DISCUSSION

Disc. 106-S07/From the January-February 2009 ACI Structural Journal, p. 60

Influence of Flexural Reinforcement on Shear Strength of Prestressed Concrete Beams. Paper by Elias I.
Saqan and Robert J. Frosch

Discussion by Miguel Fernández Ruiz, Stefano Campana, and Aurelio Muttoni

École Polytechnique Fédérale de Lausanne, Switzerland

The paper presents the experimental results of a series of higher on average than the shear cracking load. The increase
nine prestressed concrete beams without stirrups failing in on the failure load with respect to the shear cracking load
flexure and in shear. Some theoretical considerations are can, in the discussers’ opinion, be explained and calculated
also proposed on the basis of a theory previously developed accounting for the different regions of Kani’s valley
by the authors with respect to shear strength in reinforced (Kani et al. 1979).
concrete members (Tureyen and Frosch 2003). The innovative Figure 12(a) shows a sketch of Kani’s valley and its two
design of the tests as well as the well-documented data governing regimes. The ascending branch (named “crack
presented by the authors have allowed the discussers to propagation” in the figure, see Point A) is due to a sudden
investigate a number of aspects with respect to shear strength propagation of a flexural crack as it develops through the
in prestressed members. A series of independent conclusions theoretical compression strut carrying shear. Such failure
and interpretations derived from this analysis may complete (disabling the teeth action as proposed by Kani) is followed
those proposed by the authors. by a total loss of load-carrying capacity of the member and
As shown by the tests of the paper, beams are developing is often named diagonal shear failure. The descending
shear (diagonal) cracking at a given load level. Such branch (named “direct struting” in the figure) has a different
cracking, however, does not lead to failure of the specimens, nature. Flexural cracks may reach the location of the theoretical
and load can be significantly increased before failure. For compression strut carrying shear and develop through it
two specimens (V-4-0 and V-4-0.93), the increase meant that (Point B in Fig. 12(a)) but they do not progress in an unstable
yielding of the flexural reinforcement was reached and manner. Instead, once such inclined cracks have developed,
bending was governing for the strength. For the other seven they can widen progressively as the load increases. A typical
specimens, failure also developed in shear, but at a load 42% crack pattern illustrating this case is plotted in Fig. 12(b)

Fig. 12—Kani’s valley: (a) failure regions; (b) crack pattern of reinforced concrete specimen
failing in shear in direct struting region; (c) idem in crack propagation region; (d) geometric
shear span in a reinforced concrete beam; (e) effective shear span for a prestressed member;
and (f) CSCT and EPSF results for Specimen V-7-2.37 compared with test results.

ACI Structural Journal/November-December 2009 907

Table 5—Shear cracking loads Vcr and failure loads VR for all specimens*
Specimen Vcr,test, kN (kips) VR,test, kN (kips) Vcr,calc (CSCT), kN (kips) Vcr,test /Vcr,calc VR,calc (EPSF), kN (kips) VR,test /VR,calc
V-4-0 205 (46.0) 244 (54.8) 200 (44.7) 1.04 204† (45.6) 1.20
V-4-0.93 265 (59.5) 334 (75.2) 250 (55.9) 1.06 300† (67.1) 1.11
V-4-2.37 296 (66.5) 367 (82.4) 301 (67.3) 0.98 367 (82.1) 1.00
V-7-0 256 (57.5) 370 (83.1) 225 (50.3) 1.14 276 (61.7) 1.34
V-7-1.84 298 (67.0) 484 (109) 295 (66.0) 1.01 354 (79.2) 1.37
V-7-2.37 305 (68.5) 428 (96.1) 307 (68.7) 0.99 363 (81.2) 1.18
V-10-0 287 (64.5) 406 (91.2) 242 (54.1) 1.19 292 (65.3) 1.39
V-10-1.51 300 (67.5) 440 (99.0) 295 (66.0) 1.02 351 (78.5) 1.25
V-10-2.37 322 (72.5) 440 (99.0) 318 (71.1) 1.01 369 (82.5) 1.19
Average 1.05 1.23
COV 0.07 0.10
*
Calculations performed with geometric and mechanical properties given in paper and assuming average values of yield strength for prestressing and ordinary reinforcement (fpy =
1690 MPa [245 ksi] and fy = 496 MPa [71.9 ksi] estimated on the basis of Mirza and MacGregor [1979]).
†
Flexural failures.

