Professional Documents
Culture Documents
The Application and Interpretation of Linear Finit
The Application and Interpretation of Linear Finit
Journal of the South African Institution of Civil Engineering • Volume 56 Number 1 April 2014 77
Table 1 A
pportionment between column and middle strip in footings expressed as a percentage Where shear effects become important (i.e.
of the total negative (hogging) moment deep beams where span/effective depth < 2)
the member can be modelled using equiva-
Column strip
Code Column strip Edge strip lent truss analogy.
width
FE analyses can be either linear or
bcol + 3d
TMH 7 1989, Code of Practice for the Design
if the width of the
non-linear. Linear FE analysis is the most
of Highway Bridges and Culverts in South 66.67% 33.33%
footing is greater commonly used type, but is limited in its
Africa, Part 3 (discussed under footings)
than 1.5(bcol + 3d) capabilities as it does not take cracking
and softening of the concrete into account
2 éD – 1é
SANS 10100 2000, The structural é é × 100 ê ê (Rombach 2004). This type of analysis is
B êD ê êB ê × 100
use of concrete, Part 1 ê + 1ê êD ê suitable for an ultimate limit state design
ëB ë ê + 1ê
ëB ë
check, but cannot be used to check service-
bcol + 3d ability deflection and cracking. Non-linear
Eurocode 2, Design of concrete if the width of the
66.67% 33.33%
FE analyses model the cracked behaviour of
structures EN 1992-1-1:2003 (E) footing is greater the concrete by means of an iterative process,
than 1.5(bcol + 3d)
but are complicated and time consuming to
bcol + 3d set up, and the software cost is significantly
BS 8110:1997 Structural use if the width of the more than a linear FE program. In practice,
66.67% 33.33%
of concrete, Part 1 footing is greater
than 1.5(bcol + 3d) flat plate type structures are generally
designed using a linear FE analysis, and
bcol = column dimension in the long direction; d = depth of the slab; serviceability compliance done with ‘rule of
D = longer plan dimension of footing; B = shorter plan dimension of footing
thumb’ span to effective depth ratio checks.
The main criticisms of linear FE analyses are
its use of elastic material properties, which
result in overestimated support moments
Edge strip Column strip Edge strip and underestimated deflections (Jones &
Morrison 2005), and an impractical required
reinforcement contour output. Figure 1
shows the typical transverse bending
Bending moment
78 Journal of the South African Institution of Civil Engineering • Volume 56 Number 1 April 2014
Y
Mxy Mx
Mxy
My
My Mx
Mxy Mxy
Journal of the South African Institution of Civil Engineering • Volume 56 Number 1 April 2014 79
The analysis of a reinforced concrete
D
1
pad footing is a multi-parameter problem.
The stiffness of the flat slab is significantly
influenced by the non-linear properties of
concrete (i.e. cracking, which in turn influ-
ences member forces and deflections), and
Critical furthermore there is the added complexity of
section soil-structure interaction. The deformation
1 200
characteristics of the soil can play a signifi-
cant role in the distribution of the pressure
and hence the load effect. For this exercise a
conservative uniform bearing pressure under
the footing was assumed, with the column
acting as a support.
The moments obtained from the FE analy-
Critical sis are also very sensitive to how the supports
section are modelled in the FEM, and to the slab
geometry. The effect of the support model
was considered by analysing a square footing
and changing the way the supports were
modelled; and the effects of geometry was
considered by analysing a combined loading
Figure 4 Critical section for the determination of the flexural design moment rectangular footing and varying the thickness.
The conventional flat slab/footing design
may be smoothed or averaged in a two-way conservative design approach is to have two method described in the South African
spanning slab. slab models, one where columns are assumed bridge code TMH 7 (1989) was used for the
The typical output from a linear FE to be pinned supports to determine the worst simplified method of design. The critical
analysis of plate elements gives bending case sagging moment, and the second where section for the design of the flexural hogging
moments in the x and y directions Mx and the column supports are fixed to determine reinforcement in the x and y directions is
My and a local twisting moment Mxy (see the worst-case hogging moments. Eurocode taken at the face of the column (Figure 4).
