You are on page 1of 17

Engineering Structures 247 (2021) 113214

Contents lists available at ScienceDirect

Engineering Structures
journal homepage: www.elsevier.com/locate/engstruct

Cyclic response of column to foundation connections of reinforced concrete


precast structures: Numerical and experimental comparisons
R. Nascimbene a, *, L. Bianco b
a
Scuola Universitaria Superiore IUSS Pavia, Piazza della Vittoria n.15, 27100 Pavia, Italy
b
PEIKKO Italia, Via Privata Giuseppe Ugolini, 30, 20125 Milano, Italy

A R T I C L E I N F O A B S T R A C T

Keywords: The cyclic response of a column to foundation connection system, based on the mechanical connection between
Precast connections steel shoes embedded into the column base and protruding steel bolts anchored into the foundation, is experi­
Column shoes mentally and numerically examined in this paper. The hysteretic behaviour and the global collapse mechanism of
Pocket foundations
the connections are studied through three full-scale specimens subjected to a constant axial load and to a quasi-
Quasi-static cyclic test
static cyclic horizontal top displacement history at increasing drift levels. Additionally, advanced three-
dimensional finite element models, including non-linearity in material and geometry, are developed and cali­
brated in order to analyze and capture the local experimental response of the connections. Due to the lack of
information and guidelines on the subject, the aim of this research is to evaluate the seismic performances of the
system in order to evaluate the real response of the welded connection between steel shoes and reinforcing bars.

1. Introduction such seismic events pointed out the weakness of connection systems for
both structural elements [20] and non-structural components [21]. As a
In the last two decades, research has shown remarkable progress in main consequence of these events [19,22,23,24], advanced numerical
the context of prevention and seismic protection of structures, in order models and analyses [25,26] and updates in seismic codes identify in the
to ensure the development of modern codes [1,2,3,4,5]. The main goal connections the weak links of the precast structures. Detailed target
of the most advanced codes is the protection against seismic risk through connection performances in terms of strength, stiffness and ductility
the application of performance-based design (PBD) and in particular the demand are consequently required. A specific column-foundation
control of the level of accepted damage at each considered limit state. connection system, classified as direct connection type, is examined in
Within the framework of performance-based seismic design of precast this research. In precast single or multi-storey buildings, a pocket
structures, significant improvement has been recently made in order to foundation is the typical column-to-foundation connection. The column
develop innovative solutions [6,7,8,9], new methodologies, and new is inserted into a special pocket arising from the base footing and the
structural systems in a way that the capacity design objective could be final joint is filled with cast-in-place concrete. The performance and
achieved. Very recent experimental and numerical results suggest that efficiency of this traditional connection have been extensively tested and
prefabricated systems equipped with steel shear keys present reliable described by Saisi and Toniolo in [27]. The main disadvantage of this
stiffness and bearing capacity and improved energy dissipation ability, column-to-foundation connection is that, in the case of multi-storey
compared with conventional systems [10,11,12,13]. large size columns, the pocket foundations become so big that their
Precast concrete systems offer many potential advantages over cast transportation turns out to be difficult and, hence, a large excavation is
in-situ structures due to the speed of construction, higher performance required. Recent earthquakes, hitting regions with moderate or
of materials, plant certification and more [41]. Unfortunately, very advanced structural engineering practice in precast construction,
recent earthquakes in Italy, and more in general in Europe, stressed the pushed the market in finding advanced solutions for column-foundation
importance of adequate seismic design and vulnerability assemblies: lapped bars, welded splices, bolted coupling or mechanical
assessment approaches as well as advanced technological solutions splices, innovative technologies and quicker assemblages to avoid in-
for precast industrial structures [14,15,16,17,18,19]. The experience of situ concreting. All these solutions are characterised by dry or semi-

* Corresponding author.
E-mail addresses: roberto.nascimbene@iusspavia.it (R. Nascimbene), lorenzo.bianco@peikko.com (L. Bianco).

https://doi.org/10.1016/j.engstruct.2021.113214
Received 17 February 2021; Received in revised form 8 August 2021; Accepted 12 September 2021
Available online 24 September 2021
0141-0296/© 2021 Elsevier Ltd. All rights reserved.
R. Nascimbene and L. Bianco Engineering Structures 247 (2021) 113214

Fig. 1. Column steel shoes and anchoring bolts in column connections and additional steel ordinary reinforcement in square columns (welded bars and lap splice
along the height).

