Cwwtt and Concrete Composites 20 (199X) 437-453

0 199X Elsevier Science Ltd. All rights reserved.

09%9465/98/$ - see front matter

PII:SO958-9465(98)00026-2
ELSEVIER

Plasticity Approach to Shear Design

L. C. Hoang” & M. P. Nielsen
Technical University of Denmark, Department of Structural Engineering and Materials, 2800 Lyngby,
Denmark

Abstract fYW yield strength of shear reinforcement

h depth of beam
The paper presents some plastic models for shear hi lever arm, i.e. the distance between the
design of reinforced concrete beams. Distinction is tensile and the compression resultant
made between two shear failure modes, namely P external load
web crushing and crack sliding. The first men- PC, cracking load
tioned mode is met in beams with large shear P” ultimate load/load-carrying capacity
reinforcement degrees. The mode of crack sliding stirrup distance
is met in non-shear reinforced beams as well as in i(h) size effect parameter
lightly shear reinforced beams. For such beams the U relative displacement in yield line
shear strength is determined by the recently V shear force
developed crack sliding model. This model is WI internal work at failure
based upon the hypothesis that cracks can be W internal work at failure per unit length
transformed into yield lines, which have lower x Horizontal projection of yield line/criti-
sliding resistance than yield lines formed in cal diagonal crack
untracked concrete. Good agreement between
theory and tests has been found. 0 1998 Elsevier Greek letters
Science Ltd. All rights resewed.

Keywords: design, plasticity, reinforced concrete, angle between yield line and displace-
reinforcement, shear, web crushing ment direction
angle of friction
effectiveness factor for web crushing
NOMENCLATURE effectiveness factor for non-shear rein-
forced beams and lightly reinforced
a shear span/half of the span length beams
alh shear span ratio effectiveness factor for crack sliding
AS longitudinal reinforcement area inclination of the diagonal compression
A SW total cross-sectional area of one stirrup field
in a horizontal section reinforcement ratio ( = A,/bh)
b web width of beam shear reinforcement ratio ( = AJbs)
C cohesion of concrete concrete stress
cohesion in a crack shear stress
;:’ uniaxial compressive strength of con- = 0.059\&
cre te shear capacity/shear strength of cracked
f tef Effective plastic tensile strength of con- concrete
crete
shear reinforcement degree ( = pvfyWlfc)
*To whom correspondence should be addressed. minimum shear reinforcement degree
438 L. C. Hoang, M. P Nielsen

INTRODUCTION strength is \ifo where 1’is an effectiveness factor

and fc is the standard uniaxial cylinder strength.
The application of plastic theory to shear in The reinforcement material is assumed to be
reinforced concrete has now been studied for rigid, perfectly plastic as well with the yield
more than 30 years. All important problems strength fY. Furthermore, the stirrups are nor-
have been touched upon. mally assumed to be closely spaced so that their
For many years the application of plastic action may be described by an equivalent
theory in practical design was restricted to Den- stirrup stress.
mark and Switzerland. Now the interest in other The plastic approach for design purposes was
countries is growing and plastic methods have developed as a lower bound approach, i.e. it
been adopted in the Eurocode 2, the future operates with statically admissible and safe
common concrete code in Europe. stress fields. However, in some cases, coinciding
The problem to be overcome is not to find upper and lower bound solutions have been
the plastic solutions, although this is not always found, see Ref. 3.
easy. The difficult problem is the pronounced
softening of unconfined concrete which renders The lower bound solution for constant shear
the perfectly plastic material model a rather force
crude one. What has been shown in the past
decades is that by introducing empirically deter- Consider a simply supported beam symmetric-
mined reduction factors, the so-called ally loaded by two concentrated forces as shown
effectiveness factors, on the concrete strengths in Fig. 1. The compression zone and the tensile
the theoretical solutions may be brought into zone are idealized as stringers carrying the
close agreement with experiments, forces C and T respectively. The shear force V
We might say that the plastic theory is an is carried by the web with the thickness b. The
extremely effective way of finding empirical for- depth of the web is the distance between the
mulas for the strength of a concrete structure stringers. In design situations, as well as when
subjected to shear. comparing with test results, the depth has been
The present ideas about the reduced strength taken as the internal moment arm hi at maxi-
of concrete in a concrete structure are mum moment.
described in the paper. Further, some important The shear force V gives rise to an average
solutions from a practical point of view will be shear stress z in the cross-section given by
given. They will mainly be related to beam
problems. V
z= - (1)
hhi

