Journal of Advanced Concrete Technology Vol. 4, No.

1, 85-97, February 2006 / Copyright © 2006 Japan Concrete Institute 85

Scientific paper

Nonlinear Analysis for Reactive Powder Concrete Beams under Rapid

Flexural Loadings
Kazunori Fujikake1, Takanori Senga2, Nobuhito Ueda3, Tomonori Ohno4 and Makoto Katagiri5

Received 27 June 2005, accepted 3 November 2005

Abstract
The aim of this study was to develop an analytical model based on a fiber model technique for representing the behavior
of a reinforced Reactive Powder Concrete (RPC) beam subjected to rapid flexural loads. In the analytical model, first, the
moment-curvature relationship of the section of the RPC beam was calculated, considering the fact that the constituent
materials, i.e., RPC and reinforcing steel, exhibit strain rate effects on mechanical properties. Then, the load-midspan
deflection relationship was obtained through the moment-curvature relationship. The analytical model was applied to the
experimental results for verification. The analytical results were in good agreement with the experimental results. Sub-
sequently, analytical investigations were performed to find out the influence of variables, such as loading rates, com-
pressive strengths, the amount of reinforcing steel and the volume fraction of steel fibers, on the behaviors of RPC beams.

1. Introduction this testing.

In recent years, finite element methods or analytical
Reactive Powder Concrete (RPC) reinforced with short methods based on a fiber model technique have been
steel fibers, which is generally classified as high-per- applied to nonlinear analysis in which the load-deflection
formance fiber reinforced cementitious composite, has relationships of reinforced concrete beams under flexural
ultra-high strength and high toughness. These excellent loads are obtained. From a practical point of view, the
properties make it very suitable for structures that need to analytical method based on the fiber model technique is
resist the impact loading that may result from crashing excellent to the extent that it can easily model the be-
vehicles, ships or airplanes, falling rocks, avalanches or haviors of reinforced concrete members through crack
explosions. occurrence to the ultimate state with little computational
To examine the structural safety of RPC members effort (Park and Paulay 1975, Poston et al. 1985, Krau-
subjected to impact loadings in design, it is essential to thammer et al. 1987, Ikeda et al. 1990, Suzuki et al. 1994,
analytically predict both the load bearing capacity and Watson et al. 1994, Ahmad and Weerakoon 1995,
the deformation capacity of the RPC members, consid- Ezeldin and Shiah 1995). The fiber model technique has
ering strain rate effects on the mechanical properties of been successfully applied not only to reinforced concrete
constituent materials, i.e., RPC and reinforcing steel. members but also to concrete-filled steel tube members
RPC, however, being a relatively new material, little (Tomii and Sakino 1979, Varma et al. 2005).
analytical work has been performed on the behavior of This study provides an analytical model to determine a
RPC subjected to impact loading. Meanwhile, we have load-midspan deflection relationship for a reinforced
already performed rapid flexural loading testing of RPC RPC beam, in which the loading rate effects of the con-
beams to investigate the influence of the loading rate and stituent materials are duly considered. The analytical
amount of longitudinal steel reinforcing bar on the me- model includes 1) determining the moment-curvature
chanical characteristics of RPC beams (Ueda et al. 2005). relationship of the section of the RPC beam based on the
The quasi-static flexural response of RPC beams under fiber model technique; and 2) evaluating the load-
constant high deformation velocity was examined during midspan deflection relationship through the moment-
curvature relationship.

1 2. Nonlinear analysis based on fiber model

Associate Professor, Department of Civil and
technique
Environmental Engineering, National Defense Academy,
Japan. This study is intended to develop an analytical model to
E-mail: [email protected] predict the load-midspan deflection relationship of a
2
Officer, Ground Self-Defense Force, Japan. reinforced RPC beam (hereafter simply referred to as
3
Research Engineer, Research & Development Center of “RPC beam”) subjected to rapid flexural loads as shown
Taiheiyo Cement Corporation, Japan. in Fig. 1. The RPC beam is simply supported over a span
4
Professor, Department of Civil and Environmental L , and loaded symmetrically at two points at the
Engineering, National Defense Academy, Japan. midspan deflection rate δ .
5
Manager, Research & Development Center of Taiheiyo
Cement Corporation, Japan.
86 K. Fujikake, T. Senga, N. Ueda, T. Ohno and M. Katagiri / Journal of Advanced Concrete Technology Vol. 4, No. 1, 85-97 2006

P P
2 2

M
(b) Moment distribution

L1 L1 δ
L δ
(a) Loading configuration (c) Deflection curve
Fig. 1 RPC beam subjected to rapid flexural loads.

yi
y0 φ ε c,i εc,i σ c,i
i-th fiber
element
ys
Ac ,i φ φ
σs
εs εs

(a) Fiber element (b) Strain and strain-rate distributions (c) Stress distributions
Fig. 2 Section analysis based on fiber element method.