dry assemblages that can be mainly classified into two main typologies: because, as highlighted by Belleri and Riva in [39] and Popa et al. in
grouted sleeve connections and grouted corrugated duct connections. [40] limited information are reported in scientific literature.
The former uses a typical grouted mechanical coupler able to provide The novelty of the system presented herein that was experimentally
continuity of steel bars between prefabricated units, in order to develop investigated is based on the mechanical connection between steel shoes
the full strength of the deformed bars [28]. Generally, in this connec­ embedded into the column base and the protruding anchor bolts
tions with bar couplers [29,30,31,32,33,34,35], the grouted sleeves are anchored into the foundation. The use of nuts and washers attached to
made of steel and filled with grouting material after the insertion of the the anchor bolts permits to control the vertical position, the height level
reinforced bars. Very recently, cyclic load experimental tests on a of the column and the fixity of the connection. Furthermore, an
column-to-foundation joint using threaded mechanical connectors were advanced numerical model has been built in order to: 1) analyze and
presented by Dal Lago et al. in [36]. The second type of connection capture the local experimental response of the connections; 2) compare
[37,38], the grouted corrugated duct splices, is characterized by some it to the behaviour of a cast-in place equivalent connection; 3) predict
bars protruding from the precast unit foundation and inserted into the the response of similar connection typologies to cyclic loading.
corrugated ducts encased in the column.
This paper presents an experimental study on a connection where the 2. Experimental investigation
forces are transferred from the column to the foundation through a
system of steel components embedded into the foundation and the col­ An experimental program focusing on full scale connections in pre­
umn and connected between them through different techniques [42]. cast concrete structures has been developed at the European Centre for
More precisely, four or more steel plates (steel shoes) are embedded into Training and Research in Earthquake Engineering (Eucentre). This
a precast reinforced concrete column and connected to the concrete paper describes the comparison between the experimental results and
foundation (Fig. 1) through a system of protruding threaded anchoring the numerical analyses of non-linear models. The most relevant aspects
bolts, nuts and washers. The nuts provided below the base of the steel that were investigated include: the evaluation of the actual response of
shoes permit to control both column plumbness and the height of the the welded connection in the steel column shoe, in order to demonstrate
void between the column and the foundation, which is filled by an that the welds remain elastic; the definition of the actual global collapse
additional injection of cement mortar [43]. The stress transfer from the mechanism and of the displacement ductility and dissipation capacity;
steel shoes to the column is based on a couple of longitudinal bars the capacity of the base section of the column to resist shear actions; the
welded at the top of each shoe and on additional overlapping bars, initial stiffness of the column compared to the
which represent the steel reinforcement of the column (Fig. 1). The stiffness of other precast structural typologies.
seismic performance of the system has been evaluated by means of a
numerical and experimental research, the main objectives of which are:
a) the evaluation of the response of the welded connection between steel 2.1. Specimens
shoes and reinforcing bars, justified by the lack of information and
guidelines; b) the evaluation of the global failure mechanism; c) the A reinforced concrete precast three-storey rectangular industrial
study of the characteristics of the seismic response; d) the establishment building was considered as a reference prototype structure in order to
of the hysteretic behaviour of such column-to-foundation connections design the specimens. The case-study is characterized by monolithic
beam-to-column connections and span lengths equal to 14 m and 8 m

2
R. Nascimbene and L. Bianco Engineering Structures 247 (2021) 113214

Fig. 2. (a) Geometrical dimensions and details of the steel reinforcement used in the connection; (b)-(c) base and mid-height side; (d) column detail of the specimens.

Fig. 3. Column to foundation details: 4 ϕ 24 mm steel pins added to resist shear action.

along the two principal directions, respectively. The total height is 12 m, anchoring bolts. The additional steel longitudinal and transversal rein­
subdivided into three interstorey heights equal to 4 m. The assumptions forcement along the height of the column was designed to resist the
of high ductility class, peak ground acceleration value of 0.25 gand soil distribution of bending moment and shear actions, respectively, without
composed of medium dense sands were made during the development of abrupt interruption of the longitudinal rebars. The amount of symmet­
the design phase. Square columns with cross sections of 40 × 40 cm were rical distribution of longitudinal reinforcement is depicted in Fig. 2. Four
designed and connected to the foundation through steel shoes and steel steel pins were designed and embedded into the column (Fig. 3) to

3
R. Nascimbene and L. Bianco Engineering Structures 247 (2021) 113214

Table 1
Compressive cubic strength of concrete.
Specimen ID Dimensions [mm] Type of element Weight [kN] Compressive strength [kN] Compressive strength [MPa]

1 150 × 150 × 150 Foundation slab - cast in place 83.00 1180 52.44
2 82.80 1200 53.33
3 82.80 1240 55.11
4 82.00 1230 54.67
1 150 × 150 × 150 Foundation slab - cast in place 80.90 1260 56.00
2 81.10 1250 55.56
3 81.10 1260 56.00
4 80.70 1250 55.56
1 150 × 150 × 150 Foundation slab - cast in place 81.00 1270 56.44
2 80.40 1260 56.00
3 81.30 1270 56.44
4 81.30 1270 56.44
1 150 × 150 × 150 Column 80.80 1360 60.44
2 82.10 1380 61.33
3 82.00 1390 61.78
1 150 × 150 × 150 Column 81.00 1370 60.89
2 81.50 1400 62.22
3 81.90 1400 62.22
1 150 × 150 × 150 Column 82.80 1460 64.89
2 82.50 1450 64.44
3 83.00 1470 65.33

Fig. 4. Specimen before the test and MTS 210 testing machine.

prevent shear demands to the anchor bolts. Alternatively, precasters use phase, to be less than 0.4, in order to avoid problems related to low
a steel profile in the centre of the section of the column, embedded into displacement ductility values. The range of values selected is in
the column and the foundation. The aim of both the steel pins and the accordance with a conservative choice;
steel profile is to impose to the anchor bolts only axial forces and avoid - Boundary conditions (base): the specimens have to be considered
shear actions. fixed at the base corresponding to the laboratory floor; the founda­
The main characteristics of the specimens can be summarized in the tion has been appropriately overdesigned and heavily reinforced in
following list: order to impose this condition and to inhibit cracking phenomena;
- Boundary conditions (top): the top of the specimen is free. It means
- Amount: 3 specimens with constant square cross section (400 mm) that the experimental configuration is representative of a column
with four steel shoes placed at the base; with height of 4.30 m subjected to double curvature, or to a column
- Height of the column (lever arm of the force at the top of the spec­ with height of 2.15 m fixed at the base and free at the top;
imen): 2.15 m, constant for all the specimens; - Materials: the values of the compressive strengths obtained from the
- Axial load: three different levels of axial load, corresponding to a tests on cubes of concrete (in accordance with the UNI provisions)
ratio to the critical compressive axial load of 0.05, 0.10, and 0.15. are summarized in Table 1. For mild steel, the yield strength is fyk =
The maximum value of this ratio should be assumed, in the design 500MPa and the corresponding design value fyd = 435MPa and the

4
R. Nascimbene and L. Bianco Engineering Structures 247 (2021) 113214

Fig. 5. Example of arrangement of a specimen with Φ25mm ribbed bar and Φ30mm anchor bolt and test setup.