THE WEB CRUSHING CRITERION A statically admissible stress field in the web
region is easily constructed. The concrete is
The basic idea of plastic theory for reinforced assumed to be in a state of plane stress and
concrete beams was formulated in 1967.’ subjected to uniaxial compression with the com-
During the 197Os, further development of the pression stress cc at an angle 0 with the
theory took place resulting in, among other horizontal x-axis (the diagonal compression
things, the establishment of the upper bound field). The horizontal components of this stress
technique. field are transferred to the stringers and the
A detailed description of the theory may be vertical components are carried by the vertical
found, for example, in Refs 2 and 3. Readers stirrups, see Fig. 1.
with interest in the historical aspects of the If we assume the stringers to be sufficiently
developments are referred to Ref. 4. In this strong, the shear capacity of the beam will be
paper, only a brief summary will be given of the exhausted in the following way: at a certain load
earlier work. level the stirrups will reach the yield stress fYw
In the plastic approach, concrete is assumed (it is assumed that the beam is not overrein-
to be a rigid, perfectly plastic material obeying forced with stirrups). Once the stirrups are
the modified Coulomb failure criterion with the yielding, increasing shear force can only be
associated flow rule. The tensile strength is nor- carried by increasing the compression stress cr,
mally put to zero and the effective compressive which at the same time rotates to smaller angles
 Plasticity approach to shear design 439

h hi

Y
t Compression stringer ,

Tensile stringer
---Wl--s+S+Ss-
Fig. 1. Diagonal compression field in the web.

0. The concrete stress gc may increase until it

- = Jgq I
reaches the effective compression strength vfc
and the beam finally fails by crushing of the web z
concrete. (5)
vfc 1
The shear stress, which can be carried if the -
stirrups are allowed to yield, is \2

The formulae in eqn (5) are known as the web

7 = $fc cot II (2) crushing criterion.
The inclination of the diagonal stress field at
Here, the degree of shear reinforcement $ is failure is given by
defined as
(6)

(3) The shear capacity z/$ versus $1~ is depicted

in Fig. 2.
This lower bound solution is not the exact
A,, being the total sectional stirrup area of one solution in the whole $/v range. The solution
stirrup in a horizontal section and s is the must be modified for small $/v values depend-
stirrup distance. ing on the shear span. For a discussion of the
The following shear stress can be carried if complete solution, see Ref. 3.
the concrete is allowed to ‘yield’: Based on the results of a large number of
tests, see Ref. 3 for references, it was found
that the effectiveness factor v with sufficient
accuracy could be determined by the simple and
well-known empirical formula

By equalizing eqns (2) and (4) the shear \I= 0.8 -

fc
-2oo ,fc in MPa (7)
capacity of the beam is found to be
440 L. C. Hoang, h4. ?! Nielsen

web crushing
/ r-

Fig. 2. The web crushing criterion. Fig. 3. Code restrictions on maximum cot 0 and mini-
mum $.

In practice the constant 0.8 has normally been

replaced by the conservative value 0.7.
Most of the tests used to confirm the web condition was in fact introduced as a service-
crushing criterion had relatively large reinforce- ability requirement to reduce the risk of
ment degrees and none of them had $/v less premature yielding but has undoubtedly com-
than about 0.05. pensated for the inability of the web crushing
The web crushing criterion does not agree criterion to cover cases with small shear rein-
well with test results when the shear reinforce- forcement degrees. The web crushing criterion
ment degree becomes small. The reasons for and the code restrictions are illustrated in
this are obvious. When Ii//v is small, the angle 0 Fig. 3.
must also be small before web crushing can take Referring to Fig. 3, the shear strength of
place. Hence, the diagonal stress field at failure beams with small shear reinforcement degrees
must cross cracks with many different direc- is lying somewhere between the straight line
tions. If a diagonal stress field with a and the web crushing criterion.
compression stress equal to vfC has to be trans- A more accurate evaluation of the shear
ferred through these cracks, it would be strength of such beams can only be carried out
necessary to supply one or another kind of con- by taking crack sliding as well as large stirrup
finement for instance in the form of extra distances into account. This will be demon-
stirrup reinforcement. Such a confinement is strated in the following.
normally not provided. The cracked web region,
therefore, may become unstable, resulting in
sliding failure along cracks before vfCis reached. THE CRACK SLIDING MODEL
The possibility of sliding in cracks will be fur-
ther enhanced by the large stirrup distances This section presents the recent developments
normally used when dealing with small rein- in the plasticity approach to shear design. The
forcement degrees. approach is an upper bound approach and is
When beams with small shear reinforcement based upon the hypothesis that cracks can be
degrees are designed by use of the web crushing transformed into yield lines, which have lower
criterion the shear strength will, for the reasons sliding resistance than yield lines formed in
given above, be overestimated. This problem untracked concrete. The plasticity interpreta-
has been partly overcome in the design codes” tion of crack sliding was introduced by Zhang in
by introducing a minimum shear reinforcement her thesis on the shear strength of non-shear
requirement* and by limiting the choice of reinforced beams.f’,7
cot II, for instance to cot 0 ~2.5 for beams with The model explains in a rational way the
constant longitudinal reinforcement. The latter shear behaviour of non-shear reinforced beams
and it can easily be extended to cover continu-
The present Danish code requires a minimum shear rein-
forcement degree &min = 0.2/\,~, fck being the ous beams, prestressed beams and beams with
characteristic concrete compressive strength. small shear reinforcement degrees.
 Plasticityapproach to shear design 441

