Professional Documents
Culture Documents
file f:lihll-e
Illecll:lllislll t.f I..-icl( 111:ISt.III-\'
It.:ltletl ill :IXi:11
Pro! Hubert K. Hi/sdorf. Civil Engineering Dept., University various attempts have been made to relate the compressiw
of minois strength of masonry to the tensile strength of the bricks 3
Since a multitude of parameters govern the compressiw
strength of brick masonry , a reasonably accurate estimate ri
the masonry compressive strength is possible only if the.
Inlroduction parameters are considered simultanf',ously , and if the failuft
mechanism of masonry is taken into account. In the followin~
Structural engineers often hesita te to employ reinforced or this failure mechanism will be studied and, based upon ti",
unreinforced brick masonry as a structural material. This is nol discussion, a procedure wilJ be developed which may enable tE
surprising. Despite lhe use af masonry over thousands af years , to predkt strength and behavior of masonry units from
its potenlials hardly have been utilized and our knowledge of characteristic properties of its constituents whieh ean be
its struclural behavior is still quite limited. The prediction of determined in simple tests.
masonry strength from known characleristics af bricks and
mortar used in a masonry unit is Tather unreliable. Therefore,
acceptance af brick masonry units aften require experimental The Slress Slale in Brick Masonry
detennination af the compressive strength af a particular type Subjecled to Uniaxial Compression
af masonry Talher lhan to predict it f TOm known properties af
its constituents. Because af these uncertainties, the allowable Brick masonry is a two-phase material , both phases 0Il
stresses for masonry are Talher conservative ; lhus , lhe use af only have different strengths but also different deformati'"
brick mason ry as a load~bearing material may then become characteristics. In general, the uniaxial compressive strengtll
uneconomical. and lhe modulus of elasticity of the morta r are considerab~
In previous investigations 1,2,for a given slenderness ratio of lower Ihan the corresponding values af the bricks. Therefo~
the test piece and given coring paU em of the bricks, the if the mortar could deform freely, ils lateral slrains would bI
masonry strength increased as the standard compressive Iarger than lhe strains in lhe bricks. This is especially true j
strength of bricks and mortar increased. However. even if the the externaI load approaches the uniaxial compressivc streng
strengths of mortar and bricks are approximately equal. the of the mortaL However, because of bond and friction betWeef
strenglh of the masonry is smaller than the strength of its brick and mortar, the mortar is confined. Thus, an interflol
constituents. Poor workmanship and low workability of the state of stress is developed which consists ofaxial compressi~
mortar generally result in masonry strength reduction. and lateral tension in lhe briek and lriaxial eompression in t~
Furthennore, thick mortar joints reduce the masonry strength, morta r (Figure 6-1). It is only beca use of Ihis triaxial stated
and masonry strength is illtluenced by the water ahsorption of compression lhat a masonry unit eao be subjected to exter[l2l
lhe bricks. However. these relationships exhibit unusually stresses whkh exceed the uniaxial compressive strength oftbl
larger scattcr. mortaL
From observations of brick faiJure mode , il was concluded Even if care is taken in laying bricks, not ali bricks will ~
that the failure of masonry loaded in compression is initiated evenly supported by their mortar bed. Then , in addition to ti><
by vertical cracking or splitting of the bricks. Therefore , externaI load and the internaI slress slate described above, t~
34
Investigation into the Failure Mechanism of Brick Masonry 35
N
Results from an Experimental Investigation
Specimen
Type 8
With this concept in mind, the author has investigated the
failure mechanism of c1ay brick rnasonry units while working
at the Laboratory for Building Materiais at the Technical 9 .2 i
University of Munich. A delailed description of this investiga- Depfh: 4 .2 in.
tion is reported in Reference 4. Deplh: 9.2 in.
Most of the masonry specimens tested in this investigation
consisted of five layers of bricks (Specimen Type B in Figure
6-2). The thickness of lhe morlar joints was 0.47 inches and
0.24 inches, respectively. The bricks , 9.2 by 4.2 by 2.7 inches,
had a standard compressive strenglh of 5300 psi according lo ',7
'" '"
lhe German Specificalion DIN 105. The flexural slrength of
Arrangement
• • • • • • • • • --Tl
~• ~• ~• .~
the bricks was 780 psi, their lensile splitting strength
I .< S
Of Strain
amounted to 410 psi. The major parameters in this study were
Iype and strength of the joint mortar. Eilher lime mortars.
Iime-cement mortars or cernent mortars were used. Their
compressive strength was determined from prismatic speci-
Measurements
'"
• • • • •
f-2.0;,-1 "~14
.- J
o
N "
N
mens 1.6 by 1.6 by 4.1 in. and ranged fTOm 21 psi for lhe lime I, 9.2 In.
mortar lo 3500 psi for the high strenglh cemenl mortars.