yield strain is εse = 0.002; the steel corresponding to the shoes is


Table 2
classified as grade S500. Furthermore, an experimental campaign
Drift and displacement levels at the top of the column.
has provided information regarding single steel rebar coupons
(Fig. 4) and the behaviour of the global steel system (shoe – welded Drift level Displacement in [mm.]
bars – overlapped bar – steel rebar coupon) subjected to axial load ± 0.40 ± 8.60
and embedded in a concrete cylinder (i.e., the single connection ± 0.80 ± 17.20
± 1.20 25.80
system), neglecting the failure of the anchor bolt (Fig. 5). ±
± 2.40 ± 51.60
± 3.60 ± 77.40
Three single steel rebars were tested (Fig. 4), consisting of one cy­ ± 4.80 ± 103.20
lindrical sleeve with external diameter equal to 80 mm, made of C40
processed steel; three ribbed steel bars with an equivalent diameter of
25 mm and a M30 threaded end, 40 mm long; four 25 mm-thick rect­ strength of the concrete and Ac the area of the cross section). In
angular plates made of Fe510 steel. The external dimensions are 200 × particular, the following requirements have been satisfied:
300 mm. For all three specimens, the failure mechanism was charac­
terized by a large elongation of the bar (about 50 mm), without any - The thickness of the concrete cover is 40 ÷ 46 mm;
damage to the connection. The maximum measured strength was 309 - The concrete is non-shrinking type;
kN. The ratio between the maximum strength and the yielding strength - The grouting mortar strength must be at least 70% of the concrete
(262 kN) was 1.18. The displacement at maximum load was 34.7 mm, grade adopted for the column;
while the ultimate displacement was 49.7 mm. The ratio between the - The transversal steel reinforcement of the lap-splices has been
maximum load displacement and the yielding displacement was 1.43. designed and verified according to Eurocode 2 (Point 9.5.3 and
The effective ultimate strain, measured conventionally on a gage length 8.7.4.1) [44].
of five diameters, was about 20%.
Regarding the global steel system, the information obtained from the 2.2. Instrumentation and test equipment
tests depicted in Fig. 5 are:
- about welds: the failure mechanism of the system is not governed by The experimental loading history consists of a series of horizontal
the response of the welds, since they did not suffer damage and, displacement cycles at the top of the column with increasing target drift
consequently, are not the weak link of the whole system. This result as reported in Table 2. The vertical distance between the axis of the
positively influences also the energy dissipation capacity, avoiding actuator and the base of the column is 2150 mm; it is expected that the
brittle responses; shear component of the displacement profile of the specimen will be
- about the steel system: the transducers used in this campaign of negligible, due to the value of the ratio between the height and the depth
tests permitted to measure the axial deformation up to 1%. In general, of the cross section of the column (corresponding to 5.37). The test
the load-deformation curves show that the longitudinal bars have arrangement is depicted in Fig. 6; furthermore the displacement shape
similar initial stiffness, whereas the lap-splice and the shoe are more corresponding to a drift level of 4.8 % (Table 2) is represented in Fig. 6
deformable (about twice as much as the longitudinal bars). Indepen­ (b). The instrumentation depicted in Fig. 6(c)-(e) has been arranged in
dently of the initial stiffness, the overlapping bars and the lap-splice order to measure relative displacements, deformations and curvatures of
yield first. the cross section at different height levels.
The connections were designed to resist bending moments of 230
kNm and 160 kNm along the two orthogonal directions (biaxial 3. Summary of the test results
bending). This investigation was restricted to three constant vertical
loads with no eccentricity imposed at the top of the column during the The main results obtained from the experimental campaign on three
tests: 200 kN, 400 kN and 600 kN, corresponding to the values of the specimens is briefly summarized as follows:
non-dimensional axial force described in the previous list (obtained by
the ratio AcNfcd , where N is the axial load, fcd the design compressive

5
R. Nascimbene and L. Bianco Engineering Structures 247 (2021) 113214

Fig. 6. (a) Global test set-up; (b) specimen at a drift of 4.8 % (displacement 103.20 mm); (c)-(e) instrumentation of the connection perpendicular and parallel to the
load direction.

- The behaviour of the anchor bolts governs the collapse mechanisms, - The equivalent viscous damping (sum of viscous and hysteretic
without any significant damage to the specimen, independently of components) is 8.5 % up to a drift level of 1.2 %, then it increases and
the axial load level imposed at the top of the column; is in the range of 16 % − 20 % for the first cycles;
- The non-linear branch of the base shear – top displacement curve - The incipient yielding condition of the columns is attained at a drift
(Fig. 7 corresponding to the three specimens) is due to the yielding of level in the range of 1 % − 1.3 %.
the anchor bolts, which are the only components of the system which
have a non-linear response, whilst the other components do not
exceed the yielding condition and are not significantly damaged;
3.1. Strength degradation
- The hysteretic behaviour, evaluated up to 4.8 % drift level, is char­
acterized by negligible strength degradation and by a global ductility
No strength degradation is evident up to a 3.6% drift level. After that,
level always greater than 4.5;
a degradation less than 5% is visible from 3.6% to 4.8% drift. These

6
R. Nascimbene and L. Bianco Engineering Structures 247 (2021) 113214

Fig. 7. Experimental capacity curve: base shear and horizontal top displacement of the specimens subjected to a vertical load of (a) 200 kN, (b) 400 kN and (c) 600
kN; damage pattern at the end of the test.

