MSTR515
FINITE ELEMENT METHOD
Topic: Review On Experimental testing and simulation of bolted beam-column
connections having thick extended endplates and multiple bolts per row
Presented By : Sudip Shrestha
Presented to: DR. Shyam Sundar Khadka
Wednesday, December 30, 2020 1
1. Introduction
2. Experiment Program
3. Experiment Result
4. Numerical Modelling
5. Numerical Result
6. Conclusion
Wednesday, December 30, 2020 2
For the efficient design of steel building systems Better understanding of the strength
and rigidity of connection regions is necessary. Commonly, In design connection of the
beam-column is assumed to be pinned or rigid.However,In reality connection is semi
rigid connection.
In the Retrofitting of the steel structure, Strengthening of connection is required. To
strengthen it ,accurate understanding of the rigidity of both existing and improved
connection is required.
To predict the behavior of beam-column connection :Eurocode has adopted the
component method
Component method considers the various the connection component(Bolt,flange, web)
Wednesday, December 30, 2020 3
1. Introduction
1.1 Component Method (EC -3)
This Component Method involve the 5 basic steps:
1. Identification of the load-path through the connection
2. Determination of individual component strength
3. Determination of individual component stiffness in the load path
4. Assembly of the individual components in series and or parallel
5. Determination of the ‘‘weakest link’’ in the load path
1.2 Literature Review
Based on the various past research analytical and experimental studies have investigated bolted beam column
connection it is found that
Connection strength is increased using endplate stiffeners, in a retrofit scenario (adding endplate stiffeners is
difficult when a concrete slab or metal deck is present.)
Addition of bolt on either side of the column web increases the connection strength.
Wednesday, December 30, 2020 4
1. Introduction
1.2 Literature Review
It is often more economical to use multiple bolts per row when wide H-sections are used. However, current code
methods, generalized to all connection configurations with only one bolt on either side of the column or beam web.(2
in row)currently based on T-stub experiments
Limited experimental data exists in the literature comparing the performance of the EC3-1.8 component method
with connection configurations having multiple bolts per row.
2. Experiment Program
In this Experiment the interactive behavior of bolted beam-column connections having thick extended endplates and
multiple bolts per row (without column web stiffeners) was tested.
Weak column strong beam situations was considered
Detailed numerical models are also created to determine detailed stress and strain distributions within the
connection regions, and to investigate techniques for simulating semi-rigid bolted connections having multiple bolts
per row
Wednesday, December 30, 2020 5
2. Experiment Program
2.1 Objective of the Experiment
Determine the connection response due to influence of bolt grouping (multiple bolts per
row) and thick endplates
Determine the different connection failure modes
Compare experimental performance with code methods & numerical modelling
The experimental specimens consist of a
column element and a beam element which
is fully welded to an endplate
GR10.9 zinc coated M20 bolts with zinc
coated GR10.9 HV nuts and standard GR4.6
washers
All bolts are pre-tightened with 480 N-m of
torque
Wednesday, December 30, 2020 6
6 beam column connections are tested
representing three different bolt
configurations (both extended and non-
extended configurations) and two column
profiles
Column profile tested in group A
(specimens 1A, 2A, and 3A) is an HE 300A
section having a web thickness of 8.5 mm
and a flange thickness of 14 mm
The column profile tested in group B is an
HE 300B section having a larger web and
flange thickness (11 mm and 19 mm
respectively).
The beam (HE 300B) is the same for all
specimens, creating a weak-column strong-
beam scenario for specimens 1A, 2A, and
3A
Wednesday, December 30, 2020 7
2.2 Test configuration, instrumentation, and loading
•Fig. 3 shows the experiment setup to investigate the response of
bolted beam-column connections during beam overloading.
•The column section rests horizontally on two supports to prevent
the column-flange contact with the ground,
•The beam extends vertically from the column connected with
horizontal actuator and displacement controlled loading is applied
•Various measuring devices attached to the specimen are used to
record local and global connection behavior
• Twelve unidirectional strain gauges and four linear variable
differential transducers (LVDTs) are attached to each specimen to
record local member strains and global connection displacements
Wednesday, December 30, 2020 8
3.Experimental Result
3.1 Observations and governing failure modes
Two failure limit levels are considered, (1) a classical
failure limit based on initial component yielding (used to
compare code methods and experimental results). (2) an
ultimate failure limit defined as a 20% reduction in
connection moment capacity
In each test, column flange capacity was observed to be the
initial limiting component, with large deformations clearly
visible within the connection tensile zone.Fig4
For beam-column connections having extended endplates,
complete bolt fracture always occurred in the extended
section on the bolt group close to the column web;
For beam-column connections with flush endplates, bolt
failure occurred in the tensile zone on the bolt group close
to the column web
Wednesday, December 30, 2020 9
3.Experimental Result
3.2 Moment-rotation behavior
Use of four bolts per row and extended endplates
increases connection moment capacity but may
reduce rotation capacity
shows the moment rotation behavior for all six beam-
column connection configurations
For beam-column connections with flush endplates, bolt
failure occurred in the tensile zone on the bolt group closest
to the column web
In this work, the connection moment is determined using a
lever-arm between the column flange-to-endplate contact
zone and the applied horizontal load. This method for
calculating the connection moment is verified as no column
web buckling was observed
Higher Moment capacity is found on the T1A and T1B
specimen(29.5% higher than T2A & T2B).
