A] -

Alexandria Engineering Journal (2013) 52, 163-173

FACULTY OF ENGINEERING ALEXANDRIA UNIVERSITY

Alexandria University Alexandria Engineering Journal

www.elsevier.com/locate/aej www.sciencedirect.com

ORIGINAL ARTICLE

Evaluation of tension stiffening effect on the crack width calculation of flexural RC members

Said M. Allam *, Mohie S. Shoukry, Gehad E. Rashad, Amal S. Hassan

Structural Eng. Dept., Faculty of Engineering, Alexandria University, Alexandria, Egypt

Received 10 November 2012; revised 13 December 2012; accepted 16 December 2012 Available online 31 January 2013

KEYWORDS

Tension stiffening; Crack width; Flexural members; Reinforced concrete; Codes' provisions

Abstract Building codes consider the tension stiffening when calculating the crack width of the flexural members. A simple analytical procedure is proposed for the determination of forces, stresses and strains acting on a reinforced concrete section subjected to flexure considering the concrete contribution in tension up to tensile concrete strain corresponding to the cracking strength of concrete. This analytical method gives the minimum value (lower bound) of tension stiffening. Also, a commercial Finite Element Program (ABAQUS 2007) was used to perform non-linear analysis in order to evaluate the total contribution of the tensioned concrete in carrying loads which may be considered as the upper bound of tension stiffening. In addition, a comparison is carried out among the different codes using four reinforced concrete rectangular models to compare and evaluate the tension stiffening with proposed analytical lower bound tension stiffening and upper bound as obtained by ABAQUS. The models include different percentages of flexural steel ratio. The comparison revealed that the codes' equations always consider tension stiffening lying between lower and upper bound of tension stiffening proposed in this study. Also, the study showed that the tension stiffening decreases with the increase of the percentage of the flexural reinforcement ratio.

© 2013 Faculty of Engineering, Alexandria University. Production and hosting by Elsevier B.V.

All rights reserved.

1. Introduction r i

reinforcement, whereas between cracks some amount of the

tensile force is transferred through bond to the surrounding The term ''tension stiffening'' is defined as the effect of concrete concrete, which results in a reduction in the reinforcement stres-acting in tension between cracks on the stress of steel reinforce- ses and strains, and causes the reinforcement strain at un-ment. At a crack, all the internal tensile force is carried by the cracked zone to be less than the reinforcement strain at the

cracked sections. However, the tension stiffening is reduced due to the creep effect and the cyclic loading, which induces an additional excessive slip between the steel and concrete. The most popular concept is that, by the long term loading, the tension stiffening value reduces to approximately half its initial value (ACI Committee 224R-01) [1].

The tension stiffening may be estimated as an empirical value; Welch and Janjua [2] calculated the average steel strain esm as

1110-0168 © 2013 Faculty of Engineering, Alexandria University. Production and hosting by Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.aej.2012.12.005

* Corresponding author. Tel.: +20 1112244066.

E-mail address: sa_allam@yahoo.com (S.M. Allam).

Peer review under responsibility of Faculty of Engineering, Alexandria

University.

eSm -(s - 20.7)/Es

The value of "20.7/E/' is the empirical value of tension stiffening which is approximately equal to "n fctmax/E", it was considered as the average contribution of the concrete in tension. Where "n" is the modular ratio — Es/Ec and iifctmax ' is the tensile strength of the concrete, N/mm2, and fs is the stress in the tension reinforcement assuming a cracked section, N/ mm2. In addition, the well-known crack width equation introduced by Gergely and Lutz [3] assumed a constant value of tension stiffening corresponding to a stress value of 3.5 N/ mm2 (5 ksi). This value of tension stiffening depends on the grade of concrete strength but the above assumption is convenient and practical to most reinforced concrete members. Ger-gerly and Lutz' equation assumes that the concrete tension stiffening is not affected by the steel stress when the stabilized crack pattern has been formed. This means that in the cracked stage, the tension stiffening is a constant value and is affected only by the concrete tensile strength and the bond nature between the steel and concrete at the tensioned region of the section, regardless the steel stress level.

