Structures and Buildings Volume 165 Issue SB3 Shear capacity of reinforced concrete corbels using mechanism analysis

Yang and Ashour

Proceedings of the Institution of Civil Engineers Structures and Buildings 165 March 2012 Issue SB3 Pages 111125 http://dx.doi.org/10.1680/stbu.2012.165.3.111 Paper 900101 Received 30/11/2009 Accepted 26/08/2010 Keywords: beams & girders/concrete structures/shear failures

ICE Publishing: All rights reserved

Shear capacity of reinforced concrete corbels using mechanism analysis

1 j Keun-Hyeok Yang MSc, PhD, Archi. Engng 2 j Ashraf F. Ashour MSc, PhD, CEng, FIStructE

Associate Professor, Department of Architectural Engineering, Kyonggi University, Suwon, Kyonggi-do, South Korea

Reader in Structural Engineering, EDT1, School of Engineering, Design and Technology, University of Bradford, Bradford, UK

1 j

2 j

A mechanism analysis is developed to predict the shear capacity of reinforced concrete corbels. Based on shear failure observed in experimental tests, kinematically admissible failure mechanisms are idealised as an assemblage of two rigid blocks separated by a failure plane of displacement discontinuity. Shear capacity predictions obtained from the developed mechanism analysis are in better agreement with corbel test results of a comprehensive database compiled from the available literature than other existing models for corbels. The developed mechanism model shows that the shear capacity of corbels generally decreases with the increase of shear span-to-depth ratio, increases with the increase of main longitudinal reinforcement up to a certain limit beyond which it remains constant, and decreases with the increase of horizontal applied loads. It also demonstrates that the smaller the shear span-tooverall depth ratio of corbels, the more effective the horizontal shear reinforcement.

1.

Introduction

Reinforced concrete (RC) corbels, generally dened as short cantilevers having shear span-to-depth ratios less than or equal to 1.0, are commonly used to transfer loads from beams to columns or walls in precast concrete construction. Corbels are primarily designed to resist vertical loads and horizontal actions owing to restrained shrinkage, thermal deformation and creep of the supported beam and/or breaking of a bridge crane. Owing to their geometric proportions, the capacity of RC corbels is governed by shear rather than exure and shear deformations are not negligible, similar to deep beams. RC corbels are identied as discontinuity regions (Schlaich et al., 1987) where strain distribution is signicantly non-linear and conventional beam theory would not be applicable. RC corbels are generally known to display several modes of failure (ASCEACI Committee 426 (ASCEACI, 1973); Kriz and Raths, 1965; Russo et al., 2006), such as anchorage failure or yielding of main longitudinal reinforcement, shear splitting at the interface between column and corbel, diagonal splitting or crushing of the concrete strut joining the loading point and bottom point of the interface, and local crushing failure under the bearing plate of the applied load. Beam-shear failure occurring between the interface and diagonal plane joining the loading point and bottom point of the interface was the most commonly

observed failure in experimental tests (Campione et al., 2007; Fattuhi, 1994a, 1994b, 1994c; Foster et al., 1996; Russo et al., 2006; Yong and Balaguru, 1994). Premature failure modes, such as anchorage failure of main longitudinal reinforcement and bearing failure (Campione et al., 2007), would be prevented by proper reinforcement detailing, as specied in ACI 318-08 (ACI, 2008) or EC 2 (BSI, 2004). On the other hand, it is not straightforward to evaluate the capacity of corbels owing to beam-shear failure as it involves various parameters such as the amount of main longitudinal and horizontal shear reinforcements, shear span-to-depth ratio and amount of horizontal load (Fattuhi, 1994a, 1994c; Yong and Balaguru, 1994). In the present study, a mechanism analysis for corbels is developed using upper-bound theorem of concrete plasticity to complement the strut-and-tie models driven from an equilibrium approach and calibrated against limited test results. Based on experimental tests carried out by many researchers (Campione et al., 2007; Fattuhi, 1994a, 1994b, 1994c; Foster et al., 1996; Mattock, 1976; Yong and Balaguru, 1994), kinematically admissible failure modes are idealised and studied. The effect of different parameters on the shear capacity of corbels is also investigated using developed mechanism analysis, existing empirical equations by Fattuhi (1994a), the shear-friction model by ACI 318-08 (ACI, 2008), a simplied strut-and-tie model by 111

Structures and Buildings Volume 165 Issue SB3

Shear capacity of reinforced concrete corbels using mechanism analysis Yang and Ashour

Russo et al. (2006), a softened strut-and-tie model by Hwang et al. (2000) and test results of a comprehensive database compiled from available literature.

2.

Review of existing models

Currently available theoretical models to evaluate the shear capacity of corbels would be classied into three categories: empirical equations (Fattuhi, 1994a) calibrated against test results, formulas (ACI 318-08, (ACI, 2008); Mattock, 1976) developed from shear-friction theory (Hermansen and Cowan, 1974; Mattock, 1976), and strut-and-tie models (Hagberg, 1983; Hwang et al., 2000; Russo et al., 2006; Siao, 1994; Solanki and Sabnis, 1987) including plastic truss models (Campione et al., 2007). Existing models for estimating the shear capacity of corbels are briey summarised below. 2.1 Empirical equation Based on extensive test results, Fattuhi (1994a) combined different parameters inuencing the shear capacity, V n , of corbels and developed the following formula:

( 1:23 f c 9 , where f c 9 is cylinder compressive strength), rst ( Ast = bd ) is ratio of main longitudinal reinforcement, Ast and fy are area and yield point stress of main longitudinal reinforcement, respectively, and N = V is the ratio between horizontal and vertical loads applied to corbels. The values of the constants, k 1 to k 8 , obtained from regression analysis of test data are 611, 0.7298, 0.3569, 0.8204, 0.5745, 0.1644, 0.0261, 0.1342, respectively. The above equation takes no account of the inuence of horizontal shear reinforcement. Although it is relatively easy to use empirical equations such as Equation 1, its application to corbels having parameters deviated from the range used for calibration is doubtful. In addition, this equation may cause overtting owing to the high number of constants used to develop it. 2.2 Formula using shear-friction theory ACI 318-08 (ACI, 2008) species the shear capacity of corbels considering the load transfer capacity of horizontal reinforcement by shear friction and exural yielding of the main longitudinal reinforcement at the interface as follows

Vn k 1 ( bd ) k 2 f ct k 3 ( a= d ) k 4 rst k 5
1:

3 (10)

k6( N =V )

f y = f cu

k7

( d = h)

k8

where b, d and h (in mm) are width, effective depth and overall depth of the interface between column and corbel, respectively, a is shear span measured from the loading point to the interface as shown in Figure 1, fct (in MPa) is indirect tensile splitting strength of concrete, fcu (in MPa) is cube compressive strength of concrete

6 7 Vn min4 A f jd N h d jd 5 st y a
2:

Ast f y Ah f yh N ,

< min0:2 f c 9 bd , 5:5 bd

a Ast

V N

d Ah

Potential failure plane in strut-and-tie model Potential failure plane in shear-friction theory

Figure 1. Potential failure planes of corbels considered in existing models

where is coefcient of friction which is taken as 1.4 for monolithic construction, Ah and fyh are total area and yield strength of horizontal shear reinforcement, and jd [ d 0:5( Ast f y N )=(0:85 f c 9 b)] is moment lever arm. The rst term of the right-hand side of Equation 2 indicates the shear friction resistance of longitudinal tensile reinforcement and horizontal shear reinforcement across the interface between corbel and column, while the second term refers to exural resistance of longitudinal tensile reinforcement. Equation 2 also assumes that applied horizontal forces can be resisted by longitudinal tensile reinforcement. The above formula based on shear-friction theory neglects the shear transfer capacity of concrete, as the critical failure section is always assumed at the interface between corbel and column, as shown in Figure 1. However, failure of corbels having a very small shear span-to-depth ratio or sufcient horizontal shear reinforcement seldom occurs along the interface, as pointed out by Hwang et al. (2000) and Hermansen and Cowan (1974). In addition, ACI 318-08 limits concrete strength of corbels to f c 9 < 27:5 MPa when the shear capacity of corbels is governed by the second part of Equation 2, which does not allow full utilisation of higher strength concrete. For corbels having shear span to depth ratio greater than 1.0, ACI 318-08 recommends the use of a strut-and-tie model described in ACI 318-08, Appendix A.

112

Structures and Buildings Volume 165 Issue SB3

Shear capacity of reinforced concrete corbels using mechanism analysis Yang and Ashour

V a T jd1 2 Fh FD v r d N

non-linear distributions of strains and stresses of corbels to estimate the moment lever arm, jd 1 , as

4:

jd 1

  k d 1 3

jd1 2

(a)

05ws cos 1

q where k [ ( nrf )2 2 nrf nrf ] is the ratio of compression zone depth to effective depth at the interface, rf [ ( Ast An )= bd ] is equivalent main longitudinal reinforcement ratio, An ( N = f y ) is area of main longitudinal reinforcement resisting the applied horizontal force, Np , n ( Es = Ec ) is modular ratio of elasticity, Es and Ec ( 4700 f c 9 in MPa) are elastic moduli of reinforcement and concrete, respectively. In Equation 3, Fh is assumed to be a function of the area, Ah , and average strain of horizontal shear reinforcement, and FD is dependent on the concrete strut width, ws , assumed to be equal to kd , angle, , and maximum allowable compressive stress, d,max , of concrete strut. Hwang et al. adopted the softened stressstrain curve of cracked concrete proposed by Zhang and Hsu (1998) to estimate d,max that is dependent on the average principal compressive (d ) and tensile (r ) strains in concrete and a softening coefcient, : The compatibility condition considered by Hwang et al. relates the average principal strains to the average horizontal (h ) and vertical (v ) strains as below 5:

r d h v

1 ws 05ws sin1

(b)

Figure 2. Typical strut-and-tie models of corbels: (a) Hwang et al. (2000); (b) Russo et al. (2006)

To avoid iterative procedure, Hwang et al. proposed reasonable values for the average horizontal (h ) and vertical (v ) strains and hence, the values of d and r can be determined using numerical analysis. As a result, the shear capacity of corbels using the softened strut-and-tie model proposed by Hwang et al. can be obtained, although several assumptions based on linear elastic beam theory are imposed. On the other hand, Russo et al. (2006) derived a simple equation using the strut-and-tie model shown in Figure 2(b). In this model, load transfer mechanisms of cracked concrete and horizontal shear reinforcement were considered and simplied as below 6:

2.3 Strut-and-tie model Considering equilibrium, compatibility and constitutive laws of cracked reinforced concrete, Hwang et al. (2000) predicted the shear capacity of corbels based on the softened strut-and-tie model shown in Figure 2(a) as follows 3:

Vn 0:8 k f c 9 cos 1 0:65rh f yh cot 1 bd

Vn FD sin Fh tan

where FD and Fh are compression force in concrete strut and tension force in horizontal shear reinforcement, respectively, [ tan1 ( jd 1 = a)] is the angle of concrete strut to the longitudinal axis of corbels as shown in Figure 2(a). Hwang et al. used the linear bending theory of reinforced concrete beams with only tensile reinforcement neglecting shear deformations and

where is a non-dimensional interpolating function to provide a single expression for softening coefcient, , given in the softened strut-and-tie model above, 1 represents angle of concrete strut to the column axis as shown in Figure 2(b), which can be obtained from the elastic beam theory and trigonometric relations, and rh ( Ah = bd ) is ratio of horizontal shear reinforcement. The principal tensile strain, r , was assumed to be f ct = Ec to obtain directly the softening coefcient, : Russo et al. also proposed that could be approximately expressed as a poly113

Structures and Buildings Volume 165 Issue SB3

Shear capacity of reinforced concrete corbels using mechanism analysis Yang and Ashour

nomial of third degree in f c 9 : Different constants used in Equation 6 were determined by calibrating the strut-and-tie model against 243 test results of corbels. Other strut-and-tie models (Hagberg, 1983; Siao, 1994; Solanki and Sabnis, 1987) were developed for corbels but they are similar in principle to those presented above. Strut-and-tie models are considered to be a good and rational tool for the design of discontinuous regions such as RC corbels as they provide a systematic load transfer and better understanding of internal forces. However, several assumptions are imposed for simplication. The width and inclination of concrete strut is evaluated from the neutral axis depth of the interface using the conventional elastic beam theory that is not applicable to deep corbels where both strains and stresses are non-linear. In addition, biaxial stresses in compressed concrete struts are chosen arbitrarily or calculated from tensile strength of concrete. In the following a mechanism analysis of corbels based on the upper-bound theorem is developed to complement the strut-and-tie models presented above.

Hyperbolic yield line Ast Rigid block II Y (Origin of O global axes)

a xe

Vn N Rigid block I r YI.C. X I.C.

