Bond Slip Effect in the Assessment of RC Structures Subje 2015 Procedia Engi
Bond Slip Effect in the Assessment of RC Structures Subje 2015 Procedia Engi
com
ScienceDirect
Procedia Engineering 114 (2015) 792 – 799
Abstract
The effect of bond-slip in the assessment of reinforced concrete structures under static or dynamic cyclic loads with numerical
models may be significant. Its relevance is discussed in this paper, by analysis of the correlation of experimental and analytical
results. The latter were obtained with a perfect bond-based model and with a fiber-section beam-column model which makes
it possible to consider the effect of bond-slip in the vicinity of exterior as well of interior joints of reinforced concrete frame
structures. The experimental results consist of a shaking table test of a reinforced concrete frame, with predominant flexural
condition, in agreement with the premisses of this research. The model with bond-slip was capable of predicting the structural
behaviour in a very satisfactory manner. Furthermore, the results of the assessment including bond-slip are significantly more
accurate than those assuming perfect bond.
©2015
© 2015TheTheAuthors.
Authors. Published
Published by Elsevier
by Elsevier Ltd.
Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of INEGI - Institute of Science and Innovation in Mechanical and Industrial Engineering.
Peer-review under responsibility of INEGI - Institute of Science and Innovation in Mechanical and Industrial Engineering
Keywords: Reinforced concrete; Bond-slip; Fiber-section model; Nonlinear dynamic analysis; Seismic assessment
1. Introduction
Not considering bond-slip and adopting Bernoulli’s hypothesis in a reinforced concrete (RC) fiber-section finite
element (FE) results in an overestimation of the stiffness of the model, especially relevant near footings and beam-col-
umn joints. Also, tension-stiffening is neglected, which, given the damage levels which normally occur in RC struc-
tures subjected to extreme cyclic actions, is usually negligible in seismic analysis. Even though they possess the
previously mentioned limitations, perfect bond-based models [1] are frequently adopted, mainly due to the simplicity
and lower computational effort required. Significant research effort has been put into the development of RC frame
element models capable of accounting for bond-slip effects, namely those of the category of fiber-section FEs. In this
group there is the model developed by Monti and Spacone [2], obtained from the combination of the fiber beam FE
for seismic analysis of RC by Spacone et al. [1] and of the FE for reinforcing bars anchored in concrete of Monti et
al. [3]. This formulation makes it possible to account for bond-slip in the end joints of RC frame elements. Also,
∗ Corresponding author.
E-mail address: [email protected]
1877-7058 © 2015 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of INEGI - Institute of Science and Innovation in Mechanical and Industrial Engineering
doi:10.1016/j.proeng.2015.08.028
Paulo Silva Lobo et al. / Procedia Engineering 114 (2015) 792 – 799 793
the model by Limkatanyiu and Spacone [4], is noted. It allows for the consideration of bond-slip and the coupling
effect of adjacent beam-column elements, and consists of a frame element and of a plane rigid-panel joint element.
This model leads to the introduction of a significant number of beam FEs to model a structural element. A direct
and simple way of considering bond-slip involves implementing nonlinear spring elements at the FE-end nodes [5].
Even though this solution is competitive regarding computational implementation, the constitutive models definition
requires ad-hoc calibration. Also, the coupling effect of adjacent structural elements cannot be accounted for with
these models. D’Amato et al. [6] used the same concept as applied by Monti and Spacone [2] for the implementa-
tion of an anchored bar formulation [7] in a concentrated plasticity FE. As in the previous case, this model does not
consider coupling effects. Solid models have been proposed [8], but the highly demanding computational implemen-
tation currently makes their utilization difficult for the assessment of structures with significant dimensions. A fiber
force-based FE with continuous anchored bars was proposed recently [9]. This model makes it possible to consider
bond-slip in interior joint regions and, being fiber-section based, accounts for the effect of the variation of the axial
force due to the overturning effects on the hysteretic behaviour of the element. The research reported herein makes
use of this model and of the model by Monti and Spacone [2] to evaluate the correlation of the predicted response con-
sidering bond-slip with experimental results. The relevance of bond-slip is assessed by comparison with the response
obtained considering the perfect bond based FE model by Spacone et al. [1].
A FE formulation of a reinforcing bar anchored in concrete with continuous bond was proposed by Monti et al.
