See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/362077684

PERFORMANCE OF EXTENDED END PLATE SMA BOLTED CONNECTIONS

UNDER CYCLIC LOADING

Conference Paper · July 2022

CITATIONS READS

0 45

4 authors, including:

Walid Hassan Shalaby

Alexandria University
2 PUBLICATIONS 0 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

A Case Study on construction repair View project

All content following this page was uploaded by Walid Hassan Shalaby on 18 July 2022.

The user has requested enhancement of the downloaded file.

 PERFORMANCE OF EXTENDED END PLATE SMA BOLTED CONNECTIONS
UNDER CYCLIC LOADING

FAHMY A. FATH EL-BAB AND MOHAMED M. EL-HEWEITY

Department of Civil Engineering, Alexandria University

Alexandria, Egypt

MAHMOUD A. ELFEKI AND WALID H. SHALABY

Department of Civil Engineering, Alexandria University

Alexandria, Egypt

ABSTRACT: This work presents a numerical study to investigate the performance of extended
end plate bolted connections using SMA bolts under cyclic loading. A model is developed to
implement the material constitutive relationships (non-linearity and super-elastic behavior of
SMA). Outputs have been compared and verified against previous experimental data that show
good agreement. NiTi bolt materials is implemented, and Auricchio’s model is used to simulate
the super-elastic behavior of the SMA bolts. Comparing ordinary steel bolted connection with the
proposed system shows major improvement in minimizing residual rotation (without losing great
magnitude of the connection capacity). A parametric study is then conducted using the verified
model. Connection capacity and residual rotation are of primary concern. Variables considered are
the SMA material physical properties (six parameters), bolt diameter, and bolt length. The
assessment of the connection performance is identified as ratio between relative residual rotation
and relative moment capacity (CI). Results indicate that the physical properties starting of
transformation loading stress and bolt diameter have significant impact, while bolt length has
small effect. Other physical properties variables have no influence on the mentioned concerns.
However, increasing the connection index (CI) reflects high residual rotation, which guide the
connection design to minimize it to achieve minimal residual rotation.

INTRODUCTION

The occurrence of destructive earthquakes and its resulting damages have clearly shown the
need for more advanced techniques to protect the building structures against earthquake
excitations. Moment Resisting Frames (MRF) are one of the most capable techniques for resisting
lateral loads Sultana and Youssef28. The mentioned resistance is developed from the structure
members and their connections, as ductile behavior and plastic deformations to dissipate the
developed energy. Above certain lateral loading magnitude, the structure cannot back to its
original status and residual deformation is existed. In most of the code provisions, AISC 20101,2

1
and Eurocode3 20056 the locations of these deformations are determined in the beginning to occur
near the connection in the column, connection panel zone, or in the beam. Depending on that,
connection configuration is essential to guide the developed deformation. Although the connection
has great performance against lateral load, it is considered as fully restrained type AISC 20101, and
it is still creating plastic hinge in the structure members as essential requirement in its design.

The weak performance of pre-Northridge earthquake steel moment resisting frames is related
to geometric nonlinearities and brittle failure of welded beam-to-column connections FEMA-350a
, 351, 352, 353, 355e, 355f.8,9,10,11,12,13,14 As a result, research work is done to investigate the
performance of bolted connections, and one of the most common types is the extended end plate
connection, produced many configurations as shown in Figure 1.

Figure 1 Examples of Extended End Plate Connections

To asses connection performance against lateral loads SAC loading protocal FEMA-355d 12
recommended by AISC. In this protocol, the interstory drift angle imposed on the test specimen is
controlled as shown in Figure 2 continues with increments in angle of 0.01, and performs two
cycles each step.

Figure 2 SAC Loading Protocol FEMA-355d12

To enhance connection performance by making super hinge or removable small deformed part
of the structure is a pressing matter. This is especially important to prevent any unexpected risks
from the happened residual deformations, saving time and financial costs. To satisfy super-elastic
demand new material is developed with the ability of re-centering, therefore the shape memory

2
alloy (SMA) is selected, to play the spring part. SMA material has super elastic behavior and
damping capability.

SMA are one of the most well-known types of smart material and they have found extensive
uses since their discovery in late '30s. A shape memory transformation was first observed in 1932
in an alloy of gold and cadmium, and then later in brass in 1938. Particularly, the concept of
thermo-elastic martensitic transformation, which explained the reversible transformation in
martensite, was introduced in Kurdyumov and Khandros18 based on experimental observations of
the thermally reversible martensitic structure.