XI.C.

(a)

a xe Ast

Vn N Rigid block I r d h YI.C.

3.

Mechanism analysis

Rigid block II I.C. Y (Origin of O global axes) XI.C. X

3.1 Failure mechanism of corbels Failure planes of corbels commonly occurred along the diagonal plane joining the inner edge of loading plate and bottom point of the interface (ASCE-ACI Committee 426, (ASCEACI, 1973); Kriz and Raths, 1965). However, few test specimens exhibited splitting cracks stemmed from a point close to the loading plate at failure. Thus, the failure mechanism can be idealised as an assemblage of two rigid blocks separated by a yield line representing the failure zone along which in-plane displacement discontinuity occurs (Nielsen, 1984). It is also assumed that the failure surface joins the base point of corbel interface and the inner corner of loading plate. It should be noted that other failure modes such as anchorage failure of main longitudinal reinforcement, bearing failure and failure along the interface between corbel and column can be mostly prevented by following proper reinforcement detailing (ACI 318-08 (ACI, 2008), EC 2 (BSI, 2004), Yong and Balaguru, 1994). Rigid block I undergoes a rotation around an instantaneous centre (I.C.), while rigid block II is considered to be xed with the supporting column as shown in Figure 3. The I.C. can be located anywhere in the vertical plane of corbel and its position identies the shape of the failure surface. Jensen (1982) proved that the optimal shape of the yield line is a hyperbola with orthogonal asymptotes at the I.C., as shown in Figure 3(a), and the yield line reduces to a straight line to achieve a stationary value of the total energy dissipation when the I.C. approaches innity. On the other hand, the yield line turns into two straight segments when the I.C. of relative rotation lies inside or on a circle whose diameter is the straight yield line joining the inner edge of loading plate and bottom point of the interface as shown in Figure 3(b). Therefore, the failure planes having either a hyperbolic yield line or two straight yield lines identify crushing of concrete strut and beam-shear failure including the yielding of main longitudinal reinforcement, respectively. For each location of I.C., an upper 114

(b)

Figure 3. Idealised failure mechanism of corbels: (a) hyperbolic yield line; (b) yield line with two straight segments

bound load to the corbel capacity can be established. However, the optimum position of I.C. corresponds to the minimum load capacity as explained below. 3.2 Material modelling Concrete is assumed to be a rigid perfectly plastic material obeying the modied Coulomb failure criteria (Nielsen, 1984) with zero tension cut-off. The effective compressive strength to , is be used in calculation, f c 7:

f 9 c e f c

where e is effectiveness factor that is introduced to account for the limited ductility of concrete and to absorb other shortcomings of applying the plasticity theory to concrete. Although there is no unied approach for evaluating the effectiveness factor of concrete, many investigations (Ashour and Morley, 1996; Brstrup,

Structures and Buildings Volume 165 Issue SB3

Shear capacity of reinforced concrete corbels using mechanism analysis Yang and Ashour

1990) clearly showed that the effectiveness factor depends on concrete strength and geometrical properties of reinforced concrete members. The measure of success of the mechanism analysis presented below would depend on the extent that the effectiveness factor is reasonably uniform and predictable across the range of corbels considered. In the present study, Nielsens model considering the effect of compressive strength of concrete modied by the equation proposed by Brstrup (1990) for the inuence of shear span-to-overall depth ratio is adopted for the evaluation of the effectiveness factor as given below

10:

Wc

e f c 9 b F ( O9) 2

where F ( O9) is a function of the position of the I.C. and, consequently, the shape of the yield line; Equations 11(a) and 11(b) below gives the value of F ( O9) for the hyperbolic and two straight segments yield lines respectively (Jensen, 1982) 11a: F O9 r1 sin h= sin for r . h=2 sin

8:

fc 9 0:8 200



1 0: 2

a h


2 2 11b: F O9 X I:C: Y I:C: for r < h=2 sin

All reinforcement is considered to carry only axial tensile and compressive stresses and its dowel action is ignored. Steel reinforcement in both tension and compression is assumed to be a rigid perfectly plastic material with yield strength, fy : Yielding of steel reinforcements is generally achieved by utilising some sort of a mechanical anchorage at corbel end. Strain hardening of steel reinforcement is ignored as it requires excessive wide cracks in concrete beyond the failure mechanism considered in the present analysis. 3.3 Work equation The upper-bound theorem is based on the energy principle, by equating the total internal energy, WI , to the external work done, WE : The total internal energy mainly depends on the position of the I.C. and the amount of internal stresses in both concrete along the yield line and reinforcement crossing the yield line. The energy ( W c )l dissipated in concrete per unit length of the yield line is written in the following general form (Nielsen, 1984)

where ( tan1 ( h= xe )) is angle of the diagonal line joining the inner edge of the loading plate and bottom point of the interface to the longitudinal axis of corbels and xe is clear shear span measured from inner edge of the loading plate to interface. In Equation 11a, both r and depend on the position of the I.C. Hence, the energy dissipated in concrete is a function of horizontal ( X I:C: ) and vertical ( YI:C: ) coordinates of the I.C. The relative displacement of reinforcement, s , can be expressed as rs , as shown in Figure 4. Therefore the energy, W s , dissipated in main longitudinal and web reinforcement crossing the yield line is calculated from
n X i 1

12:

Ws

As, i f y, i rs, i cos s, i

9:

W c l

e f c 9 b1 sin 2

where is relative displacement of rigid block I and is the angle between the relative displacement at the midpoint of the chord and yield line chord as shown in Figure 3. The relative displacement, , can be expressed as r, where r is distance between the midpoint of yield line chord and the I.C. and is rotation of rigid block I. For a yield line with two straight segments as shown in Figure 3(b), it should be noted that the two segments intersect at the I.C. of relative rotation, indicating that one segment of the yield line is under compression and the other is in tension owing to the relative rotation between the two blocks. Therefore, the yield line segment under tensile stresses with pure separation ( =2) cannot contribute to the internal energy distribution owing to the assumption of zero tensile concrete strength. If the origin of global coordinates is set to be at the bottom point of the interface as shown in Figure 3, the total internal energy, W c , dissipated in concrete along the hyperbolic or two straight yield line is (Jensen, 1982)

where n is number of reinforcing bars crossing the yield line, As, i and f y, i are area and yield strength of reinforcing bar, i, crossing the yield line, respectively, rs, i is the distance between the intersection point of reinforcing bar, i, with the yield line and the I.C. and s, i is angle between the relative displacement, s, i , and the reinforcing bar, i, crossing the yield line. In case of horizontal
Yield line Rigid block I (xi, yi) Rigid block II s,i

) rs,i = [(XI.C. xi 2 05 ] ) y (Y i
I.C.

