Professional Documents
Culture Documents
(5)
( ) ( )( ) ( )
s p s s m
m s m
m
A A E E
E
,
2
2
2 2
,
1 1 + +
(6)
where E
s
is the elastic modulus of steel bars. Assuming that both E
m,s
and are a function of the
compressive strain
m
, then Equation 6 may be considered as the - law of the confined masonry.
The above equations hold until the stress in the tensioned bars does not exceed the steel yield
stress f
y
, When this occurs, from Equation 1 it results:
p s y x c
A A f /
,
(7)
where the steel yield stress f
y
has to be taken negative. By substituting Equation 7 into Equation 2
and considering that
c,y
= 0, the stress in the load direction becomes:
( ) ( ) ( )
2 2
,
1 / +
p s y m s m m
A A f E (8)
At the same time, from Equation 7, the stress acting in steel bars must fulfil the condition:
y s p x c s
f A A
,
(9)
6 2003 ABAQUS Users Conference
where
c,x
is given by Equation 5. Equations 8 and 9 allow the wall behaviour in the steel post-
elastic range to be easily described.
3.3 Application of the method
For the above method to be applied, appropriate functions for E
m,s
and have to be assigned. The
secant modulus E
m,s
can be obtained starting from a suitable - relationship for plain masonry. A
number of - relationships susceptible to be used exist in the technical literature, derived from
experimental tests on either plain concrete, or fitted directly on masonry specimens. In this
procedure, a model derived from the Saenzs law for concrete has been proposed (Sargin, 1971):
( )
( )
,
_
,
_
,
_
,
_
,
_
z
u m
m
u m
m
u m u m
u m
m
u m u m
u m
m
E
E
k
, , , ,
, , ,
,
2
/
1
/
(10)
where E,
m,u
and
m,u
are the initial elastic modulus, the ultimate compressive stress and the
corresponding strain, respectively. Such a model represents a slight variation of that proposed in
(Mandara,1998) and is able to provide a good description of both pre- and post-collapse
compressive behavior. There are some differences in the model presented herein, compared to the
original formulation of Saenzs law: 1) a strength enhancement factor k due to confinement has
been introduced in order to take into account the increase of masonry resistance produced by the
combined state of stress; 2) the coefficient z have been introduced instead of a numerical factor
equal to 2, in order to have a more accurate reproduction of the softening branch of the -
relationship. Assuming values of z other than 2 causes the actual maximum of the - curve to be
slightly different from
m,u
, but gives a much better approximation of the material post-collapse
behavior. As shown in the next section, both k and z have been found being rather dependent on
both mechanical and geometrical properties of the masonry wall. An appropriate expression for k
can be put into the form:
( )
m y c m x c
k
, ,
1 + + (11)
where is a numerical coefficient to be fitted experimentally, which depends on the masonry
features. The
c,x
/
m
ratio can be evaluated from Equation 5 - or from Equations 7 and 8 after the
bar has yielded - as a function of , E
m,s
, E
s
, A
p
and A
s
. Similarly, from Equation 4 it is easy to
show that:
( )
m x c m y c
, ,
1+ (12)
The expression of secant modulus can be easily deducted from Equation 10 remembering that E
m,s
=
m
/
m
:
2003 ABAQUS Users Conference 7
z
u m
m
u m
m
u m u m
m
m
s m
c
E
E k
E
,
_
,
_
,
_
, , , ,
,
1
(13)
In the evaluation of
c,x
/
m
and
c,y
/
m
a trial-and-error procedure would be necessary for their
calculation. In fact, since
c,x
depends on E
m,s
through
m
, its value depends on k itself, too.
Because of Equation 12 this happens for
c,y
as well. Nevertheless, it can be observed that, in the
view of design calculations, the terms of Equation 11 may be computed in an approximate way
assuming k = 1 in the evaluation of E
m,s
, independently of the bar yielding. The related inaccuracy
can then be covered by an appropriate value of the parameter .
