Ii th INTERNA TIONAL BRICKlBLOCK MASONR Y CONFERENCE

TONGJI UNIVERSITY, SHANGHAI, CHiNA, 14 - 16 OCTOBER 1997

AUSTRALIAN DESIGN MANUAL FOR DIAPHRAGM

WALLS

G. Simundic 1 and A.W. page2

1. ABSTRACT

The use of diaphragm and other walls of geometric cross section was pioneered in the
United Kingdom and has gained wide acceptance in that country. Diaphragm walls are
particularly suited for applications where tall walls are required in structures sue h as sports
halls and other light commercial and industrial buildings. There is scope in Australia for the
use of diaphragm walls not only in the manner developed in the United Kingdom, but also
for domestic construction in high wind areas. This paper gives an overview of a study
which has investigated the potential for the use of diaphragm walls in Australia and the
development of an industry design manual. Several areas of difference between Australia
and the United Kingdom practice are identified and problems resolved. A computer
program for the design of propped cantilever diaphragm walls has been developed and used
to produce a series of design curves for wall selection and incorporation into the industry
design manual.

2. INTRODUCTION

In recent years there has been a major shift away from the use of masonry for commercial
and light industrial structures in Australia because of the impact of precast cOQcrete and tilt-
up construction. The disadvantage of conventional masonry walls in these structures is the
need to provide additional support for the walls in the form of mullions or frarning support

Keywords: Diapbragm; Design Manual; Masonry; Walls

1 Laboratory Manager, Department of Civil, Surveying and EnvironmentaJ Engineering, The University of
Newcastle, NSW, 2308, Australia
2 CBPI Professor in Structural Clay Brickwork, Departrnent of Civil, Surveying and EnvironmentaJ
Engineering, The University ofNewcastle, NSW, 2308, Australia

971
because of their height. The use of masonry diaphragm walls offers aviable altemative.
Engineers in Australia have used the concept of diaphragm walls in isolated instances for
many years, with the design being from first principIes rather than from code provisions.
Specific provisions for diaphragm walls were included in the SAA Masonry Code in 1988
(1), but their use has not been widespread due to the novel nature of the technique and the
ignorance of designers of the potential of the system.

The Australian clay masonry industry has recently recognised the potential of walls of
geometric section, and in collaboration with the University of Newcastle is producing a
design guide for diaphragm walls for the use of structural engineers and architects. This
guide will include both architectural aspects and structural design information with a series
of design curves of lateral loading capacity being provided. A range of wall geometries
have been considered, for use in both domestic or commercial construction.

This paper gives an overview of the design manual and describes a series of experiments
which were carried out as a background to the study. In particular the tests related to the
differences between Australian and British practice with regard to shear connector design
and water penetration performance as described. Tests on shear connectors revealed major
differences with previous British research. The reasons for these differences are identified
and discussed. The water penetration properties of the diaphragm walls have been also
investigated with the emphasis on thin diaphragm walls which are unique to Australia.

3. DESIGN OF DIAPHRAGM WALLS

A detailed review of the design of walls of geometric section for Australian conditions has
been produced by Phipps and Page (2). The design of diaphragm walls will be rarely
govemed by the compressive capacity of the wall, with the flexural resistance usually being
the controlling factor influencing the overall wall dimensions as well as the spacing of the
diaphragms.

The Australian Masonry Standard AS3700 allows walls to be designed for the ultimate
limit state utilising the flexural tensile strength of the masonry. For transient loads a value
of 0.20 MPa can be assumed as the characteristic flexural tensile strength of masonry f'mt,
except at membrane type damp-proof courses and at the interface between masonry and
other materials when it is taken as zero. Because of the possibility of low bond strength (or
the loss of bond strength due to previous cracks) the British approach is to neglect bond
strength and base ultimate strength calculation on the formation of a collapse mechanism
with hinges forrrung at points of maximum bending momento This is an inherently safe
approach and will usually give a capacity which is lower than that predicted if the bond
strength is mobilised (as is the case in this study).