[3]. The relation ΔS = KΔu between nodal displacements, u, and stresses, S, was presented (a bold symbol is used
to represent a vector or a matrix). K is the stiffness matrix given by the contribution of the rebar element and of the
bond forces. Monti and Spacone [2] revised the previous model, making it possible to implement it in a fiber-section
FE [1]. The resulting formulation is defined by the anchorage, modelled by a series of n rebar anchored in concrete
FEs [3], and the rebar fiber with length LIP (see Fig. 1a). This length should be considered equal to the plastic hinge
length of the element [6,9], which may be given by the equation proposed by Bae and Bayrak [10].
The anchorage is assimilated to a spring attached to the rebar fiber, resulting in the anchorage slip law
kaa kan Δua Δσa
= (1)
kna knn Δun 0
where the subscript n refers to the anchorage FEs nodes other than the node of the anchorage/rebar fiber interface, in
which case the subscript a is adopted. kaa , kan , kna and knn are terms in the anchored bar stiffness matrix, u refers to
the nodal displacements of the anchorage FEs and σa is the stress at the anchorage-end, equal to the stress along the
rebar fiber, σ s+a , in which the subscript is a reference to steel bar plus anchorage. ua is equal to εa LIP . The anchorage
Fig. 1: Anchored rebars with bond-slip in beam column joints; (a) and (b) in the image caption
plus rebar fiber series system may then be described by (E/LIP · ka )/(E/LIP + ka ) · Δu s+a = Δσ s+a , where E is the
elasticity modulus of steel and ka = kaa − kan k−1 nn kna . The relation between the normal stress and the strain of the
fiber is obtained knowing that u s+a = ε s+a LIP , resulting in E s+a Δε s+a = Δσ s+a . For column-foundation or exterior
beam-column joints, the previous formulation can be implemented directly in a fiber-section FE model, such as the
one proposed by Spacone et al. [1].
794 Paulo Silva Lobo et al. / Procedia Engineering 114 (2015) 792 – 799
A FE capable of modelling bond-slip in interior joints of RC elements was recently proposed [9]. The model of
the continuous anchored rebar with bond-slip, which integrates the mentioned FE, is a generalization of the anchored
element formulation addressed previously. It consists of three springs in series (see Fig. 1b), in which the anchorage
is modelled by a series of n + 1 bars with continuous bond [3]. The anchorage tangent stiffness matrix results in
ka,a ka,n Δua Δσa
= (2)
kn,a kn,n Δun 0
where Δua = {Δua1 Δua2 }T and Δσa = {Δσa1 Δσa2 }T . The 1 and 2 subscripts indicate the rebar fibers on each side of
the joint anchorage. With ka = ka,a − ka,n k−1
n,n kn,a , adding the rebar fibers to the anchorage, yields
k12 −k12 Δu s+a Δσ s+a E/L1 0
= with k12 = (3)
−k12 ka + k12 Δua 0 0 E/L2
where Δu s+a = {Δu s+a,1 Δu s+a,2 }T and Δσ s+a = {Δσ s+a,1 Δσ s+a,2 }T . The previous equation results in k12 ka /(k12 + ka ) ·
Δu s+a = Δσ s+a . Because the displacements u s+a,1 = ε s+a,1 L1 and u s+a,2 = ε s+a,2 L2 are obtained at the cross-sections
state determination level, the unknowns are the stresses σ s+a,1 and σ s+a,2 , applied at the rebar fibers-end.