Together with , in the calculation of
m
a suitable function for the Poissons modulus (
m
) has
to be assigned in order to evaluate the - relationship of confined masonry. An accurate
estimation of is extremely important for the accuracy of the model, that is for a correct
evaluation of the confinement effect, as the confinement mainly depends on itself.
Unfortunately, the meaning of the Poissons modulus in masonry is not exactly the same as in an
elastic continuum, in particular when the collapse load is approached. The transverse expansion of
the masonry is, in fact, strongly influenced by the onset of cracks along the load direction. As a
consequence, when the masonry wall is regarded as a whole, specific allowance should be made
for cracks in the evaluation of the expansion ratio. Also, the actual masonry texture, that is the
block size and configuration as well as the mortar properties, should be considered in the
assumption of a (
m
) function. In this view, it is clear that a direct evaluation of the transverse
expansion ratio leads to a so-called apparent Poissons modulus, whose mechanical meaning
is far different from the one of an elastic, isotropic continuum. Some existing tests (Faella, 1993),
in fact, indicate apparent values of the coefficient at the ultimate strength equal to or higher than
1.5 2, depending on the masonry features.
Such values, clearly incompatible with the physical meaning of , can not be assumed in case of
confined masonry, as the development of cracks is greatly contrasted by confining ties. In this
case, cracks occur as well, but with a rather different aspect from the case of unconfined masonry.
In general, depending on the value of confining stress, the concept of apparent expansion drops
most of its meaning, being downsized by an unknown extent. In absence of reliable data on
under combined stress conditions, assuming that in such a case a reduction of the void volume due
to the local crushing of the masonry could take place, it seems more appropriate to assign a law for
reaching values not higher than 0.5 (no volume change) in the large displacement range. This is
more complying with the assumption of isotropic continuum made for deriving the - law. At the
same time, the condition that = 0 for
m
= 0 should be fulfilled, this corresponding to what
commonly observed in tested specimens. As being stated, a possible law for can be put into the
form:
( ) ( ) ( ) ( )
2
,
2
, ,
/ / / /
u m m u m m u m m
b a + (14)
8 2003 ABAQUS Users Conference
The coefficients , a and b have to be fitted in order to adequately reproduce the results coming
from either numerical simulation or direct experimentation. Since the Poissons ratio is
responsible for the lateral expansion of the masonry and, consequently, for the confining pressure
of the steel plates, the coefficients a and b can be fitted in such a way the bar yielding occurs at the
same point as in the reality. On the contrary the factor, taking into account the effect of
confinement on the masonry strength, can be evaluated by considering the actual value of ultimate
load bearing capacity, as deducted by experimentation or by F.E.M. analysis.
4. Parametric analysis
An extensive parametric analysis has been carried out by means of the F.E.M. model described in
the previous chapter, assuming several material properties and uniformly confined walls (Figure
8b). The analysis has been concerned with the case of both uniformly and partially confined
masonry walls, in order to investigate all possible failure modes. Three masonry types have been
considered, whose main mechanical parameters are shown in Table 1, together with relevant
values of z, a, b and , as found from a best fitting procedure of theoretical curves against F.E.M.
results. A corresponding *CONCRETE material model has been used. Values of wall thickness
equal to 300 and 480mm have been considered. Different values of A
s
, represented through the bar
diameter , have been assumed, so as to emphasize the effect of confining steel area. A yielding
stress f
y
= 600 N/mm
2
has been considered in the analysis.
In general, a good agreement between the proposed method and the numerical simulation is found,
in particular in the estimation of ultimate load bearing capacity (Figure 9). Some minor
discrepancies with respect to the F.E.M. results exist only around the knee point and in the
softening branch of the - curve. This is partially to be related to the - relationship assumed the
unconfined material. It can be observed that, owing to the geometrical symmetry of the masonry
panel, a certain amount of confinement in the wall plane (y-direction) does exist even without
plates in the x-direction. This results in the ultimate load and the corresponding strain of the
unconfined wall being higher than those of the plain masonry. This aspect is well caught by the
analytical model when the position A
s
= 0 is made.
