See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/29460053

Evaluation and Experimenal Work for Concrete Walls with Openings in One
and Two-way action

Article · January 2004

Source: OAI

CITATIONS READS

2 273

2 authors, including:

Jeung Hwan Doh

Griffith University
98 PUBLICATIONS   840 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Climate change adaptation of concrete buildings using Building Information Modelling (BIM) View project

Utilize red mud and glass powder to synthesize fly ash-based geopolymer concrete View project

All content following this page was uploaded by Jeung Hwan Doh on 11 August 2014.

The user has requested enhancement of the downloaded file.

technical paper 1

Evaluation of experimental work on concrete walls

in one and two-way action*
JH Doh & S Fragomeni
School of Engineering, Griffith University Gold Coast Campus, PMB 50 Gold Coast Mail Centre,
Queensland 9726, Australia

SUMMARY: Currently, the design of reinforced concrete wall panels under eccentric axial
loading only is carried out using empirical or semi-empirical methods. These methods involve
approximations, which are not always reliable and being empirical, their scope of application is
limited. In particular walls supported on all sides (i.e. two-way action behaviour), made of high
strength concrete strength or high slenderness ratios have been given little treatment in research.
In view of this, the authors have undertaken an extensive investigation on the development of a
more reliable and accurate wall design equation. This paper initially presents a brief overview of
an experimental program undertaken. The experimental results are compared to the proposed
design equation. Also the proposed design equation is compared to results from previous studies
on this topic, confirming that the new equation is accurate and reliable.

NOTATION steel to compressive strength of

concrete
Ac = total area of concrete section M*xx , M*yy = amplified moment in x-axis and y-axis,
Ag = Ltw, gross area of wall panel section
respectably
Asv, Ash = area of vertical and horizontal steel in
Mxi , Myi = internal resisting moment per unit
wall section, respectively
width in x and y directions
π 2 (1 / L + L)2 (h / b)2 respectively
Nx , Ny = in-plane load per unit width in x and
B = 6ε 0 (1 − ρ) where h and
y direction, respectively
b is wall thickness and width of a cross-section, N* = ultimate axial design load.
respectively Nu = axial load capacity of wall or ultimate
c = 0.8352(L/t)-0.0052(L/t)2 axial capacity = 0.85f’cLtw
d = horizontal centre to centre distance P = axial unit stress at failure
between restrained edges of the shear
wall or the horizontal distance of the kE
free edge to the restrained edge of the Pcr =
(L / t ) 2
shear wall
where k = 0.8 for walls braced top and
e = eccentricity of axial force from a
bottom against lateral translation and
centroid axis
restrained against rotation at one or
ex, ey = eccentricity of in-plane load in x and both ends = 1.0 for walls unrestrained
y direction, respectively against rotation both ends
Ec, Esec = modulus of elasticity of concrete, R =
secant modulus at peak stress
f’c = characteristic compressive cylinder A g f c' + A sv f yv {1 + [( A sh f yh ) / A sv ffv )]}
strength of concrete
(L / t w )
fy, fsy, fsy = yield strength of steel
H = unsupported height of a wall tw = thickness of a wall
Hwe = effective height of a wall ε0 = initial strain vector or the concrete
L = length of a wall strain corresponding to the concrete
ultimate strength, f’c
m = fy/f’c, the ratio of yield strength of
φ = strength reduction factor
* Paper S24/921 submitted 3/3/2004 φxo, φyo = maximum curvature in x and y
Paper accepted for publication 15/9/2004 direction, respectively

© Institution of Engineers, Australia 2004 Australian Journal of Structural Engineering, Vol 6, No 1

2 “Evaluation of experimental work on concrete walls…” - Doh & Fragomeni

ν = Poisson’s ratio The main objective of this paper therefore is to

ρm = ratio of steel to gross concrete area investigate the behaviour of axially loaded walls by
evaluating the results of an extensive experimental
1. INTRODUCTION program on the walls with varying support
conditions and concrete strengths. Since only the
Today, reinforced concrete load bearing walls are axial load capacities of concrete wall panels are of
commonly used: a) in core wall systems; b) to resist interest in this study, no horizontal in-plane nor
lateral loads arising from wind or seismic effects; or lateral forces were applied. Using the test results, a
c) as external walls resisting combinations of new and more sophisticated design equation was
horizontal and in-plane vertical forces. developed and also presented. It was compared work
from previous research and found to be accurate and
National concrete codes of practice such as the reliable.
Australian Concrete Standard (AS3600-01) and the
American Concrete Institute Code (ACI318-02) 2. CURRENT CODE METHODS (AS3600-01)
devote separate chapters to wall design, providing
simplified equations to predict load capacity. These Currently, the design of reinforced concrete walls in
simplified wall design equations are only intended Australia is undertaken using the guidelines given
for load bearing walls supported at top and bottom. in the Australian Concrete Standard (AS3600-01). It
The codes fail to recognise any contribution to load gives two methods for the design of concrete walls:
capacity from restraints on the side edges, which is a Section 11 specifies a simplified equation to calculate
common feature of core walls and shear walls. Walls the axial load capacity of walls when certain loading
restrained top and bottom only, with free vertical and bracing restrictions are met. The code also allows
edges, subjected to vertical loads, behave in one-way for any wall to be designed as a column using the
action. This is characterised by uniaxial curvature provisions of Section 10. Although higher capacities
in the direction of loading (see Figure 1(a)). However, can be obtained using this more accurate method, a
in practice, axially loaded walls with side restraints minimum vertical reinforcement content of 1%
can also behave in two-way action, where curvatures placed in two layers is required (only 0.15% is
occur in the directions parallel and perpendicular to required for the simplified equation).
that of loading (see Figure 1(b)).
For the simplified design method, the ultimate design
axial strength per unit length (in N/mm) of a braced
wall in compression is given by the following
formula.
φN u = φ (t w − 1.2e − 2e a )0.6fc'
20 ≤ f′c ≤ 65 MPa (1)