7
R. Nascimbene and L. Bianco Engineering Structures 247 (2021) 113214

Fig. 8. Loading branches of the cycles at each drift level obtained from the experimental hysteretic response curve.

results are compatible with a ductile behaviour of a flexural mechanism. 3.3. Stiffness degradation
Also, the strength degradation between the first and the other cycles is
less than 5%. The specimens reach the maximum strength without evident
strength degradation for each cycle during the entire duration of the
3.2. Ductility tests, although the loading stiffness of the second and third cycles is
lower than the stiffness of the first cycle for each drift level (Figs. 7 and
The tests were ended at a drift level equal to 4.8% because higher 8).
displacements were not considered useful for design purposes. For this In the force based design (FBD) methodology, a structure is usually
reason, the flexural response observed during the experimental tests, defined using the initial stiffness of its equivalent single degree of
which is associated to a displacement ductility level of 4.5, can be sus­ freedom (SDOF) system. In the direct displacement
tained by the system. In any case, a displacement ductility level of 4.5 is based design (DDBD) methodology, a structure is well defined ac­
a very good option in order to design these type of connections. The most cording to the secant stiffness of its equivalent SDOF system [45]. As a
important results in terms of attaining high ductility values, is the small consequence, based on the experimental data, the initial and secant
incipient yielding displacement (1.0% − 1.3%), when compared to stiffness values have been considered for further investigation. In order
typical values of 1.8% − 2.0% [8] that are usually associated to precast to evaluate the initial stiffness at each cycle, a number of calculations
column-foundation connections. were required to quantify the stiffness degradation of the specimen upon
repeated cycling. The loading excursions at each cycle were extracted
from the complete hysteretic force–displacement curve and then

8
R. Nascimbene and L. Bianco Engineering Structures 247 (2021) 113214

13752
14000

12025
Cycle 1

11019
Cycle 2

Energy dissipation by each cycle


12000

9667
Cycle 3
10000

7305
8000

6459
5946
6000

3369
2867
4000

1146
2000

682
548

586
297
232

242
130
111
0
0.4 0.8 1.2 2.4 3.6 4.8
Drift (%)
(a)

31291
35000

Cycle 1
Cycle 2

23808
Cumulative Energy dissipation

28000

21284
Cycle 3

17539
21000

11783
10265
14000
7872
4478

7000
3806
1926
1109
939
780
427
353
232
130
111

0
0.4 0.8 1.2 2.4 3.6 4.8
Drift (%)
(b)
Fig. 9. (a) Energy dissipation and (b) cumulative energy dissipation by the tested specimen.

overlain to start from the same zero displacement and force, and then
finally compared, as shown in Fig. 8. Fm

3.4. Equivalent viscous damping Ah = Shaded area

The area enclosed by a complete lateral force–displacement cycle Δm


represents the amount of energy dissipated by the system. Fig. 9 in­
dicates the energy dissipated and cumulative energy dissipated by cycles Displacement
at different drift levels of the tested specimen, respectively. It illustrates
that even though during the tests the column specimen essentially re­
mains in the elastic range through the first three levels of drift, a Fig. 10. Hysteretic area for damping calculation.
comparatively lesser amount of energy has been dissipated through the
cracking of concrete. Furthermore, it indicates that energy dissipation
where Ah is the area of a complete cycle of force–displacement response,
reduces with the increase of the number of constant amplitude cycles for
Fm and Δm are the maximum force and the displacement occurred in the
a given drift level.
complete cycle respectively. The values obtained from Eq. (1) are then
It is worth to represent the dissipated energy by means of the
decreased using the correction factors proposed in [47] as a function of
equivalent viscous damping. Traditionally, the equivalent viscous
the area and of the ductility level. Figs. 10 and 11 show the graphical
damping is calculated using the area based approach [46,47] in which
representation of these parameters and Fig. 12 illustrates the equivalent
damping is calculated by equating the energy absorbed by the hysteretic
viscous damping obtained from test specimen, for cycles at different
steady-state cyclic response at a given displacement level as expressed
drift levels.
by:
The global response is almost linear if the values of the drift level are
ξhys =
Ah
(1) lower than 1.2%, then it becomes non-linear with a ductile mechanism,
2πFm Δm without strength degradation up to the end of the test (drift equal to

9
R. Nascimbene and L. Bianco Engineering Structures 247 (2021) 113214

247,892 and the total number of degrees of freedom is 866058. Figs. 14


and 15 show the arrangement of the four shoes at the base of the column
and the embedded reinforcement bars in the longitudinal and trans­
versal directions.

4.2. Material description: concrete and steel

Unlike the elastic material, plastic material exhibits permanent


deformation during loading and unloading [48,49]. In order to reflect
such properties, material models should be capable to simulate the
elastic and plastic behavior. Therefore, the total strain crack model and
the Von Mises plasticity model have been used in the numerical model to
simulate the elastic and plastic behavior of the materials considered.
The total strain crack model has been developed along the lines of the
modified compression field theory, originally proposed by Vecchio and
Collins [50] and then extended to the three-dimensional case by Selby
and Vecchio [51]. Two main approaches can be selected: a rotating
crack model in which the stress–strain relationships are evaluated in the
Fig. 11. Correction factor for area-based equivalent damping [18]
principal directions of the strain vector and a fixed crack model in which
the stress–strain equations are evaluated in a fixed coordinate system
4.8%). For that reason, a low damping is associated with the first 9 which is fixed upon cracking. In this research this second approach was
cycles. For the next cycles up to the end of the test, the damping strongly used, because it can be considered to be more appropriate for the
increases to 15% − 17%. The test data clearly indicate that the equiv­ physical nature of cracking [52]. In a fixed stress–strain approach
alent viscous damping reduces as the number of constant amplitude
cycles at a given drift level increases. The reduction is dominant between
the first and the second cycle. A maximum reduction of 42 % is observed
at the 2.4 % drift level while a minimum reduction of 8 % is observed at
the 4.8 % drift level.

4. Numerical investigation

4.1. Geometry

For the detailed investigation of the performance of the precast


column-to-base connection that was experimentally tested under cyclic
loading, a refined finite element model (FEM) has been built using
tetrahedral four-node element. The numerical model has been divided
into main groups: concrete column, basement foundation including the
embedded reinforcement bars, four steel shoes, layer of mortar, longi­
tudinal and transversal mild steel along the column. The material non­
linearities of concrete and steel have been defined respectively using the
total strain crack model and the Von Mises yield criterion. A detailed
description of the material models have been reported in the following
sections. A refined mesh is used in the region where the large inelastic Fig. 13. Global finite element model of the experimentally tested specimen:
behavior is expected (Fig. 13): the total number of brick elements is column, shoes, mortar and basement foundation.