Non-shear reinforced beams the critical diagonal crack is transformed into a

yield line.
Simply supported beams The transformation of a crack into a yield
line has been demonstrated clearly by Muttoni’
Consider the case of two symmetrical point who measured the relative displacements along
loads on a non-shear reinforced beam with rect- a crack. When the crack is formed the relative
angular cross-section as shown in Fig. 4. The displacement is mainly perpendicular to the
total depth of the beam is h. crack. When the crack is transformed into a
The crack pattern at the state of failure is yield line there will be a displacement compo-
schematically shown in the figure. The first nent parallel to the crack.
cracks are normally formed in the region with In order to calculate the shear strength of the
maximum moment. These cracks are vertical. beam, we must answer the following questions:
Then, gradually, diagonal cracks appear in the What is the sliding resistance of a crack, and
shear span. These cracks are formed closer and how can we determine the position of the criti-
closer to the support, and will, if prolonged, cal diagonal crack?
approximately intersect the top face at the load- To answer the first question, we will make
ing point. the simplified assumption that diagonal cracks
It is often observed in experiments that shear are developed following straight lines from the
failure takes place as sliding along the last bottom face to the loading point, see Fig. 5.
formed crack. This crack is called the critical Thus, the diagonal cracks may be described by
diagonal crack. In terms of the plastic theory, their horizontal projection X.

P
t a-
Fig. 4. Non-shear reinforced beam.

Fig. 5. Beam with idealized diagonal cracks.

442 L. C. Hoang, M. IT Nielsen

Further, we will assume that the beam is The factor v() also includes an effect of the
overreinforced in the longitudinal direction, absolute scale h and an effect of the longitu-
imposing the condition on the relative displace- dinal reinforcement ratio p.
ment u along the critical diagonal crack to be Since sin CI= x/,= the total internal work
vertically directed. Wr dissipated in the crack is found to be
It has been shown by Zhang” that the modi-
fied Coulomb failure criterion may also be
applied to a crack. The tensile strength along
the crack is naturally zero. By means of a
micromechanical model, Zhang has demon-
strated that the angle of friction p and the
cohesion c’ along a crack may be taken as
Inserting eqn (12) into the work equation, P,u
being the external work, we find the following
> load-carrying capacity P, as a function of X:
(8)

Here, c is the cohesion of untracked concrete

and vs is a sliding reduction factor which for P, = ; v,v,,f,hh [/I+( ;r- ;] (13)
normal strength concrete (fc I 50 MPa) was
found to be
Since this is an upper bound solution, one
1 would expect that eqn (13) should be minimized
v,= -
2
(9) with respect to X. In that case x should be put
equal to a. However, a crack sliding solution
For concrete with relatively weak aggregate corresponding to x = a is only valid if the hori-
materials, where cracks may pass through the zontal projection of the critical diagonal crack is
aggregate particles resulting in aggregate frac- a. This is normally not the case. Thus, x must be
turing, the factor v, may be less than 0.5.“’ determined in another way.
Here, only beams of normal strength concrete In order to determine the horizontal projec-
are considered. tion x of the critical diagonal crack, we must
For cracks transformed into yield lines the determine the load level necessary to form a
internal work per unit length may now be writ- diagonal crack with the horizontal projection x
ten as and the length Y’h2 +x2. According to plastic
theory, a crack may develop when the effective
tensile strength ftef of concrete is reached along
the crack path. Hence, referring to Fig. 6, the
(10) cracking load Per may be determined by a
moment equation around the loading point 0.
Here CIis the angle between the yield line and
This is equivalent to a failure mechanism
the displacement direction. Notice that the con-
involving rotation around point 0. After the
straint on CIis due to the fact that crack sliding
formation of a crack the longitudinal reinforce-
is a plane strain problem.
ment will be activated to maintain equilibrium.
Owing to softening it is necessary to intro-
The cracking load is given by
duce an effectiveness factor vo, which may be
taken to be

0.88 The effective tensile strength, see Ref. 6, may

vo = (1+26p),vo11 (11)
;x be taken to be

ftef=O.l56f:“s(h) (15)
Here fc is in megapascals, h in metres and p is
the longitudinal reinforcement ratio. Here, s(h) is a size effect factor:
 Plasticityapproach to shear design 443

Fig. 6. Stress distribution at formation of a crack.

The procedure of calculation is as follows:

first the equation P, = PCr is solved with respect
-0.X to X. Thereupon x is inserted into eqn (13) to
,h in metres (16) determine the shear capacity.
If x > a then a must be used when calculating
the shear capacity. Further, sliding is only pos-
The variation of P, and PC, versus x is shown
sible for x/h 20.75. This is due to the constraint,
in Fig. 7. By following the load increments from
imposed by the normality condition, on the
zero to failure, Fig. 7 provides a simple and
angle between the relative displacement and the
clear explanation of the shear failure mechan-
yield line.
ism.
Some additional effects and conditions due to
At lower load levels only steeper cracks can
the fact that cracks are formed with finite dis-
be formed. These cracks are indicated in Fig. 5
tances have been discussed by Zhang” to which
as the dotted cracks. Sliding failure along these
the reader is referred.
cracks cannot take place since their sliding
strength P,, is larger than the load needed to
Simplified calculation
form them. This is true as long as the PC,-curve
is lying below the P,-curve. Sliding failure along
The calculation may be strongly simplified if the
a crack immediately after it has been formed is
correct formula for the internal work (eqn (12))
only possible if the load needed to form the
is replaced with the following approximation,
crack equals the sliding strength of the same
see Refs 11 and 12:
crack. Thus, the shear strength of the beam as
well as the horizontal projection of the critical
diagonal crack is determined by the intersection w,=2 25 hhu (17)
point of the curves representing P,, and PC,. This x/h
is illustrated in Fig. 7.
where

z, = O.O59vJ,. (IQ
P,, shear capacity curve
Using this approximation, P, turns into