where tw is the wall thickness (mm), e is the load

eccentricity (mm) which has a minimum of 0.05tw, f’c
(MPa) is concrete strength and ea =Hwe2/(2500tw). Hwe
is the effective height of the wall and is the lesser of
kH or kL. Also k is 0.75 for walls restrained against
(a) One-way action (b) Two-way action rotation at both ends and 1 for walls not restrained
against rotation at ends. The strength reduction factor
Figure 1. Walls with or without side’s support φ is 0.6.

Further, the codes are restricted to walls of normal Eq. 1 applies to walls where the slenderness ratio,
strength concretes only. That is, AS3600-01 is Hwe/tw ≤ 30 (if the ultimate design axial force, N* ≤
restricted to concrete strengths of between 20 and 65 0.03f’cAg then Hwe/ tw ≤ 50). The walls are required to
MPa even though designers are using the have minimum reinforcement ratios of 0.0015
recommended methods to analyse axial load vertically, ρv and 0.0025 horizontally, ρh.
capacities of wall panels with concrete strengths in
excess of 65 MPa. Although ACI318-02 does not By appropriate substitution and manipulation using
stipulate a restriction on the maximum concrete the minimum required eccentricity equal to 0.05tw,
strength, the empirical design equations prescribed Eq. (1) can be rewritten as;
in the code are based on tests carried out on normal
  H 2 
strength concrete walls. Thus a better understanding
of wall behaviour, particularly of high strength
φN u = φ 0.564 f t  1 −  we  
'
c w (2)
concrete walls, is desirable so that more sophisticated   34.4t w  
design guidelines can be formulated.
This is very similar to the ACI 318-02 empirical wall

Australian Journal of Structural Engineering Vol 6, No 1

 “Evaluation of experimental work on concrete walls…” - Doh & Fragomeni 3

design formula given by of the panels as indicated in Figure 2. The walls were
loaded in increments up to failure. At each load
  kH  2  increment, crack patterns and deflections were
φN u = φ 0.55f t  1 − 
'
c w   (3) recorded. The latter allowed the load-deflection
  32t w   history to be accurately traced.
where φ taken as 0.7. Here k equals 0.8 for walls
braced top and bottom against lateral translation and
restrained against rotation at one or both ends, and
k equals 1.0 for walls unrestrained against rotation
both ends.

Eq. 3 applies to walls where H/tw ≤ 25 or L/tw ≤ 25

whichever is less for load bearing walls. The
minimum allowable thickness is 100mm. The
resultant load must be in the ‘middle third’ of the
overall thickness of the wall allowing for a maximum
eccentricity allowance of tw/6. The walls are required
to have minimum reinforcement ratios of 0.0015
vertically, ρv and 0.0025 horizontally, ρh. These values
can be reduced to 0.0012 and 0.002 respectively if bars
Figure 2: Test frame arrangement and side
are less than 16mm diameter or if mesh is used.
supports
It should be noted that Eqs. 1 and 3 are intended for
normal concrete strength and only consider
behaviour in one-way action. Hence the current code
practices available to calculate failure loads for
reinforced concrete walls are limited in scope.

3. EXPERIMENTAL PROGRAM

3.1 Test-rig setup

An overview of the test set-up is shown in Figure 2.

The test frame was designed to support three
independent hydraulic jacks each of 80 tonne
capacity. The jacks were required to transmit a Figure 3: Details of top hinged edge
uniformly distributed load across the top through a
loading beam at an eccentricity of tw/6. The top and
bottom hinged support conditions were each 3.2 Test Specimens and Crack Patterns
simulated by placing a 23 mm diameter high strength
steel rod on a 50 mm thick steel plate of 150 mm A total of 18 wall panels were cast and tested. The
width and varying lengths which corresponded to dimensions and material properties of the walls are
the different test panel dimensions. Two 20 mm x summarised in Table 1. The panels designated as
20 mm angle sections were clamped to the thick plate OWNS and OWHS were tested in one-way action
by bolts and the 23 mm diameter rod was welded and panels TWNS and TWHS, in two-way action.
along the length of the plate. Details of the simply Further, NS refers to normal strength specimens and
supported top hinged edge are shown in Figure 3. HS refers to high strength ones. TA refers to the
To achieve the hinged support conditions for two- panels with varying aspect ratio (H/L).
way action, the edges of the panels had to be
effectively stiffened so that rotation about the x-axis A local ready-mix concrete supplied normal and high
was prevented while they were free to rotate about strength concrete for all the specimens. The
the y-axis. To achieve this, two parallel flanged maximum aggregate size was 10 mm. The high
channel sections extending along the height of each strength concrete mix consisted of silica fume and
of the wall panels were placed on both sides of the super-plasticiser in addition to the normal
panel. They were separated by a square hollow constituents. No admixtures were used in the normal
section as shown in Figure 2. strength mix. Three test cylinders for each test panel
were cast to obtain the average compressive strength
Dial gauges were used to record the out-of-plane of concrete. All wall panels were reinforced with a
displacements or deflections. They were placed at single F41 mesh, placed centrally in the panel cross-
quarter and three-quarter heights, and at the centre section. The F41 mesh had a design yield strength of

