Académique Documents
Professionnel Documents
Culture Documents
3 REINFORCED CONCRETE
Design of steel fibre reinforced concrete using the (~-w method: principles and applications
The text presented hereafteris a drafifor general consideration. Comments should be sent to the TC Chairlady: Prof. dr. ir. Lucie Vandewalle, K.U. Leuven,
Department Burgerlijke Bouwkunde, Kasteelpark Arenberg 40, 3001 Heverlee, Belgium. Fax: +32 16 321976; e-mail:
lucie.vandewalle@bwk,kuleuven.ac,be, by 31 October2002.
2.2 Fracture mechanics for concrete (i) In the originalformulation of the FCM, the process zone and the fictitious
crack are synonyms; here however two different mechanisms are associated with
Fracture mechanics deals with the mechanics of crack the fictitious crack: micro-cracking in the process zone and aggregate interlock
formation and propagation in materials. When structures (along with fibre bridging - in the case of tension softening SFR C).
(2) According to linear elastic analysis of cracks infinitely large stresses - stress
are made from softening materials, crack formation often singularities - always exist at sharp crack tips. In linear elastic fracture
governs the structural behaviour. mechanics these singularities are characterized through so-called stress intensity
Fracture mechanics of concrete is a relatively young factors which depend only on geometry and loading. Crack propagation is
scientific and engineering field compared e.g. to fracture assumed to take place when the stress intensity factor reaches a critical value.
263
Materials and Structures/Mat~riaux et Constructions, Vol. 35, June 2002
lch EGF
= f2 (4)
The characteristic length can be interpreted as the
ratio between fracture energy per crack area and elastic
energy per volume for a given material. The characteris-
tic length can be used to characterize the brittleness of a
given material but can also be used to define a dimen-
sionless structural brittleness number B by taking the
ratio of a representative structural dimension L to the
characteristic length of the material:
Fig. 3 - A crack in SFRC and the essential features: zone with
B - Lf/ (5) fibre bridging, the process zone and aggregate interlock together
EG F with the FCM.
264
TC 162-TDF
3.00
2.00
~.~gh
Low strength concrete,
fiber content (1 vol. %)
3.00
2.00
1
/Total response
Fig. 4 - Typical stress-crack open-
ing relationships (a) obtained
from experimental measurements
on steel fibre reinforced concrete
1.QO
containing 0.3 and I vol. % of
~High strength concrete, I
1.00
hooked end steel fibres in high
ber content (0.3 vol. %)
and low strength concrete, respec-
tively. In (b) is shown a concep-
0.00 I I tual theoretical modelling of the
0.0 0.1 0.2 0.3 0.00
0.01 0.10 1.00 10.00 relationship, following [18],
w (ram) w (mm) showing the concrete and the
(a) (b) fibre contributions, (Note the log-
arithmic length scale).
265
Materials and Structures/Mat6riaux et Constructions, Vol. 35, June 2002
f, 1
f,
o.5f, p = 0.2
>
0 wI w~ w >
0 W* W
Fig. 6 - Illustration of the bi-linear stress-crack opening relationship.
Fig. 8 - Illustration of the free-form stress-crack opening rela-
tionship and the influence of the shape parameterp for a fLxed
value of w*.
3.4 Drop-constant relationship
An even more simple representation which has obvi-
ous advantages from a design point of view is the so- Crw(W) - f' (9)
called drop-constant relationship:
[ ffor
t w-- 0
(8) Again, two material parameters are associated with
O'w(W) ml(~y for W<Wma x
the model. The free-form stress-crack opening relation-
defining a residual strength crr which characterizes the rela- ship is illustrated in Fig. 8.
tionship up to a certain maximum crack opening Wmax.
In this approach a total of two (or three - counting
Wmax) material parameters are required to describe the 3.6 Characteristic values and safety factors
stress-crack opening relationship.
The drop-constant stress-crack opening relationship The analysis of SFRC structures based on the mater-
is illustrated in Fig. 7. ial parameters described above should be carried out
according to the principles known from standard con-
crete structural analysis and design.
