EXPERIMENTAL DATA OF FRICTION COEFFICIENT FOR SOME

TYPES OF MASONRY AND ITS CORRELATION WITH AN INDEX OF

QUALITY MASONRY (IQM).

Manuela MARINO1, Fabio NERI2, Alessandro DE MARIA3, Antonio BORRI4

ABSTRACT

This study constitutes the results of an experimental investigation into the description of certain
types of wall masonry present in traditional southern Italy buildings. The parameter applied is that of
the friction coefficient between the joints of the stone blocks which constitute the masonry. The wall-
to-wall friction is as important as other mechanical parameters in that it acts on the structural
behaviour of masonry with regard to in plane mechanisms and mixed mechanisms (both in plane and
out of plane). The friction coefficient provides a useful indication in evaluating the type of masonry
and its behaviour to be used in seismic areas. Similar friction values for all types of masonry are
reported in literature, while in effect these values vary according to the type of masonry and materials
used. The aim of this study is not to provide certain friction values, but a range of acceptable values
in relation to the type of masonry considered in the study itself. Experimental investigations which
allow the estimation of the friction coefficient associated with failure through sliding, are carried out
with a hydraulic jack, coherent with the Mohr-Coulomb failure criterion. Shear strength depend by
compression in the wall, so it is necessary evaluate ultimate shear strength (cohesion) in the absence
of normal compression. In order to obtain this result, it is necessary to carry out two tests, on two
walls belonging to the same masonry type, but subject to differing values of compression. The critical
interpretation of the tests provides, for each sample, the experimental values of the coefficients of
friction and cohesion as well as will be shown later in this work. A further interesting aspect of this
study is the verification of the existence of a correlation between experimental results, and a
qualitative judgement expressed through observation of the masonry. To this end, a method well
established in literature has been used for the measurement of the quality of the masonry. An index of
quality masonry (IQM) is evaluated on the basis of the survey of predetermined parameters that
indicate respect for the rule of the art of building (size of the stones, regularity of the texture, quality
of the mortar, etc..). This methodology has been applied to the masonry samples used in the
experimental investigation, in such a way as to be able to compare the results of the two procedures.
This comparison has provided encouraging results, establishing a satisfactory correlation between
the qualitative evaluation of the masonry and experimental data.

1
PhD, Department of Structural and Environmental Engineering, University of Catania, Italy,
[email protected]
2
Professor, Department of Structural and Environmental Engineering, University of Catania, Italy,
[email protected]
3
Dott. Ing., Service Control Construction and Civil Protection, Perugia, Italy.
4
Professor, Department of Engineering, University of Perugia, Italy,
antonio.borri@ unipg.it

1
INTRODUCTION

A description of the mechanical parameters of historical buildings is not simple: this masonry
parameter consists, in fact, is a heterogeneous mixture of stone blocks and mortar, whose
characteristics are dissimilar to those of the individual materials, and which depend on the masonry
textures and their correct application. The behaviour of a wall during seismic activity on its own plane
is strongly connected with the friction resistance created along the joints between blocks where lesions
appear. Friction resistance depends not only on the masonry texture and the wall cross section but also
on the materials used. As has been shown in recent studies on the seismic vulnerability of traditional
buildings (Neri and Marino, 2012), the precise measurement of friction resistance produced in a wall
subject to mixed mechanisms (both in plane and out of plane) is necessary in reliably measuring the
precise capacity of the wall. It is emphasised that mixed mechanisms can only occur in buildings with
good connections between the walls. The conventional values given in literature to wall-wall friction
are measured as 0.4÷0.45, for all types of walls. However, existing wall types may vary greatly, so
that it is hardly reasonable to assume that a single value of friction coefficient can be universally
valid. The aim of this study is, through a series of experimental measurements, to identify a range of
acceptable values of friction coefficient of some wall types of traditional Sicilian buildings.
Furthermore, the study aims at showing a correlation between friction coefficients identified in the
experimental investigations and masonry quality as defined on the basis of specific parameters. The
mechanical resistance of a wall is strongly dependent on some qualitative elements, tied to the rules of
construction when followed “to the letter”. Numerous studies exist in literature providing qualitative
information relating to wall masonry in quantitative parameters, with the aim of giving an indication
of wall quality characteristics rather than supplying definite behavioural information. It is therefore
necessary to establish correspondence between two different approaches, the first based on the
qualitative analysis of building principles “to the letter”, and the other on direct physical
measurements. In this way it is possible to establish a correlation between some factors relating to “to
the letter” building principles, and mechanical parameters which govern the behaviour of the masonry
with respect to vertical action, horizontal action in the plane of the wall, and horizontal actions out to
the plane. This approach can also be extended to the description of friction which develops along the
wall in the case of mixed mechanisms.

