PLASTIC HINGE LENGTHS OF

NORMAL AND HIGH-STRENGTH

CONCRETE IN FLEXURE
P. Mendis
Dept. of Civil and Environmental Engineering, The University of Melbourne, Australia

Abstract : Full-range analysis methods are becoming popular in design of reinforced

concrete structures. These methods require a knowledge of the behaviour of plastic
hinges up to advanced curvatures. Concrete sections characteristically soften beyond
the plastic phase. To analyse a strain-softening structure, many researchers have used
a finite hinge length. In this paper, existing formulae are re-examined and the effects
of different variables on hinge length are discussed. Experimentally measured values
are compared with the values predicted by using these formulae. It is shown that the
upper and lower bounds suggested by the ACI committee 428 provide reliable
estimates of hinge lengths for both normal and high-strength concrete flexural hinges
up to 80 MPa.
Keywords: Hinge length, Inelastic analysis, Reinforced Concrete, High-strength concrete

1. INTRODUCTION has been experimentally established (Bazant, 1984) that

In actual tests of concrete beams at advanced curvature strain softening frequently is distributed over finite-size
in flexure, yielding of reinforcement and crushing of regions of a heterogeneous material. It is used as the
concrete occur over a finite length around the critical finite length during softening in the mathematical model
moment section. It is referred to herein as the hinge for collapse load analysis (Darvall and Mendis, 1985),
length. Several researchers [e.g. Cranston (1965b)] have in contrast to the point hinge assumption in the plastic
noted that plastic deformation was confined to a small range. The rotation dq of a point plastic hinge is found
length of a member, even in areas of constant bending by multiplying hinge length, lp, by a curvature change
moments. Concrete sections characteristically soften, (df) (Park and Paulay, 1975). lp is defined to one side of
that is the bending moment decreases with increasing a maximum moment section. There is no satisfactory
curvature after the end of the plastic plateau. Several theoretical prediction of hinge length in the plastic or
researchers such as Rosenblueth and de Cossio (1964), softening regions. Similarly, it is not yet possible to
Barnard (1964), Cranston (1965a) and Darvall and determine theoretically the hinge length at advanced
Mendis (1985) have used a finite length over which curvatures. Calculations for hinge length can best be
curvature increased while moment decreased in based on test observations. Several empirical formulae
computations including softening. Somes (1966) are available to calculate the hinge length, lp. Those
observed that in prestressed concrete beams at advanced formulae are reviewed in this paper.
curvatures the greater part of the beam began to The major advantage of including the inelastic
straighten out while a finite zone continued to soften. behaviour of the members in structural design is that the
Bazant (1976), using energy concepts, proved the moment redistribution which occurs can increase
softening region to be localised. Using a random two- appreciably the strength of the structure above the
dimensional network, whose joints are placed at random strengths calculated by the elastic theory. If inelastic
locations over a rectangular domain, Burt and Dougill flexural deformations occur in a reinforced concrete
(1977) obtained a finite size strain-softening region. It structure as a result of gravity and lateral force

Advances in Structural Engineering Vol. 4 No. 4 2001 189

 Plastic Hinge Lengths of Normal and High–Strength Concrete in Flexure

overloads, they will normally be concentrated in a finite Baker and Amarakone (1964)
hinge length. Hinge length is an important parameter in On the basis of experimental evidence, including beams
full-range analysis of normal strength and high-strength with lateral reinforcement, Baker and Amarakone
concrete structures including elastic, plastic and (1964) proposed the following formula.
softening phases (Mendis, 2000). A comprehensive
review of formulae available to calculate the hinge (2)
length is presented in this paper. Although presented
way back in 1968, it is shown in the paper by k1, k3 and z are defined as for the I.C.E. formula. kud is
considering experimental results, that the formulae the neutral axis depth at collapse.
suggested by ACI 428 committee can be used as lower
and upper bounds in inelastic analysis of normal and Cohn and Petcu (1963)
high-strength concrete structures. However the ACI During their tests on two span beams, Cohn and Petcu
formulae are independent of longitudinal and lateral (1963) observed that the length to one side of the plastic
reinforcement percentages. An empirical formula is zone at the intermediate support is between 0.3d
presented in this paper to include these variables. and 0.9d.

2. RE-EXAMINATION OF HINGE Sawyer (1964)

LENGTH FORMULAE The approach of Sawyer (1964) is illustrated in
Figure 1 for a length of beam between two points of
Chan (1955) contraflexure.
Using a semi-empirical method, Chan (1955) showed
that the hinge length did not vary appreciably with the
steel ratio and took an average value of 0.4z for the
hinge length.

