This action might not be possible to undo. Are you sure you want to continue?
Micromechanics of crack bridging in fibre-reinforced concrete
V. C. LI
Advanced Civil Engineering Materials Research Laboratory, Department of Civil and Environmental Engineering, University of Michigan, Ann Arbor, MI 48109-2125, USA
H. S T A N G , H. K R E N C H E L
Department of Structural Engineering, Technical University of Denmark, DK 2800, Lyngby, Denmark The stress-crack width relationship has been determined experimentally for concretes reinforced with two types of fibres, steel and polypropylene, of various fibre volume fractions. A micromechanics-based theoretical model is proposed which captures the essential features of the stress-crack width relationships at small crack widths (less than 0.3 mm). Micromechanisms accounted for include the bridging actions due to aggregates and fibres, Cook-Gordon interface debonding and fibre pre-stress. The fibre bridging action involves interfilce slip-dependent friction as well as snubbing friction for fibres bridging at inclined angles. Theoretical predictions based on independent parametric inputs compare favourably with experimental measurements of the stress-crack width relationship. Findings in this research provide confidence in the use of the proposed model for materials engineering targeted at prescribed structural performance.
1. I N T R O D U C T I O N Crack width control is of primary importance in many reinforced concrete structures, since it is believed that there is a close relationship between mean or maximum crack widths and the durability of the structure. Furthermore, when the concrete structure in question acts as a containment vessel for fluids, leakage must be minimized or might not be tolerated at all. Consequently, in the serviceability limit state a mean or maximum crack width less than about 0.1-0.3 mm is usually prescribed. It has been found [I] that there is a close relationship between the tensile load capacity of the concrete cracks (the so-called stress-crack width relationship) and the crack widths for a given loading on a given concrete structure. Furthermore, the ultimate load-carrying capacity of a cracked unreinforced concrete structure was found to depend on the tensile load capacity across the cracks, especially at small crack widths (e.g. ). For these reasons the stress-crack width relationship for small crack widths (less than about 0.1-0.3 mm) is of fundamental importance in concrete structure design in both serviceability and ultimate limit state. For example, control of the stress-crack width relationship can reduce the amount of conventional reinforcement and therefore reduce labour costs for a given acceptable crack width [33. In plain concrete, crack bridging is provided by aggregate locking action. Although the detailed physics has not been fully understood in quantitative terms, aggregate bridging action can be well represented analytically . When fibres are added to concrete, an additional bridging action is brought into effect, superimposing on the aggregate bridging effect. The combined 0025-5432/93 ~ RILEM
bridging action in a fibre-reinforced concrete (FRC) can be very beneficial to increase the stress carried across the crack. In addition, fibres have been added to concrete to create pseudo-strain-hardening [4,5]. The load transfer capacity across a matrix crack is critical in determining the amount of fibres necessary in achieving pseudo-strainhardening. In this paper, the post-cracking stress-crack width relationship (ar of the composite is studied from a micromechanics point of view, particularly with reference to fibre bridging action combined with the well-known C o o k - G o r d o n debonding effect . In addition, fibre pre-stressing and slip-dependent interfacial bond strength are accounted for in the model. Analytical results are used to predict stress-crack width relationships for steel fibre and polypropylene fibre reinforced concretes, for various fibre volume fractions. Comparisons between model predictions and experimentally measured ac-w relations confirm many of the salient features in the ac-w curves associated with the various mieromechanisms. A companion paper [7-] will describe the use of the proposed analytical model in design and structural applications of fibre-reinforced concrete.