3.1 Design for Compressive Forces

Diaphragm walls are usually used to enhance the lateral load capacity of the masonry.
However diaphragm walls can also be successfully and economically used to support heavy
axialloads, particularly for tall walls. The geometry of the cross section provides increased

972
n'sistance to buckling, with higher compressive capacities than would be applicable to an
equivalent solid wall (3). Since the current Australian Masonry Code AS3700 does not
have specific allowances for geometric sections in its compression provisions, diaphragm
walls are usually conservatively designed by calculating the slenderness ratio using the
overall depth of the wall. Alternatively a "first principies" approach can be used by
ca1culating the equivalent thickness of a solid wall with the same radius of gyration as the
hollow wall. If the simplified method is used, the wall is considered as an equivalent solid
section to obtain the compressive reduction factor, with the actual cross-sectional area of
the wall being used to calculate the compressive capacity. Overall lirnits on sIenderness
ratio are also applied to ensure a rninimum leveI of "robustness".

3.2 Design for Lateral Loads

The analysis of a wall spanning vertically between a footing and a roof and subjected to a
lateralload has been previously presented by Phipps and Page (2), and is described briefly
here. As shown in Figure 1 the peak moments and crack locations are at the walI base and
near rnid-height. If the dead load effects of roof are ignored, assurning for the worst case
that the roof uplift force is balanced by the mass of the capping beam and transrnitted
directly to the footings by a tie down system, the axial load on the wall is its self weight
(P). The value of the moment of resistance at the base MA' will be directly related to the
weight of the wall, and given by (see Figure 1):
T
MA = P x - ,where T =overall depth of the wall. [1]
2
B e ndin g M o m e nt
w per Di ag ra m Befo re

--
u nit l e n g th

-----
:
R oof Collap s e
C
-,
I hinge
, ,

l
I
I

-----
h,
I
A I A

-----
-----
-- 1-
I
I
I
ma x M X
h in g c
M= H

----- 1
b
I
I -
-----
-----
I
I
I
I
/ A
I
I
c...!..- hing e
hb

~
Ln l era l I .na d V////~ MA
~
M oment ofresistance
Colla p se M echanism

at base due to wall

s elf weight
Figure 1. Bending Moment and Idealised Collapse Mechanism for Propped Cantilever Wall

If the flexural strength of the masonry is utilised (the "strength method"), as perrnitted by
the Australian Code, the lateralload capacity is given by:
W=~(MA +2Mmax +2~(Mmax +MJMm.,J [2]
H
where: M max = Cm f~1 Zd +fd.B Zd' f A • = 1jIA", , Cm=capacity reductian factar
and Zd = section modulus.
If the flexural strength of the masonry is ignored (the "mechanism method"), as used in the

973
British approach, the lateralload capacity is given by (see Figure 1):
1.6 ç A", T where : x =!l.
w= li (l-x,) , H'
[3]
ç = density ofmasonry, Aro = cross-sectional area of diaphragm wall.
The spacing of diaphragms will be govemed by ability of the flange to span between them.
The horizontal bending capacity isenhanced by partial two-actipn in some cases (when the
flange is also supported along its bottom horizontal edge), and also by in-plane arching
between -the diaphragms. This enhancement is usually ignored in the designo Any
discontinuity such as vertical control joints must also be considered. Note that if the
diaphragms are widely spaced, not alI ' of the flange is inc1uded ln the cross section
considered for ·vertical bending due to shear lag effects.

3.3. Design for Shear

Walls of geometric section can fail in shear by sliding at the base, by web failure (diagonal
tension cracking) or by failure at the flange-web interface. Consideration of these aspects
have been previously presented (4). One area whichhas not been fully explained is the
behavior of metal c;onnectors across an interface when they are being used to transfer shear
forces (as, for example,between the flange and the web). Figur~ 2 shows a cross section
through a wall in which the leaves are connected to the webs with steel connectors.

v.