P,=2 z, J,h
(19)
xlh

P,,, cracking load curve

Now, by equalizing eqns (14) and (19), we
L ‘I *x obtain the following cubic equation rendering
Fig. 7. Cracking load and shear capacity versus x. x/h :
444 L. C. Hoang, M. E Nielsen

other end, see Fig. 8. The beam is assumed to

have the same constant reinforcement at top
and bottom. As before, we assume overrein-
forcement in the longitudinal direction. Thus,
no moment redistribution will take place prior
to failure.
Thus, the shear strength may be found by
The crack pattern, observed in tests, is sche-
inserting the solution of eqn (20) into eqn (19).
matically shown in the figure. Sliding failure will
Alternatively, the shear strength may also be
take place in a diagonal crack emerging from
expressed by the average shear stress z, = P,l
the fixed support due to the relatively larger
bh:
shear force in this region.
7,
The shear strength, expressed as the average
z,=2 - (21) shear stress at the fixed support z, = V/bh, is
x/h identical to eqn (21). This is easily verified by
writing down the work equation with Vu as the
Solutions for prestressed beams and beams
external work.
with uniform loading may be found in Refs 6
The cracking load, however, will differ from
and 7. Application of the model to prestressed
the case of a simply supported beam due to the
hollow-core slabs is explored in Ref. 13. For all
contribution from the bending moment at the
the cases treated good agreement with experi-
fixed end. This moment is considered as an
ments has been obtained.
external load. The cracking load may be found
Although rather rough, the suggested method
by a moment equation around point S, see Fig.
of determining the cracking load seems to work
8, or equivalently by considering a rotation
well. Improvements on this point require a frac-
mechanism around point S.
ture mechanics approach. Gustafsson’4
Expressions for the cracking load for this and
determined the cracking load by the Hillerborg
other cases of statically indeterminate beams
model using FEM-calculations. He also
may be found in Ref. 11. Here we only state the
observed that the position of the crack and the
cubic equation rendering the x/h.
load-carrying capacity may be determined by
the intersection point of two curves as shown in
Fig. 7.
3
7, a*
Continuous beams ; +;_” - -=o
(22)
( i 11 .ftef h
The crack sliding model presented is also cap-
able of treating continuous beams. This will be Notice that the shear span in this case is
demonstrated for the case of a point load in the denoted a*.
middle of a beam with rectangular cross-section, With eqns (20) and (22) in mind, we may
fixed in one end and simply supported in the pose the important question: What is required

P
-a*

Fig. 8. Half of two-span continuous beam.

 Plasticityapproach to shear design 445

for the two types of beam considered to have with rectangular cross-section, see Fig. 9. The
the same shear strength? If we assume that the beam is assumed to be overreinforced in the
material properties are identical, then the con- longitudinal direction.
dition to be fulfilled is that the solution of eqns Sliding failure in a diagonal crack with the
(20) and (22) must be identical. We see that horizontal projection x is considered. For the
this condition requires the following relation: time being, we consider the stirrups to be
closely spaced in the web (smeared approach).
a 6 a* The shear capacity, as a function of x, may then
--- -
(23) immediately be written down:
h-11 h

which is equivalent to the condition

MI MZ
-=- (24) (25)
Q,h Qd
Here, M, and Q, denote the internal forces at which, rewritten using an average shear stress,
the loading point of the simply supported beam, takes the form
and iV2 and Q2 denote the internal forces at the
fixed support for the beam shown in Fig. 8.
It is interesting to ascertain that eqn (24) is in (26)
fact an empirical rule (derived from test results)
given by Leonhardt et a1.15 for the purpose of The cracking load may be calculated using
answering the posed question. eqn (14), thereby neglecting the influence of the
Thus, the model presented has reproduced shear reinforcement on the cracking load.
an old empirical rule. This is indeed a triumph The cracking load curve and the shear capa-
of plastic theory. city curve are schematically shown in Fig. 10.
Unlike the case of a non-shear reinforced beam,
Lightly reinforced beams the shear capacity curves now display a mini-
mum value. Thus, if the cracking load curve
Smeared approach intersects the shear capacity curve for a value of
x smaller than that corresponding to the mini-
The crack sliding problem is also met in beams mum value, the load-carrying capacity will
with small shear reinforcement degrees. Such correspond to the intersection point. If, how-
beams are termed lightly reinforced beams. In ever, the minimum value is reached at an
what follows, we shall demonstrate that the x-value smaller than that corresponding to the
crack sliding model is capable of treating lightly intersection point, the load-carrying capacity
reinforced beams as well.‘” will be determined by the minimum value of the
Again, we consider the case of two symmetri- shear capacity curve. The two situations are
cal point loads on a simply supported beam illustrated in Figs 10(a) and (b).

t
h

Fig. 9. Simply supported lightly reinforced beam.