⎧
2.1 Section analysis ⎪ f tf ,d for 0 ≤ wc ≤ w1
⎪
2.1.1 Basic assumptions ⎪ f tf ,d − σ 2
In section analysis based on a fiber model technique, σ (wc ) = ⎨ f tf ,d − (wc − w1 ) for w1 < wc ≤ w2
⎪ w2 − w1
the section of the RPC beam was divided into a number
⎪ σ2
of discrete fiber elements, as shown in Fig. 2. The fol- ⎪σ 2 − w − w (wc − w2 ) for w2 < wc ≤ w3
lowing assumptions were made to calculate the moment- ⎩ 3 2
curvature relationship. (1)
1) Plane sections of the RPC beam before bending
remain plane after bending. where wc = crack opening, f tf , d = dynamic tensile
2) Stress and strain within each discrete fiber element strength, σ 2 = 4.8 MPa; w1 = 0.4 mm, w2 = 2.0 mm,
are constant over the element, and these values are and w3 = 4.4 mm. The dynamic tensile strength f tf , d
calculated at the centroid of the fiber element. was given as the function of strain rate ε .
3) Additional deformation due to shearing force is
⎧
1.95
⎡ ⎛ ε ⎞⎤
ignored. ⎪ ⎛ ε ⎞
0.0013 ⎢ Log ⎜⎜
⎣⎢ ⎝ εst
⎟⎥
⎟
⎠ ⎦⎥
4) A perfect bond exists between RPC and reinforcing ⎪ f tf , s ⎜⎜ ⎟⎟ for ε ≥ εst (2)
steel. f tf ,d =⎨ ⎝ ε st ⎠
5) The stress-strain curves for RPC and reinforcing ⎪
steel are known. ⎪
⎩ f tf , s for ε < εst
6) To incorporate the rate dependences of the con-
stituent materials, the curvature varies with the where f tf , s = static tensile strength = 10.8 MPa, and ε st
constant curvature rate φ . = strain rate corresponding to static loading = 1.0×10-6 /s.
For the section analysis using the fiber model tech-
2.1.2 Stress-strain relationship of RPC with nique, the tensile stress-crack opening relationship has to
strain rate effects be transformed into the tensile stress-strain relationship
(1) In tension expressed in an average strain within a crack spacing Lc
Fujikake et al. (2006) have proposed the tensile stress- as shown in Fig. 3. Based on the fictitious crack model
crack opening relationship of RPC with rate effects as by Hillerborg et al. (1976), a deformation at any stress
follows: after cracking within the crack spacing Lc consists of an
elastic deformation for the concrete between cracks and a
crack opening wc at the crack. The average tensile
strain within the crack spacing is calculated as the total
 K. Fujikake, T. Senga, N. Ueda, T. Ohno and M. Katagiri / Journal of Advanced Concrete Technology Vol. 4, No. 1, 85-97 2006 87

σ criterion to predict the maximum strengths of RPC, in

which the effects of steel fiber volume fractions,
f tf ,d
Lc Lc strain-rates and triaxial compressive stresses were con-
σ σ sidered. Thus, for RPC with a steel fiber volume fraction
wc wc Ec of 2%, the dynamic increased ratio of dynamic com-
ε tf ,d ε pressive strength f cf′ ,d to static compressive strength
(a) Crack spacing (b) Stress-strain relationship f cf′ , s can be calculated by the dynamic failure criterion.
without crack plane The calculation process, however, is rather complicated.
Hence, from a regression analysis on the calculated dy-
σ σ namic increased ratios f cf′ ,d f cf′ , s , the following rela-
f tf ,d f tf ,d tionship is obtained.

⎧
0.951
⎡ ⎛ ε ⎞⎤
σ2 σ2 0.0055 ⎢ Log ⎜⎜ ⎟⎟ ⎥
⎪⎛ ε ⎞ ⎣⎢ ⎝ εsc ⎠ ⎦⎥
w3 εt2 f cf′ ,d ⎪⎜ ⎟⎟ for ε ≥ εsc
wc = ⎨⎜⎝ ε sc
w1 w2 ε tf ,d ε t1 εt3 ε (9)
⎠
(c) Stress-crack opening rela- (d) Stress-average strain re- f cf′ ,s ⎪
tionship lationship ⎪
Fig. 3 Tensile stress-strain relationship for RPC. ⎩1.0 for ε < εsc

where εsc = 1.2 × 10 −5 (1/s).

deformation divided by the crack spacing Lc . Hence,
the tensile stress-average strain relationship is expressed 2.1.3 Stress-strain relationship of reinforcing
as: steel with strain-rate effect
For reinforcing steel, a bilinear stress-strain relationship
⎧Ec ⋅ ε for ε ≤ ε tf,d is adopted as shown in Fig. 5. It is assumed that the
⎪f for ε tf , d < ε ≤ ε t1 elastic modulus E s and the strain hardening modulus
⎪ tf ,d
⎪ f tf ,d − σ 2 E sp are independent of loading rates. The effect of the
σ (ε ) = ⎨ f tf ,d − ε − ε (ε − ε t1 )
⎪ for ε t1 < ε ≤ ε t 2 loading rate on the yield strength is considered. For the
⎪
t2 t1 dynamic yield strength ( f sy , d ), the following formula-
σ2
⎪σ 2 − (ε − ε t 2 ) for ε t 2 < ε ≤ ε t 3 tion proposed by Takahashi is adopted (JSCE 1993):
⎪ ε t3 − ε t 2
⎩⎪0 for ε t 3 < ε
(3)
Stress
where ε tf , d , ε t1 , ε t 2 and ε t 3 are given as:
ε tf ,d = f tf ,d E c (4) f cf′ , d

ε t1 = ε tf ,d + w1 Lc = f tf ,d E c + w1 Lc (5)

ε t 2 = σ 2 E c + w2 Lc (6)

ε t 3 = w3 Lc (7) Ec
ε cf′ ,d
(2) In compression Fig. 4 Compressive stress-strain relationship for RPC.
RPC subjected to rapid compressive loading does not
exhibit significant nonlinear behavior (Fujikake et al. Stress
(2002)). As shown in Fig. 4, the compressive stress-
strain relationship of RPC can be assumed as linear
elastic because of its simplicity, in which the dynamic f sy ,d Esp
uniaxial compressive strength f cf′ , d with the rate effect
is defined as an ultimate state. Thus, the compressive
stress-strain relationship can be given as:
σ = Ec ⋅ ε (8) Es
where ε ≤ f cf′ ,d E c . ε sy,d Strain
Fujikake et al. (2002) have proposed a dynamic failure
Fig. 5 Stress-strain relationship for reinforcing steel.
88 K. Fujikake, T. Senga, N. Ueda, T. Ohno and M. Katagiri / Journal of Advanced Concrete Technology Vol. 4, No. 1, 85-97 2006