0.2
0.173

Cycle 1
0.159
0.158
0.151

0.150

Cycle 2
Equivalent Viscous Damping

0.15 Cycle 3
0.125
0.113
0.088
0.088

0.1
0.077
0.072
0.064
0.050
0.044

0.044
0.038
0.036
0.029

0.05

0
0.4 0.8 1.2 2.4 3.6 4.8
Drift (%)

Fig. 12. Equivalent viscous damping obtained from experimental data.

10
R. Nascimbene and L. Bianco Engineering Structures 247 (2021) 113214

Fig. 14. Finite element model of the shoes connection between the column and the basement foundation.

Fig. 15. Finite element model of the embedded reinforcement in the concrete column: longitudinal and transversal direction.

normal stresses and the shear stresses exist on the crack surfaces, resistance at the crack locations. Due to the lateral confinement (i.e.
therefore the total strain crack model used in the present study requires compressive stresses) concrete shows a pressure-dependent behavior (i.
three stress–strain relationships for tensile, compressive and shear e. the strength and ductility increase with increasing isotropic stress) so
behavior. This means that, during loading, the concrete is subjected to the compressive behavior is modified to incorporate the effects of the
both tensile and compressive stress which can result in cracking and increased isotropic stress. The Thorenfeldt et al. [53] model is used.
crushing of the material and that the shear behavior is modeled The material nonlinearity of steel is represented by the classical Von
explicitly with a bilinear relationship between the shear stress and the Mises yield condition that can be considered a smooth approximation of
shear strain. The stress–strain relationship for the tensile behavior is the Tresca condition using a circular cylinder in the principal stress
simulated using an exponential softening decay curve of the tensile space.

11
R. Nascimbene and L. Bianco Engineering Structures 247 (2021) 113214

Fig. 16. Contact finite element model of the connection between: shoe and mortar, shoe and nut and concrete column and mortar.

4.3. Surface contact elements


150

1 The experimental pinching behavior that resulted from losing con­


100
tact between the nut and the shoe during the unloading and the
reloading, cannot be predicted by a fully continuous connection. In
50
Base shear (kN)

order to properly capture this phenomenon, discrete cracking contact


0
2 surface elements have to be introduced in the finite element model.
-150 -100 -50 0 50 100 150 Thus, contact surface elements have been introduced at each interface
-50
3 between the nut and the shoe and between the shoe and the mortar.
Furthermore, surface elements have been also introduced at the inter­
-100 face between the concrete column and the mortar layer as shown in
Fig. 16.
4
-150 A detailed explanation on how the combination of those contacts can
Top displacement (mm) represent the phenomenon of pinching (Fig. 7) shown by the experi­
mental base shear-top displacement hysteretic response will follow. For
Fig. 17. Base shear-top displacement experimental hysteretic response: key this purpose, several key points are selected from the base shear-top
points to describe interface behavior at the connection level. displacement hysteretic response as indicated by the red dots in
Fig. 17. As the lateral displacement increases from the unloaded con­
dition to the value corresponding to point 1, the compressive forces on
the left side of the contact elements between the nuts and the shoes and
the tensile forces on the contact elements on the same side between the

2 MPa 2 MPa
1

1 Displacemen Displacemen

(a) Interface between the nut and (b) Interface between the shoe and the
the shoe in left side of the system mortar layer in left side of the
Stress

Stress

2 MPa 2 MPa
1

Displacemen Displacemen
1
t

(c) Interface between the nut and (d) Interface between the shoe and the
the shoe in right side of the mortar layer in right side of the
Fig. 18. State of deformation, in term of displacement field, of each contact elements corresponding to point 1 in Fig. 17.

12
R. Nascimbene and L. Bianco Engineering Structures 247 (2021) 113214

2
2 MPa 2 MPa

Stress

Stress
2 1

Displacement Displacement
1

(a) Interface between the nut and the (b) Interface between the shoe and the
shoe in left side of the system mortar layer in left side of the system
Stress

Stress
2
2 MPa 2 MPa
1

Displacement Displacement
2 1

(c) Interface between the nut and the (d) Interface between the shoe and the
shoe in right side of the system mortar layer in right side of the system
Fig. 19. State of deformation, in term of displacement field, of each contact elements corresponding to point 2 in Fig. 17.
Stress

Stress

2
2 MPa 2 MPa
3 1
2

1 Displacement 3 Displacement

(a) Interface between the nut and the (b) Interface between the shoe and the
shoe in left side of the system mortar layer in left side of the system
Stress

Stress

2
2 MPa 2 MPa
1 3

Displacement 1 Displacement
2
3
(c) Interface between the nut and the (d) Interface between the shoe and the
shoe in right side of the system mortar layer in right side of the system
Fig. 20. State of deformation, in term of displacement field, of each contact elements corresponding to point 3 in Fig. 17.

shoes and the mortar layer are increased. When the tensile forces reach deformed shape model at the connection region corresponding to the
their peak resistance values, the contact between the shoes and the same point 1. It is very important to highlight that, even though the
mortar layer will start to disappear. In contrast to the behavior of the left three sets of contact elements remain elastic during the loading excur­
side, on the right side the elements between the nuts and the shoes, and sion as shown in Fig. 18, the yielding of the anchor bolts in tension
between the shoes and the mortar layer, tensile and compressive loads causes a drastic reduction of the stiffness at the overall displacement
rise, respectively. Fig. 18 reports the state of deformation of each contact level and creates a significant gap between the shoe and the mortar
with reference to point 1 (Fig. 17) in the global experimental hysteretic layer, as shown in Fig. 22(b).
response while Fig. 22 (b) shows an enlarged view of the finite element During the unloading excursion from point 1 to point 2 (Fig. 17), the

13
R. Nascimbene and L. Bianco Engineering Structures 247 (2021) 113214

Stress

Stress
2
2 MPa 2 MPa
3 1
4 3 2

1 Displacement 4 Displacement

(a) Interface between the nut and the (b) Interface between the shoe and the
shoe in left side of the system mortar layer in left side of the system
Stress

Stress
2
2 MPa 2 MPa
1 3
4
Displacement 1 Displacement
2
3
4
(c) Interface between the nut and the (d) Interface between the shoe and the
shoe in right side of the system mortar layer in right side of the system
Fig. 21. State of deformation, in term of displacement field, of each contact elements corresponding to point 4 in Fig. 17.