Australian Journal of Structural Engineering Vol 6, No 1

4 “Evaluation of experimental work on concrete walls…” - Doh & Fragomeni

Table 1: Reinforced concrete wall panel dimensions and concrete strengths

Axial
Failure
f'c strength
MODEL H L tw H/t H/L Load2
(MPa) ratio
(kN)
(Nu/f’cLtw)
OWNS2 1200 1200 40 30 1 35.7 253.1 0.151
OWNS3 1400 1400 40 35 1 52.0 426.7 0.147
OWNS4 1600 1600 40 40 1 51.0 441.5 0.135
OWHS2 1200 1200 40 30 1 78.2 482.7 0.129
OWHS3 1400 1400 40 35 1 63.0 441.5 0.125
OWHS4 1600 1600 40 40 1 75.9 455.8 0.094
TWNS1 1000 1200 40 25 1 45.4 765.2 0.422
TWNS2 1200 1200 40 30 1 37.0 735.8 0.415
TWNS3 1400 1400 40 35 1 51.0 1177.21 0.412
TWNS4 1600 1600 40 40 1 45.8 1177.21 0.402
TWHS1 1000 1200 40 25 1 68.7 1147.8 0.418
TWHS2 1200 1200 40 30 1 64.8 1177.21 0.379
TWHS3 1400 1400 40 35 1 60.1 1250.8 0.372
TWHS4 1600 1600 40 40 1 70.2 1648.1 0.367
TAHS1 1600 1400 40 40 1.14 77.8 1618.7 0.372
TAHS2 1400 1000 40 35 1.40 77.8 1118.3 0.299
TAHS3 1600 1200 40 40 1.33 73.8 1265.5 0.429
TAHS4 1600 1000 40 40 1.60 77.8 1442.1 0.463

1
Notes: Hydraulic jacks measured 40 tonnes for these specimens (40×3×9.81kN).
The accuracy of jack was ± 1 tonne ( = 9.81 kN).
2
Load eccentricity = tw/6

450MPa and the minimum tensile strength was 500 were quite symmetrical, both in the horizontal and
MPa. The reinforcement ratios rv and rh were 0.0031 vertical directions. Also sudden and explosive types
for all panels, satisfying the minimum requirements of failure were observed, and some twisting moment
in the Australian Standard and the ACI code. near the corners of the specimens caused minor
torsional cracks near the corner edges. The TAHS
The crack patterns for five of the panels are indicated panels, which were designed to have varying aspect
in Plates 1 to 10 for illustrative purposes. Plates 1 to ratios (H/L), also had cracks developed in two-way
4 showing that walls in one-way action characterised biaxial curvature with some evidence of half wave
by horizontal cracking at midspan at failure, while action (see Plates 9 and 10).
walls in two-way action feature biaxial cracking as
shown in Plates 5 to 10. 3.3 Test results of one-way panels

A comparison of OWNS2 and OWHS4 crack patterns Table 1 also gives failure loads of all test panels,
illustrates differences in failure modes of one-way giving rise to the following observations. The axial
action panels. The crack pattern of OWNS2 on the strength ratios (Nu/f’cLtw) for both normal (OWNS)
tension face (see Plate 1), was horizontal and high strength (OWHS) concrete panels were
(perpendicular to the loading direction) with failure found to decrease gradually with increasing
occurring near the centre of the panel, signifying slenderness ratios. That is the reduction in axial
bending failure being intensified by buckling. In strength ratios in the normal strength concrete
contrast, high strength panel OWHS4 developed a (OWNS) panels was approximately 10.6% when H/tw
single large crack, commencing at the tension face increased from 30 to 40. Also the reduction in the
splitting in two separate parts (Plate 3). This indicates axial strength ratios in the high strength concrete
that the high strength concrete panels possessed a (OWHS) panels was approximately 27.1% when H/t«
more brittle failure mode, with some yielding of increased from 30 to 40.
reinforcement taking place before concrete failure.
This suggests that the use of very slender and high In comparison, the reduction in axial strength ratios
strength concrete wall panels may become dangerous found in testing by Saheb and Desayi (1989) was
in practice, when only minimum reinforcement is about 41% when H/tw was increased from 9 to 27. It
provided, as abrupt failure may occur. should be noted that Saheb and Desayi’s (1989)
panels were only of moderate slenderness ratios and
All of the two-way action panels developed a biaxial of normal strength concrete. It seems that for higher
curvature crack pattern (see Plates 5 to 8). The cracks slenderness ratios (H/tw > 30), the axial strength

Australian Journal of Structural Engineering Vol 6, No 1

 “Evaluation of experimental work on concrete walls…” - Doh & Fragomeni 5

Plate 1: Wall OWNS2 tension face Plate 2: Wall OWNS2 compression face

Plate 3: Wall OWHS4 tension face Plate 4: Wall OWHS4 compression face

Plate 5: Wall TWNS1 tension face Plate 6: Wall TWNS1 compression face

Australian Journal of Structural Engineering Vol 6, No 1

6 “Evaluation of experimental work on concrete walls…” - Doh & Fragomeni

Plate 7: Wall TWHS2 tension face Plate 8: Wall TWHS2 compression face

Plate 9: Wall TAHS4 tension face Plate 10: Wall TAHS4 compression face

ratios decrease gradually with increasing slenderness may need to be modified for high strength concrete
ratios. panels.