1. MASONRY TYPES ANALYZED IN THIS STUDY

The subjects of this study are two different types of masonry characteristic of traditional buildings
in the south of Italy: square cut blocks in soft stone, and rough cut blocks in lava stone. Ten samples
were made of each type of stone, and for each sample there was:
− n.1 survey of masonry composition by removing of plaster and survey of cross section of the
wall, when it was possible. Observation of masonry texture and wall section are necessary to
establish compliance with the rules of construction followed “to the letter”.
− n.2 sliding tests at the point where the sample was taken and on a other wall of the same type, but
subject to differing normal compression with respect to the former;
− n.2 measurement of the compression stress at the points where the cutting tests have been carried
out.
Compression stress was measured through tests with flat jacks or through analysis of load.
Critical interpretation of the above tests has provided, for each sample, experimental values of the
coefficients of friction and of cohesion. Some examples of representative samples of each masonry
type are reported below (figg. 1-2)

2
 3

Figure 1–Some samples of square cut blocks in soft stone masonry analyzed in this study:(a) masonry texture in
a school near Siracusa (SR); (b,c,d) masonry textures and cross section in “Concordia Theater”, Ragusa (RG);
(e) masonry texture and cross section in a office building, Modica, (RG).

Figure 2 – Some samples of rough cut blocks in lava stone masonry analyzed in this study: (a -d) masonry
textures and cross sections, in a school in Paternò (CT); e) masonry texture in an housing for undergraduate
students, Catania; (f-g) masonry texture and cross section, in a municipal building, S. Gregorio, (CT).
2. EVALUATION OF MASONRY QUALITY INDEX

Rules of construction followed “to the letter” constitute the overall building measures which, if
carried out during the construction of a wall, guarantee its reliability and ensure it is compact and
structurally solid. This is derived from an age-old building practice, and from the direct observation of
the behaviour of masonry. On observing characteristic parameters of correct and efficient masonry
construction, it is possible to establish a masonry quality index for each of the possible stress
directions on a wall panel. An index of wall quality is therefore obtained for vertical loads, horizontal
actions out of plane and finally for horizontal actions in the plane. The degree of application of each
parameter “to the letter” is evaluated through observation of the masonry, and a value is given to each
according to the direction of the stress action. Building requirements which together constitute “to the
letter” are listed below:
High Quality Mortar/ effective contact between elements (M.A.) This condition, which is
necessary for the uniform transmission of loads and the discharge of energy on to the ground, is
obtained either by direct contact between square-cut blocks or through the mortar. Besides levelling
the contact between the blocks, high quality mortar ensures the wall possesses cohesive resistance.
Transversal Locking /presence of diatones(P.D.). This condition prevents the wall dividing into
various sections simply positioned one next to the other, and furthermore, permits the load to be
distributed along the entire width of the wall even when the load bearing is on the wall edge (e.g. a
floor lying only on the internal part). This condition can be met thanks to “diatones”, or rather, blocks
crossing the entire width of the wall.
Form of Resistant Elements (F.E.L.). The presence of blocks with a regular shape ensures the
activation of the friction force, to which a large part of a wall’s ability to resist horizontal actions is
due. In fact, friction is principally activated through the effect of weight from masonry lying above the
sliding surface, and is maximised for horizontal sliding. It can thus be seen that this is one of the
conditions necessary in obtaining efficient locking between the wall elements.
Dimension of the Elements (D.E.L.). Resistant elements of large dimensions with respect to the
width of the wall, ensure a high degree of stability to the wall, as do the diatones. Furthermore, thanks
to their elevated dimensions, they are extremely heavy with good locking between them.
Staggering of Vertical Joints/In Plane Locking (S.G.). This condition allows for the application of
a further resistance element in the wall: wedging between resistant elements (also known as “locking
effect”) which, together with friction, guarantees resistance to co-plane action.
Horizontal Rows (O.R.). This condition makes for good distribution of vertical load in that regular
support is obtained. However, the horizontal aspect also assumes importance during seismic activity as
it allows for oscillation around horizontal cylindric hinges without damaging the masonry.
High Quality Resistant Elements (RE.EL). Stone elements must be resistant and in good
condition: intrinsically weak (e.g. mud bricks used in certain areas of the world) or heavily degraded
elements invalidate other “to the letter” conditions.
In order to arrive at an judgement of masonry quality, as has been said above, it is indispensable to
evaluate (analytically or qualitatively) to what degree the seven parameters defining “to the letter”
have been respected. Only following this preliminary step will it be possible to put the seven partial
considerations together into a complete evaluation. It is often easy to recognise the two differing
stances, to respect (R.) or not (N.R.) a given parameter “to the letter”. Field observations on masonry
in real buildings have, however, shown that middle-range stances exist as well, which neither fall into
or outside the “to the letter” definition. Thus, for each parameter in question, a category of judgement
defined as “partial respect” (P.R.) of the “to the letter” has been introduced. After having evaluated
the degree of respect of each of the above parameters, a certain number of points can be awarded to
each, described in table 1. The points reported in the table are then inserted into the formula below,
therefore arriving at a total number of points for each type of stress action.

