THE INFUENCE OF THE EFFECTIVE MOMENT OF

INERTIA ON THE DEFLECTION OF FRP

REINFORCED CONCRETE MEMBERS

Enio Deneko1, Anjeza Gjini2

1
Polytenchnic University of Tirana, Faculty of Civil Engineering, Tirana, Albania
[email protected]
2 Polytenchnic University of Tirana, Faculty of Civil Engineering, Tirana, Albania
[email protected]

Abstract. Fiber reinforced polymer (FRP) bars, beacause of their non-corrosive

properties, emerged as an alternative to steel bars in RC structures in agressive
environments. Mechanical characteristics of FRP materials, such as lower
elastic modulus, lower ratio between Young’s modulus and the tensile strength,
lower bond strength of FRP bars and concrete, compared to steel reinforcement,
can lead to large deflections and crack width, so this make that serviceability
design determine the design of FRP reinforced concrete, based on the
serviceability requirements. Different parameters can influence on the stresses
in materials, maximum crack width and the allowed deflections, but one of the
most important especially for the deflection design, is the effective moment of
inertia. In this paper, reference is made ACI, EC2 and other models, regarding
deflections of FRP reinforcement concrete beams comparing with each other
and also with steel reinforcement concrete formulas. Concrete beams reinforced
with glass-fiber (GFRP) bars, exhibit large deflections compared to steel
reinforced concrete beams, because of low GFRP bars elasticity modulus. For
this purpose we have used equations to estimate the effective moment of inertia
of FRP-reinforced concrete beams, based on the genetic algorithm, known as
the Branson’s equation. In the last two decades, a number of researchers
adjusted the Branson’s equation based on a lot of test results. In the paper, is
elaborated a numerical example to check the deflection of a FRP-RC based on
these equations of different code provisions and models, used for predicting the
deflection of FRP-reinforced concrete beams.

Keywords: FRP bars, reinforced concrete beams, serviceability, deflection,

effective moment of inertia, modulus of elasticity.

1 Introduction

FRP bars can be effectively used in corrosive environments because of their corrosion
resistant property. The problems seem similar to steel RC, but solutions, limits and
analytic models are different, because of the very large band of FRP bars on the
market, with a large variety of mechanical characteristics. There are many types of
fibers including glass (GFRP), carbon (CFRP) and aramid (AFRP), with different
grades of tensile strength and modulus of elasticity. The behavior of FRP-RC beams
differs from steel RC beams, because FRP bars display a linear elastic behavior up to
the point of failure and do not demonstrate ductility. Also the bond strength of FRP
bars and concrete is lower than that of steel bars, leading to an increase in the depth of
cracking, a decrease of stiffening effect, and so an increase of the deflection of FRP-
RC beams for an equivalent cross-section of reinforcement of steel reinforced
concrete beams.
GFRP-RC beams have lower elastic modulus quite 20-25% of that of steel bars.
Because of it, the deflection criterion may control the design of long FRP-RC beams.
Consequently a method is needed in order to know the expected service load
deflections with a high degree of accuracy. Only some countries have developed a
code in FRP-RC design that is still in preparation phase, so is very difficult to operate
in this new field.

2 Methodology

FRP are anisotropic materials and are characterized by high tensile strength with no
yielding in the direction of the reinforcing fibers. An FRP-RC member is designed on
its required strength and then checked for serviceability and ultimate state criteria
(e.g., deflection, crack width, fatigue and creep rapture). In most instances,
serviceability criteria will control the design so safety checks for FRP-RC are more
important than that for steel-RC.
A well-known deflection model for steel-RC is proposed by ACI and EC2, later
modified for FRP-RC. This deflection model simulates the real behavior of the
structure by taking into account cracking, but not the tension-stiffening effect of
concrete.
ACI 318-95 is based on Branson [1968-1977] formulae of the effective moment of
inertia for Steel-RC:
(1)
where, Ie is the effective moment of inertia, Mcr is the cracked section moment, Ig is
the total moment of inertia, Ma is the maximum moment in member at the deflection
stage, Icr is the cracked section moment of inertia.
ACI 440R-96 proposed new formulae of the effective moment of inertia for FRP-RC:
(2)
where, βd is a reduction coefficient related to the reduced tension-stiffening exhibited
by FRP-RC members. Based on evaluation of experimental results
. (3)
This equation is valid only if Ma ≥ Mcr. If Ma ≤ Mcr, than Ie = Ig and if Ma ≈ Mcr, or
slightly less, than Ie = Icr, because shrinkage and temperature may cause section
cracking.
Some other models based on ACI give this formula:
 (4)
where, α and β are coefficients of bond properties of FRP and
(5)
where, is a coefficient that increases with the bond quality. When no experimental
results are given than = 0, 5 and α = 1.
EuroCode 2 propose a simplified model for Steel-RC
· + · where . (6)
Here, v1 and v2 are calculated taking into account that the moment of inertia of the
section is constant and respectively I1 and I2 (or Ig and Icr). For steel bars the
coefficients β and m, are to be taken as given in the table.

