most probably later on local falling out of
masonry.

(a) (b) (c)

Figure 6. Principal tensile strains for =0.518: (a) W-Y4; (b) W-Y6; (c) faade wall W-X1.

Similar damage pattern has been obtained by Frame by Macro Elements as well (see Figure 7). The
basement which is made out of concrete remains undamaged during the application of the load. For
more details see Ademovi, 2012, Ademovi el. al.,2013).
 (a) (b) (c) (d)
Figure 7. Damage in the walls in "+Y" direction at the failure stage: (a) W-Y4; (b) W-Y6; (c) W-X1;
(d) legend.

This kind of damage was observed on a structure of a similar type after the earthquake that struck Skopje
in 1963, as shown in Figure 8. The concentration of damage is located on the ground floor.

Figure 8. Concentrated damage at the ground floor (WHE, W. H. (2002)).

2.1. Time History Analysis

The structure was exposed to the Petrovac short-period earthquake, recorded during the earthquake at
Petrovac on the April 15, 1979, (Montenegro).

Comparison of the capacity curves from the pushover and THA for the earthquake action of ag=0.1g;
0.2g; 0.3g and 0.43g , as shown in Figure 9, gives information regarding the maximum load coefficient
and displacement, however it is important to take into account the maximum inter-storey drift during
assessment and verification of the structure. It is also possible to observe the dynamic behavior causes
more pronounced stiffness degradation due to cyclic loading.

Figure 9. Part of hysteresis curves

The large drift imposes severe deformation and ductility demand at locations of the lower floors. The
distribution of the storey drift over the height of the structure depends on how much the structure can
deform into the inelastic range. Basically, it greatly depends upon the ductility factor, a measure of the
degree of inelastic behavior. During the earthquake action of ag=0.1g maximum load coefficient of 0.23
was reached while the maximum displacement was 11.38mm and the entire hysteresis curve for the
control node 44014 during the earthquake action of ag=0.1g. is shown in Figure 10(a). Displacement is
rather small so it can be stated that the structure is safe under this kind of seismic action. However,
during the earthquake action of ag=0.2g the maximum inter-storey drift amounts to 5.77%, which to a
large extend overcomes the allowable maximum inter-storey drift of 1% as defined in Eurocode 8, see
Figure 10 (b). For more details see (Ademovi, 2012).

(a) (b)
Figure 10 (a) Hysteresis curve of the relationship between the horizontal displacement at the top floor (Control
Node 44014) and the seismic coefficient in the Y direction Inter-story drift (b) Inter story-drift at ag=0.2g

Regarding the global behavior of this case study, distribution of cracks due to earthquake action as
shown in Figure 12 (a), it can generally be correlated with the experimental investigations conducted by
Tomaevi in 1991 where it has been seen that there is a concentration of damage in the first floor as
seen in Figure 12 (b). In this case, the ground floor can be looked at as the first floor due to high stiffness
of the basement made out of reinforced concrete. This indicates that this type of masonry structures can
be modeled as a storey mechanism mode-shear wall with pier action (Tomaevi, 1987).

(a) (b)
Figure 12. (a) Crack pattern during 0.1 g at 23.87 s (b) Experiments on 4 storey masonry building (Tomaevi, et
al., 1991)
2.1. Conclusion
On the basis of the results obtained from the THA it can be seen that the structure has a typical shear
behavior. The walls parallel to the load experience diagonal cracks caused by shear, and due to the cyclic
loading, an evident diagonal "X" type cracks are formed. At the location of the openings the
concentration of the damage is evident due to the concentration of the stresses. As it can be noticed at the
end of the earthquake action major damage, besides the load-bearing walls that are governing the
behavior of the structure, are seen at the non-load bearing walls at the lower levels. The major
concentration of the damage is located between the basement and the ground floor, which can be
connected with the discontinuity and large difference of the stiffness. Large damage is observed in the
lower floors where the largest inter-story drift was observed imposing severe deformation and ductility
demand at these. This kind of behavior has been identified by the previous earthquakes on a similar
structure in Skopje and by experiments conducted by Tomaevi in Slovenia. The propagation of the
damage slowly expands to the upper floors on the faade walls. The earthquake action, as a cyclic
loading causes the degradation of the stiffness of the structure.

3. POSSIBLE STRENGTHENING METHODS

Which type of strengthening method will be applied depends on type of the structure, quality of the
material of the existing structure, availability of adequate equipment and workers' skills and knowledge,
and above all its seismic resistance. The state of the structure before the intervention greatly influences
the selection of the technical procedure, its applicability and effectiveness. It is important to ensure a
good global behavior of the structure due to earthquake action. So, the entire structure has to possess an
adequate resistance, ductility and energy dissipation. The correct application of strengthening method is
not only of a technical issue but of an economic justification as well.

The major deficiency of these types of structures is the lack of load-bearing structures in the longitudinal
direction.

One possibility would be to make additional walls in the longitudinal direction, the location of these
walls should be chosen such that a uniform distribution is obtained in plane and in height, in order to
avoid unfavorable torsion effects. As the building is occupied, this would be rather difficult to perform,
especially construction of internal walls.

