Professional Documents
Culture Documents
The Use of Characteristic Strengths in Masonry Design
The Use of Characteristic Strengths in Masonry Design
The Use of Characteristic Strengths in Masonry Design
S Lawrence
Centre for Infrastructure Performance and Reliability,
School of Engineering, University of Newcastle
SUMMARY: Despite recent advances in our understanding of masonry behaviour, the design of
masonry structures in Australia in accordance with AS 3700 is often problematic. The recommended
methods of calculation and the relevant material properties do not produce designs commensurate
with some common practices. These common practices have stood the test of time, but in some
cases cannot be justified by present design methods. In this respect, Australian design standards
are no different from their overseas counterparts. This paper examines the design approaches used
by AS 3700 for masonry bending and for wall ties. In particular, the use of characteristic strengths
and their validity in terms of the models of behaviour are examined. Alternative approaches are
proposed, which produce designs more commensurate with common practice and that, it is argued,
more accurately represent the actual structural behaviour.
1 INTRODUCTION
methods of analysis and current assumptions about
Like other structural materials, the design of masonry
material properties. These sections of masonry
in Australia in accordance with the Masonry
appear to have adequate strength and do not exhibit
Structures standard AS 3700 (Standards Australia,
a history of failures, although it could be argued
2001) follows a limit states approach and uses lower
that in some cases they have not been subjected to
five-percentile values of characteristic strengths. In
their full design load. Similarly, common practice
principle, this is intended to ensure an acceptably
for the installation of wall ties is difficult to justify
low probability of failure and, by appropriate
by calculation in accordance with the current
choice of capacity reduction factors, some degree
standard. These tie configurations have been in use
of consistency between designs using the various
for some considerable time and, with the exception
structural materials.
of the Newcastle earthquake in 1989, where specific
However, the approach taken can sometimes lead installation and corrosion resistance faults were
to unintended conservatism, which affects the identified, have not exhibited a history of failure.
economies of usage of masonry. It is commonly These so-called common practices continue to
acknowledged that there are problems justifying be used, either on the basis of deemed-to-satisfy
by calculation some of the common forms of regulations or because of long-standing traditional
construction used for masonry in houses and other methods of construction.
small buildings. Specifically, sections of unreinforced
It is suggested that one of the reasons for this
masonry spanning one-way between top and bottom
apparent conservatism of masonry design standards
supports, for example between the roof structure
is that we are too reliant on characteristic strengths
and the supporting footing or slab, often cannot
of individual components, particularly where our
be justified for even low wind loads using current
behaviour models are somewhat uncertain. This has
* Paper S07-963 submitted 8/02/07; accepted for led to a mismatch between (supposedly rational)
publication after review and revision 14/03/07. designs according to the standard and common
practice. This paper examines possible reasons for
Corresponding author Dr Stephen Lawrence can this apparent conservatism in the current design
be contacted at spl@bigpond.net.au. approach for two areas: bending of unreinforced
masonry and design of wall ties. It recommends comprising eight individual masonry joints, and
rational modifications that will lead to improved the test result was the average strength of these
design standards. beams (0.28 MPa), whereas the strength adopted
for design was the characteristic value (lower five-
There are three areas where conservatism can enter
percentile) of individual joint strengths (0.20 MPa).
the design process. These are the assumed material
The relationship between beam and joint strengths
properties, the capacity reduction factor ( factor)
is set out in AS 3700 Clause D7.2 and the conversion
applied to the material properties and the method
from mean to characteristic strength was based on
of analysis or the assumptions about behaviour of
an assumed coefficient of variation (CV) for joint
the structural element. The paper discusses these in
strength of 0.30. The adjustment ensured that the
relation to design for simple vertical bending and
levels of design strength in the two codes were
wall ties, and explores ways in which designs may
equivalent. Subsequent developments have led to the
be made more realistic by using a design strength
bond wrench test now used in AS 3700 to measure
that is no longer simply equal to the characteristic
the strengths of individual joints directly.
strength of the material. These adjustments are based
on consideration of the structural behaviour models From the above it can be seen that the assumption of
for each type of action. Taking this approach avoids a basic level of flexural tensile strength for individual
the problems of redefining characteristic strengths, masonry joints equivalent to 0.20 MPa (characteristic)
for which a large amount of data are available, has been in use in Australia since 1969. This strength is
and providing different factors for different supported as a minimum level by ample experimental
circumstances, such as vertical and horizontal evidence (McNeilly et al, 1996; Lawrence et al; 2000).
bending. Much masonry construction is known to have higher
strengths than this, and such variation and its effect
In the introductory sections, this paper gives
on the probability of failure should ideally be taken
an outline of the historical development of the
into account in a rational analysis to determine the
current design regime for masonry bending and
factor for use in design. This work is planned as
wall ties in Australia. The factors contributing to
an extension of recent work on compressive strength
conservatism outlined above are then examined
design (Stewart & Lawrence, 2007). It is therefore not
and recommendations for changes to AS 3700 are
considered appropriate at the present time to propose
given.
any variation to the material strength for bending.