f sy ,d fiber element, and y s = distance from the extreme com-

= 1.202 + 0.040 × log 10 ε ≥ 1.0 (10) pression fiber to the tensile reinforcing steel. In the beam
f sy , s
member, as no axial force is acted on the section, the
where f sy , d = dynamic yield strength at a strain rate ε , condition of N = 0 must be satisfied.
and f sy , s = static yield strength. From a linear distribution of strains across the section
Finally the stress-strain relationship for reinforcing depth in the assumptions, the strains of the i-th concrete
steel is given as: fiber element ( ε c,i ) and of the reinforcing steel element
( ε s ) are expressed as:
⎧E s ⋅ ε for ε ≤ ε sy ,d
σ (ε ) = ⎨ (11) ε c ,i = ( y 0 − y i ) ⋅ φ (16)
f
⎩ sy ,d + E sp (ε − ε sy ,d ) for ε > ε sy ,d
ε s = ( y0 − ys ) ⋅ φ (17)
where ε sy ,d (= f sy , d E s ) = strain corresponding to the
dynamic yield strength, and E sp = strain hadening where y 0 = distance from the extreme compression
modulus = 0.01E s . fiber to the neutral axis, and φ = curvature of the sec-
tion.
2.1.4 Relationship between midspan deflection From the basic assumptions, strain rates vary linearly
rate and curvature rate with depth at the section as well. Once the curvature rate
When section analysis is performed for an RPC beam φ is given, the strain rate at each fiber element is de-
that is simply supported and subjected to two point loads termined as:
as shown in Fig. 1, the relationship between the midspan
deflection rate ( δ ) of the beam member and the curva- εc ,i = ( y0 − yi ) ⋅ φ for concrete fiber elements (18)
ture rate ( φ ) at the section is required to calculate the
strain rate at each fiber element in this analysis. Based on εs = ( y 0 − y s ) ⋅ φ for reinforcing steel (19)
the linear elastic theory, the relationship between the
midspan deflection δ and the curvature φ as shown in Therefore, the stresses of any concrete fiber element and
Fig. 1 can be given as (Gere 2003): of the reinforcing steel element ( σ c,i ，σ s ) are calculated
with the strains ( ε c,i ， ε s ) and the strain rates ( εc,i ， εs )
M 24 (12) at the corresponding elements. Figure 6 shows the cal-
φ= = 2 δ
EI 3L − 4 L1 2
Input material properties, sectional details
In a similar manner, it is simply assumed that the fol- and midspan deflection rate δ
lowing relationship exists between the curvature rate φ
and the midspan deflection rate δ . Calculate curvature rate φ by Eq. (13)

24
φ = 2
δ (13) Set curvature φ = φ + ∆φ
3L − 4 L12

Assume neutral axis position y0

This assumption seems to be reasonable because the
influence of loading rate on the mechanical properties of Calculate strain distribution and
RPC is expressed as a function of the logarithmic value strain-rate distribution on the section
by Eq. (16), (17), (18), (19)
of the strain rate.
Next step
Calculate each concrete fiber stress
2.1.5 Analytical moment-curvature relationship and re-bar stress by Eq. (3), (8), (11)
Dividing the reinforced concrete section into n-fiber
elements as shown in Fig. 2, an axial load and bending Calculate axial force by Eq. (14)

moment acting on the section are given as:

No
n Equilibrium of axial force
N = ∫ σdA = ∑ σ c ,i Ac ,i + σ s As (14)
Yes
A
i =1
Calculate bending moment by Eq. (15)
n
M = ∫ σydA = ∑ σ c ,i y i Ac ,i + σ s y s As (15)
A
i =1 Does the strain of extreme No
compression fiber reach
where σ c,i = stress acting on the i-th concrete fiber ele- ultimate state?

ment, Ac ,i = area of the i-th concrete element, σ s = Yes

stress of the tensile reinforcing steel, As = area of the End
tensile reinforcing steel, yi = distance from the extreme
compression fiber to the centroid of the i-th concrete Fig. 6 Calculation flow for section analysis.
 K. Fujikake, T. Senga, N. Ueda, T. Ohno and M. Katagiri / Journal of Advanced Concrete Technology Vol. 4, No. 1, 85-97 2006 89

culation flow for the moment-curvature relationship. M cm , the total load P acting on the beam is calculated
as:
2.2 Load-midspan deflection curves of RPC
beams subjected to rapid flexural loads 4 M cm (21)
P=
The midspan deflection of an RPC beam subjected to L − Lcm
rapid flexural loading at the midspan deflection rate δ
can be calculated by integrating the curvature distribu- The bending moment at any point x j (j=0~m) is given
tion over the half-length of the beam, which is deter- as:
mined based on the moment-curvature relationship ob-
j ⋅ ∆x P
tained from the section analysis at the curvature rate φ Mj = ⋅ M cm = j ⋅ ∆x ⋅ (22)
given by Eq. (13), taking the boundary conditions into (L − Lcm ) 2 2
consideration.
As shown in Fig. 7, the half-length of the beam is Once the bending moment M j at each point is calcu-
considered in the calculation. The half-length of the span lated, the corresponding curvature φ j is determined
without the pure bending moment region is divided into from the moment-curvature relationship. Hence, the
m equal segments. The coordinate at each division point midspan deflection can be calculated as:
is follows: ∆x
δ = ∫ 2 φxdx = ∑ (φ j x j + φ j −1 x j −1 )
L m
L
+ φcm (2 L − Lcm ) cm
x j = j ⋅ ∆x ( j = 0,1," m) (20) 0
j =1 2 8
(23)
where ∆x = (L − Lcm ) 2m , and Lcm = length of a pure
bending moment region.
Assuming that the bending moment M cm acts on the 3. Verification of nonlinear analysis based
pure bending moment region, from the relationship be- on fiber model technique
tween the load P 2 at x m and the bending moment
3.1 Comparison of analytical results with rapid
flexural loading test results
L2
m ⋅ ∆x Lcm 2
To verify the proposed nonlinear analysis based on the
fiber model technique, the proposed analysis was applied
P to the experimental results obtained by Ueda et al. (2005).
2
x0 Ueda et al. have performed the rapid flexural loading test
x1 x2 x j −1 x j xm for RPC beams. In the test, the RPC beams were simply
( j −1) ⋅ ∆x supported over a span of 1200 mm. Figure 8 shows the
j ⋅ ∆x
cross-section of an RPC beam. Two different amounts of
(a) Calculation points longitudinal steel reinforcement with an yield strength of
295 MPa were provided. They were designated as FM13
and FM19, respectively. Three numbers of deformed bars
M j −1 with a diameter of 13 mm for FM13 and 19 mm for
Mj FM19 were used. The tensile reinforcement areas pro-
(b) Moment distribution vided for the FM13 and FM19 were 380 and 859 mm2,
respectively. The RPC beams were loaded symmetrically
φ j −1 φ at two points spaced 150 mm apart at the midspan de-
j flection rates of 1.4×10-4, 4.5×10-2, 4.9×10-1 and 2.1×100
m/s. The midspan deflection rate of 1.4×10-4m/s is re-
(c) Curvature distribution