In order to consider collapse mechanisms other than yielding of steel bars, partial confinement
with uniformly spaced plates has been also considered. The corresponding F.E.M. model, where
due account of symmetry has been taken, is shown in Figure 10. In such a case, depending on both
plate spacing and A
s
/A
p
ratio, collapse by either local crushing or shear-tension may occur. For a
given A
s
/A
p
value, the corresponding i/t value has been assumed as relevant parameter to establish
whether the wall collapse occurs due to bar yielding, masonry punching or shear-tension failure of
masonry between plates. As long as the i/t ratio increases, the corresponding wall collapse load
decreases, according to the trend shown in Figure 11, where the ratio between confined and
unconfined strength R
c
/R
nc
is plotted against i/t ratio, for A
s
/A
p
= 50, 100 and 400. In general, for i/t
1.5, the effect of confinement vanishes completely. Correspondingly, the failure type moves
from bar yielding to punching or shear/tension depending on the value of A
s
/A
p
value. Both
experimental and numerical results show that the value of i/t and A
s
/A
p
ratio determine the collapse
to occur by shear-tension or punching, respectively. A synopsis view of all possible failure
conditions is given in Figure 12, where the relevant collapse mechanisms are also indicated.
2003 ABAQUS Users Conference 9
Accordingly, a border line between them has been traced, which can give useful indication from
the design point of view. In practice, collapse conditions other than that involved by bar yielding
should be avoided, in that they cause confinement ineffectiveness and/or local crushing of
masonry, and hence, a brittle behavior at collapse for the wall.
Because of the relatively large number of simulation tests referred to in Figure 12, it can be
considered as a helpful tool in orienting design choices. Nevertheless, concerning practical
calculations, an easy tool for the prediction of collapse load for a given value of i/t ratio is needed.
As a matter of fact, the solution to this problem would involve the definition of a very complex
mechanical model, taking into account all relevant aspects of the collapse mechanism. Such
difficulty arises due to the fact that in case of bar yielding or local punching the wall collapse
occurs due to masonry crushing between confining plates, whereas in case of shear-tension failure
it is predominantly a matter of local instability of outer masonry leaves. To this purpose, suitable
interaction models should be used to represent the actual collapse phenomenology. As a more
direct approach is needed for design calculations, a simplified procedure is proposed herein, based
on the combined use of results provided by both F.E.M. simulation and theoretical model for
uniformly confined masonry. In this case, results coming from the theoretical model can be used
for i/t = 0, whereas F.E.M.-based data can be exploited to reproduce the variation of the wall
strength as long as the i/t ratio increases. Curves in Figure 11 can be used to this purpose, in order
to fit a simple relationship relating R
c
/R
nc
ratio to i/t ratio. The following equation is proposed:
( ) [ ]
( )
1
1
1 . 0
1
67 . 0
0
0
+
1
]
1
t i
nc c
nc c
nc
c
R R
R R
R
R
(15)
whose results are plotted in Figure 11 as well. Such equation can be used to predict the resistance
of confined masonry for a given value of the i/t ratio when the resistance ratio of uniformly
confined (i/t = 0) to unconfined masonry (R
c
/R
nc
)
0
is known. Since a certain degree of confinement
exists in walls also without ties, such value is to be calculated via the procedure illustrated in the
previous section. Information on the possible collapse mechanism is then obtained from Figure 12.
5. Conclusions
The non-linear F.E.M. code ABAQUS has been used to simulate the inelastic response of confined
masonry walls up to collapse. This has led to a thorough understanding of global behavior of
masonry in such loading conditions, highlighting all relevant collapse mechanisms. The results of
the wide parametric analysis carried out in this paper have been used to calibrate a purposely
conceived theoretical model. With respect to existing models, mostly concerned with confined
concrete in compression, the procedure discussed herein is based on a reduced number of
parameters, to be fitted on the basis of either experimentation or numerical simulation. With a
suitable choice of these factors, the model exhibits a satisfying degree of accuracy.