On the other hand, many researchers (e.g. [4]) proposed empirical functions to estimate the tension stiffening. Leonhardt [4] presented a model for computing the mean strains. The average strain over the entire length; esm, is less than the bare bar strain; es which is the strain developed by the steel alone especially after cracking. The differences between esm and es is referred to as ''tension stiffening'' as shown in Fig. 1.

If the cracking strain of the concrete; ecm is ignored as being very small, em may be approximated by:

em esm es

directly after cracking [5].

Eurocode2 (2004) [6] and Egyptian Code ECP 203-2007 [7] followed Leonhardt [4] equation with different modifications parameters.

In this paper, an evaluation of the tension stiffening effect on the crack width calculation of the flexural reinforced concrete members is presented. The tension stiffening was evaluated using two different methods representing lower and upper bound. In the first method, an analytical approach

was considered to evaluate analytically the tension stiffening by considering the contribution of the concrete of the tension zone lying below the neutral axis however, the contribution of the tension between cracks was ignored. In the second method, a Finite Element Analysis using a commercial Finite Element Program (ABAQUS 2007) [8] was used to perform non-linear analysis to evaluate the total tension stiffening which may be considered as the upper bound. The lower and upper bound of tension stiffening was compared by corresponding tension stiffening terms used by different codes' equations while calculating crack width values for flexural reinforced concrete members. The comparison includes four reinforced concrete theoretical models representing different percentages of flex-ural reinforcement ratio.

2. Tension stiffening terms in some codes' equations

2.1. Eurocode2 (2004)

Eurocode2-2004 [6] gives the following equation for the mean tensile strain (esm — ecm) for calculating the crack width of a flexural member

KífC,eff(1+npeff)\\

^ Peff >> p 0.6f

where fs is the stress in the tension reinforcement assuming a cracked section, N/mm2, es — fs/As, and fscr2 is the steel stress pf —

where Kt is the factor expressing the duration of loading: Kt — 0.6 for short term loading and Kt — 0.4 for long term loading, fs is the stress in the tension reinforcement computed on the basis of a cracked section, n is the modular ratio Es,

and fcteff is the mean value of tensile strength of the concrete effective at the time when the cracks may first be expected to occur,

Aceff = effective tension area, is the area of concrete surrounding the tension reinforcement.

The term part.

fcteff(1^~nPeff) qeff

represents the tension stiffening

2.2. Egyptian code; ECP 203-2007

The Egyptian Code ECP 203-2007 [7] gives the the mean strain

£s Bare Bar Strain tension stiffening

Strain

Figure 1 Mean and bare bar steel stress-strain relationship.

- EH 1 - «=

where fs is the stress in the tension reinforcement calculated on the basis of a cracked section, N/mm2, fscr2 is the stress in the tension longitudinal reinforcement computed on the basis of a cracked section under loading conditions that cause the first crack, N/mm2, b1 is the a coefficient accounting for bar bond characteristics, and is equal to 0.8 for deformed bars and 0.5 for plain smooth bars, b2 is the a coefficient accounting for load duration; is equal to 1.0 for single short-term loading and 0.5 for sustained or cyclic loading, and Es is the Modulus of elasticity of the reinforcement, N/mm2. (biftff)2)

The term ^-e- represents the tension stiffening part.

2.3. British standards BS 8110-1997

According to BS 8110-1997 [9], the average strain at the level where the cracking is being considered is given as:

b(h — x)(a' — x) 3EsAs(d — x)

where 8! is the strain at the level considered, calculated ignoring the stiffening effect of the concrete in the tension zone, b is the width of the section at the centroid of the tension steel, and a' is the distance from the compression face to the point at which the crack width is being calculated.

Also, according to BS 8110-1997 [9] in assessing the strains, the modulus of elasticity of the concrete should be taken as half the instantaneous values.