I.C. (XI.C., YI.C.)

s,i

Reinforcing bar i

Figure 4. Reinforcing bar crossing yield line

115

Structures and Buildings Volume 165 Issue SB3

Shear capacity of reinforced concrete corbels using mechanism analysis Yang and Ashour

reinforcement, cos s, i j YI:C: yi j= rs, i , where yi is vertical coordinate of the intersection point of reinforcing bar, i, and the yield line, as shown in Figure 4. The external work, W E , done by the vertical load, Vn , and horizontal load, N , on rigid block, I, is (see Figure 3) 13:

W E V n j a X I:C: j N j h YI:C: j

beams and showed that the position of the I.C. of relative rotation commonly depends on the amount of main longitudinal reinforcement and shear span-to-overall depth ratio. The two straight yield lines clearly indicate yielding and participation of the main longitudinal top reinforcement of corbels, whereas in case of hyperbolic yield line, the I.C. is more likely to be located along the main longitudinal top reinforcements and consequently, these reinforcements do not contribute to the internal energy dissipation. For the case of corbels, the position of the I.C. would also be inuenced by the amount of the applied horizontal force. When shear reinforcement is not provided in concrete corbels, the determination of the minimum shear capacity could be achieved by considering the two differential equations, (@ Vn )=(@ X I:C: ) 0 and (@ Vn )=(@ YI:C: ) 0: In addition, for corbels with sufciently strong main longitudinal reinforcement or subjected to high horizontal load N, the vertical coordinate YI:C: of the I.C. would be located at the level of main longitudinal reinforcement (Ashour, 2000) or horizontal applied load. As a result, the optimisation process is further simplied as the shear capacity turns to be a function of only X I:C: : A numerical example explaining the solution procedure of shear capacity calculation of specimen PA 2 tested by Foster et al. (1996) using the mechanism approach is presented in the Appendix.

In Equation 13 above, the horizontal load, N, is assumed to act at the level of the corbel top surface. However, this could be easily adjusted to other levels as the case in some experiments (Fattuhi, 1994a, 1994c; Yong and Balaguru, 1994). The horizontal load, N, applied to corbels can be normalised with respect to the yielding force, Ast fy , of the main longitudinal reinforcement to simplify the calculation of the vertical coordinate, YI:C: , of the I.C. Equating the total internal energy dissipated in concrete and reinforcement to external work done, the shear capacity, Vn , can be obtained as below

2 bhf 9c 4 e F O9 st j d YI:C: j j h Y1:C j Vn j a X I:C: j 2 h

14:
n X i 1

# s, i rs, i cos s, i

4.

Comparisons and discussions

9 )) is the main longitudinal reinforcement where st ( ( Ast f y = bhf c index, ( N = Ast f ) is the ratio of horizontal load to the yield force of main longitudinal reinforcement and s, i ( ( As, i f y, i = bhf c 9 )) is the web reinforcement index for each individual reinforcing bar, i, crossing the yield line. Equation 14 implies that the horizontal load is one of various parameters inuencing the vertical coordinate, YI:C: , of the I.C. 3.4 Solution procedure According to the upper-bound theorem, the collapse occurs at the least strength. The shear capacity of corbels is implicitly expressed as a function of the position of the I.C. ( X I:C: , YI:C: ), as given by Equation 14. The horizontal ( X I:C: ) and vertical ( YI:C: ) coordinates of the I.C. are repeatedly tuned until the minimum shear capacity is obtained. For each position of the I.C., the shape of the yield line can be identied, the energy dissipated in concrete along the yield line can be estimated using Equations 10 and 11a or Equations 10 and 11b for the hyperbolic and two straight segments yield lines, respectively and nally the corresponding shear capacity is calculated using Equation 14. The process of adjusting the position of I.C. is achieved by reliable numerical optimisation procedures provided in Matlab software (Chapman, 2004). However, if such routines are not available, a simple algorithm has to be developed to survey a wide range of locations of I.C. to achieve the minimum load capacity of corbels. Ashour (2000) developed a mechanism analysis for RC deep 116

4.1 Database of reinforced concrete corbels Test results of 265 RC corbels were compiled from different experimental investigations carried out by Abdul-Wahab (1989), Campione et al. (2007), Chakrabarti et al. (1989), Fattuhi and Hughes (1989a, 1989b), Fattuhi (1994a, 1994b, 1994c), Foster et al. (1996), Kriz and Raths (1965), Mattock et al. (1976), and Yong and Balaguru (1994), and other test results originally collected by Kriz and Raths (1965). All corbels were reported to fail in shear owing to a major diagonal crack within the corbel shear span. Distribution of different parameters in the database of the 265 corbels is listed in Table 1. It should be noted that corbel specimens with bre reinforced concrete (Fattuhi, 1990, 1994b) are excluded as the effective strength of bre reinforced concrete struts is signicantly affected by the volume fraction and aspect ratio of bres. The test specimens in the database were made of concrete having a relatively low compressive strength of 15.2 MPa and very high compressive strength of 105 MPa, however, 73.5% of the corbel specimens in the database had compressive strength between 20 MPa and 40 MPa. The shear span-to-overall depth ratio of corbels ranged from 0.1 to 1.0. The main longitudinal reinforcement index st varied between 0.023 and 0.49, whereas the horizontal load to yield force of main longitudinal reinforcement ratio ranged from 0 to 1.5. 74.3% and 77.4% of the corbel specimens in the database had no horizontal shear reinforcement or horizontal force, respectively. All test specimens had no vertical shear reinforcement. 4.2 Comparison of shear load capacity Table 2 gives the mean, standard deviation, cs,s , and coefcient of variation, cs,v , of the ratio between measured and predicted shear

bw 250,300 32 30,40 52 0.3,0.4 50 0.075,0.1 45 0.01,0.025 19 0.3,0.4 3 0.4,0.5 14 0.5,0.6 14 0.6,0.7 8 0.7,1.0 9 0.025,0.05 21 0.05,0.075 14 0.075,0.1 3 0.1,0.125 4 0.125,0.15 4 1.0,1.25 5 0.1,0.125 48 0.125,0.15 25 0.15,0.2 28 0.2,0.25 22 0.25,0.3 10 0.4,0.5 31 0.5,0.6 36 0.6,0.7 30 0.7,0.8 16 0.8,0.9 6 0.9,1.0 17 0.3,0.49 18 0.15,0.23 2 1.25,1.5 1 40,50 29 50,60 7 60,70 10 70,80 11 80,90 90,105 8 300,400 9 400,500 72 500,600 20 600,700 62 700,800 7 800,940 11