The axial strength ratios of high strength concrete The conservative nature of the code formulae was
panels were smaller than that for the identical normal also evident in Figure 4. For example, when H/tw >
strength concrete panels. This result seems to indicate 30, there was a distinct load capacity recorded from
that the wall failure load does not increase linearly the test results whereas the code formulae gave no
with concrete strength. For example for the OWHS4 strength. It should be noted that the AS3600-01 and
wall, a 48.8% increase in concrete strength (from the the ACI318-02 formulae do not apply to walls with
OWNS4 wall) led to only a 3.3% increase in failure slenderness ratios greater than 30 and 25 respectively
load. This result suggests that prediction formulae as shown in Figure 4.

Australian Journal of Structural Engineering Vol 6, No 1

 “Evaluation of experimental work on concrete walls…” - Doh & Fragomeni 7

0.45
TWNS
0.44
TWHS
0.43

0.42

0.41

0.4

0.39

0.38

0.37

0.36

0.35
25 30 35 40
H/tw
Figure 4: Axial strength ratio versus slenderness Figure 5: Axial strength ratio versus slenderness
ratio for walls in one-way action ratio for walls in two-way action

3.4 Test results of two-way panels A comparison of the failure loads of identical panels
tested in both one- and two-way action, showed that
Table 1 also indicated that the axial strength ratios the failure loads increased, by an overall average of
for both normal (TWNS) and high strength (TWHS) 2.9 times when walls were supported on four sides
concrete panels were found to decrease gradually as compared to only two sides. This implies that the
with increasing slenderness ratios. This is also shown side support caused an increase in wall strength.
in Figure 5. The reduction in the axial strength ratio Interestingly, Timoshenko and Gere’s plate theory
for the normal strength concrete (TWNS) panels was (1961) anticipates the increase in strength is about 4
approximately 4.7% when H/tw was increased from times, when a square isotropic thin plate is submitted
25 to 40. Also the reduction in the axial strength ratios to axial loading with these same support conditions.
for the high strength concrete (TWHS) panels was
approximately 12.2% when H/tw was increased from 4. SIMPLIFIED DESIGN METHOD FOR BOTH
25 to 40. The axial strength ratios of the high strength ONE- AND TWO-WAY ACTION
concrete panels were generally smaller than that of
the normal strength panels (a similar result was The Simplified wall design equation proposed
observed for on-way action panels). Also when the incorporates the behaviour of both one-way action
concrete strengths were increased by an overall and two-way action walls, allowing for higher
average of 49.4%, there was an average 7.0% decrease slenderness ratios, and various concrete strengths.
in the axial strength ratios. Comparatively, It has been derived using the experimental data
Fragomeni et al. (1994) found that for an average presented from previous research, and by
concrete strength increase of 90.4% the average considering relevant parameters prescribed by
decrease in axial strength ratios was 39%. However, AS3600-01 (Eq.1).
his panels had lower slenderness ratios (H/tw = 15
and 12). These findings further emphasise the need The new empirical formula proposed herein for the
for an adjusted formula for more slender and higher axial strength prediction of walls becomes:
strength concrete walls.

In observing varying aspect ratios in Table 1, the axial φN u = φ 2.0 fc07 ( t w = 1.2e − 2e a ) (4)
strength ratios of these panels were found to increase
linearly with an increase in the aspect ratios, except where all variables were defined earlier. Except that
for the TAHS2 panel. Wall TAHS2 had a very low in the calculation of ea, the variable Hwe is determined
failure load (1118.34 kN) and this may be primarily below.
due to a greater eccentricity than tw/6 noted during
testing. The increase in the axial strength ratio was 4.1 Determination of Effective Height (Hwe)
approximately 26.2% for an increase of H/L from 1
to 1.6, when H/tw was kept constant and equal to 40. The results obtained from the current experiments
Similarly it was found for walls TWHS3 and TAHS3 of walls with H/tw between 25 and 40 indicate that
that there was an increase of 15.3% in axial strength the buckling capacity of concrete walls in two-way
ratio for an increase of H/L from 1 to 1.4, when H/tw action is approximately 2.9 times that of one-way
was equal to 35. action walls, when e = tw/6.

Australian Journal of Structural Engineering Vol 6, No 1

8 “Evaluation of experimental work on concrete walls…” - Doh & Fragomeni

As a result, Hwe is modified to compensate for the this slenderness.

increased load capacity of two-way action walls by
the incorporation of an additional eccentricity The proposed Eq. 4 gives a safe estimate of failure
parameter (a) in the effective height factor β that was load for all panels with a mean (predicted/test) of
suggested by Fragomeni and Mendis (1996). That is 0.89 and standard deviation of 0.09. Even though
the effective height equation is given by: Eq. 9 has a mean close to unity, the standard
deviation is too high leading to over-predictions. This
Hwe = βH (5) suggests that Eq. 4 is more reliable due to the lower
standard deviation.
With the following effective length factors defined:
The proposed equation is deemed acceptable because
β=1 for H/tw < 27, simply supported top of the higher strengths obtained in the testing and
and bottom only the application of the capacity reduction factor in
design (φ = 0.6) assuring that all the experimental
18 test results fall well above the predicted design load.
β= 0.88
for H/tw ≥ 27, simply supported top
 H
  Figure 6 shows the design axial strength ratios that
 tw  were obtained with concrete strengths of 30, 50 and
and bottom only 80MPa using the proposed equation (Eq. 4) with
respect to H/tw. Also an eccentricity equal to tw/6
1 was used in all cases. These values were chosen so
β =α 2
for H ≤ L when all four all sides that close comparisons could be undertaken with test
 H results.
1+  
 L
As seen in Figure 6, the proposed Eq. 4 lines agree
are restrained well with the experimental data, with the
L experimental results following the proposed formula
β =α for H > L when all four all sides are trends. The effect of concrete strength on axial
2H strength ratio is evident. The axial strength ratios (at
restrained a fixed slenderness) are higher when the concrete
where α is an eccentricity parameter and is equal to: strength is at a lower value. Previous design methods
did not consider this matter.
1
α= for H/tw < 27
e Figure 6 also shows that when H/tw > 30, the axial
1− strength ratio obtained from the modified formula
tw
is almost linear. This is to compensate for the walls
of high slenderness exhibiting a buckling failure
1 18
α= × 0.88
for H/tw ≥ 27 mode rather than a crushing failure. This result again
1−
e  H illustrates the ineffectiveness of the current simplified
tw   code formulae (Eq. 1 and 3) and other simplified
 tw  equations that give no axial capacity for H/tw > 30.