All NiTi SMAs exhibit two distinct crystal structures martensite and austenite. This
transformations have been identified in two main transformations. Start loading, martensite is
stress-induced, resulting in a stress plateau in stress–strain curves, as shown in Figure 3. The
martensite becomes unstable upon unloading, and it transforms back to austenite with restoring the
initial un-deformed shape. The recoverable strain of SMAs can be as high as 8–10%, this is called
super-elastic effect (SE).

Figure 3 SMA, Superelastic Effect (SE) Figure 4 SMA, Shape Memory Effect (SME)

Upon loading above a certain strain, the de-twinning process of martensite occurs and the
material is easily deformed to several percent strain, as shown in Figure 4. Unloading results in a
residual deformation, and by heating can be recover to its initial shape, this is called the shape
memory effect (SME).

Unusual SMA macro behavior and properties imply non-conventional constitutive models. The
traditional inelastic models do not provide an adequate description of complex behavior of this
material, so a new class of inelastic models is the generalized plasticity based on an internal-
variable formalism. This generality makes the model an adequate tool for representing constitutive
materials with complex behavior Lubliner and Auricchio19, Taylor and Auricchio30, Simulia,
UMAT27.

Many connections are developed to avoid residual deformation, or at least to minimize it. This
is occurred by the meaning of super elastic hinge creation with moderated energy dissipation
3
capacity. Hongwei, Cho and Wilkinson16, proposed retrofitted rigid extended stiffened end-plate
bolted connection with long shank SMA bolts and specific connection details, to prevent beam
buckling. The results show that the deformations of the connection at peak load are recoverable
upon unloading. The SMA connection shows moderate energy dissipating capacity, and shows
obvious semi-rigid characteristic and its ductility is significantly influenced by the bolt length.
Fang, et al.7, explored the feasibility of SMA bolts used for extended end-plate connections
through experimental studies. It was mainly observed that the connection with the SMA bolts had
a very good re-centring ability with moderate energy dissipation. Ductility of the connections was
less satisfactory due to the relatively early fracture of the furthest bolt in the threaded area. In
addition, slight degradation of rotational stiffness was observed as the cycle number increased.
Wang, et al.31, proposed connection is integrating super-elastic SMA tendons with steel tendons.
Test results showed that connections equipped with SMA tendons exhibit moderate energy
dissipation, double-flag-shaped hysteresis loops and excellent re-centering capability after being
subjected to cyclic loads up to 6% inter-story drift angle.

The objective of this study is to achieve super elastic hinge performance, by using SMA bolts
in the extended end plate connection to reach minimal residual deformation. Based on this
objective, to evaluate SMA bolts contribution in existing connection three steps process has been
conducted. Firstly, FE model has been proposed to simulate the connection behavior under cyclic
loading, where models have been verified by comparing analysis outputs with the experimental
results. Secondly, SMA bolts replaced steel bolts in the verified models, where connection
capacity and residual rotation have been compared with the original connection. Thirdly,
parametric study has been conducted to obtain the effective parameters that affect both the
capacity and residual rotation of the connection.

FINITE ELEMENT MODELLING

Modeling 3D steel connection has gained attention and continuous study from researchers.
Sherbourne and Bahaari 23, Bahaari and Sherbourne 3, developed a three dimensional finite
element model for simulating endplate connections by using solid (brick) elements. The model
was assumed to have a continuous connection between the nodes of the bolt head and nut, and the
nodes of end plates. As a consequence, the relative motions between bolt, column flange and end
plates were numerically simplified. The bolt shank behavior was represented using truss elements
instead of solid (brick) elements, which prevents the numerical model reproducing properly the
bearing action between bolts and bolt holes, because the interface between the bolt shank and the
hole boundary was neglected. So the stresses distribution in the end plate and around the bolts are
not reveal the true status.

Mashaly, et al.20, developed three dimensional model with using shell elements for the column,
beam, endplates, and stiffener and solid elements for the bolts components (head, shank and nut).
The interaction between connection’s components has been defined by gap element” interface
element”. The mechanical properties of the gap element are to reflect the normal and tangential
behavior of the interaction. New bolt technique is presented, a bolt head and nut is simulated as
shell element and connected by link tension element (bolt shank), and results are accurate and
significantly saving analysis time.

Hu, et al.17, and Gerami, et al.15 used solid elements for all of the connection components,
consider the welded components as one part, and simulate the interaction between attached parts
with surface to surface contact element. FE model results show good agreement with experimental

4
results, and bolts arrangement variation, in general has insignificant effect on the moment capacity
of the connection.