Range (mm) Frequency

50,100 16

100,125 10

125,150 15

150,175 76

175,200

200,225 128

225,250

250,275 14

300,360 6

Range (mm) Frequency

140,200 36

200,250 16

Structures and Buildings Volume 165 Issue SB3

fc 9

Range (MPa) Frequency

15,20 5

20,30 143

a= h

Range Frequency

0.1,0.2 34

0.2,0.3 45

st

Range Frequency

0.023,0.05 24

0.05,0.075 45

Range Frequency

0 197

0.001,0.01 1

Range Frequency

0 205

0.12,0.3 6

Table 1. Distribution of different parameters in the database of 265 corbels

Fattuhi (Empirical equation)

ACI 318-08 (Shear-friction theory)

Hwang et al. (Softened strut-and-tie model)

Russo et al. (Strut-and-tie model)

This study (Mechanism analysis)

W/O W/H W/N W/HN Total W/O W/H W/N W/HN Total W/O W/H W/N W/HN Total W/O W/H W/N W/HN Total W/O W/H W/N W/HN Total

Shear capacity of reinforced concrete corbels using mechanism analysis Yang and Ashour

cs,m cs,s cs,v

0.83 1.23 0.86 0.50 0.86 1.43 1.67 1.07 1.32 1.41 1.16 1.17 1.02 1.07 1.14 1.0 1.02 0.89 0.89 0.98 1.03 0.96 1.05 0.96 1.01 0.26 0.24 0.21 0.15 0.31 0.57 0.48 0.24 0.35 0.53 0.27 0.15 0.37 0.26 0.27 0.24 0.19 0.32 0.25 0.25 0.22 0.18 0.22 0.19 0.21 0.31 0.19 0.25 0.29 0.36 0.4 0.28 0.23 0.27 0.34 0.23 0.13 0.36 0.25 0.24 0.24 0.18 0.36 0.28 0.25 0.21 0.19 0.21 0.19 0.21

Note : cs,m , cs,s and cs,v indicate the mean, standard deviation and coefcient of variation of cs ( Vn )Exp: =( Vn )Pre: , respectively W/O, W/H, W/N, and W/HN refer to 164 corbels without both horizontal shear reinforcement and horizontal load, 41 corbels with horizontal shear reinforcement and without horizontal load, 33 corbels without horizontal shear reinforcement and with horizontal load, and 27 corbels with both horizontal shear reinforcement and horizontal load, respectively

Table 2. Comparison of experimental and predicted shear capacities of corbels in the database using different methods

117

Structures and Buildings Volume 165 Issue SB3

Shear capacity of reinforced concrete corbels using mechanism analysis Yang and Ashour

capacities, cs ( Vn )Exp: =( Vn )Pre: , of corbels with different horizontal shear reinforcement and horizontal load. The distribution of cs of the total specimens in the database against the shear span-to-overall depth ratio, a= h, is also shown in Figure 5; Figure 5(a) for the empirical equation (Equation 1) proposed by Fattuhi, Figure 5(b) for the ACI 318-08 equation (Equation 2) based on shear-friction theory, Figure 5(c) for the simplied softened strutand-tie model by Hwang et al., Figure 5(d) for the strut-and-tie model by Russo et al., and Figure 5(e) for the current mechanism analysis. Fattuhis equation largely overestimates the shear capa30 25

city of corbels without both horizontal shear reinforcement and horizontal load when a= h < 0:3, whereas it underestimates that of corbels with horizontal shear reinforcement and without horizontal load regardless of a= h, as the contribution of horizontal shear reinforcement to shear transfer capacity is not included in this method. The largest standard deviation and coefcient of variation of all models appear in the equations specied in ACI 318-08 based on the shear-friction theory. In particular, the ACI 318-08 equations are occasionally unconservative for corbels without horizontal shear reinforcement and subjected to horizontal load, as
30 25

(Vn)Exp. / (Vn)Pre.

(Vn)Exp. / (Vn)Pre.

20 15 10 05 0 0 01 02 03 04 05 06 07 08 Shear span-to-overall depth ratio, a/h (a) 09 10

20 15 10 05 0 0 01 02 03 04 05 06 07 08 Shear span-to-overall depth ratio, a/h (b) 09 10

30 25

30 25

(Vn)Exp. / (Vn)Pre.

(Vn)Exp. / (Vn)Pre.

20 15 10 05 0 0 01 02 03 04 05 06 07 08 Shear span-to-overall depth ratio, a/h (c) 09 10

20 15 10 05 0 0 01 02 03 04 05 06 07 08 Shear span-to-overall depth ratio, a/h (d) 09 10

30 25

(Vn)Exp. / (Vn)Pre.

20 15 10 05 0 0 01 02 03 04 05 06 07 08 Shear span-to-overall depth ratio, a/h (e) 09 10

Without horizontal shear reinforcement and horizontal load With horizontal shear reinforcement and without horizontal load Without horizontal shear reinforcement and with horizontal load With horizontal shear reinforcement and horizontal load

Figure 5. Comparisons of measured and predicted shear capacities: (a) Fattuhis equation; (b) ACI 318-08; (c) softened strut-and-tie model by Hwang et al. (2000).; (d) strut-and-tie model by Russo et al. (2006); (e) mechanism analysis

118

Structures and Buildings Volume 165 Issue SB3

Shear capacity of reinforced concrete corbels using mechanism analysis Yang and Ashour

shown in Figure 5(b) and indicated in Table 2, where the lowest cs,m is recorded for such a case. The softened strut-and-tie model proposed by Hwang et al. generally underestimates the shear capacity of corbels, especially those without horizontal shear reinforcement. Russo et al.s strut-and-tie model shows reasonable agreement with test results, but high deviation exhibited by few specimens without horizontal shear reinforcement or with horizontal load as shown in Figure 5(d). On the other hand, predictions obtained from the developed mechanism analysis are in better agreement with test results than other models regardless of shear span-to-overall depth ratio, horizontal shear reinforcement and horizontal load, as given in Table 2 and Figure 5(e).