3.5 Free-form relationship Thus, based on materials testing, an average and a
characteristic response should be identified, see below.
In previous work, [32], it has been shown that the o- Analysis in the serviceability hmit state should be based on
w relationship of the form (9) can be used to model a assumed or measured characteristic values, while analysis
large variety of SFRC materials using empirical varia- and design in the ultimate limit state should be based on
tions of the parameters w* (reference crack width) and p characteristic values modified with safety factors. At this
(shape parameter): stage only limited experience has been gained with struc-
266
TC 162-TDF
267
Materials and Structures/Mat6riaux et Constructions, Vol. 35, June 2002
268
TC 162-TDF
hl-
9 a~
-I
(N,M)
t-}"
r,,(w)
N,
from Equation (15):
N,(~) + Ns(%Xo)- Nc(IP,x0) = N (19)
Given Ip and N the position of the neutral axis can be
determined from Equation (19).
Now the moment M relative to the center line of the
cross section can be determined. The moment is given by:
f, is i
Wcmod
stress (h xol N +(h+x o 2alN +(h_alNf +My (20)
distribution v=tT-T) < tg T - T ) ' t,2 )
Fig. 12 - The analysis according to the simplified approach by where a is determined by Equations (12) and (15). Thus
Pedersen. The non-linear hinge subjected to axial force and
moment is shown along with the assumed stress distribution and
the result of the calculation is corresponding values of
the resulting forces in the different regions. angular deformation, axial force, and bending moment:
M= MOp,N ) (21)
Now, the distribution of normal stresses within the and the solution constitutes the characteristics of the
cracked zone is given by the stress-crack opening rela- non-linear hinge.
tionship, Ow(W) and the resulting force per width of the Furthermore, results are obtained for fictitious crack
cracked zone and the bending moment per width of the length a, crack mouth opening Wcmodand stresses.
cracked zone taken at the crack tip, Nf and My respec- The non-linear Equation (19) can be solved for x 0
tively, can be obtained by integrating: using a simple numerical iteration technique, e.g. bisec-
tion. Furthermore, the involved integrations in
Ul = 4f.o ~176r (13) Equations (13) and (14) can be carried out using a
~o numerical integration scheme allowing for the use of any
stress-crack opening relationship Ow(W).
1
- jo',Oo',.wt.juau (14)
Simplifiedapproachby Casanova
Casanova and Rossi, [7], proposed a model which
It is now required that the crack opening angle corre- adopts the assumptions shown on Fig. 11 (a) with respect
sponds to the total angular deformation of the non-linear to the angle at the faces of the non-linear hinge and the
hinge. With the given assumptions the depth of the tensile crack opening angle. In the model it is assumed that the
zone, h - x0, (see Fig. 12) in the non-linear hinge may now angle formed by the crack varies according to the crack
be related to the crack mouth opening by: mouth opening and the depth of the crack:
h - X o = l ( f t S+Wcmod) (15)
q~= Wc'~ (22)
a
q)\E
The length s of the non-linear hinge in Casanova's
where s is the length of the non-linear hinge. model varies with the depth of the crack such that:
The resulting force per width of the elastic compres-
sion zone is denoted N< and the resulting force per width s = 2a (23)
of the elastic tensile zone is denoted Nt: The final assumption is related to the internal kine-
matics of the hinge. Two curvatures are considered, the
Uc = ~pEx~ (16) elastic curvature of the un-cracked part of the hinge, K1,
2s and the curvature in the cracked zone, ~c2. The elastic
curvature is given by:
N,= (f')2s (17) 12M (24)
2~pE 1~ 1 -- Eh 3
Equation (16) describes how N< is related to Ip and x0: where M is the moment per unit width in the beam.
N< ({P,Xo).Equation (17) describes how N t is related to The curvature in the cracked zone is given by:
qJ:Nt (q~). Equation (13) describes how Nf is related to ~p
and Wcmod:Nf ((p, Wcmod). le2 = -~c- (25)
Denoting the resulting axial force per width of the xo
element N, for a given angular deformation ~p, the equi- where ec is the strain at the extreme fibre in compres-
librium of the section is written in the following way in sion, and x 0 is the depth of the neutral axis at the crack.
