Infrastructures: Feasibility of Reusing Damaged Steel Beams in Temporary Structures

infrastructures
Article
Feasibility of Reusing Damaged Steel Beams in
Temporary Structures
Bilal El-Ariss 1, * and Said Elkholy 1,2
1 College of Engineering, United Arab Emirates University, Al Ain 15551, United Arab Emirates;
selkhouly@uaeu.ac.ae
2 Faculty of Engineering, Fayoum University, Faiyum 63514, Egypt
* Correspondence: bilal.elariss@uaeu.ac.ae
Abstract: This study addresses the feasibility of reusing pre-damaged steel beams in temporary
structures. The extensive structural investigation of notch-damaged, unrepaired, and laterally
unsupported steel beams was performed experimentally and numerically. The simply supported
specimens were tested in two-point loading with the study parameters being the location and size
of the notch. Some beams had one notch on one edge of the tension flange at different locations,
and some beams had two notches on both edges of the tension flange. Three-dimensional numerical
models were generated to simulate the behavior of the test beams. After verifying the model, the
numerical analysis was extended to cover additional different notch depths and widths. The study
showed that the capacity of beams with single notch was more influenced by the notch depth increase
than it was by the increase in the notch width. Beams with double notches exhibited an even more
pronounced and distinct decrease in the capacity as the notch depth and width increased. This
investigation supports the feasibility of reusing pre-damaged steel beams in temporary structures
under service loads and certain levels of damage, where the behavior of such beams is within the
elastic range and the beam maximum defection is less than the allowable one.
Keywords: reused damaged steel I-beam; notch size and location; load-carrying capacity; stiffness;

deflection; FE simulations
Citation: El-Ariss, B.; Elkholy, S.
Feasibility of Reusing Damaged Steel
Beams in Temporary Structures.
Infrastructures 2021, 6, 69. https:// 1. Introduction
doi.org/10.3390/infrastructures6050069
Several shapes of structural steel members are used in structures and bridges. Some
of the most commonly used steel shapes are I-sections, built-up sections, and wide flange
Received: 31 March 2021
Accepted: 30 April 2021
members which, when used in flexure, develop the highest bending stresses in the flanges,
Published: 3 May 2021
making them a critical part of the member cross-section. Therefore, any damage to the
flanges such as area loss in the flange area, due to removing a localized corrosion or due
Publisher’s Note: MDPI stays neutral
to minor collisions during construction or because of low clearance of structures, reduces
with regard to jurisdictional claims in
the flexural capacity of the beam. This might give rise to a limited or whole failure of
published maps and institutional affil- the structure. Fiber-reinforced plastic (FRP) composite materials have been utilized in
iations. retrofitting concrete structures as well as wood, masonry, and steel structural members.
The composite materials may be glass or carbon-FRP materials (GFRP or CFRP) and they
are used in the form of flexible sheets or rigid plates.
Hmidan et al. [1] tested notch-damaged steel I-beams strengthened with CFRP sheets.
Copyright: © 2021 by the authors.
They stated that the level of initial damage in the CFRP-strengthened beams affected
Licensee MDPI, Basel, Switzerland.
the behavior of the plastic region in the vicinity of the notch tip, the CFRP debonding
This article is an open access article
initiation rate, and the beam mode of failure. Shaat and Fam [2–4] presented the results of
distributed under the terms and an experimental work on retrofitting artificially damaged steel–concrete composite beams
conditions of the Creative Commons by saw-cutting part of the steel beam tension flange. The tested beams were repaired
Attribution (CC BY) license (https:// using adhesively bonded CFRP sheets. The findings of their study showed that adhesively
creativecommons.org/licenses/by/ bonded CFRP sheets can be used to repair damaged steel–concrete composite beams.
4.0/). Nozaka et al. [5] compared the test results of various CFRP laminate/adhesive schemes
Infrastructures 2021, 6, 69. https://doi.org/10.3390/infrastructures6050069 https://www.mdpi.com/journal/infrastructures