Figure 3). This twisting moment takes the 2 does not prescribe a specific analysis or If the width of the footing is greater than
three-dimensional behaviour of a flat slab dictate how to interpret FE method load 1.5(bcol + 3d) the code requires the slab to
into account. However, it does not act in the effects, which are open to a wide range of be split into column and middle strips, and
direction of the reinforcement. A popular interpretations depending on how the col- designed and detailed accordingly, where
method of including the twisting moment is umn supports are modelled. Most commonly bcol is the column width and d the effective
known as Wood-Armer moments, and most used FE packages give no clear directive on depth to the tension reinforcement of the
design software will automatically calculate how to detail the reinforcement for flat slabs slab. The width of the column strip (w) is
the Wood-Armer moments for the user. The designed using FE. In general it is accepted thus governed by the width of the column
Wood-Armer moments were developed to that the design engineer will use the required and the depth of the slab using the equa-
take complex loading into account, where the reinforcement contour plots to decide how to tion w = bcol + 3d. The column strip is then
twisting moment Mxy needs to be considered place the slab reinforcement. It is, however, designed to resist two thirds of the total
(Denton & Burgoyne 1996). There are four obvious that if the FE reinforcement con- bending moment, and the middle strip is
components – top (hogging) moments in the tours are followed exactly this would lead to designed to resist one third of the total bend-
x and y directions MxT and MyT , and bottom a very impractical reinforcement layout. ing moment.
(sagging) moments in each direction MxB From the above it is clear that when
and MyB . This method is conservative and designing and detailing, using FE analy- Finite element method: the effect
these moments form an upper limit envelope sis, a great deal is left up to engineering of how the supports are modelled
of the worst-case design moments. The four judgement. To get an indication of how the support
components can be used to calculate the model effects FE moments, a square footing
required reinforcement for each of the rein- was analysed with the different support con-
forcement layers in a flat slab type structure. FLEXURAL CAPACITY OF ditions described by Rombach (2004), and
(Brooker 2006) FOOTING CALCULATED BY the following noted for each, as summarised
Modern codes allow for nonlinear analy- THE SIMPLIFIED METHOD AND in Table 2 and shown Figure 5:
sis of reinforced concrete structures, but in FINITE ELEMENT METHODS ■■ peak My axis moment (FE Mpeak)
practice such a complex analysis is seldom In order to assess load effects in a linear ■■ peak Wood and Armer moment (FE
justified due to the large amount of work elastic FE model, the design of a simple Wood and Armer Mpeak)
required and the cost of suitable software. foundation pad footing was undertaken using ■■ total My axis moment – the sum under
Designs are usually based on linear-elastic simplified design methods (SD), and then the moment curve at the face of the col-
material behaviour, assuming that the ductile compared with the results of a linear elastic umn (FE Mtotal)
properties of reinforced concrete allow for FE model of the same footing. The moment ■■ total Wood and Armer moment – the
a limited redistribution of forces. Rombach variation at the critical design section (i.e. at sum under the moment curve at the face
(2004) states that the accuracy of such a the face of the column) for both methods of of the column (FE Wood and Armer
simplified approach is generally sufficient. A analysis was compared. Mtotal)
80 Journal of the South African Institution of Civil Engineering • Volume 56 Number 1 April 2014
Table 2 S ummary of footing analysis parameters
Y
Item Description
Pad
Flat plate type
foundation
40
(c) Rigid supports – edges in South Africa (see Figure 5). It consisted of
30 (d) Rigid supports – corners square 0.025 m x 0.025 m plate elements. The
elements used to analyse the pad foundations
(e) Encased – fixed at column edges
20 are discreet Kirchoff-Mindlin quadrilaterals
(f) Springs
which provide good results for both thick and
(g) Springs – edges thin plates and are free from shear locking.
10
(h) Rigid – rigid links Shear deformations are not considered here,
0 (i) Point support as the cantilever span to depth ratio was rela-
0 0.2 0.4 0.6 0.8 1.0 1.2 tively low. The following is a summary of the
Footing width (m) results shown in Figure 6 and Table 3:
■■ For each different support model the total
Figure 6 Bending moment comparison at the face of the column using different support conditions FE M x or My moments were the same as
the total SD moment.