2. CRACK BRIDGING MODEL 2.1 Aggregate bridging action
The post-peak tension-softening of concrete has been modelled as the coalescence of regularly spaced microcracks [8-10]. Although these models are convenient to use, the lack of correspondence between the real physical structure of concrete and that of the model makes it difficult to determine physical parameters. For lack of a
060 cb ~a . . 0 0 . . . The parameter p describes the shape of the softening process with increasing crack opening. . . Equation 1 is shown to describe the two concretes very well. whereas the one reinforced with steel fibres is a high-strength concrete with a~ = 5. and has been determined to be close to unity for most concrete tested to date. . . and where where a~mis the maximum bridging stress due to aggregate action at w = 0. . E m and Vmstands respectively for the Young's modulus and volume fraction of the matrix material which contains the fibres. .00E-5 4. .00E-5 8.OOE+O 2.0 MPa. . as specified in Equations 3. In deriving Equations 4 and 6. 61(LUT). .. Substitution of Equations 3 into Equation 2 and assuming three-dimensional uniform randomness for the probability density functions p(qS) and p(z) gives the pre-peak fibre bridging stress : F C ooL/3kJ'12 - 3 73] 3- . . . . . . MPo ~ oo a where P(6) is the bridging force exerted by a single fibre with an embedded length (the shorter of the two on either side of the matrix crack) of l at an orientation angle ~: P(5) = ~ [(1 + r/)Efd~5] v2 e i* for b <3o (3a) 0. ~ . Equation 1 is shown in Fig.412"c/[(1 + r/)Efdf] and tl = gfEf/VmEm. xzldf(e 3-_1 ~)o)eS 4.3*)] 2 for 3* < 3 < 1 (6) Based on the concept of debonding against a frictional strength of r and on the concept of an inclined fibre acting as a flexible rope passing over a frictional pulley against a snubbing coefficient f . . . . good physical model. The snubbing factor 9 in Equation 4 is defined in terms of the snubbing coefficient f: 2 9 -= 4 + f ~ (1 + e =y12) (5) The post-peak stress versus crack opening relationship can be derived from Equation 2 by using only Equation 3b and the result is  of(3) = eat1 . . .o~ . . In this model the aggregate bridging stress o-a is expressed as a function of the crack opening w: ~r~ - P(6) = 0 for 30 + l_< 6 (3c) where bo . . . . . z may be expressed as a function of local slippage 6'. . . (4) u O"m 1 + (W/Wo)p (1) where o"o . . . . . .Materials and Structures Model I O0 487 bridging for aggregate be expressed as a function of cracking opening 3: b 0. . . Li  derived the fibre bridging stress by integrating the individual contribution of fibres located at various centroidal distances (z) from the matrix crack and at various orientations (~b) relative to the tensile loading direction. . . . .0 MPa) used in the steel FRC series.~ = 3.07 ram). the bridging stress of may Equation 6 has been found to compare favourably with experimental data for both steel and polymeric FRC materials .00E-5 w (m) Fig. .(L. In the above. Experimental data include both the normal-strength concrete used in the polypropylene FRC series (a~ = 3.015 ram. The one reinforced with polypropylene fibres is a normal-strength concrete with ~. .80 o ~. 2. 1 Post-crack G-w relation for unreinforced concrete (maximum aggregate size = 8 mm) shown together with a fitted curve based on Equation 1.40 o~ ..4 MPa.0 MPo u O'f(6) 4Vf F~12 ~lLr. .9z v.4 MPa) and the high-strength concrete (cr~ = 5. . . .o o ~ ~ ~ ~ ~ P(6)= Plain c o n c r e t e 0 .2 and wo = 0. . and Ef stands for the Young's modulus of the fibre. 0 _< 3 _< 6 . . . . .00E-5 6. . . care must be taken to discount the contribution of those fibres which have slipped out of the matrix. . . . . Two concrete types with a maximum aggregate size of 8 mm were used in the fibre-reinforced concretes. at least for the range of w indicated (w < 0. . . . . we shall adopt here an empirical model proposed by Stang  and which fits a wide range of experimental data extremely well. This amounts to eliminating those fibres whose embedment length l is less than the crack opening 6 in the integration process in Equation 2. . .)12. . . .20 ~~ ~ O o oli . . . The parameter wo corresponds to the crack opening when the stress has dropped to half of cry. .la. . . 1 with p = 1. . . .'2)c~162 = -P(6)p(qS)p(z) dz dq5 ~d? d4b=O (2) dz=O -.' ~ o 0. . . for 3o < 3 <_ 3 o + 1 (3b) O.2 Fibre bridging action (1 + r/)Ef \ d f J corresponds to the maximum attainable value of 6o (normalized by Lf/2) for the fibre with the longest possible embedment length of Lf/2.(3 . . . In the event that the interracial bond hardens with slip distance  or weakens with slip distance [13-15].o. For a composite with fibre volume fraction Vf of fibres of length Lf and diameter d r.r= 5. .