I.C.E. Formula (1962)

Based on A.L.L. Baker’s recommendations, the I.C.E.
committee (1962) specified the following formula for
the equivalent plastic hinge length, lp. lp was defined as
the equivalent length over which plastic rotation
occurred with constant curvature in that length.

k1 = 0.7 for mild steel

= 0.9 for cold worked steel
Figure 1. Approach of Sawyer to computation of hinge length
(1)
It is assumed that the maximum moment in the member
k2 =
is equal to the ultimate moment, Mu, that the yield

moment, M y, is related to Mu by , and that

k3 =
the zone of yielding spreads d/4 in each direction past
the section at which the bending moment is equal to My,
where d is the effective depth of the beam. The moment-
(Assuming f’ c = 0.85 x cube strength of concrete)
curvature relationship is assumed to be bilinear to the
f’c is in MPa
maximum moment point, the curve being defined by
z - distance of critical section to point of contraflexure.
points (0,0), (My, fy) and (M u, fu). The total inelastic
P - axial load on the member
rotation to one side of the critical moment section is
Pu - ultimate capacity of the member under axial load
given by,
when no bending moment acts.

190 Advances in Structural Engineering Vol. 4 No. 4 2001

P. Mendis

when he plotted the measured values of as a

where fu is the curvature corresponding to M u and l’ p is function of numerous possible variables. In view of this
the equivalent hinge length without considering spread scatter he disputed the validity of a least squares fit. He
of plasticity beyond the point of yield moment. proposed a simple expression for computing lp.

Thus l’ p = 0.075z and lp = 0.075z + 0.25d. (4)

Mattock (1964) where both d and z are in inches.

Using his test results, Mattock (1964) concluded that
the extent to which plasticity extends beyond the Modified Mattock formula (1967)
distance d/2 from the section of maximum moment By plotting measured value of lp against z, obtained in
depends primarily on the ratio of the distance z from the investigations at the Portland Cement Association,
point of maximum moment to the point of contraflexure Mattock (1967) showed that even though there is
to the effective depth, d. To a lesser extent it depends on considerable scatter, the trend of the data can be
the amount of flexural reinforcement in the section. He represented reasonably well by the following
proposed the following empirical formula for the length expression.
of the plastic region, based on a least squares fit.
(5)

(3)
ACI Limits (1968)
The ACI-ASCE Committee 428 on Limit Design (1968)
proposed envelopes, or upper and lower limits, rather
than a single expression for hinge length. The length
q = tension reinforcement index, along a member from the section of maximum moment,
over which the inelastic curvatures are assumed to
occur, lp, shall exceed the lesser of these two values:

(6)

q’ = tension reinforcement index,

but not greater than (7)

qb = tension reinforcement index for balanced in which; (8)

ultimate strength conditions,

and (9)

pb = steel ratio at balanced ultimate strength

conditions in a beam without compression
reinforcement, (10)
d is in inches. f’c is in psi.
where;
Corley (1966) w = uniformly distributed load at a section of
Corley (1966) found a pronounced experimental scatter maximum moment