5.

Further experimental verication

The inuence of various parameters on the shear capacity of corbels is evaluated using the empirical equations proposed by Fattuhi, the equation specied in ACI 318-08 based on shearfriction theory (SFT), strut-and-tie models developed by Hwang et al. and Russo et al., the current mechanism analysis, and appropriate experimental results in the database. In this parametric study, the width and overall depth of section in the interface between columns and corbels, and compressive strength of concrete are selected to be xed at 150 mm, 600 mm, and 30 MPa, respectively. The effective depth of section and width of loading plate are also assumed to be 0.9 h and 0.1 h, respectively. In addition, the shear capacity of corbels is normalised by bhf c 9: In a few cases, when a value of corbel parameter used in the parametric study does not match that in the database, a range of such parameters is extracted from the database and presented in the following gures to validate the trend predicted by different theoretical models. 5.1 Shear span-to-overall depth ratio Figure 6 shows the inuence of the shear span to depth ratio, a= h, on the normalised shear capacity, Vn = bhf c 9 , of corbels without both horizontal shear reinforcement and horizontal load. Calculations obtained from ACI 318-08 remain constant when a= h < 0:6 as the shear-friction theory considers only the shear transfer of reinforcement crossing the failure plane at the interface, independent of a= h: On the other hand, normalised shear capacities of corbels obtained from experimental results, empirical equations, strut-and-tie models and mechanism analysis generally decrease with the increase of a= h: However, the strutand-tie models give higher predictions than mechanism analysis, empirical equation and experimental results when a= h > 0:5, especially for st 0:05: When a= h , 0:3, the strut-and-tie models underestimate the normalised shear capacity, especially when st 0:1, whereas the empirical equation by Fattuhi highly overestimates the normalised shear capacity. The predictions from ACI 318-08 remain constant up to a= h 0:6, beyond which the predictions decrease with the increase of a= h: As a result, ACI 318-08 generally underestimates the normalised shear capacity when a= h , 0:5, but overestimates test results when a= h . 0:7: On the other hand, the proposed mechanism analysis is closer to the test results, regardless of the variation of a= h:

5.2 Main longitudinal reinforcement The inuence of the main longitudinal reinforcement index, st , on the normalised shear capacity, Vn = bhf c 9 , of corbels without horizontal shear reinforcement and horizontal load for two different shear span-to-overall depth ratios is presented in Figure 7. The prediction obtained from the mechanism analysis increases with the increase of st up to a certain limit beyond which the normalised shear capacity remains constant, agreeing with the test results. When the main longitudinal reinforcement reaches this limit, it becomes strong enough not to yield and hence the I.C. lies at the level of main longitudinal reinforcement. In this case, main longitudinal reinforcement would have no contribution to the shear capacity of corbels as depicted in Figure 7. The shear capacity predicted from ACI 318-08 is governed by the upper limit specied by Equation 2 with the increase of st : For the mechanism analysis and ACI 318-08, the limit to the st depends on a= h, indicating that the value of st to achieve the peak point decreases with the decrease of a= h: On the other hand, the prediction obtained from empirical equation, and strut-and-tie models increases with the increase of st without any limits, exhibiting large overestimation of test results for large st , especially for a= h 0:9: 5.3 Horizontal load Figure 8 shows the effect of the ratio, , of the applied horizontal load to yield force of main longitudinal reinforcement on normalised shear capacity, Vn = bhf c 9 , of corbels without horizontal shear reinforcement for two different shear span to overall depth ratios, a= h: The normalised shear capacity of corbels steadily decreases with the increase of as predicted by all theoretical models and experimental results. However, the predictions obtained from the mechanism analysis are closer to the experimental results than other models. In particular, strut-and-tie models largely overestimate the effect of on the normalised shear capacity for a= h 0.56. 5.4 Horizontal shear reinforcement The effect of horizontal shear reinforcement index, h ( ( Ah f yh = bhf c 9 )), on the normalised shear capacity, Vn = bhf c 9, of corbels without horizontal load is shown in Figure 9. The normalised shear capacity of corbels increases with the increase of h as predicted by the mechanism analysis and strut-and-tie models and also supported by the limited experimental results available. The effect of horizontal shear reinforcement on the shear capacity of corbels is more prominent for low values of a= h, namely, a higher increasing rate is developed in corbels with a= h of 0.3 than corbels with a= h of 0.9. However, prediction obtained from Fattuhis equations does not account for h , and the ACI 318-08 prediction is also independent of h for a= h 0.9 and slightly increases with the increase of h for a= h 0.3.

6.

Conclusions

A mechanism analysis based on upper-bound theorem is proposed to predict the shear capacity of RC corbels. The effect of different 119

Structures and Buildings Volume 165 Issue SB3

Shear capacity of reinforced concrete corbels using mechanism analysis Yang and Ashour

025

Foster et al., 1996 (004 st 006) Fattuhi and Hughes, 1989a (st 007) Fattuhi, 1994c (003 st 004) Kriz and Raths, 1965 (003 st 007) Fattuhi ACI 318-08 STM (Hwang et al.) STM (Russo et al.) Mechanism analysis

020

Vn /(bhf c)

015

010

005

0 01

02

03

04 05 06 07 Shear span-to-overall depth ratio a/h (a)

08

09

10

025

020

Vn /(bhf c)

015 Fattuhi 010 ACI 318-08 STM (Hwang et al.) STM (Russo et al.) 005 Mechanism analysis Fattuhi, 1994a (008 st 011) Kriz and Raths, 1965 (008 st 012) 0 01 02 03 04 05 06 07 Shear span-to-overall depth ratio a/h (b) 08 09 10

Figure 6. Inuence of a/h on shear capacity of corbels: (a) jst 0.05; (b) jst 0.1

parameters on the shear capacity of corbels is also investigated using the developed mechanism analysis, empirical equations proposed by Fattuhi, ACI 318-08 based on shear-friction theory, strut-and-tie models developed by Hwang et al. and Russo et al. and test results in a comprehensive database collected from the available literature. Various analytical models developed in the literature and present study consider only beam-shear failure mode including yielding of main longitudinal reinforcement and crushing of concrete strut as other modes of failure such as bearing failure or anchorage failure of longitudinal reinforcement can be prevented by following reinforcement details specied in codes. The following conclusions may be drawn. 120