Infrastructures 2021, 6, 69 2 of 24
that were used to retrofit notched steel beams. The experimental results showed that an
adhesive with fairly large ductility was essential to effectively redistribute the stresses in
the adhesive layer. Deng et al. [6,7] proposed the utilization of mixed-mode cohesive law
to mimic notch-damaged steel beams strengthened with CFRP plates. They stated that
even though the bearing capacity of the beams was increased, the ductility decreased as the
notch size increased and the premature debonding mode of failure took place. Siwowski
and Siwowska [8] and Siwowska and Siwowski [9] experimentally compared the flexural
behavior of steel beams strengthened with CFRP plates. Four different configurations
of strengthened steel plates with a center notch under fatigue load (adhesive-bonded
non-prestressed plates and adhesive-bonded prestressed plates) were considered. They
demonstrated the possibility of extending the fatigue life of steel plates nearly four times
with non-prestressed plates and, in some cases, completely arresting the crack growth
by using prestressed plates. Elchalakani [10] presented the experimental results of CFRP-
rehabilitated corroded steel circular hollow sections (CHS) tested under quasi-static large
deformations. The results showed that the increase percentage in the combined flexural
and bearing strength of the CHS was generally affected by the corrosion level. For the
CFRP-rehabilitated CHS with 20% corrosion level, the stiffness and capacity were fully
restored; however, for CFRP-rehabilitated CHS with higher corrosion levels of 40–80%, such
stiffness and capacity were not entirely recovered. Chen et al. [11] examined the fatigue
behavior of cracked rectangular hollow section (RHS) steel beams. They revealed details
of strengthening arrangements using CFRP plates with high modulus and conducted
fatigue tests on the strengthened RHS beams. The experimental results indicated that high
modulus CFRP increased the beam fatigue life. Ochi et al. [12] experimentally explored the
strengthening effects of high modulus CFRP strips bonded to the tensile flange of I-shaped
steel girders. Six beam specimens with different adhesive lengths were used and the beams
were tested in flexure. The authors suggested a scheme to prevent debonding of the strips
and experimentally tested it. The experimental results were discussed to highlight the
effectiveness of the strengthening CFRP strips and the suggested debonding prevention
scheme. Ghafoori and Motavalli [13] developed a retrofitting system by prestressing CFRP
plates and attaching them to the steel beams using only a pair of mechanical locks that
function in friction. Design considerations of the locks were proposed. The retrofitted
beams were statically tested to failure and the test results showed an increase in beam
yielding and ultimate capacities. Li el al. [14] designed end anchorage systems to be
used with prestressed CFRP plates to rehabilitate notch-damaged steel beams. They
experimentally studied the rehabilitated beam specimens for flexural behavior. It was
concluded that prestressed CFRP strengthening plates could restore the bearing capacity
and ductility of the notch-damaged steel specimens, delay debonding propagation, and
prevent premature failure. However, the improvement of the rehabilitated beam stiffness
was not evident.
Zhou et al. [15] experimentally and numerically investigated nine notch-damaged
wide flange steel beam specimens. The specimens were retrofitted using carbon fiber rein-
forced polymer (CFRP) laminates and a recently developed carbon-fiber hybrid-polymeric
matrix composite (CHMC) that has been termed CarbonFlex. The peak load deflections
of the CarbonFlex-retrofitted beams were higher than their CFRP-retrofitted counterparts.
Both experimental and numerical results indicated the effectiveness of CarbonFlex, as a
candidate retrofitting material, for damaged steel structures. Hmidan et al. [16] proposed
a modeling approach to simulate debonding of the CFRP sheets and crack propagation
in crack-dependent behavior of CFRP-strengthened steel members subjected to axial ten-
sion. The modeling approach was experimentally validated and implemented into three-
dimensional FE models to examine the effect of various notch sizes on the behavior of
CFRP-strengthened steel members. Chiew et al. [17] tested three types of FRP-bonded steel
joints to investigate the bond failure mechanism of FRP. Numerical simulations of strain
distributions along the bond line were also carried out to verify the experimental work. Us-
ing the equivalent strain energy density based on the failure criterion a bond failure model
was proposed. Tavakkolizadeh and Saadatmanesh [18,19] and Photiou et al. [20] studied
the behavior of damaged steel–concrete composite steel girders repaired with CFRP sheets
under static loading. Three scenarios of 25%, 50%, and 100% loss of tension flange area
of the W-shape cross-section were investigated. Test results revealed that epoxy-bonded
CFRP sheets could restore the stiffness and ultimate load-carrying capacity and of repaired
girders. Comparison between experimental and numerical modeling showed that the
traditional methods of analysis of composite beams were conservative. Wang and Wu [21]
developed two approaches termed symmetrical and asymmetrical prediction approaches to
predict the fatigue crack spreading of FRP-strengthened double-edged cracked steel beams
using finite element (FE) modelling. These approaches were verified experimentally. The
comparative results concluded that the prediction of the fatigue crack propagation of the
double-edged cracked FRP-strengthened steel members could approximately be projected.
Colombi and Fava [22] developed curves for predicting crack growth in CFRP-strengthened
steel beams. They conducted nine fatigue tests on CFRP-strengthened cracked steel beams.
The authors concluded that their developed curves showed that CFRP strengthening strips
prolonged the fatigue life of the beams and reduced the fatigue crack propagation. They
also stated that debonding of the CFRP strips had an unfavorable effect on the efficiency of
the strengthening technique.
A simple analytical approach was developed by Kim and Brunell [23] and Kim and
Harries [24] for estimating the bond length of the various repair strategies. Different CFRP
repair strategies for steel beams with notch damages were examined. The influence of the
notch configurations was studied using finite element (FE) analysis to obtain the interaction
between CFRP-repair and the initial damage of steel beams loaded in flexure. Kadhim [25]
numerically examined the performance of continuous steel beams strengthened with CFRP
plates and focused on the effect of plate length on the steel beam behavior. The numerical
results indicated that when the length of strengthening CFRP plates in sagging and hogging
regions of the beam exceeded 40% and 60% of the beam span length, respectively, the
ultimate strength increase rate decreased. Haeri el al. [26] numerically studied the crack
propagation in beams with double internal cracks tested in a three-point flexural test. They
concluded that when two cracks overlapped, they combined in the failure progression with
decreased stress values when the cracks were positioned at the midpoint of the beam span.
Yu et al. [27] developed a bond failure model to predict the bond failure load of steel beams
strengthened with FRP laminates. They studied the effects of laminate thickness, bond
length, and adhesive thickness on the bond strength. Their study concluded that reducing
the laminate thickness and modulus, reducing the adhesive thickness, or increasing the
bond length can increase the bond strength between the strengthening FRP laminates and
steel beams. Previous studies by Zhou et al. [15], Abambres and Arruda [28], and Ulger
and Okeil [29] examined the effects of initial imperfection on the behavior of strengthened
steel beams. They first executed the preliminary FE models to extract the Eigen buckling
mode shapes, which were later impregnated into the preliminary FE models to impose
initial imperfections in the final FE model. These studies established that using separate
or a combination of mode shapes imposed initial imperfection in the FE modeling of the
control steel beam and the strengthened steel beams had an insignificant effect on the
final results, and that it was pointless to consider initial imperfections in the FE models.
Moreover, Gonçalves [30] stated that if local/distortional buckling is not significant, it
would be recommended to use one-dimensional finite elements as they provide sufficiently
accurate results with a considerably lower computational time and cost.
Mohammed et al. [31] presented a critical review of the existing jacketing techniques
to repair and strengthen declining infrastructure. The review identified present practices
and techniques, highlighted design guidelines to use FRP composite jackets in repairing
structures, and underlined the challenges in future research to increase the acceptance of
using composite repair schemes in infrastructure. The structural performance of three new
proposed railway composite sleepers were investigated experimentally and numerically
by Ferdous et al. [32]. The findings of the investigation indicated that the proposed
composite sleepers were economical and could structurally replace the current railway
track timber sleepers. A review by Tingley et al. [33] and case study by Pongiglione and
Calderini [34] have recommended the reuse of structural steel to reduce embodied carbon
and raw material consumption, despite some barriers such as the quality of the reused
steel, additional cost, and availability of the desired sections. However, reusing structural
steel members with damages was not addressed. A new publication (SCI P427—Structural
Steel Reuse: assessment, testing and design principles) recommends using reclaimed steel
members that were not used in old construction and have not been subjected to extreme
loads such as fire or impact. Further, the publication recommends to avoid using reclaimed
members in structures where they could be subjected to fatigue, such as bridges. However,
this publication did not provide design recommendations for reclaimed members with
damages in temporary infrastructure work.
The literature review demonstrates that researchers have mostly investigated beams
with damages in the tension side and proposed rehabilitation and strengthening techniques
to notch-damaged structural steel members using composite FRP plates and sheets with
some unstrengthened notched beams as reference and reveals the lack of research on
unrepaired steel beams that have notch damage and are laterally unsupported. The
objective of this study is to experimentally and numerically examine the feasibility of
reusing steel beams with notch damage in the tension in temporary structural applications
without being repaired and laterally supported.
2. Experimental Program
2.1. Specimen Specifics
A total of twelve bare I-shaped steel beams of W 203 × 203 × 46.1 sections were
prepared, instrumented, and experimentally tested to examine their flexural performance.
One of the specimens, control beam, was kept undamaged while the other eleven specimens
were pre-damaged by cutting a U-shaped notch in their tension (bottom) flange and
were left unrepaired. The American Institute of Steel Construction manual (AISC) was
utilized in this work to design the control beam. Accordingly, the span and section of the
laterally unsupported control beam were selected to ensure that it would not undergo
lateral torsional buckling. The AISC manual provides design recommendations for tension
members with bolted connections where the bolt holes can be considered as a damage to
the member. Additionally, the AISC manual covers the design of beams with web openings
and beams with missing flange. All twelve laterally unsupported I-shaped specimens were
tested in two-point loadings for flexure and the parameters considered in the study were
the location and the size of the notches. The notch size (depth) considered in this study
was 30 mm, 40 mm, and 60 mm in depth, while the notch opening width was kept constant
at 20 mm in all the damaged specimens, as shown Figure 1. All beams had a total length
of 2000 mm and an actual span of 1800 mm between supports, as shown in Figure 1. The
location of the notch selected in this study was at one-quarter, one-third, and the middle of
the specimen clear span. All pre-damaged steel specimens had notches on one side of the
tension flange with respect to the beam web except for two specimens where two notches
were cut at the middle of the specimen clear span on both sides of the tension flange and
had two different depths, 40 mm and 60 mm.
Infrastructures2021,
Infrastructures 2021,6,6,69
69 55of
of25
25
Figure 1. Dimensionsof
of damagedbeam
beam specimens.
Figure1.1.Dimensions
Figure Dimensions ofdamaged
damaged beamspecimens.
specimens.
All
Alltested
All testedI-shaped
I-shapedsteel
steelbeam
beamspecimens
specimensininthisthisstudy
studyhad hadthe
thesame
samesection,
section, WW 203
203××
Alltested
testedI-shaped
I-shapedsteel
steelbeam
beamspecimens
specimensin inthis
thisstudy
studyhad hadthe
thesame
samesection,
section,W W203
203××
203 × 46.1. The
203 × 46.1.The
203 flange and
The flangeand web specifics
and webspecifics are listed
specifics arelisted Table 1. Figure
listed Table1.1.Figure 2 shows
Figure 2 showsthethe front view
the frontview of
view
203××46.1.
46.1. Theflange
flange andwebweb specificsareare listedTable
Table 1. Figure22shows shows thefront front view
aoftypical
ofa beam
atypical
typical specimen
beam
beam during
specimen
specimen testtest
during
during setup.
test Vertical
setup.
setup. stiffeners
Vertical
Vertical werewere
stiffeners
stiffeners installed
were at theat
installed
installed atspecimen
thespec-
the spec-
of a typical beam specimen during test setup. Vertical stiffeners were installed at the spec-
support and loading
imen support
imen point locations,
and loading as shown
point locations, in Figures
as shown 1 and 12,and
in Figures to avoid web crippling.
2, toavoid
avoid web crip-
imensupport
supportand
andloading
loadingpoint
pointlocations,
locations,asasshown
shownin inFigures
Figures11and and2,2,toto avoidweb webcrip-
crip-
All damaged
pling.
pling. All specimens,
All damaged
damaged as well as
specimens,
specimens, asthe
as control
well
well as the
as specimen,
the were laterally
control specimen,
control specimen, were
were unsupported and
laterally unsup-
laterally unsup-
pling. All damaged specimens, as well as the control specimen, were laterally unsup-
unrepaired.
ported andunrepaired.
ported unrepaired.
portedandand unrepaired.
Table
Table
Table Typical
1. 1. details
Typical ofof
details thethe
examined I-beam
examined (W-shape:
I-beam WW
(W-shape: 203 × ×203
203 203××46.1).
46.1).
Table1.1.Typical
Typicaldetails
detailsof
ofthe
theexamined
examinedI-beam
I-beam(W-shape:
(W-shape:WW203
203××203
203××46.1).
46.1).
Product
Product
Product Dimensions
Dimensions
Dimensions
Product Dimensions
Mass per
Mass Web Flange Web Flange Root
Massper
per Web
Web Flange
Flange Web
Web Flange
Flange Root
Root
Mass per
meter
meter Web
height
height Flange
width
width Web
thickness
thickness Flange
thickness
thickness Root
radius
radius
meter
meter height
height width
width thickness thickness
thickness thickness radius
radius
Designation
Designation
Designation GG
GG
hh b
bb twtw tftf rr
Designation hh b tw
tw tftf rr
Inch × inch × lbs/ft mm × mm × kg/m kg/m
kg/m
kg/m mmmm
mm mm
mm
mm mm
mm
mm mm
mm
mm mm
mm
mm
Inchxinchxlbs/ft
Inchxinchxlbs/ft mmxmmxkg/m
mmxmmxkg/m kg/m mm mm mm mm mm
Inchxinchxlbs/ft
W 8 × 8 × 31 Wmmxmmxkg/m
203 × 203 × 46.1 46.1 203 203 7.2 11.0 10
W 8 × 8 × 31
W W 203××203
203 × 46.1 46.1 203 203 7.2 11.0 10
W88××88××31
31 W
W203
203 × 203××46.1
46.1 46.1
46.1 203
203 203
203 7.2
7.2 11.0
11.0 10
10
Figure 2. Beamtest
Figure test setup.
Figure2.2.Beam
Beam testsetup.
setup.
Figure 2. Beam test setup.
2.2.Specimen
2.2.
2.2. SpecimenCategorizing
Categorizing
2.2.Specimen
SpecimenCategorizing
Categorizing
Elevenbeams
Eleven
Eleven beamsoutoutof
ofthe
thetwelve
twelvetested
testedsteel
steelbeam
beamspecimens
specimenshad hadan anartificially
artificiallymade
made
Elevenbeams
beamsoutoutof
ofthe
thetwelve
twelvetested
testedsteel
steelbeam
beamspecimens
specimenshad hadan anartificially
artificiallymade
made
U-shaped
U-shaped
U-shaped notch
notch damage
damage of
of different
different depths
depths and
and different
different locations,
locations, while
while the
the width
width of
of the
the
U-shapednotch
notchdamage
damageof ofdifferent
differentdepths
depthsand
anddifferent
differentlocations,
locations,while
whilethe
thewidth
widthofofthethe
notches
notches
notches was
was kept
kept constant.
constant. The
The notch
notch layout
layout was
was marked
marked on
on the
the flange
flange to
to make
make sure
sure of
of
notcheswaswaskept
keptconstant.
constant. The
The notch
notch layout
layout was
wasmarked
marked on onthe
theflange
flangetotomake
makesuresureofof
the
the cut dimensions and the flange was then saw-cut. One beam had no artificial damage
thecut
cutdimensions
dimensionsand
andthe
theflange
flangewas
wasthen
thensaw-cut.
saw-cut.One
Onebeam
beamhadhadno noartificial
artificialdamage
damage
the cut dimensions and the flange was then saw-cut. One beam had no artificial damage
and was denoted as a control beam. Nine of the eleven notch-damaged beams had a single
notch (denoted as S) on one side of their tensile flange at different locations, while two
damaged specimens had double notches (denoted as D) on both sides of the tension flange
at the middle of the beam. The eleven damaged specimens were categorized into three
sets, as presented in Table 2, depending on the location of the notch: group Q refers to
notch located at one quarter of the beam clear span; group T refers to notch located at
one-third of the beam clear span; and group M refers to notch located at the beam clear
mid-span. The damaged beams were labeled as Q-S-X, T-S-X, M-S-X, or M-D-X, as shown
in Table 2, where the X represents the notch depth of 30 mm, 40 mm, and 60 mm, while
the notch opening width was kept constant at 20 mm in all damaged beams. The control
beam was labeled C-NG-0 (control, no groove/notch, zero notch depth). The notch depth
from 0 to 60 mm, representing up to 60% of the flange free width, was selected to simulate
minor to moderate damage in the beam and to avoid any sudden flipping of the laterally
unsupported damaged steel specimen during testing.
Table 2. Test matrix—specimen categorizing.
Set (Location of Notch) * Notch Depth (mm) Specimen **