The term

b(h — x)(a' — x) 3EsAs(d — x)

represents the tension stiffening

3. Analytical method for the calculations of stresses and strains of cracked sections (lower bound of tension stiffening)

The tensile strain of concrete is usually neglected for the calculation of stresses in a reinforced concrete beam, even though concrete continues to carry tensile stress between the cracks. Moreover, the concrete contributes in carrying tension even at the crack position because the crack does not extend to the neutral axis under service loading. The contribution of concrete at crack position is considered as the lower bound of the concrete contribution in carrying tensile stresses. This contribution of the concrete in tension affects the member's stiffness after cracking and hence the width of the cracks under service loads.

A simple procedure is proposed for the determination of forces, stresses and strains acting on a concrete section prior and after first crack occurrence for a reinforced concrete member subjected to bending. In addition, the crack length can be determined at different values of strain. In this approach, the steel stresses is calculated without neglecting the concrete in tension after the formation of first crack, and considering the non-linear nature of the stress-strain relationship of concrete in compression. Strains obtained by the analysis are calculated using strain compatibility, while forces are calculated using equilibrium conditions.

3.1. Assumptions considered in the analytical method

The following assumptions are made for the present procedures:

i. Plane sections before bending remain plane after bending.

ii. Steel and concrete are assumed to be homogeneous and isotropic.

iii. Concrete stresses due to shrinkage and temperature changes are negligible.

iv. The tensile strength of the concrete is considered.

v. The stress-strain curve for concrete is known.

Figs. 2 and 3 show the idealized stress-strain relationships for steel reinforcement and concrete; respectively. For concrete in compression, the non-linear stress-strain relationship is

Stress fy

Strain

Strain

Stress

Figure 2 Idealized steel stress-strain relationship.

Stress

Equation ,fc= fa_ Ec(2Eci-Ec) of parabola £ci2

Strain

Figure 3 Idealized stress-strain relationship for concrete.

considered as a parabolic up to strain ec1 = 0.002 and a linear from ec1 = 0.002 to ec2 = 0.003 however fc1 = 0.67 fcu as given by the Egyptian Code ECP 203-2007.

3.2. Forces acting on an uncracked section

For the analysis of uncracked section subjected to flexure, the strain and stress distributions at the section are as shown in Fig. 4, where the tensile concrete stress at bottom fiber is less than that required to form a crack (i.e. the concrete stress fct at the tensioned face is less than the concrete tensile strength;

fctmax).

For any given value of concrete strain at the extreme compression fiber; ecc, the compressive steel strain esc, the tensile steel strain est, and the concrete strain at the extreme tensioned fiber ect could be simply calculated as follows:

a — W

est — ---ecc (7)

1 + w-

where a is an unknown factor representing the ratio of neutral axis depth and the section effective depth and W is a known variable defined by the geometric properties of the section; dc = W d, and consequently h = (1 + W) d.

Subsequently, the reinforcement stresses are determined based on the stress-strain curve of the steel, shown in Fig. 2, while the concrete stresses are determined based on the stress-strain relationship cited by Park and Paulay [10], Fig. 3. The equations expressing stresses across the section are functioned in ecc and a, theses equations may be written as follows:

Figure 4 Strain and stress distribution for an uncracked section.

fsc - esc * Es

fst - est * Es

fct sct * fctmax =sctmax

(9) (10) (11)

where ectmax is the concrete tensile strain corresponds to the tensile concrete strength.

The internal forces acting on the section are determined from the stress values obtained above. The compression force carried by concrete is determined as the integration of the stress-strain relationship. By this step, the equations representing internal forces across section are functions in ecc and a and may be written as follows:

C _fc1

Cc - 2

Cs - l'bdfsc Ts - ibdfst

si Sc1

ad — b

Tc - 0.5fct(1 + W - a)bd The equilibrium equation is:

Cc ^ Cs Tc ^ Ts

Applying the equilibrium equation, for any given value of concrete strain at the extreme compression fiber; ecc, the depth of neutral axis (ad) is calculated and so, all the section strains and forces are determined. The obtained concrete tensile stress is checked to be less than the maximum tensile strength of the concrete in order to apply the same procedure on the next value of ecc.