150
50 50 50
(d) Deflection curve
20
10

M cm
100

Mj
120
200

M j −1
30 30 10
60

φ j −1 φ j φm φcm
(e) Moment-curvature relationship at φ 30 45
150
45 30
Unit: mm

Fig. 7 Illustration for analytical procedure. Fig. 8 Cross-section detail by Ueda et al. (2005).
90 K. Fujikake, T. Senga, N. Ueda, T. Ohno and M. Katagiri / Journal of Advanced Concrete Technology Vol. 4, No. 1, 85-97 2006

ferred to as static loading. 75 mm) determined from the analysis consists with the
First the analytical results are compared with the test half length of the pure bending moment region, the ana-
data obtained from the static flexural loading test of RPC lytical results seem to be reasonable. However, to ex-
beams. The material parameters used in the analysis are a
static compressive strength of 214.7 MPa and an elastic 250
modulus of 55.0 GPa for RPC, and a static yield strength Analysis

of 295 MPa and an elastic modulus of 200 GPa for re- 200
Test data Lc= 50 mm
Lc= 75 mm

inforcing steel. In the analysis, the crack spacing length Lc=100 mm

Lc=125 mm
for evaluating the tensile stress-strain relationship is

Load (kN)
150
selected to be 50, 75, 100 or 125 mm. The cross-section
100
of the RPC beam in Fig. 8 is divided into 200 discrete
fiber elements. The half-length of the span without the 50
pure bending moment region is divided into 100 equal
segments as well. The analysis is terminated when the 0
0 5 10 15 20 25
stress of the extreme compression fiber of RPC reaches Midspan deflection (mm)
the compressive strength. (a) FM13 specimen
Figure 9 shows the experimental and the analytical 400
results. As can be seen, when the crack spacing Lc is 75 350
Analysis
Lc= 50 mm
mm, the analytical results are in good agreement with the 300
Test result
Lc= 75 mm
experimental results. However, the initial stiffness of the Lc= 100 mm
250

Load (kN)
experimental result seems to be lower than that of the Lc= 125 mm

200
analytical result due the extraneous deformations due to
150
settlement and crushing at the supports. The analytical
results clearly show that the value of the crack spacing 100

Lc affects the post-peak behavior in the load-midspan 50

deflection curve. The larger the crack spacing, the 0

0 5 10 15 20 25 30
steeper the descending curve. Thus, the load-midspan Midspan deflection (mm)
deflection curves tend to be more brittle with increases in (b) FM19 specimen
the crack spacing Lc . In the test, one major crack was
formed at the center of the pure bending moment region Fig. 9 Load-midspan deflection relationships for RPC
with a length of 150 mm. Since the crack spacing ( Lc = beams subjected to static loading.

250 250 300

Analysis Analysis
250
200 Analysis 200 Test results Test results
Test results
200
Load (kN)

Load (kN)

Load (kN)

150 150
150
100 100
100

50 50
50

0 0 0
0 5 10 15 20 25 0 5 10 15 20 25 0 5 10 15 20 25
Midspan deflection (mm) Midspan deflection (mm) Midspan deflection (mm)

(a) δ = 4.5×10-2 m/s (b) δ = 4.9×10-1 m/s (c) δ = 2.1×100 m/s

Fig. 10 Comparison of proposed model with test results for FM13.

500 500 500

Test results Test results Test results

400 400 Analysis
400
Analysis Analysis
Load (kN)

Load (kN)
Load (kN)

300 300 300

200 200 200

100 100 100

0 0 0
0 5 10 15 20 25 30 0 5 10 15 20 25 30 0 5 10 15 20 25 30
Midspan deflection (mm) Midspan deflection (mm) Midspan deflection (mm)

(a) δ = 4.5×10-2 m/s (b) δ = 4.9×10-1 m/s (c) δ = 2.1×100 m/s

Fig. 11 Comparison of proposed model with test results for FM19.
 K. Fujikake, T. Senga, N. Ueda, T. Ohno and M. Katagiri / Journal of Advanced Concrete Technology Vol. 4, No. 1, 85-97 2006 91

amine the crack spacing precisely, another flexural D correspond to the crack occurrence, the first yielding
loading test for the RPC beams with different lengths of of the reinforcing steel, the maximum load, and the ul-
pure bending region will be required. timate state, respectively.
For the FM13 and the FM19 beams with different In the stress distributions, the position of the neutral
amounts of reinforcing steel, comparison of the analyti- axis moves upward with increases in curvature. When
cal results with the test results was executed. Figures 10 the moment reaches the maximum value, the stresses in
and 11 show the load-midspan deflection curves at the tension are almost equal to the maximum tensile strength
midspan deflection rates of 4.5×10-2, 4.9×10-1 and and uniformly distributed. After the maximum moment,
2.1×100 m/s. In the analysis, a crack spacing of 75 mm the tensile stress at the bottom fiber decreases to zero at
was selected, based on the analytical results under static the ultimate state.
loading. The analytical results can be seen to be in good
agreement with the test results at each loading rate. Thus, 4.3 Strain rates over a cross-section of RPC
the proposed nonlinear analysis can be used to represent beam
the behavior of RPC beams under dynamic loading, For the FM13 and FM19 beams loaded at the midspan
unless the RPC beams exhibit shear failure. The RPC deflection rate of 2.1×100 m/s, Fig. 17 shows the rela-
beams in the test still have a certain level of load bearing tionships between the strain rates both at the extreme
capacity over the ultimate midspan deflection obtained compression fiber of RPC and at the centroid of tensile
from the proposed nonlinear analysis. In design, RPC
beams subjected to impact loads are strongly required to 200
possess a certain level of redundancy to prevent catas- Test
trophic failure. From that perspective, the ultimate state 150 Analysis