The adopted fiber-section FE is composed of two force-based beam-column elements and three nodes. The interior
joint is modelled by the node common to both beam-column FEs, in which the anchored rebar fibers follow the
previously addressed model. Furthermore, here the dimensions of the rigid joint are neglected, for simplicity. The
constitutive relations of the cross-sections of the three-node element are computed similarly to two-node force-based
FEs for 0 ≤ x3(1) < L(1) and 0 < x3(2) ≤ L(2) , where x3 is the longitudinal axis of the element, L is the length of the RC
elements, and (1) and (2) are references to the elements at each side of the interior joint. For x3(1) = L(1) and x3(2) = 0,
the constitutive relations of both middle-node control-sections cannot be uncoupled, resulting in
⎧ ⎫ ⎡ (1,1) ⎤⎧ ⎫
⎨Δs(L) ⎪
⎪
⎪ (1)
⎬ ⎢⎢⎢⎢ Ω E AdΩ Ω E AdΩ⎥⎥⎥⎥ ⎪
⎪
(1,2)
⎨Δe(L) ⎪
⎪ (1)
⎪
⎬
⎪ ⎪ = ⎢ ⎥ ⎪ ⎪
⎭ ⎢⎣ E (2,1) AdΩ E (2,2) AdΩ⎥⎦ ⎪
⎪ (4)
⎩ Δs(0)(2) ⎪ ⎩ Δe(0)(2) ⎪
⎭
Ω Ω
s(x3 ) = {N(x3 ) M1 (x3 ) M2 (x3 )}T is the vector of internal forces at the control-sections, in which Mi is the bending
moment about xi axis. x1 and x2 form the axes system of the cross-sections. e(x3 ) = {εo (x3 ) χ1 (x3 ) χ2 (x3 )}T represents
the deformation vector of the cross-sections, in which εo is the axial strain at the cross-section coordinate system
origin and χi is the curvature about xi . Ω refers to the cross-section domain and A is composed of terms on x1 and x2 ,
making it possible to obtain the stiffness terms of the cross-section. E (1,1) and E (2,2) are the tangent elasticities of the
fibers of the cross-sections of the elements (1) and (2) , and E (1,2) and E (2,1) are the crossed terms due to the fiber model
based on continuous anchored bars, given by adaptation of the equation relating Δu s+a with Δσ s+a , resulting in
(1,1) (1,2)
E E Δε s+a,1 Δσ s+a,1
= (5)
E (2,1) E (2,2) Δε s+a,2 Δσ s+a,2
The equations which relate the independent forces, q, with the internal forces at the cross-sections of each partial
element on each side of the interior joint are obtained with an equilibrium transformation matrix, delivering Δs(x3 ) =
B̄Δq, in which the matrix B̄ contains force-interpolation functions. Substitution of (4), which may include twisting
moments, in this equation yields {Δe(1) Δe(2) }T = Ā−1 B̄ · {Δq(1) Δq(2) }T , for x3(1) = L(1) and x3(2) = 0. The tangent
flexibility matrix of the three-node FE can be computed with the increments of the independent deformations, Δv,
defined according to the virtual force principle. Adopting the Gauss-Lobatto integration scheme, with the exception
of the control-sections which model the interior joint, the tangent stiffnesses of the cross-sections are determined as
usual. In the former case, computation is performed considering the previous equation, delivering
⎧ (1) ⎫ ⎧ (1) ⎫
⎪
⎪ Δq ⎪ ⎪ ⎪
⎪ ⎪
⎪ ⎪ ⎨Δq ⎪
⎪ ⎪
n−1 n
⎨ ⎬ ⎬
Δv(1) = wgti L(1) f1(1) Δq(1) + wgtn L(1) f2(1) ⎪
⎪ ⎪
⎪ and Δv(2) = wgti L(2) f1(2) Δq(2) + wgt1 L(2) f2(2) ⎪
⎪ ⎪ (6)
i=1
⎪
⎩Δq ⎭
(2) ⎪
i=2 ⎩Δq ⎪
⎪ (2) ⎪
⎭
in which wgti is the ith integration point weight. Adding both equations, results {Δv(1) Δv(2) }T = f · {Δq(1) Δq(2) }T , in
which f is the material flexibility matrix of the three-node FE. The linear compatibility matrix for the adopted FE is
Paulo Silva Lobo et al. / Procedia Engineering 114 (2015) 792 – 799 795
given by {v(1) v(2) }T = a(1,2) ū(1,2) , where a(1,2) is the compatibility matrix and ū(1,2) is the vector of nodal displacements.
Any level of geometrical nonlinearity consistent with the small-deformations/large-displacement theory may be added
to the model. The analysis reported herein was of the P-Δ type. The element state determination may be computed
with a residual displacements based method such as the Spacone et al. [1] method.
