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

net/publication/261510227

Impact Loads in Steel Connections

Conference Paper in IABSE Symposium Report · February 2013

DOI: 10.2749/222137813807054601

CITATIONS READS

4 4,853

3 authors:

Pedro Barata Constança Rigueiro

University of Coimbra Polytechnic Institute of Castelo Branco
17 PUBLICATIONS 136 CITATIONS 99 PUBLICATIONS 528 CITATIONS

SEE PROFILE SEE PROFILE

Aldina Santiago
University of Coimbra
168 PUBLICATIONS 1,891 CITATIONS

SEE PROFILE

All content following this page was uploaded by Constança Rigueiro on 06 April 2015.

The user has requested enhancement of the downloaded file.

 Impact loads in steel connections

Pedro BARATA Constança RIGUEIRO Aldina SANTIAGO

PhD Student, ISISE Assistant Professor, ISISE Assistant Professor, ISISE
University of Coimbra Polytechnic Castelo Branco University of Coimbra
Coimbra, Portugal Coimbra, Portugal Coimbra, Portugal
[email protected] [email protected] [email protected]

João Paulo RODRIGUES

Assistant Professor, ISISE
University of Coimbra
Coimbra, Portugal
[email protected]

Summary
After attacks on World Trade Center (2001), Madrid (2004), London (2005) and Mumbai (2008),
special attention was given to the study of robust structures subjected at different accidental loads,
allowing localized failure without being damaged to an extent disproportionate to the original
cause. The collapse of the World Trade Center (WTC) showed problems in the design of structural
elements: columns collapse, beams buckling and brittle failure of connections. Concerning this last
topic, it was realised that the joint structural details plays very significant role behaviour in struc-
tures subjected to accidentals loads. The accidental loads may result from an object impact, blast,
explosions, earthquake and fire.
The work presented in this paper corresponds to the first work package of the research project Im-
pactfire, currently in development at the University of Coimbra; the behaviour of steel beam-
column connections against accidental impact loading is the main objective of the project. The cur-
rent paper is focused in: i) characterization of the impact scenarios in steel structures, ii) influence
of strain rate sensitivity steel material model, and iii) previous research studies on the behaviour of
steel beam-column connections against impact loading. At the end, the experimental programme of
the research project Impactfire is presented.
Keywords: connections, design standards, high strain rate, impact scenarios, robustness, steel
structures.
1. Introduction
The loads associated to an accidental event are normally with severe intensity and resulting in ex-
traordinary consequences. Examples of accidental loads are: fire, blasts, impact, earthquake, ava-
lanches, landslides, and so on; moreover, the combination of these scenarios must also be consid-
ered such as fire after impact or fire after earthquake. These topics are addressed in the strategic
research agenda of the European Steel Technology Platforms [1], which cites the need for safety in
the design, manufacture and performance of steel structures, especially against natural hazards and
accidental loading.Thus, it is important to know the type and magnitude of the loads applied in the
structure, when an extreme event occurs, in order to evaluate the corresponding response and the
role played by each structural component to prevent progressive collapse of the structure.
The Eurocode standards present some specific parts for the design of structures in case of accidental
actions such is fire (parts 1.2 of the Eurocodes) and earthquake (Eurocode 8). Additionally, part 1-7
of Eurocode 1[2] gives some guidelines and application rules for the assessment of accidental ac-
tions on buildings and bridges with the identified and unidentified accidental actions. The unidenti-
fied actions are related to robustness requirements. For the identified actions, impact and explosions
are under the scope of this standard. This type of loads can be defined as impulsive, characterized
by: i) short-term high intensity load that does not change the direction of its action; ii) has not more
than one maximum, during the constant action and iii) normally induced elevated strain rates in the
structure.
However, because this type of impulsive loading can be induced dinamic stress and keen stress gra-
dients that result local damages of the structural elements; a robust design has to be achieved in
order to avoid the progressive collapse of the structure.
2. Impact loads
The strain rates for different types of impulsive loading are presented in Fig. 1. The research project
Impactfire deals with strain rates corresponding to those produced by impact load.