In addition, collapse mechanisms other than t hat involved by bar yielding have been investigated
by means of numerical simulation, leading to the attainment of some general conclusions on the
range of geometrical and mechanical properties to be adopted in practice. At the same time,
10 2003 ABAQUS Users Conference
starting from results obtained through the theoretical model, F.E.M. results have been further
exploited to set a simplified procedure for the evaluation of the wall load bearing capacity as a
function of confinement ratios i/t and A
p
/A
s
.
In conclusion, an exhausting framing of the problem has been reached, at least referring to the case
of wall confinement (one-direction confinement). As a further step of the research, an extension to
the case of two-directional confinement is planned, in order to cover application to columns,
pillars, and other vertical masonry elements. To this purpose, the theoretical model has been
already formulated, even though not reported herein, and F.E.M. simulation is about to start.
6. References
1. ABAQUS, User and Theory Manual, Version 6.2, Hibbitt, Karlsson & Sorensen Inc., 2002.
2. Ballio, G., Calvi, G.M., Strengthening of Masonry Structures by Lateral Confinement, Proc.
of IABSE Symp. Structural Preservation of the Architectural Heritage, Rome, 1993.
3. Faella, G., Manfredi G., Realfonzo, R., Stress-Strain Relationships for Tuff Masonry: Ex-
perimental Results and Analytical Formulations, Masonry Int., Vol. 7, No. 2, 1993.
4. Mandara A., Strengthening Techniques for Buildings, in Refurbishment of Buildings and
Bridges (F.M. Mazzolani & M. Ivanyi Eds), Springer Verlag, Wien-New York, 2002.
5. Mandara, A., Mazzolani, F.M., Confining of Masonry Walls with Steel Elements, Proc. Of
Int. IABSE Conference Save Buildings in Central and Eastern Europe, Berlin, 1998.
6. Mazzolani, F.M. Strengthening Options in Rehabilitation by means of Steelwork, Proc. of
SSRC International Colloquium on Structural Stability, Rio de Janeiro, 1996.
7. Mazzolani, F.M., Principles and Design Criteria for Consolidation and Rehabilitation, in
Refurbishment of Buildings and Bridges (F.M. Mazzolani & M. Ivanyi Eds), Springer Ver-
lag, Wien-New York, 2002.
8. Mazzolani, F.M., Mandara A., Modern Trends in the Use of Special Metals for the Im-
provement of Historical and Monumental Structures, Jour. of Struct. Eng. 24, Elsevier, 2002.
9. Nemat-Nasser, S., Hori, M., Micromechanics: Overall Properties of Heterogeneous Materi-
als, 2
nd
Ed., Elsevier, Amsterdam, 1999.
10. Sargin, M., Stress-Strain Relationship for Concrete and Analysis of Structural Concrete Sec-
tions, Study n. 4, Solid Mechanics Division, University of Waterloo, Canada, 1971.
7. Acknowledgements
This research initially started within the project Metal Systems for the Consolidation of Structures
(resp. F.M. Mazzolani), in the framework of Progetto Finalizzato Beni Culturali issued by the Italian
National Research Council (C.N.R.). The ongoing development is now framed within the project
Innovative Metal Materials in the Seismic Strengthening of Masonry Structures (resp. A. Mandara),
which is a part of the project Diagnostic and Safeguard of Architectonic Works, sponsored by Italian
National Research Council (C.N.R.) with funds granted by Italian Ministry of University and Research
(MIUR) (L. 449/97).
2003 ABAQUS Users Conference 11
Table 1. Synopsis of wall mechanical properties assumed in the parametric
analysis and relevant material calibration parameters (z, a, b and ).