5. COMPARISON OF PROPOSED EQUATION More comparisons of failure loads for walls tested
WITH TEST RESULTS in one-way action using the proposed formula and
previously recommended equations, against test
5.1 Walls in one-way action results by previous researchers was also undertaken
and given in Table 4. (Details of dimensions and
In this section, the calculated strengths from Eq. 4 failure loads of test panels by previous researchers
and the two other applicable prediction equations are shown in Appendix, Table A.1).
given in Table 2 below were compared with the
experimental test results of the six normal and high The results indicate that the mean ratio of the
strength concrete walls from the current one-way proposed equation (Eq. 4) and test results is 0.84 and
action test panels. Note that the walls tested had a standard deviation of 0.20. The standard deviation
H/tw ratios between 25 and 40. The mean of the is lowest amongst all predicted methods. While there
ratio N u /N* and its standard deviation were are some ratios that are greater than 1 (over-
calculated and presented in Table 3. Eq. 7 (Pillai & estimation), there generally is a good agreement
Parthasarathy, 1977) and Eq. 9 (Zielinski et al., 1982, between the experimental test results and Eq. 4. The
1983) are included in Table 3 for comparison as these results also indicate that AS3600-01 gives a more
equations give strength values for walls H/tw > 30 conservative result (mean 0.7 and standard deviation
whereas the other formulae give no capacities past 0.27) than Eq. 4.

Australian Journal of Structural Engineering Vol 6, No 1

 “Evaluation of experimental work on concrete walls…” - Doh & Fragomeni 9

Table 2: Design axial strength formulae for one-way action

Equation Design axial strength expression
Oberlender/Everard(1973) N u = 0.60f c’A g [1 − ( H / 30 t w ) 2 ]
Eq. 6
Pillai/Parthasarathy(1977) N u = 0.57f c’A g [1 − ( H / 50 t w ) 2 ]
Eq. 7
Kripanarayanan (1977) N u = 0.55f c’A g [1 − ( H / 32 t w ) 2 ]
Eq. 8
Zielinski et al. (1982,83) Pu = 0.55f c’ A g [1 − ( H / 40t w ) 2 ][1 + ρ m ( m − 1)]
Eq. 9
Saheb/Desayi (1989)
  H 2    H 
Eq. 10a N u = 0.55[A f + ( f y − f ) A sv ]1 − 
’
g c
’
c   1.20 −  
  32 t w     10 L 
for H/L < 2.0
Eq. 10b   H   2

N u = 0.55[A g f c’ + (f y − f c’ ) A sv ]1 −   
  32 t w  
for H/L ≥ 2.0
Fragomeni/Mendis (1996)
Eq.11a N u = ( t w − 1.2e − 2e a )0.6f c’ 20 ≤ f’c ≤ 50
Eq. 11b N u = ( t w − 1.2e − 2e a )30[1 + (f c’ − 50) / 80]
50 ≤ f’c ≤ 80

Table 3: Comparison of concrete test panel results with design equations

Wall Eq. 4 Eq. 7 Eq. 9

Panels N*(Test) N*(Test) N*(Test)

OWNS2 0.99 2.47 1.63

OWNS3 0.80 1.98 0.88
OWNS4 0.78 1.52 0.00
OWHS2 0.90 2.84 1.87
OWHS3 0.88 2.32 1.03
OWHS4 1.00 2.19 0.00
Mean 0.89 2.22 0.90
Standard Deviation 0.09 0.45 0.79

5.2 Walls in two-way action calculated, and are presented in Table 6.

In this section, the calculated strengths from Eq. 4 and The proposed Eq. 4 seems to give a good conservative
other selected equations are compared with the test mean at approximately 0.92 and the lowest standard
results of the 12 normal and high strength concrete deviation. This mean value ensures that most of test
wall panels from the current test data. Other two-way results are safely predicted, except for wall TAHS2.
action formulae from previous research are presented The TAHS2 panel failed prematurely a problem that
in Table 5, and the more applicable empirical was discussed before in Section 3.4.
formulae, such as Eqs. 14, 15 and 18, are compared to
Eq. 4. The equations of Ernest (1952), Swartz et al. Even though Eq. 18 has a better mean (0.98 compared
(1974), Aghayere and MacGreor (1990), Attard (1994) to 0.92 of Eq. 4), the standard deviation is slightly
and Raviskanthan (1997) are not considered in this higher than Eq. 4. All results obtained from Eqs. 14
comparison due to the lack of information on the and 15 show over-estimated failure loads, which
material properties as well as the complicated iterative could be due to the fact that those equations do not
procedures that would have been required. The mean consider the effect of higher slenderness ratios.
of the ratio φNu/N* and its standard deviation were

