Available via license: CC BY-NC-ND 4.0
Content may be subject to copyright.
45
DR WIBKE DE VI LLIERS is a senio r lecturer at
Stellenbosch University, where she obtained
her BEng (20 06), MSc Eng (2008) and Ph D
(2019), the latter in the d evelopment of
performance-based specications for
alternati ve masonry uni ts. As part o f the
Unit of Cons truction Ma terials, her rese arch
interests lie in alternative and sustainable
constru ction materi als, and their str uctural and t hermal perf ormance. She
has super vised a number of p ostgraduate s tudents and au thored or
co-autho red several pub lications over t he past nine year s. Prior to this,
she spent t wo years at Aurecon a s a structur al design engin eer.
Contact details:
Department of Civil Engineering, Stellenbosch Universi ty
Private B ag X1, Matieland 7602, Stellenb osch, South Africa
T: +27 21 808 4072, E: wdv@sun .ac.za
PROF GIDEO N VAN ZIJL (Pr Eng) is a prof essor
in structural engineering at Stellenbosch
University. He obtained his Bachelor’s (1986)
and Master ’s (1990) degrees in Ci vil
Engineering from Stellenbosch Universit y,
and his PhD (200 0) from Delf t Universit y in
the Nethe rlands. As Dire ctor of the Cen tre for
Development of Sustainable Infrastruc ture,
his researc h interests a re 3D construc tion printi ng, struct ural and
computational mechanics, including the development, charac terisation, and
constit utive and durab ility modell ing of advanced co nstruct ion materials.
Contact details:
Department of Civil Engineering, Stellenbosch Universi ty
Private B ag X1, Matieland 7602, Stellenbosch, South Africa
T: +27 21 808 4436, E: gvanzi jl@sun.ac.z a
PROF BILLY BOSH OFF (Pr Eng) is a prof essor in
civil engineering at the University of Pretoria.
He was head of the Structural Engineering
Division at Stellenbosch University for seven
years, and se rved as the pr esident of the
Concrete S ociety of So uthern Afri ca from
2012 to 2014. He completed h is PhD at
Stellenbosch University in the eld of
bre-rei nforced concr ete in 2007, and has publishe d over 100 paper s in
journals an d conference pro ceedings. Hi s research inter est is in the eld o f
construction materials, but more specically bre-reinforced concrete,
concrete c racking in the f resh state, a s well as unconventi onal and
eco-friendly construc tion materials.
Contact details:
Address at t he time of prepar ing this paper:
Department of Civil Engineering, Stellenbosch Universi ty
Private B ag X1, Matieland 7602, Stellenb osch, South Af rica
Current address:
Faculty of Engineering, Built Environment and Information Technology
University of Pretoria
Private B ag X20, Hateld 0 028, South Afri ca
T: +27 12 420 2746, E: billy.bo sho@up.ac.z a
Keywords: low-income housing, National Building Regulations,
SANS 10400, S outh African load ing code, concrete
masonry, simplied micro-model
De Villie rs WI, Van Zijl GPAG, Bosho WP. Review of com patibility be tween SANS 10400 d eemed-to- satisfy maso nry wall provis ions and loading co de.
J. S. Afr. Inst. C iv. Eng. 2021:63(1), Art. #1062, 16 pages. h ttp://dx.doi.org /10.17159/2309- 8775/2021/v63n1a5
TECHNICAL PAPER
JOURNAL OF THE SOUTH AFRICAN
INSTITUTION OF CIVIL ENGINEERING
ISSN 1021-2019
Vol 63 No 1, March 2021, Pages 45–60, Paper 1062
INTRODUCTION
The South African government has pro-
vided nearly 3 million subsidised housing
units since 1994 (Department of Human
Settlements 2017). However, a backlog of
over 2 million units persists (Sisulu 2016).
Government-subsidised housing units are
typically a stand-alone dwelling of 40 m2
(Laubscher 2014), containing a kitchen,
living area, two bedrooms and a bathroom
(see Figure 1).
The structural design of housing
in South Africa is regulated by the
Application of the National Building
Regulations, based on the National
Building Regulations and Standards Act of
1977 (RSA 1977). The standard was first
published in 1985 but has since been updat-
ed several times to the current edition
SANS 10400 (SANS 2010a). On a practical
level, all housing construction in South
Africa is regulated by the National Home
Builders Registration Council (NHBRC),
the establishment of which is enshrined
in The Housing Consumers Protection
Measures Act (RSA 1995; NHBRC 2015).
The NHBRC stipulates general home build-
ing technical requirements and guidelines
in the form of the Home Building Manual
and Guide, which is based on SANS 10400.
Both SANS 10400 and the Home
Building Manual rely on normative refer-
ence standards for the material-specific
design aspects, which have typically been
prescriptive in nature. For masonry, this
reference standard is SANS 10164 The struc-
tural use of masonry (prescriptive-based),
as well as the recently adopted SANS 51996
Design of masonry structures (performance-
based). SANS 10400 is performance-based
in nature but contains extensive deemed-to-
satisfy solutions, the typical mixed approach
taken in transitioning from prescriptive to
performance-based regulation.
A number of developments over the
past decade or two warrant a reconsidera-
tion of these deemed-to-satisfy solutions
in SANS 10400, specifically with regard to
masonry walling solutions:
QLoading: The South African loading
code SANS 10160 2011 (SANS 2011) has
been revised, in the form of an adaption
Review of compatibility
between SANS 10400
deemed-to-satisfy
masonry wall provisions
and loading code
W I de Villiers, G P A G van Zijl, W P Bosho
South Africa has a housing shortage estimated in excess of 2 million units. This backlog is being
addressed predominantly with the construction of 40m2 low-income, single-storey, detached,
state-subsidised houses built with conventional concrete masonry units, regulated by the
Application of the National Building Regulations, SANS 10400. However, several developments
warrant a reconsideration of SANS 10400 deemed-to-satisfy masonry wall provisions. Two
critical configurations of single-storey, unreinforced, bonded masonry walls are generated,
based on these deemed-to-satisfy provisions. Subsequently, a simplified micro-scale finite
element model is used to analyse these configurations under serviceability and ultimate limit
state loading conditions. Characterisation tests of the concrete masonry material, together with
numerical fitting to test data and data taken from literature, generate the necessary parametric
input. The numerical analyses reveal that in half of the load cases, the walls’ resistances
failed to achieve the design load as required by the South African loading code. A significant
shortfall was found for the out-of-plane resistance against the wind load, as well as structural
vulnerability under seismic loading due to the geometric layout permitted by the deemed-to-
satisfy rules. This indicates that the SANS 10400 provisions for masonry wall panel geometries
require reconsideration, especially given the recent revision of the South African loading code.
Volume 63 Number 1 March 2021 Journal of the South African Institution of Civil Engineering46
of EN 1990 (EN 1990) and EN 1991 (EN
1991), with notable changes and addi-
tions. The design of single-storey mason-
ry structures must take seismic loading
into account more comprehensively in
certain areas of the country, and signifi-
cant improvements have been made to
the South African wind datamap.
QCategory1 buildings: The Application
of the National Building Regulations
(SANS 2010b) has been revised with
significant changes, including the intro-
duction of Category1 buildings, specifi-
cally aimed at drawing more low-income
structures into a regulatory framework.
QAdoption of Eurocode 6: The South
African masonry industry is in the
process of updating its suite of masonry
standards to the EN approach, Eurocode
6 Design of Masonry Structures, mark-
ing a transition from prescriptive to
performance-based standards.
QAdvances in numerical analyses: A
significant amount of research has been
conducted internationally, using finite
element modelling, to better understand
the discontinuous behaviour of masonry
structures, and these advances need to
be taken into consideration in the speci-
fications for masonry.
QOutdated mechanical limits: Current
mechanical limits set in the South
African prescriptive standards of
conventional masonry units are largely
based on yield line theory analysis (JSD
1995), taken from the withdrawn British
Standard BS 5628-1 (BS 1978).
This paper therefore investigates the
response of conventional concrete masonry
walls in the context of South African low-
income housing (LIH) by means of finite
element (FE) analysis. The analyses are
performed on single-storey, unreinforced,
single-leaf, external masonry walls, which
conform to the deemed-to-satisf y solutions
of SANS 10400. Two critical wall layouts
are identified (W1 and W2), modelled in
DIANA FE analysis software and subjected
to three load conditions as required by SANS
10160: the serviceability limit state (SLS) and
the ultimate limit state for wind (ULS-W)
and seismic (ULS-S) actions. The results of
the analyses provide insight into the in-plane
and out-of-plane structural behaviour of
conventional concrete masonry walls of LIH
housing, relative to the expected loading.