C 0 C-NG-0
30 Q-S-30
Q 40 Q-S-40
60 Q-S-60
30 T-S-30
T 40 T-S-40
60 T-S-60
30 M-S-30
M 40 M-S-40
60 M-S-60
MD 30 M-D-30
MD 40 M-D-40
* Location of notch: C = control specimen; Q = notch at quarter of specimen length; T = notch at one-third of
specimen length; M = notch in middle of specimen; S = single notch on one edge of tension flange; D = double
notches on both edges of tension flange. ** NG-0 = zero groove/notch; specimen example: QS-30 = single notch
30 mm depth on one edge of tension flange at quarter of specimen length; TS-40 = single notch 40 mm depth on
one edge of tension flange at one-third of specimen length; MD-40 = double notches 40 mm depth on both edges
of tension flange in the middle of specimen.
2.3. Material Properties

The steel I-shaped specimens tested in this study were made of W 203 × 203 × 46.1
mild steel. The mechanical properties of the steel of the specimens were acquired from
standard tensile coupon test. Twelve coupon samples were cut form the top left flange, the
bottom right flange, and the middle of the web of the test steel specimens. The coupon
samples were prepared and tested in line with ASTM E8 2016 standard. The average
ultimate and yielding strengths achieved were 431 ± 12 MPa and 289 ± 8 MPa, respectively
(according to the manufacturer’s catalogue, the ultimate strength is 420 MPa and the
yielding strength is ≥250 MPa).
2.4. Instrumentation of Test Setup

The steel beam specimens were tested in a two-point loading test. The clear span
of all specimens was 1800 mm. The loading points were 200 mm to the left and right of
the specimen middle section and 700 mm away from the specimen supports, as shown
in Figure 1. The damaged beam specimens were left unrepaired and vertical stiffeners
were installed at specimen support and loading point locations to avoid web crippling, as
shown in Figures 1 and 2. The laterally unsupported beam specimens were loaded in a
displacement control mode at a rate of 2 mm/min.
shown in Figures 1 and 2. The laterally unsupported beam specimens were loaded in
displacement control mode at a rate of 2 mm/min.
All tested beam specimens were prepared and instrumented with linear variable di
Infrastructures 2021, 6, 69 ferential transformers (LVDTs) and strain gauges for deflection and strain7 of readings.
24 Th
strain gauges were attached to the external surface of bottom and top flanges and the we
of the specimens at different locations, as displayed in Figure 3. The LVDTs were place
under the bottom flange at the loading point locations and at the middle section of th
All tested beamasspecimens
specimen, were prepared
shown in Figure 3. and instrumented with linear variable
differential transformers
The experimental setup was selectedfor
(LVDTs) and strain gauges deflection
in light of theand strain readings.
literature The
review presented in th
strain gauges were attached to the external surface of bottom and top flanges and the web
paper, in which most of the cited references address members with notches in the tensio
of the specimens at different
side of the beam. Itlocations,
should beas displayed
noted in Figure
that since 3. The LVDTs
the damaged were is
steel beam placed
to be reused,
under thecanbottom flange at the loading point locations and at the middle section
be rotated and flipped around in the site to properly position the notch in of thethe tensio
specimen,zone
as shown
of theinbeam.
Figure 3.
Stain gauges LVDTs
Strain gauges placed symmetrically at top and bottom flanges and in web.
Strain gauges bonded to flange at midspan, different cross-sections and notch location.
Figure
Figure 3. Beam specimen 3. Beam specimen instrumentation.
instrumentation.
The experimental
3. Discussionsetup was selectedResults
of Experimental in light of the literature review presented in the
paper, in whichThe
most of the cited
purpose references
of the address
experimental members
testing was with
mainlynotches in the
to trace thetension
load–deflectio
side of thecurve
beam.and
It should be noted that since the damaged steel beam is to
to collect readings from the strain gauges bonded to the specimenbe reused, it up to th
can be rotated and flipped around in the site to properly position the notch in the tension
post-yielding zone. However, since the steel beam specimens were laterally unsupporte
zone of theextra
beam.
caution was required during testing in terms of an immediate halt to the test to tr
to prevent
3. Discussion the actuator
of Experimental in the testing frame from tilting when buckling/lateral rotatio
Results
occurred in the specimens. Accordingly, some specimen tests took a longer time tha
The purpose of the experimental testing was mainly to trace the load–deflection
other tests, and consequently more data were gathered, depending on how soon the buck
curve and to collect readings from the strain gauges bonded to the specimen up to the
ling initiated causing the actuator to tilt and the test to stop. It should be noted that th
post-yielding zone. However, since the steel beam specimens were laterally unsupported,
test was originally planned to have a maximum beam deflection equal to the beam clea
extra caution was required during testing in terms of an immediate halt to the test to try
span over 120 (1800/120 = 15 mm). However, the early unanticipated twisting of the actu
to prevent the actuator in the testing frame from tilting when buckling/lateral rotation
ator forced the test to be stopped earlier and before reaching the beam intended deflectio
occurred in the specimens. Accordingly, some specimen tests took a longer time than other
tests, and consequently more data were gathered, depending on how soon the buckling
initiated causing the actuator to tilt and the test to stop. It should be noted that the test was
originally planned to have a maximum beam deflection equal to the beam clear span over
120 (1800/120 = 15 mm). However, the early unanticipated twisting of the actuator forced
the test to be stopped earlier and before reaching the beam intended deflection.
3.1.
3.1.Failure Mode
Failure Mode
The steel
The steelspecimens inin
specimens this study
this were
study not
were tested
not toto
tested ultimate since
ultimate they
since were
they laterally
were laterally
unsupported to simulate their applications in temporary structures. All tested specimens
unsupported to simulate their applications in temporary structures. All tested specimens
exhibited the same failure mode of flange buckling causing the actuator to twist, as shown
exhibited the same failure mode of flange buckling causing the actuator to twist, as shown
in Figure 4, and the test had to stop to avoid any hazardous consequences. As a result,
in Figure 4, and the test had to stop to avoid any hazardous consequences. As a result, the
the tests did not reach the ultimate load and the descending branch of the load–deflection
tests did not reach the ultimate load and the descending branch of the load–deflection
curve could not be obtained because directly after a buckling in the compression flange
curve could not be obtained because directly after a buckling in the compression flange
developed, the actuator tilted instantly and the test had to be stopped immediately for
developed, the actuator tilted instantly and the test had to be stopped immediately for
safety reasons and to avoid unreliable test readings. Hence, the peak load was determined
safety reasons and to avoid unreliable test readings. Hence, the peak load was determined
as the load at which the compression flange buckled causing the actuator to suddenly tilt.
as the load at which the compression flange buckled causing the actuator to suddenly tilt.
Tilting of actuator and spreader beam. Buckling of specimen and yielding of notch surrounding area.
Figure
Figure 4. 4. Failure
Failure mode
mode of of specimens.
specimens.
3.2. Applied Load versus Mid-Span Deflection

3.2. Applied Load versus Mid-Span Deflection
All the notch-damaged steel specimens had their notches in the tension flange located
All the notch-damaged steel specimens had their notches in the tension flange located
at a quarter
at a quarter (Q) (Q) and
and third
third (T)(T)
ofof
thethe clear
clear span
span from
from thethe right
right support
support andandat at
the the midspan
midspan
(M).
(M). The
The applied
applied load–midspandeflection
load–midspan deflectionrelations
relationsfor forallall tested
tested beams
beams are
are depicted
depicted inin
Figure 5. Initially, all the beams had the same stiffness and the deflection
Figure 5. Initially, all the beams had the same stiffness and the deflection increased linearly increased line-
atarly at therate
the same same rate
until theuntil the specimens
specimens started tostarted to yield.
yield. After After the
yielding, yielding,
specimen the stiffness
specimen
stiffness decreased, causing the deflection to increase at a faster rate.
decreased, causing the deflection to increase at a faster rate. It is clear from Figure 5 thatIt is clear from Figure
all
tested specimens had almost the same load-carrying capacity and that their behavior had be-
5 that all tested specimens had almost the same load-carrying capacity and that their a
haviorpattern.
similar had a similar pattern.
However, FigureHowever,
5a shows Figure 5a shows
that the absence that
of the absenceand
symmetry of symmetry
closeness of and
closeness
the notch toof thethe notch appeared
support to the support
to be appeared to be the more
the more significant factors significant
affecting factors affecting
the stability of
the stability of the beam than the size of the notch. As indicated above,
the beam than the size of the notch. As indicated above, the tests stopped when the actuator the tests stopped
when the
twisted andactuator twistedwere
the specimens and unloaded.
the specimens It is were
worthunloaded.
noting that It isthis
worth
studynoting that this
considered
study considered only two specimens with double
only two specimens with double notches, M-D-30 and M-D-40. notches, M-D-30 and M-D-40.
69 99ofof25
24
-500 -500
-450 -450
-400 -400
Applied load P(kN)
Applied load P(kN)

-350 -350
-300 -300
-250 -250
-200 -200
Q−S−30 T−S−30
-150 Q−S−40 -150
T−S−40
-100 Q−S−60 -100 T−S−60
-50 Control beam -50 Control beam
0 0
0 -5 -10 -15 -20 -25 -30 0 -5 -10 -15 -20 -25 -30
Midspan deflection (mm) Midspan deflection (mm)
(a) Control beam and beams with single notch in tension (b) Control beam and beams with single notch in tension flange
flange at quarter span. at third span.
-500
-500
-450
-450
-400
Applied load P(kN)
-400
Applied load P(kN)
-350
-350
-300
-300
-250
-250
-200
M−S−30 -200
-150
M−S−40 -150 M−D−30
-100 M−S−60 -100 M−D−40
-50 Control beam Control beam
-50
0
0
0 -5 -10 -15 -20 -25 -30
0 -5 -10 -15 -20 -25 -30
Midspan deflection (mm)
(c) Control beam and beams with single notch in tension (d) Control beam and beams with double notches in tension
flange at mid span. flange at mid span.
Figure
Figure5.5.Experimental
Experimentalapplied
appliedload–midspan
load–midspandeflection
deflectionrelations.
relations.
3.3. Applied Load versus Strain