and (b) leads to an additional crack opening •r due to elastic stretching of the fibre segment ~ in addition to that 3 associated with interface frictional debonding. hereafter labelled as the C o o k . For the present purpose. Cook and G o r d o n  predicted that a crack of finite root radius in an elastic solid under remote tensile load will create a crack tip stress field with a crack-plane-parallel tensile component which reaches a maximum at a distance of the radius of the crack tip.G o r d o n effect. Thus a matrix crack approaching an isolated fibre can cause interface debonding before the 7zdgEf (9) This P-3 relationship for a single fibre may be substituted into Equation 2 to obtain 5~g 4c~ VfEf crf (10) so that the total crack width w is approximately given as w = 3 + 3r (11) (a) with 5 related to at by Equations 4 and 6.G o r d o n effect by means of in situ SEM.488 A convenient polynomial form has been employed previously and appears to describe the experimental data for fibre pull-out reasonably well: r(3') = r o + a t 3 ' + a2(3') 2 Li. based on the assumption of an elastic bond.18] of the C o o k . For a three-dimensional uniform random distribution of fibres. respectively. 2. and is likely to depend on the inhomogeneity of the interfacial microstructure. the fibre is already in a stressed state. it would seem appropriate to employ r/0 = I. For a single fibre. typically around 0. The level of pre-stressing tr~ may be estimated from the load-sharing between fibre and matrix at the matrix cracking strain. As a consequence. it should be recognized that at the formation of a matrix crack. Because the local slippage prior to complete debonding is typically very small. since we intend to use Equation 12 for the post-crack contribution of fibres. and processing conditions. and 6cg given by Equation 10 above. For the efficiency factor related to finite fibre length. Interfacial debonding has been identified by SEM observations in fibre-reinforced ceramics to act as local events precursory to first cracking [ 16]. This C o o k Gordon effect is illustrated schematically in Fig. Laws  found that ~h may be expressed in the following form: rh = 1 emu Ef df 4Lr'c o (13) . Thus Equation 7 should be rewritten as "c(5) = Zo "C(3) = "C + a13 + a2 •2 O for for 6 < 5" 3 > 6" (8) We now consider the additional bridging compliance as a result of the C o o k . Cox  found that rt0 = 1/6 for unconstrained deformation. This leads to  0 aps = rlo?hernuE f Vf (12) w= 3+ ~cg ? (b) Fig. and Krenchel  found that q0 = 1/5 for the case when deformation is constrained to the direction of applied tensile loading. Finally. and since in this stage all bridging fibre segments crossing the matrix crack will be reorientated along the load direction. 2 The Cook-Gordon effect (a) induces fibre-matrix separation due to the tensile stress in the horizontal direction associated with the elastic crack tip field of the approaching matrix crack. where emu is the matrix cracking strain. Direct observation [17. al and a 2 must be experimentally determined for each specific combination of fibre type. This procedure allows us to calculate the fibre bridging stress af indirectly as a function of the total crack width w.02~o for cementitious materials. and prior to any crack opening. suggests that the debond zone can extend for 200 gm to over 1000 lam.G o r d o n parameter. we expected an additional displacement 6cg related to the elastic stretching of the debonded length ~. matrix. so that the local slippage is assumed to be equal to the fibre end slippage after complete debonding. the additional displacement may be estimated as follows: 4~P 3cg -- (7) The constants "co. Stang and K r e n c h e l crack tip reaches the fibre-matrix interface if the interfacial strength is adequately weak. in a steel fibre-reinforced cement paste. and qo and ql are orientation and length efficiency factors. it would be useful to simplify Equation 7 by equating 3 to 6'.
and for maximum stiffness.1 was used. 1. [23.1 Independently measured or deduced parameters The aggregate bridging parameters p = 1. Data were recorded by the testing machine and transferred to a PC.Materials and Structures Fibre prestressing may be expected to be reduced by the interfacial debonding process. The deformation was measured using two standard Instron extensometers (type 2620-602) with 12. 4. This occurs when b has reached 3*.4 MPa and for the high-strength u concrete am = 5. This fixture consists of two interchangeable steel blocks on to which the specimens were glued. When the prestressing effect is included. so that 0 ap~(w) = ap~(W* . 3 Set-up for the uniaxial tension test of FRC.w)/w* 489 for for w < w* w > w* (14) aps(W) = 0 where W* = 3 " + 6r = ao) (15) In a fibre-reinforced concrete. has been obtained directly from the peak uniaxial tensile load of these matrices. E X P E R I M E N T A L D E T E R M I N A T I O N O F STRESS-CRACK W I D T H R E L A T I O N S H I P In order to determine the stress-crack width relationship experimentally. and the prestress should be completely eliminated when the fibre is fully debonded. After the peak was reached the rate was slowly increased to 6. as shown in Fig.25 x 10 -2 mm min -1. The experimental set-up is shown in Fig.0 MPa. deformation-controlled tensile tests were conducted on notched specimens with thickness 40 mm. C O M P A R I S O N B E T W E E N M O D E L P R E D I C T I O N AND E X P E R I M E N T A L DATA 4.25 x 10 .5 mm gauge length mounted across each of the two 10 mm deep and 3 mm wide notches. and especially for a small crack opening. Technical details on testing procedure and type of adhesion used can be found elsewhere . the total composite bridging stress at(w) is then given by a~(w) = a~(w) + af(w) + aps(w) (16) with aa(w) as given by Equation 1 and af(w) and pps(W) as just described. the aggregate bridging effect and the fibre bridging effect may be assumed to operate simultaneously. The same normal-strength concrete matrix has been used in all the polypropylene FRC samples. The ultimate aggregate bridging stress a~. mixing procedures and curing can be found elsewhere for the polypropylene fibre FRC  and the steel fibre FRC . load.2 and wo = 0. To eliminate the prestressing inevitably introduced in the specimens when using conventional grips. The stress-crack width relationship was determined experimentally on two sets of FRC materials: polypropylene fibres in normal-strength concrete and steel fibres in high-strength concrete. and the same high-strength concrete matrix has been used in all the steel FRC ones. we assume that the prestress level diminishes linearly with crack opening. 3. The raw data consisted of time. width 50 mm and height 55 mm.015 mm have been obtained by fitting Equation 1 to experimental aa--W curves obtained with the two concrete matrices. displacement reading from each extensometer. The steel blocks were both fixed in advance to the frame and the crosshead of a 100 kN 6025 Instron testing machine equipped for closed-loop testing.3 mm min. As a first approximation. For the normal-strength a~ = 3. Details of recipes. material outside the crack unloaded linearly with a stiffness corresponding to the initial stiffness observed in uniaxial tension (see e. The crack opening was extracted from the total average displacement across the crack by subtracting the total elastic and inelastic deformation outside the crack. Equation 16 is used to predict composite bridging stress as a function of crack opening for two families of FRCs to be described below. This was done by assuming that the Fig. and average displacement. 3.g. a special specimen fixture was developed.24]). The maximum aggregate size was 8 mm in both concretes. The test was conducted as a prescribed deformation-rate test using the average signal from the extensometer as feedback. for improved alignment. An initial crosshead rate of 1. We make the assumption that the aggregate bridging parameters are .