FINITE ELEMENT MODEL
Modelling approach
Significant advances in numerical methods
and computational capabilities in recent
decades have altered the way in which
masonry is analysed. For masonry finite
element modelling, two main approaches
have been established, namely macro- and
micro-modelling, with the level of abstrac-
tion directly related to the complexity
and size of the problem to be analysed,
(Gia mbanco et al 2001; Reyes et al 2008;
Roca et al 2010; Abdulla et al 2017).
Macro-modelling assumes a smeared
continuum approach, where the unit, mortar
and unit-mortar interface behaviours are
combined in a representative continuous
material. In contrast, micro-modelling repre-
sents a high degree of detail where the unit,
mortar and unit-mortar interface are mod-
elled distinctly. Simplified micro-modelling
(SMM) is a subset of micro-modelling as its
name implies, wherein the units are modelled
as expanded elements, with solely elastic
material properties, to encompass the volume
of the unit and the mortar in order to main-
tain the overall geometry. The relatively weak
mortar joint and unit-mortar interface are
combined into a single zero-thickness inter-
face element in which the nonlinear material
behaviour is concentrated (Lourenço 1996).
The SMM approach is considered the most
appropriate for this study in terms of the
wall scale to be analysed and the computing
requirements (Figure2refers).
Constituent material model
The constitutive material model chosen is
the Combined Cracking-Shearing-Crushing
(CCSC) model, implemented in DIANA. The
plasticit y-based model is defined by a multi-
surface yield function, shown in Figure3,
consisting of a tension cut-off, a Coulomb
Figure 1 Typical 40 m
2
government subsidised concrete masonry house: (a) plan (CMA 2011,
reproduced with the kind permission of the CMA) and (b) under construction
(a)
(b)
Journal of the South African Institution of Civil Engineering Volume 63 Number 1 March 2021 47
friction criterion and an elliptical compres-
sion cap. The depiction of the composite yield
surface in 3D in Figure 3(b) does not include
the elliptical compression cap, but it has been
implemented in DIANA (2017).
WALL CONFIGURATIONS
In an effort to make buildings that meet
regulatory requirements more affordable
to low-income communities and to reduce
the immense health and safety risk that
unregulated informal structures present
(Watermeyer 2004), the Joint Structural
Division (JSD) of the South African
Institution of Civil Engineering (SAICE)
developed a new category of buildings in
2000 (Watermeyer & Milford 2003). This
Category1 building type was introduced
in The Application of the National Building
Regulations: SANS 10400 in2004.
These buildings are restricted in size and
occupancy class. The floor area is limited
to 80 m2, wall lengths to less than 6 m
between lateral supports, and the structure
to a single storey with no basement. The
wall thickness can be as little as 90 mm,
compared to 140mm in non-Category1
buildings (SANS2011b). Occupancy classes
are restricted to places of instruction or
worship, small shops, offices, dormitories,
domestic residences and dwelling houses.
Category1 buildings also make allowance
for lower performance levels with regard to
size and serviceability limits, but no distinc-
tion is made between the categories regard-
ing seismic and wind loading.
Deemed-to-satisfy requirements
Society’s expectation of a wall ’s perfor-
mance is implicitly represented by the
deemed-to-satisf y masonry wall solutions
in SANS 10400-K (SANS 2011b). Therefore,
these solutions are used to identify appropri-
ate wall configurations for the LIH context.
Additional recommendations and limitations
regarding the geometry, roof configuration,
energy use and seismic loading, as set out in
SANS 10400-A (SANS2010b), SANS10400-L
(SANS2011c), SANS10400-XA (SANS2011d)
and SANS10160-4 (SANS 2017) respectively,
are also taken into consideration. Two differ-
ent wall configurations, panel wall W1 and
gable wall W2, are derived by selecting the
most extreme and critical combinations of
distance between lateral supports, wall height
and openings. These two wall configurations
are detailed in Figure 4, where the hatched
areas designate positions of return walls
which provide lateral support.
The selection criteria for these two
representative single-leaf wall layouts are
detailed in Table1. Although only the walls
are modelled, the walls are set within the
context of representative 40m2 Category1
houses to derive loading and support con-
ditions. The roof construction is assumed
to be timber, with metal sheetcovering.
Typical reinforcement according to
SANS10400-K (SANS 2011b) of 5.6 mm
steel rods and 2.8 mm brickforce (Figure5)
Figure 2 SMM approach: (a) in 2D with expanded unit elements (adapted from Lourenço & Rots 1997) and (b) in 3D with solid brick elements and
2Dinterface elements (adapted from Macorini & Izzuddin 2011)
h
zero thickness
h
h
h+h
16-noded
interface
element for
mortar joints
20-noded solid
element for bricks
20-noded solid
element for bricks
16- noded
interface
element for
mortar joints
hu
hm
hm
Potential crack
in the unit
Interface
elements (joints)
hu + hm
Zero thickness
Continuum elements (units) (b)(a)
Figure 3 Combined Cracking-Shearing-Crushing yield criterion: (a) in 2D (adapted from Lourenço
1996) and (b) in 3D (adapted from Van Zijl 2000)
σ
co
ft
co
τt
τt
τs
τs
cap
mode
initial
yield surface
intermediate
yield surface
residual
yield surface
coulomb
friction mode
|
τ
|
σ
tension
mode
Tension mode
Residual yield surfaceInitial yield surface
Cap mode
Intermediate
yield surface
Coulomb
friction mode |τ|
σ
σ
ft
τt
τt
co
co
τs
(b)
(a)
Volume 63 Number 1 March 2021 Journal of the South African Institution of Civil Engineering48
is modelled in the bed joints above the open-
ings, instead of concrete lintels. The rod
reinforcement yield strength is taken as the
required proof stress of rod reinforcement
by SANS 10400-K (SANS2011b), namely
485N/ m m2. Whilst brickforce proof stress
is not specified in SANS 10400, tensile tests
conducted by Talocchino (2005) on typical
South African brickforce found a proof yield
stress of 500 N/mm2. Therefore, the brick-
force yield strength is taken as 485N/mm2 as
well. The elastic modulus of both reinforce-
ment types is taken as 200000N/ mm2.
Boundary conditions
In modelling the support conditions of the
wall models, a number of assumptions need
to be made. The foundations are modelled
as fixed, as depicted in Figure 6, hence fully
supported with no potential for differential
settlement. Pin supports on the short return
Figure 4 Wall W1 (top) and W2 (bottom) layout and dimensions
2 650 mm
300 mm 900 mm 450 mm 3 000 mm
630 mm630 mm1 010 mm1 010 mm
300 mm
4 950 mm
2 646 mm
756 mm
6 000 mm
1 008 mm882 mm
750 mm 900 mm 600 mm 750 mm3 000 mm
Table 1 Selection criteria for representative houses and wall layouts
Layout Selection Clause Standard
Wall effective
thickness W1 140 mm* B.3.3 (a) SANS 10160-4 (2017)
Wall length
W1 6.0 m 3.6 (c) SANS 10400-A (2010b)
Table 1, Panel C SANS 10400-K (2011b)
W2 5.0 m 3.6 (c) SANS 10400-A (2010b)
Tables 5 & 6 SANS 10400-K (2011b)
Wall height
W1 2.7 m** Table 1, Panel C SANS 10400-K (2011b)
B.3.3 (b) SANS 10160-4 (2017)
W2 2.6 m** Fig 4 SANS 10400-K (2011b)
B.3.3 (b) SANS 10160-4 (2017)
Roof slope 15° 4.2.2.1 SANS 10400-L (2011c)
Truss spacing 1.2 m Table 4 SANS 10400-L (2011c)
Openings Various
Fig 6(a), Table 7 SANS 10400 -K (2011b)
6.2.2 SANS 10160-4 (2017)
4.4.4 SANS 10400-XA (2011d)
Reinforcement 5.6 mm rods Tables 20, 21 & 23, Fig 27 SANS 10400-K (2011b)
2.8 mm brickforce B.3.3 (d), (f) SANS 10160-4 (2017)
Vertical control joint none Tabl e 19 SANS 10400-K (2011b)
* shear wall teff requirement of 190 mm is not met (SANS 10160-4 (2017) B.3.3 (a))
** shear wall heff /teff < 17 requirement is not met (SANS 10160-4 (2017) B.3.3 (b))
Secondary: 5.6 mm × 2
Primary: 5.6 mm × 3
(equivalent of)
Brickforce: 2.8 mm × 2
Brickforce: 2.8 mm × 2
Brickforce: 2.8 mm × 2
Primary: 5.6 mm × 3
Primary: 5.6 mm × 2
(equivalent of)
Secondary: 5.6 mm × 2
Brickforce: 2.8 mm × 2
Brickforce: 2.8 mm × 2
Primary: 5.6 mm × 2
Figure 5 Bed joint reinforcement above
openings for wall W1 (top) and W2
(bottom)
Journal of the South African Institution of Civil Engineering Volume 63 Number 1 March 2021 49
walls provide lateral support to the modelled
walls, whilst allowing for some rotation of
these joints. It is assumed that the roof truss
system does not provide substantial lateral
load transfer, based on the type of connec-
tion and the typical poor quality of the con-
nection between the roof and walling in LIH.