where it can be noted that inclined cracking affects only a A more accurate analysis of the results can, on the basis of
limited region of the compression strut. Failure happens the previous considerations, be performed accounting for
when bending becomes governing (for rather short shear both regimes and determining the characteristic points (refer
spans) or when the opening of these inclined cracks is such to Points B and C in Fig. 12(a)). This can, for instance, be
that the strength of the compression strut is severely limited accomplished using some theoretical models proposed by
failing in shear (Point C in Fig. 12(a)). Large scatter in the the discussers, such as the critical shear crack theory (CSCT)
strength of members governed by direct struting is typically (Muttoni and Fernández Ruiz 2008) for the crack propagation
observed. This is due to the fact that strength depends regime and the continuous elastic-plastic stress fields
primarily on the positions of the cracks affecting the (EPSF) (Fernández Ruiz and Muttoni 2007) for the direct
compression strut, which are rather random. struting regime. The results of analyses using these two theories
Looking at the experimental results of the paper, the are shown in Table 5. It can be noted that pretty accurate
discussers think that the tested specimens were, in fact, estimates of the shear force leading to crack propagation is
governed by the second type of behavior (direct struting). obtained using the CSCT, with an average value of the
This conclusion is supported by two facts: measured-to-predicted load of 1.05 and a coefficient of
1. It was experimentally observed that propagation of variation (COV) of 7%. The fitting to test results is even
shear cracks was not unstable, and load could be increased better than with the approach followed by the authors
significantly after diagonal cracking; and (Table 4), which seems to be (if discussers have correctly
understood the theory proposed by the authors) a simplified
2. The beams had a geometric shear span of 3.3 according
design formula that neglects some phenomena (such as size
to the authors. This value corresponds approximately, for
effect) that may have a significant influence on actual
ordinary reinforced beams, to the limit between unstable
strength. Regarding strength, a significant increase beyond
crack propagation and direct struting regions. In prestressed
diagonal cracking load is obtained using EPSF, with an
members, however, the effective shear span is smaller than
average increase of 20% in the failure load for specimens
the geometric one (Muttoni and Fernández Ruiz 2008). This
failing in shear (Fig. 12(f)) and predicting flexural failures for
is due to the fact that prestressing does not allow flexural
Specimens V-4-0 and V-4-0.93 (as observed in the tests).
cracks to develop close to the support region, as shown in
The estimates of the failure load are safer than for first shear
Fig. 12(e). The effective shear span aeff can be calculated in
cracking. This is logical because scatter in this failure regime is
this case (eccentric prestressing) as
significantly larger as previously discussed and this is
accounted for in the EPSF (see also value of COV).
P To conclude, the discussers would like to highlight with
a eff = a – ---z
V this discussion that shear failures have to be carefully
investigated accounting for their various failure regions.
This is particularly important in prestressed members as the
where a is the geometric shear span, P is the prestressing
amount and layout of prestressing may significantly shorten
force, V is the shear force, and z is the flexural lever arm of
the effective shear span.
the member. For preliminary estimates, this equation can be
simplified further by replacing z with d (effective depth).
REFERENCES
The theoretical model used by the authors to investigate Fernández Ruiz, M., and Muttoni, A., 2007, “On Development of Suitable
the test results seems, in the discussers’ opinion, only valid Stress Fields for Structural Concrete,” ACI Structural Journal, V. 104,
for members failing in the unstable crack propagation No. 4, July-Aug., pp. 495-502.
regime. In fact, this is what the authors are doing by Kani, M. W.; Huggins, M. W.; and Wittkopp, R. R., 1979, “Kani on
Shear in Reinforced Concrete,” Department of Civil Engineering, University
comparing their predictions to the load leading to development of Toronto, Toronto, ON, Canada, 97 pp.
of first diagonal cracking. This approach leads to reasonable Mirza, S. E., and MacGregor, J. G., 1979, “Variability of Mechanical
(rather safe) estimates for diagonal cracking load (term Vc of Properties of Reinforcing Bars,” Proceedings of the American Society of
ACI 318-08) but clearly underestimates the actual failure Civil Engineers, V. 105, No. ST5, May, pp. 921-937.
Muttoni, A., and Fernández Ruiz, M., 2008, “Shear Strength of Members
load. In this sense, it should be noted that first diagonal without Transverse Reinforcement as a Function of the Critical Shear
cracking load is not directly related to the actual failure load. Crack Width,” ACI Structural Journal, V. 105, No. 2, Mar.-Apr., pp. 163-172.

908 ACI Structural Journal/November-December 2009

Disc. 106-S07/From the January-February 2009 ACI Structural Journal, p. 60

Influence of Flexural Reinforcement on Shear Strength of Prestressed Concrete Beams. Paper by Elias I.
Saqan and Robert J. Frosch

Discussion by Andor Windisch

ACI Member, PhD, Karlsfeld, Germany.