Table 1: Characteristic strengths of Type A veneer conditions to which only very few ties in the wall
and cavity ties. will be subjected.
Characteristic strength (kN) A further important point about the tie rating system
Rating is that the ratings are determined in most cases by
Tension Compression the stiffness of the tie, not by its pullout or buckling
Light duty 0.20 0.24 strength. A typical load-deflection relationship
for a tie test shows an initial almost linear section,
Medium duty 0.40 0.48
followed by an extended non-linear section, with little
Heavy duty 1.00 1.20 reduction in load (see figure 1). In the test, the load
corresponding to a deformation of 1.5 mm is taken
values consistent with the lower 5% characteristic as the strength of the tie (shown by a dotted line in
strengths, ensuring that the product ratings of light, figure 1). In other words, the tie is typically behaving
medium and heavy duty remained unchanged since much like a spring and can withstand relatively large
the scheme was first implemented in 1984. deflections while maintaining a high proportion of
The test procedure used to establish tie ratings is its load capacity. Note in relation to figure 1 that the
based on the worst case of loading that the ties will be test was terminated at a deflection just greater than
subjected to in service. Whereas ties are intended to 1.5 mm, although the load capacity of the tie was still
perform their function when aligned straight across a below its ultimate strength. This behaviour provides
cavity (and the water transfer test in AS/NZS 2699.1 considerable scope for redistribution of load among
ensures that they will not transfer water when aligned the ties present in a wall panel.
horizontally), the strength test requires a relative From the above discussion, it can be seen that the
misalignment of the ends of the tie before load is tie ratings have remained essentially unchanged
applied in tension or compression. This is intended since 1984. Because of the ramifications for the tie
to cover an extreme case of misalignment and/or manufacturing industry, it would be inappropriate
relative movement between the leaf/frame on each to propose to change this rating system (either the
side of the cavity, arising from the effects of brick criteria or the test procedure) in order to achieve more
expansion or shrinkage and timber frame shrinkage realistic wall tie designs.
(in the case of brick veneer). As a consequence, the
design strengths quoted in AS/NZS 2699.1 (see table
1) are lower 5% characteristic strengths determined 3 CAPACITY REDUCTION FACTORS
for the worst combination of in-service conditions.
It should be noted that, where relative movements
do occur between the ends of ties in a wall, the 3.1 Bending
movements are not uniform, but range from zero
at the base to a maximum at the top corners of the The design of masonry to AS 3700 involves the
wall. All ties are therefore rated according to the application of (sometimes quite severe) capacity
lower 5% characteristic flexural tensile strength of 1 MPa are possible in good quality masonry
f ‘mt of 0.20 MPa and assumes that failure occurs and it could be argued that averaging is more
when the bed joint that is subjected to the greatest likely with lower strength masonry (which is
bending moment reaches its capacity. The bending represented by an f ‘mt value of 0.20 MPa) than
moment is calculated based on simple bending with higher strength masonry, which would
behaviour (usually wL2/8). Because the f ‘mt value behave in a more brittle way. The effect of any
is a characteristic strength of individual bed joints, averaging of the strengths of adjacent joints
this approach means that the capacity of the whole would be to raise the strength of the panel
bed joint across the full width of the section under above that predicted by the current analysis
consideration is governed by the weakest individual method.
bed joint along its length.