Therefore the roof line is modelled as unsup-
ported laterally for both wall configurations.
MATERIAL INPUT PARAMETERS
Material characterisation
Solid conventional concrete masonry units
form the basis of the masonry analysed,
together with 10 mm mortar joints. The
blocks have a length of 290 mm, width of
140 mm and height of 116 mm. Extensive
material parameter characterisation is
required for the selected constituent mate-
rial model, CCSC, that was selected in
DIANA. Table 2 contains the material input
parameters that were used to define the
concrete masonry, as well as the method by
which they were determined. Experiments
were conducted by Fourie (2017) to deter-
mine several of the unit, crack and joint
interface parameters, indicated by EXP
in Table 2 under Method. Finite element
analysis (FEA) was used to determine the
tensile strength of the crack interface and
the compressive fracture energ y and equiva-
lent plastic relative displacement of the
joint interface by numerically fitting data to
experimental data. Suitable literary sources
(LIT) were used to determine the remain-
ing parameters. For further details on the
material input parameters or the process to
obtain them see De Villiers et al (2 018).
Characteristic values of the current
material input data cannot be established
since the data is statistically insufficient.
Figure 6 Boundary conditions for wall W1 (left) and W2 (right), inner perspective
Table 2 Input parameters for wall numerical analyses
Parameter Symbol DIANA Method Value Unit
Unit
Density ρu–EXP 2090 kg/m3
E-modulus EuYOUNG EXP 177 0 0 N/mm2
Poisson’s ratio νuPOISSON LIT 0 .16 –
Crack interface
Tensile strength ft,c TENSTR FEA 0.66 N/mm2
Mode I fracture energy GIf,c GF EXP 0.047 N/mm
Cohesion ccCOHESI LIT 1.0 N/mm2
Friction angle φcPHI LIT 37 °
Dilatancy coefficient ψcPSI LIT 0 °
Mode II fracture energy GIIf,c MO2VA L LIT 0.47 N/mm
Compressive strength fc,c COMSTR EXP 12.1 N/mm2
Shear traction contrib Css,c CS LIT 1.0x10-3 –
Compr fracture energy Gc,c GC LIT 10.0 N/mm
Eq plastic relative displ κp,c DUPEAK LIT 0.030 mm/mm
Tangential stiffness ks,c DSSX/Y LIT 763x10-3 N/mm3
Normal stiffness kn,c DSNZ LIT 1770 x10 -3 N/mm3
Joint interface
Tensile strength ft,j TENSTR LIT 0.12 N/mm2
Mode I fracture energy GIf, j GF LIT 0.005 N/mm
Cohesion cjCOHESI EXP 0.17 N/mm2
Friction angle φjPHI EXP 49.5 °
Dilatancy coefficient ψjPSI LIT 0 °
Mode II fracture energy GIIf,j M O2VAL LIT 0.05-0.08σ N/mm
Compressive strength fc,j COMSTR EXP 5.5 N/mm2
Shear traction contrib Css,j CS LIT 0.7 –
Compr fracture energy Gc,j GC FEA 18 .0 N/mm
Eq plastic relative displ κp, j DUPEAK FEA 0.030 mm/mm
Tangential stiffness ks,j DSSX /Y LIT 214 N/mm3
Normal stiffness kn, j DSNZ LIT 520 N/mm3
Volume 63 Number 1 March 2021 Journal of the South African Institution of Civil Engineering50
The use of nominal values is allowed in
such instances according to SANS 10160-1
(SANS 2018a), and it requires the use of
mean values for the parameters govern-
ing the structural stiffness. The choice of
input parameters, as well as the modelling
approach, was validated experimentally
in a separate process for both the in-plane
and out-of-plane behaviour by comparing
numerical results to experimental data
of dedicated large- and medium-scale
masonry wall tests (see De Villiers 2019).
Material partial factors
Material partial factors make a significant
contribution to the design process in limit
states design. The main considerations in
both SANS 51996-1-1 (SANS 2018c) and
SANS 10164-1 (SANS 1989) in determining
partial factors for materials are manufactur-
ing and construction or execution control.
The greater the certainty regarding the unit
and mortar strength and the manner in
which they are combined on site, the greater
the reward is in terms of the load magnitude
that the masonry is allowedtobear.
Given the inadequate guidance for
the execution control classification in
Eurocode6 (Sýkora & Holický 2010), and the
disparity in classifications, a direct compari-
son of the material partial safety factors is
not reasonable. However, in broad terms, the
SANS 51996-1-1 (least conservative) ranges
from 1.5 to 3.0, the UK National Annex to
Eurocode 6 of 2005 (BSI2005) from 2.3 to
3.0, and SANS10164-1 (most conservative)
from 2.9 to 3.5.
Despite their importance, material par-
tial factors are not included in this study.
Admittedly, this may result in a less con-
clusive evaluation of the concrete masonry
walls and the design loads applied to them,
which are adjusted with partial factors.
However, including the material partial fac-
tors renders the finite element output less
clear, impacting in particular the fracture
behaviour of the concrete masonry as well
as the investigation into the relative impor-
tance of the parameters.
DESIGN LOADS
SANS 10400-B (SANS 2012) requires
the structural strength and stability of
the structure to be assessed by loading it
with the relevant actions as determined
according to the South African load-
ing code, SANS 10160 (SANS 2011). All
relevant design situations are taken into
consideration to arrive at the most critical
combinations of permanent, imposed, wind
and seismic actions for the two wall con-
figurations under consideration. The three
load cases considered are the ser viceability
limit state (SLS) and the ultimate limit
states under wind (ULS-W), based on basic
fundamental wind speeds of 44, 40 and
36 m/s respectively, and seismic (ULS-S)
actions. The latter are determined by means
of the equivalent lateral static force method.
Table3 summarises the factored loads
applied in the numerical analyses for the
vertical loads on the roof and the horizontal
out-of-plane (OP) and in-plane (IP) loads on
the walls, whereas Table 4 details the partial
load factors for the three load cases, accord-
ing to SANS10160-1 (SANS2018a).
Load assumptions
The assumptions made in determining
these critical load cases are detailed in the
following three sub-sections according to
self-weight and imposed load, wind load
and seismic load.
Self-weight and imposed load
The self-weight of the walls is based on
the density determined experimentally,
as detailed in Table 2. The roof assembly
consists of six bay Howe type trusses,
assuming a timber density of 5 000 N/m3
according to Table A.4 of SANS 10160-2
(SANS 2011a) for the structural pine,
and 0.5 mm metal sheeting with a self-
weight of 39.5 N/m2 according to Table
A.5 of SANS 10160-2 (SANS 2011a). The
roof is classified as an inaccessible roof
according to Table 5 of SANS 10160-2
(SANS 2011a) and loads for normal main-
tenance and repair of 400 N/m2 would be
included. However, since an additional
compressive load on the walls is favour-
able, the load combination nullifies the
roof-imposedload.
Wind load
The loads due to wind actions are deter-
mined according to SANS 10160-3 (SANS
2018b). The pertinent parameters are
summarised in Table 5 and assumptions
discussed thereafter.
In most instances, the parameter result-
ing in the most critical load is selected.