determined by the proposed nonlinear analysis seems to

be reasonable. Load (kN)
100

3.2 Experimental results of Kakei et al. (2003)

Kakei et al. (2003) have executed the static flexural 50

loading test of a RPC beam with the same cross-section

shown in Fig. 8. In the test, the RPC beam was simply 0
0 5 10 15 20 25
supported over a span of 1350 mm and loaded symmet- Midspan deflection (mm)
rically at two points spaced 150 mm apart. The RPC had Fig. 12 Comparison of proposed model with test result of
a compressive strength of 224.3 MPa and the reinforcing Kakei et al. (2003).
steel had a yield strength of 345 MPa. Figure 12 shows
both the experimental and the analytical load-midspan 300
deflection curves. The analytical curve is generally in 2 . 0 × 10 0 m/s

agreement with the experimental curve. 250 4 . 9 × 10 − 1 m/s

4 . 5 × 10 − 2 m/s
200
4. Analytical investigation 1 .4 × 10 − 4 m/s
Load (kN)

150
4.1 Influence of loading rate on load-midspan 100
deflection
Figure 13 shows the load-midspan deflection curves 50
calculated at the midspan deflection rates of 1.4×10-4, 0
4.5×10-2, 4.9×10-1 and 2.1×100 m/s for the FM13 and 0 5 10 15 20
Midspan deflection (mm)
FM19 beams. As can be seen, the maximum load in-
creases with increases in the loading rate while the slope (a) FM13 specimen
400
of the descending branch of the load-midspan deflection 2 . 0 × 10 0 m/s
350
curve is much steeper with increases in the loading rate. 4 . 9 × 10 − 1 m/s

The ultimate deflection increases with increasing the 300

loading rate, while the deflections corresponding to the 250

Load (kN)

maximum loads are almost independent of the loading 200 4 .5 × 10 − 2 m/s

rates. 150
1 . 4 × 10 − 4 m/s

100
4.2 Normal stress distributions over a
50
cross-section of RPC beam
0
The moment-curvature relationships calculated at the 0 2 4 6 8 10 12 14
midspan deflection rates of 1.4×10-4 and 2.1×100 m/s for Midspan deflection (mm)
the FM13 beam are shown in Fig. 14, and the stress dis- (b) FM19 specimen
tributions over the cross-section at the four points of A to Fig. 13 Influence of loading rate on load-midspan deflec-
D in the figure are shown in Figs. 15 and 16. Points A to tion relationship.
92 K. Fujikake, T. Senga, N. Ueda, T. Ohno and M. Katagiri / Journal of Advanced Concrete Technology Vol. 4, No. 1, 85-97 2006

200 200

50
150 150

Height (mm)

Height (mm)
40 C 100 100
Moment (kN.m)

B 50 50
30
D
0 0
-50 0 50 100 150 200 250 300 -50 0 50 100 150 200 250 300
20
Stress (MPa) Stress (MPa)

(a) At point A (b) At point B

10 A
(Cracking moment) (Yielding moment)
200 200
0
-4 -4 -4 -4
0 1.0 10 2.0 10 3.0 10 4.0 10
150 150
Curvature (1/mm)

Height (mm)

Height (mm)
-4
(a) At the deflection rate of 1.4×10 m/s 100 100
80
50 50

C
60 0 0
Moment (kN.m)

-50 0 50 100 150 200 250 300 -50 0 50 100 150 200 250 300
B Stress (MPa) Stress (MPa)

40 (c) At point C (d) At point D

D
(Max. moment) (Ultimate moment)
20
A Fig. 16 Stress distribution for FM13 at the deflection rate
0
of 2.1×10 m/s.
0 -4 -4 -4 -4
0 1.0 10 2.0 10 3.0 10 4.0 10 3.5
Curvature (1/mm) At extreme compression fiber of RPC
3
(b) At the deflection rate of 2.1×100 m/s At reinforcing steel

Fig. 14 Moment-curvature relationship for FM13. 2.5

Strain rate (1/s)

2
200 200
1.5
150 150
Height (mm)

Height (mm)

1
100 100
0.5
50 50
0
-4 -4 -4 -4 -4
0 0
0 1 10 2 10 3 10 4 10 5 10
-50 0 50 100 150 200 250 -50 0 50 100 150 200 250 Curvature (1/mm)
Stress (MPa) Stress (MPa)
(a) FM13 beam
(a) At point A (b) At point B 3.5
(Cracking moment) (Yielding moment)
At extreme compression fiber of RPC
200 200 3 At reinforcing steel

150 150 2.5

Strain rate (1/s)
Height (mm)

Height (mm)

100 100 2

50 50
1.5

1
0 0
-50 0 50 100 150 200 250 -50 0 50 100 150 200 250
Stress (MPa) Stress (MPa) 0.5
(c) At point C (d) At point D
0
(Max. moment) (Ultimate moment) 0 5 10
-5
1 10
-4
2 10 2 10
-4 -4 -4
3 10
Fig. 15 Stress distribution for FM13 at the deflection rate Curvature (1/mm)
-4
of 1.4×10 m/s. (b) FM19 beam
Fig. 17 Strain-rate variation at a section at the midspan
0
reinforcing steel and the curvatures. At the initial state of deflection rate of 2.1×10 m/s.
loading, the strain rates at the extreme compression fiber
and at the reinforcing steel exhibit almost certain values,
i.e., 1.8 and 0.8 /s, respectively. However, while the cur- 4.4 Influence of loading rate on yielding load
vatures increase after crack formation, the strain rates at and maximum load
the extreme compression fiber and at the reinforcing steel Figure 18 shows the influence of loading rates on both
are varied because the neutral axis position moves up- the yielding load and the maximum load for the FM13
wards, even though the curvature rate at the cross-section and FM19 beams. The yield load and the maximum load
remains constant. When the curvature increases up to the can be seen to increase with increases in the loading rate.
ultimate value, the strain rate at the extreme compression The yielding load and the maximum load at each loading
fiber decreases to approximately one fifth of its initial rate for the FM13 beams are determined to be approxi-
value, while the strain rate at the reinforcing steel in- mately 65% of those values for the FM19 beams, while
creases to three times its initial value. the ratios of the maximum load to the yielding load at
each loading rate for the FM13 and FM19 beams are
 K. Fujikake, T. Senga, N. Ueda, T. Ohno and M. Katagiri / Journal of Advanced Concrete Technology Vol. 4, No. 1, 85-97 2006 93