Advances in Structural Engineering Vol. 4 No. 4 2001 191

 Plastic Hinge Lengths of Normal and High–Strength Concrete in Flexure

(Kip/in.), Paulay and Priestley (1992) have suggested the

Vz = shear adjacent to a concentrated load or following formula for the length of a plastic hinge
reaction at a section of maximum moment, formed at the bottom of a cantilever column. This value
Mm = maximum moment in a length of member, includes an allowance for the elongation of bars beyond
Me = elastic-limit resisting moment, the theoretical base (tensile strain penetration), which
Mu = ultimate resisting moment concurrent with Pu, leads to additional deflection and rotation.
Pu = ultimate resisting axial load,
ecue = elastic component of ecu, either calculated or lp = 0.08l + 0.022 d b f y
assumed in the range 0.001 to 0.002,
ecu = maximum compressive strain in concrete at Where l is the height of the cantilever column. They
Mu and Pu, and suggested that, for typical column proportion, the above
ecuo = basic maximum compressive strain in equation results in values of lp < 0.5h, where h is the
concrete (neglecting possible amplifying section depth.
influences of confinement, loading rate and strain
gradients) to which a value in the range 0.003 to 0.004 Suggestion of Warner and Yeo (1984)
shall be assigned. By considering the flow of forces in a hinging region at
Re restricts the range of total inelastic rotation by an intermediate support, Warner and Yeo (1984)
providing reduced limits on lengths for the greater concluded that the region of steel yield spreads
assumed values of inelastic strains and curvatures, and outwards from the face of the support in steps of size s,
increased limits on lengths for the smaller assumed where s is the stirrup spacing. They suggested that a
strain values ACI-ASCE (1968) . reasonable estimate of the final length of the hinging
region to one side of the support was that multiple of s
Park (1982) which is less than or equal to the effective beam depth d.
Based on tests on four full size reinforced concrete
columns with a 550 mm square cross section, Park et al. 3. EXPERIMENTAL RESULTS OF
(1982) showed that the experimentally measured MENDIS
equivalent plastic hinge lengths are comparatively Mendis (1986) conducted tests on simply supported
insensitive to axial load level and had an average value beams loaded at mid-span to investigate the effects of
of 0.42h, in which h is the overall section depth. For the the following parameters on the hinge length:
columns tested, h=550 mm and d =488 mm. (i) Relative amounts of compression and tension
reinforcement- this was examined by varying the
Therefore, amounts of compression and tension steel.
(ii) Transverse reinforcement- this effect was
The equivalent plastic hinge length, lp, was calculated evaluated by changing the spacing between
by replacing the inelastic area at the ultimate stage by an the stirrups.
equivalent rectangle of height (fu-fy) and width lp. The (iii) High shear forces - this was evaluated by testing
following implicit formula was derived beams with different shear span ratios.
for lp. (iv) Axial loads - this effect was examined by applying
an axial force in addition to the shear and bending
moment from transverse forces. Axial forces equal
to 50,100 and 175 kN were selected for the
dp = inelastic deflection measured beyond the experiments. The balanced point axial load, Nub,
first yield displacement. on the axial load-moment interaction diagram for
(fp-fy) = plastic curvature measured beyond the first the cross section was approximately 100 kN. Core
yield curvature over the 100mm gauge length dimensions, reinforcement details and concrete
adjacent to the central stub. properties for the test beams are summarised in
Blakeley and Park (1971), Park and Thompson Table 1. Complete dimensions and reinforcement
(1977) and Thompson and Park (1980) used a technique layout with a longitudinal section for Beam A1 is
similar to Park et al. in measuring the equivalent plastic presented in Figure2.
hinge lengths of prestressed concrete beams. They
obtained equivalent plastic hinge lengths approximately
equal to one-half of the overall depth of the beam, i.e.,
lp values independent of prestress levels.

192 Advances in Structural Engineering Vol. 4 No. 4 2001

P. Mendis

Table 1 - Details of Beams tested by Mendis It was observed that the hinge length increased with
the shear span ratio and the reinforcement percentage,
but decreased with the lateral reinforcement percentage.
In the tests conducted, long span beams J2 and K1
showed a longer hinge length, as predicted. Except for
Beam H2, the hinge length remained constant around
0.40d for beams tested with axial loads. It is
complimentary to the findings of Park et al. (1982), who
showed that the hinge length ratio is independent of the
axial load ratio.
The experimental results were compared with the
hinge lengths predicted by the formulae presented
before. These results are plotted in Figure 3. ACI, lower
and upper bounds are also plotted in the figure with
Note: Beam J1 was not included as it was tested in
double-point loading.
solid lines. It is clearly seen that the experimental results
are within the ACI limits. The 95% confidence interval
calculated from a t-distribution for the ACI lower limit
is (1.079, 1.411): i.e. we are more than 95% confident
that the ACI lower limit underestimates the
hinge length.
As seen from Figure 3, Mattock, Corley and Mod.
Mattock formulae over-estimate the hinge lengths by a
significant margin. Sawyer’s formula over-estimate the
hinge lengths in many beams. Although the ICE
formula over-estimate the hinge lengths, the predictions
are closer to the experimental values. Park et al. formula
makes reasonable predictions for beams with axial
loads. Baker and Amerakone formula under-estimates
the hinge lengths of all the beams.