Australian Journal of Structural Engineering Vol 6, No 1

10 “Evaluation of experimental work on concrete walls…” - Doh & Fragomeni

Table 4: Comparison of predicted/test ratios of walls in one-way action

Model Eq. 1 Eq. 3 Eq. 4 Eq. 6 Eq. 7 Eq. 8 Eq. 9 Eq. 10 Eq. 11
N* N* N* N* N* N* N* N* N*
Tests by Pillai &
Parthasarathy

A5 0.54 0.62 0.72 0.67 0.68 0.62 0.66 0.63 0.54

(1977)

A6 0.63 0.72 0.91 0.79 0.76 0.72 0.73 0.72 0.63

B5 0.57 0.65 0.74 0.70 0.72 0.65 0.68 0.65 0.57
B6 0.73 0.84 1.07 0.91 0.88 0.84 0.84 0.84 0.73
WAR-1 0.18 0.22 0.96 0.19 0.44 0.22 0.34 0.21 0.18
Tests by Saheb & Desayi (1989)

WAR-2 1.05 1.21 0.98 1.29 1.37 1.21 1.31 1.24 1.05
WAR-3 1.11 1.28 1.04 1.36 1.45 1.28 1.39 1.31 1.11
WAR-4 1.00 1.15 1.05 1.22 1.30 1.15 1.24 1.18 1.00
WAR-4 0.75 0.86 1.03 0.92 0.98 0.86 0.91 0.86 0.75
WSR-1 0.54 0.61 0.78 0.66 0.67 0.61 0.63 0.61 0.54
WSR-2 0.56 0.65 0.79 0.69 0.73 0.65 0.68 0.65 0.56
WSR-3 0.57 0.66 0.80 0.67 0.87 0.66 0.77 0.66 0.57
WSR-4 0.27 0.33 0.38 0.24 0.84 0.33 0.63 0.33 0.27
Tests by Fragomeni et al. (1994)

2a 0.79 0.92 0.86 0.91 1.31 0.92 1.13 0.80 0.79

4a 0.39 0.46 0.44 0.40 0.92 0.46 0.72 0.40 0.39
6a 0.70 0.81 0.78 0.85 0.98 0.81 0.89 0.73 0.70
8a 0.76 0.87 0.84 0.93 0.99 0.87 0.94 0.89 0.76
2b 1.07 1.25 1.02 1.24 1.78 1.25 1.56 1.10 0.98
4b 0.54 0.64 0.54 0.55 1.28 0.64 1.02 0.56 0.52
6b 1.13 1.30 1.06 1.36 1.57 1.30 1.46 1.19 1.02
8b 0.87 1.00 0.82 1.07 1.14 1.00 1.08 1.02 0.79
Mean 0.70 0.81 0.84 0.84 1.03 0.81 0.93 0.79 0.69

Australian Journal of Structural Engineering Vol 6, No 1

 “Evaluation of experimental work on concrete walls…” - Doh & Fragomeni 11
Figure 7 shows the design axial strength ratios
obtained by Eq. 4 with respect to H/tw. Saheb and (LFEM) and two-way beam-column method
Desayi’s (1990), Fragomeni et al. (1994) and the (WASTABT), both developed and fully described by
current experimental two-way action test results Doh (2003).
were included for comparison. (Details of di-
mensions and failure loads of the test panels are As seen in Figure 7, the proposed equation safely
shown in Appendix, Table A.2). Figure 6 also predicts and follows the general trend of the test
includes results from a Linear Finite Element Method results. Similar to the comparison of walls in one-

Table 5: Design axial strength formulae for two-way action

Equation Design axial strength expression
Ernest (1952) f c’
Eq. 12 P =
1 + cf c’ / Pcr
Swartz et al. (1974)
Eq. 13 f cr = 0.425f c’B[ − B + (4 + B 2 ) 0.5 ]
Saheb & Desayi (1990)
Eq. 14 N u = 0.67φf c’A g {1 − [ L /(120t w )]2 }{1 + 0.12( H / L)}
Eq. 15
N u = φcR
Aghayere/MacGreor
(1990)
 N b 2φ 
Eq. 16 N y e y = (a 2 / b 2 ) M xi − x 2 x 0 − N x e x  +
 π 
 N y a φ y0
2
 a
M yi −  +   M xi M yi
 π 2
  b 
Attard (1994)
0.4E sec
Eq. 17a M *xx = δ b N x 0.00625t w π 2 {1 + ν}
Ec
Eq. 17b
M *yy = δ b N x 0.0025t w π 2 {1 + ν}
Fragomeni/Mendis (1996)
Eq. 18a N u = 0.7f c’ ( t w − 1.2e − 2e a ) 20 ≤ f’c ≤ 50
Eq. 18b N u = 35(t w − 1.2e − 2e a )[1 + (f c’ − 50) / 80] 50 < f’c ≤ 80

Table 6: Comparison of concrete test panel results with design equations

Wall Eq. 4 Eq. 14 Eq. 15 Eq. 18
Panels N*(Test) N*(Test) N*(Test) N*(Test)
TWNS1 0.94 1.70 1.78 1.12
TWNS2 0.96 1.70 1.76 1.05
TWNS3 0.87 1.67 1.67 0.94
TWNS4 0.91 1.66 1.66 0.84
TWHS1 0.83 1.72 1.76 1.02
TWHS2 0.89 1.86 1.87 1.05
TWHS3 0.88 1.85 1.84 0.98
TWHS4 0.86 1.82 1.78 0.82
TAHS1 0.91 1.88 1.83 0.91
TAHS2 1.21 2.43 2.36 1.26
TAHS3 0.88 1.75 1.72 0.97
TAHS4 0.83 1.65 1.59 0.88