Quasi-Static Earthquake Impact Blast

-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6
10 10 10 10 10 10 10 10 10 10 10 10 10
-1
Strain Rate (s )
Fig. 1: Strain rate for different types of loads

The impact load can be a result of collision of an object, vehicle or flying debris with a structure or
swinging objects during construction. Impact load are characterized as an external force applied
suddenly to a structure, system or component. The duration of the impact is very short, small than
the response period of the target. The Eurocode 1 part 1-7 defines impact as an interaction phe-
nomenon between a moving object and a structure, in which the kinetic energy of the object is sud-
denly transformed into energy of deformation. The properties of object and structure are taken into
account to calculate the dynamic interaction forces; the main parameters that influence the internal
forces and the structural response during the impact are: i) the velocity and the mass of object; ii)
the deformation characteristics of the object and the structure (damping, elastic limit, stiffness) and
iii) the surface and angle of impact.
The most common example of impact load is the accident resulting of vehicular collision with the
buildings (or viaducts). Road traffic is a threat to the structures that are located along traffic routes;
the possibility of occurrence of an impact from a vehicle due to an accident is real, they occur
mainly in urban areas and are result of the vehicles leaving the roadway. An example of the varia-
tion of the force on a column due to the impact of a vehicle is depicted in Fig. 2 .
For bridges piers, barrages, dock walls and gates, the impact loading by floating objects must be
considered, as well the impacts caused by ships. The action of water on structures depends on the
dynamic characteristic of the loaded structure, the effects of waves (on breakers, sea structures etc.)
are also be considered as impact loadings. However, if the ratio of natural frequency of the structure
to the duration of the load is higher than 0.4, the loading cannot be considered as a static in terms of
the overall structural response.
In some regions, where there is a danger of falling stones, structures are required to protect the traf-
fic, as well in mountain the impact loading caused by avalanches must be considered for the protec-
tion structures and also for the bridges [3]. Moreover, accidents during the construction works
should also be considered as impact loading
 Force (kN)
667

445

222

25 50 75 Time (ms)
Fig. 2: Force on column due to a vehicular impact [4].

3. Design Standards
3.1 Eurocode 1, part 1-7
A strategy to avoid or to control the consequences of accident scenarios was introduced in the
Eurocode 1 part 1-7. This standard defines two types of actions: unidentified actions and identified
actions. The unidentified actions are associated to robustness and the standard prescribes some
strategies to design such are: i) key element design; ii) enhanced redundancy of the structure; iii)
prescriptive rules: tying forces (Annex A). These design strategies should be adapted to the conse-
quences class classification of the structure defined by the Annex B of the Eurocode 0 [5]. Note that
the Annex A only specifies guidance for consequences class 2. For the identified actions, the Euro-
code 1 part 1-7 considers accidental situations due the followings events: i) explosions and ii) im-
pact (from road vehicles; forklift trucks; trains; ships and helicopters on roofs). For structural de-
sign, the actions are represented by an equivalent static force giving the equivalent action affects in
the structure, and may be assumed that impacting body absorbs all the energy; this assumption
gives conservative results. Strain rate effects should be considered on the description of the material
properties of the impactor and structure.
According Eurocode 1 part 1-7 the impact can be hard impact or either soft impact. A simplified
model to calculate the equivalent static forces developed during hard impact is presented. This
model assumes that the structure is rigid and the impacting object deforms linearly during the im-
pact:
(1)
Where: is the object velocity at impact; k is the equivalent elastic stiffness of the object and m is
the mass of the impacting object.
The maximum dynamic force on the outer surface of the structure is given by the equation 1. How-
ever, within the structure these forces may increase due the dynamics effects. The upper bound of
these effects can be determined assuming: the structure responds elastically and the load may be
considered as a rectangular pulse. In that case the dynamic amplification factor dyn is 2.0.
In case of soft impact, the structure is designed to absorb the impact through plastic deformations
and should be provided requirements in the structure to have enough ductility to absorb the total
kinetic energy of the impacting object. In the limit case of rigid-plastic response of the structure the
work performed by the plastic deformation would be equal to the kinetic energy.
(2)
Where: m is the mass of the impacting object; vr is the object velocity at impact; F0 is the plastic
strength of the structure; y0 is the displacement of the impact point that the structure can undergo.
For the specific case of impact from road vehicles, the values of the equivalent static design forces
depend on the category of traffic and the location of the building (Table 1).
 Table 1 – Indicative equivalent static design forces according Eurocode 1 part 1-7
Category of traffic Force Fdx (kN) Force Fdy (kN)
Motorways, country national roads and main roads 1000 500
Country roads in rural areas 750 375
Roads in rural areas 500 250
Parking garages and courtyards with access to:
- Cars 50 25
- Lorries 150 75
Note: Fdy: force acting in the perpendicular direction of the impactor and F dx: force acting in the direction of the impac-
tor.
It is recommended that Fdy does not act simultaneously with Fdx in the structure.