IQM= RE.EL.×(OR.+ P.D.+ F.EL.+S.G.+D.EL.+MA.) (1 )

This procedure, therefore, leads to three values of IQM between 0 and 10: one for each direction
of stress. What defines each IQM value is the various weight attributed to the “to the letter”
evaluation for the three stress charges.

4
 5

Table 1. Value given to each parameter depending on degree of respect “to the letter”

Vertical loads Horizontal load out of plane Horizontal load in plane

Building principles “to the letter” 
N.R P.R R. N.R P.R. R. N.R P.R. R.
O.R.  Horizontal Rows       0  1  2  0  1  2  0  0,5  1 
P.D. Transversal Locking   0  1  1  0  1,5  3  0  1  2 
F.EL. Form of Resistant Elements     0  1,5  3  0  1  2  0  1  2 
S.G. In Plane Locking  0  0,5  1  0  0,5  1  0  1  2 
D.EL. Dimension of the Elements    0  0,5  1  0  0,5  1  0  0,5  1 
MA. High Quality Mortar/   0  0,5  2  0  0,5  1  0  1  2 
RE.EL. High Quality Resistant Elements    0,3  0,7  1  0,5  0,7  1  0,3  0,7  1 

The weights attributed indicate the importance of a certain parameter for a good reaction of
the wall under a certain type of stress action. The conventional aspect of the method is therefore
emphasized, whose aim – evidently unrealistically– is not that of establishing a certain
characterization of the mechanical behavior of a masonry wall, but that of providing a general
realistic indication. The resistance of a wall on its plane is due to following characteristics: cohesion
attributed to the masonry by high quality mortar; friction optimized for sliding on horizontal surfaces
of elements resisting contact with each other; vertical joints staggered in order to provide adequate
friction resistance. Given the above, it is implicit that the three parameters “to the letter” held to be
fundamental in the development of friction resistance and in the provision of high resistance to in-
plane action, are the presence of square cut blocks, staggering of vertical joints and the quality of the
mortar (F.EL.,S.G. and M.A.). Greater weight has also been given to the presence of diatones (P.D.).
These are important as they allow horizontal in-plane action to work over the entire thickness of the
wall. Masonry of a quality index below 4 does not generally manage to develop monolithic behaviour,
therefore friction resistance cannot be calculated. Some examples are reported below relating the
samples analysed, and their relative IQM. (fig 3)

Figure 3 –Evaluation of IQM in a square cut blocks in soft stone masonry. (N.R.: not respected parameter;
R.: respected parameter; P.R.: Partially respected parameter; D.: Diatones).
 Figure 4 – Evaluation of IQM in a rough cut blocks in lava stone masonry. (R.: respected parameter;
P.R.: Partially respected parameter;).