3.3. Applied Load versus Strain
Figure 6 presents the applied load–strain relations in all tested steel specimens. Fig-
Figure 6 presents the applied load–strain relations in all tested steel specimens.
ure
Figureshows
6a 6a showsthe response
the responseof theof extreme top (compression)
the extreme top (compression) and bottom (tension)(tension)
and bottom strains
atstrains
the midspan of the control beam. The strains in Figure 6b–e
at the midspan of the control beam. The strains in Figure 6b–e are the top and are the top and bottom
strains
bottomatstrains
cross-sections passing through
at cross-sections passingthe notch. the
through Figure 3 shows
notch. Figurethe3strain
shows gauges at a
the strain
section passing through the notch. For the control beam, it is apparent
gauges at a section passing through the notch. For the control beam, it is apparent from from Figure 6a that
the top and
Figure bottom
6a that the top strain
andresponses
bottom strain wereresponses
almost identical,
were almost which is expected
identical, which asisthe strain
expected
gauges were located at the section of maximum moment of
as the strain gauges were located at the section of maximum moment of the undamagedthe undamaged beam. It can
also
beam.be noticed thatbe
It can also atnoticed
a load of 420atkN,
that the of
a load top420 strain
kN,response showed
the top strain a clearshowed
response yieldinga
point
clear beyond
yieldingwhich
point the strain
beyond demonstrated
which the strain ademonstrated
yielding plateau behavior
a yielding with behavior
plateau a minor
increase in the applied load; whereas the bottom strain response
with a minor increase in the applied load; whereas the bottom strain response showed showed a change in the
rate of the increasing strain at a load magnitude of 280 kN, and
a change in the rate of the increasing strain at a load magnitude of 280 kN, and as theas the load was increased,
the
loadbottom strain increased
was increased, at a faster
the bottom strain rate until the
increased at astrain
fasterreached
rate untilthetheyield
strainplateau
reached zone
the
atyield
a load of 420
plateau kN.atAll
zone of Figure
a load of 4206b–e show
kN. All that the6b–e
of Figure top show
and bottom
that thestrain
top andresponses
bottom
presented similar presented
strain responses pattern as that similarof the control
pattern as beam
that ofbut thewith different
control beamstrain values.
but with Fig-
different
ure 6b,cvalues.
strain show that for notches
Figure 6b,c show located at one-quarter
that for notches located and one-third of theand
at one-quarter specimen
one-thirdspan,of
including
the specimen the different notch sizes
span, including considered
the different notch in sizes
this study,
consideredthe beam
in thistop andthe
study, bottom
beam
strain
top andvalues at thestrain
bottom notchvalues
sectionatdid thenot significantly
notch section did vary.notIt significantly
can be concluded vary.that if the
It can be
notch, with the
concluded thatsizes
if theexamined
notch, with in this
thestudy, is located in
sizes examined anywhere
this study,within one-third
is located of the
anywhere
beam
within clear span, of
one-third thethe beam
beam behavior
clear span,wasthe not sensitive
beam to the
behavior wasnotch location.toHowever,
not sensitive the notch
location.
when theHowever, when
notch is at the the notch is
midspan ofatthe
thebeam
midspan of the
(Figure beam
6d,e), the(Figure
top and6d,e), the top
bottom and
strains
bottom strains increased significantly, indicating that the beams exhibited ductile behavior
increased significantly, indicating that the beams exhibited ductile behavior and were able
and were able to undergo larger deflections. As a result, when the notch is at the midspan
to undergo larger deflections. As a result, when the notch is at the midspan of the beam,
of the beam, the beam becomes more resilient. This change in behavior can be attributed
the beam becomes more resilient. This change in behavior can be attributed to the location
to the location of the notch. When the notches were at one-quarter and one-third of the
of the notch. When the notches were at one-quarter and one-third of the clear span, the
clear span, the steel beams were not symmetrical, causing the vertical deflected shape to
steel beams were not symmetrical, causing the vertical deflected shape to be unsymmet-
be unsymmetrical as well. As the load was increased, the unsymmetrical cross-section of
rical as well. As the load was increased, the unsymmetrical cross-section of the beams at
the beams at the notch location made the beam to buckle unsymmetrically out of plane.
the notch location made the beam to buckle unsymmetrically out of plane. The simulta-
The simultaneous effects of unsymmetrical vertical deflection and out of plane buckling
neous effects of unsymmetrical vertical deflection and out of plane buckling expedited the
expedited the torsion of the beams at small strain responses, whereas for beams with
torsion of the beams at small strain responses, whereas for beams with notches at their
notches at their centerlines, the beams were symmetrical, causing a symmetrical vertical
centerlines, the beams were symmetrical, causing a symmetrical vertical deflected shape
deflected shape and the symmetrical out of plane buckling of the beam. As a result of this
and the
dual symmetrical
symmetry out of plane
in deflection buckling
and out of the
of plane beam. the
buckling, As abeams
resultshowed
of this dual
more symmetry
resilient
in deflection
responses as and out of
the load plane
was buckling,
increased and,theeventually,
beams showed more resilient
experienced larger responses as the
top and bottom
load was
strains. increased and, eventually, experienced larger top and bottom strains.
500
400
Applied Load
300
Control beam
(kN)
200
100
0
-25,000 -20,000 -15,000 -10,000 -5000 0 5000 10,000 15,000 20,000 25,000
Microstarin με (at midspan
(a)
-500
Applied Load (kN)
-400
-300
Q−S−30
-200 Q−S−40
-100 Q−S−60
0
-25,000 -20,000 -15,000 -10,000 -5000 0 5000 10,000 15,000 20,000 25,000
Microstarin με (at quarter of span)
(b)
-500
Applied Load (kN)
-400
-300 T−S−30
T−S−40
-200 T−S−60
-100
0
-25,000 -20,000 -15,000 -10,000 -5000 0 5000 10,000 15,000 20,000 25,000
Microstarin με (at third of span)
(c)
Figure 6. Cont.
Infrastructures 2021, 6,
Infrastructures 2021, 6, 69
69 11
11 of
of 24
25
-500
-400
Applied Load
-300 M−S−30
(kN)
-200 M−S−40
M−S−60
-100
0
-25,000 -20,000 -15,000 -10,000 -5000 0 5000 10,000 15,000 20,000 25,000
Microstarin με (at midspan with single notch in bottom flange)
(d)
-500
-400
Applied Load
-300
(kN)
-200 M−D−40
M−D−30
-100
0
-25,000 -20,000 -15,000 -10,000 -5000 0 5000 10,000 15,000 20,000 25,000
Microstarin με (at midspan with double notches on both sides of bottom flange)
(e)
Figure 6.
Figure 6. Applied load–strain response
Applied load–strain response of
of specimens.
specimens.
3.4.
3.4. Deflection-Strain
Deflection-Strain Response
Response
Figure
Figure 7 shows the midspan
7 shows the midspan deflection–strain
deflection–strain responses
responses in in all
all tested
tested steel
steel specimens.
specimens.
Figure
Figure 7a,b
7a,b show
show that
that the the top
top and bottom strains
and bottom strains at the location
at the location of the notch
of the notch were
were very
very
small and somewhat the same as the deflection increased. However,
small and somewhat the same as the deflection increased. However, when the notch was when the notch was
at the midspan of the beam, the top and bottom strains increased
at the midspan of the beam, the top and bottom strains increased significantly as deflection significantly as deflec-
tion increased,
increased, as shown
as shown in Figure
in Figure 7c,d.7c,d.
TheThelatterlatter figures
figures alsoalso indicate
indicate thatthat as the
as the deflec-
deflection
tion
reached 5 mm, the bottom flange started to yield while the top flange remained elastic.elas-
reached 5 mm, the bottom flange started to yield while the top flange remained As
tic.
the As the deflection
deflection increased,increased, the bottom
the bottom strain increased
strain increased muchthan
much faster faster thethan
topthe top strain.
strain. When
When the deflection
the deflection was almostwas almost
twice as twice
much as much
(about(about
10 mm), 10 the
mm), topthe top flange
flange startedstarted
to yield to
yield while the bottom flange had experienced large inelastic
while the bottom flange had experienced large inelastic deformation. This is attributed todeformation. This is at-
tributed
the beam’s to neutral
the beam’saxis neutral axis being
being shifted shiftedthe
up towards uptoptowards
flangetheduetop flange
to the notchdue to the
being in
notch being in the bottom flange. It can be established from Figure
the bottom flange. It can be established from Figure 7c,d that when the deflection reached 7c,d that when the
deflection
almost 10 mm, reachedthe almost
top strain10 mm, the top
response was strain
linear response
and it waswasnot
linear and it to
sensitive wasthenot sensi-
location
tive to notch.
of the the location
On the ofother
the notch.
hand,On thethe other strain
bottom hand, response
the bottom strain
was response affected
significantly was signif- by
icantly affected
the location bynotch.
of the the location of the
Therefore, thenotch.
locationTherefore, the location
of the notch of thea notch
did not have major did not
impact
have
on thea deflection
major impact of the onbeam
the deflection
but it did of the beam
accelerate thebut it did accelerate
yielding and plasticthe yielding
behavior ofand
the
plastic behavior of the tension part of the beam way before
tension part of the beam way before it did for the compression part of the beam. it did for the compression part
of the beam.
-30
Midspan Deflection
-25
-20 Q−S−30
(mm)
-15 Q−S−40
-10 Q−S−60
-5
0
-25,000 -20,000 -15,000 -10,000 -5000 0 5000 10,000 15,000 20,000 25,000
Microstarin με (at quarter of span)
(a)
-30
Midspan Deflection
-25
-20
(mm)
T−S−30
-15
T−S−40
-10 T−S−60
-5
0
-25,000 -20,000 -15,000 -10,000 -5000 0 5000 10,000 15,000 20,000 25,000
Microstarin με (at third of span)
(b)
-30
Midspan Deflection
-25
-20 M−S−30
(mm)
-15 M−S−40
-10 M−S−60
-5
0
-25,000 -20,000 -15,000 -10,000 -5000 0 5000 10,000 15,000 20,000 25,000
Microstarin me (at midspan with single notch in bottom flange)
(c)
-30
Midspan Deflection
-25
-20
(mm)
-15
-10 M−D−40
-5 M−D−30
0
-25,000 -20,000 -15,000 -10,000 -5000 0 5000 10,000 15,000 20,000 25,000
Microstarin με (at midspan with double notches on both sides of bottom flange
(d)
Figure 7.
Figure 7. Midspan
Midspan deflection–strain
deflection–strain response
response of
of specimens.
specimens.
test results
The test results in
inFigures
Figures5–7
5–7show
showthat
thatthe
theworst-case
worst-case scenario
scenario in in terms
terms of defor-
of deforma-
mations
tions waswas
whenwhen
the the specimens
specimens had had a notch
a notch damage
damage atmidspan.
at the the midspan.
Based Based onoutcome,
on this this out-
come,
the the comparison
comparison betweenbetween the numerical
the numerical and
and test test results
results focused
focused only
only on on specimens
specimens with
with notch
notch damages
damages at theatmidspan
the midspan andother
and the the other specimens
specimens were were excluded.
excluded.
4. FE Modeling (ATENA-3D)
To further explore the flexural performance of the steel I-shaped beams, all tested
4. FE Modeling
specimens (ATENA-3D)
were modelled using a finite-element software ATENA-3D, version 5.6.0 (Cer-
venkaTo et further
al. [35]).explore
The authors have used
the flexural ATENA of
performance in the
previous studies [36,37]
steel I-shaped beams,due to its
all tested
specimens
high werecapability.
simulation modelled Moreover,
using a finite-element
the accuracysoftware ATENA-3D,
of numerical version
results in 5.6.0 con-
this study (Cer-
venka et
firmed theal.validity
[35]). Theofauthors have used
this software. TheATENA in previous
specimen models studies [36,37] due to in
were accomplished its high
two
simulation
steps: capability.and
pre-processing Moreover, the accuracy
post processing. of numerical
Material selection,results in this dimensions
geometrical study confirmedand
the validity of this software. The specimen models were accomplished
shapes, and boundary conditions, and mesh generation of the specimens were identifiedin two steps: pre-
processing and post processing.
in the pre-processing phase. Material selection, geometrical dimensions and shapes,
and boundary conditions, and mesh generation of the specimens were identified in the
pre-processing
4.1. phase.
Material Modeling
The material stress–strain relationship of the W 203 × 203 × 46.1 steel beam specimens
4.1. Material Modeling
was represented by a bilinear elastic–perfectly plastic model with a yielding strength of
The elastic
280 MPa, material stress–strain
modulus of 200relationship
GPa, and aofhardening × 203 × 46.1
the W 203modulus of 0 steel
GPa.beam specimens
The steel spec-
was represented by a bilinear elastic–perfectly plastic model with a yielding strength
imens in this study were not tested to ultimate since they were laterally unsupported to of
280 MPa, elastic modulus of 200 GPa, and a hardening modulus of 0 GPa. The steel
simulate their applications in temporary structures. Using the elastic perfectly plastic
specimens in this study were not tested to ultimate since they were laterally unsupported
stress–strain relationship for the steel material would be more representative of the test
to simulate their applications in temporary structures. Using the elastic perfectly plastic
results. Therefore, this relationship was adopted in this study. The constitutive stress–
stress–strain relationship for the steel material would be more representative of the test
strain relation of the utilized loading and bearing steel plates was denoted as the elasto-
results. Therefore, this relationship was adopted in this study. The constitutive stress–strain
isotropic model and these steel plates were assumed to remain elastic with a modulus of
relation of the utilized loading and bearing steel plates was denoted as the elasto-isotropic
elasticity of 200 GPa. Von Mises yield criterion was considered as the failure criterion in
model and these steel plates were assumed to remain elastic with a modulus of elasticity of
the FE analysis.
200 GPa. Von Mises yield criterion was considered as the failure criterion in the FE analysis.
4.2.
4.2.Mesh
MeshGeneration
Generation
ATENA-3D
ATENA-3Dsoftwaresoftwarehashasthree
threedifferent
differenttypes
typesofof3D
3Dsolid
solidelements:
elements:tetrahedral
tetrahedralele-ele-
ments 3D (CCIsoTetra), brick elements (CCIsoBrick), and wedge elements
ments 3D (CCIsoTetra), brick elements (CCIsoBrick), and wedge elements (CCIsoWedge). (CCIsoWedge).
In
In this
this study, thethree
study, the threeelement
elementtypes
types were
were examined
examined to tune
to tune themesh,
the FE FE mesh,
and itand
wasitfound
was
found
that thethat the tetrahedral
tetrahedral 3D (CCIsoTetra)
3D (T3D) (T3D) (CCIsoTetra) solid element
solid element was thewas
mostthe most appropriate
appropriate element
element
that bestthat best explained
explained the behavior
the behavior of the specimens
of the specimens and, therefore,
and, therefore, wasin
was used used
thisinstudy
this
study to generate the FE models. Hence, the maximum T3D element
to generate the FE models. Hence, the maximum T3D element size of 25 mm and up to size of 25 mm and
up to 16500
16500 T3D T3D elements
elements usedused in steel
in the the steel specimen
specimen modeling
modeling were
were found
found to to
bebe adequate
adequate to
to produce
produce acceptable
acceptable results
results within
within a balanced
a balanced analysis
analysis running
running time.
time. Figure
Figure 8 shows
8 shows a
a FE
FE
meshmesh of the
of the W-shaped
W-shaped steelsteel
beam beam
withwith a notch
a notch in thein the bottom
bottom tensiontension
flange flange at the
at the middle
middle of the span.
of the span.
Figure
Figure8.8.ATENA
ATENA3D
3DFE
FEmodel—mesh
model—meshgeneration.
generation.
69 14of
14 of24
25
4.3. Analysis Method