approximately a constant value of 1.10 regardless of the proximately 6.5 at each loading rate, while that for the
amount of reinforcing steel. FM19 is 3.5 at each loading rate. In the ductility corre-
sponding to the ultimate state, these analytical results
4.5 Influence of loading rate on absorbed en- suggest that the FM19 beams exhibit more brittle failure
ergy than the FM13 beams.
The absorbed energies corresponding to the yielding load,
the maximum load and the ultimate load are calculated 4.7 Crack distribution
by integrating the load-midspan deflection curve ob- The proposed analytical method assumes that the crack
tained by the analysis. Figure 19 shows the relationship begins to form when the stresses at the cross-section
between the absorbed energies and the loading rates. The reach the maximum tensile strength. Thus, in the section
absorbed energies increase with increases in the loading analysis, the crack depth corresponding to the given
rate. The energies absorbed up to the yielding load and up curvature can be easily estimated by checking the value
to the maximum load for the FM19 beams are approxi- of tensile stresses over the cross-section. Finally the
mately twice those for the FM13 beams. On the other crack depth-moment relationship is calculated from both
hand, the ultimate energy, absorbed up to the defined the moment-curvature relationship and the crack
ultimate state at which the stress of the extreme com- depth-curvature relationship obtained from the section
pression fiber of RPC reaches dynamic compressive analysis. The crack depth induced at any position in the
strength f cf′ ,d , for the FM19 beam is slightly larger than RPC beam can be evaluated based on the calculated
that for the FM13 beam at each loading rate and the crack depth-moment relationship. Figure 21 shows the
comparison shows no significant differences after all. crack distributions analytically estimated for the FM13
beam and the FM19 beam at the rates of 1.4×10-4 and
4.6 Influence of loading rate on ductility
To investigate the structural performance of RPC beams 600

in the analytical load-midspan deflection relationships, 500

FM13 beam
FM19 beam
ductility is defined as the ratio of the deflection at the
maximum load or at the ultimate state to the deflection at 400
Energy (J)

the yielding load. Figure 20 shows the ductility calcu- 300

lated for the FM13 and FM19 beams. The ductility cor-
200
responding to the maximum load for both the FM13 and
FM19 beams is determined to be approximately 2.0. At 100
the ultimate state, the ductility for the FM13 is ap-
0
1.4x10-4 4.5x10-2 4.9x10-1 2.1x100
400 Midspan deflection rate (m/s)
FM13 beam (a) At yielding moment
350
FM19 beam 1400
300
Yielding load (kN)

FM13 beam
1200
250 FM19 beam
1000
200
Energy (J)

800
150
100 600

50 400

0 -4 -2 -1 0
200
1.4x10 4.5x10 4.9x10 2.1x10
0
Midspan deflection rate (m/s) 1.4x10-4 4.5x10-2 4.9x10-1 2.1x100
(a) Yielding load Midspan deflection rate (m/s)
400 (b) At max. moment
FM13 beam 3500
350
FM19 beam FM13 beam
300 3000
FM19 beam
Max. load (kN)

250 2500
Energy (J)

200 2000

150 1500

100 1000
50 500
0 -4 -2 -1 0 0
1.4x10 4.5x10 4.9x10 2.1x10 1.4x10-4 4.5x10-2 4.9x10-1 2.1x100
Midspan deflection rate (m/s) Midspan deflection rate (m/s)
(b) Max. load (c) At ultimate moment
Fig. 18 Yielding load and max. load at each loading rate. Fig. 19 Energy obtained by analytical model.
94 K. Fujikake, T. Senga, N. Ueda, T. Ohno and M. Katagiri / Journal of Advanced Concrete Technology Vol. 4, No. 1, 85-97 2006

10 tion length at the rate of 2.1×100 m/s is slightly smaller

At max. load
At ultimate
than that at the rate of 1.4×10-4 m/s. The crack distribu-
8
tion lengths for the FM13 and the FM19 beams were
6
actually 800 mm and 1000 mm in the test, respectively
Ductility

(Ueda et al. 2005). The influence of the loading rate on

4 the length of the crack region has not been observed. The
length of the crack region estimated by the analysis al-
2 most agrees with that obtained from the rapid loading
test.
0
1.4x10-4 4.5x10-2 4.9x10-1 2.1x100
Midspan deflection rate (m/s) 4.8 Influence of compressive strength on
(a) FM13 load-midspan deflection relationships
10
To examine the influence of the magnitude of compres-
At max. load
8 At ultimate
sive strength on the load-midspan deflection relation-
ships, in the analysis for the FM13 specimen subjected to
6 static loading, the load-midspan deflection curves were
Ductility

calculated with the compressive strengths of 150, 175,

4 200, 225 and 700 MPa. Figure 22 shows the analytical
results. Each of the load-midspan deflection curves can
2
be seen to follow the same path, while the compressive
0
strengths are different. The magnitude of the compres-
1.4x10-4 4.5x10-2 4.9x10-1 2.1x100 sive strength clearly affects the ultimate midspan de-
Midspan deflection rate (m/s)
flection. The ultimate midspan deflection increases with
(b) FM19 increases in compressive strength.
Fig. 20 Relation between ductility and midspan deflection
rate.
200
4.9 Influence of steel fibers on behaviors of RPC
beams
The responses of the beams using ultra-high strength
150
concrete with or without steel fibers are examined by
Crack depth (mm)

using the proposed nonlinear analysis. The stress-strain

100 relationships used in the analysis for the ultra-high
strength concrete with or without steel fibers are shown
50
-4
1.4x10 (m/s)
in Fig. 23. The difference in the stress-strain relation-
0
2.1x10 (m/s) ships for the ultra-high strength concrete with or without
0 steel fibers is only in the tensile behaviors after the
0 200 400 600 800 1000 1200
Location (mm)
maximum tensile strength. The ultra-high strength con-
(a) FM13 crete without steel fibers is assumed to behave like an
200 elastic-brittle material under tensile loading, so that in
the tensile stress-strain relationship, the tensile stress
150
suddenly drops to zero after reaching the maximum ten-
sile strength.
Crack depth (mm)