10dr for steel fibre in neat cement paste. and f = 0 has been assumed for this type of fibre. Both the fibre and the concrete matrix were similar to those used in the present experiment. We have assumed here that the bond strength remains constant at ~o. 1. We note here that the polypropylene fibre used has a rectangular cross-section. are of the same order of magnitude as the parameters reported previously .2 0. The results reported [l 5] show quite large variations in the bond parameters for relatively small variations in matrix composition.0 0 (MPa ram. Stang and K r e n c h e l Interracial parameters 1:0 at a2 f (GPa) Steel Polypropylene 210 11. For both matrices E m "~" 3 0 G P a . 4a-e shows the predicted Gc-w curves for the steel FRC with Vf = 0.0 0. The values used here. From an independent test series employing continuous aligned fibres of the same kind. we expect the actual bond strength for the present FRCs (with short random fibres) processed by conventional vibration casting to be in the low end of this range. which we again assume not to be affected by the presence of the low volume of fibres used in present series of tests. with a weak dependence on Vf. The 0 ~ data and curve are included to show that the model remains valid even when the fibre volume fraction approaches zero. 4.490 Table I Fibre and interfacial parametric values used as model input Fibre types Fibre parameters Ef Lf df Li. .over point of the uniaxiat stress-strain curve and on an analytical model originated by Aveston et al. We have adopted ~ = 15dr for lack of more precise experimental evidence.5. The maximum stress at w = 0 increases by a small amount.75 has been adopted for the present model. the snubbing coefficient f. The interracial parameters required in the model consist of the interfacial bond strength v as expressed in Equation 7. The slip-dependent bond strength has been measured by Glavind  for steel fibre by means of pull-out tests. Finally. A value of 0. An average value of 0.8 MPa has been adopted for use in the present model. Because the aligned fibre specimens employ an extremely well-packed neat cement matrix processed by a pultrusion technique.72 not affected by the presence of the low volume of fibres used in the present study.048 (MPa) 4.0. and an effective fibre diameter has been employed: df 2bt b+t (18) where b and t are the width and thickness of the fibre. which would suggest that e should fall in the range of one to several tens of the fibre diameter. Bentur  has suggested ~ = 2df . Equation 18 has taken into account the change in surface contact area per fibre and the change in the number of fibres due to the rectangular cross-section for a given fibre volume fraction. we use here u O"m emu -- (17) Em where E m has been determined from the initial slope of the tensile stress-strain curve.5 and 1. It would seem not unreasonable to gauge its value from the magnitude of the distance between adjoining cracks observed to zig-zag at fibre sites . They inferred from the maximum post-peak stress of a test series by Visalvanich and Naaman  that f lies somewhere between 0. together with the experimental data obtained from the uniaxial test procedure described in section 3.0 0 0. The snubbing coefficient f for steel fibres in a concrete matrix has been studied by Li et al. and the C o o k G o r d o n parameter ~. respectively.mu for the normal-strength concrete is 0.G o r d o n parameter ~ presents the greatest uncertainty. Only the fibre volume is changed.9 (ram) 25 12 (ram) 0. although slip hardening has been previously reported  in a polyethylene fibre. The magnitude .1) 1.1. these ac-w curves are characterized by an initial load drop followed by a stress rise and a subsequent close to linear decay. Stang  inferred Vo to be in the range of 0. 0.0. For steel fibre in a more heterogeneous concrete matrix.4 to 1. Assuming a linear response up to failure. we may expect larger values of ~.017%. Values for these parameters for the steel fibres and polypropylene fibres used can be found in Table 1. . No corresponding data exist for polypropylene fibre.5 and 2. as shown in Table 1. In all five plots. There are no similar measurements available for the polypropylene fibres. All fibre parameters El. so that e.013')/o and that for the high-strength concrete is 0. In both experimental data and theoretical predictions.i) -4.2 Model results Fig. 1.0~.8 (MPa ram.75 0 (mm) 6.4 0. Lf and d~ are measurable with good accuracy. We discuss each of these parameters in the following. The only other concrete matrix parameter needed in the calculation of fibre prestressing level (Equation t2) is the failure strain emu.8 MPa. the C o o k . based on the bend. the fibre and concrete matrix data for model input are fixed as described above in the section 4.