Shear force ·on connector Bending of shear connector

Figure 2. Shear Transfer in Diaphragm Walls

The shear force Vc acting on an individual connector is:

~,'= t d X I. x s, [4]
where: s, = vertical spacing of connectors, td :!: diaphragm thickness and Iv = shear stress.
n, is assumed in equation [4] that the connectors transfer alI the shear force ªcross the
weblflange interface with the contribution of the vertical /mortar joint being ignored .

.J;iritish research (5) has established the mechanism of shear transfer across the interface.
The mechanism is as illustrated in Figure 2, with the connectors deforming elastic"ally until
two plastic hinges are formed a distance j apart. British tests have indicated that the two
plastic hinges can be taken to be at 'a distance apart of six times the connector thicklless.
One of the aims of the Australian investigation was to confrrin the validity of this
relationship.

974
4. EXPERIMENTAL PROGRAM

In conjunction with the preparation of the design manual, two sets of tests were carried out
to confirm overseas design recommendations for wall profiles designed specifically for
masonry housing in high wind areas and unique to Australian conditions. The:;e walls
typically consist of two skins of conventional 110 mm brickwork connected by a 50 mm
thick diaphragm (resulting in an overall wall thickness of 270 mm, the same as tbat for a
conventional cavity wall). The major advantage of this configuration is that composite
action is achieved between each leaf, with a consequent significant increase in wall strength.
and stiffness. The diaphragm consists of 30 mm thick paving units, 230 mm wide and 160
mm high, with a 10 mm mortar, collar joint between the paver and the inner and outer skins.
The geometry of the paver is such that the top of the paver coincides with every second
mortar joint, allowing the inclusion of metal shear connectors (usually conventional cavity
ties) at these locations. These act as both shear connectors andas ties holding the two
leaves together. This form of construction is unique to Australia, and is particularly suitable
to masonry housing in cyclonic areas. The Jwo aspects of ' design which required
clarification were the effectiveness of the ties acting as shear connectors, and the
effectiveness of this type of section (and similar).to resist waterpenetration.

4.1 Shear Connector Tests

The shear tests were aimed at checking the British relationships for design of connectors
across a mortar joint between a diaphragm and a flange. To check this relationship two sets
of specimens were constructed and tested in shear, one to simulate th~ normal connection
of a flange to a web (Figure 3, specimens type A,B, C and W), the other to simulate the
thin wall housing detail described earlier (Figure 3, specimens type W). In both cases a
symm«.tric double shear arrangement was ' used for the tests. Initially 15 specimens were
built and tested (referred to as Stage 1). To more clearly define the shear mechanism a
further tests were subsequently performed (referred to as Stage 2).
Sh•• rLo.d

!J peclmens Type A, B, C and W Speclmen8 Type D

Figure 3. Shear Connector Test-Schematic Testing Arrangement

All specimens were of "r' cross section and were built· in the two different configurations.
For the wide wall specimens, a vertical damp proof membrane was incorporated between
the web and the flange to eliminate any, shear contribution from the mortar joint, thus
ensuring that ali shear forces were transfe~ed by the shear connectors.For the narrow wall

975
specimens, no membrane was included, as the performance of the composite section (as
built) was to be checked (in this case the wall ties possibly act as both connectors and
tension ties preventing the two skins from spreading apart). To stabilise the arrangement a
preload of 2.6 kN was applied to each flange representing a stress orO.05 MPa (typical of
that which may be present in practice).