To verify the analytical method for the ultra-high

100
strength concrete without steel fibers, the analytical
model was applied to the experimental results of Kakei et
50 -4
1.4x10 (m/s) al. (2003). Kakei et al. have provided the static flexural
bending test on a reinforced ultra-high strength concrete
0
2.1x10 (m/s)

0 beam with the same cross-section of the FM13 and with a

0 200 400 600 800 1000 1200
Location (mm) shear reinforcement ratio of 2.0%. In the test, the beam
(b) FM19 was simply supported over a span of 1350 mm, and was
Fig. 21 Crack distributions. loaded symmetrically at two points spaced 150 mm apart.
The compressive strength of the ultra-high strength con-
crete used was 209.4 MPa, and the yield strength of the
2.1×100 m/s. The crack distribution lengths along the reinforcing steel used was 345 MPa as well. Figure 24
beam axis for the FM13 are approximately 900 mm at the shows both the experimental and the analytical results.
rate of 1.4×10-4 m/s and approximately 800 mm at the As can be seen, the analytical result agrees well with the
rate of 2.1×100 m/s, while those for the FM19 are ap- experimental result. However, Kakei et al. have not
proximately 1000 mm at the rate of 1.4×10-4 m/s and clearly touched on whether the ultra-high strength con-
approximately 900 mm at the rate of 2.1×100 m/s. crete beam reached the ultimate state or not, so that the
In the FM13 and the FM19 beams, the crack distribu-
 K. Fujikake, T. Senga, N. Ueda, T. Ohno and M. Katagiri / Journal of Advanced Concrete Technology Vol. 4, No. 1, 85-97 2006 95

200 120
U.P: Ultimate Point
U.P. for f'c = 150 MPa
U.P. for f'c = 175 MPa 100
150
80
Load (kN)

Load (kN)
100 Test result
U.P. for f'c = 225 MPa 60
U.P. for f'c = 200 MPa
Analysis
Compressive strength
150 MPa
50 175 MPa
40
200 MPa
225 MPa
700 MPa 20
0
0 5 10 15 20 25 30 0
Midspan deflection (mm) 0 5 10 15 20 25 30 35 40
Fig. 22 Influence of compressive strength on load- Midspan deflection (mm)
midspan deflection relation. Fig. 24 Comparison of analytical model with test result of
Kakei et al. (2003).
Stress 50
Compression side With steel fibers

Bending Moment (kN.m)

40 Without steel fibers
f cf′ , d
30

20

Ec
10
ε t3 εt 2 ε t1 ε tf ,d
0
σ2 ε cf′ ,d Strain 0 0.0001 0.0002 0.0003 0.0004 0.0005
ftf , d Curvature (1/mm)
(a) At 1.4×10-4 m/s
Tension side 80
(a) With steel fibers (RPC) 70 With steel fibers
Stress
Bending Moment (kN.m)

Compression side 60 Without steel fibers

50
f cf′ , d
40
30
20
10
Ec
0
0 0.0001 0.0002 0.0003 0.0004 0.0005 0.0006
ε tf ,d
Curvature (1/mm)
σ2 ε cf′ ,d Strain (b) At 2.1×100 m/s
ftf , d Fig. 25 Influence of steel fibers on moment-curvature
relationship (FM13).
Tension side
(b) Without steel fibers
Fig. 23 Stress-strain relations with or without steel fibers. observed after the maximum moment for the beam with
steel fibers, while hardening behavior was observed up to
analytical ultimate midspan deflection cannot be simply the maximum moment after the yielding moment for the
beam without steel fibers.
compared with the experimental one. Thus further ex-
The influence of steel fibers on the yielding moment
perimental research will be required to check analytical
was analytically investigated for the FM13 and FM19
performance in terms of the ultimate midspan deflection.
beams. Figure 26 shows the analytical results. The
To analytically investigate the behaviors of ultra-high
strength concrete beams with and without steel fibers and FM13 and FM19 beams with steel fibers have yield
with the same shape and size, reinforcing steel ratio and moment values twice and one-and-a-half times those of
FM13 and FM19 beams without steel fibers, respectively.
loading condition as those of the FM13 beam described
It is clear from the results that the yield moment increases
in the previous section, the moment-curvature relation-
by mixing steel fibers into ultra-high strength concrete.
ships were calculated at the midspan deflection rates of
The smaller the reinforcement ratios, the greater the
1.4×10-4 and 2.1×100 m/s. The analytical results are
shown in Fig. 25. The steel fibers clearly affect the efficiency of steel fibers in improving the yield moment.
moment-curvature relationships. Softening behavior was
96 K. Fujikake, T. Senga, N. Ueda, T. Ohno and M. Katagiri / Journal of Advanced Concrete Technology Vol. 4, No. 1, 85-97 2006

70 7. The ductility at the maximum load was determined

With steel fibers
60
Without steel fibers
to be approximately 2.0 regardless of the amount of
Yielding moment (kN.m)

50
reinforcing steel at each loading rate, while the
ductility at the ultimate state decreased with an in-
40
crease in the amount of reinforcing steel at each
30 loading rate.
20 8. At the ultimate state, the FM13 beam had a ductility
10
of approximately 6.5, while the FM19 beam had a
ductility of approximately 3.5.
0
1.4x10-4 4.5x10-2 4.9x10-1 2.1x100 9. The crack distribution length obtained by the pro-
Midspan deflection rate (m/s) posed analytical method almost completely agreed
(a) FM13 with that obtained from the rapid flexural loading
100
With steel fibers
test.
Without steel fibers 10. The ultimate midspan deflection increased with
Yielding moment (kN.m)