3. EXPERIMENTAL METHOD FOR FRICTION EVALUATION.

The methodology for the identification of the experimental friction coefficient uses semi-destructive
tests which have been the subject of recent studies. The mechanisms of failure of a masonry wall
stressed by in plane action are associated with either the sliding of the stone along the mortar joints, or
with a break in the normal direction of isostatic traction, which usually coincides with the diagonal
line of the wall panel. A semi-destructive test procedure will be illustrated which allows the
evaluation of the friction coefficient associated with failure through sliding, along mortar joints. The
test is carried out with a hydraulic jack, coherent with the Mohr-Coulomb failure criterion. A block of
masonry is isolated through the removal of mortar on one side and the adjacent block on the other. The

6
 7

block is kept up through its upper, lower and posterior sides, and is therefore subject to the normal
compression acting on the wall. A hydraulic jack is inserted into the empty cavity of the removed
block (fig.5 ). The test consists in applying horizontal pressure on the previously isolated block, which
is increased at a steady rate, resulting in the contemporary movement of the block, until the block
begins to slide. The horizontal movement of the block is measured with a suitable means, rigidly
attached to the wall at an initial distance of about 200 mm, so that the axis of the bases is directly next
to the jack inserted into the masonry. The measurement of the bases, or rather the position relative to
the centres of the disks can be obtained through electronic transductors of millimetric accuracy; the
pressure is measured through digital meters of 0.1 bar. The measurement of the weight applied to the
block is estimated from the pressure in the hydraulic circuit connected to the jack, and correlates with
the deformation of the block itself until sliding occurs.

Figura 5. Semi-destructive test carried out with a hydraulic jack: (a) removal of a masonry block; (b) insertion
of hydraulic jack in the empty cavity.
It is reasonable to assume that sliding occurs along the horizontal mortar joints, subject to normal
compression (fig. 6), whose value can be obtained on the basis of load, or rather on the basis of the
results of a test with a single flat jack; normal compression along the back of the block is inexistent.
Given a simplified hypothesis of elastic material, which is perfectly plastic, isotropic and
homogeneous, and in general suitable to calculation models, the plastic phase equilibrium of the block
is expressed as:

Fu = As (τ 0 + μ ⋅ σ ) + Avτ 0 (2)

As Total area of horizontal mortar joints along which sliding occurs;

Av Area of vertical posterior sliding surface;
τ0 Ultimate shear strength in the absence of normal compression (cohesion);
μ Friction coefficient;
σ normal compression stress;

Figure 6: stress state on the stone subject to sliding.

In most investigations carried out, it has been noted that the posterior side of the block had no mortar,
and therefore its relative contribution to sliding was negligible. Equation 2 has two unknown factors:
the ultimate strength in the absence of normal compression, and the angle of friction. Shear strength
depend by compression in the wall, so it is necessary evaluate ultimate shear strength (cohesion) in
the absence of normal compression. In order to obtain this result and to calculate a single experimental
value of friction coefficient it is therefore necessary to carry out two tests, on two walls belonging to
the same masonry type, but subject to differing values of compression; the two walls may be on
different floors of the building in case of same masonry type on two different levels, or there may be a
central wall and a side wall in t case of differing masonry types on different levels (fig. 7).

Figure 7. Example of location of on-site tests. The samples may be on different floors of the building in the
case of the same masonry type on two different levels (a), or alternatively there may be a central wall and a side
wall in the case of differing masonry types on different levels (b). Slide tests are indicated with code SiA and
SiB; flat jacks tests with code MiA and MiB, where -i is the number of sample.