It is proposed that the first two of these possible
There are three possible errors in this approach: errors are difficult to account for in design and
1. The maximum bending moment in a section will counteract each other to some extent, and
of masonry spanning one-way between a roof they should therefore be ignored. In the case of the
or ceiling structure and a footing might be less first, it is difficult to quantify the possible extent of
than that given by the simple bending formula rotational restraint at the ends in the large number of
wL2/8. As a consequence of partial rotational possible configurations. Similarly, the second effect
restraint at either top or bottom, or both, the depends on the height (number of courses) in the
maximum bending moment might be closer to masonry section and it would therefore be difficult
wL2/10, which would indicate a conservatism to implement in design.
of up to 25% in the current approach. Even the The third effect, that of averaging adjacent masonry
self-weight of the masonry will result in some joint strengths, is proposed as something that could
partial rotational restraint at the base, but this be easily considered in design, simply by using a
is usually ignored. modified (equivalent) flexural strength, and that
2. It has been shown previously (Baker & this would go some way to provide a rational basis
Franken, 1976; Lawrence, 1991) that a weakest- to justify many common construction practices. The
link type of behaviour occurs in masonry following sections outline this proposal and the effect
members subjected to simple vertical bending it would have on design to AS 3700.
across the bed joints. By this theory, it will
not be the weakest course that fails, unless 4.3 Effect of averaging
all courses are subjected to the same bending
moment. Rather, when the courses have
various levels of bending moment applied, The effect of averaging adjacent bed joint strengths
the one with the lowest ratio of strength can be assessed as follows:
to applied bending moment will fail first, 1. Consider a population of individual joint
causing failure of the whole member. This is strengths with a lower 5% characteristic value
the basis of the correction factor applied to of 0.20 MPa and a coefficient of variation
the test strengths of bond beams in AS 3700 (CV) of 0.30, corresponding to the generally
Appendix D. The effect of this phenomenon accepted level of variation in masonry bond
would be to reduce the strength of a wall strength (AS 3700 Appendix B). The lower
panel below that calculated by the current 5% characteristic value, assuming a Normal
method of analysis. distribution, is given by:
3. It has been argued (Baker & Franken, 1976) f ′ = X − 1.645σ (1)
that sections of masonry subjected to bending
across the bed joints will fail at a stress level where X is the mean and is the standard
corresponding to some degree of averaging deviation ( = X × CV ). Therefore, we can
between adjacent masonry units. Values of derive X = 0.395 MPa and = 0.199 MPa.
three or five adjacent units were suggested. 2. For the average of three adjacent joint
However, other work (Lawrence, 1991) strengths, the mean strength will also be
examined this approach using test results for 0.395 MPa but the standard deviation will be
narrow sections of masonry and presented an given by:
argument that the weakest of the individual 0.119
bed joints across the width of the section σ3 = = 0.069 MPa (2)
would govern its strength. The results overall 3
are not conclusive and it remains likely that The lower 5% characteristic strength for the
some level of averaging will occur, especially average of three joints will therefore be
in sections of masonry with a width greater f 3′ = 0.395 − 1.645σ3 = 0.282 MPa (3)
than five masonry units. It is commonly
known that flexural tensile strengths in excess 3. Similarly, for the case of averaging five
adjacent joints, we can derive a characteristic relevant wind class (Standards Australia, 2006). The
strength: last column in the table shows whether the design is
f 5′ = 0.395 − 1.645σ5 = 0.307 MPa (4) justified (ie. extreme fibre stress under applied load
is less than factored strength). Each case has been
calculated for f ‘mt equal to 0.20 MPa (the AS 3700
Baker and Franken (1976) suggested that the degree default) and 0.28 MPa (the equivalent value proposed
of averaging of adjacent joints would extend in section 4.3). The latter results are shaded for ease
from three to five masonry units in width. It is of comparison. The AS 3700 factor for bending (0.6)
suggested that, to be conservative, and in view of the has been used.
uncertainty surrounding the extent of averaging of
adjacent joints, the case of averaging over three joints Table 2 shows that two cases currently included in
should be taken. This would result in an equivalent the BCA cannot be justified with f ‘mt equal to 0.20 MPa
characteristic flexural tensile strength for vertical and the AS 3700 factor of 0.6. A 90 mm thick wall
bending design of 0.28 MPa. with 290 x 290 piers at 1650 mm centres, cannot be
justified for N3 wind. However, using f ‘mt of 0.28 MPa
this case can be justified. Similarly, the case of a
4.4 Analysis of common cases
110 mm wall with 350 x 350 piers at 1800 mm centres
cannot be justified, with f ‘mt equal to 0.20 MPa, but
To explore the effects of this proposal, various cases can be justified with f ‘mt of 0.28 MPa. The last row
have been analysed for the effect of averaging three in table 2 shows an additional case (not currently
adjacent joint strengths. The cases include those in the BCA) that can be justified for N3 wind up to
permitted in the BCA Volume 2 for single leaf garage a height of 3000 mm and would be useful in some
walls with engaged piers (Australian Building Codes situations.