Figure 2. Longitudinal and cross-sections of Beam A1

After testing, the unloaded specimens were used to

measure the lengths within which plastic rotations were
concentrated. A reference line was drawn at a height of
d/2 or 75 mm from the top of the beam. When a
specimen is unloaded the regions beyond the hinge
unload along the elastic slope. The reference profile
within this region is a straight line. When these lines are Figure 3. Comparison of experimental results with
extended the reference profile is seen to deviate at two existing formulae
particular points either side of the midspan. The region
between these points is taken as the total hinge length. A nonlinear regression analysis was performed to
Enlarged photographs were used to verify these lengths. find a suitable equation to fit the experimental results.
Table 1 gives the measured hinge lengths for the beams The regression parameters were found by IMSL
tested. In some of the specimens the length of the hinge subroutine ZXSSQ (IMSL, 1982) . An equation (Eqn.
on one side of midspan is different from that on the 11) was derived for the ratio between hinge length,lp,
other side. This is due to the non homogeneity of and effective depth,d, for beams without axial loads.
concrete. An average of the two side values was
selected for the hinge length. For beams without axial loads,

Advances in Structural Engineering Vol. 4 No. 4 2001 193

 Plastic Hinge Lengths of Normal and High–Strength Concrete in Flexure

Hinge length = (11)

Where

RFT =

SPR =

PER = Percentage ratio of volume of stirrups to

volume of concrete core measured to outside of Figure 4. Experimental results of Kovacic and Pendyala
the stirrups.
5. CONCLUSIONS
4. WORK ON HIGH-STRENGTH 1. Hinge length is an important parameter in inelastic
CONCRETE analysis of reinforced concrete structures.
As part of a long-term investigation on behaviour of 2. The hinge length is very difficult to be evaluated
high-strength concrete structures at the University of theoretically. Formulae based on experimental results
Melbourne, Pendyala (1997) and Kovacic (1995) are more reliable in estimating the hinge length.
conducted experimental and theoretical investigations on 3. As seen from the experimental results and the
full-range behaviour of high-strength concrete (up to 80 computer experiments, the hinge lengths predicted
MPa) beams and columns with low axial load ratios by most of the formulae lie within the limits
respectively. The details of their specimens and measured suggested by the ACI Committee 428 in 1968.
hinge lengths are given in Table 2. ACI lower and upper Also, a specified range is more suitable for hinge
bounds for hinge lengths are plotted with experimental length than a single expression. Based on the results of
results in Figure 4. As seen from this figure, all the this study, the envelope defined by ACI limits is
experimental results, except for D1 and D4 specimens of recommended for the hinge length. As shown by
Kovacic, are located within the ACI limits, justifying the Mendis (1986), the critical softening slope for a
use of these formulae to estimate the hinge lengths of softening structure is inversely proportional to the hinge
high-strength concrete beams (up to 80 MPa) and high- length ratio, m. m was defined as the ratio between the
strength concrete columns with low axial loads. More total length of a member and the hinge length. Thus the
work is required to extend these observations to very effects of softening are more significant for shorter
high-strength concrete members. hinge lengths. The ACI lower limit for hinge length is
therefore an appropriate input value for softening
Table 2 - Details of Beams tested by analysis. Softening analysis of normal strength and high
Kovac ic and Pendyal a strength concrete structures is described elsewhere
(Mendis, 2000).
4. The ACI formulae give reliable estimates of hinge
lengths for high strength concrete beams (up to 80 MPa)
and columns with low axial loads. More work is
required to confirm this observation for columns with
high axial loads and for very high strength concretes.

REFERENCES
1. Baker A.L.L. and Amarakone A.M.N. (1964), “Inelastic
Hyperstatic Frame Analysis”, Proceedings International
Symposium on the Flexural Mechanics of Reinforced Concrete,
Miami, ACI SP-12, 85-142.
2. ACI-ASCE Committee 428 (1968), “Progress Report On Code
Clauses for Limit Design”, ACI Journal, 65, No. 9, 713-715.
3. Barnard P.R. (1964), “The Collapse of Reinforced Concrete
Beams”, Proceedings International Symposium on the Flexural
Mechanics of Reinforced Concrete, Miami ACI SP-12, 501-520.
4. Bazant Z.P. (1984), “Continuum Theory For Strain-Softening”,