In the presented FE model, since beam web could be buckled during loading, no symmetry has
been taken into consideration.

Connection material is steel. It is ductile material which has two main phases to be considered.
Elastic phase which is defined by modulus of elasticity and Poisson’s ratio, Then plastic phase is
defined by using yield stress and plastic strain in tubular data. For the cyclic loading combined
hardening is preferable. Yield stress, plastic strain tubular data has specified form, which is true
stress against logarithmic plastic strain with the next formulas:

(1)

(2)

(3)

Where is true plastic strain, is true total strain, is true elastic strain, is true stress, is
Young’s modulus

The selected element for bolted steel connections is the reduced integration brick element
C3D8R. In order to control the hourglass modes Simulia24, a very dense mesh finite element model
has been set up.

Interaction definitions between parts are two types. For the welded parts interaction is
simulated as tie constraints for the welded contact surfaces. For the attached parts (End plate and
column flange), using gap element could lead to analysis diversion due to the produced sliding
between column flange and end plate. So it is not preferable in cyclic loading. Using contact pairs
interaction is executed by generating surfaces and contact pairs for the attached elements with
specified tolerance, Simulia26. In some cases master surface is penetrate the slave surface and lead
to diversion analysis. The robust alternative is the general contact, which defines all the parts
surfaces of the model and strengthens the slave surface in the sliding regions to avoid penetration.

Interaction properties is defined for each surfaces. For the attached elements interaction
properties are “Hard” contact for the normal behavior using Penalty standard, and for Tangential
behavior is “Penalty” with friction coefficient 0.3.

Meshing procedure is executed with more care for the end plate region. End plate thickness
divided into four layers of elements. Beam flange and web thickness have two layers of elements.
This process has been done because the solid element has no rotation on its nodes. Applying cyclic
loading is generating at the same region of the end plate and column flange two main shapes. First
shape, when the load is down (in respect to the top and bottom beam flange location) so lower
region of the end plate is in compression and upper bolts are in tension. Second shape, when the
load is moving up the compression stresses are generated in the upper region and lower bolts are in
tension. This causing instability in the model. This defect has been solved on two steps. First by
making the meshing in the face of the end plate and column flange in this region are identical in
position as shown in Figure 5. Second, when the end plate is deflected, the mesh points are not
identical in position, using the general interaction, is performing constraint to avoid penetration
between surfaces.

5
 Bolt, solid elements,
End plate and column
meshing
flange meshing view

FE model surfaces

Figure 5 Meshing Nodes in the End plate Region of the connection

Since the test is executed in horizontal alignment, no gravity loading is considered. At initial
step, the two ends of the column are hinged by holding ends surface from moving in the three
direction to give partially restrained behavior. For the beams lateral support at four points are
added. Two points are on the top flange, and other two points on the bottom flange. At step labeled
“pre”, Pretension loading for the bolt is applied by bolt load with linear amplitude. This procedure
is executed for all the bolts in the mentioned step. Cyclic loading protocol is performed at next
step named “cyclic”. On the middle beam tip (as in Lab. test) vertical displacement is applied with
tabular amplitude to produce the required rotations. Instead of simulating stiffener under the
loading point which make the model more complicated, MPC constraint is used. This make all of
the points on the surface slave points to main loading point Simulia25.

VERIFICATION OF THE MODEL

The experimental specimens of the chosen connections were single-sided beam-to-column

assembly that is representative of exterior beam-to-column connections, experiments (4E-1.25-1.5-
24) and (4W-1.25-1-30) by Suminer29, and experiment (4ES-1.0-1.0-24) by Blumenbaum4. Each
experiment with its components has been summarized in Table 1.

Table 1 Chosen Connection Properties

Experiment Id. (4E-1.25-1.5-24) (4W-1.25-1-30) (4ES-1.0-1.0-24)

Column Section W14x120 W14x193 Web 3/8 x 18 1/8, Flange 1x8

Column Length 18’ 2 1/2’’ 22’ 2 1/2’’ 22’ 2/12”

Beam Section W24x68 W30x99 Web 3/8 x 22 ¾, Flange 5/8x 7

Beam Length 6’ 14’’ 9’ 4” 16’ 5 ½”

Column; Beam material A572 Gr.50 A572 Gr.50 A572 Gr.50

6
End Plate thick. 1 ½ inch 1 inch 1 inch

Doubler Plate thick. 3/8 inch 3/8 inch 5/8 inch

Continuity Plate Dim 5/8 inch 3/8 inch 5/8 inch

Plates material A36 A36 A36

Bolts Number, grade 4, A490 8, A325 4, A490

Finite Element Model Results

When the models are completed, analysis job has been executed for each. Output results have been
shown in Table 2. A comparison between experimental results and FE model results, which show
good agreement.