3.2 TM5-1300
TM5-1300 [6] is the main code used for the design of structures to resist blast and impact loads, in
the United Sates. It is provided the procedures, construction techniques and detailing to prevent
propagation of explosion (from one structure, or part of, to another), as well as the protection for
personnel and valuable equipment. The code considers that the structural system are divided in
three types of systems: i) donor system, includes the type and amount of the explosive and the ma-
terials that become part of the damaging output; ii) receptor system, personnel, equipment or even
explosives that require protection and iii) protection system, the structural components, (distance,
shelters, containment structures and barriers) necessary to shield or reduce the hazardous effects.
The capacity of the protective structure should take into account the category of protection needed
by the acceptor system. The type and capacity of the protective structure are selected to produce a
balanced design with the degree of protection required by the acceptor and the hazardous output of
the donor. The TM5-1300 considers that in steel structures the shapes are considerably slender and
as result, the effect of local or global instability, rather than the ultimate capacity, is an important
aspect to take into account in the design. It is recommended to take special care in steel design for
connections integrity and to avoid brittle failure modes, especially in the welded zones. The design
procedures according this code are based in the the ductility ratios, µ (ratio of the maximum deflec-
tion to the elastic deflection); the calculation of the dynamic design stresses for different ductility
ratios, up to 10 and greater than 10, are presented. For values higher than 20, sufficient bracing is
necessary. The deformation criteria are defined according type of system and typology of structural
elements.
4. Previous studies on steel connections under impact loading
The first reported studies on structural steel connections subjected to impulsive loads were per-
formed in 1999 by Krathammer [7]. The main purpose of these studies was to evaluate the applica-
bility of the procedure prescribed in TM5-1300. Fully restrained 3D welded beam-to-column con-
nections (designed for seismic conditions) were numerically simulated against an impulsive load
equivalent to that result from the explosion of a TNT charge in center of the structure floor. Two
numerical models were performed: two-dimensional (external joint defined by one column and one
beam) and three-dimensional (internal joint defined by one column and two beams). Additionally,
in order to consider the strain-rate effect on the steel properties, the numerical simulations were
carried out with and without consideration of dynamic increasing factors (DIF).
The numerical results showed in case of full strength connections, flexural capacities of the adjacent
structural members rules the rotation capacity; otherwise, for semi-rigid connections the rotation
capacity is governed by its internal resistance and deformations. The full strength connection (with-
out DIF) failed catastrophically by fracture at the weld, while the same connection with DIF pre-
sented large flexural rotations. However, 3D model (internal joint) showed large horizontal defor-
mations of the beam members, highlighting the possible unsafe application of TM 5-1300, and the
strains, stresses and rotations exceeded the failure conditions and the limits set by TM 5-1300.
Moreover, the models with dead load included presented two times larger rotational deformation
values than the cases without dead load. The authors concluded that TM5-1300 overestimated the
rotation deformation capacity of the system and might not be adequate for extreme dynamic prob-
lems.
The previous work was continued in 2005 by Sabuwala et al. [8]; fully restrained steel connections
subjected to blast loads were numerically studied using FEM package ABAQUS (explicit solver).
Tests previously performed in the AISC Northridge Connection Test Program were used to validate
and calibrate the models. The numerical results showed that reinforced connections (with cover
plates for the beam flanges and stiffeners for the column web) had a better performance than the
unreinforced connections and presented lower rotations, stresses and displacements; moreover lo-
calized stresses in welds indicating failure were observed in unreinforced connections. Comparing
the numerical results with the design values, it was concluded that, for this connections typology,
the TM5-1300 overdesign the structural members for the blast loads
In order to study and improve the structural robustness of the buildings when subjected to an excep-
tional event of loading, the University of Sheffield [9] had performed a series of 33 tests in steel
beam-to-column joints. Three types of connections were tested: i) flexible end plate (8 and 10 mm),
ii) web-cleats and iii) fin plate. All joints were built, using the S275 steel grade and assembling us-
ing bolt M20 8.8. The joints were tested under different loading rates from quasi-static to fast dy-
namic reaching the maximum value of applied load at t = 5 ms. The results of the tests showed that
the connections response becomes stiffer, reaches a higher resistance peak and are less ductile as
the loading increased from static to dynamic. For the 8 mm flexible end plate connections, failures
in the endplate tearing along the edge of the weld due to tension, shear or a combination of two
were observed. The increase of the thickness of the end plate induced greater rotation capacity and
improvement of the resistance to tearing. Additionally, for the 8 mm flexible end plate connections
a variety of different of failures modes were observed: i) bolt failure, ii) weld tearing away from the
endplate and iii) beam web crush or pulled out along the weld. The web cleat connections failed by
block shear in the beam web and the fin plate failed due to bolt shear.
Recently, Urgessa [10] developed a complex eight story steel frame model using ABAQUS (ex-
plicit solver) under a dynamic pressure equivalent to the effect of explosion of 90.7 kg of TNT;
three different types of connections were employed in beam-to-column joints: SidePlate; TA con-
nection; welded connection. The results showed that the SidePlate connection type provides a better
performance against blast loading than welded connection. The TA connection type, which in-
cluded horizontal plates intended to shift plastic zones away from columns, also performs good be-
haviour against impact loading.
5. STRAIN RATE SENSITIVITY STEEL MATERIAL BEHAVIOUR
5.1 Strain rate sensitive material models
Strain rate ( ), is a dimensionless parameter which describes the amount of deformation which oc-
curs in a piece of material when it is loaded. Most of the ductile materials have strength properties
which are function of the loading speed. Some constitutive models has been presented during the
last decades: Malvern [11] (equation 1); Cowper-Symond [12] (equation 2); Johnson-Cook [13]
(equation 3); SCI [14] (equation 4) and ASTM standard E1820 [15] (equation 5). The Cowper-
Symond and the Johnson-Cook models are commonly used in mainly computer finite-element pro-
grams.
(1)