269
Materials and Structures/Mat~riaux et Constructions, Vol. 35, June 2002
I
Equation (10)) of the hinge by assuming a parabolic vari- 41 f,
ation of the curvature along the non-linear hinge: ~I~
a ili aw(w)
2~:1 + K2 (26)
]~m -- -
3
-
270
TC 162-TDF
Phase
0 1 II II1 g = 4(1_30t +3or a 1 _a___~3qt
2) 0 (41)
a= l-q*: - 1--~-b- l-V2 -V2+? (40) Application of the non-linear hinge in structural
20 0 1 14/2) analysis
Structural calculations of e.g. beams, [6, 7, 26, 28, 33-
together with: 35], beams on elastic foundation [26] and pipes, [28], can
be done by introducing a non-linear hinge in the struc-
ture, prescribing the angular deformation on the non-lin-
ear hinge, using the non-linear hinge solution to solve for
the generalized stresses at the non-linear hinge, establish-
271
Materials and Structures/MaMriauxet Constructions, Vol. 35, June 2002
ing the linear elastic solution for the structure given the able. However, attention should be drawn to the recent
generalized stresses at the non-linear hinge and finally formulation of a so-called adaptive hinge model [25],
solving for the applied load and the total deformations. As which essentially is a non-linear hinge (section 4.3) that
an example, see section 5 and the appendix in case of a takes into account both the propagation of the fictitious
simply supported beam subjected to 3-point loading. crack and the de-bonding between re-bars and SFRC.
The length of the non-linear hinge s has to be con- The reason for calling the model the adaptive hinge is
sidered a fitting parameter for the calculations. In general that the width of the hinge adapts itself to the de-bonded
the optimal length will depend on the type of structural region on both sides of the bending crack. The model
element. The hinge width s in a beam has previously also provides information about crack opening and aver-
been assessed ([28, 35]) using non-linear hinge models age curvature of the adaptive hinge. However, the model
for plain as well as fibre reinforced concrete and it has has not yet been experimentally verified. Therefore,
been shown that s = h/2 is an adequate choice. until such information is available, crack spacing has to
Comparisons between the explicit model by Olesen, the be evaluated based on reasonable assumptions. The for-
simplified approach by [28] and the approach by [6] for mulation proposed in Eurocode 2, [10], for the average
the bi-linear stress-crack opening relationship show very crack spacing and described in Test and Design Methods
similar results. Furthermore, all models compare favor- for Steel Fibre Reinforced Concrete. Recommendations
ably with a non-linear FEM analysis using the same bi- for the r~-a design method [2] gives an upper limit.
linear stress-crack opening relationship, see the appen- Experimental evidence indicates that crack spacing in
dix. This seems to indicate that, for a large number of SFRC members smaller than the member depth is
stress-crack opening relationships, the simplified observed, both in flexure and in direct tension. It is
assumptions that the fictitious crack faces remain plane therefore proposed to adopt the following assumptions
and that the deformation of the non-linear hinge is pri- for crack spacing, srm:
marily due to the crack opening are reasonable.
Srm= 50 + 0.25klk 2 q~ (49)
9r
4.4 Cross-section with conventional where (~is the bar diameter in mm whereas 9, the effec-
reinforcement tive reinforcement ratio, and parameters k 1 and k2 are
defined in [10].