446 L. C. Hoang, M. t? Nielsen

P P
shear capacity curve
cracking load curve \f-
-V

cracking load curve

shear capacity curve

l load carrying capacity

Fig. 10. Cracking load and shear capacity versus x for beam with shear reinforcement.

In the first situation, failure will take place lead to quite unsafe results. To take into
immediately after the critical crack has been account the effect of finite stirrup distances, the
formed, whereas the load in the latter situation shear capacity curve must be drawn by counting
must, when the critical crack has formed, the number of stirrups crossing the cracks. If
increase until it reaches the minimum value of the stirrup distance is s and the shear reinforce-
the shear capacity curve before failure takes ment degree is $, each stirrup crossed adds
place. It turns out that the situation illustrated $f&h to the shear capacity z. An example of a
in Fig. 10(b) is already topical at very small shear capacity curve constructed as described is
shear reinforcement degrees, see Ref. 10. In the shown in Fig. 11. The curve will of course be
following, only the situation in Fig. 10(b) will be discontinuous.
treated in detail. The shear capacity curve corresponding to
The minimum value of the shear capacity the smeared stirrup approach has also been
curve is easily determined by minimizing eqn shown in the figure (the dotted line). Notice
(26) with respect toxlh. The result is that the two curves touch whenever x is a mul-
tiple of s (i.e. x = s, 2, 3s, etc.).
--X -2%
h- de-c (27)
In cases where the minimum value of the two
shear capacity curves is localized at the same
value of x, which hereby must assume a value
and given by a multiple of s, the effect of finite
stirrup distances may be incorporated in the
solution in eqn (29) in a very simple way;
(28) namely by deducting the contribution t,bfG/h
from eqn (29). This is illustrated in Fig. 11
where the shear capacity drops from point A to
Inserting eqn (18) into the right-hand side of
point B. (The figure illustrates a situation where
eqn (28) we find the following simple expres-
the minimum value of the two shear capacity
sion:
curves, at A and B, is localized at x = s).
The solution taking into account finite stirrup
7” ti
- = 11.64 - (29) distances will hereby take the simple form
7, J vo
T” ti
- =11.64 -16.95 - $ (30)
Effect of finite stirrup distances 7, vo

So far, the effect of finite stirrup distances has Naturally, it is more an exception than a rule
been neglected. However, when dealing with that the minimum values of both shear capacity
large stirrup distances, the solution derived may curves are localized at the same value of x. It
 Plasticity approach to shear design 447

shear capacity

Fig. 11. Shear capacity curves according to smeared approach and discrete approach

may be shown, however, that the difference In cases where x/h given by eqn (27) is larger
between the minimum values of the two curves than the shear span a, an additional condition
will never exceed a value corresponding to $f‘sl must be introduced. Thus, there is a lower limit
h. Thus, the solution in eqn (30) is generally a of G/v,, below which eqn (30) cannot be used.
safe and reasonable approach to take into The reader is referred to Ref. 10.
account the effect of finite stirrup distances. The solution in eqn (32) is shown in Fig. 12.
Owing to the normality condition the solution The reduction due to the finite stirrup distances
derived is valid only when x/h given by eqn (27) appears to be very drastic for large stirrup dis-
is larger than 0.75. By requiring x/h 20.75, the tances combined with large shear reinforcement
following condition for the validity of eqn (30) degrees.
is found: However, such reductions will not appear in
reality if the beams are designed using normal
$ practice. Designers usually choose to operate
- 10.21 (31) with one or two types of stirrup. $ is thus
\‘o
increased by decreasing s. In such cases, the
When $/v,, > 0.21 the shear capacity is cal- reduction will be constant ( =A,,&,/bh).
culated by inserting x/h = 0.75 into eqn (26) For comparison, the web crushing criterion
rendering a linear variation of the shear capa- has also been drawn in Fig. 12. To use the same
city versus $. (non-dimensional) nominal shear stress, the
When incorporating the effect of finite web crushing criterion given by eqn (5) must be
stirrup distances in the case where rl/l~~,> 0.21, multiplied by hi/h and divided by z,. By drawing
the shear capacity formula reads: the web crushing curve in Fig. 12, it has been
assumed that hi/h = 0.8 (which is a typical
value). Further, it has been assumed that v = vo.
ti $ s T <021
1 1.64 - This is obtained for the following typical values:
--T” - 1’0 -‘6.95 -7
k’o “I) fC = 25 MPa, h = 300 mm and p = 1.4%.
r< In Fig. 12 we notice that even in the case of
2.67 + 16.95 -Ic/ Jk >0.21 closely spaced stirrups (s/h = 0) crack sliding
\‘I) \‘(I
should be taken into account for shear rein-
forcement degrees up until $/v,, Y 0.25.
(32) Strictly speaking, the crack sliding criterion
Notice that if s/h > 0.75 when $/v,, > 0.21 the should always be applied if the corresponding
stirrups will give a negative contribution. This shear strength lies below the one belonging to
result, which of course is not true, is, however, the web crushing criterion. However, as stated
not of practical interest. Designers usually above, large reinforcement degrees will (nor-
choose stirrup distances to be less than 0.75h, mally) be accompanied with closely spaced
especially when $/v. > 0.21, which is a relatively stirrups, which makes the assumption of
large shear reinforcement degree. smeared stirrups more realistic. Thus, the trans-
448 L. C. Hoang, M. P Nielsen