Each I section was loaded symmetrically in shear by applying a uniform verticalload to the
web of the specimen using a hydraulic jack. The effects of flange prestress on the shear
behaviour has been reported by Phipps (6), and this was considered in the assessment of the
results. During the test the relative displacement between the flange and the web was
continuously monitored at four locations by potentiometric transducers. The load was
applied until a large shear displacement (in the order of 30 mm) was obtained. For each
test, load-deflection graphs was plotted for every position and the average of ali four
positions obtat,·n_e'-'d_._ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _-.
LOAD-DEFLECTION GRAPH
SPECIMEN A2 (AVERAGE)
50~-----------------~

Z
6
o<{
40
30 .. ........ ... .. .. .... ..::1:: :::: ::
20
9
10
o
O 2 3 4 5
DEFLECTION (mm)

Figure 4. Shear Connector Test - Typical Load - Average Deflection Graph

A typicalload-deflection curve is shown in Figure 4. Loads corresponding to the change in

slope of the curve (forresponding to the start of hinge formation) and the ultimate load
where the curve reaches a plateau (corresponding to the final hinge mechanism) are
suminarised in Table 1. For purposes of comparison with the British shear results, the
distance between the nominal plastic hinges j was back-ca1culated from the equations:
f x r x u'
j = " for a connector of rectangular cross section, [5]
2 x V;
1.33 x f x d'
or j = " for a connector of circular cross section. [6]
V;
It can be seen from Table 1 that the calculated values for j are significantly different to the
suggested British value of six times the connector thickness, with the value varying from
1.2 to 2.5 for the rectangular cross section connectors, and 0.5 to 0.9 times the diameter for
the circular cross section wire ties. The actual distance between the plastic hinges in the
tests could not be deterrnined with any accuracy b.ecause of the large shear displacement
that was imposed on the specimens during each test. The reason for this large difference
between the obtained value and that used in British design practice was unclear and a
second series of tests were performed to more clearly define this mechanism (Stage 2).

976
Stage 2 Tests
A smaller second series of tests were performed in an attempt to more dearly define the
distance between the plastic hinges on the shear connectors. Five specimens were
constructed for the second stage, one specimen for each wall type. The specimens were
made and cured in the sanle way as the specimens tested in the first stage. The testing set
up and the testing procedures were the same as in the first stage except that the relative
displacement between the flange and the web was carefully monitored and the test stopped
at a shear deflection of about 10 mm. Stopping the test at this smaller displacement allowed
the distance between plastic hinges 011 the shear connectors j to be measured on each
connector after the completion of the test.

The failure loads for each of the tests are summarised in Table 1. The design distance
between the plastic hinges j was back-calculated from the equations 5 and 6 and the actual
distance j was measured from the shear connectors. It can be seen from Table 1 that the
measured values for j are very dose to the suggested British value of six times the
connector thickness. If the distance j is back-calculated from the British Design Equations
(5 and 6) using the observed ultimate load, the calculated values are significantly different,
with the value varying from 0.4 to 2.2 (similar to the first stage results). Conversely this
means that the predicted ultimate load using 6u rather than a smaller value is conservative.
s
Table 1. hear c onnector Test Results - s taqes 1 an d2
SPECIMEN DIMENSIONS FIRST HINGE FINAL MECHANISM DISTANCE
OF SHEAR j
CONNECTORS (SC) LOAD DISTANCE ULTIMATE DISTANCE (measured)
OR WALL TIES (WT) j LOAD j
(width x thicknes x length) (back calculated) (back calculated)
(mm~ (kN) (mm) (kN) (mm) (mm)
AveraC,Je A: se 25x3x120 37 .3 6.2 = 2.1 u 43.3 5.4 - 1.8 u -
AveraC,Je8: se 20x3x120 25.3 7.35 - 2.5 u 31 5.9 - 2.0 u -
AveraC,Je e: se 10x3x120 20 4.68 - 1.6 u 25.3 3.7-1.2u -
AveraC,Je W: WT d = 3.85 mm 16.7 3.40 - 0.9 d 31.3 1.83 = 0.5 d -
AveraC,Je O: WT d = 3.85 mm 180 * 180 * -
A1 se 25x3x120 35 6.6 - 2.2 u 65 3.6 - 1.2 u 18.2 = 6.1 u
81 se 20x3x120 30 6.2 = 2.1 u 70 2.6 - 0.9 u 17 - 5.7 u
e1 se 10x3x120 15 6.2 = 2.1 u 35 2.7=0.9u 18 = 6 u
W1 WT d = 3.85 mm 17 3.3 = 0.9 u 34 1.7-0.4u 21 - 5.5 iJ
(*) Could nol be calculaled beca use of mfluence of lhe mortar Joml on lhe shear slrenglh .
d = diameler of conneclor
u = Ihickness of conneclor