The models presented in Section 4.3 were aimed at The relationships presented in Section 4.3 for unre-
predicting the behaviour of cross-sections reinforced inforced concrete are applicable to members reinforced
only with fibres. In applications where conventional with conventional reinforcement. In this case the length
reinforcement is used with fibre concrete, additional of the non-linear hinge defined in Section 4.3 still
assumptions related to the length of the non-linear applies to determine the response of the section.
hinge, the stress in the reinforcement at the crack, and However in computing the deflection, [20] showed that
the average curvature must be adopted. more than one crack may govern the member behaviour
In presence of reinforcement, SFRC members and that a length equal to the member depth but not less
exhibit multiple cracking until reinforcement yielding. than s or not less than the value given by Equation (49)
At that point, one crack generally governs the member should be used. They also showed that adopting the
behaviour due to the softening nature of fibre concrete. variable non-linear hinge length presented by [7] leads to
The assumptions must enable modelling the member reasonable results as compared to experimental results.
response at small and moderate crack opening. At small
crack opening under service loads, both the maximum Reinforcement strains in the cracked region
crack width and the stress level limits in the reinforce- Adopting Eurocode representation of the interaction
ment can have a significant importance (e.g. for durabil- between concrete and reinforcement allows us to con-
ity purposes or for the fatigue of the reinforcement, [5]). sider reinforced SFRC member behaviour as a weighted
In general, the ultimate load of a conventionally rein- average between the cracked and uncracked regions.
forced SFRC member corresponds to the situation when Therefore the strain in conventional reinforcement in
the conventional reinforcement yields. At this stage the the cracked zone is given by:
fibre contribution to the load carrying capacity can be
significant. In both cases it is therefore essential for the
assumptions adopted to be realistic and representative of
the actual behaviour. where d is the position of the reinforcement with respect
to the compression face. This assumption is applicable
Crack spacing with the simplified methods presented in Section 4.3.
In direct tension and in flexure, the observed crack
spacing in the presence of fibres is less than in identical Average curvature
members without fibres [9, 22, 36]. Although formula- The average curvature can be determined according
tions have been proposed to determine crack spacing in to the recommendation [10], averaging the curvature at
the presence of fibres [22], further research has to be car- the crack and between cracks:
ried out in this area before a general formulation is avail-
272
TC 162-TDF
I
"crack ;L J 5. BEAM ANALYSIS: LOAD-
DEFLECTION BEHAVIOR
273
Materials and Structures/Mat6riaux et Constructions, Vol. 35, June2002
12sM (56)
(Pe- Eth 3 --
while cp is the total angular deformation of the non-lin-
ear hinge prescribed in the cross-sectional analysis. The .,: z d
deflection due to the rigid-body rotation ~ may be writ-
ten in the form:
r 450
!
L (57) Wm
Uc
In the elastic case the deflection is given by u = u e. Fig. 17 - Assumed crack geometry in a beam with conventional,
When a crack has developed the deflection is given by u longitudinalreinforcementloadedto the ultimate shear loading
capacity,VRd3. The crackis assumed to extend under 45~ and the
= u e + u c. P is related to L and M by: crack opening at the re-bar is limited to w m.
p=4M(~)
L (58) Considering a rectangular cross-section with width b,
effective depth d (distance from the top of the beam to
In this case the elastic deflection can be obtained the reinforcing bars) and inner lever arm z = 0.9d, see
from Equations (54) and (58). Fig. 17, the fibre contribution Vf,d is calculated from the
The entire load deflection diagram for a beam in design stress-crack opening relationship Ow,d(W) in the
three point bending can be determined by prescribing following way:
angular deformations cp > %rackof the non-linear hinge,
calculating the load on the beam from Equation (58) and Vf, d = bZ-~p,d(Wm) (61)
the two parts of the deformation from Equations (54)
and (57). It follows that the load which initiates cracking with:
Pcrackis given by:
a( w ]= ~W m J Ofwm ow ,a(U) d u (62)
P~,~ck= f, 2h2---~t (59) p, \ m]
3L
The quantity -~p,d(Wm) is called the mean design resid-
In Equation (58) any of the moment-rotation rela-
tionships described above can be applied. Furthermore, ual stress at the crack width w m and represents the mean
it was shown in [34] how the above analysis can be value of the post-cracking stress between zero and wm.
extended to cover the case of a beam with a notch. A definition of w m is necessary to quantify the ulti-
Alternative approaches based on integration of the cur- mate load-carrying capacity of the beam failing in shear.