The basic fundamental wind speed is taken
as the highest value for any area in South
Africa of 44 m/s; however, loads based on
Table 3 Critical design loads for SLS, ULS-W and ULS-S to SANS 10160
(N/mm2)
Roof OP
IP
Self-Weight Wind Zone A Zone B
SLS W1–44 –10.1 × 10–3 43.9 × 10–3 1.4 × 10–3 1.1 × 10–3 24.4 × 10–3
SLS W1–40 –10.1 × 10–3 36.3 × 10–3 1.2 × 10–3 0.9 × 10–3 20.2 × 10–3
SLS W1–36 –10.1 × 10–3 29.4 × 10–3 0.9 × 10–3 0.7 × 10–3 16.3 × 10–3
SLS W2–44 – – 2.3 × 10–3 1.0 × 10–3 17.6 × 10–3
SLS W2–40 – – 1.9 × 10–3 0.8 × 10–3 14.5 × 10–3
SLS W2–36 – – 1.5 × 10–3 0.7 × 10–3 11.8 × 10–3
ULS–W W1–44 –9.0 × 10–3 117.1 × 10–3 3.7 × 10–3 2.8 × 10–3 65.1 × 10–3
ULS–W W1– 40 –9.0 × 10–3 96.8 × 10–3 3.1 × 10–3 2.3 × 10–3 53.8 × 10–3
ULS–W W1–36 –9.0 × 10–3 78.4 × 10–3 2.5 × 10–3 1.9 × 10–3 43.6 × 10–3
ULS–W W2–44 – – 6.0 × 10–3 2.7 × 10–3 46.9 × 10–3
ULS–W W2–40 – – 5.0 × 10–3 2.2 × 10–3 38.7 × 10–3
ULS–W W2–36 – – 4.0 × 10–3 1.8 × 10–3 31.4 × 10–3
ULS–S W1 – – 0.8 × 10–3 0.8 × 10–3 53.9 × 10–3
ULS–S W2 – – 1.0 × 10–3 1.0 × 10–3 56.4 × 10–3
Table 4 Load combination partial factors according to SANS 10160-1 (SANS 2018a)
Load case Self-weight Roof imposed Wind Seismic
SLS 1.0 0.0 0.6 –
ULS-W 0.9 0.0 1.6 –
ULS-S 1.0 0.0 –1.0
Journal of the South African Institution of Civil Engineering Volume 63 Number 1 March 2021 51
basic fundamental wind speeds of 40 and
36 m/s are also included in the results for
comparative purposes. The terrain cat-
egory is chosen as the most likely scenario
for single-storey residential structures
in a suburban or peri-urban setting. The
default topography factor is chosen, on
the assumption that it is unlikely that low-
income housing is developed on extreme
terrain, which is costly to construct on.
The highest air density value is chosen, to
result in the highest critical load.
Additionally, SANS 10400-B (SANS2012)
specifies minimum wind pressures to be
applied to housing structural systems of
370N/m2 and to housing structural elements
of 450 N/m2. The peak wind pressures
determined according to SANS10160-3
(1213N/ m2) (2018) exceed these minimum
load requirements.
Seismic load
The loads due to seismic actions are
determined according to SANS 10160-4
(SANS 2017). The pertinent parameters are
summarised in Table 6 and assumptions
discussed thereafter.
The highest peak ground acceleration
for natural seismicity in South Africa is
selected. The most unfavourable of ground
types is chosen and the selected build-
ing importance factor is commensurate
with a typical residential structure. The
behaviour factor for unreinforced masonry
is used, given that minimum detailing and
reinforcement requirements are adhered
to. The fundamental period of vibration is
chosen based on structural system type.
The reliability redundancy factor is not
present in the parent standard, Eurocode8
(EN 1998), but was taken from the Uniform
Building Code (ICBO 1997) and introduced
in the South African loading code to com-
pensate for a lower nominal peak ground
acceleration of 0.1 g (Wium 2010). However,
the UBC permits a reliability redundancy
factor range of 1.0 to 1.5, compared to
a range of 1.2 to 1.5 in SANS 10160-4
(SANS2017). The lower limit of 1.2 was
set to compensate for the higher behaviour
factors for reinforced concrete shear walls
used in the UBC (ICBO 1997) compared to
Eurocode 8 (EN 1998; Wium2010).
For determining the seismic design
load, the reliability redundancy factor is
chosen as the lower limit of the allowable
range (1.2 to 1.5), hence less conservative,
for two reasons. First, a higher peak ground
acceleration of 0.15 g was selected for the
analyses, not 0.1 g. Second, the lower limit
of 1.2 in SANS 10160-4 (SANS 2017) was
introduced to compensate for the higher
behaviour factors of reinforced concrete
shear walls. This discrepancy in behaviour
factors is less relevant for this study, given
that a consistent behaviour factor for unre-
inforced masonry of 1.5 is used. It would
hence be justifiable to use a reliability
redundancy factor of 1.0. However, compli-
ance with SANS 10160-4 (SANS2017)
is considered salient and a factor of 1.2
isused.
Load application
The critical design loads, detailed in
Table3, are applied to the two wall con-
figurations, as shown in Figure7 for the
SLS and ULS-W, and in Figure8 for the
ULS-S. Loads that act out-of-plane (OP) of
the wall, are applied as a uniform distrib-
uted load over the entire wall, including
the wind or seismic load, as applicable.
Under wind action, the most critical
load case is the modelled wall acting as
side wall in the context of a 40 m2 house
Table 5 Wind load parameters to SANS 10160-3 (SANS 2018b)
Parameter Symbol Value Clause
Fundamental value of basic wind speed νb44 m/s 7.2.2
Terrain category - C Tab le 2
Terrain roughness factor cr(z) 0.73 7.3.2, Table 3
Topography factor c0(z) 17.3.3
Air density ρ1.2 kg/m3Ta ble 4
Peak wind pressure qp(z) 1213 N /m27.4, Equation 6
Table 6 Seismic load parameters to SANS 10160-4 (SANS 2017)
Parameter Symbol Value Clause
Peak ground acceleration ag0.15 g 5.2, Figure 1
Ground type – 4 5.1.2, Figure 2, Table 2
Building importance factor γ11.0 7.3, Table 3
Reliability redundancy factor ρ1.2 7.3, Equation 6
Behaviour factor q1.5 8.2, Table 4
Fundamental period of vibration factor CT0.05 8 .5.2.1
Figure 7 Load applications for wall W1 (left) and W2 (right) for SLS and ULS-W (N/mm
2
)
IP IP
Roof wind
OP Zone B
OP Zone A
OP Zone B
OP Zone A
Volume 63 Number 1 March 2021 Journal of the South African Institution of Civil Engineering52
structure. The OP wind loads are there-
fore differentiated into Zones A and B,
according to SANS 10160-3 (SANS2018b).
The total OP force (N) is applied uniform-
ly over the masonry portions of the model
walls to account for the lack of surface
area over the wall openings.
The horizontal in-plane (IP) load is dis-
tributed over the full height and thickness
of the wall. The load arises from the lateral
loads on the walls adjacent to the wall
modelled, and includes either the wind or
seismic load, as applicable. The IP load var-
ies linearly for the wind load case, with the
maximum (presented in Table 3) applied
at the top of the wall. Using the principle
that lateral loads are applied at the location
of the mass, the IP load for the seismic
load case is distributed uniformly over the
height of the wall.
The vertical loads transferred from the
roof to the wall are applied at each truss
support point and distributed over the width
of masonry block to prevent stress concen-
trations at these points. The roof self-weight
and the wind load, if applicable, are included
in this load. Vertical uplift is indicated by a
positive value, whereas a compressive force
is indicated by a negative value. Since W2 is
a gable wall configuration, which does not
support trusses, the roof load is only applied
to the W1 configuration.
In the past, numerical and experimental
studies on unreinforced masonry walls have
been focussed on the effect of either OP or
IP loading. In more recent research, the sig-
nificance of the interaction of these two load
conditions has gained prominence, but still
focused on masonry infill walls. Few numer-
ical investigations have taken into account
the combined effect of IP and OP actions
on load-bearing unreinforced masonry, and
even fewer experimental studies (Milani
2008; Agnihotri et al 2013; Najafgholipour
et al 2013; Dolatshahi et al 2015). Typically
the findings are that the IP load may have a
crucial effect on the OP capacity of the wall.
The wall slenderness and aspect ratios large-
ly determine the interaction level. Therefore,
this study applies the simultaneous action of
OP and IP loading.
RESU LTS
Results overview
Figure 9 provides an interpretation
key for results discussed later. The OP
deflections provided in Figures 12 and 17
are measured at the top midspan position,
as indicated in Figure 9. The IP displace-
ments provided in Figures 13 and 18 are
measured at the top left corner of each
respective wall. Typical crack positions
found in the numerical analyses are also
identified in Figure 9, for the later discus-
sion on crack damage classification and
crack width in Figure 14.
Typical failure modes are presented in
Figure 10 for the SLS and ULS-W and in
Figure 11 for the ULS-S. Compressive fail-
ure or crushing is not identified in either of
the wall configurations for both OP and IP
failures. This is not remarkable given the
small structure size and low vertical loads.
The contours in Figure 10 indicate the OP
deflections and OP failure dominates for
both the SLS and ULS-W, since most of the
total load applied is in the lateral direction.
The ‘Base 1’ crack, located in the lowest
bed joint of the wall, indicates tensile fail-
ure, together with shear failure in columns
adjacent to the door openings (‘Door’) for
wall configurations W1 and W2 for both
load cases. Subsequent tensile cracks also
form in the pier adjacent to the window
opening (‘Pier 1’ and ‘Pier 2’).