Additional dala are summarized in Table 6-1.
Numerous strain measurements over a gage length of 2 in. Figure 6-2. Types oI specimens and arrangemenl oI slrain
at the surfaces of several bricks were carried out to study the measurements.
Latera l Slresses
"ITlI IIlll state af stress acting in the bricks if the masonry unit was
subjected lo an axial Ioad (Figure 6-2).
In evaluating the strain measurements it was assumed that
the stresses in the bricks acting in the horizontal x-direction
+
are line.rly distribu!ed over the height af the brick. Then the
anCk' < F==zzzzzj stress 0x can be separated into tWQ components: (a) the
flexural stresses a xM and the normal stresses 0xN which are
+ constant over the height of a brick. The stresses a x N can be
attributed to differences in lateral strains between the mortar
Mortar< and the bricks. The flexural slresses may be caused by an
Joints + uneven support of the bricks as described earlier.
For two specirnens the distribution af lateral stresses 0xN
~ZZZZZZzzzj and 0xM and of the longitudinal stresses 0y across the width
of a brick is given in Figures 6-3 and 64. These resuI!s confirm
J-..
the original concept af the stress state in masonry. Lateral
tension, whose magnitude increases as the externaI load is
! !! !
" f 1 I II
Longitud ina l Stresses, u-y
increased, was observed in the bricks. Despite the extreme care
taken in manufacturing the masonry samples, the flexural
stresses and an almost random distribution of the stresses in
the direction of lhe externalload could nol be avoided.
) " Vertical cracking was observed in the bricks at loads
considerably below the failure load. In most cases initial
Figure 6-1. Idealized slress dislribulion in a masonry unit cracking occurred in sections in which the sum of the stresses
fUbjected to concentric compression. 0x = 0xM + 0xN reached a maximum.
36 Designing, Engineering, and Constrncting with Masonry Products
Table 6-1
Summary of Test Results
lfI A lime 410 5.1 x 105 5300 450 2_2 x 106 2210 - -
cement
2 B lime 410 5.1 x 10 5 5300 450 2_2 x 106 2930 2760 L58
cernent
3/1 3460 - -
3/2 B cemenl 2550 3.1 x 106 5300 450 2.2 x 106 3100 2860 L53
4/1 3600 - -
4/2 B cemenl 3500 3.5 x 106 5300 450 2.2 x 106 4850 3480 1.35
S/I 1060 - -
5/2 B lime 21 9.4 x 104 5300 450 2.2 x 106 1370 1610 2.18
6/1 1570 - -
6/2 B' lime 21 9.4 x 104 5300 450 2.2 x 106 1490 1910 1.85
To express analylically lhe nonunifonnily of lhe slresses in mortar strength increases lhe coefficient U decreases
lhe direclion of lhe exlernal load ay, lhe nonuniformily 6-6). It is likely thal U is also a function of the
coefficient , U, was introduced. This coefficient is the ratio the mortar. It should be larger for mortars with I'"
between the maximum normal stress observed within ooe workabilily Ihan for mortars with high workability, si"~
brick lo lhe average normal slress acling on lhe masonry. The workable mortars result in more uniform joints and, lhus, in'
coefficienl of nonuniformily is a funclion of lhe applied load. more unifonn suppart af lhe brick. However, this tenden~
At low stresses, U decreases as the externaI load increases: cannot be c1early deduced frem the available data.
local "yielding" or crushing of lhe morlar ai poinls of high lf the loca1 stress concentrations in the bricks were lhe on~
stress concentration results in a more even distribution af parameler affecting the compressive slrenglh of vari""
slresses. As failure appreaches, lhe nonunifonnily coefficienl masonzy unils made frem lhe same Iypes of bricks,
rapidly increases (Figure 6-5). [ailure should Dccur as soon as the maximum local stress
The coefficient of nonuniformity a1 failuTe is also a U • a x m exceeds lhe compressive strength of lhe brick, j<4
funclion of slrenglh and workability of the mortar. As lhe U is lhe nonuniformily coefficient and a ym is the appl
lnvestigation into the Failure Mechanism of Brick Masonry 37
E
~
"'-
~
-400 --
Wldth of Brick
An AnaIytical Procedure to Predict the Strength
of Clay Brick Masonry Loaded in Compression
1500
Figure 6-8 shows the developrnent Df stresses as they may
oceur in a single brick within a masonry uoH subjected to axial
compression. It is assumed that the lateral tensile stresses in
lhe x and z directions, a x and azo are equa1. In Figure 6-8
these stresses are given as a function of the local maximwn
stresses, 0y' which act in the direction of the externalload.
1000
500 r--;,;
"" I'-.. ,./ r---...