To cover the entire loading range, this procedure is repeated for values of ecc ranging from zero to a value of concrete compressive strain equals to ec1, until concrete tensile stress reaches the maximum concrete tensile strength. If the reinforcement stress reaches the steel yield stress, the analysis is terminated. If the concrete tensile stress reaches the maximum concrete tensile strength, the cracked section analysis is applied as given in the following section.

3.3. Forces acting on a cracked section

For the analysis of a cracked section subjected to flexure, the strain and stress distributions at the section are shown in Fig. 5, where the tensile concrete stress at bottom fiber is larger than that required to form a flexural crack (i.e. the concrete stress fct at the tensioned face P fctmax).

The analysis of cracked section follows the same steps applied for the uncracked section analysis except that there is a section below the neutral axis has a value of concrete tensile strain equals to the concrete tensile strength. Below this section, concrete is neglected in tension while, above it, concrete contributes in tension. The distance measured from that section to the extreme tensioned fiber is assumed to be the crack length. With increasing the value of ecc, this section is moving upward and thus length of the crack increases and the tensile force carried by concrete decreases.

In addition to the variables a and W, the position of the section below the neutral axis at which the crack ends, is expressed as a function of a*, where a* is the ratio of the distance of that section from the neutral axis to the section effective depth; d.

Based on Eqs. (6) and (7), at somewhere below the neutral axis,

ect ecc

For a value of ect (Eq. (17)) equals to the maximum tensile concrete strain; ectmax the relation between a, and a* is known. Thus, the above equations are functions in only two unknowns: ecc, and a.

The concrete and steel stresses are determined from the stress-strain relationships using Eqs. (9) and (10) where:

fct - fct

Next, the internal forces acting on the section are determined using Eqs. (12)-(14) however, the tensile force in concrete can be written as:

strain distribution diagram

stress distribution diagram

Figure 5 Strain and stress distribution for a cracked section.

Tc = 0.5 fctmaxa* bd (19)

Finally, by applying the equilibrium equation (Eq. (16)) for any given value of concrete strain at the extreme compression fiber; ecc, the depth of neutral axis; ad is known and so all the section strains, stresses and forces are determined. To cover the entire loading range, this procedure is repeated for different values of ecc until the concrete crushes or the reinforcement yields.

With the calculation of the concrete and steel stresses along the section and the determination of the extended crack length under different loading stages, the flexural capacity of the section could be calculated easily. Although this analysis takes into account the tensile force carried by concrete after cracking, it should be noted that the analysis gives the lower bound of the concrete contribution since it does not consider the un-cracked concrete between cracks which confines the steel elongation.

4. Finite element analysis

Numerical non-linear analysis, using the Finite Element Method (FEM), was carried out to investigate the concrete contribution in tension after flexural cracking of reinforced concrete beams. The objective was to compare the FEM results with those obtained from lower bound analytical method and

the equations proposed by Egyptian Code 203-2007 [7], BS 8110-1997 [9] and Eurocode2-2004 [6].

A commercial FEM software; Program ABAQUS version 6.7 - 2007 [8], was used for the present analysis. Concrete was modeled using 3-dimensional, 8-node solid elements; C3D8, with three degrees of freedom for each node; translations u, v, and w in the three orthogonal directions; x, y and z, respectively. Steel reinforcement was modeled as discrete 3-D, 2-node truss elements with three degree of freedom per each node: u, v and w. Perfect bond was assumed between concrete and steel. Smeared cracking approach was chosen to represent the discontinuous macro-cracking brittle behavior of concrete. Fully integration scheme was chosen to integrate the element's internal forces and stiffness. Incremental/iterative procedure was carried out with the use of modified New-ton-Raphson Method for the non-linear analysis.

4.1. Material modeling

The material modeling was as follows:

1. Steel reinforcement was modeled as linear elastic-perfectly plastic material, as shown in Fig. 2. The input data includes: the yield stress; fy, modulus of elasticity; Es and Poisson's ratio; us (us = 0.3).