(2)

(3)

(4)

(5)
Where: s and b are material constants; is the equivalent plastic strain rate; is the dynamic
yield strain; is the viscosity parameter and is the strain rate hardening parameter; is the

effective plastic strain; is the strain rate used to determine A,B and N; is the
homologous temperature, is the melting temperature, is the reference temperature when de-
termining A, B and N; is there is mass density and is specific heat; A is the
yield stress; B and N are the effects of strain hardening; C is the strain rate constant; M is the ther-
mal softening parameter; is the yield stress of steel; D and q are parameters related to steel; T
temperature of the coupon and t is load time (ms).

5.2 Hopkinson bar experiments

Included in the research project ImpactFire, the material behaviour of mild steel S355 subject to
elevated strain rates were assessed using Hopkinson bar test and compared with the corresponding
results obtained by quasi-static tensile tests. In the Hopkinson bar test, a short cylindrical specimen
is sandwiched between two long elastic bars. Typically a striker bar is fired into the end of the inci-
dent bar, generating a compressive stress pulse.
The steel coupons were obtained from hot rolled IPE 300 (flange and web), SHS200x8, SHS200x10
and plate with 15 mm of thickness. For each one, three coupons were tested (T1, T2 and T3) under
a strain rate of 900 s-1 (approximately). For coupons from flange of IPE 300 and SHS200x10, the
true stress-strain relationships are presented in Fig. 3; for elevated strain rates, it was observed that:
i) the yield stress increased; ii) the elastic module remains unchanged and iii) the ultimate plastic
strain decreased. The dynamic increase factor (DIF) for the yield stress of IPE300 flange and
SHS200x10 were about 1.5 and 1.7 respectively.

a) b)
Fig.3 –True-stress versus true-strain curves:a) IPE 300 - flange; b) SHS200x10

Finally, the experimental results were used to assess the applicability of theoretical models to pre-
dict the strengh of the steel subject to elevated strain rates (Table 3).
 Table 3: Test results analysis
SR σys σyd_exp σyd ASTM σyd C-S σyd SCI Variation Variation Variation
Specimen
[s-1] [MPa] [MPa] [MPa] [MPa] [MPa] ASTM [%] C-S [%] SCI [%]