■■ The total FE Wood and Armer moment
Journal of the South African Institution of Civil Engineering • Volume 56 Number 1 April 2014 81
Table 3 Effect of support conditions on FE peak and total moment
FE
FE W&A SD Mcol FE FE W&A FE W&A FE
FE Mpeak SD Mcol Mpeak / SD Mtotal FE Mtotal
Support type Mpeak inner Mpeak / Mtotal Mtotal / Mtotal /
(kNm/m) (kNm/m) SD Mcol (kNm) (kNm)
(kNm/m) (kNm/m) SD Mcol (kNm) SD Mtotal SD Mtotal
inner
a)
Full 3D
38.86 39.24 33.1 44.12 1.17 0.88 29.68 33.03 29.78 1.11 1.003
continuum
model
b)
Rigid
supports 43.87 47.74 33.1 44.12 1.325 0.99 29.68 32.44 29.78 1.09 1.003
(column
area)
c)
Rigid
50.8 52.6 33.1 44.12 1.53 1.15 29.68 32.03 29.78 1.08 1.003
supports
(edges)
d)
Rigid
57.6 59.4 33.1 44.12 1.74 1.31 29.68 31.98 29.78 1.08 1.003
supports
(corners)
e)
Encased
supports
52.85 73.8 33.1 44.12 1.6 1.20 29.68 33.12 28.47 1.12 0.96
(fixed at
column
edges)
f)
Spring 37.05 42.08 33.1 44.12 1.12 0.84 29.68 33.10 29.78 1.12 1.003
supports
g)
Spring
39.74 39.75 33.1 44.12 1.2 0.90 29.68 32.79 29.78 1.10 1.003
supports at
edges
h)
Rigid
column 52.85 73.84 33.1 44.12 1.6 1.20 29.68 33.12 28.47 1.12 0.96
head
(rigid links)
i)
Point 33.48 42.74 33.1 44.12 1.01 0.76 29.68 35.72 29.68 1.20 1.003
support
82 Journal of the South African Institution of Civil Engineering • Volume 56 Number 1 April 2014
modelling the columns support as a 3D
Y X
continuum and modelling the edge of the
–0.1
0.0
column with springs. The peak moment
5.0 is within 5% of the inner column strip-
10.0
15.0 stepped SD method moment.
20.0 ■■ For both spring models the sensitivity
25.0
30.0 of the model to support models was
35.0 checked. In general a more realistic
40.0
45.0 moment distribution was obtained as the
50.0 stiffness of the spring decreased.
51.1
■■ Ignoring the stiffness of the column, and
modelling the support as pinned over the
footprint of the column, show reverse cur-
vature in the column area (see Figures 3
and 6). This reduction in moment over the
column could be attributed to the fact that
fixing the translational degrees of freedom
Z on the column footprint prevents the
movement at the nodes, while curvature
Figure 8 Principal axis Mx moment contours (kNm/m) within the element still occurs and there-
fore a reduced moment over the footprint
of the column is observed. If the rotational
Y X
degrees of freedom are also fixed in the
0.0
5.0 column footprint, the moment over the
10.0 column reduces to almost zero, which in
15.0
20.0 reality is impossible.
25.0
30.0
35.0
Adequacy of the simplified method
40.0 The pad foundation requires bottom rein-
45.0
forcement in the transverse (x) and longitu-
50.0
51.6 dinal (y) directions to resist the My and M x
hogging moments respectively.
The Simplified Design (SD) method
of analysis results in a constant moment,
which is then split into a column strip
moment and an edge strip moment for the
pad foundation. The column strip can be
stepped again by concentrating two thirds
Z of the column strip moment into half of
the column strip width to form an inner
Figure 9 Wood and Armer Mx moment contours (kNm/m) and outer column strip, according to the
detailing rules of SANS 10100 Cl 4.6.5.4 (see
45 Figure 7).
Considering the results of the FE analyses
40 with different support constraints, the
supports modelled with full 3D continuum
35
are the most realistic and will be used for
30 comparison with the SD analysis. The linear
SD Mdesign FE moment outputs are the M x and My
25 moments and the commonly used Wood-
Moment (kNm/m)
Concentrated SD
FE Mdesign Armer moments, which include the twisting
20
moment M xy. Figure 8 shows the FE M x
FE W&A moments
15 moment contours for the hogging moments
in the pad foundation. Figure 9 shows the FE
10 Wood and Armer moment contours for the
M xT and MyT hogging moments in the pad
5
foundation.