194 Advances in Structural Engineering Vol. 4 No. 4 2001

P. Mendis

Eng. Mech. J., ASCE, 110, No. EM12, 1666-1692. Miami, ACI SP-12, 143-181.
5. Bazant Z.P. (1976), “Instability, Ductility and Size Effect in 19. Mendis P.A. (1986), “Softening of Reinforced Concrete
Strain-Softening Concrete”, Eng. Mech. J, ASCE, 102, No. EM2, Structures”, Ph.D. Thesis, Monash University.
331-344. 20. Mendis P.A. (2000), “Softening of Normal strength and High
6. Blakeley R.W.G. and Park R. (1971), “Seismic Resistance of strength Concrete Frames”, Advances in Structural Engineering
Prestressed Concrete Beam-Column Assemblies”, ACI Journal, Journal, Vol. 3, No.2, 2000, pp. 109-117.
68, No. 9, 677-692. 21. Park R. and Paulay T. (1975), Reinforced Concrete Structures,
7. Burt N.J. AND Dougill J.W. (1977), “Progressive Failure in a John Wiley & Sons.
Model Heterogeneous Medium”, Eng. Mech. J., ASCE, 103, No. 22. Park R. and Thompson K.J. (1977), “Cyclic Load Tests on
EM3, 365-376. Prestressed and Partially Prestressed Concrete Beam-Column
8. Chan W.W.L. (1955), “The Ultimate Strength and Deformation of Joints”, PCI Journal, 22, No. 5, 84-110.
Hinges in Reinforced Concrete Frameworks”, Magazine of 23. Park R. (1982), “Ductility of Square-Confined Concrete
Concrete Research, 7, No. 21, 121-132. Columns”, Struct. J., ASCE, 108, No. ST4, 929-950.
9. Cohn M.Z. and Petcu V.A. (1963), “Moment Redistribution and 24. Paulay T. and Priestley M.J.N. (1992), “Seismic Design of
Rotation Capacity of Plastic Hinges in Redundant Reinforced Reinforced Concrete and Masonry Structures”, John Wiley &
Concrete Beams”, Indian Concrete Journal, 37, No. 8, 282-290. Sons, Inc.
10. Corley W.G. (1966), “Rotational Capacity of Reinforced 25. Pendyala R. (1997), “Flexural Behaviour of High-strength
Concrete Beams”, Struct. J., ASCE, 92, No. ST5, 121-126. Concrete Members”, Ph.D. Thesis, The University of Melbourne.
11. Cranston W.B.A. (1965a), “Computer Method for Inelastic 26. Rosenblueth E., and Diaz De Cossio R. (1964), “Instability
Analysis of Plane Frames”, Technical Report TRA/386, Cement Considerations in Limit Design of Concrete Frames”,
and Concrete Association, London. Proceedings International Symposium on the Flexural Mechanics
12. Cranston W.B. (1965b), “Tests on Reinforced Concrete Frames; of Reinforced Concrete, Miami, ACI SP-12, 439-463.
1: Pinned Portal Frames”, Technical Report TRA/392, Cement 27. Sawyer H.A. (1968), “Comments on Model Code Clauses”, ACI
and Concrete Association, London. Journal, 65, No. 9, 715-719.
13. Darvall P.LEP. and Mendis P.A. (1985), “Elastic-Plastic- 28. Sawyer H.A. (1964), “Design of Concrete Frames for Two
Softening Analysis of Plane Frames”, Struct. J, ASCE, 111, Failure Stages”, Proceedings International Symposium on the
No. 4, 871-888. Flexural Mechanics of Reinforced Concrete, Miami, ACI SP-12,
14. I.C.E. Research Committee (1962), “Ultimate Load Design of 405-431.
Concrete Structures”, Proc. I.C.E. (London), 21, No. 2, 399-442. 29. Somes N.F. (1966), “Moment-Rotation Characteristics of
15. IMSL International Mathematical and Statistical Libraries Inc., Prestressed Concrete Members; Stage 1: Rectangular Sections”,
(1982), Reference Manuals, Vol.4. Technical Report No.TRA/398, Cement and Concrete
16. Kovacic D. (1995), “Design of High-strength Concrete Association.
Columns”, M.Eng.Thesis, The University of Melbourne. 30. Thompson K.J. and Park R. (1980), “Ductility of Prestressed and
17. Mattock A.H. (1967), Discussion of “Rotational Capacity of Partially Prestressed Concrete Beam Sections”, PCI Journal, 25,
Reinforced Concrete Beams,’’ by W.G.Corley, Struct. J., ASCE, No. 2, 47-70.
93, No. ST2, 519-522. 31. Warner R.F. and Yeo M.F. (1984), “Collapse Behaviour of
18. Mattock A.H. (1964), “Rotational Capacity of Hinging Regions Concrete Structures with Limited Ductility”, Report No. R61,
in Reinforced Concrete Beams”, Proceedings International Dept. of Civil Engineering, University of Adelaide.
Symposium on the Flexural Mechanics of Reinforced Concrete,

Priyan Mendis is an Associate Professor and Reader at the University of Melbourne. His primary research
interests include design of concrete structures, high- strength/high- performance concrete, seismic
design of concrete structures and tall buildings. He is the chairman of the international sub- committee of
ACI 363 committee on high- strength concrete.

Advances in Structural Engineering Vol. 4 No. 4 2001 195