The reference results adopted in this section are the test results of the structure assessed by Clough and Gidwani
[11]. These researchers chose to study a structure with predominant flexural conditions, which is in agreement with the
conditions also considered in this research. This structure was a model of a typical apartment or office building, built to
a length scale of 0.7, consisting of a two-storey, one bay framed structure, with the ground accelerations applied along
the direction of the frame. It was designed in order to obtain maximum ductility. With the objective of simulating the
mass of a typical building and to obtain representative values of the period of vibration, concrete weights were added
to each floor (see Fig. 2a). The construction of two identical, parallel and connected frames allowed the stabilization
Fig. 2: Geometry, arrangement on shaking table and reinforcement of the test structure (dimensions in meters); (a) and (b) in the image caption
of the direction perpendicular to the ground motion. Also, the inclusion of the slab on the top of the girders allowed the
simulation of the flange action conferred by these elements. The rebar detailing is presented in Fig. 2b. The average
values of the characteristics of the materials obtained from tests were reported [11]. The reinforcement consists of
deformed steel bars with the exception of the rebars used for hoops and stirrups. The relevant values obtained for steel
are presented in Table 1. σy is the yielding stress, ε sh is the limit strain of the yield plateau, εu is the ultimate strain
Table 1: Average test values of the characteristics of the reinforcement bars of the test structure [11].
φ E σy ε sh εu σu
(mm) (GPa) (MPa) (‰) (‰) (MPa)
9.525 195.8 358.5 28.9 193 500.6
12.7 193.1 386.8 17.6 140 578.5
15.875 205.5 286.1 9.4 113 499.9
796 Paulo Silva Lobo et al. / Procedia Engineering 114 (2015) 792 – 799
and σu is the ultimate stress. The average values of the unconfined concrete compressive strength, fco = 30.3 MPa,
and of the longitudinal strain at unconfined concrete peak stress, εco = 3.35‰ , were obtained just before the shaking
table test was conducted, with a loading rate of the cylinders of about 0.252 MPa/s.
The ground motion record used is the N69W accelerogram recorded at Taft during the Arvin-Tehachapi earthquake
in 1952. The structure was subjected to a low intensity shake, with peak ground acceleration of 9.7% g, with the
objective of inducing a normal degree of cracking of common real structures. A second high intensity shake with a
peak ground acceleration of 57% g, capable of causing extensive damage, was simulated [11].
Given the structural symmetry, only one frame was numerically assessed. A scheme of the model of the structure
is depicted in Fig. 3. The gray squares represent the nodes of the model. Elements 1, 2, 5, 7, 8, 9, 10 and 12 were
(a) Discretization and geometry (b) Vertical loads and reactions (c) Lumped masses
Fig. 3: Scheme of the model of the test structure; (a), (b) and (c) in the image caption
modelled as two-node FEs with bond-slip in the exterior joints. Elements 3 and 4 are three-node FEs with bond-slip in
the interior beam-column joints. Elements 6 and 11 are fiber-section force-based FEs with perfect bond. The bottom
column-ends at the footings were considered to be fixed and the columns and girders axis were considered at the
centroid of the gross concrete cross-sections. The loads due to the weight of the elements were lumped together with
the loads due to the added concrete masses and at the girders exterior nodes. The mass due to self-weight was lumped
at the corresponding nodes and, as visible in Fig. 3c, elements were added to allow the allocation of the concrete
block masses at their real position. The added trusses, which modelled the support conditions of the concrete blocks,
are composed of massless linear elastic elements with high axial stiffness, pinned at both their ends. The beams and
columns were simulated with four control-sections for each beam-column, defined according to the Gauss-Lobatto
integration scheme. In the case of the three-node FEs, a total of eight control sections per element was considered.
All rebars pertaining to exterior joints were modelled following the Monti and Spacone model [2] and for the interior
joints the continuous anchored bar fiber model [9] was considered. The concrete cross-sections were discretized into
fibers, following the Gauss-Lobatto scheme. In the case of the columns, 10 integration points were considered, and
regarding the beams, the flange was modelled with five control points and the web with 10 fibers.
The Monti and Nuti [12] model was considered for steel. For the column rebars, the kinematic strain hardening
ratio was taken equal to 0.035 , while for the girder rebars a negligible value fo this parameter was assumed. The
model by Martı́nez-Rueda and Elnashai [13] was used for concrete and the concrete tensile strength was neglected.
The confined concrete properties, according to the work by Mander et al. [14], were considered for the fibers inside
the hoops. Moreover, the average strain rate of the cylinder tests previously mentioned was close to the quasi-static
strength. As such, the dynamic values of strength, elastic modulus and strain at peak stress were obtained considering
a strain rate of 1.67 × 10−2 s−1 [14]. Bond-slip was modelled following the Eligehausen et al. proposal [15], with
constants obtained with the empirical formulas developed by Monti et al. [16]. Hooked bars bond-slip was modelled
following the proposal by Eligehausen et al. [17]. All anchorages, either interior or exterior, where modelled with
five FEs, with four Gauss-Lobatto integration points for every FE.