A
8-

6 --

0 0.2 0.4 0.6

Fig. 12. Crack sliding criterion and web crushing criterion versus shear reinforcement degree.

ition point between crack sliding and web f,~[13-48 MPa] (average strength 30 MPa)
crushing may, in practice, be taken at the shear
i,GE [0.006-O. 1741 (average $ 0.043)
reinforcement degree $/v. = 0.25.”
Crack sliding solutions for lightly reinforced
Comparison with experiments beams subjected to uniform loading and for
continuous beams may be found in Ref. 10. For
The solution in eqn (32) has been compared these cases, good agreement with experiments
with some test series. Only a few of the test was found too.
series reported in the literature had varying
shear reinforcement degree with constant
stirrup distance. One small series was reported THE EFFECTIVENESS FACTORS
by Regan & Placas in Ref. 17.” The results are
shown in Fig. 13. It can be seen that the test The use of reduced concrete strengths through
results are in good agreement with the crack a multiplication with an effectiveness factor on
sliding solution. fc has been, and still is, a subject of debate
In Fig. 14 the results of 106 beam tests from among researchers. Researchers with different
five test series have been compared with the attitudes to practical engineering design take
calculated shear strength. It can be seen that different stands in the debate. But, no matter
the test results are in fairly good agreement what the opinions may be, the necessity of
with the theory. The mean value of the ratio working with effectiveness factors is indispen-
rtcst/rttle0ry is 0.97 and the standard deviation is sable when the rigid and perfectly plastic
0.17. material model is applied.
The concrete strength and the shear rein- The first effectiveness factor was introduced
forcement degree of these tests varied between: in a discussion on the web crushing criterion.22
Based on a test series of Leonhardt &
*This agrees with the experimental observations in Ref. Walther2’ it was found that v = 0.85. The inter-
16 where yielding of stirrups at failure was not detected pretation of the effectiveness factor at that time
when I) became larger than about 0.2. This indicates that was that only the part of the web concrete con-
crack sliding did not take place.
*Information concerning the stirrup distances was not
fined by the stirrups is active (the concrete
given in Ref. 17 but provided by a personal communica- cover must remain unstressed to match equili-
tion with the authors. brium requirements). This interpretation thus
 Plasticityapproach to shear design 449

I I I , I I I

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35

Fig. 13. Tests by Regan & Placas” (s/h = 0.5 and average value of v/v,~= 0.95).

addressed the ‘problem’ to the geometry of the understanding of the physical meaning of the
stirrups and not to the properties of concrete! effectiveness factors and why different problems
Much has been learned since then. require different values.
The last decades of research on the proper- At present it is believed that the effectiveness
ties of concrete and on the behaviour of factors take into account the following main
concrete structures has brought us to a better effects:

5.0 ~ ,
MC (test)
4.5 . _ _. . .

4.0

.: 0 0. Moretto [45.1]

1 0 A. P. Clark [51.1]

2.5
8n Krefeld &
-1 Thurston [66.1]

-1 0 Regan &
Placas [71.1]
. . . . . . . . . . . . . _ . . . .
1q Bresler &
Scordelis [63.1]
. . . . . . . .,. . . . . -4
j z,/T; (theory) :
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
Fig. 14. Comparison of theory with 106 test results: 0, Moretto;lX 0, Clark;” A, Krefeld & Thurston;“’ l, Regan &
Placas; ” a, Bresler & Scordelis.*’
4.50 L. C. Hoang, M. l? Nielsen

1. the softening post-peak behaviour of uncon- fC and for large structures. (Therefore v()
fined concrete; decreases with increasingf, and h.)
2. cracking of concrete; Cracking. The effect of cracking, which we
3. local damage of concrete prior to failure due have been dealing with in this paper, will enter
to stressed transverse reinforcement. in all problems involving crack sliding failure.
The strength reduction due to cracking has
Softening. The effect of softening will enter been taken as a constant (rS = 0.5). It is to be
in all problems with inhomogeneous strain expected that the validity of this assumption
fields. Owing to softening the stresses along a depends upon the roughness of the crack sur-
failure surface do not at every point assume the face as well as the width of the crack. The
maximum value at the same time, see the illus- roughness of the crack surface may possibly be
trations in Fig. 15. Thus, instead of using fC, we taken into account by expressing V, as a function
must deal with a reduced strength which of for instance a relative strength and a relative
expresses the average stress maintained in the size of the aggregate materials. The influence of
fully developed failure surface. The effect of the crack width on V, may be taken into account
softening is more pronounced for high values of by the parameters which directly or indirectly