(a) Compared with the existing models, predictions obtained from the developed mechanism analysis are in better agreement with test results regardless of shear span-to-overall depth ratio, horizontal shear reinforcement and horizontal load. (b) The largest standard deviation and coefcient of variation of the ratio between measured and predicted shear capacities of corbels are shown by ACI 318-08. In addition, ACI 318-08 is unconservative for corbels having a shear span-to-overall depth ratio of more than 0.5. (c) The normalised shear capacities Vn = bhf c 9 of corbels obtained from experimental results, empirical equations, strut-and-tie

Structures and Buildings Volume 165 Issue SB3

Shear capacity of reinforced concrete corbels using mechanism analysis Yang and Ashour

035

030

025

Foster et al., 1996 (a/h 03) Fattuhi and Hughes, 1989a (a/h 035) Kriz and Raths, 1965 (026 a/h 033) Fattuhi ACI 318-08 STM (Hwang et al.) STM (Russo et al.) Mechanism analysis

Vn /(bhf c)

020

015

010

005

005

010

015

020

025

030

Main longitudinal reinforcement index st (a) 020 018 016 014

Vn /(bhf c)

012 010 008 006 004 002 0 0 005 010 015 Mattock, 1976 (a/h 09) Fattuhi, 1990 (a/h 09) Campione et al., 2007 (a/h 09) Fattuhi ACI 318-08 STM (Hwang et al.) STM (Russo et al.) Mechanism analysis 020 025 030 035 040 Main longitudinal reinforcement index st (b)

Figure 7. Inuence of jst on shear capacity of corbels: (a) a/h 0.3; (b) a/h 0.9

models and mechanism analysis generally decrease with the increase of shear span-to-overall depth ratio. (d ) The shear capacity obtained from the mechanism analysis increases with the increase of the main longitudinal reinforcement index up to a certain value beyond which it remains constant, agreeing with test results. (e) The normalised shear capacity, Vn = bhf c 9 , of corbels decreases with the increase of the horizontal load; the decreasing rate obtained from the mechanism analysis is similar to that of test results.

( f ) The effect of horizontal shear reinforcement on shear capacity is more prominent in corbels having small shear span-to-overall depth ratios.

Acknowledgements
This work was supported by the National Research Institute of Cultural Heritage and the Regional Research Centers Program (Bio-housing Research Institute), granted by the Korean Ministry of Education & Human Resources Development. 121

Structures and Buildings Volume 165 Issue SB3

Shear capacity of reinforced concrete corbels using mechanism analysis Yang and Ashour

030

025

020

Vn /(bhf c)

015 Fattuhi ACI 318-08 STM (Hwang et al.) STM (Russo et al.) Mechanism analysis Kriz and Raths, 1965 (007 st 017) (021 a/h 027) 0 02 04 06 08 Ratio of horizontal load to yield force of the main reinforcement, (a) 10

010

005

018 016 014 012

Vn /(bhf c)

010 008 Fattuhi 006 004 002 0 ACI 318-08 STM (Hwang et al.) STM (Russo et al.) Mechanism analysis Kriz and Raths, 1965 (a/h 056, 007 st 019) 0 02 04 06 08 Ratio of horizontal load to yield force of the main reinforcement, (b) 10

Figure 8. Inuence of on shear capacity of corbels (jst 0.12, jh 0): (a) a/h 0.25; (b) a/h 0.56

APPENDIX: Numerical example of mechanism analysis

The following numerical example presents the shear capacity calculation of specimen PA 2 tested by Foster et al. (1996) using the mechanism approach. The material and geometrical properties of the specimen PA 2 are as follows: b 150 mm, h 600 mm, fc 9 53 MPa, d 500 mm, a 300 mm, xe 250 mm, Ast 1884 mm2 (Y20), fy 450 MPa, Ah 157 mm2 (R10) fyh 360 MPa and sh 85 mm. There was no horizontal load applied to 122

the specimen. The measured shear capacity, Vn , was 800 kN. The following steps summarise the procedure: (a) Calculate the effectiveness factor for concrete, e , using Equation 8; ve 0.4815. (b) Calculate the main longitudinal reinforcement index, st ( Ast f y = bhf c 9 ), and the web reinforcement index, s, i ( Ah, i f yh, i = bhf c 9 ), for each individual bar crossing the yield line, st 0:1777, s, i 0:0118:

Structures and Buildings Volume 165 Issue SB3

Shear capacity of reinforced concrete corbels using mechanism analysis Yang and Ashour

025

020

Vn /(bhf c)

015 Foster et al., 1996 (a/h 03, st 008) 010 Yong and Balaguru, 1994 (a/h 03, st 007) Kriz and Raths, 1965 (026 a/h 053, 009 st 01) Fattuhi ACI 318-08 005 STM (Hwang et al.) STM (Russo et al.) Mechanism analysis 0 0 001 002 003 004 Horizontal shear reinforcement index h (a) 005 006

025

Campione et al., 2007 (a/h 09, 008 st 013) Fattuhi ACI 318-08 STM (Hwang et al.) STM (Russo et al.) Mechanism analysis

020

Vn /(bhf c)

015

010

005

001

002 003 004 Horizontal shear reinforcement index h (b)

005

006

Figure 9. Inuence of jh on shear capacity of corbels (jst 0.1, 0): (a) a/h 0.3; (b) a/h 0.9

(c) Determine the angle (tan1 ( h= xe )) of the diagonal line joining the inner edge of the loading plate and bottom point of the interface to the longitudinal axis of the corbel; 67.388. (d ) Different positions of the I.C. ( X I:C: , YI:C: ) have to be examined using the Matlab software optimiser (Chapman, 2004). For each position, the shape of the yield line has to be identied using Equation 11. The internal energy dissipated in concrete has to be estimated using Equations 10 and 11. For example, for XI:C: 150 mm and YI:C: 150 mm,

r 152 mm and h=(2 sin ) 325 mm , ( r < h=(2 sin )), the yield line is then identied to be two straight segments and F ( O9) is then obtained from Equation 11b, F(O9) 45 000 mm2 ; however for XI:C: 400 mm and YI:C : 500 mm, r 340 mm ( r . h=(2 sin )), the hyperbolic yield line occurred and F ( O9) is then obtained from Equation 11a as F(O9) 32 273 mm2 (e) The minimum shear capacity, Vn , of 784 kN is achieved when XI:C: 1950 mm and YI:C: 500 mm (i.e. along the main longitudinal reinforcement). This indicates that the main 123

Structures and Buildings Volume 165 Issue SB3

Shear capacity of reinforced concrete corbels using mechanism analysis Yang and Ashour