IPE300 Web 1118.0 408.1 726.05 727.12 1023.17 778.14 -0.15 -40.92 -7.17
IPE300flange 943.0 404.2 619.70 703.82 944.25 736.11 -13.57 -52.37 -18.78
SHS200x8 1011.1 406.4 686.23 871.92 1392.29 911.69 2.71 -55.35 -1.72
SHS200x10 828.4 461.7 816.05 792.95 1154.86 812.03 2.83 -41.52 0.49
Plate 15 609.9 380.0 640.95 695.97 884.73 687.06 -8.58 -38.03 -7.19
Where: SR- applied strain rate; σys – static yield stress; σyd – dynamic yield stress; σyd ASTM – dy-
namic yield stress according ASTM E1820; σyd C-S – dynamic yield stress obtained by Cowper-
Symonds model; σyd SCI – dynamic yield stress obtained by SCI model.
The results for the models ASTM and SCI showed a good correlation with tests results. The results
obtained for the Cowper-Symonds model overestimates the strength of the steel. This could be due
to the constant materials values (suggested by [16]) may not fit to the material behaviour.
6. Current work on-going in the University of Coimbra
The experimental workpackage, performed in the framework of the research project ImpactFire
PTDC/ECM/110807/2009, includes a series of experimental tests on steel joints under impact load-
ing. Special attention is also paid to the combined scenario of fire after impact (observed, for exam-
ple, during the World Trade Center attack). The experimental tests comprise two campaigns: i) a
first campaign in steel connections components (t-stub and reverse channel in tension) and ii) a sec-
ond campaign in steel beam-to-column connections (flushed end plate and reverse channel in bend-
ing).
In the first campaign, the component tests are performed under quasi-static conditions at ambient
temperature and at elevated temperatures (400ºC, 500ºC and 600ºC) and also under dynamic condi-
tions. Three different thicknesses of specimens were selected: T-stub (10, 15 and 20 mm), reverse
channel (8, 10 and 12 mm). In the second campaign, the quasi-static tests are limited to ambient
temperature, but dynamic tests are also predicted. The specimens to be tested include: flushed end
plate, with 10 and 15 mm of thickness and reverse channel with 8 and 10 mm.
The quasi-static tests are performed with a hydraulic actuator with displacement control up to fail-
ure. For the dynamic tests a custom made high pressure pneumatic actuator, working with nitrogen,
applies a short duration force in the specimens. The force and the strain rate can be changed by
regulating the pressure of the nitrogen. The strain rate applied during these tests is 500 s-1 approxi-
mately.
7. Conclusion
The Eurocode 1 part 1-7 provides rules and guidance on the design for robustness with relatively
simple measures to prevent disproportionate damage. This standard also provides methods for the
calculation of the forces in the structure in case of an accidental impact. However, the design of
steel connections under this kind of accidental action is not include. From previous studies, it is
known that the strain rate enhancement of yield stress in steel connections could be very significant