Because the nonlinear material behaviour is modelled according to the materials constitutive relations, and because
for structures with highly nonlinear response viscous damping is of small importance when compared to hysteretic
damping, in this analysis Rayleigh damping was considered, for numerical stability sake, with a negligible value,
ζ = 0.1%, for the first two natural vibration modes, with α M and αK determined for the frequency values 3.13 and
Paulo Silva Lobo et al. / Procedia Engineering 114 (2015) 792 – 799 797
8.70 Hz measured in the reported snap tests [11]. For the numerical integration of the equations of motion, the Hilber,
Hughes and Taylor α-method was used with α = −0.05. A time step Δt = 0.005 s was adopted.
The perfect bond frame model was developed with the exact same characteristics of the bond-slip numerical model
described above, using the fiber beam FE for seismic analysis of RC by Spacone et al. [1].
The correlation between the results of the model previously described and the experimental results was made in
terms of bottom-storey and top-storey displacements. For evaluation of the influence of bond-slip, the results of
the analysis of the model with perfect bond are also presented. The bottom-storey and top-storey displacements
comparison with experimental results are depicted in Fig. 4 and 5. Good agreement between experimental and
analytical results can be seen. The period of vibration and the waveform obtained with the numerical model correlate
well with the test results. Also, the prediction of peak-displacements is very satisfactory.
For evaluation of the effect of bond-slip on the response, the displacements obtained with the models with and
without bond-slip are depicted in Fig. 6 and 7. The results of the model without bond-slip correspond in general to
smaller displacements and the results present a significantly inferior correlation with the experimental data. Thus, the
influence of bond-slip is significant and its consideration made it possible to obtain a significantly better response
prediction.
For further analysis of the results, the shear-displacement relation with and without bond-slip of the left column
of the bottom-storey are depicted in Fig. 8. As expected, the energy dissipation obtained with the consideration of
perfect bond is significantly higher than that of the more sophisticated model, including rebars bond-slip behaviour.
798 Paulo Silva Lobo et al. / Procedia Engineering 114 (2015) 792 – 799
Fig. 6: Bond-slip effect on the bottom-storey displacement response of the test structure
Fig. 7: Bond-slip effect on the top-storey displacement response of the test structure
The markedly more pinched response of the bond-slip model is noted, consistent with the results for static cyclic loads
of the applied models previously reported [2,9]
Fig. 8: Analytical bottom-storey shear-displacement relation of the left-side column of the test structure; (a) and (b) in the image caption
Paulo Silva Lobo et al. / Procedia Engineering 114 (2015) 792 – 799 799
4. Conclusions
Fiber-section models for nonlinear analysis of structures subjected to strong cyclic actions were used to assess the
response of a RC structure. The objective of this research was to evaluate the effect of bond-slip in the prediction
of the response of RC frame structures, thus two models were considered: a frame FE model considering perfect
bond and a model considering bond-slip in both exterior and interior joint regions. The numerical analyses performed
are of the nonlinear dynamic type, considering nonlinear geometric and material behaviour, and the reference results
adopted are from a shaking table test of a two-storey RC frame structure. As the correlation of experimental and
analytical results show, the model with bond-slip in the vicinity of interior and exterior joints is capable of predicting
the structural behaviour of RC frame structures subjected to dynamic loads in a very satisfactory manner. It is stressed
that the model was developed based only on the reported material and geometric properties of the reference test frame,
without any calibration procedures. The influence of bond-slip on the response was also assessed by comparison of
the results of the more sophisticated model with the results of the model with perfect bond. This comparison makes
it possible to conclude that the model with bond-slip is significantly more accurate than the model with perfect
bond. Also, the response in terms of shear-storey displacement was assessed, showing that the pinching effect in
the hysteretic cycles, an effect which is observable in experimental tests of RC structures and that has significant
influence on the overall structural response as it reduces the energy dissipated, is noticeable in the response obtained
with the bond-slip model, unlike for the perfect bond model. The numerical model robustness and efficiency, as well
as accuracy, was shown with the assessed example, adding to previously reported applications.