Fig. 22. (a) Comparison between experimental and numerical model and (b)-(e) deformed shape corresponding to the key points (Fig. 17).

compressive forces on the left side of the contact elements between the 2.
nuts and the shoes, and the tensile forces between the shoes and the During the unloading excursion from points 2 to 3 (Fig. 17), the gap
mortar layer, decrease. Note that the unloading branch of the created between the shoe and the mortar layer in the left side due to the
compressive forces on the left side follows the elastic stiffness degra­ elongation of the anchor rods during previous loading excursion must
dation while the unloading branch of the tensile forces on the left side close before the two surfaces come into contact. From Fig. 22(d), it is
follows the secant stiffness (Fig. 19). Similar to the loading excursion, clear that no connectivity exists between the nuts and the shoes and
the right side contact elements behave opposite to the left side. Fig. 22 between the shoes and the mortar layer in the left side of the system
(c) shows the deformed shape of the connection with reference to point when unloading from points 2 to 3. As a consequence of that, the system

14
R. Nascimbene and L. Bianco Engineering Structures 247 (2021) 113214

Fig. 23. Comparison between experimental and numerical model: crack evolution.

displays almost zero lateral stiffness and strength in the unloading column’s section with similar extensions, indicating a global unique
branch. Fig. 20 shows the deformation states of the interfaces with mechanism in the three specimens. The first cracks developed sym­
reference to the point 3. metrically with respect to the vertical axis of the column (Fig. 23(a))
Loading the column in the opposite direction (Fig. 21), the stiffness is after the target displacement of 17.20 mm (corresponding to a 0.8%
regained because the contacts between the shoes and the mortar layer in drift). The first horizontal crack with a length of 300 mm was noticed at
the left side are reactivated in compression after closing the gap. In large a height of 320 mm from the base of the column (Fig. 23(b)), after a
displacement analysis, the stiffness of the system reduces due to yielding target displacement of 51.6 mm (corresponding to a 2.4% drift). After a
of the anchor rods in the right side of the system. As mentioned before, 3.6 % lateral drift the bolt lost his pretension and the washer became
the yielding of the rods creates a significant gap between the shoes and oval; several local horizontal cracks formed at the two corners at the
the mortar layer in the right side as shown in Fig. 22(e). The comparison column base under compression (Fig. 23(c)).
between experimental and numerical data is depicted in Fig. 22(a).
During the loading phases, the three specimens (both experimentally 5. Conclusion and final remarks
and numerically), show similar crack patterns and failure modes, as
shown in Fig. 23. The damage was concentrated at the bottom part of the The cyclic response of a column-to-foundation connection system,