Table 1. Coefficient β and m for steel bars

Model β m
EC2 1,0 2,0
CEB 0,8 1,0

For FRP bars, these coefficients must be reconsidered through experimental tests.
EC2 hasn’t proposed yet a model for FRP-RC deflection design, but the Italian Code
CNR-DT 203/2006, based on experimental tests, shows that the model for EC2 can be
suitable for FRP-RC too. Therefore the EC2 equations to compute deflection “f” must
be reconsidered:
(7)
f1 = deflection of uncracked section
f2 = deflection of cracked section
β1 = 0,5 coefficient of bond properties of FRP bars
β2=coefficient of the duration of loading: β2=1 for short-time loads and β2=0,5 for
long-time cycling loads.
Mmax = the maximum moment acting on the examined element
Mcr = the cracking moment calculated at the same cross section of Mmax
m=2
In the last two decades a number of researchers tried to adjust the Branson’s equation,
comparing to experimental results of FRP-RC member tests. The experimental results
show that Branson’s equation overestimated the moment of inertia Ie and
underestimates the deflection, because the Branson’s equation was calibrated for RC
beams where Ig/Icr ≤ 3 [Bischoff 2005], but not for most members that have
[Bischoff 2009]. Also the bond behavior of FRP bars and concrete
differs from the bond behavior of steel and concrete.
Benmokrane [1996], added two reduction factors and adjusted this equation:

(8)
From experimental data α=0.84 and β=7, because of the nature of FRP reinforcement,
with larger deflection and greater reduction of compressed concrete section when
applied Mcr.
Faza and Gangarao [1992], proposed a model for two concentrated point loads based
on the assumption that a concrete section between the point loads is fully cracked,
while the end sections are partially cracked.
where (9)
The maximum deflection is calculated as follows:
. (10)
Toutanji and Saafi [2000] adjusted the ratio Mcr/Ma to take into account the modulus
of elasticity of FRP bars (Efrp) and the reinforcement ratio (ρf). They took a set of 13
GFRP-RC beams with a ratio . The model proposed was:

. (11)

where: if and if
Brown and Bartholomew [1996], used the same model based on tests of two GFRP-
RC beams with the ratio and while Al-Sayed [2000] proposed

ACI 440.1 R-01 adopted the modification proposed by GAO [1998]:

. (12)
where, βd = 0,6 based on Masmoudi (1998) and Theriault & Benmokrane [1998]
studies. They recommended,
(13)
αb is a bond dependent coefficient: αd =0.5 for GFRP [GAO 1998] and later based on
experimental tests of 48 GFRP-RC and the amount of the longitudinal reinforcement:
. (14)
Recently, Rafi and Nadjai [2009], introduced γ factor, that reduces the portion of
cracked moment of inertia:
. (15)

. (16)

. (17)
ISIS Canada [2001], based on Ghali and Azarnejad [1999], when service load level
is less that cracked moment, Mcr, the immediate deflection can be evaluated using the
transformed moment of inertia, It, instead of effective moment of inertia, Ie , used
when service moments exceed the cracked moment.
Mota [2006] examined a number of the suggested formulations for Ie and found an
equation that provided the most conservative results over the entire range of
experimental results of test specimens.
 . (18)

Where, It is the moment of inertia of a non-cracked concret section, and

. (19)

. (20)

. (22)
“b” represents the width of cross-section (mm) and “d” the depth to FRP layer (mm).
CAN/CSA-S806 [2002] used Razqapur methodology which assumes that tension
stiffening is insignificant in cracked regions on FRP-RC beams, using Ec·Ig when Ma
< Mcr, and Ec·Icr when Ma > Mcr, to integrate the curvature M/EI along the beam span.
This leads to a simple expression for beam deflection δmax, for a four-point bending
configuration with two point loads at a distance a from the supports, assuming Lg, the
distance that the beam is uncracked:
. (23)
Bischoff [2005], Bischoff [2007a], Bischoff and Scanlon [2007], proposed an
equation derived from integration of curvatures along the beam taking into account
the tension-stiffening effect:
. (24)

Abdalla, El-Badry and Rizkalla introduced a model similar to EC2-CEB, suggesting

α=0.85 and β=0.5
. (25)
But Abdalla [2002] gave also a model based on ACI:
. (26)

where . The coefficient of 1.15 (or better 1/0.85), takes into account the
reduction of tension-stiffening effect in the fully cracked FRP concrete section.