4.3. Analysis Method
The Newton–Raphson procedure was utilized by the software when the numerical
The Newton–Raphson procedure was utilized by the software when the numerical
models of all specimens were executed. The specimens were subjected to displacement-
models of all specimens were executed. The specimens were subjected to displacement-
control loading that would allow displacements to continue increasing even after the spec-
control loading that would allow displacements to continue increasing even after the
imen fails.
specimen fails.
5. Discussion
5. Discussion of of Test
Test and
andNumerical
NumericalOutcomes
Outcomes
The load–deflection
The load–deflection curves curves diddid not
not reach
reach thethe peak
peak loads
loads because
because thethe compression
compression
flange of the test steel specimens buckled. Once the local buckling
flange of the test steel specimens buckled. Once the local buckling in the compression in the compression
flange developed,
flange developed, the the actuator
actuator tilted
tilted instantly
instantly andand thethe test
test was
was stopped
stopped immediately
immediately for for
safety reasons. Therefore, the curves could not reach
safety reasons. Therefore, the curves could not reach the peak loads. the peak loads.
In this
In this study,
study,the theultimate
ultimateloadloadwaswasdetermined
determinedas asthe
the load
load at at which
which the
the compression
compression
flangebuckled,
flange buckled,causing
causingthe theactuator
actuator toto suddenly
suddenly tilttilt
andand forcing
forcing the the authors
authors to stop
to stop the
the test
test immediately
immediately for safety
for safety reasons.
reasons. ThisThis
is theisload
the load
that that caused
caused the beginning
the beginning of perma-
of permanent
nent deformation
deformation in theintest
thespecimens.
test specimens.
Figure 9a–f show a
Figure 9a–f show a comparison comparison between
between the the experimental
experimental and and numerical
numerical FE FE load–
load–
deflection responses of
deflection of each
eachindividual
individualspecimen
specimenwith withnotches
notches at the midspan.
at the midspan. It is clear
It is
that the
clear thatproposed
the proposed FE model predicted
FE model load–deflection
predicted load–deflection responses are identical
responses and com-
are identical and
parable to their
comparable experimental
to their experimental counterparts.
counterparts. Figure 9b,f 9b,f
Figure showshow that that
the numerically
the numerically pre-
dicted stiffness
predicted stiffnessandandpeakpeakload were
load werealmost
almostthethesamesameas those
as those of experimental
of experimental work for
work
specimens
for specimens with single
with andand
single double
double notches,
notches, M-S-30
M-S-30 andandM-D-40,
M-D-40, whereas
whereas forfor
thetherest of
rest
thethe
of specimens,
specimens, thethe
numerically
numerically predicted
predicted stiffness
stiffnessandand peak
peakload were
load wereon onthethe
average
averagebe-
tween 2–4%
between 2–4%higher
higherthanthanthetheexperimentally
experimentallyobtainedobtained ones.
ones. This
This slight discrepancy
discrepancy could could
be
be attributed
attributed to to misalignment
misalignment of of the
the specimen
specimen or or the
the deflection
deflection measuring
measuring devicedevice during
during
testing.
testing. These
Theseidentical
identicalresults
resultsbetween
betweenthe experimental
the experimental andandnumerical
numerical FE FEload–deflection
load–deflec-
responses
tion responses ensure that that
ensure the proposed
the proposed FE model
FE modelcan can
be utilized for further
be utilized investigation
for further investigation of
the behavior of damaged steel beams with notch sizes and locations
of the behavior of damaged steel beams with notch sizes and locations other than the ones other than the ones
considered
considered in in the
the experimental
experimental program
programof ofthis
thisstudy.
study.
-500 -500
-450 Control Beam -450
-400 M–S–30
-400
Applied load P (kN)
Applied load P (kN)
-350 -350
-300 -300
-250 -250
-200 -200
-150 FE -150 FE
-100 Exp. -100 Exp.
-50 -50
0 0
0 -10 -20 0 -5 -10 -15 -20
(a) (b)
Figure 9. Cont.
-500 -500
-500
-450 -500
-450
-450 M–S–40 M–S–60
-400 M–S–40 -450
-400
P (kN)
P (kN)
-400
-350 -400 M–S–60
-350
P (kN)
P (kN)
-350
-300 -350
-300
-300
load
-250 -300
load
-250
load
-250
load
-200 -250
-200
Applied
Applied
-200
-150 FE -200
-150 FE
Applied
Applied
-150
-100 Exp.
FE -150
-100 FE
Exp.
-100
-50 Exp. -100 Exp.
-50
-50
0 -50
0
0 0 -5 -10 -15 -20 0 0 -10 -20
0 -5
Midspan -10
deflection -15
(mm) -20 0 -10
Midspan deflection (mm) -20
Midspan deflection
(c) (mm) Midspan
(d) deflection (mm)
-500 (c) -500 (d)
-500
-450 M–D–30 -500
-450 M–D–40
-450
-400 M–D–30 -450
-400 M–D–40
P (kN)
P (kN)
-400
-350 -400
-350
P (kN)
P (kN)
-350
-300 -350
-300
-300
-250 -300
-250
load
load
-250 -250
-200
load
-200
load
Applied
-200 -200
-150 FE
Applied
-150 FE
Applied
-150
-100 FE
Applied
-150
-100 FE Exp.
Exp. -100
-50
-100
-50 Exp.
Exp. -50
-50
0 0
0 0 -5 -10 -15 -20 0 0 -10 -20
0 -5
Midspan -10
deflection -15
(mm) -20 0 Midspan deflection (mm) -20
-10
(e) (f)
(e) (f)
Figure 9. Comparison between test and numerical load–deflection relations of specimens with notch at midspan.
Figure9.
Figure 9. Comparison
Comparison between
between test
testand
andnumerical
numericalload–deflection
load–deflectionrelations
relationsof
ofspecimens
specimenswith
withnotch
notchatatmidspan.
midspan.
Figure 10 shows similar failure modes in tests and FE models (buckling in the com-
Figure
Figure
pression 10
10 shows
shows
flange). similar
similar failure
It is apparent from modes
failure modes
the in
in tests
figure tests and FE
FE models
andpredicted
that the models (buckling in
in the
(bucklingfailure
numerical com-
themode
com-
pression
pression flange).
flange). It
It is
is apparent
apparent from
from the
the figure
figure that
thatthe
the predicted
predicted numerical
numerical
was in agreement with the test failure mode. This similarity between the numerical and failure
failure mode
mode
was
wasin
test inagreement
failure modeswith
agreement with the
confirms test
the failure
test
the mode.
failure
accuracy ofThis
mode. similarity
This
the between
similarity
numerical the numerical
between
simulation the and
numerical
generated in test
and
this
failure modes confirms the accuracy of the numerical simulation generated
test failure modes confirms the accuracy of the numerical simulation generated in this
study. in this study.
study.
Figure 10. Cont.