r
0

Fig. 15. Response to flexural failure of a beam of softening materiaLz4

 Plasticityapproach to shear design 4.51

control the crack width, e.g. the reinforcement cracks cannot take place. In this case vxc,rtis the
ratio p. only effectiveness factor and it may be found
Local damage. Local damage of concrete using eqn (11). When sliding in cracks may take
prior to failure reduces the area of the final place, for instance for larger shear spans, it has
failure surface or the area carrying the load, been demonstrated in this paper that the effec-
thus resulting in a reduced load-carrying capa- tiveness factor may be taken as \tslid\ls<,rt
= \I~\‘~),
vs
city. The local damage may take the form of being 0.5 and v. determined by eqn (11).
cover spalling or local punching around the In this context the formula in eqn (7) valid
reinforcement bars. These damages are due to for the effectiveness factor for shear reinforced
the bursting stresses transferred from the beams with large reinforcement degrees belongs
stressed reinforcement bars to the concrete. to the type \‘hUTht.For such beams crack sliding
Studies on the effect of local damage may be may be disregarded and it seems that no vsOftis
found in Ref. 2.5 where effectiveness factors for present. One may hope that this is a general
this effect may also be found. Figure 16 shows trend, i.e. softening effects may be neglected
examples of local damage. when we are dealing with concrete reinforced
While the individual effects are now much with a uniform reinforcement in two directions
better understood, it is still an open question as and possibly a concentrated reinforcement nor-
to how to find the resulting effectiveness factor. mally placed along the boundaries.
There is some evidence that the resulting for- Thus, there is some evidence that the most
mula may be obtained by simple multiplication general formula for the effectiveness factor is
of the individual effectiveness factors.
Consider first homogeneous strain fields. If a 1’= ~‘slid\‘scd’t \‘hurst (33)

yield line in unreinforced concrete is formed in

a crack the effectiveness factor will be \Islid= v,. One or more terms may be missing as
If the concrete is reinforced so that there will explained above. For normal strength concrete
be stressed transverse reinforcement in the yield \‘siid= vs may always be taken as 0.5. For non-
line, the effectiveness factor will be vhUTht.If the shear reinforced beams and lightly reinforced
yield line in the latter case is formed in a crack, beams in shear vhu,.stmay be neglected and vsc,rt
evidently the effectiveness factor iS VslidVhurht. may be determined by eqn (11). For beams with
When dealing with inhomogeneous strain very small shear spans only vsc,frremains. It may
fields in unreinforced concrete and with no be determined by eqn (11). For beams in shear
yield lines in cracks, there is an effectiveness with large reinforcement degrees both Vslidand
factor v,,,r, due to softening. An example is a vstlft may be neglected and I’hurst may be cal-
corbel with small shear span so that sliding in culated by eqn (7).

I II II-II

effective efkctive
I II dlicktl~ failure surface
Fig. 16. Local damage due to bursting stresses.
452 L. C. Hoang, M. E Nielsen