15
R. Nascimbene and L. Bianco Engineering Structures 247 (2021) 113214

characterized by steel components (steel plates embedded into the col­ improved the research.
umn connected to the anchor bolts, embedded into the foundation) has
been numerically and experimentally investigated and compared. The References
research work was focused on the evaluation of the capacity of the
system to resist seismic loading. [1] ACI (American Concrete Institute), Building Code Requirements for Structural
Concrete (ACI 318-19) and Commentary (ACI 318R-19), 2019.
The experimental results highlighted that the strength and the [2] AIJ (Architectural Institute of Japan) Japanese Architectural Standard
ductility levels are effective to resist seismic actions. In particular, the Specification JASS 10, Precast Concrete Work, 2013, 346p.
strength degradation is not particularly evident until displacement [3] CEN (European Committee for Standardisation), Eurocode 8: Design of structures
for earthquake resistance - Part 1: General rules, seismic actions and rules for
ductility levels equal to 5. Conversely, the stiffness degradation deserves buildings, prEN 1998-1, 2009.
more detailed explanations. The stiffness partially decreases if the drift [4] Standards New Zealand, Concrete Structures Standard, NZS 3101:2006,
level increases and if more than one cycle with the same amplitude is Wellington, New Zealand, 2006.
[5] Mpampatsikos V, Nascimbene R, Petrini L. A critical review of the R.C. frame
imposed to the system. Moreover, a global behaviour was observed existing building assessment procedure according to EUROCODE 8 and Italian
characterized by a particular unloading curve, which determines a Seismic Code. J Earthquake Eng 2008;12(sup1):52–82.
“pinching” effect. The stiffness degradation is due to the relative dis­ [6] EngIekrk RE. Design-Construction of The Paramount – A 39-Story Precast
Prestressed Concrete Apartment Building. PCI J 2002;47(4):56–71.
placements between the surfaces of the bottom plate of the steel shoe
[7] Pampanin S. Emerging Solutions for High Seismic Performance of Precast/
and the nut at the top of the anchor bolt, which govern a particular local Prestressed Concrete Buildings. J Adv Concr Technol 2005;3(2):207–23.
rocking mechanism at the base of the column (between concrete and [8] Mpampatsikos V, Bressanelli ME, Belleri A, Nascimbene R. A non-dimensional
mortar layer). parametric approach for the design of PT tendons and mild steel dissipaters in
precast rocking walls. Eng Struct June 2020;212(1):110513. https://doi.org/
The results obtained from the tests highlighted that the collapse 10.1016/j.engstruct.2020.110513.
mechanism is governed by the behaviour of the anchor bolts without any [9] Smith, J.W., Sullivan, T.J., Nascimbene, R., Precast concrete spreader-walls to
significant damage to the specimen, independently of the axial load level improve the reparability of RC frame buildings, Earthquake Engineering and
Structural Dynamics, in press, DOI: 10.1002/eqe.3374.
imposed to the top of the column: [10] Shen S-D, Cui Y, Pan P, Gong R-H, Miao Q-S, Li W-F. Experimental Study of Rc
Prefabricated Shear Walls with Shear Keys Affected by a Slotted Floor Slab.
- The non-linear branch of the base shear – top displacement curve is J Aerosp Eng 2019;32(3):04019013. https://doi.org/10.1061/(ASCE)AS.1943-
5525.0001000.
due to the yielding of the anchor bolts, which are the only compo­ [11] Shen S-D, Pan P, Miao Q-S, Li W-F, Gong R-H. Behaviour of wall segments and floor
nents of the system to reach the plastic branch, while the other slabs in precast reinforced concrete shear walls assembled using steel shear keys
components do not exceed yielding and are not significantly (SSKW). Struct Control Health Monitor 2019;26(10). https://doi.org/10.1002/stc.
v26.1010.1002/stc.2418.
damaged; [12] Shen S-D, Pan P, Miao Q-S, Li W-F, Gong R-H. Test and analysis of reinforced
- The hysteretic behaviour has been evaluated up to a drift level equal concrete (RC) precast shear wall assembled using steel shear key (SSK). Earthquake
to 4.8%, since higher displacement values have not been considered Eng Struct Dyn 2019;48(14):1595–612.
[13] Shen S-D, Cui Y, Pan P, Ren J-Y. Development of prefabricated composite energy-
to be significant. It is characterized by negligible strength degrada­
dissipating slotted shear wall. Eng Struct 2019;199:109577. https://doi.org/
tion and by a global ductility level always greater than 5; 10.1016/j.engstruct.2019.109577.
- The total damping, given by the sum of the viscous (considered equal [14] Sezen, H., Elwood, K., Whittaker, A., Mosalam, K., Wallace, J., and Stanton, J.,
to 2%) and the hysteretic (calculated as a function of the ratio be­ “Structural Engineering Reconnaissance of the August 17, 1999, Kocaeli (Izmit),
Turkey Earthquake”, 2000.
tween the dissipated energy and the elastic energy of each cycle) [15] Toniolo G, Colombo A. Precast concrete structures: The lessons learned from the
damping is about 8.5% in the elastic branch (up to a drift level equal L’Aquila earthquake. Struct Concrete 2012;13(2):73–83.
to 1.2%), then it increases and is in the 16% − 20% range for the first [16] Magliulo G, Bellotti D, Cimmino M, Nascimbene R. Modeling and seismic response
analysis of RC precast Italian code-conforming buildings. J Earthquake Eng 2018;
cycles of the subsequent drift levels; 22(sup2):140–67.
- The incipient yielding condition of the columns examined is attained [17] Magliulo G, Ercolino M, Petrone C, Coppola O, Manfredi G. The Emilia earthquake:
at a drift level in the 1% − 1.3% range; in addition, since the column seismic performance of precast reinforced concrete buildings. Earthquake Spectra
2014;30(2):891–912.
does not exceed the elastic limit and it is not subjected to significant [18] Ozden S, Akpinar E, Erdogan H, Atalay HM. Performance of precast concrete
damage, a more rational design than the case of traditional RC pre­ structures in October 2011 Van earthquake, Turkey. Magazine Concrete Res. 2014;
cast structures characterized by monolithic columns and pinned 66(11):543–52.
[19] Belleri, A., Brunesi, E., Nascimbene, R., Pagani, M., Riva, P., Seismic performance
beams (yielding condition at 2% drift level), can be obtained. of precast industrial facilities following major earthquakes in the Italian territory,
Journal of Performance of Constructed Facilities, Vol. 29, Issue 5, 1 October 2015,
CRediT authorship contribution statement Article number 04014135, DOI: 10.1061/(ASCE)CF.1943-5509.0000617.
[20] Casotto, C., Silva, V., Crowley, H., Nascimbene, R., Pinho, R., Seismic fragility of
Italian RC precast industrial structures, Engineering Structures,Vol. 94, July 01,
R. Nascimbene: Conceptualization, Methodology, Software, 2015, Pages 122-136.
Writing – original draft. L. Bianco: Data curation, Investigation. [21] Perrone D, Calvi PM, Nascimbene R, Fischer EC, Magliulo G. Seismic performance
of non-structural elements during the 2016 Central Italy earthquake. Bull Earthq
Eng 2019;17(10):5655–77.
Declaration of Competing Interest [22] Minghini F, Ongaretto E, Ligabue V, Savoia M, Tullini N. Observational failure
analysis of precast buildings after the 2012 Emilia earthquakes. Earthq Struct
The authors declare that they have no known competing financial 2016;11(2):327–46.
[23] Savoia M, Buratti N, Vincenzi L. Damages and collapses in industrial precast
interests or personal relationships that could have appeared to influence buildings after the 2012. Emilia earthquake. Eng Struct 2017;137:162–80.
the work reported in this paper. [24] Buratti N, Minghini F, Ongaretto E, Savoia M, Tullini N. Empirical seismic fragility
for the precast RC industrial buildings damaged by the 2012 Emilia (Italy)
earthquakes. Earthq Eng Struct Dyn 2017;46(14):2317–35.
Acknowledgements [25] Bovo M, Savoia M. Numerical simulation of seismic-induced failure of a precast
structure during the Emilia earthquake. J Perform Constr Facil 2018;32(1):
The authors would like to express gratitude to Peikko Oy for having 04017119. https://doi.org/10.1061/(ASCE)CF.1943-5509.0001086.
[26] Demartino C, Vanzi I, Monti G, Sulpizio C. Precast industrial buildings in Southern
financed the experimental and numerical research carried out by Europe: loss of support at frictional beam-to-column connections under seismic
Eucentre in Pavia. Particular mention is due to Ing. Davide Bolognini, actions. Bull Earthq Eng 2018;16(1):259–94.
Ing. Davide Bellotti, Ing. Silvia Santagati, Ing. Emanuele Brunesi and Dr. [27] Saisi A, Toniolo G. Precast r.c. columns under cyclic loading: an experimental
programme oriented to EC8. Studies and researches, F.lli Pesenti Master school.
Kushan Wijesundara for their contributions to the development of the Politecnico di Milano 1998;19:373–414.
research project, suggestions and helpful comments. The authors also [28] Fan J-J, Fenga D-C, Gang Wu, Hou S, Yong Lu. Experimental study of prefabricated
thank Ing. Lorenzo Bianco, whose comments and observations generated RC column-foundation assemblies with two different connection methods and
using large-diameter reinforcing bars. Eng Struct 2020;205:110075.
many interesting discussions, the outcome of which has significantly