IP IP
OP OP
Figure 8 Load applications for wall W1 (left) and W2 (right) for ULS-S (N/mm
2
)
Base 2
Door
IP displacement OP deflection
Pier 3
Base 1
Pier 3
Pier 4
Pier 1
Base 1
Lintel
IP displacement
OP deflection
Door
Pier 2 Pier 4
Base 2
Pier 2
Pier 1 Lintel
Figure 9 Crack position and deflection/displacement measurement legend (W1 left, W2 right)
Journal of the South African Institution of Civil Engineering Volume 63 Number 1 March 2021 53
The contours in Figure 11 indicate the
IP displacements. IP failure dominates for
the ULS-S load case, due to shear action
in the walls bearing most of the seismic
load. Tensile/flexural cracks dominate,
starting with ‘Base 2’ next to the door
and ‘Pier 1’ to ‘Pier 3’ surrounding the
window opening. Crack onset at ‘Base 2’
is tensile, and progresses to sliding shear,
whereas ‘Pier 4’ is a combination of shear
and tensile stepped cracks. The W2 wall
configuration is made particularly vulner-
able by the slender column to the left of
the door opening.
As an overview, the design loads and
numerical wall resistance determined
through the numerical analyses are pre-
sented in Table 7 for the three load cases,
namely SLS, ULS-W (for three basic funda-
mental wind speeds of 44, 40 and 36 m/s)
and ULS-S. For each instance of the critical
load direction, the ratio of the design
load to the numerical wall resistance is
included. Failure is therefore indicated by a
ratio of greater than 1.0.
Out-of-plane response
The OP load/deflection responses for W1
and W2 are depicted in Figure 12 for the
three load cases. The South African load-
ing code (SANS 10160 2011) OP design
loads for each of the three load cases are
also included to contextualise the results.
In Table 3 the design loads were provided
in the form of pressures (N/mm2) but are
converted to forces (N) to facilitate the
comparison of the design loads and the
resistance capacities of the walls.
Of the six analyses presented in
Figure12, in half of them the OP design
load significantly exceeds the OP
load-carrying capacity of the walls, namely
in W1 ULS-W, W2 ULS-W (both for all
three basic fundamental wind speeds) and
W2 ULS-S. It is important to note that,
in line with recent findings in literature
(Vaculik 2012; Derakhshan et al 2018), the
OP response for the ULS-S is inadequate
and that OP behaviour of unreinforced
masonry cannot be disregarded under
seismic action. SANS 10400-B (SANS 2012)
imposes a 1:175 deflection limit on such
building walls, which is well above the
Figure 10 Typical failure for SLS and ULS-W for walls W1 (left) and W2 (right)
Figure 11 Typical failure for ULS-S for walls W1 (left) and W2 (right)
Table 7 Summary of OP SLS and ULS-W loads and IP ULS-S loads
SLS/ULS OP/
IP
SANS 10160
design load Numerical resistance Design load /
numerical resistance
[N] [N[ [N] [N] [-] [-]
W1 W2 W1 W2 W1 W2
SLS-44 OP 12482 11530 1608 6 21430 0.8 0.5
SLS-40 OP 10315 9 529 16 086 21 430 0.6 0.4
SLS-36 OP 8356 7 718 160 86 21 430 0.5 0.4
ULS-W-44 OP 33285 30746 15 597 16543 2.1 1.9
ULS-W-40 OP 27508 25 410 15597 16 543 1.8 1.5
ULS-W-36 OP 22 281 20 582 15 59 7 16 5 43 1.4 1.2
ULS-S IP 19 613 20 539 25964 11295 0.8 1.8
Volume 63 Number 1 March 2021 Journal of the South African Institution of Civil Engineering54
OP deflection range encountered in these
analyses of between 2.5 mm and 7 mm.
The gable of the W2 configuration was
not buttressed, contrary to the specifica-
tions of SANS 10160-4 (SANS 2017). It was
presumed that this lack of lateral support
to the gable would cause instabilities in the
analyses, but this element was noncriti-
cal in the OP loading conditions under
consideration. This is most probably due
to the presence of more vulnerable, slender
elements in the wall in other locations. A
wall with better-proportioned openings
may well cause the gable to become critical,
requiring buttressing.
In-plane response
The IP load/displacement responses for W1
and W2 are depicted in Figure 13 for the
three load cases. The South African load-
ing code (SANS 10160 2011) IP design loads
for each of the three load cases are also
included to contextualise the results.
Reflective of the OP response, in three
of the six analyses presented in Figure 13,
the IP design load significantly exceeds
the IP load-carrying capacity of the walls,
namely in W1 ULS-W, W2 ULS-W (both
for all three basic fundamental wind
speeds) and W2 ULS-S. Notably, the
seismic IP capacity of W1 is 2.3 times
greater than W2. This is explained in part
by the particularly weak geometry derived
forW2.
The IP displacement of both of the
wall configurations is not significant (less
than 1 mm). Arguably, the IP displace-
ment would be greater if the load/dis-
placement path is continued numerically,
but the laborious work of overcoming
the post-peak divergence prevents this
pursuit. There are no limitations specified
for IP displacement in SANS 10400 as
they are for IP def lections. However, the
pronounced reduction in load-carrying
capacity, together with the negligible IP
displacement, is typical of the extremely
brittle behaviour of masonry.
Crack damage classification
The crack damage is classified and pre-
sented in Figure 14 for W1 and W2. Only
the dominant crack for each load case
combination is included for clarit y. The
frame of reference for the crack widths
is taken from the damage categories and
maximum crack widths in SANS 10400-B
(SANS 2012) and the South African Home
Building Manual (NHBRC 2015). The
damage categories vary from less than
0.25 mm, classified as negligible, to greater
than 25 mm, classified as very severe.
These classifications were developed by
Watermeyer and Tromp (1992) to serve
as serviceability performance criteria for
masonry walls and were subsequently
included in SANS10400.
Several t ypical crack positions were
identified in Figures 9, 10 and 11 for the
SLS and ULS-W and ULS-S load cases
respectively. The most dominant crack
is identified for each of the analyses per-
formed and plotted against the OP loads for
W1 and W2 in Figure 14. Cracks occurred
in a number of typical positions for the
IP and OP-dominant loading conditions,
as illustrated in Figure 10, as well as in
Figure11 for the ULS-W and in Figure12
for the ULS-S. For each analysis performed,
the most dominant crack is identified and
plotted against the OP load in Figure14
(left) for W1 and (right) for W2. For both
W1 and W2, the ‘Base’ cracks as well as
cracks around the window openings in the
‘Piers’ are prolific. Most cracks measured
in the numerical analyses of the concrete
masonry walls fall below the ‘negligible’
(less than 0.25 mm) classification. With
further development of the walls’ post-peak
responses, the cracks would undoubtedly
widen. However, in the numerical analyses
performed, these initial cracks suffice to
demonstrate crack development and to
produce a significant reduction in the load-
carrying capacity, which is typical of the
brittle nature of masonry.
W1 OP load (N)
35 000
30 000
25 000
20 000
15 000
10 000
5 000
0
W1 OP deflec tion (mm)
3210
SANS 10160 ULS-W 44 m/s
SANS 10160 ULS-W 40 m/s
SANS 10160 ULS-W 36 m/s
SANS 10160 ULS-S
NUM SLS
NUM ULS-W
NUM ULS-S
Figure 12 Out-of-plane response of walls W1 (left) and W2 (right) for SLS, ULS-W and ULS-S
W2 OP load (N)
35 000
30 000
25 000
20 000
15 000
10 000
5 000
W2 OP deflection (mm)
742
0
0 631 5
NUM SLS
SANS 10160 ULS-W 44 m/s
SANS 10160 ULS-W 40 m/s
SANS 10160 ULS-W 36 m/s
SANS 10160 SLS 44 m/s
SANS 10160 SLS 40 m/s
SANS 10160 SLS 36 m/s
SANS 10160 ULS-S
NUM ULS-S
NUM ULS-W
SANS 10160 SLS 40 m/s
SANS 10160 SLS 36 m/s
SANS 10160 SLS 44 m/s
Journal of the South African Institution of Civil Engineering Volume 63 Number 1 March 2021 55
Summary
The progression of most of the analyses
was limited by divergence. Several steps
were taken to move the analyses past these
points of diversion, such as adjusting the
step size, increasing the tolerance of the
convergence criteria, employing other
iterative procedures, etc, and the arc length
method was used throughout. However,
convergence was rarely achieved. It is clear,
though, from the load-displacement or
load-deflection trace, that the linear-elastic
0.05
W1 IP load (N)
30 000
25 000
20 000
15 000
10 000
5 000
0
W1 IP displacement (mm)
0.200.150.100
SANS 10160 ULS-W 44 m/s
SANS 10160 ULS-W 40 m/s
SANS 10160 ULS-W 36 m/s
SANS 10160 SLS 44 m/s
SANS 10160 SLS 36 m/s
SANS 10160 ULS-S
NUM SLS
NUM ULS-W
NUM ULS-S
Figure 13 In-plane response of walls W1 (left) and W2 (right) for SLS, ULS-W and ULS-S
W2 IP load (N)
25 000
20 000
15 000
10 000
5 000
W2 IP displacement (mm)
0.80.2
0
00.6
0.4
NUM SLS
SANS 10160 ULS-W 44 m/s
SANS 10160 ULS-W 36 m/s
SANS 10160 SLS 44 m/s
SANS 10160 SLS 36 m/s
SANS 10160 ULS-S
NUM ULS-S
NUM ULS-W
SANS 10160 ULS-W 40 m/s
SANS 10160 SLS 40 m/s
SANS 10160 SLS 40 m/s
W1 OP load (N)
35 000
30 000
25 000
20 000
15 000
10 000
5 000
0
W1 crack width (mm)
0.30.20.10
SANS 10160 ULS-W 44 m/s
SANS 10160 ULS-W 40 m/s
SANS 10160 ULS-W 36 m/s
SANS 10160 SLS 44 m/s
SANS 10160 SLS 40 m/s
SANS 10160 SLS 36 m/s
SANS 10160 ULS-S
NUM SLS Base 1
NUM ULS-W Base 1
NUM ULS-S Pier 4
Figure 14 Crack damage classification for walls W1 (left) and W2 (right) for SLS, ULS-W and ULS-S
W2 OP load (N)
35 000
30 000
25 000
20 000
15 000
10 000
5 000
W2 crack width (mm)
0.3
0
00.20.1
NUM SLS Base 1
SANS 10160 ULS-W 44 m/s
SANS 10160 ULS-W 40 m/s
SANS 10160 SLS 44 m/s
SANS 10160 SLS 40 m/s
SANS 10160 SLS 36 m/s
SANS 10160 ULS-S
NUM ULS-S Pier 2
NUM ULS-W Pier 1
Negligible
0.25 mm
SANS 10160 ULS-W 36 m/s
Negligible
0.25 mm
Volume 63 Number 1 March 2021 Journal of the South African Institution of Civil Engineering56
region has been surpassed and that post-
peak global stiffening is improbable.