8
 9

Every pair of values obtained in the tests is recorded on a diagram, having values of the shear stress
and values of compression stress. In this way two points and a straight line are unequivocally
obtained: the slant of the line indicates the angle of friction, while the interception with the vertical
axis indicates cohesion. Values of shear stress and of compression stress obtained from sample n. 2
are summarised in table 2, where the letter S indicating the horizontal sliding tests, and the letter M
indicating the jack tests for the calculation of the compression load. The diagrams of the tests carried
out on wall sample n. 2 and their correlation are illustrated in figure 8.

Table 2- Experimental values obtained from tests on sample n.2.

sample compression stressσ [MPa] tangenzial stress τu [MPa] attrito μ Coesione τ0 [MPa]
S2A − 0,30
M2A 0,25 −
− 0,24 0,67 0,13
S2B
M2B 0,16 −

Figure 8. Diagrams obtained from tests carried on sample n.2: (a) sliding test S2A; (b) flat jack test M2A; (c)
sliding test S2B; (d) jack test M2A; (e) correlation of values obtained by tests; (f) sample S2A.
Once you have two pairs of values for compression stress and for ultimate shear strength, the friction
coefficient μ is calculated as following:

(τ uSiA − τ uSiB )
μ= (3)
(σ iA − σ iB )
τ uSiA , σ iA values of ultimate shear strength and respective compression stress obtained from
sliding tests SiA and flat jack test MiA, where -i is the number of the sample;
τ uSiB , σ iB values of ultimate shear strength and respective compression stress obtained from sliding
tests SiB and flat jack test MiB where i is the number of the sample;

4. TEST RESULTS AND CORRELATION BETWEEN IQM AND EXPERIMENTAL

FRICTION

Results of the investigation are reported in the Tables 3. The values of the experimental friction and
the values of the index of masonry quality on the plane have been obtained for each type of stone
block:

Table 3 – Results of experimental tests.

n° sample masonry typology τ uSiA τ uSiB σ iA σ iB μ IQM

1 Square cut blocks in soft stone 0,27 0,21 0,18 0,08 0,60 5,25
2 Square cut blocks in soft stone 0,30 0,24 0,25 0,16 0,67 6,3
3 Square cut blocks in soft stone 0,30 0,23 0,29 0,20 0,78 7,0
4 Square cut blocks in soft stone 0,27 0,24 0,23 0,18 0,60 5,6
5 Square cut blocks in soft stone 0,32 0,28 0,23 0,153 0,52 5,6
6 Square cut blocks in soft stone 0,25 0,21 0,19 0,13 0,67 5,6
7 Square cut blocks in soft stone 0,15 0,11 0,22 0,156 0,63 4,9
8 Square cut blocks in soft stone 0,16 0,13 0,28 0,22 0,54 4,9
9 Square cut blocks in soft stone 0,23 0,32 0,18 0,36 0,50 4,9
10 Square cut blocks in soft stone 0,24 0,15 0,287 0,15 0,65 6,3
11 Rough cut blocks in lava stone 0,22 0,11 0,20 0,07 0,88 8,0
12 Rough cut blocks in lava stone 0,20 0,13 0,19 0,10 0,81 7,0
13 Rough cut blocks in lava stone 0,427 0,29 0,27 0,12 0,89 8,0
14 Rough cut blocks in lava stone 0,52 0,39 0,24 0,10 0,92 8,5
15 Rough cut blocks in lava stone 0,44 0,37 0,39 0,29 0,70 7,5
16 Rough cut blocks in lava stone 0,87 0,63 0,72 0,46 0,94 10,0
17 Rough cut blocks in lava stone 1,30 1,14 0,322 0,18 1,10 10,0
18 Rough cut blocks in lava stone 0,66 0,58 0,36 0,26 0,76 8,0
19 Rough cut blocks in lava stone 0,71 0,60 0,44 0,29 0,75 7,5
20 Rough cut blocks in lava stone 0,65 0,61 0,23 0,17 0,67 7,0

Early interpretation of the results confirms the assumptions made during the initial phase of the
study, which are that friction coefficient values in literature cannot be assumed for all types of
masonry. In fact, they are decidedly underestimated for the type of masonry walls examined with
respect to the experimental data. Collected data has been used to identify interval values of the
friction coefficient regarding the type of masonry structure in the object of this study. The aim is not
to identify a certain unequivocal parameter, but to give information regarding the behaviour of
masonry. The acceptable range of the friction coefficient values has been calculated for each type of
wall masonry, eliminating the highest and lowest values, subtracting and adding standard deviation to
the average of these values.