Table 2. Different design models

Reference Models for

ACI 318 R -95 (1995)

Branson formulae

Benmokrane (1996)
 ACI 440.1R -03 (2003)
, = 0,5 and α = 1

Yost (2003) based on ACI

ACI 440.1R -06 (2006)

Rafi & Nadjai (2009)

(GPa)

EC2-CEB, Italian Code

CNR-DT 203/2006

Toutanji & Saafi (2000) if so m = 5.98

and if

Alsayed Model A (2000)

for
Alsayed Model B (2000)
for 1<

Bischoff (2005,2007)

& Scanlon (2007)

Abdalla based on ACI

(2002)

Abdalla, Rizkalla & El α = 0,85 dhe β = 0,5

Badry (EC2)
, ,

ISIS Canada (2001) &

Mota (2006)
 Hall & Ghali (2000)

3 The Numerical Example

We have made calculations for a simply supported, normal weight interior beam with
a span length and . It is designed to carry a service live load of
and a superimposed service dead load of . The cross
section of the beam is to be taken as 250 mm x 400 mm. 4 Ø 16 GFRP bars are
selected as main beam reinforcement and Ø 9.5 GFRP bars are selected as shear beam
reinforcement. Material properties of GFRP bars are: tensile strength = 320 MPa,
rupture strain =0.014 and Modulus of elasticity Ef = 44 800 MPa.

4 The Result Comparison

All the analysis results taken from different methods are put in Table 3.
Reference
ACI 318R-95 (1995), Branson
Benmokrane (1996)
ACI 440.1R-03 (2003)
Yost (2003) based on ACI
ACI 440.1R -06 (2006)
Rafi & Nadjai (2009)
EC2-CEB, Italian Code CNR-DT 203/2006
Toutanji & Saafi (2000)
Alsayed Model A (2000)
Alsayed Model B (2000)
Bischoff (2005,2007) & Scanlon (2007)
Abdalla based on ACI (2002)
Abdalla, Rizkalla & El Badry (EC2)
ISIS Canada (2001) & Mota (2006)
Hall & Ghali (2000)
Is calculated the mid span displacements of a GFRP reinforced concrete beam, based
on different equations for effective moments of inertia , using different levels of
loading (only different service dead loads, while the service live load is maintained
constant), in normal reinforced ratio ( ) and in high reinforced ratio
( ). The results of the most used methods are included in comparative charts
in order to find out some theoretical conclusions.

DEFLECTION OF FRP REINFORCED ACI 2003

CONCRETE BEAM
20 ACI 2006
Mid span displacements

15 Yost

10 Rafi &
(mm)

Nadjai
EC2-CEB
5
ISIS
0
0 5 10 15 20 Branson
Service dead load (kN)
Fig. 1. Relation between service dead load and mid span displacement for

DEFLECTION OF FRP REINFORCED ACI 2003

20 CONCRETE BEAM
Mid span displacements (mm)

ACI 2006
15 Yost

10 Rafi &
Nadjai
EC2-CEB
5
ISIS
0 Branson
0 10 20 30 40
Service dead load (kN)
Fig. 2. Relation between service dead load and mid span displacement for

5 Summary and conclusions

Based on the numerical examples we used and the charts above, we can see that the
most reliable methods, those with the most consistent theoretical results, are: ACI
440.1R-06 (2006), Yost (2003), Rafi & Nadjai (2009) based on ACI and ISIS Canada
(2001& Canadian Code (CAN/CSA-S806 2002) a similar method, but at the same
time some different to ACI Code. We can see also that the original Branson’s
equation is just an envelope for the other methods, for short beams we get smaller
deflections than other methods and for longer ones we get bigger displacement than
other methods. This because Branson’s equation doesn’t takes into account the type of
FRP used as reinforcement, so it hasn’t used any reduction factor based on the
ratios, or even on the quantity of the reinforcement used so the ratios, as other
methods do. So this equation is more generalized and conservative. Some other
methods use only one of these coefficients so they are conservative too.
Based on a lot of experimental results, the reinforcement ratio and the elastic modulus
of FRP bars are the most significant variables to calculate deflections, especially the
ratio, because the ratio doesn’t have a significant effect. Yost [2003] based on
ACI, Rafi & Nadjai [2009] and also ACI 440.1R -06 [2006] use both reduction
factors taking into account these variables and give good results, so this three
methods, based on ACI code, gives more accurate results for beams with low and
high reinforcement ratios. This is also shown in the charts where ACI 440.1R -06
[2006] preserve the same load ratios for the same deflection for and