Mean 0.92 1.81 1.80 0.98

Standard
0.10 0.21 0.19 0.12
Deviation

Australian Journal of Structural Engineering Vol 6, No 1

12 “Evaluation of experimental work on concrete walls…” - Doh & Fragomeni

way action, axial strength ratios (at a fixed Further comparison studies were carried out with
slenderness) are higher when the concrete strength previous experimental test results in two-way action
is at a lower value. Note that the previous design and presented in Table 7. This comparison included
methods do not consider this matter. Further when the equations proposed by Saheb and Desayi (1990)
H/tw > 30, the axial strength ratio obtained from the and Fragomeni et al. (1994) who tested panels with
proposed formula shows an almost linear trend to slenderness ratios between 5 and 27. Details of
compensate for the fact that walls of higher dimensions and failure loads of test panels are shown
slenderness have a buckling failure mode rather than in Table A.2.
crushing failure.

Table 7: Comparison of failure loads using modified formula in two-way action

Eq. 4 Eq. 14 Eq. 15 Eq. 18

Model N*(Test) N*(Test) N*(Test) N*(Test)

WAR-1(P) 0.90 1.02 1.13 0.75

WAR-2(P) 0.83 0.96 1.06 0.69
Saheb & Desayi (1990)

WAR-3(P) 0.83 0.99 1.07 0.69

Tests by

WAR-4(P) 0.76 0.94 1.00 0.63

WSR-1(P) 0.62 0.76 0.81 0.51
WSR-2(P) 0.67 0.79 0.86 0.55
WSR-3(P) 0.73 0.83 0.92 0.61
WSR-4(P) 0.90 1.03 1.17 0.76
11a 0.81 1.37 1.38 0.84
Fragomeni et
al. (1994)

Tests by

11b 0.83 1.69 1.61 0.93

12a 0.70 1.18 1.22 0.73
12b 0.87 1.78 1.73 0.97
Mean 0.79 1.11 1.16 0.72
Standard deviation 0.09 0.32 0.27 0.13

Australian Journal of Structural Engineering Vol 6, No 1

 “Evaluation of experimental work on concrete walls…” - Doh & Fragomeni 13

The results indicate that the ratios of the proposed

formula to test results had a mean of 0.79 and a Ernest, G. C. (1952), “Stability of Thin-Shelled
standard deviation of 0.09, indicating good Structures” ACI Structural Journal, Vol.49, No.20,
agreement. Eqs. 14 & 15 are less conservative than 1952, pp.227-292.
Eq. 4. Eq. 18 gives a slightly more conservative mean
than Eq. 4, but a higher standard deviation. Fragomeni, S., Mendis, P.A. and Grayson, W.R.
(1994), “Axial load tests on Normal and High
strength concrete Wall Panels”, Australian Civil
6. CONCLUSIONS Engineering Transactions, The Institution of
Engineers, Australia, Vol.36, No.3, Aug. 1994, pp.257-
An experimental program has been undertaken to 263.
investigate the applicability of concrete wall design
equations in current codes of practise. A comparative Fragomeni, S. and Mendis, P.A. (1996), “Improved
study indicates that the design formulas re- axial load formulae for normal and high strength
commended in the Australian Standard (AS3600-01), reinforced concrete walls”, Australian Civil/
the ACI code (ACI318-02) and also other available Structural Engineering Transactions, Institution of
design methods proposed by previous researchers Engineers, Vol.38, No.2, 3 and 4, 1996, pp.71-81.
are inadequate in that their strength predictions
particularly for walls with H/tw ratios greater than Kripanarayanan, K. M. (1977), “Interesting Aspect
30. In such cases the code formulas yield negative of the Empirical Wall Design Equation.” ACI
strength values which indicate zero load-bearing Structural Journal, Vol.74, No.5, May 1977, pp.204-
capacity, which was found to be incorrect in this 207.
experimental evaluation.
Mendis, P.A. (2002), “Design of high strength
concrete members: State of the art”, Engineers
Consequently this paper has provided recom-
Australia, Crows Nest, NSW, Australia, 2002.
mendations to modify the current simplified AS3600
wall design equation. The special feature of this
Oberlender, G.D. and Everard, N.J. (1977),
unified formula is its applicability to both normal
“Investigation of Reinforced Concrete Walls”, ACI
and high strength concrete walls behaving in one-
Journal, Vol.74, Jun. 1977, pp.256-263.
and two-way action. It also caters for the effect of
various slenderness ratios (0 ≤ H/tw ≤ 40) and aspect
Pillai, S.U. and. Parthasarathy, C.V. (1977),“Ultimate
ratios (1.0 ≤ H/L ≤ 1.6). Comparisons with test data
strength and design of concrete walls”, Building and
and previous test results confirmed that the new
Environment, London, Vol.12, 1977, pp.25-29.
formula is accurate and reliable.
Saheb, S.M., and Desayi, P. (1989), “Ultimate Strength
of R.C. Wall Panels in One-way in-plane action”,
7. REFERENCES Journal of Structural Engineering, ASCE, Vol.115,
No.10, 1989, pp.2617-2630.
ACI318-(2002), “Building Code Requirements for
Reinforced concrete”, American Concrete Institute, Saheb, S.M., and Desayi, P. (1990), “Ultimate Strength
Detroit, 2002. of R.C. Wall Panels in Two-way in-plane action”,
Journal of Structural Engineering, ASCE, Vol.116,
AS3600-(2001), “Concrete Structures”, Standards No5, 1990, pp.1384-1402.
Association of Australia, North Sydney, NSW,
Australia, 2001. Swartz, S.E., Rosebraugh, V.H. and Berman, M.Y.
(1974), “Buckling Tests on Rectangular Concrete
Aghayere, A.O. and MacGregor, J.G. (1990), “Tests Panels”, ACI Structural Journal, Vol.71, Jan, 1974, pp.
of Reinforced Concrete Plates under Combined In- 33-39.
Plane and Transverse Loads”, ACI Structural journal,
Vol.87, No.6, Nov-Dec. 1990, pp.615-622. Zielinski, Z.A., Troitski, M.S., and Christodoulou, H.
(1982), “Full-scale Bearing Strength Investigation of
Attard, M.M. (1994), “Buckling of Reinforced Thin Wall-Ribbed Reinforced Concrete Panels”, ACI
Concrete Walls”, School of Civil Engineering, Journal, Vol.79, Jul-Aug. 1982, pp.313-321.
University of New South Wales. UNICIV Report No.
R-339, Sep 1994. Zielinski, Z.A., Troitski, M.S. and El-Chakieh, E.
(1983), “Bearing Capacity Tests on Precast Concrete
Doh, J. (2003), “Experimental and theoretical studies Thin-Wall Ribbed Panels”, Prestressed Concrete
of normal and high strength concrete wall panels Institute Journal, Chicago, Vol.28, No.3, May-Jun.
(with emphasis on two-way action)”, Ph.D Thesis, 1983, pp.89-103.
Griffith University, Gold Coast, Australia, 2003.