Board, 2006).
The cases included as deemed-to-satisfy solutions
The effect of engaged piers has been considered
in the BCA reflect common practice and are not
by calculating a section modulus for the combined
known to present a history of problems. Whereas
section, based on the thickness, width and spacing
they cannot be justified by the approach taken in the
of engaged piers. The worst case, of tension on the
current AS 3700, the modified approach proposed
pier side, has been taken. The arrangement is shown
here allows them to be justified by calculation and
in figure 2. After finding the neutral axis position,
also allows other cases to be rationally designed.
the extreme fibre tension stress at the face of the
The proposed change could easily be implemented
piers governs the design. From the overall section
in AS 3700 Section 12 by permitting an equivalent
modulus, an equivalent thickness for a simple
bending strength f ‘mt equal to 0.28 MPa for vertical
rectangular section can be calculated and these have
bending design in small buildings. At the time of
been shown in table 2 for the cases examined.
writing, a new standard (which is expected to be
The pressure capacity of the wall has then been designated AS 4773.1) is under development for
calculated using simple bending theory, taking the design of masonry in small buildings and it is
account of the self-weight at the mid-height of the proposed that this standard should also incorporate
wall. The results are shown in table 2, with the the proposed approach.
5 BEHAVIOUR OF WALL TIES load prior to failure will lead to more realistic tie
designs in the future.
In the design of wall ties, because of their spring-
like behaviour, there is considerable scope for 5.1 Current basis of design
redistribution of stresses prior to failure. The
treatment of behaviour as a brittle phenomenon The basis for design of wall ties in AS 3700 is that
may therefore be too conservative. Lateral load every tie is designed to resist a certain level of load.
testing of walls has shown (Lawrence, 1978) that This means that the strength of a wall containing
the ties in a typical cavity wall behave elastically as perhaps up to 40 or 50 ties is determined by the single
springs throughout the load history until failure of tie with the highest applied load. In other words, the
the wall. Testing of individual ties (see section 2.2) problem is treated as an elastic-brittle failure type.
shows that they retain a high load capacity through a Given the high level of variability in both tie strength
relatively high range of deformation, allowing them and loads to which the ties are subjected, this seems
to redistribute forces from the more heavily loaded unrealistic.
ties to those less heavily loaded.
The current basis of design for wall ties in AS 3700 has
Furthermore, an assessment of wall test results in the been based on research conducted at the University of
UK (Edgell & West, 1985) showed that reducing the Newcastle (Page et al, 1996). For veneer walls, there
number of ties in a cavity wall with a conventional are two cases considered (see AS 3700 section 7.7).
50 mm cavity did not significantly reduce the lateral The case of a veneer wall with a flexible backing is
load capacity of the wall. The ties were found the most common, where the frame supporting the
to deform while still maintaining sufficient load veneer is constructed of timber or metal framing. In
resistance to distribute load up to the complete failure this case, it is assumed that the masonry veneer will
of the wall. By inference, the same would apply for crack under load, as the framing deflects, and that as
lighter grade ties used at the same spacings (for a consequence, the highest tie forces will occur at one
example, substituting light-duty ties for medium- or two rows of ties near mid-height. However, all the
duty ties at the same spacings). ties in the wall must be designed to resist the same
force and no allowance is made for redistribution of
There have been several investigations of wall tie
forces between ties as they deflect. Considering that
forces, both experimentally and by modelling,
the ties behave as non-linear springs, with their rated
most recently summarised by Yi et al (2003). Recent
force capacity determined (usually) at a deflection
experimental work aims to extend this understanding
of 1.5 mm and having considerable further capacity
(Page et al, 2007). However, these investigations have
for deflection under load, this seems an extremely
generally not considered the ties as springs having
conservative approach.