0
A section taken through the SD and FE
0 0.2 0.4 0.6 0.8 1.0 1.2 bending moment diagrams at the face of the
Footing width (m) column in the x and y directions is shown
in Figure 10. Both the peak FE My and M x
Figure 10 S implified design method moment compared to FE design moments hogging moments occur at the face of the
(W&A in legend = Wood and Armer) column as there is a cantilever on both sides.
Journal of the South African Institution of Civil Engineering • Volume 56 Number 1 April 2014 83
Table 4 Summary of footing analysis parameters
9 000
2 250 4 500 2 250 Item Description
Pad
Flat plate type
foundation
Full 3D
Support in FE analysis
continuum
2 250
2 500
Varies from
Thickness, h 0.4 m to
5 000
500
1.3 m
2 250
Concrete Young’s Modulus, Ec 28 GPa
Mx My
d
X
Z
Figure 12 T hree-dimensional footing model
showing moment sign convention
84 Journal of the South African Institution of Civil Engineering • Volume 56 Number 1 April 2014
in the longitudinal (y) direction to resist the
–1 550.1 Y X
–1 400.0 M x sagging moment.
–1 200.0 The linear FE moment outputs used were
–1 000.0 the M x and My moments and the Wood and
–800.0 Armer moments. Figures 14 and 15 show
–600.0 the moment contours in the x and y direc-
–400.0
tions for a pad foundation depth of 700 mm.
–200.0
0.0 Figures 16 and 17 show the FE Wood and
5.1 Armer moment contours for the M xT and
MyT hogging moments in the pad foundation
with a depth of 700 mm. The effect of the
twisting M xy moment at the constraints (col-
Z umn) can be seen in the Wood and Armer
moment contours.
Figure 14 Principal axis My moment contours (kNm/m) A section taken through the SD and FE
bending moment diagrams at the face of
the column in the x and y directions for pad
Y X
foundation depth of 700 mm is shown in
Figure 18. Both the FE M x and My moments
0.0
and the FE Wood and Armer moment are
200.0
shown in the graph.
400.0
600.0 The peak FE My hogging moment occurs
800.0 on the cantilever side of the column, whereas
1 000.0 the peak FE M x hogging moment is mirrored
1 200.0 about the pad foundation centreline, as there
1 400.0 is a cantilever on both sides on the column.
1 550.7 The Wood and Armer design moments are
greater than the SD design moments, as
Z these moments include the twisting M xy
moment. Because of the unique solution and
Figure 15 Hogging My Wood and Armer moment contours (kNm/m) – bottom rebar optimisation requirement, the Wood and
Armer moment is always greater than the
applied moment (Denton & Burgoyne 1996).
–1 208.8 Y X
–1 200.0 The peak FE moment Mpeak (maximum
–1 100.0 FE hogging moment) was affected by the
–1 000.0 curvature of the pad foundation. As the
–900.0
–800.0 stiffness of the footing decreased the peak
–700.0 FE moment increased. The SD method
–600.0 of analysis results in a constant moment
–500.0
–400.0
which is split into a column strip and edge
–300.0 strip moment. The column strip require-
–200.0 ment ensures that as the FE peak moment
–100.0
increases with a footing depth decrease, the
0.0
100.0 SD column strip reduces, resulting in an
176.8 Z increased SD moment, thus ensuring that
the increase in curvature is provided for.