The factors affecting the location of the plastic hinges have been reported by Phipps and
Montague (6) who indicate that the design equation ignores the influence of factors such as
prestress (prestress enhances connector strength), the mortar around the connector (the
mortar contributes to shear strength), and the unit strength (the higher the unit strength the
greater the shear strength). The design equation is thus conservative as it only considers the
contribution of the connector itself, and ignores the above effects. With this conservative
approach, the value of j is 6u. If the other factors are considered, the value of j approaches
the observed 'value of (0.4 to 2.2) u. The connector carries the shear force across the joint
in bending, making local contact in the masonry joint, and deforming elastically until two

977
plastic hinges are fQnned (at the position of maximum moments) j apart. The connector
beaÍ-s on portions of the s~ounding masonry in a "prying" mechanism, producing bending
in the connector. This results in the maximum moment in the connector being produced at a
point inside the shear interface. The distance between two plastic hinges j is thus greater
than the thickness of the vertical mortar joint, with the distance being :influenced by the
factors mentioned above.

4.2 Water Penetration Tests

rbe water penetration tests were aimed at checking the water penetration characteristic of
the diaphragm walls. Five typical wall configurations were selected from the list of the
walls to be inc1uded in the Design Manual. The results of these tests have been reported
elsewhere (7). From the test results for thin walls in a severe exposure conditions it would
be advisable to use a membrane on the interface of the flange and the web. For other
exposures, and for thickerwalls', water penetration appears not to be a problem. The results
further highlighted the ' dependence of water penetration on workmanship and the shear
connector detail. '

5. DESIGN MANUAL

In order to encourage the use of diaphragm walls by designers, the Australian c1ay masonry
industry is in the process of preparing a "user friendly" desígn manual. The research
described in this paper provides background to the manual.

5.1 Overview of the Manual

The Design Manual wiU be used by structural, engineers and architects, and will therefore
cover both architectural and structural aspects. The Design Manual wiU contain infonnation
on the following:

, -1\RCHITECTURAL DESIGN
Facade Options; Brick Type and Bond Pattem; Web DetaiIing; Openings - Wiridow and
Door Detailing; Roof / Wall Junction; Control Joints - Movement Gaps, Articulation Joints

- CONSTRUCTION ASPECTS
Foundations and Footings; Material SeIection; Construction Sequence, Propping, Weather
Protection; Bond Beam and Roof Tie-Down; Roofing; Finishes and Cleaning

- PERFORMANCE ASPECTS
Rain Resistance; Fire Resistance LeveIs; Acoustical Properties; Thennal PerfonniUlce

- STRUCTURAL ASPECTS
Conceptual Behaviour; Cross-Sectional Properties; Design for Compression; Design for
Lateral Loads; Design for Shear
- DESIGN CHARTS
Outline ofWalI Types; Design Charts; Worked Examples

978
There are ten suggested wall types, with the wall thickness varying from 230mm to 590
mm. The walls are made of stan,dard or modular bricks with the diaphragms engaged,
ladder (every second course engaged with no units in the courses between), or tied (see
Figure 5).

The design curves are presented for a range of wall types. For every wall type four curves
are presented: one based on the mechanism method; and three on the strength method (for
different values of fmt of 0.2 , 0.3 and 0.4 MPa). Each graph includes three or four curves
corresponding to different distances between diaphragms. This allows the designer some
flexibility in choosing the bonding pattem.
c ro 55 - 'cc rio n .Ellla..1i2Jl.