the parametric values used in the model are fixed and described in the previous section. . i . . .. .. .. . ..50E-4 2. .00E-5 1.00E-5 1.. o6) o 2...50s .0.. O 8. . .00 o o oo ~ o Steel fibers in high s t r e n g t h V~=O. The fibre volume percentages are 0..00 .0. . 4 Comparison of ao-w curves obtained from uniaxial tension tests and from model predictions of steel FRC for (a) Vr = 0. .. 2.w curves indicates a smoother decay in comparison to those of steel fibre-reinforced concrete.00 o o 00~ o o ~ ~ o ~ o ~176 o ~ o ~o o o o ~ " ~ ' ~ . 0 . . .00E-4 (m) Model versus experimentel results 8.. . Fig.O 1 b 6. 1. ~ . .00 491 results concrete 0 13. . . . 2~ Steel fibers in high strength concrete ) Vf=O. .00E-4 1.. 0 0 .O05 4.OOE+O 5. .. . . Again. .. . . ..00 ~ o ~ o ~ o Oooo % ~ -~= Ooo o oooO 00o o _ 0 u 0 o 0 4. ..=O. . . .00 ~o~ 8 4. .. ./-o~~176176 o 0 2.00 O.. .00E-4 1.0 and 3... .. 1. and only the fibre volume fraction is changed for model input for these figures. . . . .. ~ .OOE+O 2... . r r r of the bumps for very small crack widths appears accentuated by increasing fibre volume fractions. . . .00E-4 . 5 a .OOE+O'' "5.00k~5 ' ' '1'..00 ~ 2 ~ o oOOOo~ o o o o _~9". . J .4 (b) Model o O_ 8.00E-4 (e) w (m) Fig.00E-4 0. . O. . ... 3.00E-4 2.0.. . .00 ~- ~. ..02 0. (c) V = 1.00 a0eOoo O o ~176 ~ ~ Oo o o ooo o o o o o~ 1 7 6 1 7 6 1 7 6 Q Steel fibers in high strength concrete V~= 0..3. . b 6.. . . ..50E-4 3. ~ .00 versus experimentol results z.00s 1. .~.00 ~0~0~00 %00 ~ o OO:oo %0 O0 0 0 O0 0 0 0 0 0 0 (~ O~ 0 4.OOE+O 5. . . . . .00 O..50E~4 ' ' '2. . .~ . .oo ~ o o ~o-Z. ..50E-4 3. but for polypropylene FRC.. 0 0 ' E . . . . 5 0 E . .... v G.0%. .O0 b 6. .5 and (e) V = 2. . . . . 1 7 .0. . ...00F"-'4" ' 2 . . .. . . . . .5. . .50E-4 2.00 Model 8. .00 w (m) versus experimentol results O 0. .~ o o o u ~ ~ o o o o o o o o o o ~ %~o ~ o o o ~ ~ o ~ ~o o ~8~~ o o Oo o o -Goo ~ _ o o o o ~o ~ ~ ~ o 2...00 L 4. 5.015 0. . .00E-4 concrete 2. . The general trend both experimental and theoretical. . .4 ' ' '.00E-4 O..4 . . .. .50E-4 (c) w (m) Model versus (d) experimentol results w (m) 0 O_ 8.00 (a) .00 o~OOoo~ 2 o ~ Oo .. . . .00 b 6..00 Steel fibers in high strength concrete V.00k~4 ' ' '1'. ~ o oo~ o~.d shows a similar set of data.3.00E. .~ ~ ~a o o aoD ~ o o O o ~ o 2.00 O.Materials and Structures Model versus experimentol High strength Vf=O. . . .00E-5 1 . . respectively. i .00 2. . . (b) V = 0.0.. . (d) Vf = 1.~ o ~ ~176 % . of the a c .OOE+O 5.. . . ... .50E-4 2..00E-4 2. .50E-4 2. . . ~ j o 0 % ~ O " oo o o g 0 0 n 00 000 O~ 0 0. . . . .00E-4 1.. ..00 b 6. . . . . .