Corbel

XI:C : mm

YI:C: : mm

Measured shear capacity: kN

Predicted shear capacity: kN

% Ratio between predicted and measured 108 98

PA1 PA2

496 1950

500 500

550 800

596 784

Table 3. Comparisons between predictions of corbels PA1 and PA2

longitudinal reinforcement does not yield at failure. The predicted shear capacity is 98% of the measured shear failure load. ( f ) Table 3 compares the shear capacity and location of I.C. of corbels PA1 and PA2 tested by Foster et al. (1996) as predicted by the mechanism analysis. Corbel PA1 was a control specimen having the same geometrical dimensions and concrete strength as corbel PA2, but no secondary horizontal shear reinforcement. The optimum location of I.C. is moved to XI:C: 496 mm and YI:C: 500 mm and the predicted shear capacity is reduced to 596 kN for corbel PA1 (108% of the measured shear capacity). Table 3 also indicates that a 32% shear capacity increase owing to the use of secondary horizontal shear reinforcement of R10 at spacing of 85 mm is predicted using the mechanism analysis.
REFERENCES

Abdul-Wahab HMS (1989) Strength of reinforced concrete

corbels with bers. ACI Structural Journal 86(1): 6066.

ACI (American Concrete Institute) (2008) ACI Committee 318:

Building code requirements for structural concrete (ACI 31808) and commentary (ACI 318R-08). ACI, Farmington Hills, Michigan, USA.
ASCEACI (American Society of Civil EngineersAmerican Concrete Institute) (1973) Joint Task Committee 426: The shear strength of reinforced concrete members. ASCE Proceedings, 99(ST): 10911187. Ashour AF (2000) Shear capacity of reinforced concrete deep beams. Journal of Structural Engineering, ASCE 126(9): 10451052. Ashour AF and Morley CT (1996) Effectiveness factor of concrete in continuous deep Beams. Journal of Structural Engineering, ASCE 122(2): 169178. BSI (2004) The European Standard EN 1992-1-1:2004 Eurocode 2: Design of concrete structures. EC 2. BSI, London, UK. Brstrup MW (1990) Shear strength prediction-plastic method. In Reinforced Concrete Deep Beams (Kong FK (ed.)). Blackie, Glasgow and London, UK, pp. 182203. Chakrabarti PR, Farahani DJ and Kashou SI (1989) Reinforced and precompressed concrete corbels-an experimental study. ACI Structural Journal 86(4): 405412. Campione G, Mendola LL and Mangiavillano ML (2007) Steel ber-reinforced concrete corbels: experimental behavior and

shear strength prediction. ACI Structural Journal 104(5): 570579. Chapman SJ (2004) Matlab Programming for Engineers. Thomson, USA. Fattuhi NI and Hughes BP (1989a) Reinforced steel ber concrete corbels with various shear span-to-depth ratio. ACI Materials Journal 86(6): 590596. Fattuhi NI and Hughes BP (1989b) Ductility of reinforced concrete corbels containing either steel ber or stirrups. ACI Structural Journal 86(6): 644651. Fattuhi NI (1990) Strength of SFRC corbels subjected to vertical load. Journal of Structural Engineering, ASCE 116(3): 701 717. Fattuhi NI (1994a) Reinforced corbels made with plain and brous concretes. ACI Structural Journal 91(5): 530536. Fattuhi NI (1994b) Strength of FRC corbels in exure. Journal of Structural Engineering, ASCE 120(2): 360377. Fattuhi NI (1994c) Reinforced corbels made with high-strength concrete and various secondary reinforcements. ACI Structural Journal 91(4): 376383. Foster SJ, Powell RE and Selim HS (1996) Performance of highstrength concrete corbels. ACI Structural Journal 93(5): 555 563. Hagberg T (1983) Design of concrete brackets: on the application of the truss analogy. ACI Structural Journal 80(1): 312. Hwang SJ, Lu WY and Lee HJ (2000) Shear strength prediction for reinforced concrete corbels. ACI Structural Journal 97(4): 543552. Hermansen BR and Cowan J (1974) Modied shear-friction theory for bracket design. ACI Journal, Proceedings 71(7): 5560. Jensen JF (1982) Contribution of An upper-bound rigid-plastic solution for the shear failure of concrete beams without shear reinforcement by K. O. Kemp and M. T. Al-Sa. Magazine of Concrete Research 34(119): 100103. Kriz LB and Raths CH (1965) Connections in precast concrete structures-strength of corbels. PCI Journal 10(1): 1661. Mattock AH (1976) Design proposals for reinforced concrete corbels. PCI Journal 21(3): 1842. Mattock AH, Chen KC and Soongswang K (1976) The behavior of reinforced concrete corbels. PCI Journal 21(2): 5277. Nielsen MP (1984) Limit Analysis and Concrete Plasticity. Prentice-Hall, Englewood Cliffs. Russo G, Venir R, Pauletta M and Somma G (2006) Reinforced

124

Structures and Buildings Volume 165 Issue SB3

Shear capacity of reinforced concrete corbels using mechanism analysis Yang and Ashour

concrete corbels-shear strength model and design formula. ACI Structural Journal 103(1): 310. fer K and Jennewein M (1987) Toward a Schlaich J, Scha consistent design of structural concrete. PCI Journal 32(3): 74150. Siao WB (1994) Shear strength of short reinforced concrete walls, corbels, and deep Beams. ACI Structural Journal 91(2): 123 132.

Solanki H and Sabnis GM (1987) Reinforced concrete corbels-

simplied. ACI Structural Journal 84(5): 428432.

Yong YK and Balaguru P (1994) Behavior of reinforced high-

strength-concrete corbels. Journal of Structural Engineering, ASCE 120(4): 11821201. Zhang LXB and Hsu TTC (1998) Behavior and analysis of 100 MPa concrete membrane elements. Journal of Structural Engineering, ASCE 124(1): 2434.

WHAT DO YOU THINK?

To discuss this paper, please email up to 500 words to the editor at journals@ice.org.uk. Your contribution will be forwarded to the author(s) for a reply and, if considered appropriate by the editorial panel, will be published as a discussion in a future issue of the journal. Proceedings journals rely entirely on contributions sent in by civil engineering professionals, academics and students. Papers should be 20005000 words long (brieng papers should be 10002000 words long), with adequate illustrations and references. You can submit your paper online via www.icevirtuallibrary.com/content/journals, where you will also nd detailed author guidelines. 125