69 16 of
16 of24
25
Principal strain distribution; yielding at end of notch and buckling of compression.
Figure 10. Similar failure modes in tests and FE models.

Figure 10. Similar failure modes in tests and FE models.
6. Application of the Verified FE Model to Beams with Notch Sizes not Considered in
6. Application of the Verified FE Model to Beams with Notch Sizes Not Considered in
Test Program
Test Program
Based on
Based on the
the verification
verification of of the
the FE
FE model
model in in the
the previous
previous Section
Section (Figures
(Figures 99 and
and 10),
10),
the numerical investigation was extended to cover further different
the numerical investigation was extended to cover further different notch depth and width notch depth and
width values,
values, with the with thedamage
notch notch damage being considered
being considered in the tension
in the tension side and side andatonly
only at the
the beam
beam midspan since this was the worst damage location scenario,
midspan since this was the worst damage location scenario, as revealed by the results as revealed by the re-
sults shown in Figures 5–7. The extended analysis examined
shown in Figures 5–7. The extended analysis examined beams with single notches (M-S) beams with single notches
(M-S)
and and double
double notches notches
(M-D)(M-D) at midspan
at midspan with 40 with
mm,40 60
mm, mm,60 and
mm,80 and
mm 80 in
mm in depth,
depth, and
20 mm, 40 mm, 60 mm, 100 mm, 150 mm, 200 mm, 250 mm, 300 mm, and 350 mm in mm
and 20 mm, 40 mm, 60 mm, 100 mm, 150 mm, 200 mm, 250 mm, 300 mm, and 350 width. in
width.Figure 11 displays the numerical load–midspan deflection relationships for the ex-
tendedFigure 11 displays
numerical the numerical
investigations. load–midspan
In comparison deflection
with the control relationships for the ex-
beam load-carrying
tended numerical investigations. In comparison with the control
capacity, the figure shows that the capacity of damaged steel specimens that have single beam load-carrying ca-
pacity, the figure shows that the capacity of damaged steel
notches at the midspan (M-S) was more significantly influenced and weakened by the specimens that have single
notches
notch at the
depth midspan
increase (40 (M-S)
mm, 60 wasmm, more
and significantly
80 mm) thaninfluenced
it was influencedand weakened by the
and weakened
notch
by thedepth
increaseincrease
in the(40 mm,width
notch 60 mm, (20andmm,8040mm) mm,than it was
60 mm, 100influenced
mm, 150 and mm,weakened
200 mm,
by the
250 mm,increase
300 mm, in the
andnotch widthThe
350 mm). (20 increase
mm, 40 mm, in the60notch
mm, 100 depthmm, 150 mm, 200
contributed mm, 250
directly to
mm, 300 mm,
decreasing and 350 of
the stiffness mm). The increase
the damaged beamsin as
themanifested
notch depth contributed
in the directlybeams
figure. Damaged to de-
creasing
with notch thedepths
stiffnessof of40the
mm damaged
(M-S-40)beamsshowed as manifested
a decrease in in the
the figure. Damaged
load-carrying beams
capacity
with notch
ranging from depths
1–8.5% ofas40opposed
mm (M-S-40) showedbeam
to the control a decrease
when the in the
notchload-carrying
width ranged capacity
from
ranging
20 mm tofrom 1–8.5%
350 mm. as opposed
Beams with notchto the control
depths of 60beam
mmwhen
(M-S-60)the notch
showed width rangedinfrom
a decrease the
20 mm to 350 mm.
load-carrying Beams
capacity with notch
ranging depthswhen
from 5–15% of 60 the
mmnotch
(M-S-60) widthshowed a decrease
stretched from 20inmm the
up to 350 mm.capacity
load-carrying Likewise, damaged
ranging from beams
5–15% with
whennotch
thedepth
notch of 80 mm
width (M-S-80)
stretched showed
from 20 mm a
decrease
up to 350inmm. the load-carrying
Likewise, damaged capacity ranging
beams with from 8–20%,
notch depth asof
opposed
80 mm to the control
(M-S-80) beama
showed
when the in
decrease notch
the width ranged from
load-carrying 20 mm
capacity to 350 from
ranging mm. 8–20%, as opposed to the control
beam when the notch width ranged from 20 mm to 350 mm.
The influence of increasing the notch depth on the load-carrying capacity was even
more pronounced and more distinct in damaged beams with double notches in the flange
at the midspan (M-D), as shown in Figure 11, where the increase in the notch width started
to considerably contribute to weakening the beam capacity as well. The increase in both
the notch depth and width in the M-D specimens contributed directly to substantially de-
creasing the stiffness and the capacity of the damaged steel specimens, as manifested in
Figure 9. Damaged beams with notch depths of 40 mm (M-D-40) showed a decrease in
load-carrying capacity ranging from 2.5–16%, as opposed to the control beam when the
notch width ranged from 20 mm to 350 mm. Beams with notch depths of 60 mm (M-D-60)
showed a decrease in the load-carrying capacity ranging from 7.5–32% when the notch
with the exception of M-D-80 beams, performed well in service load conditions with de-
flections less than the allowable deflection, the load capacity of the beams at the allowable
deflection was considerably influenced by the notch depth and width. As a result, all dam-
aged beams would not require any immediate action in service load conditions except
those of M-D-80, where immediate and urgent action would be needed, such as replacing
or retrofitting the beams. However, the decision for any needed repair in terms of urgency
depends on the load capacity of the damaged steel beams at the allowable deflection.
-500 -500
M–S–40 M–D–40
Allawable Deflection
-450 -450
-400 -400
-350 -350
Applied Load P (kN)

Applied Load P (kN)
-300 -300
-250 Service Load -250 Service Load
-200 -200
Control Beam Width 20 Control Beam Width 20
-150 Width 40 Width 60 -150 Width 40 Width 60
Width 200 Width 250 Width 200 Width 250
0 0
0 -5 -10 -15 0 -5 -10 -15
-500 -500
M–S–60 M–D–60
-450 -450
-400 -400
Applied Load P (kN)
-350 -350
Applied Load P (kN)
-300 -300
-250 Service Load
-250 Service Load
-200 -200
Control Beam Width 20 Control Beam Width 20
-100 Width 200 Width 250 -100
Width 200 Width 300
-50 Width 300 Width 350 -50 Width 350
0 0
0 -5 -10 -15 0 -5 -10 -15
-500 -500
M–S–80 M–D–80
-450 -450
-400 -400
-350 -350
Applied Load P (kN)
Applied Load P (kN)
-300 -300
-250 Service Load -250 Service Load
-200 -200
-150 Control Beam Width 20 -150 Control Beam Width 20
-100 width 100 Width 150 -100 Width 100 Width 150
-50 Width 300 Width 350 -50
Width 300 Width 350
0 0
0 -5 -10 -15 0 -5 -10 -15
Figure 11.
Figure 11. Numerical
Numerical applied
applied load–midspan
load–midspan deflection
deflection relations.
relations.
Figure
The 11 indicates
influence that when
of increasing thethe depth
notch of the
depth onnotch increased, thecapacity
the load-carrying load-carrying
was evenca-
pacitypronounced
more of the beams andwas significantly
more distinct inaffected.
damagedThis canwith
beams be explained by thein
double notches observation
the flange
that
at thewhen the notch
midspan (M-D),depth increased,
as shown the 11,
in Figure notch damage
where got closer
the increase to notch
in the the beam
widthweb and,
started
therefore,
to the stress
considerably concentration
contribute at the end
to weakening theof the notch
beam became
capacity within
as well. Thethe vicinity
increase inof the
both
beam web, leading to an increase in the web strains and stresses. The increase in the web
strain and stress led to a decrease in the beam stiffness and capacity and to an increase in
the defection rate. This analysis is clearly confirmed by Figures 12 and 13, where for M-S
and M-D beams with notch depths of 80 mm, the stresses and strains were concentrated
the notch depth and width in the M-D specimens contributed directly to substantially
decreasing the stiffness and the capacity of the damaged steel specimens, as manifested
in Figure 9. Damaged beams with notch depths of 40 mm (M-D-40) showed a decrease in
load-carrying capacity ranging from 2.5–16%, as opposed to the control beam when the
notch width ranged from 20 mm to 350 mm. Beams with notch depths of 60 mm (M-D-60)
showed a decrease in the load-carrying capacity ranging from 7.5–32% when the notch
width stretched from 20 mm up to 350 mm. Likewise, damaged beams with notch depths
of 80 mm (M-D-80) showed a decrease in the load-carrying capacity ranging from 14–48%,
as opposed to the control beam when the notch width ranged from 20 mm to 350 mm. One
observation from Figure 11 worthy of highlighting is that when the notch width in all the
numerically examined damaged beams was equal to half the distance between the applied
loads, i.e., 200 mm or more, the notch width influence on the capacity of the beams became
insignificant.
To relate the above findings to practical everyday applications, the service load was
assumed to be equal to half the maximum applied load, i.e., 455 kN/2 = 228 kN. For
steel beams, deflection due to service live load is of concern in the finished structures and,
therefore, the allowable deflection (i.e., deflection limit) considered in the discussion of the
numerical result findings was set to be equal to δ = L/360 = 1800 mm/360 = 5 mm, where
L is the beam span length. Hence, horizontal and vertical lines showing the level of the
service load and the allowable deflection, respectively, were drawn in Figure 10. With these
two lines established, it is clear from the figure that although all damaged beams, with the
exception of M-D-80 beams, performed well in service load conditions with deflections less
than the allowable deflection, the load capacity of the beams at the allowable deflection
was considerably influenced by the notch depth and width. As a result, all damaged beams
would not require any immediate action in service load conditions except those of M-D-80,
where immediate and urgent action would be needed, such as replacing or retrofitting the
beams. However, the decision for any needed repair in terms of urgency depends on the
load capacity of the damaged steel beams at the allowable deflection.
Figure 11 indicates that when the depth of the notch increased, the load-carrying
capacity of the beams was significantly affected. This can be explained by the observation
that when the notch depth increased, the notch damage got closer to the beam web and,
therefore, the stress concentration at the end of the notch became within the vicinity of the
beam web, leading to an increase in the web strains and stresses. The increase in the web
strain and stress led to a decrease in the beam stiffness and capacity and to an increase in
the defection rate. This analysis is clearly confirmed by Figures 12 and 13, where for M-S
and M-D beams with notch depths of 80 mm, the stresses and strains were concentrated
and high not only in the beam flange but also in the beam web, due to the nearby notch.
On the other hand, as the width of the notch increased, these high stresses and strains
spread, affecting a larger area of the web, as revealed in Figures 12 and 13, and caused a
wider portion of the beam to yield. As a result, the stiffness of the beam decreased, the
beam capacity decreased, and the beam deflection rate increased rapidly, which accelerated
the occurrence of permanent deformations. The figures show that the damage affected
the stiffness of the beams, resulting in an increase in deflections at design loads. This
impact of the damage on the steel beam deflection should be addressed in design codes for
serviceability requirements.
M−S, width 60mm, depth 80mm. M−D, width 60mm, depth 80mm.
M−S, width 200 depth 80mm. M−D, width 200mm, depth 80mm.
M−S, width 200 depth 80mm. M−D, width 200mm, depth 80mm.
Figure 12.
Figure 12. Stress
Stress distribution
distribution with
with the
the middle
middlethird
thirdsegment
segmentof
ofthe
thebeams.
beams.
Figure 12. Stress distribution with the middle third segment of the beams.
Figure 13. Cont.