Australian Journal of Structural Engineering Vol 6, No 1

14 “Evaluation of experimental work on concrete walls…” - Doh & Fragomeni

APPENDIX

Table A.1: Comparison of failure loads using various experimental test models in
one-way action (e = tw/6)

H××L×
× tw Experimental
f’c
Model H/tw Ultimate load
(mm) (MPa)
(kN)

A5 800×700×80 20.8 10 932.0

Tests by Pillai &
Parthasarathy

A6 400×700×80 15.2 5 647.5

(1977)

B5 800×700×80 22.7 10 971.2

B6 400×700×80 15.6 5 559.2
C2 1200×500×48 20.6 25 343.4
WAR-1 600×900×50 17.9 12 484.3
Tests by Saheb & Desayi (1989)

WAR-2 600×600×50 17.9 12 314.8

WAR-3 600×400×50 17.9 12 198.3
WAR-4 600×300×50 17.9 12 147.4
WAR-4 600×300×50 17.3 12 147.4
WSR-1 450×300×50 17.9 9 214.2
WSR-2 600×400×50 17.3 12 254.1
WSR-3 900×600×50 17.3 18 298.0
WSR-4 1350×900×50 17.3 27 373.7
2a 1000×300×50 42.4 20 231.8
4a
Tests by Fragomeni et .al

1000×300×40 35.7 25 198.6

6a 600×200×40 38.3 15 162.6
8a 420×210×35 39.6 12 158.2
(1994)

2b 1000×300×50 65.4 20 263.5

4b 1000×300×40 54.0 25 216.8
6b 600×200×40 67.4 15 178.0
8b 420×210×35 67.4 12 233.0

Australian Journal of Structural Engineering Vol 6, No 1

 “Evaluation of experimental work on concrete walls…” - Doh & Fragomeni 15

Table A.2: Comparison of failure loads using various experimental tests models in
two-way action

Experimental
H × L × tw f’c Ecc
Model H/tw Ultimate Load
(mm) (MPa) (e)
(kN)

600×900×50
Tests by Saheb & Desayi (1990)

WAR-1(P) 17.9 tw/6 12 555.9

WAR-2(P) 600×600×50 17.9 tw/6 12 416.5
WAR-3(P) 600×400×50 17.9 tw/6 12 284.9
WAR-4(P) 600×300×50 17.9 tw/6 12 235.2
WSR-1(P) 450×300×50 17.3 tw/6 9 283.9
WSR-2(P) 600×400×50 17.3 tw/6 12 346.7
WSR-3(P) 900×600×50 17.3 tw/6 18 463.3
WSR-4(P) 1350×900×50 17.3 tw/6 27 534.0
11a 600×200×40 37.3 tw/6 15 198.2
Fragomeni et
al. (1994)

Tests by

11b 600×200×40 68.7 tw/6 15 296.2

12a 420×210×35 37.3 tw/6 12 208.8
12b 420×210×35 73.3 tw/6 12 271.2

JEUNG-HWAN DOH

Jeung-Hwan Doh is a Associate Lecturer in the School of Engineering at Griffith

University, Gold Coast Campus, Australia. He obtained Bachelor of Honours and
Master of Honours in his civil engineering degree from University of Wollongong
and obtained Ph.D. from Griffith University. His current research focuses on the
reinforced concrete walls with and without openings using normal and high strength
concrete.

SAM FRAGOMENI

Sam Fragomeni is a Senior Lecturer in the School of Engineering at Griffith University

Gold Coast, Australia. He obtained his PhD from The University of Melbourne,
Australia in the area of concrete structures. His research interests include behaviour
of concrete wall panels, high performance concrete and high-rise buildings. He is
the author of over 60 journal/conference papers in these areas.

Australian Journal of Structural Engineering Vol 6, No 1

View publication stats