ENGAGED

LADDER

~c:::::::::J'D'r--'
Lo- IL--==r:::==::c::l
I

TIED

Figure 5. Typical Example ofWall Types

The designer is left to choose the type of the wall írom a list of suggested wall geometries.
Once the wall geometry has been chosen, either the Mechanism Method or tqe Strength
Method for calculating the lateralload capacity must be norninated. If the Strength Method
is chosen, an appropriate value for f~t must also be norninated, depending on the expected .
leveI of supervision on the designed building and the confidence in achieving the required
bond strength. The designer can then select the appropriate chart, and for the required ,
height of the diaphragm wall determine the lateral load capacity of the wall (in kPa) and the
corresponding value of the clear distance between diaphragms d.

5.2 Design Curves

The design curves for lateral load capacity were generated from a ,computer program
written for the design of propped cantilever diaphragm walls using the- provisions of the
Australian Masonry Code, AS3700 (l). Both the "strength" and "mechanism" methods of
determining the lateral load capacity were included in the program, with design curves
being generated for both methods. The choice of method in the Design Manual is left' to the
designer. The design curves for lateral load capacity have been plotted from the results of
the analyses of each wall type. The intention of presenting design curves is 'to allow the
designer to easily choose the lateralload capacity for a given wall geometry from the list of
the suggested diaphragm wali geoIIletries presented. These curves ,cover wall heights from
2 to 8 metres. Some of the curves exhibit discontinuities in slope in some locations because

979
of a change in the failure mechanism governing the designo Usually for low height walls
shear resistance is the governing factor in assessing the lateral load capacity. For taller
walls lateralload capacity governs with the capacity being governed by the vertical bending
requirement. This is in turn influenced by the effective flange width which is a direct
function of the height of the wall.

6. SUMMARY AND CONCLUSIONS

Diaphragm walls are a viable alternative to precast or tilt-up construction for tall walls or
walls subjected to high lateralloads. Although this type of wall has been used in the United
Kingdom for some time, the concept has not gained wide acceptance in Australia, mainly
due to the ignorance on the part of engineers and architects of the effectiveness of the
system. To overcame this ignorance and encourage greater use of masonry, an industry
based design manual for diaphragm walls has been prepared. This paper has given an
overview of the design manual and described some background research aimed at assessing
the viability of the concept for Australian applications.

7. ACKNOWLEDGMENTS

The research described in this paper was partially funded by the Clay Brick and Paver
Institute of New South Wales (Australia). Their support, and donation of material by its
member companies (particularly Boral Bricks NSW Pty Ltd and PGH Bricks NSW), is
gratefully acknowledged.

8. REFERENCES

1. The Standards Association of Australia, "Masonry in Buildings", AS3700-1988,

Sydney, Australia.
2. Phipps, M.E. & Page A.W., "Developments in Masonry, Part 1 - Walls of
Geometric Cross-Section," Research Report No. 103.09.1994, The University of
Newcastle, Australia, 1994.
3. Curtin, W.G. & Shaw, G., "Brick Diaphragm Walls in Tall Single-Storey
Buildings," The Brick Development Association, London, November 1977.
4. Simundic, G., & Page, A-W. , "Australian Developments in the Use of Walls of
Geometric Section", Proceedings of the 7th North American Masonry Conference, The
University of Notre Dame, United States, 1966, pp. 1007-1018.
5. Phipps, M.E., "Diaphragm and Other Walls of Geometric Cross Section in the
UK," Proceedings of the 3rd National Masonry Serninar, Brisbane, Australia, 1994, pp.
24.1-24.10
6. Phipps, M.E., & Montague, T. L, "The Behaviour and Design of Steel Shear
Connectors in Plain and Prestressed Masonry", Proceedings of the 7th North American
Masonry Conference, The University ofNotre Dame, Indiana, 1966, pp. 789-798.
7. Simundic, G., "DiaphragnÍ Walls", Master of Engineering Thesis, The University
of Newcastle, Australia, 1997.

980