Concrete Tensile Stress

Concrete Tensile Stress

__—,__

Ecr 3£cr 5.875 Ecr 10.5 Ed £ma*ot= 16 Ecr Concrete Tensile Strain

Figure 7 Modified tension softening curve.

Reinforcing bar —

ySectior^

860 mm

910 mm

Ec - 300 KN/mm1 feu = 40 N/mm2 Sc2 = 0.0035 £cl= 0.00222

Figure 8 Dimensions and loading for the beam tested by Managat and Elgarf [12].

2. Concrete in tension was modeled as linear elastic brittle material with strain softening. Tension stiffening is allowed by modifying the concrete softening behavior. Post-cracking stress-strain relationship was as suggested by Massi-cotte et al. [11] and is shown in Fig. 6. This relationship assumed that the strain softening after cracking reduces the stress to zero at a total strain of about 16 times the strain at first cracking. The curve, shown in Fig. 6, was softened to permit a relatively gradual response behavior and consequently to decrease the convergence problems, as shown in Fig. 7. The input data includes: the concrete tensile strength; fctmax, the strain at first crack; ecr and the strain softening curve.

3. Concrete in compression was modeled as elastic-plastic model. The input data includes: the concrete compressive strength; fcUi modulus of elasticity; Ec, Poisson's ratio; oc (tc = 0.2), stress-plastic strain relationship, and the following failure ratios:

• The ratio of the ultimate biaxial compressive stress to the ultimate uniaxial compressive stress; this ratio was taken as 1.16.

• The absolute value of the ratio of the uniaxial tensile stress at failure to the ultimate uniaxial compressive stress; this ratio was taken as 0.109.

• The ratio of the magnitude of a principal component of plastic strain at ultimate stress in biaxial compression to the plastic strain at ultimate stress in uniaxial compression; this ratio was taken as 1.25.

• The ratio of the tensile principal stress at cracking, in plane stress, when the other principal stress is at the ultimate compressive value, to the tensile cracking stress under uniaxial tension; this ratio was taken as 0.2.

4.2. Verification of the FEM Modeling

First, the FEM modeling was verified against the results of one of the control beams tested by Managat and Elgarf [12]. They carried out an experimental study on 111 under-reinforced beams to determine their residual flexural capacity after undergoing different degrees of reinforcement corrosion. Their study included experimental testing of control beams that can be used for calibration of the FEM model. The under-reinforced concrete beam specimens were 910 mm long and had a rectangular cross section of 150 mm depth, 100 mm width. Fig. 8 shows the beam dimensions and its reinforcement. The beam was tested under four-point bending. The FEM analysis was

carried out assigning the same material properties and dimensions of the tested beam, as shown in Fig. 8.

Fig. 9 shows the load-deflection relationship obtained experimentally together with that obtained from the FEM analysis for the beam. The figure indicates that the finite element model matches well with the experimental results.

5. Investigation of tension stiffening

A group of four reinforced concrete beams was investigated using ABAQUS software. The beam properties and parameters examined are given in Table 1. Each beam model has a cross section of 150 mm width and 550 mm height with a span of 5.5 m, as shown in Fig. 10. The variable studied was the diameter of the reinforcing steel bar, consequently the reinforcement ratio, which ranged from 0.54% to 1.31%. The beam was subjected to two point loads applied at each third of the span. Due to symmetry of the beam dimensions and loading, only one half of the beam was analyzed.

The material properties of the models were as follows:

• fy = 360 N/mm2

• Es = 200 kN/mm2

• fctmax = 3.28 N/mm2

• ecr = 0.0001125 (concrete strain at cracking)

• fcu = 30 N/mm2

• Ec = 29.15 kN/mm2

Figs. 11-14 show the steel stress-strain relationships for the beam models described in Table 1 as obtained from the following:

80 70 60

-Expe Omental esults -

F.E. results

012345678 Deflection (mm)

Figure 9 Load-deflection curve of the RC beam tested by Managat and Elgarf [12] with F.E. results [12].

Table 1 Properties of the studied model sections.