Figure 13.
Figure 13. Strain
Strain distribution
distribution with
with the
the middle
middle third
third segment
segmentof
ofthe
thebeams.
beams.
numerical stiffness
The numerical stiffness versus
versus numerical
numerical midspan
midspan deflection
deflection relation
relation ofof the
the control
control
represented in
beam was represented in Figure
Figure 14,
14, and
and numerical
numerical relative
relative stiffness
stiffness of
of the
the damaged
damaged steel
steel
Infrastructures 2021, 6, 69 beams with respect
respecttotothe
thecontrol
controlbeam
beamnumerical stiffness
numerical versus
stiffness numerical
versus numericalmidspan21 of de-
midspan 25
flections were
deflections drawn
were drawn in in
Figure 15.15.
Figure
Figure 14 of the control beam identifies the start of the beam yielding, which occurred
at 120
a deflection a little less than the allowable deflection of 5 mm, as shown in the figure.
The figure also identified the start of the plastic zone/behavior of the beam. It is evident
100Figure 14 that the stiffness dropped rapidly and immediately after the yielding was
from
reached, causing the beam to undergo inelastic deformation until the stiffness became as-
80
ymptotic, indicating stiffness softening or plastic behavior with no increase in the capac-
ity. The beginning of the plasticInelastic Zone
zone occurred at a deflection equal to almost three times
Stiffness
60 as the yielding deflection, as shown in the figure.

as much
Elastic Zone
40 Plastic Zone
20
0
0 -5 -10 -15
Figure
Figure14.
14.Numerical
Numericalstiffness–midspan
stiffness–midspandeflection
deflectionrelation
relationof
ofthe
thecontrol
controlbeam.
beam.
The numerical relative stiffness of the damaged steel beams versus the numerical
midspan deflections are drawn in Figure 15. The figure demonstrates that the yielding of
the M-S beams was reached faster as the depth of the notch increased while increasing the
width of the notch did not sizably affect the beginning of the plastic behavior. On the other
results highlight the applicability of using the finite element technique to investigate the
maximum acceptable damage size with respect to the reuse of any damaged beam based
on its configuration.
Therefore, this investigation supports the feasibility of reusing pre-damaged steel
Infrastructures 2021, 6, 69 beams in temporary structures under service loads and certain levels of damage, 21 where
of 24
the behavior of such beams is within the elastic range and the maximum defection is less
than the allowable one.
120 120
M–S–40 Control Beam M–D–40 Control Beam
Width 20 Width 20
100 100
Width 40 Width 40
Initial stiffness (%)