The authors tested nine specimens to evaluate the influence lower bound estimation, Vtest > Vcalc. Nevertheless, the
of the area of prestressing and mild reinforcement on the neutral axis depth belonging to Vtest is always greater than
shear behavior of structural concrete members. Similar to the depth belonging to Vcalc, that is, the development of the
ACI 318-08 (and to the originators of the shear code neutral axis depth and of the concrete contribution goes in
provisions, Zwoyer and Siess [1954] and Sozen et al. opposite directions. This means that Eq. (4) is valid in a
[1959]), they assume that the concrete contribution of shear single point only. How can the proper neutral axis depth be
is that shear force when the principal tensile stress reaches found? Moreover, in two cases (V-4-0 and V-4-0.93), the
the tensile strength of concrete. They propose a lower bound longitudinal reinforcement yielded prior to the formation of
equation (Eq. (4)) that consists of the depth of the neutral the primary shear crack. It is not clear how the proper neutral
axis. The analysis is performed on a cracked section. It is the axis depth shall be calculated. Is this assuming yielding of
discusser’s opinion that it is contradictory that a force reinforcement or not? The authors should point out the
causing cracking be calculated on a cracked section. impact of the different bond characteristics of the
Reporting on the structural behavior of the specimens, prestressing steel (strand) and of the mild reinforcing bars.
the authors write “...the structure transitioned from behaving The authors interpret the failures as diagonal tension
as a flexure beam to a tied arch...As evident from the failure. How can an arch fail in diagonal tension? How may
cracking patterns, the structure resembled a strut-and-tie a diagonal tension failure mode be characterized with the
model.” The discusser would like to ask the following questions: neutral axis depth?
1. What do the strut-and-tie models look like in case of the The discusser appreciates that, according to Eq. (4), the
crack patterns shown in Fig. 4 at beams without transverse concrete contribution of shear is related to the neutral axis,
reinforcement, that is, without transverse ties? and that is, the concrete compression zone that must be considered
as the real source of the shear contribution. In 2009, we must
2. Can the strut-and-tie models and the tied arches coexist?
finally move away from the interpretation of the originators
Figure 6 performs the development of the neutral axis in the 1950s. The formation of the primary shear crack might
depth as a function of the moment, where the values of be a safe lower bound value, nevertheless, it is physically not
neutral axis corresponding to the measured strength at the sound. The failure patterns shown in Fig. 4 reveal unambiguously
formation of the primary crack (Vtest) and the calculated that the failures occurred as compression-shear failures of the
shear strength (Vcalc), respectively, as it is due from a fair compression zones.

Disc. 106-S07/From the January-February 2009 ACI Structural Journal, p. 60

Influence of Flexural Reinforcement on Shear Strength of Prestressed Concrete Beams. Paper by Elias I.
Saqan and Robert J. Frosch

Discussion by W. L. Gamble
FACI, Professor Emeritus of Civil and Environmental Engineering, University of Illinois, Urbana, IL.

Equation (2) is attributed to the wrong sources. The data In this implementation, Mcr is the cracking moment in
behind it are mostly in the cited two papers, but not the excess of the dead load moment, and M and V are the live
specific equation. The equation, without the Vd term, was load moment and shear at the section considered, respectively.
first published in a discussion on the ACI Committee 326 This is consistent with the 2008 ACI Code definitions.
report on shear and diagonal tension (Sozen and Hawkins The d/2 term was dropped when the equation was introduced
1962). Further information on the development of the equation into the 1971 ACI Code, probably for the sake of simplicity.
is given by Olesen et al. (1967). The consequence of this varies along a span, and it is probably
Equation (2) is actually a simplified form of the one somewhat different for cases of uniformly distributed loads
suggested by Olesen et al. (1967) and moving concentrated loads (trucks).
With respect to Eq. (4), a clarification is needed. Is c the
M cr
V ci = 0.6b w d f c ′ + -------------
- + Vd elastic or plastic (ultimate) neutral axis depth? They can be
M d quite different, especially for lightly reinforced members.
----- – ---
V 2 The discussion that cites the low axial stiffness of fiber-