Model Dimensions Reinforcement steel used Concrete strength, Steel yield stress, Reinforcement

number and Bar diameter, mm N/mm2 N/mm2 ratio; 1, %

Width; b, Height; h, Effective

mm mm depth; d, mm

A1 150 550 500 2016 30 360 0.54

A2 150 550 500 2018 30 360 0.68

A3 150 550 500 2022 30 360 1.01

A4 150 550 500 2025 30 360 1.31

Longitudinal Section

150 mm

ni » '

lâ'W-Section A-A

Figure 10 Beam models analyzed by FEM.

- The FEM obtained from ABAQUS.

- The analytical method.

- The bare bar.

- Eq. (4) of the Egyptian Code ECP 203-2007.

- Eq. (5) of British Standards BS 8110-1997 and

- Eq. (3) of Eurocode2-2004. It should be noted that the results obtained by Eurocode2-2004 consider the limitation of the value of em not to be less than 0.6fs/Es.

Figs. 11-14 indicate that the analytical method represents a lower bound of the tension stiffening or a minimum tension stiffening (min TS), however the results obtained by the FE analysis using ABAQUS represent the upper bound for tension stiffening or full tension stiffening (Full TS). The results of FE analysis depend greatly on the shape of the tension softening curve considered in the analysis, as shown in Figs. 6 and 7. The tension-softening relationship depends on the nature of bond between concrete and the reinforcement bars at the crack. In reality, the magnitude of the bond is affected by steel stress, concrete cover, bar spacing, transverse reinforcement (stirrups), lateral pressure, and size of bar deformations. Therefore, the adopted tension softening curve is considered merely as a reference and indicator for the tension stiffening

value. The average strain obtained from different codes are lying between lower and upper tension stiffening represented by analytical method and the Finite Element model obtained from (ABAQUS). The values of mean steel strain as proposed by the ECP 203-2007 (Eq. (4)) and BS 8110-1997 (Eq. (5)) are conservative when compared with those obtained by Euro-code2-2004 (Eq. (3)). As shown in the figures, the values of the mean steel strain as obtained from Eurocode2-2004 are close to the FE results for high reinforcing ratios especially at high level of loading (i.e. with the increase of the steel stress). It should be noted that when using a different tension-softening curve, Eurocode2-2004 results would be considered critical or overestimating the value of tension stiffening especially in sections with relatively high reinforcing ratio.

Table 2 gives the average strain and the tension stiffening obtained from FEM, analytical method and different codes corresponding to steel stress level of 200 N/mm2. The values of tension stiffening were obtained as the difference between steel strain of bare bar and steel strain corresponding to each case. The analytical method represents the concrete contribution in carrying tension at the crack position because the crack does not extend to the neutral axis under service loading. Although the analytical method takes into account the tensile

300 250

â 100 <U

0 0.0002 0.0004 0.0006 0.0008 0.001 0.0012 0.0014 0.0016

Strain

Figure 11 Steel stress versus strain for beam model A1.

<u 100

--Analytical min.T.S.

- - - "EGYPTIAN CODE' - - - EURO CODE

------BS

-Bare bar

-FEM - ABAQUS

0 0.0002 0.0004 0.0006 0.0008 0.001 0.0012 0.0014 0.0016

Strain

Figure 12 Steel stress versus strain for beam model A2.

E 200 E

/s ■U'

s » - / X — — Analytical min.T.S. — - - "EGYPTIAN CODE" — - - - EURO CODE ------BS -Bare bar

/ s^ -K y /

/ FEM - ABAQUS

0 0.0002 0.0004 0.0006 0.0008 0.001 0.0012 0.0014 0.0016

Strain

Figure 13 Steel stress versus strain for beam model A3.

(A M a

K 100-

' ' s; .'•A/

I,**'/

.-'J- y /

— — Analytical min.T.S. --- "EGYPTIAN CODE"

/ .... .. .. EURO CODE BS Bare bar FEM - ABAQUS

0.0002 0.0004 0.0006 0.0008 0.001 0.0012 0.0014 0.0016 Strain

Figure 14 Steel stress versus strain for beam model A4.