Given the conventional strength and
stiffness range of the concrete masonry
studied, as well as the typically conservative
nature of deemed-to-satisfy solutions, it was
anticipated prior to the study that the resis-
tance of these conventional masonry walls
would exceed the design loads in both ULS
load cases. Contrary to this expectation,
both wall configurations failed to resist the
ULS-W design loads for basic fundamental
wind speeds of 44, 40 and 36 m/s, as did W2
for the ULS-S design loads, and by a large
margin. Plausible origins of these failure
are the applied design load, the material
input parameters, the derived geometry of
the wall configurations and the assumed
boundary conditions. These aspects are
discussed in the following section.
DISCUSSION
Design load
Most of the assumptions or selections
detailed for determining the wind load
were made to achieve the most critical
wind loading, not the most likely. However,
even design wind loads based on 40 and
36 m/s basic fundamental wind speeds,
which account for the majority of areas in
South Africa, exceed the walls’ capacities in
mostinstances.
It is also noteworthy that the wind pres-
sure determined in this study (1 213 N/m2,
Table 5) is over three times the minimum
wind pressure specified in SANS 10400-B
(SANS 2012) (370 N/m2). The substantially
higher design load for the ULS-W case is in
part due to the recent revision of the wind
loading code, SANS 10160-3 (SANS 2018b).
The wind load partial factor has
increased from 1.3 to 1.6 and the highest
fundamental basic wind speed from 36 m/s
to 44 m/s. These two changes combined
result in a 1.5 times higher load than
would have been the case before these revi-
sions. However, the ratios of design load
to resistance are 2.1 and 1.9 for original
configurations of W1 and W2, respectively.
These revisions alone do not account for the
discrepancy, and they were implemented for
good reason. The reliability performance
of the original wind load partial factor of
1.3 was found to be inadequate (Botha et
al 2018) and the South African wind map
has improved due to, in part, a seven-fold
increase in the historical extreme wind data
available in South Africa (Kruger et al 2 017).
Material
To investigate the potential increase in OP
load-carrying capacity due to improved
material properties, reasonable maximum
values for the three most influential OP
parameters (joint tensile strength, joint
cohesion and joint modeI fracture energy)
were sought in literature. Thereafter,
ULS-W load analyses were performed on
W1 and W2 with these three adjusted
joint parameters. Experimental data on
the joint tensile properties is scarce, but
reasonable maximum values for modeI
fracture energy and cohesion were found
in literature for normal density concrete
blocks with general purpose mortar joints,
conducted by Van Der Pluijm (1999),
and are detailed in Table 8. All other
parameters, as provided in Table 2, are
kept constant.
The outcome of these analyses (Table9
and Figure 17) shows an increase in the
load-carrying capacity of 53% for W1 and
11% for W2, due to the improved material
properties. This reduces the discrepancy
to the most critical design load by 47%
and 13% for W1 and W2, respectively.
Increasing the critical joint parameters
to reasonable maximum values does not
increase W2’s resistance sufficiently to
withstand any of the wind design loads.
In the case of W1, the wall ’s resistance is
increased such that the ULS-W load case
based on the 36 m/s basic fundamental
wind speed can be resisted, but the 40 and
44m/s not.
Table 8 CON adjusted joint parameters
Joint interface parameters Symbol Baseline Adjusted Unit
Tensile strength ft,j TENSTR 0.12 0.84 N /mm2
Mode I fracture energy GIf,j GF 0.005 0 . 0 11 N/mm
Cohesion cjCOHESI 0.17 1.17 N/mm2
1 200 mm 900 mm 1 200 mm 1 500 mm 1 200 mm
756 mm1 008 mm882 mm
2 646 mm
6 000 mm
2 650 mm
900 mm 750 mm
630 mm630 mm1 010 mm1 010 mm
4 950 mm
900 mm 1 500 mm 900 mm
Figure 15 Wall layouts and dimensions for reduced window openings for W1 (top) and W2 (bottom)
Journal of the South African Institution of Civil Engineering Volume 63 Number 1 March 2021 57
It is important to recall that in all the
analyses of W1 and W2, mean material
parameter values are used, and the material
resistance has not been reduced by means of
material partial safety factors. Applying this
necessary reduction for ULS-based design
would further widen the discrepancy.
Geometry
The limitations on wall panel sizes and
openings set out in the SANS 10400-K
deemed-to-satisfy solutions are taken from
the JSD Code of Practice: Foundations and
Superstructures for Single Storey Residential
Buildings of Masonry Construction (JSD
1995). Different wall panel configurations
were analysed using the yield line approach
to derive the panel sizes, and the then
current South African masonry structural
design code SABS 0164-1 (SABS 1980) was
applied to the respective elements to derive
the opening limitations (Watermeyer 1996).
The total area of openings for both
W1 and W2 falls within the specifications
of seismic design principles set out in the
loading code, SANS 10160-4 (SANS 2017)
Clause 6.2.2, of being less than one third
of the overall wall area. The openings
are positioned “as uniformly as possible”,
but given the large opening length of 3 m
permitted in the deemed-to-satisfy solu-
tions of SANS 10400-K (SANS 2011b), it
does result in large openings at both wall
ends, which is undesirable according to
the seismic design principles of the loading
code.
To investigate the potential improve-
ment in OP resistance of both walls under
ULS-W loading and IP resistance for W2
under ULS-S loading due to more robust
geometry, the original window opening
length is halved to 1 500 mm, and the door
and window openings are positioned in less
extreme positions in the wall, as illustrated
in Figure 15 for W1 and W2. All other origi-
nal dimensions of the walls aremaintained.
The outcome of the OP ULS-W
analyses (Table 9 and Figure 17) shows
an increase in the load-carrying capacity
of 20% for W1 and 15% for W2, due to
the reduced window opening and the less
extreme positions of the openings. This
reduces the discrepancy to the most critical
design load by 17% for both W1 and W2.
The outcome of the IP ULS-S analyses
on W2 (Table 9 and Figure18) shows an
increase in the load-carrying capacity of
160% for W2, due to the reduced window
opening and the less extreme positions
of the openings. The IP resistance of W2
now exceeds the seismic design load by
40%. Reducing the window opening by half
does not increase the walls’ OP resistances
sufficiently to withstand any of the wind
design loads. However, this mitigation
strategy significantly increases W2’s IP
load-carrying capacity to successfully resist
the full seismic design load.
Boundary conditions
The conservative assumption was made
that the timber truss system provides
negligible lateral support to the top of the
walls. The effect of this assumption could
be meaningful, but its validity is sustained
given the similarly weak OP resistance
of the opposite wall, which is meant to
provide the additional lateral resistance, as
well as the typically poor quality of con-
nection between truss and wall in LIH.