ACI Structural Journal/November-December 2009 909

reinforced polymer materials would seem to imply that this as Fig. 6 shows that the neutral axis depth corresponding to
is an elastic neutral axis depth because Young’s modulus Vtest is less than the neutral axis depth corresponding to Vcalc
ordinarily is not a factor in the plastic neutral axis position. in all cases. Therefore, the measured shear that caused the
If it is elastic, it needs a different notation than c, such as kd, primary shear to form was higher than that calculated. Vcalc
which was used with allowable stress design. It would be fair was calculated based on a lower bound value of K = 5. As
to note that if c = 0.4d, which is a fairly common value for discussed in the paper, the applied load was increased in the
steel-reinforced members, Eq. (4) reverts to the current ACI 318 analytical model until the applied shear at any location along
shear value. the beam was greater than the strength calculated by Eq. (4).
Therefore, the neutral axis depth and resulting shear strength
REFERENCES are determined as part of the analysis procedure. The analysis
Sozen, M. A., and Hawkins, N. M., 1962, discussion of “Shear and procedure also determines if the reinforcement (mild and
Diagonal Tension,” by Joint ACI-ASCE Committee 326, ACI JOURNAL, prestressed) is yielded or not. It should be noted that perfect
Proceedings V. 59, No. 9, Sept., pp. 1341-1347.
bond is assumed in the cracked section analysis for both the
mild and prestressing reinforcement. Further details on the
AUTHORS’ CLOSURE analysis and design of prestressed members using this
Closure to discussion by Fernández Ruiz et al.
The authors would like to thank the discussers for their approach are available in Wolf and Frosch (2007).
kind words regarding the testing program. They provide The authors agree with the discusser that the final failure
excellent discussion and insight regarding the third stage of mode was compression-shear, and this was stated in the
behavior for loading beyond the formation of the critical paper. The classification of the failure mode as diagonal
shear crack. While this stage was not the focus of the paper tension was intended for formation of the critical shear
as previously discussed, this discussion definitely crack. While the final mode of failure was compression-
complements the work. The authors also agree that the shear and did occur at loads significantly higher than that at
effective shear span for prestressed members is different formation of the primary shear crack, there are substantial
than that of reinforced members because prestressing physical reasons why the shear strength associated with this
increases the region of uncracked concrete from the support. event was considered important in the 1950s and should be
This difference greatly influences overall structural behavior considered today. As discussed in the paper, a significant
and must be carefully considered in developing testing change in overall behavior and in particular stiffness of the
programs evaluating shear strength. beam is observed following formation of the primary shear
crack. In addition, significant widening of the shear crack
Closure to discussion by Windisch was observed, which would not be considered desirable
The authors would like to thank the discusser for his structural behavior. Finally, if stirrups were present, it is
comments and hope that this response provides clarification. expected that they would contribute significantly to the
First, it is questioned how a force causing cracking can be strength following formation of the shear crack. While
calculated on a cracked section. There appears to be confusion behavior beyond the formation of the primary shear crack is
between the shear force that causes flexural cracking at the of importance, it is for these reasons that the concrete
section under consideration, which is the basis of Eq. (11-10) contribution of shear strength, Vc , was considered at the
in ACI 318-08 (Eq. (2)), and the shear force that forms the formation of the primary shear crack rather than at the failure
primary shear crack as discussed in the paper. As explained, load of the beam.
the beams experienced three distinct stages. The first stage is
up to flexural cracking where the beam behaves linear- Closure to discussion by Gamble
elastic, and the section is uncracked. In the second stage, the The authors would like to thank the discusser for providing
beam is cracked, and calculations are based on a cracked a clearer view of the historical development of Eq. (2). With
section analysis. This stage ends at the formation of a respect to Eq. (4), this equation was introduced in an earlier
primary shear crack. Failure is assumed to occur when the paper by Tureyen and Frosch (2003), and the details behind
principal stress in the compression zone of the cracked its development are available in this paper. To clarify, c is the
section reaches the tensile strength of the concrete. This depth of the neutral axis at the section considered and is
concept was first introduced in a paper by Tureyen and calculated based on a cracked section analysis. For a reinforced
Frosch (2003). Therefore, the shear force discussed in the concrete beam, the neutral axis is calculated based on the
paper is not the shear that causes the section to crack but elastic cracked section for which it is appropriate to use c =
rather the shear force that causes a crack to form in the kd where k is a function of the reinforcement percentage ρ
compression zone of a cracked section. Although the beams and the modular ratio n. It should be noted that, along the
were able to carry higher loads beyond the formation of the beam length, the depth of the neutral axis is constant
primary shear crack, the authors believe that the behavior of considering the elastic cracked section. As mentioned by the
the beams after formation of the primary shear crack discusser, Eq. (4) reverts to the current ACI 318 shear value
changed significantly and that the concrete contribution to of Vc = 2 f c ′ bwd for the case where k = 0.4. For values of k
shear should be limited to the load-causing formation of the less than 0.4, however, Eq. (4) suggests that shear strengths
primary shear crack. In the third stage, the beams behaved as lower than that provided by Vc = 2 f c ′ bwd may result. For
a tied arch. The statement in the paper regarding resembling a a prestressed concrete beam, the depth of the neutral axis c is
strut-and-tie model was in referral to the tied arch where the not constant along the beam length but rather varies with the
compression flows from the load to the supports with the magnitude of the applied moment at the section considered.
tension reinforcement serving as the tie. In fact, the neutral axis depth c for prestressed members can
The discusser points out that in Fig. 6, the neutral axis vary significantly depending on the amount of steel present
depth belonging to Vtest is always greater than the depth (prestressed and mild) and the effective steel stress resulting
belonging to Vcalc. It appears that the discusser is confused, in significant changes in the shear strength.

910 ACI Structural Journal/November-December 2009