Table 2 Average steel strain and tension stiffening values predicted by codes' equations and both analytical and finite element methods at steel stress level = 200 N/mm2.

Model Average strain, es x10~4 Tension stiffening, x10~4

FEM ECP-2007 Eurocode2-2004 BS 8110 Analytical Bare bar FEM ABAQUS ECP-2007 Eurocode2-2004 BS 8110 Analytical

ABAQUS Full TS min TS

A1 1.57 5.00 6.00 7.25 7.76 10 8.43 5.00 4.00 2.75 2.24

A2 3.47 6.59 6.00 7.90 8.00 10 6.53 3.41 4.00 2.10 2.00

A3 5.93 8.45 6.94 8.69 9.10 10 4.07 1.55 3.06 1.31 0.90

A4 6.78 9.00 7.49 9.03 9.32 10 3.22 1.00 2.51 0.97 0.680

force carried by concrete after cracking, it should be noted that the analysis gives the minimum value of the tension stiffening since it considers the contribution of the uncracked concrete under the neutral axis at the crack position only without considering the contribution of the uncracked concrete between cracks. The full tension stiffening or the upper bound of tension stiffening is obtained from Finite Element model since it considers the concrete contribution in carrying tension between cracks and at the crack position. Therefore, as given in Table 2, all codes equations give tension stiffening values between lower and upper tension stiffening values. It is clear from the table that the tension stiffening values obtained from the Eurocode2-2004 are always higher than those of other codes for high ratios of steel reinforcement, however the limitation of the value of em not to be less than 0.6fs/Es limits the tension stiffening for lower steel ratios. On the other hand, BS 8110-1997 shows the least tension stiffening values among all the presented codes' equations for all ratios of steel reinforcement.

Figs. 15 and 16 show the analytical (lower bound) and the FEM (upper bound) of steel stress-strain relationships for different flexural reinforcement ratios represented by models A1, A2, A3 and A4. These figures along with Figs. 11-14 and Table 2 reflect the effect of flexural reinforcement ratio on the tension stiffening. It is clear that as the percentage of the flex-ural reinforcement increased, the tension stiffening values decreased. The average strain from different codes follows the

same trend, i.e. decreasing tension stiffening with the increase of flexural reinforcement ratio. For model A4 with i = 1.31% (Fig. 14), with the largest reinforcement ratio, the lowest tension stiffening values were obtained. Generally, the values of the steel strain as obtained from the Finite Element analysis were smaller than those obtained from the bare bar analysis by 73-32%. The highest reduction in strain was obtained at low level of loading just after first cracking while the lowest value was obtained at the service stress level of 200 N/mm2 indicating a reduction in concrete contribution in tension. When compared to the values obtained using ECP (2007), the above ratios were 54-10%, which indicate less concrete contribution compared to the Finite Element analysis. Also, using Euro-code2-2004, the above ratios were 40-25%, which indicate overestimating the value of the concrete contribution at service loading. Furthermore, using BS 8110-1997, the above ratios were 21-10%, which show conservative estimating of tension stiffening values. Table 2 also displays the effect of the flexural reinforcement ratio on the tension stiffening at steel stress level of 200 N/mm2. The percentages of tension stiffening obtained by ECP equation to full tension stiffening obtained by FEM at service load varied from 59% to 31%. Also, those percentages obtained by Eurocode2 varied from 78% to 47%. However, the percentages of tension stiffening obtained by BS varied from 33% to 30%. The variation in the above percentages confirm that the increase in the flexural reinforcement ratios decreases the tension stiffening values. On the other hand, the

E 200 E

a> 100

- - - - A3

- Bare bar

0 0.0002 0.0004 0.0006 0.0008 0.001 0.0012 0.0014 0.0016

Strain

Figure 15 Analytical steel stress-strain for different models.

Strain

Figure 16 FEM steel stress-strain for different models.

most conservative percentages were obtained by the BS equation which indicate less concrete contribution compared to the other Codes' equations.