Figure 16 Principal axis Mx moment contours (kNm/m) The integration of the area under the
moment diagram gives the total SD and FE
load effect. The total SD design moment
0.0 Y X
does not vary with the change in pad foun-
100.0
200.0 dation depth, as the self-weight of the pad
300.0 foundation does not have an effect on the
400.0 applied load effect. The total FE M x and My
500.0 moments are also constant with respect to
600.0 change in footing depth, and are the same
700.0
as the SD design moments. The FE Wood
800.0
and Armer moments are affected by the
900.0
1 000.0 twisting moment, which in turn is affected
1 100.0 by how the constraints are modelled, and
1 200.0 slab geometry. An increase in slab stiff-
1 209.3 Z ness leads to an increase in the twisting
M xy moment. A comparison of the total
Figure 17 Hogging Mx Wood and Armer moment contours (kNm/m) – bottom rebar FE M x and My moments (equal to total SD
Journal of the South African Institution of Civil Engineering • Volume 56 Number 1 April 2014 85
1 800 1 800
1 600 1 600
1 400 1 400
Moment Mx (kNm/m)
Moment My (kNm/m)
1 200 1 200
1 000 1 000
800 800
600 600
400 400
200 200
0 0
0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 0 1 2 4 4 5 6 7 8 9
Position along footing width (m) Position along footing length (m)
Column edge FE W&A moment Column edge FE W&A moment
SD moment FE moment SD moment FE moment
Mx My
Figure 18 Simplified design method moment compared to FE design moment (d = 700 mm)
9 000 3 500
8 500 3 250
8 000 3 000
Moment (kNm/m)
Moment (kNm/m)
7 500 2 750
7 000 2 500
Column strip
Column strip
6 500 2 250
6 000 2 000
400 500 600 700 800 900 1 000 1 100 1 200 1 300 400 500 600 700 800 900 1 000 1 100 1 200 1 300
Footing depth (mm) Footing depth (mm)
Total FE W&A moment Total FE principal axis moment Total FE W&A moment Total FE principal axis moment
Figure 19 T otal FE Wood and Armer (W&A) moment compared to Total Figure 20 T otal FE Wood and Armer (W&A) moment compared to Total
FE principal axis moment (My) FE principal axis moment (Mx)
moment) to the total FE Wood and Armer Peak load effects (singularities) The M x FE Mpeak moment remained
moment for My and M x is shown in Figures in linear FEM constant as the footing depth increased.
19 and 20. Figure 21 shows the change in the SD The My FE Mpeak moment decreased with
The total FE Wood and Armer moment moment, column strip and middle strip, the increase in footing depth, until a footing
My (cantilever) was approximately 5.6% compared to the FE peak moment (Mpeak), depth of 1 100 mm, and then levelled out.
greater than the total SD moment at a depth as the depth of the pad foundation (h) Figure 22 shows a comparison of the
of 400 mm, increasing to approximately 6.9% varies. Both the M x and My moments and ratio of the FE Mpeak to the SD column strip
at a depth of 1 300 mm, i.e. a 23% increase. Wood and Armer moments were plotted, moment and the SD concentrated column
The principal axis moment was the same as and the SD requirement of differentiating strip moment.
the SD moment. between the column and middle strip Both the M x and My FE Mpeak / SD
The total FE Wood and Armer moment is shown. For a pad foundation width of Mcolumn ratios approach one as the stiffness
M x (beam) was approximately 10.5% greater greater than 1.5 x (bcol + 3h), the transverse of the slab decreases, and levels out to a
than the total SD moment at a depth of distribution of curvature has reduced constant value at a depth consistent with the
400 mm, increasing to approximately 16.5% sufficiently so as to not warrant the limit for the column strip of the code. Again,
at a depth of 1 300 mm. The principal axis differentiation between the column and showing that the SD column strip require-
moments were the same as the SD moment. middle strip. ment ensures that as the FE peak moment
86 Journal of the South African Institution of Civil Engineering • Volume 56 Number 1 April 2014
1 600 1 600
1 400 1 400
1 200 1 200
Moment Mx (kNm/m)
Moment My (kNm/m)
1 000 1 000
800 800
600 600
400 400
Column strip
Column strip
200 200
0 0
400 500 600 700 800 900 1 000 1 100 1 200 1 300 400 500 600 700 800 900 1 000 1 100 1 200 1 300
Footing depth (mm) Footing depth (mm)
SD My column FE W&A Mpeak SD Mx column FE W&A Mpeak
SD My mid FE Mpeak SD Mx mid FE Mpeak
Mx My
2.2 2.2
2.0 2.0
1.8 1.8
Mx FE peak / SD column strip
1.6 1.6
1.4 1.4
1.2 1.2
1.0 1.0
0.8 0.8
0.6 0.6
Column strip
Column strip
0.4 0.4
0.2 0.2
0 0
400 500 600 700 800 900 1 000 1 100 1 200 1 300 400 500 600 700 800 900 1 000 1 100 1 200 1 300
Footing depth (mm) Footing depth (mm)
FE / SD FE / Concentrated SD FE / SD FE / Concentrated SD
FE W&A / SD FE W&A / Concentrated SD FE W&A / SD FE W&A / Concentrated SD
Mx My
increases with a footing depth decrease, the provided adequate and similar flexural ■■ The peak and total Wood and Armer
SD column strip reduces, resulting in an capacity. moments obtained from a linear FE
increased SD moment, thus ensuring that ■■ The FE peak Mx or My moment can analysis is affected by the change in plate
the increase in curvature is provided for. exceed the column strip moment by a thickness – this is because of the change
significant amount, depending on how in the twisting moment.