Table 2: Results Comparison between Experimental and FE Model

Experiment Id. Item Experiment FE Results Error %
Results
(4E-1.25-1.5-24) Max. Applied Load (kips) 68.94 65.92 4.4%

Max. Applied Moment (kips. In) 11,703 11,191 4.4%

Max. rotation (rad) 0.061 0.060 1.1%

Max. inelastic rotation (rad) 0.044 0.047 6.8%

(4W-1.25-1-30) Max. Applied Load (kips) 76.77 77.59 1.06%

Max. Applied Moment (kips. In) 18,521 18,719 1.06%

Max. rotation (rad) 0.04 0.042 4.76%

Max. inelastic rotation (rad) 0.021 0.023 8.70%

(4ES-1.0-1.0-24) Max. Applied Moment (kips. In) 768 777 1.16%

Max. rotation (rad) 0.050 0.0497 0.60%

Max. inelastic rotation (rad) 0.0285 0.0297 4.04%

For the chosen experiments, during running the FE models, hysteretic curves have been
conducted between applied moment and connection rotation, and have been compared with the
experiments data sheet. Results show matching behavior, mainly for the maximum and minimum
values as shown in Figure 6, Figure 8, and Figure 10 respectively.

Failure models have been discovered for the mentioned experiments. Firstly (4E-1.25-1.5-24)
experiment in Figure 7, it shows clearly when flange buckling and torsional web buckling
happened in (0.04 rad) as mentioned in the lab. Sheet (Suminer 2003), and deformed shape
(experimental and FE model). Secondly (4W-1.25-1-30) experiment, fracture in top bolts which
7
observed from the stresses in the bolt cross section as shown in Figure 9 as mentioned in the lab.
Sheet (Suminer 2003). Thirdly (4ES-1.0-1.0-24) experiment, it shows showing clearly when flange
buckling and torsional web buckling happened in (0.04 rad) as mentioned in the lab. Sheet
(Blumenbaum 2003), and deformed shape shown in Figure 11.

FE Model Results Exprimental Results

Figure 6 Moment Rotation relation for connection (4E-1.25-1.5-24)

Figure 7 Connection (4E-1.25-1.5-24) after testing, Experimental and FE model

8
 FE Model Results Exprimental Results

Figure 8 Moment Rotation relation Connection (8E-4W-1.25-1-30)

Figure 9 Connection (8E-4W-1.25-1-30) after testing, Bolt Failure

a) FE Model Results b) Exprimental Results

Figure 10 Moment Rotation relation connection (4ES-1.0-1.0-24)

9
 Figure 11 Connection (4ES-1.0-1.0-24) after testing, Experimental and FE model

IMPLEMENTATION SMA BOLTS IN THE VERIFIED CONNECTIONS

Discovering SMA bolts effect in the connection behavior is required to evaluate the use of it.
Using the verified aforementioned connections, as platform to implement SMA bolts in these
connections is quite suitable tool. Putting into consideration the type’s variety of the verified
connections can give robust decision for the effect magnitude for each.

The super elastic behavior is defined on the uniaxial behavior in loading and unloading cycles
as shown in Figure 12. This material is included in the model by using the over cited “USER
MATERIAL” Mercuri21.

Figure 12 Super elastic behavior based on the uni-axial tension test (Simulia, UMAT27)

For the verified previous connections (4E-1.25.1.5-24), (8E-4W-1.25-1-30), and (4ES-1.0-1.0-

24) SMA bolts are implemented. SMA mechanical properties have been taken as the used in
Hongwei, Cho and Wilkinson16. Bolts have the same diameter and same length of the original steel
10
bolts. A certain pretension is applied to the SMA bolts to develop stress just below the martensite
transformation starting stress.

For connection (4E-1.25.1.5-24), results shows dramatic changes in the connection behavior.
The connection capacity has been considerably decreased, and residual rotation have been sharply
decreased as shown from hysteretic charts Figure 13, and Figure 14.

Figure 13 Connection (4E-1.25-1.5-24) FE Figure 14 Connection (4E-1.25-1.5-24) FE

model, Original bolts Configuration model, SMA system

By separating the residual rotation and erecting relation between it and applied rotation as
shown in Figure 15, SMA bolts connection reduce the residual rotation significantly along the
cyclic loading history. This reduction is observed in loading and unloading for the both directions
(up and down). Results for the connection with SMA system show degradation in moment
capacity to be 72%, and residual rotation has been decreased to be 9% in the SMA system.