A better solution would be to construct a steel skeleton around the building and in that way improve the
resistance of the structure. This kind of strengthening has been done on several structures of historical
importance in Italy (Tomasi, et.al. 2007), and in Poland (Berkowski, et.al., 2001). Construction of steel
braces allows for the introduction of shear walls with lattice scheme. In this way the internal rigidity in
respect to the shear center will be balanced, avoiding big torsional effects and at the same time the
resistance and the ductility of the structure to lateral loads would be increased.

Secondly, strengthening of the faade walls could be done by reinforced cement coating, as proposed by
Tomaevi (WHE, 2002), for buildings of the same typology in Slovenia. Application of this method
would improve the lateral resistance and energy dissipation capacity of the system.

Possibilities of using ferro-cement, carbon fiber or polymer grids are another option which is being
investigated and in particular polymer grids have been recognized by Eurocode 8 in 2003 (Ramiro,
2004).

As the floors are rigid, another possibility would be to introduce vertical confining elements
(tie-columns) at all corners and wall intersections. In this way the integrity of the structural system
would be upgraded, improving the resistance and energy dissipation capacity. This could be done in
combination with other presented methods.
5. OVERALL CONCLUSION

Before applying any of the intervention methods to the structure it is important to take into consideration
all the changes in the structural behavior that it can induce. It is necessary to conduct numerical
calculations and experimental tests to prove the effectiveness of the strengthening method in a particular
case. As well it is important to keep in mind that the structure is occupied, so the works which are to be
conducted for the strengthening should have a minimum impact on the occupants. Availability of
equipment and skillfulness of the workers to conduct certain strengthening techniques should be taken
into account. Financial aspects as well as aesthetic implications that the strengthening would have on the
structure and its surrounding should be carefully considered.

It can be clearly seen that the seismic assessment of a structure is a multidisciplinary approach where
many researchers from different fields have to work together in order to have a good and effective
results and solutions.

REFERENCES

Ademovi, N. (2011). Structural and Seismic Behavior of Typical Masonry Buildings from Bosnia and
Herzegovina. Advanced Masters in Structural Analysis of Monuments and Historical Constructions.
University of Minho, Portugal.
Ademovi, N. (2012). Ponaanje zidanih konstrukcija u BiH pri dejstvu zemljotresa sa stanovita savremenih
teoretskih i eksperimentalnih saznanja, Ph Thesis, Graevinski fakultet Univerziteta u Sarajevu, Bosnia and
Herzegovina.
Ademovi, N., Hrasnica M. and Oliveira D.V. (2013). Pushover analysis and failure pattern of a typical masonry
residential building in Bosnia and Herzegovina, Engineering Strutcures Vol 50, pp.13-29.
Berkowski, P., & Dmochowski, G. (2001). Examples of strengthening of main walls in old buildings. Historical
Constructions, pp. 1033-1037.
CEN Eurocode 8. (2003). Design of Structures for Earthquake Resistance, Part 1: General rules, sesmic action, and
rules for buildings. prEN 1998-1-1. European Committee for Standardization,Management Centre in
Brussels. Brussels.
DIANA 9.4. (2009). TNO DIsplacement method ANAlyser 9.4-Finite element analysis. Users Manual, release
9.4. Netherlands.
Eurocode8. (2003). Design of Structures for Earthquake Resistance, Part 1: General rules, sesmic action, and rules
for buildings; prEN 1998-1-1. Brussels: CEN.
Mendes, N. and Lourenco, P. B. (2010). Sesimic Assessment of Masonry "Gaioleiro" Buildig in Lisbon, Portugal
Journal of Earthquake Engineering, pp.80-101.
Peuli, . (2002). Konstruktivni elementi zgrada, Croatiaknjiga, Zagreb.
Ramiro, S. A. (2004). Seismic Strengthening of Masonry in Buildings and Culural Heritage. 6th National
Congress of Seismology and Engineering Sesmicity, pp. 82-100.
S.T.A.-DATA 3 Muri. (2010). Manual , STA DATA srl, Torino.
Tomasi, R., Pezzo, I. M., & Piazza, M. (2007). Rehabilitation of an Historical Theatre in Italy. Ann. Mus. civ.
Rovereto. Vol. 23., pp. 89-102
Tomaevi, M. (1987). Dynamic Modelling of Masonry Buildings: Storey Mechanism Model as a Simple
Alternative. Earthquake Engineering and Structural Dynamics. Vol 15, 731-749.
Tomaevi, M. (1999). Earthquake-Resistant Design of Masonry Building, Imperial College Press, Tom. I.
London.
Tomaevi, M. and Weiss, P. Sesimic Behavior of Masonry Buildings. (1991). Reinforced versus Unreinforced
Masonry. 9th International Brick/Block Masonry Conference, Berlin. Vol. 1, pp. 552-559.
WHE, W. H. (2002). Unreinforced Brick Masonry Apartment Building. Ljubljana: Housing Report.