Wednesday, December 30, 2020 10
4. Numerical Modeling
Finite Element method is used to investigate parameters not measured experimentally such continuous stress
in connection region.
Connection/Numerical Modeling techniques involves : element selection, mesh refinement, simulated bolted
connection, boundary conditions, and material properties
4.1 Finite Element Method
3-dimensional finite elements, nonlinear multi-directional springs, and various boundary conditions are
used to simulate the connection test setup.
Shell Element is used to model all the component of both beam and column to determine the stress and strain
states throughout the connection assembly
Pinned boundary conditions at the both column ends simulate the column-to-floor attachment in the
test ;an applied displacement at the beam tip simulates the horizontal actuator.
Welded connection between the beam and end-plate is assumed fixed.
Bolted connection between the end-plate and column is simulated using nonlinear multi-directional
springs, where the axial and shear resistance of the bolt is provided
Wednesday, December 30, 2020 11
4. Numerical Modeling
4.1 Finite Element Method
To prevent local stress
concentrations at the shell-
spring junction, the spring force
is distributed over a shell area
equal to the area of the bolt
diameter using rigid nodal
constraint
Steel material behavior in the
beam, column, and end-plate is
simulated using a bi-linear
material hardening model
Wednesday, December 30, 2020 12
4. Numerical Modeling
4.2 Component Method
EC3-1.8 component method can be used to predict the onset of material yielding.
Numerical Analysis tool called NASCON is used which is based on EC 3-1.8
Allows for prediction after yielding and thus post-yield stiffness and ductility of the joints
Resistance of each connection configuration is determined by the EC3-1.8 component method, as well as a
modified version of the EC3-1.8 method ( called NASCON method)
Allows for prediction after yielding and thus post-yield stiffness and ductility of the joints
The EC3- 1.8 component method considers a bilinear, elastic–plastic, connection response
While the NASCON method accounts for post-yield connection stiffness and potential nonlinear connection
behavior
Wednesday, December 30, 2020 13
5. Numerical Result
5.1 Comparison between FEA,
experimental response, and
component methods
Fig shows the moment-rotation comparison
between the finite element analysis,
experimental measurements, NASCON
method, and EC3-1.8 component method.
The finite element simulations are able to
capture the global inelastic moment-rotation
behavior observed in the experimental tests,
while the NASCON component method
significantly underpredicts the observed
strength and post-elastic resistance.
Wednesday, December 30, 2020 14
5. Numerical Result
5.1 Comparison between FEA, experimental response, and component methods
Table below presents the rotational stiffness, peak rotation, comparative strength values at 0.02 rad, and the
different failure limits for all method
Wednesday, December 30, 2020 15
5. Numerical Result
5.1 Comparison between FEA, experimental
response, and component methods
The value of 0.02 rad was chosen due to the early
failure of specimen T3B
In general, the finite element simulations fit closely
with the experiments.
The NASCON method usually leads to lower strength
values than the finite element simulations, except for
specimen T2A
The EC3-1.8 component method values are always on
the safe side by at least 11% (note that it includes
resistance factors).
In addition to the global connection behavior, eight
strain gauges located on both sides of the column webs
allow for validation of the finite element techniques
through comparison of localized member strains
Local member strains cannot be determined using the EC3-
1.8 or NASCON component methods. Strain values from
both FEM and experiment were taken at the same
geometric location
Wednesday, December 30, 2020 16
6. CONCLUSION
6 beam column connections with 3 different bolt configurations and section dimensions are tested representing
three different bolt configurations (both extended and non-extended configurations) and two column profiles to
failure under static pushover loading
Global connection behavior from the experimental tests and local connection behavior from analytical
simulations were compared with current code methods
Deformations in the column section govern the failure of equal-column equal-beam or weak-column strong-beam
connections
As bolt locations move toward the column web centerline, bolt prying forces increase at lower connection
rotations due to column flange distortions
Configurations having multiple bolts per row increase connection moment capacity
Closer inner bolt spacing (relative to the column web) decreases rotation capacity in weak-column strong beam
scenarios due to increased bolt prying forces and earlier bolt failure from column flange distortions
Finite element simulation with shell elements and nonlinear springs is reasonably accurate for determining post-
elastic global and local response of bolted beam-column connections having thick endplates and multiple bolts per
row. The EC3-1.8 method provides a conservative estimation of connection strength for connections having thick
endplates and multiple bolts per row
In this paper FEM formulation well validated by experiment
Wednesday, December 30, 2020 17
Wednesday, December 30, 2020 18