Masonry
type
E
(MPa)
m,u
(MPa)
m,u
t
(mm)
(mm)
z a b a
1 3300 3.5 0.0025 480 8,12,16 1.5 1 2.1 0.2
2 2300 2.5 0.0025 300 8,10,12 1.8 3 2 0.5
3 660 2.5 0.007 300 4,6,8 2 3 1.8 0.5
Figure 1. Examples of masonry members confined by steel tied elements.
a) b) c)
Figure 2. Collapse types: a) bar yielding; b) masonry crushing in the confined
area; c) shear-tension of masonry in unconfined areas.
12 2003 ABAQUS Users Conference
Figure 3. Experimental set-up referred to in (Ballio, 1993).
Figure 4. The ABAQUS F.E.M. model of specimens tested in (Ballio, 1993).
0
1
2
3
4
5
6
0.000 0.005 0.010 0.015 0.020 0.025 0.030
e
s (Mpa)
h)
b)
a)
g)
0
1
2
3
4
5
6
0.000 0.005 0.010 0.015 0.020 0.025 0.030
e
s (Mpa)
f)
d)
c)
e)
l )
i)
Figure 5. The masonry - laws of specimens tested in (Ballio, 1993).
2003 ABAQUS Users Conference 13
0
1
2
3
4
5
6
0 0.005 0.01 0.015
(Mpa)
h) b)
g)
a)
0
1
2
3
4
5
6
0 0.005 0.01 0.015 0.02 0.025 0.03 0.035
(Mpa)
l)
f)
d)
c) i)
e)
Figure 6. Calibration of numerical F.E.M. model (dotted line) against tests.
Collapse by bar yielding Collapse by masonry punching Collapse by shear-tension
Figure 7. Contour of Von Mises equivalent stress (top) and horizontal
displacement (bottom) for observed collapse mechanisms.
14 2003 ABAQUS Users Conference
Ap
A
s
c,x
c,x
m
t
Figure 8. The mechanical model of uniformly confined masonry (a) and the
corresponding F.E.M. idealization (b).
0.0
0.5
1.0
1.5
2.0
2.5
0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0
em/em,u
s m/s mu
F = 16mm
F = 12mm
F = 8mm
Unconfined
Plain
masonry
Masonry type 1
0.0
0.5
1.0
1.5
2.0
2.5
0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0
em/emu
s m/s mu
F = 12mm
F = 10mm
F = 8mm Unconfined
Plain
masonry
Masonry type 2
0.0
0.5
1.0
1.5
2.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
e
m/e
mu
sm/s mu
F = 8mm
F = 6mm
F = 4mm
Unconfined
Plain
masonry
Masonry type 3
Figure 9. Comparison between theoretical and F.E.M. results (dotted line) for
uniform confinement.
a) b)
2003 ABAQUS Users Conference 15
0
1
2
3
4
5
6
7
8
9
10
11
12
0 1 2 3 4 5 6 7 8 9 10
i = 100mm
i = 150mm
i = 200mm
i = 250mm
i = 300mm
i = 500mm
i = 700mm
s
m(MPa)
d(mm)
A
p
/A
s
= 50
Figure 10. The F.E.M. model in case of partial confinement (left) and the
corresponding force-displacement relationship.
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
0.0 0.5 1.0 1.5 2.0 2.5
i/t
Rc/R nc
Ap /As = 50
400
100
Masonry type 1
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
0.0 0.5 1.0 1.5 2.0 2.5
i/t
R c/R nc
Ap /As = 50
400
100
Masonry type 2
Figure 11. Influence of i/t ratio on the resistance of partially confined walls:
F.E.M. analysis (dotted line) versus Equation 15 (full line).
0
1
2
3
4
5
0 1 2 3 4
(A
p
/A
s
)*(s
mu
/f
y
)
i/t
COLLAPSE BY BAR YIELDING
COLLAPSE MECHANISM BY SHEAR-TENSION
COLLAPSE MECHANISM BY PUNCHING
Figure 12. Synopsis view of possible collapse conditions.