on the global structural behaviour. This paper shows the current research work developed at the
University of Coimbra concerning this problematic. Tests to evaluate the effect of the strain rate on
the material behaviour were performed and presented; currently, tests under impact loads in joint
components and beam-to-column joints are in development.
8. Acknowledgement
The authors acknowledge financial support from Ministério da Educação e da Ciência (Fundação
para a Ciência e a Tecnologia) research project PTDC/ECM/110807/2009.
 9. References
[1] EUROPEAN C., European steel technology platform: from a strategic research agenda to
implementation 2006, Brussels: European Communities.
[2] EN 1991-1-7: 2006, Eurocode 1: Actions on structures. Part 1-7, General actions - acciden-
tal actions, European committee for standardization, July 2006.
[3] STRUCK W. and VOGGENREITER W., Examples of impact and impulsive loading in the
field of civil engineering. Materials and Structures, Vol. 8 (2), 1975, pp. 81-87.
[4] ELLINGWOOD B.R., SMILOWITZ R., DUSENBERRY D O., DUTHINH D., LEW H. S.
and CARINO N. J., Best Practices for Reducing the Potential for Progressive Collapse in
Buildings 2007, NISTIR 7396.
[5] EN 1990: 2002, Eurocode 0: Basis of structural design, European committee for standardi-
zation, April 2002.
[6] TM 5-1300 - Structures to resist the effects of accidental explosions, Departments of the
Army, Navy, and Air Force, 1991, USA.
[7] KRAUTHAMMER T., Blast-resistant structural concrete and steel connections. Interna-
tional Journal of Impact Engineering, Vol. 22 (9-10), 1999, pp. 887-910.
[8] SABUWALA T., LINZELL D. and KRAUTHAMMER T., Finite element analysis of steel
beam to column connections subjected to blast loads. International Journal of Impact Engi-
neering, Vol 31 (7), 2005, pp. 861-876.
[9] DAVIDSON, J.B, Investigating the Robustness of Steel Beam-to-Column Connections.10th
International Conference on Steel, Space and Composite Structures. 2011. North Cyprus.
[10] URGESSA G.S. and ARCISZEWSKI T., Blast response comparison of multiple steel frame
connections. Finite Elements in Analysis and Design, 2011, Vol 47 (7), pp. 668-675.
[11] PERZYNA P., Fundamental problems in viscoplasticity. Advances in Applied Mechanics 9,
1966, pp. 243-377.
[12] COWPER G.R. and SYMONDS P.S., Strain-hardening and strain-rate effects in the impact
loading of cantilever beams, 1957, Brown University Providence R I.
[13] SCHWER L., Johnson-Cook model: parameter identification and algorithm verification, in
6th European LS-DYNA Conference, 2007, Gothenburg, Sweden.
[14] STEEL Construction Institute Elevated Temperature and High Strain Rate Properties of Off-
shore Steels. Offshore Technology Report Health and Safety Executive, January 2001.
[15] ASTM, Standard Test for Measurement of Fracture Toughness, ASTM E 1820-01, August
2001, United States
[16] BODNER S.R. and SYMONDS P.S., Plastic deformations in impact and impulsive loading
of beams, 2nd Symposium Naval Structural Mechanics, 1960, Pergamon.

View publication stats