Observations from analysis the support constraints are modelled. It is, ■■ By concentrating two thirds of the col-
From the above numerical analyses the fol- however, commonly assumed that this peak umn strip reinforcement into an inner
lowing can be concluded about the simplified is reduced by cracking of the concrete and column strip the SD design moment
method overall strength, finite element peak yielding of the reinforcement. approached the FE peak moment.
values and detailing according to linear FE ■■ The peak and total Wood and Armer
methods: moments obtained from a linear FE analy-
■■ The total FE M x and My moments are the sis are significantly influenced by how the EXPERIMENTAL INVESTIGATION
same as the total SD moment. supports are modelled. This is because of Preliminary experiments carried out at the
■■ Both the simplified method and finite ele- the change in the twisting moment with Department of Civil Engineering at the
ment analysis and reinforcement layouts the support/constraint model. University of Pretoria support the above
Journal of the South African Institution of Civil Engineering • Volume 56 Number 1 April 2014 87
1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 9
1 200
A A A A
150 150 120 120 120 120 120 150 150 100 200 150 80 70 70 80 150 200 100
250
30 120
150
150
475 250 475 475 250 475
1 200 1 200
Section A-A Section A-A
Springs
Support (k = 2 500
400
kN/m)
1.2 m x
Plan dimensions
400
1.2 m
Thickness, h 0.15 m
1 200
1 200
88 Journal of the South African Institution of Civil Engineering • Volume 56 Number 1 April 2014
450 450
400 400
350 350
Applied load (kN)
250 250
200 200
150 150
Yield strain
Yield strain
100 100
50 50
0 0
0 1 000 2 000 3 000 4 000 5 000 6 000 0 1 000 2 000 3 000 4 000 5 000 6 000
Reinforcement strain (μm/m) Reinforcement strain (μm/m)
1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 9
(a) SD (b) FE
Figure 25 Strain in reinforcement with applied load for (a) SD and (b) FE footings
250 250
Column
Column
200 200
Reinforcement strain (μm/m)
150 150
100 100
50 50
0 0
0 200 400 600 800 1 000 1 200 0 200 400 600 800 1 000 1 200
Position along footing width (mm) Position along footing width (mm)
50 kN 75 kN 100 kN 125 kN 150 kN 175 kN 50 kN 75 kN 100 kN 125 kN 150 kN 175 kN
(a) SD (b) FE
Figure 26 Strain in reinforcement prior to cracking for (a) SD and (b) FE along section AA
numerical analysis. Hossell (2012) undertook the footing. The strain in the reinforcement reinforcement strain in both footings was
tests on reinforced concrete footings sup- across the footing was logged at a rate of well below the yield strain. The yield strain
ported on springs where two specimens were 1 Hz. The change or variation in strain is was calculated using the 0.2% proof stress
designed and reinforced, one according to the shown in Figures 25, 26 and 27; the footing method described in TMH 7 (1989).
SD method and the other according to the displacement is shown in Figure 28; and a
linear FE method. The influence of the rein- summary of the footings response to the Transverse variation in reinforcement
forcement layout on the response of the foot- load is included in Table 6. It should be strain prior to cracking
ing to ultimate limit state and serviceability noted that, as a result of using the Wood and Figure 26 shows that prior to the concrete
limit state characteristics was observed. Armer moments to calculate the FE rein- cracking the reinforcement at the face of the
The spring-supported footing test setup is forcement, the FE footing required slightly footings (i.e. design section) is strained the
shown in Photo 1, with the springs simulat- more reinforcement than the SM footing. most. The SD footing had a greater variation
ing the founding support conditions. The in strain between the reinforcement under
test parameters are shown in Table 5 and the Strain in reinforcement with load the column and the reinforcement at the edge,
reinforcement and spring support layouts Figure 25 shows that the first crack (sudden than the reinforcement in the FE footing. The
are shown in Figures 23 and 24. Footing (a) “jump” in strain) in both footings occurred reinforcement strain in the FE footing appears
was reinforced according to an SD analysis, at very similar loads, as this is primarily to be more uniform across the footing width
and Footing (b) according to an FE analysis. dependent on the tensile strength of the than when compared with the SM footing.