16
R. Nascimbene and L. Bianco Engineering Structures 247 (2021) 113214

[29] Ling JH, Abd. Rahman AB, Ibrahim IS. Feasibility study of grouted splice connector [41] Bellotti D, Bolognini D, Nascimbene R. Response of Traditional RC Precast
under tensile load. Constr Build Mater 2014;50:530–9. Structures under cyclic loading. Environ Semeiot 2009;2(2):63–79.
[30] Yuan H, Zhenggeng Z, Naito CJ, Weijian Yi. Tensile behavior of half grouted sleeve [42] Vimmr V, Sharif Z. Fixing of precast concrete columns to foundations, fib
connections: experimental study and analytical modeling. Constr Build Mater Symposium. Session 2B–9: Construction Technol 2011;607–610:Prague.
2017;152:96–104. [43] Bianco L, Santagati S, Bolognini D, Nascimbene R. Seismic response of columns
[31] Liu Y, Zhou B, Cai J, Sang-Hoon Lee D, Deng X, Feng J. Experimental study on connected to the foundation through a fastening technique. In: “Peikko News”;
seismic behavior of precast concrete column with grouted sleeve connections 2009. p. 13–20.
considering ratios of longitudinal reinforcement and stirrups. Bull Earthq Eng. [44] CEN (European Committee for Standardisation), Eurocode 2: Design of concrete
https ://doi.org/10.1007/s1051 8-018-0414-9, 2018. structures, 2004.
[32] Yan Q, Chen T, Xie Z. Seismic experimental study on a precast concrete beam- [45] Priestley MJN, Calvi GM, Kowalsky MJ. Displacement-Based Seismic Design of
column connection with grout sleeves. Eng Struct 2018;155:330–44. Structures. Pavia: IUSS Press; 2007. p. 721.
[33] Ameli MJ, Pantelides CP. Seismic analysis of precast concrete bridge columns [46] Wijesundara KK, Nascimbene R, Sullivan T. Equivalent viscous damping for steel
connected with grouted splice sleeve connectors. J Struct Eng 2017;143(2): concentrically braced frame structures. Bull Earthq Eng 2011;9(5):1535–58.
04016176. https://doi.org/10.1061/(ASCE)ST.1943-541X.0001678. [47] Jacobsen L. S., Damping in composite structures, Proceeding, 2nd World
[34] Ameli MJ, Brown DN, Parks JE, Pantelides CP. Seismic column-to-footing Conference on Earthquake Engineering, Tokyo and Kyoto, Japan, 2 ,1029-1044,
connections using grouted splice sleeves. ACI Struct J 2016;113(5):1021–30. 1960.
[35] Haber ZB, Saiidi MS, David HS. Seismic performance of precast columns with [48] Venini P, Nascimbene R. A new fixed-point algorithm for hardening plasticity
mechanically spliced column-footing connections. ACI Struct J 111:639–650. https based on nonlinear mixed variational inequalities. Int J Numer Meth Eng 2003;57
://doi.org/10.14359 /51686 624, 2014. (1):83–102.
[36] Dal Lago B, Toniolo G, Lamperti Tornaghi M. Influence of different mechanical [49] Della CL, Venini P, Nascimbene R. Numerical Simulation of an Elastoplastic Plate
column-foundation connection devices on the seismic behaviour of precast via Mixed Finite Elements. J Eng Math 2003;46:69–86.
structures. Bull Earthq Eng 2016;14(12):3485–508. [50] Vecchio FJ, Collins MP. The modified compression field theory for reinforced
[37] Tullini N, Minghini F. Cyclic test on a precast reinforced concrete column-to- concrete elements subjected to shear. ACI J 1986;83(22):219–31.
foundation grouted duct connection. Bull Earthq Eng 2020;18(4):1657–91. [51] Selby R. G. and Vecchio F. J., Three-dimensional Constitutive Relations for
[38] Rave-Arango JF, Blandón CA, Restrepo JI, Carmona F. Seismic performance of Reinforced Concrete, Tech. Rep. 93-02, Univ. Toronto, dept. Civil Eng., Toronto,
precast concrete column-to-column lap-splice connections. Eng Struct 2018;172: Canada, 1993.
687–99. [52] MIDAS (2020) Nonlinear and Detail FE Analysis System for Civil Structures - FEA
[39] Belleri A, Riva P. Seismic performance and retrofit of precast concrete grouted Analysis and Algorithm Manual (www.cspfea.net), 2020.
sleeve connections. PCI J 2012;57(1):97–109. [53] Thorenfeldt E., Tomaszewicz A. and Jensen J. J., Mechanical properties of high-
[40] Popa V, Papurcu A, Cotofana D, Pascu R. Experimental testing on emulative strength concrete and applications in design. In Proc. Symp. Utilization of High-
connections for precast columns using grouted corrugated steel sleeves. Bull Earthq Strength Concrete (Stavanger, Norway) (Trondheim, 1987), Tapir.
Eng 2015;13(8):2429–47.

17

You might also like