Width 60 Width 60
80 80
Width 100 Width 100
Width 150 Width 150
60 60
Width 200 Width 200
Width 250 Width 250
40 40
Width 300 Width 300
Width 350 Width 350
20 20
Infrastructures 2021, 6, 69
0 0 22 of 25
0 -5 -10 -15 0 -5 -10 -15
120 120
M–S–60 Control Beam
M–D–60 Control Beam
100 Width 20 100 Width 20
Width 40 Width 40
Width 60 Width 60
80 80
Width 100
Width 100
Width 150
60 60 Width 150
Width 200
Width 250 Width 200
Width 300 40 Width 300
40
Width 350 Width 350
20 20
0 0
0 -5 -10 -15 0 -5 -10 -15
120 120
M–S–80 Control Beam M–D–80 Control Beam
100 Width 20 100 Width 20
Width 40 Width 40
Width 60 Width 60
80 80
width 100 Width 100
Width 150 Width 150
60 60 Width 200
Width 200
Width 250 Width 250
40 Width 300
40 Width 300
Width 350
Width 350
20 20
0 0
0 -5 -10 -15 0 -5 -10 -15
Figure 15. Numerical stiffness–midspan deflection relations of control and damaged beams.
Figure 15. Numerical stiffness–midspan deflection relations of control and damaged beams.
Figure 14 of the control beam identifies the start of the beam yielding, which occurred
7.
at Conclusions
a deflection a little less than the allowable deflection of 5 mm, as shown in the figure. The
figure also
The identifiedofthe
feasibility start of
reusing the plastic zone/behavior
pre-damaged steel beams inof the beam.structures
temporary It is evident from
such as
Figure
the 14 that the
excavation stiffness dropped
of temporary rapidly
supporting and immediately
systems, and waler after
beams,thewas
yielding was reached,
addressed in this
causingThe
study. the extensive
beam to undergo inelastic
structural deformation
investigation until the stiffness
of notch-damaged becameunsupported
laterally asymptotic,
steel beams with various sizes of U-shaped notch damage in the tension flange was carried
out experimentally and numerically. Notch damages on one side and on two sides of the
tension flange were considered in the study. The load-carrying capacity of the majority of
the investigated damaged beams was not affected by the deflection while in service load
indicating stiffness softening or plastic behavior with no increase in the capacity. The
beginning of the plastic zone occurred at a deflection equal to almost three times as much
as the yielding deflection, as shown in the figure.
The numerical relative stiffness of the damaged steel beams versus the numerical
midspan deflections are drawn in Figure 15. The figure demonstrates that the yielding of
the M-S beams was reached faster as the depth of the notch increased while increasing
the width of the notch did not sizably affect the beginning of the plastic behavior. On the
other hand, M-D beams demonstrated that a quicker yielding of the beam was achieved at
small deflection values as the depth of the notch increased. In addition, as the width of the
notch increased, the softening of the beam (plastic behavior) was reached at much smaller
deflections than those in the M-S beams. This is due to the spreading of the yielding stress
over a larger web area, as confirmed by the stress and strains distributions in Figures 12
and 13. Figure 15 reveals an interesting observation and that is the curves of the damaged
beams are shifted to the left as if the beams were already loaded and deformed. As a result,
the beams reached the allowable deformation and plastic behavior at lower loads. The
results highlight the applicability of using the finite element technique to investigate the
maximum acceptable damage size with respect to the reuse of any damaged beam based
on its configuration.
Therefore, this investigation supports the feasibility of reusing pre-damaged steel
beams in temporary structures under service loads and certain levels of damage, where the
behavior of such beams is within the elastic range and the maximum defection is less than
the allowable one.
7. Conclusions
The feasibility of reusing pre-damaged steel beams in temporary structures such as
the excavation of temporary supporting systems, and waler beams, was addressed in this
study. The extensive structural investigation of notch-damaged laterally unsupported steel
beams with various sizes of U-shaped notch damage in the tension flange was carried
out experimentally and numerically. Notch damages on one side and on two sides of the
tension flange were considered in the study. The load-carrying capacity of the majority of
the investigated damaged beams was not affected by the deflection while in service load
conditions. In contrast, when the deflection of the damaged beams exceeded the allowable
deflection limit, the capacity reduced, and the damaged beams failed due to buckling of
the compression flange. It was observed in double notched beams with notch width depths
of 80 mm and widths excessing 40 mm the beams behaved inelastically under allowable
service loads, which jeopardizes the feasibility of further reusing the damaged beam.
The beams with notch damage at midspan were the most critical ones because the
notch existed within the beam maximum flexural moment zone. The load-carrying capacity
of the considered damaged beams with a single notch at the midspan was more significantly
influenced by the notch depth increase than it was influenced by the increase in the notch
width. On the other hand, the influence of increasing the notch depth on the load-carrying
capacity was even more pronounced and more distinct in beams with double notches in
the flange at the midspan of the beam, where the increase in the notch width considerably
contributed to reducing the beam capacity as well. This can be explained by the observation
that when the notch depth increased, the notch damage got closer to the beam web and,
therefore, the stress concentration at the end of the notch became within the vicinity of the
beam web and penetrated it, leading to an increase in the web strains and stresses. The
increase in the web strains and stresses led to a decrease in the beam stiffness and load
capacity and to an increase in the defection rate. Likewise, the increase in the notch width
on one side or both sides of the flange allowed the spreading of the stresses and strains
into a larger area of the beam web leading to a further reduction in the load capacity and
larger deflections.
This study revealed that all damaged beams, with the exception of M-D-80 beams,
performed elastically in service load conditions when their deflections were less than the
allowable limit. The load capacity of the beams at the allowable deflection was consid-
erably influenced by the notch depth and width. However, for the M-D-80 beams, the
load capacities were below the service load limit due to the fact that as the notch depth
increased, the effect of increasing its width on the load capacity became more noticeable
and evident, leading to a significant decrease in the beam load capacity at the allowable
deflection. Therefore, the decision for any needed repair of U-shaped-notch-damaged
laterally unsupported beams in terms of urgency and repair method depends on how close
the notch depth is to the beam web and on the load capacity of the damaged beams at the
allowable deflection. This investigation supports the feasibility of reusing pre-damaged
beams in temporary structures under service loads and certain levels of damage, where the
behavior of such beams is within the elastic range and the maximum defection is less than
the allowable one. The FE can be used to highlight the upper limit of the notch size that
allows the damaged beam to be reused without repair.
It is recommended, for future study, to generate reduction factors for the strength
and stiffness of the damaged steel beams based on the damage size and different beam
configurations. These reduction factors could be presented either as charts or equations
to facilitate the design of the damaged beams and explore the possibility of reusing them
or not. Additionally, damaged beams with notches in both flanges (or a notch in the
compression flange) should be investigated to examine the feasibility of reusing them in
temporary structures.
Author Contributions: Conceptualization, B.E.-A. and S.E.; methodology, B.E.-A. and S.E.; software,
S.E.; validation, S.E.; formal analysis, B.E.-A. and S.E.; investigation, B.E.-A. and S.E.; resources,
B.E.-A. and S.E.; data curation, S.E.; writing—original draft preparation, B.E.-A.; writing—review and
editing, B.E.-A.; visualization, B.E.-A. and S.E.; supervision, B.E.-A. and S.E.; project administration,
B.E.-A. and S.E.; funding acquisition, B.E.-A. and S.E. All authors have read and agreed to the
published version of the manuscript.
Funding: This research was funded by the United Arab Emirates University, grant numbers
G00002738 (31N345) and 31N349.
Data Availability Statement: The data that support the findings of this study are available from the
authors upon reasonable request.
Acknowledgments: The authors also would like to acknowledge the lab technicians and to acknowl-
edge Zakaria El Karaouf for his contribution in the FE simulations.
Conflicts of Interest: The authors declare no conflict of interest.
References
1. Hmidan, A.; Kim, Y.J.; Yazdani, S. CFRP repair of steel beams with various initial crack configurations. J. Compos. Constr. 2011, 15,
952–962. [CrossRef]
2. Shaat, A.; Fam, A. Axial loading tests on short and long hollow structural steel columns retrofitted using carbon fibre reinforced
polymers. Can. J. Civ. Eng. 2006, 33, 458–470. [CrossRef]
3. Shaat, A.; Fam, A. Repair of cracked steel girders connected to concrete slabs using carbon-fiber-reinforced polymer sheets. J.
Compos. Constr. 2008, 12, 650–659. [CrossRef]
4. Shaat, A.; Fam, A.Z. Slender steel columns strengthened using high-modulus CFRP plates for buckling control. J. Compos. Constr.
2009, 13, 2–12. [CrossRef]
5. Nozaka, K.; Shield, C.K.; Hajjar, J.F. Effective bond length of carbon-fiber-reinforced polymer strips bonded to fatigued steel
bridge I-girders. J. Bridge Eng. 2005, 10, 195–205. [CrossRef]
6. Deng, J.; Jia, Y.; Zheng, H. Theoretical and experimental study on notched steel beams strengthened with CFRP plate. Compos.
Struct. 2016, 136, 450–459. [CrossRef]
7. Deng, J.; Li, J.; Wang, Y.; Xie, W. Numerical study on notched steel beams strengthened by CFRP plates. Constr. Build. Mater. 2018,
163, 622–633. [CrossRef]
8. Siwowski, T.W.; Siwowska, P. Experimental study on CFRP-strengthened steel beams. Compos. Part B Eng. 2018, 149, 12–21.
[CrossRef]
9. Siwowska, P.; Siwowski, T. Strength evaluation of steel beams prestressed with CFRP plates. In Proceedings of the Eurosteel 2017,
Copenhagen, Denmark, 13–15 September 2017.
10. Elchalakani, M. Rehabilitation of corroded steel CHS under combined bending and bearing using CFRP. J. Constr. Steel Res. 2016,
125, 26–42. [CrossRef]
11. Chen, T.; Wang, X.; Qi, M. Fatigue improvements of cracked rectangular hollow section steel beams strengthened with CFRP
plates. Thin-Walled Struct. 2018, 122, 371–377. [CrossRef]
12. Ochi, N.; Matsumura, M.; Hisabe, N. Experimental study on strengthening effect of high modulus CFRP strips with different
adhesive length installed onto the lower flange plate of I shaped steel girder. Procedia Eng. 2011, 14, 506–512. [CrossRef]
13. Ghafoori, E.; Motavalli, M. Innovative CFRP-prestressing system for strengthening metallic structures. J. Compos. Constr. 2015, 19,
04015006. [CrossRef]
14. Li, J.; Wang, Y.; Deng, J.; Jia, Y. Experimental study on the flexural behaviour of notched steel beams strengthened by prestressed
CFRP plate with an end plate anchorage system. Eng. Struct. 2018, 171, 29–39. [CrossRef]
15. Zhou, H.; Attard, T.L.; Wang, Y.; Wang, J.A.; Ren, F. Rehabilitation of notch damaged steel beams using a carbon fiber reinforced
hybrid polymeric-matrix composite. Compos. Struct. 2013, 106, 690–702. [CrossRef]
16. Hmidan, A.; Kim, Y.J.; Yazdani, S. Crack-dependent response of steel elements strengthened with CFRP sheets. Constr. Build.
Mater. 2013, 49, 110–120. [CrossRef]
17. Chiew, S.; Yu, Y.; Lee, C. Bond failure of steel beams strengthened with FRP laminates—Part 1: Model development. Compos. Part
B Eng. 2011, 42, 1114–1121. [CrossRef]
18. Tavakkolizadeh, M.; Saadatmanesh, H. Strengthening of steel-concrete composite girders using carbon fiber reinforced polymers
sheets. J. Struct. Eng. 2003, 129, 30–40. [CrossRef]
19. Tavakkolizadeh, M.; Saadatmanesh, H. Repair of damaged steel-concrete composite girders using carbon fiber-reinforced polymer
sheets. J. Compos. Constr. 2003, 7, 311–322. [CrossRef]
20. Photiou, N.; Hollaway, L.; Chryssanthopoulos, M.K. Strengthening of an artificially degraded steel beam utilising a carbon/glass
composite system. In Advanced Polymer Composites for Structural Applications in Construction; Woodhead Publishing: Cambridge,
UK, 2004; pp. 274–283.
21. Wang, H.-T.; Wu, G. Crack propagation prediction of double-edged cracked steel beams strengthened with FRP plates. Thin-Walled
Struct. 2018, 127, 459–468. [CrossRef]
22. Colombi, P.; Fava, G. Experimental study on the fatigue behaviour of cracked steel beams repaired with CFRP plates. Eng. Fract.
Mech. 2015, 145, 128–142. [CrossRef]
23. Kim, Y.J.; Brunell, G. Interaction between CFRP-repair and initial damage of wide-flange steel beams subjected to three-point
bending. Compos. Struct. 2011, 93, 1986–1996. [CrossRef]
24. Kim, Y.J.; Harries, K.A. Predictive response of notched steel beams repaired with CFRP strips including bond-slip behavior. Int. J.
Struct. Stab. Dyn. 2012, 12, 1–21. [CrossRef]
25. Kadhim, M.M.A. Effect of CFRP plate length strengthening continuous steel beam. Constr. Build. Mater. 2012, 28, 648–652.
[CrossRef]
26. Haeri, H.; Sarfarazi, V.; Zhu, Z.; Hedayat, A.; Nezamabadi, M.F.; Karbala, M. Simulation of crack initiation and propagation in
three point bending test using PFC2D. Struct. Eng. Mech. 2018, 66, 453–463. [CrossRef]
27. Yu, Y.; Chiew, S.; Lee, C. Bond failure of steel beams strengthened with FRP laminates—Part 2: Verification. Compos. Part B Eng.
2011, 42, 1122–1134. [CrossRef]
28. Abambres, M.; Arruda, M.R. Finite element analysis of steel structures—A review of useful guidelines. Int. J. Struct. Integr. 2016.
[CrossRef]
29. Ulger, T.; Okeil, A.M. Analysis of thin-walled steel beams retrofitted by bonding GFRP stiffeners: Numerical model and
investigation of design parameters. Eng. Struct. 2017, 153, 166–179. [CrossRef]
30. Goncalves, R. An assessment of the lateral-torsional buckling and post-buckling behaviour of steel I-section beams using a
geometrically exact beam finite element. Thin-Walled Struct. 2019, 143, 106222. [CrossRef]
31. Mohammed, A.A.; Manalo, A.C.; Ferdous, W.; Zhuge, Y.; Vijay, P.V.; Alkinani, A.Q.; Fam, A. State-of-the-art of prefabricated FRP
composite jackets for structural repair. Eng. Sci. Technol. Int. J. 2020, 23, 1244–1258. [CrossRef]
32. Ferdous, W.; Manalo, A.; AlAjarmeh, O.; Mohammed, A.A.; Salih, C.; Yu, P.; Khotbehsara, M.M.; Schubel, P. Static behaviour of
glass fibre reinforced novel composite sleepers for mainline railway track. Eng. Struct. 2021, 229, 111627. [CrossRef]
33. Tingley, D.D.; Cooper, S.; Cullen, J. Understanding and overcoming the barriers to structural steel reuse, a UK perspective. J.
Clean. Prod. 2017, 148, 642–652. [CrossRef]
34. Pongiglione, M.; Calderini, C. Material savings through structural steel reuse: A case study in Genoa. Resour. Conserv. Recycl.
2014, 86, 87–92. [CrossRef]
35. Cervenka, V.; Jendele, L. ATENA Program Documentation—Part 1: Theory, Cervenka Consulting; Červenka Consulting s.r.o.: Prague,
Czech Republic, 2012.
36. Mansour, M.; El-Ariss, B.; El-Maaddawy, T. Effect of Support Conditions on Performance of Continuous Reinforced Concrete
Deep Beams. Buildings 2020, 10, 212. [CrossRef]
37. Alkhalil, J.; El-Maaddawy, T. Finite element modelling and testing of two-span concrete slab strips strengthened by externally-
bonded composites and mechanical anchors. Eng. Struct. 2017, 147, 45–61. [CrossRef]

Infrastructures: Feasibility of Reusing Damaged Steel Beams in Temporary Structures

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Infrastructures: Feasibility of Reusing Damaged Steel Beams in Temporary Structures

Uploaded by

Copyright:

Available Formats

infrastructures

Infrastructures 2021, 6, 69. https://doi.org/10.3390/infrastructures6050069 https://www.mdpi.com/journal/infrastructures

Table 2. Test matrix—specimen categorizing.

Set (Location of Notch) * Notch Depth (mm) Specimen **

2.3. Material Properties

2.4. Instrumentation of Test Setup

Stain gauges LVDTs

3.2. Applied Load versus Mid-Span Deflection

Applied load P(kN)

3.3. Applied Load versus Strain

4.3. Analysis Method

Figure 10. Cont.

Principal strain distribution; yielding at end of notch and buckling of compression.

Figure 10. Similar failure modes in tests and FE models.

Applied Load P (kN)

Figure 13. Cont.

60 as the yielding deflection, as shown in the figure.

Initial stiffness (%)

You might also like