Δx = x ±
(x − x )i
2

n −1 (4)

10
 11

where:
x average value of experimental tests.
xi value obtained from each test
n number of tests.

Below is a summary table of the values of friction coefficient μr  thus identified (tab. 4).

Table 4. Acceptable values of the friction coefficient μr  for the types of masonry in object.

Types of masonry μr average  Standard deviation μr acceptable values 

max 0,65
Cut soft stone 0,61 0,06
min 0,55
max 0,90
Rough-cut lava stone 0,83 0,09
min 0,70

The experimental test results have been used to calculate a correlation curve with identifying
values of the wall quality. To this end, the values of experimental friction have been reported in
a diagram together with the relative value of IQM on the plane. For each sample, an x-axis
represents the IQM value and y-axis represents the friction coefficient value measured in the
test. On the diagram the correlation is represented by the curve where the distance between the
points representing the wall samples is shortest. The correlation thus established is shown in the
exponential curve of the equation:
y = 0,303 ⋅ e 0 ,124 x (5)
The correlation curve has been obtained using all data available, therefore on the basis of a
series made up of 20 values, an satisfying curve (fig.9) is observed. The correlation curve
allows for the friction value to be identified, starting with the quality index on the plane. For
example, having given an IQM value equal to 8 to a specific masonry wall, and then applying
this value to the curve, a friction value of 0.83 will be obtained.

Figure 9. correlation curve between friction coefficient, obtained from experimental data, and IQM for horizontal
in plane action stress.
The R2 indicator (the maximum value assumed equal to 1) represents the measurement as far as the
curve approximates the point values. It is calculated as follow:

SSE
R2 = 1− ; (6)
SST

SSE = ∑ (Yi − YPi )

2
(7)

) − (∑ X )
2

SST = ∑ X ( 2
i
n
i
(8)

Yi: value of experimental friction for the i-st sample;

Ypi: value of friction foreseen in the correlation curve
Xi: value of IQM for the i-st sample;
n number of values available.

Conclusion

The contribution this study makes is that of providing a concrete indication for the evaluation of
friction resistance relative to some characteristic Sicilian masonry types, in that, as has been shown,
the values used in literature for the friction coefficient are usually underestimated. Experimental
measurements have provided minimum and maximum intervals of values relative to the friction
coefficients assumed by the type of objects being studied. Thus the study provides not a certain
parameter but rather a behavior indicator, coherent with the approach of present-day technical norms
on existing masonry. Furthermore, the aim is to demonstrate a correspondence between two different
approaches which have until recently been seen as distinct: that based on qualitative analysis of wall
masonry, and that of direct physical experimentation through mechanical tests. It has been shown
how the correlation between the experimental friction coefficient and an evaluation regarding the
quality of wall masonry is able to quite accurately predict experimental results on existing wall
panels. Further experimental tests will be carried out to confirm the early results and to extend them to
other wall types.

REFERENCES

Borri A, (2009) “L'indice di qualità muraria (IQM): Evoluzione ed Applicazione nell'ambito delle Norme
Tecniche per le Costruzioni 2008”. Proceedings of the 13th Italian Earthquake Engineering Congress
ANIDIS, Bologna, Italy, 28 June -2 July.
Borri A, De Maria A, (2009) “ Indice di qualità muraria IQM e sue applicazioni nell'ambito delle NTC 2008.”
L'Edilizia, Ed. De Lettera, Milano, Italy
Caliò I, (2011), “La prova di scorrimento con martinetto piatto”, Proceedings of the 14th Italian Earthquake
Engineering Congress ANIDIS, Bari, Italy, 18-22 September.
Neri F, Marino M, (2012) “ A simplified procedure for the evaluation of the energy dissipation in out of plane
mechanisms of masonry building”, Proceedings of the 15th World Conference Earthquake Engineering,
Lisbona, Portugal, 24-28 September.

12