.00E-4 1.00s (c) w (m) (d) w (m) Fig.. ..00 O.00 13 I0 4. 0 0 .00E-4 1. . . ...00E-4 2. ~ : ~ llil~m~oooo~. . .50s 2... . . Indeed.01 3. Steel fibres in high-strength concrete. concrete . . . . . . concrete . . ..00E-5 1.00 results concrete 0_ z. .00 versus experimentol results 5. . I .. . .00E-4 . .. .00 O. Pp. ... . .02 b 6.00s 5.50E-4 2. . I . . .50E-4 3. .00 ooooo o o oO . . . . ::":'- ~ ...492 Model ~'O 5.00 of the model Steel fibers in high strength concrete %=0.. .. fibers it" normal 400 1%=0. fibers in normol %=002 0 .. . 4. ... . .. .. . .oo ~..00 concrete b Pp. . . .. .0 and (d) Vf = 3.50E-4 3. . . .. ... .. The more gentle fibre bridging action of the lower-modulus polypropylene fibres in turn explains the smoother a~-w curves observed when testing the corresponding FRC.00E-5 1.Aggregate bridging -. . .00 Pp.. .. .OOE+O / Fiber prestress ... . the fibre bridging action and the fibre prestress (using V = 2% steel fibre as illustration). strength . . 5.o o dP ~ c~ d~ c9 c~ ~ @ o OOOoo ~ o o o o c ~ 1 7 6 1 7o o 6 o o ~ o o o oO ~ 1.. ...OOE+O 5. fibers in normal %=0.00s (m) 5. .00s (m) (a) Model 0 13.00E-5 Fig.. . .. as expected. .50E-4 W 3.. .r">-... .00 o ~ 3. These three bridging actions control the broad picture of the shape of the a c .... 6 the individual contributions of the aggregate bridging action. .... . respectively. . at smaller crack opening due to the smaller 6*. .00 ~ 1 ~ : : ~ - ooooOOO o: ooo 0o7 oo oooOOO -- oo oooo Oooo o o c 2.. . .00 Normol strength Vr=O O0 3.00 2. J .. ::: .03 0. .00 ~ 1. . . .00 o o o o o o o o o O" o" O' d 8 o o o o o o o ~ o o o ~ ^ ~ o o ~ oo -~~o~o~%~%~%-%:%%% ~ o~ o~ 0. . 6 and 14. . .. O. . . represented by r Equations 1. ..OOE+O . . .--. . . (c) Vr= 2. .OOE-4 1. .00 O. the stiffer steel fibres create a stronger bump in the ac-w curve.. . .. . : ..-~o o o o ~ o o g o ~ o % a % o ~ 1 7 6 1 7 6 1 7 6o o o - ~ 1~ 7~6 1 7o6 1 7 6o Oo ~ - o o 8L fl ~ oo%Oo: 2. . . .00 .w curves obtained from uniaxial tension tests and from model predictions of polypropylene FRC for (a) V~= 0. . . (b) l~} = 1. b 4. .7.0. 6 indicates that the elastic modulus of the fibre dominates the location of the peak (at 6*) in the fibre bridging action. strength ... . . .00E-4 0.50E-4 W 3.0. .00E-4 1. . . . .5.. ..w curves.. ...00E-4 1. .. . . ..w curves. . . . . DISCUSSION OF M E C H A N I S M S G O V E R N I N G THE crc-w CURVES To gain physical insight into the governing mechanisms of the a c .00 ~ o ~ 1 7 6 o o o o . . Thus... . .. . ] . .00E-4 2. ...~[' / ~ 0. . .0~ . .._.00 v Li.~'~~ota I response ! ] ~ .50E-4 W 3. . ::. .00E-4 (m) (b) 0.50[-4 2. . The prestressing mechanism is 0 13- Decomposition 8... . . 6 Individual contributions of aggregate bridging action and fibre bridging action to a~-w curves. .50E-4 200E-4 2. .00E-5 1. . Illustrated also is the fibre pre-stressing effect. ... .OOE+O 5.. . . .0% 5. and Fig. . .00E-4 2. .50E-4 2. 5 Comparison of a ~ ... . .. . o 0o008@~o00080~8~8~ ~ cx~Oc~oc~ OOooo o o o o o o o o o o o o o o o o o o( ooooooooooooooooooooooooo ~ 1.. ..00E-5 1..00 ~--~ .00 O 13- Model versus experimentol results 4.. 2. ... . . . . . .. we have separately plotted in Fig. I . the major difference between the ac-w relations of steel and polypropylene FRCs lies in the fibre bridging action.. Stang and Krenchel versus experimentol results ] Model versus experimentol strength "6"o_ 5.. Vr = 2.