- :; f- ,;'
1\1/ -----
"-
'\
J-
The stresses 0y can be computed from the average masonry E
g ~
o
Width of Brick
400
z
,"
E
~
V ""-..
--
o
-- -- / - 2.5
~
-200
Width of 8rick 1cim.n JWith
Lime Mortar
,
b
"''---- ~ /
/ )ceciCement
m.n Martar
2 Wlth
/'
Width of 8rick
~
I--- 7 ---
V
-3000 500 1,000 1,500 2,000 2,500 3,000
O"ym , psi
-2000
/\ V ""-..
Figure 6-5. Nonunijarmity coefficient as functian af applied
average masonry stress, a ym'
o", -1000
r--;,; _ w
N - ~
w
;.-
E ...L!!l.. stresses, a ym , and the coefficient of nonuniformity U. Line A
g in Figure 6-8 represents the failure criterion for the triaxial
o
Width of 8r ick strength of bricks and indicates the combinations of com-
pressive stresses ar. and lateral tensile stresses 0x and a z which
Figure 6·3. Distribution of stresses in a brick within specimen will cause local failure or cracking Df the brick. For a x = a z =
no. 2. Joints are lime--cement morta,. 0ym = 2400 psi. O, 0y is equal to the uniaxial compressive strength of the
38 Designing, Engineering, and Constructing with Masonry Products
r I I
MorlofS of LCM' 'tY<rlobilily
increasing externa110ad ar increasing local stress, the minimUtn
lateral tensile stress increases.
If the exlernal load is increased beyond lhe load aI fint
cracking, then stresses in an uncracked section af the brid
õ
"r'<"- ,
~
v- Mortors of HiQh
Workob~ity
may develop along Line B2 · As soon as line B2 inlersecl, Ih.
failure criterion Line A, the second crack is forrned. TI!>
process af cracking continues, and the brick may be split intn
• small elemeots. Excessive cracking also results in an increase of
~ lhe coefficienl of nonuniforrnily (see lesl results in Figure 6-5.
Under lhe best condilions, failure of lhe masonry oceu.
oo when the laleral lensile strength of lhe brick is smaller Ih"
KlOO 2000 3000 4000 5000 the stress which is necessary to sufficiently confine the mortar
Compressive Slrengh of Morlar, f'), psi
Therefore, the intersection of the faBure criterioo Line A and
lhe minirnum laleral stress Line C corresponds to lhe ultimal'
Figure 6-6. Variation of nonuniformity eoeffteient wilh type
load of lhe masonry uni!.
ofmorlar. In Figure 6-8 the paramelers which are known lo affecl I'"
compressive streogth of brick masoory are considered:
1. The uniaxial compressive slrenglh of lhe brick.
8000 2. The biaxial lensile slrenglh of lhe brick.
3. The failure crilerion for bricks under a Iriaxial slale <Í
, stresses as represenled by Line A.
b 4. The uniaxial compressive strenglh of lhe mortar whicl
,; 6000
1 corresponds lo lhe onsel of Line C in Figure 6-8.
'ª " , b
" S. The behavior of lhe mor lar under a slale of lriaxii
"'••
õ
400
'/
~ o
eompression, determining the shape and inclination of Line C.
6. The coefficient of nonuniformity.
•
in °f Failure of the masonry may oceur before Line A intersecu
wilh Lirte C: lhe stress si ale ai lhe surface of lhe brick and Ih<
"
u
o
-' 2000
6
mortar joints is only biaxial, so the mortar is not sufficientl)
,
E confined; therefore, spalling of lhe surface layers of lhe brickt
E may occur at lower stresses and may result in a rapidl)
~
00 1000 2000
i
3000 4000
progressing failure.
The following equation is an attempl lo express the failwr
Compressive Slrength of Mortor, t'j ,psi criterion described above in an analylical formo No informo
tion 00 the behavior of bricks under triaxial stresses WI
available. Therefore, il was assumed lha I lhe failure crileri'"
Figure 6-7. Maximum local stress in briek aI failure.
Line A as shown in Figure 6-S corresponds lo a slraighl !in<
Then, Line A can be expressed by lhe following equalion:
strength of concreteS found that the triaxial strength of <, ' <,
concrete ean be approximated by the following expression:
Cf ac~ i n 9
(6 -2a)
f"Oilur e ( "'erlan 01 Sro ck-L ,n e A
I (Gy
4.T - f).') (6-2b) Figure 6-8. Failure criterion of brick masonry.