80
increases in compressive strength.
60 11. The load-midspan deflection curves for the RPC
beams with different compressive strengths fol-
40 lowed the same path under static loading.
12. The yielding moment increases by mixing steel
20
fibers into ultra-high strength concrete. The smaller
0
the reinforcement ratio, the greater the efficiency of
1.4x10-4 4.5x10-2 4.9x10-1 2.1x100 steel fibers in improving the yield moment.
Midspan deflection rate (m/s)
(b) FM19
References
Fig. 26 Influence of steel fibers on yielding moment.
Ahmad, S. H. and Weerakoon, S. L. (1995). “Model for
Behavior of Slender Reinforced Concrete Columns
5. Conclusions under Biaxial Bending.” ACI Structural Journal, 92
(2), 188-198.
Based on the results presented in this paper, the follow- Ezeldin, A. S. and Shiah, T. W. (1995). “Analytical
ing conclusions can be drawn. Immediate and Long-Term Deflections of
1. The proposed analytical model based on the fiber Fiber-Reinforced Concrete Beams.” Journal of
model technique, in which the strain rate effects of Structural Engineering, ASCE, 121 (4), 727-738.
RPC and reinforcing steel were duly considered, is Fujikake, K., Uebayashi, K., Ohno, T., Shimoyama, Y.
in good agreement with the experimental results. and Katagiri, M. (2002). “Dynamic Properties of Steel
2. In the analytical load-midspan deflection relation- Fiber Reinforced Mortar Under High-rates of
ships, the increase in loading rate results in 1) an Loadings and Triaxial Stress States.” In: N. Jones, C.
increase in maximum load, 2) an increase in de- A. Brebbia, and A. M. Rajendran, Eds., Proceedings of
scending slope after the maximum load and 3) an the 7th International Conference on Structures Under
increase in ultimate deflection. However, the de- Shock and Impact, Montreal, WIT Press, 437-446.
flection corresponding to the maximum load is Fujikake, K., Senga, T., Ueda, N., Ohno, T. and Katagiri,
almost unaffected by the loading rate. M. (2006). “Effects of Strain Rate on Tensile Behavior
3. In the stress distribution over the cross-section of of Reactive Powder Concrete.” Journal of Advanced
RPC beams, at the maximum moment, the stresses Concrete Technology, 4 (1), 79-84.
in tension were almost equal to the maximum ten- Gere, J. M. (2003). “Mechanics of Materials.” 6th
sile strength and were uniformly distributed. Edition, Brooks/Cole Publishing Company.
4. The strain rates induced over the cross-section of Hillerborg, A., Modeer, M. and Petersson, P. E. (1976).
RPC beams were examined using the proposed “Analysis of Crack Formation and Crack Growth in
analytical method. A change in strain rates occurs Concrete by Means of Fracture Mechanics and Finite
following the start of crack formation at the section. Elements.” Cement and Concrete Research, 6,
5. The yielding load and the maximum load at each 773-782.
loading rate for the FM13 beam were determined to Ikeda, S., Yamaguchi, T. and Goto, Y. (1990). “Study on
be approximately 65% of those for the FM19 beam. the Flexural Behavior of Reinforced Concrete Beams
6. Compared with the analytical results for the FM13 with Low Reinforcement.” Concrete Research and
and FM19 beams, the increase in the amount of Technology, Japan Concrete Institute, 1 (1), 51-59. (in
reinforcing steel improved the yielding load and the Japanese)
maximum load at each loading rate. The ultimate JSCE (1993). “Impact Behavior and Design of
energy, however, was almost identical regardless of Structures.” Structural Engineering Series 6. (in
tension reinforcement. Japanese)
 K. Fujikake, T. Senga, N. Ueda, T. Ohno and M. Katagiri / Journal of Advanced Concrete Technology Vol. 4, No. 1, 85-97 2006 97

Kakei, T., Kawaguchi, T., Niwa, J. and Hyodo, H. (2003). Tomii, M. and Sakino, K. (1979). “Elasto-Plastic
“Fracture Mechanics Property of High Performance Behavior of Concrete Filled Square Steel Tubular
Fiber Reinforced Cement Composite.” Japan Cement Beam-Columns.” Transactions of the Architectural
Association 57th Cement Technical Congress, 230-231. Institute of Japan, (280), 111-120.
(in Japanese) Ueda, N., Fujikake, K., Ohno, T. and Katagiri, M. (2005).
Krauthammer, T., Shahriar, S. and Shanaa, H. M. (1987). “Dynamic flexural behaviors of RPC beams under
“Analysis of Reinforced Concrete Beams Subjected to high-rate loadings. [CD-ROM]” In: N. Banthia, T.
Severe Concentrated Loads.” ACI Structural Journal, Uomoto, A. Bentur, and S. P. Shah, Eds., Proceedings
87 (6), 473-480. of the 3rd International Conference on Construction
Park, R. and Paulay, T. (1975). Reinforced Concrete Materials, Vancouver, The University of British
Structures, John Wiley & Sons. Columbia.
Poston, R. W., Breen, J. E. and Roesset, J. M. (1985). Varma, A. H., Sause, R., Ricles, J. M. and Li, Q. (2005).
“Analysis of Nonprismatic or Hollow Slender “Development and Validation of Fiber Model for
Concrete Bridge Piers.” ACI Journal, 82 (5), 731-739. High-Strength Square Concrete-Filled Steel Tube
Suzuki, M., Ichikawa, K., Sato, J. and Ozaka, Y. (1994). Beam-Columns.” ACI Structural Journal, 102 (1),
“Evaluation of Flexural Deflection and Minimum 73-84.
Tension Bar Ratio of RC Beams.” Concrete Research Watson, S., Zahn, F. A. and Park, R. (1994). “Confining
and Technology, Japan Concrete Institute, 5 (1), 51-63. Reinforcement for Concrete Columns.” Journal of
(in Japanese) Structural Engineering, 120 (6), 1798-1824.