A potential source of error could be
excessive rotation of the short return walls,
which provide lateral restraint to the walls.
The pinned modelling of the walls could
underestimate the rotational restraint
that a full-length return wall would offer,
thereby allowing greater OP deflection.
To investigate the effect of this, the trans-
lational restraint on the return walls is
applied to all nodes in the boundary plane,
as opposed to just the central row of nodes,
as shown in Figure 16.
The outcome of the OP ULS-W
analyses (Table 9 and Figure 17) shows
an increase in the load-carrying capacity
of 15% for W1 and 10% for W2, due to
the adjusted boundary conditions and
increased rotational restraint. This reduces
the discrepancy to the most critical design
load by 14% for W1 and 12% for W2.
Increasing the rotational restraint provided
by the return walls does not increase the
walls’ resistances sufficiently to withstand
any of the wind design loads.
CONCLUSIONS
The FE analyses were executed success-
fully for the two wall configurations under
three load cases. The failure modes can
be broadly classified as tensile for the
OP-dominant cases and a combination of
tensile/shear failure for the IP-dominant
cases. The analyses revealed the wall
Figure 16 Baseline (left) and adjusted (right) boundary conditions for return walls
Table 9 Peak OP resistance to ULS-W and IP to ULS-S for adjusted material parameters, geometry
and boundary conditions
Peak resistance [N] OP W1 OP W2 IP W2
Baseline 15597 16543 11 29 5
Adjusted material parameters 23835 18373 –
Adjusted geometry 18687 18 9 46 29534
Adjusted boundary conditions 18 0 11 18197 –
Volume 63 Number 1 March 2021 Journal of the South African Institution of Civil Engineering58
configurations’ failure to resist the design
loads in most instances, and significantly
so in the OP response to the ULS-W load
case. This is in part, but not exclusively,
due to recent increases to two important
parameters in the wind loading code.
Three mitigation strategies were
employed, namely improving the tensile
performance by increasing three critical
joint material parameters, improving the
distribution of openings, and increas-
ing the rotational restraint of the return
walls. None of these strategies improved
the resistance of the walls to the point of
successfully resisting the full design wind
load. However, reducing the length and
improving the distribution of openings
significantly increased the IP resistance of
W2 to seismic loading.
Additionally, it is important to recall
that these analyses were performed without
the use of material partial safety factors.
Current standardised partial factors for
conventional masonry materials range
between 1.5 and 3.5. Reducing the material
strength by these factors, as is required by
limit states design, would further signifi-
cantly increase the discrepancy between
the walls’ resistances and the design loads.
This raises the issue of possible incon-
sistencies between the deemed-to-satisfy
provisions in SANS 10400-K for wall
W1 OP load (N)
35 000
30 000
25 000
20 000
15 000
10 000
5 000
0
W1 OP deflec tion (mm)
4320
SANS 10160 ULS-W 44 m/s
SANS 10160 ULS-W 40 m/s
SANS 10160 ULS-W 36 m/s
Baseline
Adjusted material
parameters
Figure 17 OP ULS-W response for adjusted material parameters, geometry and boundary conditions for W1 (left) and W2 (right)
W2 OP load (N)
35 000
30 000
25 000
20 000
15 000
10 000
5 000
W2 OP deflection (mm)
3
0
0 2
1
SANS 10160 ULS-W 44 m/s
SANS 10160 ULS-W 40 m/s
SANS 10160 ULS-W 36 m/s
1
Adjusted
geometry
Adjusted boundary
conditions
Baseline
Adjusted
geometry
Adjusted material
parameters
Adjusted boundary
conditions
W2 IP load (N)
35 000
30 000
25 000
20 000
15 000
10 000
5 000
0
W2 IP displacement (mm)
0.80.20.10
Figure 18 IP ULS-S response for reduced window opening for W2
0.60.4
Baseline
Adjusted
geometry
SANS 10160 ULS-S
Journal of the South African Institution of Civil Engineering Volume 63 Number 1 March 2021 59
panel and opening sizes and the current
EC-based loading code, both for wind, i.e.
SANS 10160-3 (SANS 2018b) and seismic,
i.e. SANS 10160-4 (SANS 2017) loads.
Based on a case study of extreme wind
loads on an inland housing development in
South Africa, Mahachi et al (2018) came to
the conclusion that a review of the techni-
cal standards in housing development is
necessary, specifically the NHBRC Home
Building Manual. Griffith (2000) reports
on a similar case of discrepancies between
the capacity of the ‘deemed-to-comply’ wall
provisions of the South Australian Housing
Code and the Australian masonry stan-
dard’s design load. It has long been found
that, within the field of masonry buildings,
low-rise, unreinforced ones with light
roofs (such as LIH), experience the most
wind damage (Sparks et al 1989), especially
non-engineered ones, relying on empirical
design procedures. This, coupled with the
significant changes in the required wind
and seismic design loads with the revision
of the South African loading code, war-
rants a review of the SANS 10400 deemed-
to-satisfy wall provisions.
It is recommended that this is done with
the preferential housing solution in mind, by
first specifying desired wall configurations,
based on constructability, typical South
African building practice and skill level, fen-
estration requirements for building energy
usage, natural lighting and ventilation,
and fire safety. Numerous such specifica-
tions are well documented in the relevant
parts of SANS 10400 and can be used as a
basis to determine the desired geometr y.
Subsequently, the derived wall configura-
tions can be structurally analysed using the
simplified micro-modelling approach.
ACKNOWLEDGEMENTS
Funding: This work was supported by the
National Research Foundation of South
Africa (grant numbers 87961, 106965).
REFERENCES
Abdulla, K, Cun ning ham, L & Gi llie, M 2 017.
Simulat ing ma sonry wall behaviou r using a
simplified micro-model approach. Engineering
Structures, 151: 349–365.
Agni hotri, P, Singhal, V & Rai, D 2013. Effect of
in-plane damage on out-of-pla ne strength of
unreinforced masonr y wal ls. Engine ering Structures,
57: 1–11 .
Botha, J, Retief, J & Viljoen, C 2018. Reliability
assessment of the South Afr ican w ind load design
formulation. Journal of the South African Institution
of Civil Engineering, 60(3): 30–40.
BS (Briti sh Standard) 1978. BS 5628-1:1978. Code of
Practice for the Use of Mas onry. Structural Use of
Unreinforced Masonry. London: Briti sh Standards
Institution.
BSI (Briti sh Stand ards Institution) 2005. NA to BS EN
1996-1-1. UK National Annex to Eurocode 6. Design
of Masonr y Stru cture s. Gene ral Rules for Reinforced
and Unreinforced Masonry Structures. London: BSI.
CMA (Concre te Manufactu rers Association) 2011.
Initiatives. http://www.cma.org.za/Initiatives/
The-CMA-House.
De Villiers, W I 2019. Computational and e xperimental
modelling of masonry walling t owards performance-
based standardisation of alternative masonr y units
for low-income housing. PhD Thesi s. Stellenbosch
University.
De Villiers, W I, Fourie, J & Boshoff, W P 2018.
Numerical model ling of a lternat ive masonry u nits.
Proceedings, International R ILEM Symposium on
Concrete Modelli ng (CONMOD18), 26–29 August
2018, Delft, The Netherlands.
Department of Human Settlements (DHS) 2017.
Deliver y of Ser viced Sites and Houses/ Units f rom
HSDG. Pretoria : DHS.
Derak hshan, H, Lucas, W, Visintin, P & Griff ith, M
2018. Out-of-plane st rengt h of existing t wo-way
spann ing sol id and cav ity un reinforced masonry
wall s. Structures, 13: 88–101.
DIANA 2017. Finite Element Analysis Use r’s Manual
10.2 . Del ft: DI ANA FEA BV.
Dolatshahi, K, A ref, A & Wh itta ker, A 2015.
Interaction curves for in-plane and out-of-plane
behaviours of unreinforced masonry wa lls. Journal
of Earthquake Eng ineer ing, 19: 60 –84.
EN (European Sta ndard) 199 0. EN 1990 (2002).
Eurocode: Basis of Structural Design. Bruss els:
Europea n Committee for St andardizat ion (CEN).
EN (European Sta ndard) 1991. EN 1991 (2002+).
Eurocode 1: Action s on Structure s. Brussels: CEN.
EN (European Sta ndard) 1998 . EN 1998-1 (2004).
Eurocode 8: Design of Structures for Earthqu ake
Resist ance. Part 1: General Seismic Acti ons and
Rules for Buildings. Brussels: CEN.
Fourie, J 2017. Characterisation and evaluation of the
mechanical prop erti es of alternative masonry units.
MEng Dissertation. Stellenbosch University.
Giamba nco, G, Rizzo, S & Spallino, R 2001. Numerical
analysis of ma sonry struc tures via interface models.