References
[1] E. Spacone, F. C. Filippou, F. F. Taucer, Fibre beam-column model for non-linear analysis of R/C frames: Part I. Formulation, Earthquake
Engineering and Structural Dynamics 25 (7) (1996) 711–725, doi:10.1002/(SICI)1096-9845(199607)25:7<711::AID-EQE576>3.0.CO;2-9.
[2] G. Monti, E. Spacone, Reinforced concrete fiber beam element with bond-slip, J. Struct. Engrg., ASCE 126 (6) (2000) 654–661,
doi:10.1061/(ASCE)0733-9445(2000)126:6(654).
[3] G. Monti, E. Spacone, F. C. Filippou, Model for anchored reinforcing bars under seismic excitations, Tech. Rep. UCB/EERC-93/08, Earthquake
Engineering Research Center, College of Engineering, University of California, Berkeley, California (December 1993).
[4] S. Limkatanyu, E. Spacone, Effects of reinforcement slippage on the non-linear response under cyclic loadings of RC frame structures, Earth-
quake Engineering and Structural Dynamics 32 (15) (2003) 2407–2424, doi:10.1002/eqe.334.
[5] N. R. Rubiano-Benavides, Prediction of the inelastic seismic response of concrete structures including shear deformations and anchorage slip,
Doctoral dissertation, University of Texas, Austin, Texas (1998).
[6] M. D’Amato, F. Braga, R. Gigliotti, S. Kunnath, M. Laterza, Validation of a modified steel bar model incorporating bond-slip for seismic
assessment of concrete structures, Journal of Structural Engineering 138 (11) (2012) 1351–1360. doi:10.1061/(ASCE)ST.1943-541X.0000588.
[7] F. Braga, R. Gigliotti, M. Laterza, M. DAmato, S. Kunnath, Modified steel bar model incorporating bond-slip for seismic assessment of concrete
structures, Journal of Structural Engineering 138 (11) (2012) 1342–1350. doi:10.1061/(ASCE)ST.1943-541X.0000587.
[8] H. G. Kwak, F. C. Filippou, A new reinforcing steel model with bond-slip, Struct. Eng. Mech. Int. J. 3 (4) (1995) 299–312, doi:10.1002/eqe.334.
[9] P. Silva Lobo, J. Almeida, RC fiber beam-column model with bond-slip in the vicinity of interior joints, Engineering Structures 96 (0) (2015)
78–87. doi:10.1016/j.engstruct.2015.04.005.
[10] S. Bae, O. Bayrak, Plastic hinge length of reinforced concrete columns, ACI Structural Journal 105 (3) (2008) 290–300, doi:10.14359/19788.
[11] R. W. Clough, J. Gidwani, Reinforced concrete frame 2: Seismic testing and analytical correlation, Tech. Rep. EERC-76/15, Earthquake
Engineering Research Center, College of Engineering, University of California, Berkeley, California (June 1976).
[12] G. Monti, C. Nuti, Nonlinear cyclic behavior of reinforcing bars including buckling, J. Struct. Engrg., ASCE 118 (12) (1992) 3268–3284,
doi:10.1061/(ASCE)0733-9445(1992)118:12(3268).
[13] J. E. Martı́nez-Rueda, A. S. Elnashai, Confined concrete model under cyclic load, Materials and Structures 30 (1997) 139–147,
doi:10.1007/BF02486385.
[14] J. B. Mander, M. J. N. Priestley, R. Park, Theoretical stress-strain model for confined concrete, J. Struct. Engrg., ASCE 114 (8) (1988) 1804–
1826, doi:10.1061/(ASCE)0733-9445(1988)114:8(1804).
[15] R. Eligehausen, E. P. Popov, V. V. Bertero, Local bond stress-slip relationships of deformed bars under generalized excitations, Tech. Rep.
UCB/EERC-83/23, Earthquake Engineering Research Center, College of Engineering, University of California, Berkeley, California (October
1983).
[16] G. Monti, A. De Sortis, C. Nuti, Problemi di scala nella sperimentazione pseudodinamica di pile da ponte in C.A., in: Workshop Danneggia-
mento, Prove Cicliche e Pseudodinamica, Napoli, Italy, 1994.
[17] R. Eligehausen, V. V. Bertero, E. P. Popov, Hysteretic behavior of reinforcing deformed hooked bars in R/C joints, in: 7th Eur. Conf. on
Earthquake Engrg., Vol. 4, Technical Chamber of Greece, Athens, 1982, pp. 171–178.