vature along the length of the beam can be adopted, [20]. Experimental studies carried out with different geome-
tries of steel FRC beams, reinforced with conventional
longitudinal re-bars, have analyzed the onset of inclined
cracks and the formation of concrete struts in compres-
6. SHEAR CAPACITY: ULTIMATE LIMIT STATE sion [6]. According to the results, the spacing of these
cracks is roughly equal to the inner lever arm of the
The shear capacity of FRC beams with conventional beam and the ultimate crack opening is proportional to
longitudinal reinforcing bars has been analyzed exten- the height of the beam. Since the crack opening is con-
sively in the literature [7] by considering the failure to trolled by the longitudinal reinforcement, it is proposed
occur due to crack propagation along known planes. that the maximum crack opening be taken as:
Only such cases will be considered in this section since w m = s,z (63)
there is no generally accepted method for the determina-
tion of the shear capacity of FRC elements without con- where 8s is the strain of the longitudinal reinforcement.
ventional reinforcement. The approach of [7] to calcu- Since Vf,d typically decreases with an increase in the
late the contribution from the fibres is described in the maximum ~rack opening, the fibre contribution should
following. be determined for a maximum allowable crack width. If
Following Eurocode 2, [10], the ultimate shear load the maximum strain in the longitudinal steel re-bar is
carrying capacity VRd3 is taken to be the sum of the con- taken to be 1%, then wm should be taken as 0.009d.
tributions of the member without shear reinforcement In order to obtain an equivalence relationship
Vc, a, of the stirrups and/or inclined bars, Vw, a, and the between conventional transverse reinforcement and
274
TC 162-TDF
w=-2u(l)=sinh()d)= -1 (71)
275
Materials and Structures/Mat~riaux et Constructions Vol.35 June2002
o151 I I I o.1~ I I I
.
I
f t = 3 MPa ~. ~ = 4.5 MPa
E = 30 GPa ' ~ E=45 GPa
Z; = 3 MPa/m ~ L - ~ = 3 MPa/m
0.10- I
0.10
I ~ = o MPahn
Fig. 19 - C r a c k opening
0.05 0.05
divided by thickness o f
slab versus residual
strength for two different
strength levels,
(a)ft = 3 MPa,
0.00 0.00
0.0 1.0 29 3.0 0.0 1.5 3.0 4.5
(b)ft = 4.5 MPa. Applied
Residual stress (MPa) Residual stress (MPa) parameter values:
h = 0.12 m, % = 0.003 MPa,
(a) (b) = 0 and 3 M P a / m .
This equation denotes the equilibrium path, and it residual strength to tensile strength ratio, larger values of
states the necessary stress to act on the crack surface in the crack opening are encountered when the tensile
order to ensure equilibrium of the slab at a certain crack strength is increased.
opening w. The stress-crack opening relationship for the A full scale test verifying the approach presented is
slab material is a function of w and is written as Ow(W). described in [27] and experimental determination of the
Thus, the crack opening which ensures equilibrium is slab-sub base interface parameters z0 and ~ can be found
established by the equation: in [24].
~ /73/
8. C O N C L U S I O N S A N D DIRECTIONS FOR
This equation represents the general solution to the FUTURE WORK
problem and is valid for any stress-crack opening rela-
tionship. Introducing the drop-constant stress-crack In the present document the stress-crack opening
opening relationship Equation (8), the following expres- relationship, Ow(W), is applied for designing steel fibre
sions for the equilibrium crack opening may be deduced: reinforced concrete structural elements involving cross-
sections subject to combination of axial force, bending
moment and shear force with or without conventional
2 -~-[~) -1 for ~ , 0 reinforcing bars. Furthermore shrinkage of slabs on
W grade is treated. The cross-sectional analysis with com-
(74) bined m o m e n t and axial force involves two main
h
assumptions regarding:
-~-2o[ ~ ) for ~=0 1. The kinematic behaviour of the cracked cross-section.