Figure 15 Residual Rotation vs Applied Rotation (4E-1.25-1.5-24), (original, SMA) systems

For connection (8E-4W-1.25-1-30), Results show noticeable changes in the connection

behavior for SMA system. The connection capacity slightly decreased, and residual rotation have
been considerably decreased as shown from hysteretic charts Figure 16, and Figure 17.

11
 Figure 16 Connection (8E-4W-1.25-1-30) FE Figure 17 Connection (8E-4W-1.25-
model, Original bolts Configuration 1-30) FE model, SMA system

Repeating the previous process, relation between residual rotation and applied rotation is
shown in Figure 18. It shows reducing in the residual rotation along the cyclic loading history for
the SMA system. This behavior is noticed in loading and unloading for the both directions (up and
down). Results for the connection with SMA system shows degradation in moment capacity to be
95% and for the residual rotation to be 48% in case of SMA system.

Figure 18 Residual Rotation vs Applied Rotation (8E-4W-1.25-1-30), (original, SMA)

systems

For connection (4ES-1.0-1.0-24), Results shows massive changes in the connection behavior.
The connection capacity and residual rotation have been dramatically decreased as shown from
hysteretic charts, Figure 19, and. Figure 20.

12
 Figure 19 Connection (4ES-1.0-1.0-24) FE model, Figure 20 Connection (4ES-1.0-1.0-
Original bolts Configuration 24) FE model, SMA system

Applying the same process as shown in Figure 21, it shows sharply in the residual rotation
along the cyclic loading history. As previous connections, reduction is noticed in loading and
unloading for the both directions (up and down). Results for the connection with SMA system
shows degradation in moment capacity to be 55% and residual rotation 0.4% in case of SMA
system.

Figure 21 Residual Rotation vs Applied Rotation (4ES-1.0-1.0-24), (original, SMA) systems

Reading in parallel the relation results of the three connections, substituting steel bolts with
SMA bolts in the strong plate connection is keeping the connection capacity and significantly
reducing the residual rotation, on the other hand in the weak plate connection, Although it reduce
the residual rotation, the connection capacity is sharply decreased.

13
PARAMETRIC STUDY

Implementing SMA bolts in the connection has changed its features. These changes are mainly
effect on the values of residual deformation and connection capacity. This is due to the
magnificent super-elastic material behavior of the SMA. Although residual deformation is
decreased, main disadvantage is the reduction in the connection capacity. Parametric Study is
conducted to find out the effect of the new introduced material on the connection. Parameter
categories mainly are, SMA material properties, and SMA element geometrical dimensions. SMA
properties have wide range for each property DesRoches, J. and Delemont5, as shown in Table 3.

Table 3 Properties of NiTi Shape Memory Alloys

NiTi SMA

Property Austenite Martensite

Physical properties

Density 6.45 g/cm3

Mechanical properties

Recoverable elongation up to 8%

Young's modulus 30-83 Gpa 21-41 Gpa

Yield strength 195-690 Mpa 70-140 Mpa

Ultimate tensile strength 897-1,900 Mpa

Elongation at failure 5-50% (typically 25%)

Poisson's ratio 0.33

Chemical properties

Corrosion performance Excellent (similar to stainless steel)

From that, four values for each studied parameter is considered. Lower and upper extremes, in
addition two middle values are taken, to assure the parameter influence. Same concept is applied
on the element geometry parameters. All of the studied parameters have been measured relatively
to the steel bolt material and its geometrical dimensions. Considered parameters are shown in
Figure 22.

14
 SMA, Modulus of elasticity
(E-sma). (Parameter, E-sma/E-st)
SMA, Transformation strain
(T-sma). (Parameter, T-sma/El-st)
SMA, Start of transformation loading (Fys-
sma). (Parameter, Fys-sma/Fy-st)
SMA, End of transformation loading (Fye-
sma). (Parameter, Fys-sma/Fu-st)
SMA, End of transformation unloading (Funs-
sma). (Parameter, Funs-sma/Fy-st)
SMA, End of transformation unloading (Fune-
sma). (Parameter, Fune-sma/Fu-st)
SMA, bolt diameter
(DIA-sma). (Parameter, DIA-sma/DIA-st)
SMA, bolt length
(LEN-sma). (Parameter, LEN-sma/LEN-st)
Figure 22 SMA Stress Strain Curve, parameters

Initial Model

The connection under consideration is single-sided beam-to-column assembly that is

representative of exterior beam-to-column connection. It has been designed according to Practice
22
. Geometrical dimension and loading condition of the connection are taken from the previous
verified connection (4E-1.25-1.5-24). Used sections in connection are also according to the
previous code as shown in Figure 23.