(6-3) .. 2000
•,
__E ~ . -r~ '0 8 f~,m~~'
/'
~-
In which 0xb = lateral tensile stress in bricks 1000
Uxj = lateral compressive stress in rnortar joint
fi = height of brick
j = thickness of joint
I
Substituting Equation 6-3 in Equation 6-2a we obtain an 2000 3000 4000 5000
expression for Line C in Figure 6-8 f~ , measured
From Equation 6-1 and 64 the magnitude of lhe maximum f'b f bt + a f'
J
local stress at failure, 0Y' ean be determined. It corresponds to f'=- (6-5)
lhe poinl af inlerseclion af lines A and C: m Uu f bt + a f'b
f bt + a f.'
)
This expression depicts the known relationships betwecn
the compressive strength af masonry and variaus pararneters:
f bt + a f b' masonry strength increases with increasing compressive
strength of bricks aod mortar, with increasing tensile strength
where a = j
4.1b of bricks, and with decreasing ratio of joiot thickness to height
of brick. It should be realized, however, that U u is not a
Using the nonuniformity coefficient at failure constant but depends on a number of pararileters including the
average masonry stress at failure ean be expressed as joint thickness and the mortar strength.
Conclusions
lhe biaxial lensile strenglh of lhe brick was equal lo ils References
uniaxial tensile slrength. Furthermore, it was assumed that the
uniaxial compressive strength of the bricks is equal to its I. Monk, C. B., "A Historieal Survey and Analysis of the
standard strength. In Figure 6-9 the masonry strength com- Compressive Strength of Brick Masonry ," Structural Clay
puted from Equation 6-5 is compared to the experiment al Products Research Foundation , Geneva, Illinois Researeh
results: the compressive strength of lhe test specimens could Reporl No. 12, 1967.
be predicted within a range of ± 20% of the actual values. For 2. Graf, O., "The Load Carrying Capaeity of Masonry,
the higher strength mortars, lh~ l:omputed values generally cspccially of One-Story High Walls ," Fortschritle und
were too low; however. they were too high for the low- Forschungen im Bauwesen, Reihe D, No. 8 , Stuttgart.
strength mortars. This may be due to an erroneous assumption 1952.
of the failure crU.erion for bricks and mortar under triaxial 3. MeBurney, J. H., "The Effect of Strength of Brick on
stress states. The relationship given by Richart , Brandtzaeg and Compressive Strength of Brick Masonry ," Proceedings.
Brown is empirical and is not necessarily the s~me for ali types American Society for Testing and MateriaIs . Vol. 28 , Part
of mortars. Nevertheless, this rati onal approach in the pre- li, 1928 , pp. 605-634.
diction of the compressive strength of masonry units appears 4. Hilsdorf, H. , "Investigation into the Failure Mechanism of
to be promising. Before a general applicability of Equation 6-5 Brick Masonry ," Material prufungsamt f.d. Bauwesen,
could be recommended, variaus characteristic properties of Technische Hochschule, Munich, Report Nr. 40, 1965.
masonry have to be established which were normally not 5. Richart, F. E., Brandtzaeg, A., Brown , R. L., "A study of
investigated in previous studies. These are: the Failure of Concrete under Combined Compressive
I. The coefficient of non-uniformity , U, as a function of (a) Stresses," BuUetin 185, University of lllinois Engineering
quality of workmanship, (b) type and compressive strength of Experiment Station , 1928.
mortar, (c) type of bricks, (d) pattern of masonry unit and
coring of bricks, and (e) thickness of joints.
2. The behavior of bricks and mortars subjected to defined
triaxial stress states.
3. A relationship between the st rength of bricks under
biaxial tensian . uniaxial tensian and flexure . '
Then an approach as outlined in the foregoing may enable
us to safely predict lhe compressive strength af masonry from
praperties of its constituents which can be easily determined
I
in standard tests.
Notations
/J
/
a Coefficient to evaluate Equation 6-5
b
E
Height of brick
Modulus of elasticity V
Ali
)
Uniaxia1 campressive strength af brick b
Biaxial lensile slrength of brick Tangent Modulus
Compressive strength of standard concrete cylin- Brick Only. (Marlor
"Infinitely· S'ift.)
der
I~ 1 I
' (,Tangent Madulus for Brick ond
Un iaxial compressive strength of joint mortar Transi l r Arei Mortor in UnioKiol Compress1on
Compressive strength of masonry unit ./
V/
~.
Compressive strength of standard concrete cylin-
der which is confined by a slress a2
Thickness of mortar joint /
a x ' a y , Oz Stresses in x, y and z-direction
Stress in brick acting in x-direction
V/
°xb
axj
a ym
Stress in mortar joint acting in x-direction
Average masonry stress in y-direction
/ Stroin {- - in./in.l
a2 Confining pressure of a concrete cylinder
U Nonuniformity coefficienl
Uu Nonuniformity coefficient at failure FIgure 6-A. Stress strain curoe.
•
Investigation into the Failure Mechanism of Brick Masonry 41