Compute r Methods in Applied Mechanics and
Engineering, 19 0: 6493 –6511.
Griffith, M 2000. Experimental Study of t he Flexural
Streng th of UR M (Bric k) Walls. Adelaide, Australia:
Universit y of Adelaide.
ICBO (Inter nationa l Conference of Bui lding O ffic ials)
19 97. Uniform Building Cod e (UBC)™ . Whittier, CA:
ICBO.
JSD (Joint St ructural Division) 1995. Code of Practice:
Foundations and Superst ruct ures for Single Storey
Reside ntial Buildings of Masonry Const ruction.
Half way House: SAICE and IStruc tE Joint
Structural Div ision.
Kruger, A, Ret ief, J & Goliger, A 2017. Development
of an update d fundamenta l basic wind speed map
for SANS 10160-3. Journal of the South African
Instit ution of Civil Engine ering , 59(4): 12–25.
Laubscher, J 2014. Reviewing chal lenges between the
need for government-subsid ised hou sing in S outh
Africa and the sustainability requirements of the
National Build ing Regulation s. Proceedings, World
SB14, 28–30 Oc tober 2014, Barcelona, Spain,
pp18–24.
Lourenço, P B 1996. Computational Strategies for
Masonry Structures. Delft, The Netherlands: Delft
University Press.
Lourenço, P B & Rots, J 1997. Multi surface interface
model for analysis of mason ry structu res. Jour nal of
Engineering Mechanics, 123: 660–669.
Macorini, L & Izzuddin, B 2011. A non-linear i nterface
element for 3D mesoscale analysis of brick-masonr y
structures. International Journal for Numerical
Methods in Engineering, 85: 1584–1608.
Mahach i, J, Bradley, R & Goliger, A 2018. Wind
storm damage to houses: Plan ning a nd design
consideration. Proceedings, Out-Of-The-Box
Conference, 24–25 Oc tober 2018, Pretoria,
pp82–92.
Mila ni, G 20 08. 3D upper bound li mit analysis of
multi-lea f masonry wa lls. International Journal of
Mechanical Sciences, 50: 817–836.
Najafgholipour, M, Maheri, M & L ourenço, P B 2 013.
Capacit y interaction in brick mas onry u nder
simultaneous in-plane and out-of-plane loads.
Const ruction and Buil ding Mate rials, 38: 619– 626.
NHBRC (National Home Builders Regist ration
Counci l) 2015. Home Building Manu al and Guide.
Johannesburg: NHBRC.
Reyes, E, C asati, M J & Gàlvez, J C 2008. Cohesive
crack model for mixed mode frac ture of brick
ma sonr y. International Journal of Fracture,
151:29–55.
Roca, P, Cervera, M, Gariup, G & Pela, L 2010. Structural
analysis of ma sonry historical constructions:
Classical and advance d approaches . Archives of
Computer Methods in Engineering, 17:299–325.
RSA (Republic of South Af rica) 1977. National Bui lding
Regulat ions and Building Sta ndards Act, 103 1977,
Government Gazette, No 145 (5640), 6 July 1977,
Pretoria: Govern ment Printer.
RSA (Republic of South Af rica) 1995. Housing
Consumers Protec tion Measures Act, 9 of 1995.
Government Gazette, 401(19418), 2 November 1995,
Pretoria: Govern ment Printer.
SABS 1980. SABS 0164-1:1980. Code of Pra ctice . The
Struc tural Use of Masonry, Part 1: Unreinforc ed
Masonry Walling. Pretoria: S outh Af rican Bureau of
Standard s.
SANS (South A fric an Nation al Sta ndard) 1989.
SANS10164-1 1989. The Str uctural Use of Ma sonry,
Volume 63 Number 1 March 2021 Journal of the South African Institution of Civil Engineering60
Part1: Unreinforced Masonry Walling. Pretoria:
SABS Sta ndards Division.
SANS 2010a. SANS 10400 2010. The Applicat ion of
the National Building Regulations, 3rd ed. Pretoria:
SABS Sta ndards Division.
SANS 2010b. S ANS 104 00 -A 2010. The Application of
the National Building Regulations. Par t A: Gene ral
Principles and Requirements, 3rd ed. Pretoria: SABS
Standard s Division .
SANS 2011. SA NS 10160 2011. Basis of Structural Design
and Actions for Buil dings and Industrial Structures,
1.1 ed. Pretor ia: SABS Standa rds Div ision.
SANS 2011a. SANS 10160-2 2011. Basis of St ruct ural
Desig n and Actions for Buildings and Industrial
Struc tures . Part 2: Self-weight and Imposed Loa ds.
Pretoria: SABS Standard s Division.
SANS 2 011b. SANS 10400-K 2011. The Applicat ion of
the National Building Regul ations. Part K : Walls,
3rd ed. Pretoria: SA BS Standards Division.
SANS 2011c. SA NS 1040 0-L 2011. The Applicat ion of
the National Building Regul ations. Part L : Roofs,
3rded. Pretoria : SABS Standard s Division.
SANS 2011d. SA NS 1040 0-XA 2011. The Application
of the National Building Regulation s. Part X:
Environmental Sustainability; Part XA : Energ y
Usage in Buildings. Pretoria : SABS Sta ndard s
Division.
SANS 2012. SANS 10400-B 2012. The Application
of the National Building Regulation s. Part B:
Structural Design, 3rd ed. Pretor ia: SABS Standa rds
Division.
SANS 2017. SAN S 10160- 4 2017. B asis of Struct ural
Desig n and Actions for Buildings and Industrial
Structures. Part 4: Seismic Act ions and General
Requirements for Buildings. Pretoria : SABS
Standard s Division.
SANS 2018a. S ANS 10160-1 2018. Basi s of Stru ctural
Desig n and Actions for Buildings and Industrial
Struc tures . Part 1: Basis of St ruct ural Design.
Pretoria: SABS Standard s Division.
SANS 2018b. SANS 10160-3 2018. Basis of Structural
Desig n and actions for Buildings and Indust rial
Structures. Part 3: Wind. Pretor ia: SABS Standa rds
Division.
SANS 2018c. S ANS 51996-1-1 2018. Eurocode 6: Des ign
of Masonr y Stru cture s. Part 1-1: General Rules for
Reinforced and Unreinforced Masonry Structure s,
1st ed. Pretoria: SA BS Standards Division .
Sisulu, L 2016. Housing B acklog at 2 .1 mil lion, says
Mini ster Sisulu. Engine ering News, 22 April 2016.
Sparks, P, Liu, H & Saffir, H 1989. Wind damage
to mason ry bui ldings. Journal of Aerospace
Engineering , 2(4): 186–198.
Sýkora, M & Holický, M 2010. Probabilistic model for
masonry strength of exist ing structu res. Engine ering
Mechanics, 174(1): 61–70.
Talocchino, G 2005. Desig n and con struction cr iteri a
for domes in low-cost housing. MEng Disser tation,
Johannesburg: Un iversity of the Witwatersra nd.
Vaculik, J 2012. Unreinforced masonry walls subjected
to out-of-plane seismic actions. PhD Thesis.
Adelaide, Australia: Universit y of Adelaide.
Van Der Pluijm, R 1999. Out-of-pl ane bending of
masonr y: Behaviour of bed joints. PhD Thesis.
Eindhoven, The Net herlands, Techni sche
Universiteit Eindhoven, pp 1849–1862. doi .
org/10.610 0/I R52 8212 .
Van Zijl, G P A G 2000. Computational modelling of
masonr y creep and shrinkage. PhD Thesis. Del ft,
The Netherlands: D elft Un iversit y of Technology.
Watermeyer, R 1996. Recent developments in prov iding
houses of masonr y const ruction in South Africa.
The Structural Engineer, 74(19): 325–331.
Watermeyer, R 200 4. The i mpact of st ructural
engineering on t he sustainabilit y of human
settlements in developing countries. Proceedings, 2nd
Internationa l Conference on Str uctural Engineering,
Mechanics and Computation, Cape Town.
Watermeyer, R & Tromp, B 1992. A systemat ic
approach to the design a nd construction of single-
storey residential masonry str uctu res on problem
soils. The Civ il Engine er in South Afric a, March:
83–96.
Watermeyer, R & Mil ford, R 2003. The u se of
performance-base d building codes t o attain
sustainable hou sing obje ctive: T he South Afric an
approach. Present ed at the Global Polic y Summit on
the Role of Perfor mance-Based Building Regul ations
in Addres sing Societal Expec tations, Internationa l
Policy, and Local Needs. Washington, DC, US.
Wium, J A 2010. Background to d raf t SANS 10160
(2009) Part 4. Sei smic load ing. Journal of the South
Afric an Institution of Civil Engineering, 52(1): 20–27.