considerable deformation capacity and randomly
variable characteristics. It is suggested that full The second case for a veneer wall is with a stiff
consideration of this potential for redistribution of backup. This applies to cases where the supporting
wall is solid concrete or another masonry leaf (as, suction. In practice, the leaves are usually supported
for example, in a brick cavity wall). For this case, the in such a way that the inner leaf is firmly attached
tie force is taken as 1.3 times the pressure acting on to the building frame or cross walls, while the outer
the tributary area of the tie, representing the most leaf derives most of its support from the wall ties. A
heavily loaded tie in the wall. In other words, the tie similar situation exists for the veneer leaf in masonry
force is based on the worst case for any tie in the wall, veneer construction. The pressure conditions in
and all ties are dimensioned to withstand the same the cavity are indeterminate, and it follows that a
force with no allowance for any redistribution as the conservative approach to determining tie forces is
ties deform. This is also an extremely conservative to consider only the forces generated by positive or
assumption of behaviour. negative pressure on the external leaf. Not only is the
effect of internal pressure on the inner leaf or backup
The design rules for ties in cavity walls also make framework neglected in this approach, but also the
provision for cases where both leaves are supported, entire load applied to the outer leaf is considered to
in which case the design tie force is that acting on the be transferred through the ties. The capacity of the
tributary area for the tie. However, most cavity walls outer masonry leaf itself to resist part of the applied
have only the inner leaf supported, and therefore force by bending action and transferring reactions
design usually follows the approach for a veneer to its supports is ignored in this approach. Both of
wall with a stiff backup. these factors introduce additional conservatism to
the design by leading to upper bound estimates of
5.2 Wind forces tie forces.
A more detailed analysis of the interaction between
A further factor that must be considered in the design the leaves through the ties, with separately applied
of ties is the appropriate wind pressure to be applied. internal and external wind pressures and appropriate
Design wind pressures on walls are usually derived stiffness for each of the components, would be
from the most severe combination of internal and required before a more accurate model could be
external pressure or suction. For wall ties, ultimate developed. Because tie strength ratings are lower
loads should be used rather than serviceability loads, in tension than in compression, the governing case
because tie failure is considered to be a structural for design is always suction on the external leaf
failure of a component and could result in collapse of producing tension in the ties.
the wall or dislodgement of the outer leaf. However,
the combined effect of internal and external pressure/ 5.3 Design of ties for cavity walls
suction is not necessarily appropriate.
For a cavity masonry wall, if both leaves were It is clear from the discussion in section 5.1 that
supported in an identical way at their edges, the there is scope for load sharing or redistribution
average tie force would be determined by only the between the ties in a wall as they deflect under
difference between internal pressure and external load. The use of a characteristic strength in these
Figure 3: Variation of equivalent characteristic strength with number and COV of ties.
Table 4: Current and proposed tie designs for cavity walls with 600 mm vertical tie spacing.
Table 5: Current and proposed tie designs for veneer walls with 600 mm vertical tie spacing.
Height 2400 (mm) Height 3000 (mm) Height 2400 (mm) Height 3000 (mm)
Wind
450 600 450 600 450 600 450 600
class
N1 L L L L L L L L
N2 L L L M L L L L
N3 L M M M L L L M
N4 M M M H M M M M
N5 H H H H M H H H
N6 H H H H H H H H
C1 M M M M L L L M
C2 M H H H M M M M
C3 H H H H M H H H
C4 H H H H H H H H
The author represents the clay brick industry Standards Association of Australia, 1974, AS 1640
(through Think Brick Australia) on the Australian 1974 Brickwork in Buildings (Metric units).
Standards committees responsible for AS 3700 and Standards Association of Australia, 1988, AS 3700
AS/NZS 2699. 1988 Masonry in Buildings (known as the SAA
Masonry Code).
REFERENCES Standards Australia, 2001, AS 3700 2001 Masonry
Structures.
Australian Building Codes Board, 2006, BCA
Standards Australia, 2006, AS 4055 2006 Wind Loads
Volume 2, Class 1 and Class 10 Buildings – Housing
for Housing.
Provisions.
Standards Association of Australia, 1969, AS CA47 Stewart, M. G & Lawrence, S. J. 2007, “Model error,
1969 Brickwork in Buildings. structural reliability and partial safety factors for
structural masonry in compression”, Masonry
Standards Australia, 2000, AS/NZS 2699.1 2000
International, (In Press).
Built-in Components for Masonry Construction Part 1:
Wall Ties. Yi, J., Laird, D., McEwen, B. and Shrive, N. G. 2003,
“Analysis of load in ties in masonry veneer walls”,
Stewart, M. G. & Lawrence, S. J. 2002, “Structural
Canadian Journal of Civil Engineering, Vol. 30, pp.
reliability of masonry walls in flexure”, Masonry
850-860.
International, Vol. 15, No. 2, pp. 48-52.
STEPHEN LAWRENCE