2. Representation of the stress-crack opening relationship.
When using Equation (74) it should be verified that With respect to the first assumption, numerical cal-
the predicted crack opening w is less than Wmax defined culations reveal that the overall structural response is
in the drop-constant stress-crack opening relationship insensitive to the kinematic assumptions.
applied (Equation (8)). With respect to the second assumption, the results of
Once the crack opening has been determined the the uniaxial tensile test should guide the choice of the
corresponding crack spacing may be found via Equation representation of the stress-crack opening relationship.
(71). The crack spacing will be in the interval [l;2/]. Note Since fibre orientation can change from one struc-
that the crack spacing increases with increasing w, thus, tural application to another, care must be taken to ensure
if we have large cracks they will be far apart, on the other that the chosen relationship is representative of the
hand, if we have narrow cracks they will be close material in the structural application under considera-
together. tion. In general, the more representative the stress-crack
In Fig. 19 the normalized crack opening w/h is shown opening relationship is, the more comprehensive the
as a function of the residual strength o .. Two different prediction of the structural behaviour will be.
slab materials are compared: one with a tensile strength The design principles set forth in this document have
of 3 MPa, the other with a 50% higher strength. The been applied and verified in a number of studies, both
strength to stiffness ratio is assumed to be constant. It is experimental and numerical. Elements of experimental
observed that if the residual stress ~. is the same in the
9 . J/
verification can be found in [6, 7, 20-22, 28].
two cases then the crack openxng xs larger when the ten- Further work is needed in the following areas:
sile strength is higher. Note also that even for the same 1. Further analyze situations involving mixed mode
276
TC 162-TDF
crack propagation and opening (shear, torsion etc.). and design of fibre reinforced concrete structural members con-
2. Assess the sensitivity of the choice of analytical repre- taining or not conventional reinforcement', Technical Report
sentation of stress-crack opening relationships. Report EPM/GCS-2001-10, Department of Civil, Geological
and Mining Engineering, l~c01e Polytechnique de Montr&l,
3. Analyze the possibility of using the bending test as 2001. In press.
quality control in connection with the Ow(W) design [21] Massicotte, B. B~langer, A. and Moffatt, K., 'Analysis and design
method. of SFRC bridge decks' in Proceedings of the Fifth International
4. Assess time dependent effects such as fatigue, creep RILEM Symposium Fibre-Reinforced Concretes (FRC),
and durability. BEFIB' 2000 (P. Rossi and G. Chanvillard, Editors), RILEM
Publications S.A.R.L. (Cachan, France) (2000) 263-272.
[22] Moffatt, K., 'Calcul des dalles de pont avec b&on de fibres
d'acier', Master's thesis, l~cole Polytechnique de Montr&l, 2001.
BIBLIOGRAPHY [23] Nanakom, P., Horii, H. and Matsuoka, S., 'A fracture mechan-
ics-based design method for SFRC tunnel linings',J. Materials,
[1] RILEM Technical Committee TDF-162 'Test and design meth- Conc. Struct., Pavements30 (532) (1996) 221-233.
ods for steel fiber reinforced concrete. Recommendations for [24] Olesen, J. F., 'FRC slabs on grade, full-scale tests', 1999.
bending test' (Chairlady L. Vandewalle), Mater. Struct. 33 (225, Department of Structural Engineering and Materials, Technical
January-February 2000) 3-5. University of Denmark, MUP2.
[2] RILEM Technical Committee TDF-162 'Test and design meth- [25] Olesen,J. F., 'Cracks in reinforced FRC beams subject to bending
ods for steel fiber reinforced concrete. Recommendations for (y- and axial load', in 'Fracture Mechanics of Concrete Structures' (R.
e design method' (Chairlady L. Vandewalle), Mater. Struct. 33 de Borst, J. Mazars, G. Pijandier-Cabot and J. G. M. van Mier,
(226, March 2000) 75-81. Editors), A. A. Balkema Publishers (2001) 1027-1033.