Strain gauges were placed on the flexural concrete. The SM footing test had to be
reinforcement bars at the critical design stopped at a load of 412 kN, before failure, Transverse variation in
section along the face of the column, and as the testing machine piston moved out of reinforcement strain after cracking
LVDTs at the centre of each support spring alignment. The FE footing failed in punching Flexural cracking occurred at an applied
were used to measure the displacement of at 480 kN. At the design load of 318 kN the load of approximately 205 kN in both
Journal of the South African Institution of Civil Engineering • Volume 56 Number 1 April 2014 89
6 000 6 000
Column
Column
5 000 5 000
Reinforcement strain (μm/m)
4 000 4 000
3 000 3 000
2 000 2 000
1 000 1 000
0 0
0 200 400 600 800 1 000 1 200 0 200 400 600 800 1 000 1 200
Position along footing width (mm) Position along footing width (mm)
200 kN 225 kN 250 kN 275 kN 300 kN 200 kN 225 kN 250 kN 275 kN 300 kN
325 kN 350 kN 375 kN 390 kN 325 kN 350 kN 375 kN 400 kN 408 kN
(a) SD (b) FE
Figure 27 Strain variation after cracking for (a) SD and (b) FE along section AA
reinforcement.
250
Once the concrete cracked, a greater
Cracking load
variation in strain was observed in the SD 200
footing, compared to the FE footing, shown
in Figure 27. With a larger variation in strain 150
the reinforcing bars beneath the column in
the SD footing strained more than the bars 100
towards the edge; indicating that fewer, but
50
larger, cracks developed when compared
to the FE footing. The FE footing showed a
0
more uniform variation in strain, indicating 0 5 10 15 20 25 30 35 40
that more cracks had formed, but because Deflection (mm)
of the lower strain levels these cracks were SD FE
smaller.
Figure 28 Deflection at centre of footing for SD and FE
Load-deflection curves
Figure 28 shows the load deflection curves at Table 6 Summary of footing test results
the centre of the two footings. Cracking and
Flat slab analysis Simplified design Finite element
flexural failure can be seen by the change in
gradient of the curves. First crack occurred Load at first crack 205 kN 205 kN
at very similar loads and deflections for both Deflection at 200 kN (centre of pad foundation) 19 mm 18.5 mm
the FE and SD footing.
412 kN (piston moved
Load at failure 480 kN (punched)
out of position)
Observations from experimental work
Deflection at 390 kN (centre of pad foundation) 35.3 mm 33.5 mm
From the above experimental work the fol-
lowing can be concluded regarding SD and
FE analysis and design: of strains across the width of the footing CONCLUSIONS
■■ Both the simplified method and finite before and after cracking occurs.
element designs provided adequate and ■■ Cracking would appear to be controlled Total resistance achieved with
similar flexural capacity. by reinforcing to follow the finite element FE design compared to that
■■ Detailing reinforcement in accordance peak moment. of traditional methods
with the variation in moments produced ■■ There is no apparent benefit in control- Both the numerical analysis and experi-
from a linear finite element analysis ling deflection by reinforcing to follow mental work support the conclusion that at
results in a more uniform distribution the FE peak moment. the ultimate limit state there is very little
90 Journal of the South African Institution of Civil Engineering • Volume 56 Number 1 April 2014
Outer Outer
column column
Edge strip strip Inner column strip strip Edge strip
Moment (kNm/m)
Support modelled as Springs at edge
full 3D continuum of column
Journal of the South African Institution of Civil Engineering • Volume 56 Number 1 April 2014 91
SANS 10100:2000 (SANS 2000). The structural use of Authorities). Code of practice for the design of Zienkiewicz, O C, Brotton, D M & Morgan, L 1976. A
concrete, Part 1. Pretoria: South African Bureau of highway bridges and culverts in South Africa, Part finite element primer for structural engineers, The
Standards. 3. Pretoria: Department of Transport, Technical Structural Engineer, 54(10):389–397.
TMH7 1989 (CSRA – Committee of State Road Manuals for Highways.
92 Journal of the South African Institution of Civil Engineering • Volume 56 Number 1 April 2014