Figure 23 Designed Connection

15
 According to Practice22, Modulus of elasticity for the all the steel grades is equal to 210 GPA.
Also Poisson’s ratio is equal to 0.3. The stress strain curves for the steel parts of the connection are
taken as Elastoplastic (bilinear curve) with strain hardening equal to 0.5% of the modulus of
elasticity between points of yield stress and ultimate stress.

In the previous verified FE models, model has great size (about 32,000 node, 20,000
elements), and simulation time is completed not less than 9 hours. To decrease the running time,
and to keep the accuracy in the acceptable range, some trials have been performed. Accordingly,
some steps can be taken as follows:

 Using dependent parts instead of independent parts, to decrease the allocated memory.
 This leads to conduct meshing procedure directly on the assembly.
 Flanges of the column and beam and web of column and beam have two layers of elements
crossing its thickness, enough to simulated buckling rotations.
 Flanges of the column and beam and web of column and beam have ten rows of elements
crossing its widths.
 Web height of beam has twenty elements.
 Meshing size for the end plate, doubler, continuity plates, and column region in this limits
is small and equal to half of the flange thickness, This has been achieved by making the
meshing seeding identical for mentioned parts and region.
 To decrease the number of nodes, linear bias is used for the beam from its welded end to
the end plate (fine mesh) to its free end where displacement load is applied.
 On the column, bias seeding has been done from the column middle region to both ends.

Above procedures and precautions helped to decrease the model size (less than 17,000 nodes,
and 11,000 elements), therefor run time has been kept under 5 hr.

Numerical Results

FE models have been run according to the mentioned parameter values. Hysteretic curves have
been conducted for the each value of the considered eight parameters as shown in figures from
Figure 24 to Figure 31. Each figure at least is assembled from minimum and maximum values of
the considered parameter, and in the rocket changes more than two have been assembled.

Hysteretic curves for parameter values can be classified in identical curves they are clear in
Figure 24, Figure 25, Figure 27, Figure 28, and Figure 29, which show that parameters are non-
effective, and varied curves in Figure 26, Figure 30, and Figure 31 which show that parameters are
effective. Upon increasing in parameter values, the residual rotation grows too, which is causes
more magnitude of dissipate energy.

Output results for each parameter have been displayed and analyzed to verify and evaluate its
influence on the moment capacity and residual rotation if any, from Figure 32 to Figure 39. Used
parameter values (x axis) in the mentioned figures are relative values, between SMA parameter
and relevant steel parameter. Figures are conducted to present three connection features in relative
format (Y axis), between connection equipped with SMA bolts (sma), and connection with steel
bolts (st). Firstly, Relative moment (M-r) which expressed of the moment capacity developed from
using SMA bolts divided by the moment capacity from the initial connection (steel bolts).
Secondly, relative residual rotation (RES-r) which expressed the residual rotation developed from

16
using SMA bolts divided by the residual rotation from the initial connection. Thirdly CI which is
ratio of (RES-r/M-r).

Modulus of elasticity for SMA (E-sma) is displayed in Figure 24 and shows steady behavior, it
can be explained that maximum magnitude for SMA modulus of elasticity is less than half of the
steel modulus of elasticity. In the same manner, SMA parameters Fye-sma in Figure 27, SMA
parameters (Funs-sma), and (Fune-sma) in Figure 28, and Figure 29 respectively.

On the hand, transformation strain (T-sma) rang is ten to twenty from the steel elongation
value, and also has constant behavior, so the connection ductility is governed by the steel
elongation, and no change for any extra value of (T-sma).

For the start of transformation loading (Fys-sma) parameter is displayed in Figure 26, where it
shows significant behavior for the relative values from 0.20 to 0.4, then rocket behavior to 0.50,
and steady up to 0.80. It was concluded that the requested residual rotation (0.0 to 1.0) has
conjugate parameter ranges from (0.2 to 0.5). For the SMA bolt diameter (DIA-sma) has rapid
change till ratio 1.1, conjugate capacity reached 0.98. LEN parameter has moderated influence as
shown which it effect only between 0.8 and 1.0 on the connection capacity.