[3] RILEM Technical Committee TDF-162 'Test and design meth- [26] Olesen, J. F., 'Fictitious crack propagation in fiber-reinforced
ods for steel fiber reinforced concrete. Recommendations for concrete beams', Journal of Engineering Mechanics 127 (3) (March
uni-axial tension test' (Chairlady L. Vandewalle), Mater. Struct. 2001) 272-280.
34 (235,January-February 2001) 3-6. [27] Olesen, J. F. and Stang, H., 'Designing FRC slabs on grade for
[4] Bart0s, P.J.M. and Duris, M,, 'Incfined tensile strength of steel temperature and shrinkage induced cracks', in Proceedings of the
fibres in a cement-based composite', Composites 25 (November Fifth International RILEM Symposium Fibre-Reinforced
1994) 945-951. Concretes (FRC), BEFIB' 2000 (P. Rossi and G. Chanvillard,
[5] B~langer, A., 'Conception de da!les de ponts avec armature Editors), RILEM Publications S.A.R.L. (Cachan, France) (2000)
r~duite et b&on de fibres d'acier', Master's thesis, t~cole 337-346.
Polytechnique de Montr&l, 2000. [28] Pedersen, C., 'New production processes, materials and calcula-
[6] Casanova, P. and Rossi, P., 'Analysis of metallic fibre-reinforced tion techniques for fiber reinforced concrete pipes', PhD thesis,
concrete beams submitted to bending', Mater. Struct. 29 (1996) Department of Structural Engineering and Materials, Technical
354-361. University of Denmark, Series R, no. 14, 1996.
[7] Casanova, P. and Rossi, P., 'Analysis and design of steel fiber rein- [29] Rossi, P., 'Steel fiber reinforced concrete (SFRC): An example of
forced concrete beams', A CI Structural[. 94 (5) (!997) 595-602. French research', ACIMat.J. 91 (3) (May-Jtme 1994) 273-279.
[8] Cotterell, B. and Mai, Y. W., 'Fracture Mechanics of [30] Rossi, P. and Harrouche, H., 'Mix design and mechanical
Cementitious Materials', Blackie Academic & Professional, behaviour of some steel-fiber reinforced concretes used in rein-
Wester Cleddens Road, Bishopbriggs, Glasgow G64 2NZ, 1996. forced concrete structures, Mater. Struct. 23 (1990) 256-266.
[9] Dzeletovic, N., 'Propri&& des dalles de ponts avec b&on de [31] Shah, S. P., Swartz, S. E. and Ouyang, C., 'Fracture Mechanics
fibres', Master's thesis, i~cole Polytechnique de Montrdal, 1998. of Concrete',John Wiley & Sons Inc., New York, 1995.
[10] ENV-1992-1-1. Eurocode 2: Design of concrete structures - [32] Sung, H. and Aarre, T., 'Evaluation of crack width in FRC with
Part 1: General rules and rules for buildings. Technical report, conventional reinforcement', Cement & Concrete Comp. 14 (2)
1992. European pre-standard. (1992) 143-154.
[11] Hillerborg, A., 'Analysis of fracture by means of the fictitious [33] Stang, H. and Olesen, J. F., 'On the interpretation of bending
crack model, particularly for fibre reinforced concrete', The Int. J. tests on FRC-materials' in Proceedings FRAMCOS-3, volume
Cem. Comp. 2 (4) (1980) 177-184. 1, Fracture Mechanics of Concrete Structures, (H. Mihashi and
[12] Hillerborg, A., Mod~er, M. and Petersson, P. E., 'Analysis of K. Rokugo, Editors) Aedificatio Publishers (D-79104 Freiburg,
crack formation and crack growth in concrete by means of frac- Germany) (1998) 511-520.
ture mechanics and finite elements', Cem. Concr. Res. 6 (6) [34] Stang, H. and Olesen, J. F., 'A fracture mechanics based design
(1976) 773-782. approach to FRC' in Proceedings of the Fifth International
[13] Karihaloo, B. L., 'Fracture Mechanics and Structural Concrete', RILEM Symposium Fibre-Reinforced Concretes (FRC),
Concrete Design and Construction Series, Longman Scientific & BEFIB' 2000 (P. Rossi and G. Chanvillard, Editors), RILEM
Technical, Harlow, Essex, England, 1995. Publications S.A.R.L. (Cachan, France) (2000) 315-324.