, (see the load ratio for the same deflection ≈ ). At

Yost and Rafi & Nadjai, this ratio is 2-2.5.
Also, Hall& Gali, ISIS Canada [2001] and the Canadian Code [CAN/CSA-S806
2002], using different equation instead of Branson’s one, where is introduced It, the
moment of inertia of a non-cracked section transformed to concrete taking into
account not only the reinforcement and modulus of elasticity ratios (but not in
form), but also a coefficient characterizing the bond properties of reinforcement bars,
gives accurate and similar results as the others above.
EC2-CEB, Italian Code CNR-DT 203/2006 gives good results that don’t depend on
the effective moment of inertia, but only on displacements of non-cracked cross
concrete section and the cracked section. It takes into account the bond properties of
reinforcement and the type of load: short time loads or long-time cycling loads, but
doesn’t depend on reinforcement and modulus of elasticity ratios.
As shown in the chart 2, if reinforcement ratios increases, especially for ,
than deflections decreases. This happens because the cracking moment of the beam
increases and few cracks appear at the same level of loading, in fact for high
reinforcement ratios , so is used instead of Ie. The compressive strength of
concrete increases too, but this effect on beam’s deflection is not considerable in high
reinforcement ratios.
In this study is used only one type of GFRP bar but exist different types of FRP
reinforcements with very large properties, so different results are taken based on
different methods. For this reason a lot of researchers, based on a large numbers of
experimental tests are modifying and optimizing Branson’s equation, so that the
predicted values of deflection approach the experimental values. New models are
going to be developed based on experimental results and elaborated genetic algorithm
used to evaluate the effects of several parameters and reevaluate the power “m” in
Branson’s equation.

References

1. Hota V. S. GangaRao, Narendra Taly, P. V. Vijay: Reinforced Concrete Design With FRP
Composites, CRC Press Taylor & Francis Group, Boca Raton, Florida (2007)
2. American Concrete Institute (ACI): Guide for the Design and Construction of Concrete
3. American Concrete Institute, ACI Committee 440, ACI 440.1R-01, Farmington Hills,MI:
Reinforced With FRP Bars (2001)
4. American Concrete Institute (ACI): Guide for the Design and Construction of Concrete
5. Reinforced With FRP Bars, ACI Committee 440, ACI 440.1R-03, Farmington Hills, MI:
American Concrete Institute (2003)
6. Branson, D.E.: Deformation of Concrete Structures, New York: McGraw-Hill (1977)
7. ISIS Canada: Intelligent Sensing for Innovative Structures, Reinforcing Concrete Structures
With Fiber-Reinforced Polymers, Design Manual No. 3, Version 2 (2007)
8. Lawrence C. Bank: Composites for construction, Structural Design with FRP Materials,
John Wiley & Sons, Inc., Hoboken, New Jersey (2006)
9. CEB-FIB Bulletin: FRP reinforcement for RC structures (2006)
10. Nanni A. and Faza S.: Design and construction of concrete reinforcement with FRP bars ,
an emerging technology (2001)
11. Consiglio Nazionale delle Ricercce: Istruzioni per la Progettazione, l’Esecuzione ed il
Controllo di Strutture di Calcestruzzo armato con Barre di Materiale Composito
Fibrorinforzato, CNR-DT 203/ (2006)
12. Benmokrane, B., Challal, O., and Masmoudi, R.: Flexural response of concrete beams
reinforced with FRP bars, ACI Struct. J., 93, 1, 46–55, January-February (1996)
13. Abdalla, H. A. : Evaluation of deflection in concrete member reinforced with fiber
reinforced polymer , FRP bars (2002)
14. Yost,J.R.,Gross,S.P.,and Dinehart,D.W.: Effective moment of inertia for glass fiber-
reinforced polymer-reinforced concrete beams” ACI Struct. J. , 100(6), 732-739. (2003)
15. Canadian Standard Association (CSA).: Design and construction of building components
with fiber-reinforced polymers, Concrete Design Handbook : Canadian Standard S806-02,
CSA, Toronto, (2002)
16. Bischoff, P.H.: Deflection calculation of FRP reinforced concrete beams on modifications
to the existing Branson’s equation. (2007)