Connection Assessment

In this study connection performance has been evaluated with respect to the moment capacity
and residual rotation. In the effective parameters figures, with increasing in the moment capacity,
residual rotation is increased also. The increasing rate for moment capacity, and residual rotation is
not linear thought the chosen parameter values, and there is optimum zone. Which holding almost
moment capacity value, with low residual rotation. To reach this zone, and to give criterion for the
connection ductility connection index (CI) has been proposed. Proposed connection index (CI) is
established to evaluate the connection performance depending on the capacity and residual rotation

(4)

Where: (5)

(6)

According to the previous studied parameters proposed connection index (CI) has range from
0.03 to 1.0 in the SMA system. The lower extreme is expressing super-elastic behavior with lower
capacity in the effective parameters. Upper extreme is reflecting capacity similar to original steel
connection configuration with its capacity and residual deformation. In non-effective parameters
connection index is constant value all over the chosen parameter values as shown in the previous
figures for SMA system. Reading connection index in parallel with capacity (M-r), and residual
(RES-r) charts, it shows, if connection index value around 0.25, it achieved more than 88% of the
connection capacity with 20% residual in SMA system. When connection index is reaching 0.5,
the connection capacity is 96% with corresponding residual 45% in SMA. It means, gaining 8%
increasing in connection capacity increasing the residual 30% more.

17
 Figure 24 Hysteretic Curves, E-sma Figure 25 Hysteretic Curves, T-sma

Figure 26 Hysteretic Curves, Fys-sma Figure 27 Hysteretic Curves, Fye-sma

Figure 28 Hysteretic Curves, Funs-sma Figure 29 Hysteretic Curves, Fune-sma

18
Figure 30 Hysteretic Curves, DIA-sma Figure 31 Hysteretic Curves, LEN-sma

Figure 32 E-sma Parameter Figure 33 T-sma Parameter

Figure 34 Fys-sma Parameter Figure 35 Fye-sma Parameter

19
 Figure 36 Funs-sma Parameter Figure 37 Fune-sma Parameter

Figure 38 DIA-sma Parameter Figure 39 LEN-sma Parameter

CONCLUSIONS

1. The first objective is to give a method to enhance the behavior of the extended end plate
connection under the cyclic loading. The verification work is not to prove the FE
modeling reliability, it is to prove the used data, and feature of the software, and how
much is the simulation process is close to reality. Multi-purpose FE software ABAQUS,
is utilized for connection FE modeling under cyclic loading. Results show good
agreement with experimental tests.
2. Material properties of the SMA have been introduced to the model. Super elastic behavior
has been presented. It is clear in case of strong plate connection, the SMA bolts
contribution is preferable. Because connection Capacity is 72% of the original capacity
and residual rotation is only 9% of the original connection. Contrary for the weak (thin)
end plate connection, the connection Capacity is 95% from the original capacity, and the
residual rotation is 48% from the original one. For the stiffened connection capacity is
55% from the original capacity and the residual rotation is 0.40 % from the original one.
3. Parametric study gives clear vision on the effective and non-effective parameters in
respect to the connection Capacity and residual deformation. Non effective parameters
20
 are, Modulus of elasticity (E-sma), Transformation strain (T-sma), End of transformation
loading (Fye-sma), End of transformation unloading (Funs-sma), End of transformation
unloading (Fune-sma). Effective parameters are, Start of transformation loading (Fys-
sma), bolt diameter (DIA-sma), and bolt length (LEN-sma) has small effect.
4. Based on the connection Capacity and the related residual rotation new connection index
is proposed. It is absolute value. (0.2 <CI<1.0). Lower boundary gives very low residual
value, and upper boundary gives the high residual value. Connection design quality is
reached with the index is near to the lower boundary.