2. V. 1992. C. Li would like to thank his colleagues at D T H for their hospitality and cooperation during his stay in Denmark. Armed with these constituent material parameters. and with four and three different fibre volume fractions each.00E-4 .. El.. Stang. C. . and Wu. or in durability design for general concrete structures.OOE+O 5 . 7 Influence of Cook-Gordon effect on the ac-w curves. 'Conditions for pseudo strain-hardening in fiber reinforced brittle matrix composites'. . Mech. Discussions with Arnon Bentur have been particularly helpful with regard to the C o o k . highly non-linear as it may be. 6.G o r d o n effect of an initial debonded segment is indirectly revealed in the Gc-w measurements. Dr Henrik Stang and Professor Herbert Krenchel acknowledge support from the Research Programme on Fibre Reinforced Cementitious Composites sponsored by The Danish Council for Scientific and Industrial Research and The Danish Ministry for Industry. 14(2) (1992) 143-154.. crack width control will be paramount in maintaining structural functionality. K. Figs 4 and 5 demonstrate that the proposed model can be applied with some confidence..00E-4 . It is clear that when the C o o k Gordon effect is absent. but specialized for handling the post-crack regime in a fibre-reinforced concrete structure.. albeit small. The influence of frictional bond decay with slippage (Equation 7) is also illustrated. for the purpose of leakage control in fluid-containing concrete vessels. .5 . E. and Leung.. ... 45(8) (1992) 390-398. to be carried out before the structure is built and put into service. This relationship is best summed up in the following form: ac = F(w. Eng. Reinhardt and A.. Thus the C o o k . with and without the C o o k . Another application of the present model.. . . Li was a Visiting Professor at the Department of Structural Engineering at The Technical University of Denmark (DTH) in June-August. (r-n) Fig. 3. 388-406. Compos. . "Theory of steady state and interface parameters) (19) where the concrete parameters consist of E m. .. W. . . Li. can be easily calculated so long as the assumptions built into the model are satisfied. V.. 1. similar to those currently used in governing the elastic or non-linear elastic behaviour of concrete. as reinforcements for a concrete composite. E. For this purpose. .G o r d o n effect. as alluded to in section 1. Again. concrete parameters. Cem. . and Aarre. 5. .. . . a~. . T. V. so that a composite with the desirable a r relationship can be designed. ..00 Or~g[no' porom:r~/~: : : :~L 2.. Fig. 'Evaluation of properties of cementitious materials'. a constraining factor in the application of fibre-reinforced concrete to practical structures is the lack of design tools to allow the structural engineer to predict the structural response. p and wo.. . . fibre and interface parameters. Friction bond decay is also illustrated. 12(6) (1986) 566-586.0%. . REFERENCES 1.. f and ~. in the form of Equation 19. With two highly different fibre types. in combination with finite-element analysis. .. the overestimated initial fibre bridging stiffness causes a rise in the curve at w _~ 0 which is not observed in any of the experimental data sets. 4. when the material is a fibrereinforced concrete without pseudo-strain-hardening..00 . Funding by D T H for the Visiting Professorship is gratefully acknowledged. .Materials and Structures 493 O r9 v &O0 Cook-Gordon a=O mm %' and slip-weakening [3. the composite ~c-w relationship. Equation 19 serves as a constitutive relation.. Further. . Li.G o r d o n effect. Vf = 2. C. Proceedings of International RILEM/ACI Workshop. Concr. . . is in the design or prediction of structural response.02 0.. H. V. and Liang. 1992) pp. and the interface parameters consist of z... the fibre parameters consist of Vf... Naaman (Spon. ASC E J. ACKNOWLEDGEMENTS This work was performed while Victor C. Equation 19. 2. Rev. Steel fibres in high-strength concrete. 1. steel and polymer. 0 0 E . 'Evaluation of crack width in FRC with conventional reinforcement' J.. 'Fracture processes in concrete and fibre-reinforced cementitious composites'. Mech.00 4. in "High Performance Fiber Reinforced Cement Composites'.. H. embodied . 7 shows the combined influence of aggregate bridging and fibre bridging. .. Li. fibre parameters. C. responsible for the definite. O... 9 b 6. C O N C L U S I O N We have presented a micromechanically based analytical model which describes the post-crack bridging stresscrack width relationship in FRC materials. . tensile strength increase with steel fibre volume fraction at w = 0. C. Stang. Appl..50E-4 2. .. J. C.00E-4 . H. would serve as an analytical tool for crack width control. Equation 19 implies that this important composite property can be engineered by controlling the concrete. edited by H. Lf and dr.50E-4 3.00 Steel fibers in high strength concrete Vf=O.. particularly those exposed to aggressive environments such as in coastal regions or parking decks. At present. Micromechanical models such as Equation 19 allow the structural designer and/or materials engineer to choose an appropriate materials receipe which will best meet the required structural performance. . Y.