[14] Kitsutaka, Y., 'Fracture parameters by polylinear tension-soften- [35] Ultkj0er,J. P., Krenk, S. and Brincker, R., 'Analytical model for
ing analysis',J. Eng. Mechanics123 (5) (1997) 444-450. fictitious crack propagation in concrete beams', ASCE, J. Eng.
[15] Kullaa, J., 'Fibre-reinforced concrete under uniaxial tension', Mech. 121 (1) (1995) 7-15.
Nordic ConcreteResearch1]1994(14) (1994) 77-90. [36] Vandewalle, L., 'Cracking behaviour of concrete beams rein-
[16] Lange-Kombak, D. and Karihaloo, B. L., 'Design of fiber-rein- forced with a combination of ordinary reinforcement and steel
forced DSP mixes for minimum brittleness', Advn. Cem. Bas. fibers', Mater. Struct. 33 (April 2000) 164-170.
Mat. 7 (1998) 89-101.
[17] Li, V. C., 'Post-crack scaling relations for fiber reinforced
cementitious composites', ASCEJ. Mat. Civil Engng. 4 (1)
(1992) 41-57.
[18] Li, V. C., Stang, H. and Krenchel, H., 'Micromechanics of crack
bridging in fiber reinforced concrete', Mater. Struct. 26 (162)
(1993) 486-494.
[19] Maalej, M. and Li, V. C., 'Flexural strength of fiber cementitious
composites', ASCEJ. of Mat. in Civil Eng. 6 (1994) 390-406.
[20] Massicotte, B., Moffatt, K. and Bastien, D., 'Behaviour, analysis
277
Materials and Structures/Mat6riaux et Constructions, Vol. 35, June 2002
,V
met of cracking
curvature over the non-linear hinge, were adopted. The 10H
deflection of the beam was obtained through the
numerical integration of the moment curvature response
in the models of [7]and [20], while the approach out-
lined in section 5 was adopted in the calculations based
on the models of [28] and [26].
Also a non-linear fracture mechanics analysis has 0 I I I i
0.0 0,2 0.4 0.6 0.8 1.(
been carried out using the finite element program
DIANA with discrete interface elements. The finite ele- Deflection (ram)
ment program is able to handle a bi-linear stress-crack Fig. 20 - The results of the calculations of the load-deflection and
opening curve. wcraoa-deflecdon diagrams for the short beam. The analytical
The analysis has been carried out on a beam with results for the load-deflection diagram are shown using the mod-
cross sectional dimensions h = 150 m m and b = 150 mm. els of [7, 20, 26, 28]. Shown is also a FEM analysis using the dis-
Two different spans were considered: L = 500 m m and crete crack approach with the program DIANA. In general the
analytical results fall slightly below the results of the FEM analy-
L = 1000 ram. In all types of analysis the same bi-linear sis. The results using the models of [20, 26, 28] are so close that
stress-crack opening relation (7) with the parameters the curves cannot be distinguished. The Wcmod-deflection diagram
given in Table 1 was used. Young's modulus E was is shown only for the FEM analysis and the model of [28]. In this
assumed to be 35 GPa. case the FEM analysis and the analytical model give very similar
results.
I
(MPa) (MPa/mm) (MPa) (MPa/mm)
15, I I 1.2
3 30 1.5 0.15 I
i.0
The results of the calculations are given in Pigs. 20
and 21, where both the P(u) and the Wcmod(u) are shown
for both beams. Overall the agreement between the dif- 0.8
ferent methods is good. As shown it is practically impos-
sible to distinguish between the three non-linear hinge
approaches of [20, 26 and 28]. The FEM analysis shows
good agreement with the non-linear hinge approach.
i M assicote etal. (2001)
Casanova & Rossi (1997)
0.6
0.4
278