REFERENCES

1. AISC, INC. 2010. Specification for Structural Steel Buildings. Chicago, Illinois.
2. AISCb. 2002. Steel Design guide Series 16, Flush and Extended Multiple-Row Moment
End-Plate Connections. 2nd 2003. Chicago: American Institute of Steel Construction,
Inc.
3. Bahaari, M., and A. Sherbourne. 1996. "Structural Behavior of End Plate Bolted
connections to Stiffened Column." Journal of Structural Engineering 926-935.
4. Blumenbaum, S. E. 2003. Response of Cyclically Loaded Extended End-Plate Moment
Connections When Used With Welded Built-Up Sections. Virginia: Virginia Polytechnic
Institute and State.
5. DesRoches, R., McCormick J., and M. Delemont. 2004. "Cyclic Properties of
Superelastic Shape Memory Alloy Wires and Bars." Journal of Structural Engineering
38-46.
6. Eurocode3. 2005. Design of Steel Structures, Part 1.8 : Design of Joints. Euripean
Committee for Standardization.
7. Fang, Ch., M. Yam, A. Lam, and L. Xie. 2013. Cyclic performance of extended end-plate
connections equipped with shape memory alloy bolts. Journal of Constructional Steel
Research.
8. FEMA-350a. 2000. Recommended Seismic Design Criteria for New Steel Moment Frame
Buildings. Washington, DC: Federal Emergency Management Agency.
9. FEMA-351. 2000. Recommended Seismic Evaluation and Upgrade Criteria for Existing
Welded Steel Moment-Frame Buildings. Washington, DC: Federal Emergency
Management Agency.
10. FEMA-352. 2000. Recommended Postearthquake Evaluation and Repair Criteria for
Welded Steel Moment-Frame Buildings. Washington, DC: Federal Emergency
Management Agency.
11. FEMA-353. 2000. Recommended Specifications and Quality Assurance Guidelines for
Steel Moment-Frame Construction for Seismic Application. Washington, DC.: Federal
Emergency Management Agency.
12. FEMA-355d. 2000. State of the Art Report on Connection Performance. Washington,
DC: Federal Emergency Managment Agency.
13. FEMA-355e. 2000. State of the Art Report on Past Performance of Steel Moment-Frame
Buildings in Earthquakes. Washington, DC: Federal Emergency Management Agency.
14. FEMA-355f. 2000. State of the Art Report on Performance and Evaluation of Steel
Moment-Frame Buildings. Washington,DC: Federal Emergency Management Agency.

21
 15. Gerami, M., H. Saberi, V. Saberi, and A. Daryan. 2011. "Cyclic behavior of bolted
connections with different arrangement of bolts." Journal of Constructional Steel
Reserach 690-705.
16. Hongwei, M., C. Cho, and T. Wilkinson. 2008. "A numerical study on bolted end-plate
connection using shape memory alloys." Materials and Structures 1419-1426.
17. Hu, Y., B. Davison, I. Burgess, and R. Plank. 2010. "Multi-Scale Modelling of Flexible
End Plate Connection Under Fire Conditions." The open Construction and Building
Technology 88-104.
18. Kurdyumov, G., and L. Khandros. 1949. On The “Thermoelastic” Equilibrium On
Martensitic Transformations.
19. Lubliner, J., and F. Auricchio. 1995. Generalized Plasticity and Shape Memory Alloys. n
Great Britain.: Solid Structures.
20. Mashaly, E., M. El-Heweity, H. Abou-Elfath, and M. Osman. 2010. "Finite element
analysis of beam-to-column joints in steel frames under cyclic loading." Alexandria
Engineering Journal 94-104.
21. Mercuri, V. 2014. Shape Memory Alloys Strands: Conventional 3D FEM Modeling and
Simplified Models. ROMA: Università degli Studi di Pavia Facoltà di Ingegneria
Dipartimento di Ingegneria Civile e Architettura.
22. Practice, Permanent Committee for the Code of. 2013. Egyptian Code of Practice for
Steel Construction and Bridges. Code No. (205). Cairo: Housing and Building National
Research Center.
23. Sherbourne, A., and M. Bahaari. 1994. "3D simulation of end-plate bolted connections."
Journal of Structural Engineering 3122-3136.
24. Simulia. 2014. ABAQUS/CAE User’s Guide. 6.14. RI, USA: Copyright Dassault
Systèmes.
25. —. 2014. Analysis User's Guide. Providence, RI, USA: Dassault Systèmes.
26. —. 2013. "Solving Contact Problems with ABAQUS." Coventry: Dassault Systèmes.
27. —. 2013. UMAT for Superelasticity and Plasticity of Shape Memory Alloys. RI, USA:
Copyright Dassault Systèmes.
28. Sultana, Papia, and Maged A. Youssef. 2016. "Seismic performance of steel moment
resisting frames utilizing superelastic shape memory alloys." Journal of Constructional
Steel Research 125: 239-251.
29. Suminer, Emmett A. 2003. Unified Design of Extended End Plate Moment Connection
Subject To Cyclic Loading. Virginia.
30. Taylor, R., and F. Auricchio. 1997. Shape-memory alloys: modelling and numerical
simulations of the finite-strain superelastic behavior. Computer methods in applied
mechanics and engineering.
31. Wang, W., T. Chan, H. Shao, and Y. Chen. 2015. Cyclic behavior of connections
equipped with NiTi shape memory alloy and steel tendons between H-shaped beam to
CHS column. Engineering Structures.

22

View publication stats