This scheme is possibly somewhat oversimpli- REFERENCES

fied, but it is in agreement with the large
number of experimental findings as they have 1. Nielsen, M. P., Om forskydningsarmering i jernbe-
tonbjaelker (On the shear reinforcement design of
been collected up to now. reinforced concrete beams). Bygningsstatiske Medde-
However, formulas like eqn (33) can never be lelser, Vol. 38, No. 2, Copenhagen, 1967.
complete. The anchorage of the reinforcement 2. Nielsen, M. P., Brsestrup, M. W., Jensen, B. C. &
Bach, F., Concrete Plasticity - Beam Shear - Shear in
along the faces may impose restrictions on the Joints - Punching Shear. Special publication. Danish
compressive stresses which can be carried. For Society for Structural Science and Engineering, Struc-
shear reinforced beams this interpretation of v t rat Research Laboratory, Technical University of
might be more to the point than an vhurst effect. IY, enmark, Lyngby, 1978.
3. Nielsen, M. P., Limit Analysis and Concrete Plasticity.
In fact, most of the tests on shear reinforced Prentice-Hall, Englewood Cliffs, NJ, 1984.
beams were done with plain bars giving rise to 4. Thiirlimann, B., Braestrup, M. W., Feddersen, B.,
much less bursting forces than deformed bars. Andreasen, B. S., Christoffersen, J., Jagd, L. & Svens-
son, E., Concrete plasticity. In A tribute to Professor
So that is where we are today. Back to the M. P. Nielsen on his 60th birthday. Bygningsstatiske
69-reasoning. Meddelelser, No. 2-3-4, December, 1994.
Much more research is needed in this field. 5. Dansk Standard, DS 411, Nom for Betonkonsttuk-
tioner (Code of Practice for Concrete Structures). In
And this research should be done irrespective
press.
of a favourable or unfavourable view on plastic 6. Zhang, J.-P., Strength of cracked concrete. Part 1 -
theory. In any FEM code such information is shear strength of conventional reinforced concrete
needed. The only effect such a program may beams, deep beams, corbels, and prestressed rein-
forced concrete beams without shear reinforcement.
deliver, at least in the near future, is the v,,,rt Technical University of Denmark, Department of
effect. Structural Engineering, Report R No. 311, Lyngby,
1994.
7. Zhang, J.-P., Diagonal cracking and shear strength of
reinforced concrete beams. Magazine of Concrete
CONCLUDING REMARKS Research, 49( 178) (1997) 55-65.
8. Muttoni, A., Die Anwendbarkeit der Plastizitats-
theorie in der Bemessung von Stahlbeton. Institut fur
The recent years of code implementation of
Baustatik un Konstruktion, ETH Zurich, Bericht Nr.
plastic methods in many countries has clearly 176, 1990.
demonstrated that the plastic theory for rein- 9. Zhang, J.-P., Strength of cracked concrete. Part 2 -
forced concrete is indeed a powerful tool for micromechanical modelling of shear failure in cement
paste and in concrete. Technical University of Den-
practical design. mark, Department of Structural Engineering and
Plastic theory provides simplicity and free- Materials, Report R No. 17, Lyngby, 1997.
dom of choice in the design phase. It supplies a IO. Hoang, C. L., Shear strength of lightly reinforced con-
crete beams. Technical University of Denmark,
clear picture of the failure mechanism in the Department of Structural Engineering and Materials,
structure considered. Lyngby (in preparation).
In the future the use of sophisticated FEM 11. Hoang, C. L., Shear strength of non-shear reinforced
concrete elements. Part 1 - statically indeterminate
programs will undoubtedly increase, as more beams. Technical University of Denmark, Department
and more large-scale complex structures will be of Structural Engineering and Materials, Report R
built. The engineer will thus be placed in a posi- No. 16, Lyngby, 1997.
tion where he/she will find it difficult, if not 12. Nielsen, M. P., Limit Analysis and Concrete Plasticity.
2nd edn. In preparation.
impossible, to form a general view of the results 13. Hoang, C. L., Shear strength of non-shear reinforced
of complex FEM analyses. The consequence of concrete elements. Part 3 - prestressed hollow-core
undetected errors may be catastrophic (one slabs. Technical University of Denmark, Department
of Structural Engineering and Materials, Report R
scary example is the SLEIPNER-platform dis- No. 30, Lyngby, 1997.
aster). More than ever, the engineer will need 14. Gustafsson, P. J., Fracture mechanics studies of non-
to mobilize his/her commonsense and familiar- yielding materials like concrete. Report TVBM-1007,
Division of Building Materials, Lund Institute of
ise himself/herself with the behaviour of the Technology, Lund, Sweden, 1985.
analysed structure. 15. Leonhardt, F., Walther, R. & Dilger, W., Schubver-
In this context, plastic methods may serve as suche an Durchlauftmgern. Deutscher Ausschuss Fur
a tool providing fast estimates, based on which Stahlbeton, Heft 163, Berlin, 1964.
16. Brazstrup, M. W., Nielsen, M. P., Bach, F. & Jensen,
the engineer can judge the output of a detailed B. C., Shear tests on reinforced concrete T-beams,
FEM analysis. serie T. Technical University of Denmark, Dept. of
This may be one of the main roles the plastic Structural Engineering, Report R No. 72, Lyngby
1976.
methods will play in the future.
 Plasticity approach to shear design 453

17. Regan, P. E. & Placas, A., Shear failure of reinforced design of reinforced concrete beams, Discussion).
concrete beams. ACI Journal Proc., 68 (1971) Bygningsstatiske Meddelelser, Vol. 40, No. 1, Copen-
763-773. hagen, 1969.
18. Moretto, O., An investigation of the strength of wel- 23. Leonhardt, F. & Walther, R., Schubversuche an Plat-
ded stirrups in reinforced concrete beams. AC1 tenbalken mit unterschiedlicher Schubbewehrung.
Journal Proc., 42 (1945) 141-162. Deutscher Ausschuss Fur Stahlbeton, Heft 156, Ber-
19. Clark, A. P., Diagonal tension in reinforced concrete lin, 1963.
beams. ACI Journal Proc., 48 (1951) 145-156. 24. Petersson, P. E., Crack growth and development of
20. Krefeld, W. J. & Thurston, C. W., Studies of the fracture zones in plain concrete and similar materials.
shear and diagonal tension strength of simply sup- Report TVBM-1006, Division of Building Materials,
ported reinforced concrete beams. AC1 Journal Proc., Lund Institute of Technology, Lund, Sweden, 1981.
63 (1966) 451-475. 25. Zhang, J.-P., Strength of cracked concrete. Part 3 -
21. Bresler, B. & Scordelis, A. C., Shear strength of rein- load carrying capacity of panels subjected to in-plane
forced concrete beams. ACZ Journal Proc., 60 (1963) stresses. Technical University of Denmark, Depart-
51-68. ment of Structural Engineering and Materials, Report
22. Nielsen, M. P., Om forskydningsarmering i jernbe- R No. 18, Lyngby, 1997.
tonbjzlker, Diskussion (On the shear reinforcement