6. Conclusions

From the proposed analytical method representing lower bound of tension stiffening and the Finite Element model obtained from ABAQUS representing the upper bound of tension stiffening, the tension stiffening part of Codes' equations of Eurocode-2-2004, BS 8110-1997 and Egyptian Code 203-2007 for different beam models with different flex-ural steel ratios were evaluated and the following conclusions were drawn:

1. The lower bound of tension stiffening can be easily obtained by a simple analytical method which takes into account the tensile force carried by concrete after cracking. This analytical method gives the lower bound of tension stiffening since it only considers the contribution of the uncracked concrete under the neutral axis at the crack position without considering the contribution of the uncracked concrete between cracks.

2. The Finite Element analysis gives the upper bound of tension stiffening since it considers the concrete contribution in carrying tension between cracks, however the input tension-softening relationship is an important factor to obtain upper tension stiffening values. The tension-softening relationship depends on the nature of bond between concrete and the reinforcement bars at the crack.

3. Building codes consider the tension stiffening when calculating the crack width of the flexural members. The tension stiffening values considered by Codes lie between lower and upper bound of tension stiffening proposed in this study.

4. Eurocode2-2004 overestimated the value of tension stiffening especially in sections with relatively high reinforcing ratios, however, BS8110-1997 code gives underestimated values. However, ECP 203-2007 code gives tension stiffening values lie between BS and Eurocode2 except for sections with low ratios of flexural reinforcement.

5. Generally, the values of the steel strain as obtained from the Finite Element analysis and different Codes' equations are smaller than those obtained from the bare bar analysis. The highest reduction in strain is obtained at low level of loading just after first cracking while the lowest value is obtained at the service stress level indicating a reduction in concrete contribution in tension.

6. The tension stiffening is strongly affected by the percentage of flexural reinforcement ratio. As the percentage of the flexural reinforcement increased, the tension stiffening values decreased. The most conservative values are obtained by the BS equation which indicate less concrete contribution compared to the other Codes' equations.

References

[1] ACI Committee 224, Control of Cracking in Concrete Structures, ACI Report 224R-01, American Concrete Institute, Farmington Hills, MI, 2001, pp. 46.

[2] G.B. Welch, M.A. Janjua, Width and Spacing of Tensile Cracks in Reinforced Concrete, UNICIV Report No R76, University of NSW, Kensington, 1971 (Cited by Warner and Rangan (1977)).

[3] P. Gergely, L.A. Lutz, Maximum Crack Width in RC Flexural Members, Causes, Mechanism and Control of Cracking in Concrete, SP-20, American Concrete Institute, Detroit, 1968, pp. 87-117.

[4] F. Leonhardt, Crack Control in Concrete Structures, IABSE Surveys, No.S4/77, International Association for Bridges and Structural Engineering, Zurich, 1977, pp. 26 (Cited by Rizkalla (1984)).

[5] A.S. Hassan, Crack Control for Reinforced Concrete Members Subjected to Flexure, M.Sc. Thesis, Alexandria University, Alexandria, Egypt, 2008.

[6] Eurocode 2: Design of Concrete Structures - Part 1: General Rules and Rules for Buildings; The European Standard EN1992-1-1, 2004.

[7] The Egyptian Code for Design and Construction of Reinforced Concrete Structures, ECP 203-2007, Ministry of Housing, Egypt.

[8] ABAQUS, version 6.7, ABAQUS, Inc., DASSAULT Systems, USA, 2007.

[9] Structural Use of Concrete, Part 2, Code of Practice for Special Circumstances BS 8110: Part 2: 1997, British Standard Institution, London, 1998.

[10] R. Park, T. Paulay, Reinforced Concrete Structures, Wiley-Inter-Science, New York, 1975, pp. 769.

[11] B. Massicotte, A.E. Elwi, J.G. MacGregor, Tension stiffening model for planar reinforced concrete members, ASCE Journal of Structural Engineering 116 (11) (1990) 3039-3058.

[12] P.S. Managat, M.S. Elgarf, Flexural strength of concrete beams with corroding reinforcement, ACI Structural Journal 96 (1) (1999) 149-159.