A. (1992) 41-57. Copenhagen. 23.. P. 9. Pure Appl. Barsoum. Ortiz. New York. Solids Struct. F. Br. 205-219. Compos. Department of Structural Engineering. Mater. 'Experimentally Determined Mechanical Properties for Fiber Concrete' (Report Series R. ASCE J. pp. 1981). Appl. and surface treatment on synthetic fiber pull-out from a cement matrix'. in 'Fracture of Concrete and Rock'. Cox. Bentur.. E. A. PhD thesis. personal communication... L. 'Evaluation of the Compressive Behavior of Fiber Reinforced High Strength Concrete'. 'The microstructure of steel fiber-cement interface'. Concr... M. Les rOsultats de cette recherche donnent confiance dans l'utilisation du modkle proposb pour des matkriaux de gbnie civil destinds • des performances structurelles prescrites.. H.. Mater. Diamond. 25. Technical University of Denmark (1992) (in preparation. Van Mier. 282A (1964) 508-520. Compos. L. Li.. in 'Properties of Fiber Composites'. University of Illinois at Chicago Circle. 'Design and structural applications of stress-crack width relations in fiber reinforced concrete'.. 'A mechanism for the control of crack propagation in all brittle systems'. ldem. l'effet Cook-Gordon b l'interface et la prkcontrainte des fibres.3 mm). and Backer. Compos. Department of Structural Engineering. C. and Nishino. J. J. Roy. Glavind. 15 (1985) 331-342. M. S. On propose un modkle thdorique bask sur la micromkcanique. A. 28.avec diffkrentes proportions en volume de fibres. Department of Structural Engineering. P. Hasegawa.. Program Report for NSF Grant E N G 77-23534 (Department of Materials Engineering. A. edited by L. Res. in Danish). Li.. Conference Proceedings (IPC Science and Technology.494 multiple cracking of random discontinuous fiber reinforced brittle matrix composites'. Shah and S. Visalvanish.. 'Fracture process and bridging zone model and influencing factors in fracture of concrete'. E. S. 'Crack Growth and Development of Fracture Zones in Plain Concrete and Similar Materials'.. 44 (1992) 257-269. Cooper. Hansen. 'The elasticity and strength of paper and other fibrous materials'. Horii. S. Geophys. V. Y. 24. Part I: Experiments and test results'. and Wang. 'The efficiency of fibrous reinforcement of brittle matrices'. 'Indirect determination of interfacial strength parameters'. 'Modeling of fiber pull-out from a cement matrix'. Aarre. Cook. Thygesen. PhD thesis. J. Wang. 305. Stang and Krenchel 17.. Phys.. Phys. H. X. 'Post-crack scaling relations for fiber reinforced cementitious composites'. Chicago. . 26. H. Sci. D. and Mindess. Y. D. Cem. 22. S. Lund Institute of Technology. A. Technical University of Denmark (1992). 30. 1993). Kangutkar. No. 137(4) (1991) 461. E. V.427. 'Design of FRC-materials and Structures'. 'Matrix crack initiation in ceramic matrix composites. Li. H. 'Fracture modelling for fiber reinforced cementitious composites'. L'effet de pontage des fibres implique un frottement en relation avec le glissement et un frottement qui relkve les fibres se croisant d angles obliques. 24 (1988) 231-250. H. and Naaman. Stang. RESUME Microm6canique de la reprise de fissurations dans le b6ton renfore6 de fbres On a dkterminO de faqon expkrimentale la relation contrainte-largeur de fissure pour des bktons renforcks par deux types de fibres . edited by S.acier et polypropylkne . and Huang. 159. Wang. 15. 12. Krenchel. 'Tensile Characteristics of FRC with Special Emphasis on its Applicability in a Continuous Pavement'. Sci. P. Bentur. and Stang. 14. Karihaloo.. A. B. E. and Reinhardt. 'Microcrack coalescence and macroscopic crack growth initiation in brittle solids'. J. Technical University of Denmark (1992). 18. 1990) p. 'Cracking process in steel fiber reinforced cement paste'. and Gordon. and Mindess. S.. Stang.. qui prend en compte les caractkristiques essentielles des relations contrainte-largeur de fissure pour des largeurs limitkes (<0.. C. H. A.. 1991) pp. Cem. Int. T. 4 (1971) 1737-1746. 1989).. PhD thesis. From Theory to Applications'.. 10(3) (1988) 143-149. J. bundling. 1982). V. 16. La comparaison entre les prkdictions th~oriques basbes sur des pararnktres indkpendants et les mesures expkrimentales de la relation contrainte-largeur de fissure est satisfaisante. 'Fibre Reinforcement' (Akademisk. H. W. 27. Technol. 118(11) (1992) 2246-2264. Petersson. 'Fiber Reinforced Cementitious Composites' (Elsevier Applied Science. Les micrombcanismes pris en consid&ation comprennent l'effet de pontage. 29. ' Material properties'. 11. Report TVBM-1006 (Division of Building Materials. C. 67127. 21... Proc. Bentur.. 10.. 8. Soc. 1964)p. V. Swartz (Springer. M. Aveston. Lightwt Concr. 'Tensile response of quasi-brittle materials'. 449. Phys. and Backer. J. A. Department of Structural Engineering.. 21(2) (1990) 132-140. and Kelly. 6. Elfgren (Chapman and Hall. 3 (1952) 72-79. 20 (1985) 3610-3620. 'Effect of inclining angle. London. J.-E. H. Engng Mech. D: Appl. S. Li. Technical University of Denmark. C. S. G. and Krenchel. G. Hordijk. J. Civil Engng 4(I).. A. W. ASCE J. V. in preparation.. 'Single and multiple fracture'. J. Int. in preparation. K. 13. 15-24. in' Fracture Mechanics of Concrete Structures.. 19. 7. Laws. J. 20. 1971) pp. Li. M.