Sensors: SHM System and A FEM Model-Based Force Analysis Assessment in Stay Cables

sensors
Article
SHM System and a FEM Model-Based Force Analysis
Assessment in Stay Cables
Jan Biliszczuk , Paweł Hawryszków * and Marco Teichgraeber *
Faculty of Civil Engineering, Wrocław University of Science and Technology, Wybrzeże Wyspiańskiego 27,
50-370 Wrocław, Poland; jan.biliszczuk@pwr.edu.pl
* Correspondence: pawel.hawryszkow@pwr.edu.pl (P.H.); marco.teichgraeber@pwr.edu.pl (M.T.);
Tel.: +48-71-320-45-62 (P.H.); +48-661-120-555 (M.T.)
Abstract: The R˛edziński Bridge in Wrocław is the biggest Polish concrete cable-stayed bridge. It is
equipped with a large structural health monitoring (SHM) system which has been collecting the
measured data since the bridge opening in the year 2011. This paper presents a comparison between
the measured data and the finite element method (FEM) calculations, while taking into account
7 years of data collection and analyses. The first part of this paper concerns the SHM application.
In the next part, which contains comparisons between forces in cables and temperature changes
throughout the structure, the measured data are presented. The third part includes SHM-based
calculations and simulations with a complex FEM model to check the measured data and to estimate
future measurements. The last part contains a durability assessment calculation for the cable stays.
Keywords: bridges; stay cables; durability; maintenance; monitoring; SHM; FEM

Citation: Biliszczuk, J.; Hawryszków,

P.; Teichgraeber, M. SHM System and 1. Introduction
a FEM Model-Based Force Analysis The structural health monitoring system (SHM) is a relatively new tool for examining
Assessment in Stay Cables. Sensors the technical condition of structures. A wide description of the continuous observation
2021, 21, 1927. https://doi.org/
methods of structures in recent years can be found, e.g., in [1]. The development over the
10.3390/s21061927
past decades has been focused mainly on fulfilling the expectations of the maintenance
administration to constantly control the structures’ behaviour. Such systems provide the
Academic Editor: Saeed
user with a range of information that is only compared with the design values. However,
Eftekhar Azam
even the most complex systems can still provide incorrect or missing information. For this
purpose, the collected data should be properly verified and checked, for example, with
Received: 26 December 2020
Accepted: 18 February 2021
the use of adequate numerical models of the monitored structure. In this paper, such a
Published: 10 March 2021
solution is presented for the longest concrete cable-stayed bridge in Poland, the R˛edziński
Bridge in Wrocław. Here, as in other cable-stayed bridges, the most important load-bearing
Publisher’s Note: MDPI stays neutral
element are the stays and their tensioning force as the effect of all loads acting on the
with regard to jurisdictional claims in
whole structure. A simple tension force analysis from the SHM system does not deliver
published maps and institutional affil- the possibility to determine which load (vehicle movement, temperature changes, wind
iations. impact, etc.) affects its value or fluctuates the most. For fatigue or reliability calculations,
a detailed model of each stay cable should be created. Advanced studies on the work of
anchorages and mutual friction of the wires in the strand can be found in [2–5]. However,
they are based on purely laboratory experiments.
Copyright: © 2021 by the authors.
Most of the studies related to SHM, which were cited in this work, present the general
Licensee MDPI, Basel, Switzerland.
methods of these system operations. The aim of the authors was to show how to use the
This article is an open access article
collected data for scientific purposes, which is also an extension of their own research
distributed under the terms and presented in [6], where the SHM system of the R˛edziński Bridge and measured quantities
conditions of the Creative Commons were described, but without statistical analysis. The idea of the separation of individual
Attribution (CC BY) license (https:// load effects is presented based on data from the SHM monitoring system and FEM models
creativecommons.org/licenses/by/ of the structure. The methods of validation of these data and the possibilities of their
4.0/). further statistical analysis are also shown. An example of their use for fatigue calculations
Sensors 2021, 21, 1927. https://doi.org/10.3390/s21061927 https://www.mdpi.com/journal/sensors

Sensors2021,
Sensors 2021,21,
21,xxFOR
FORPEER
PEERREVIEW
REVIEW 22ofof29
29
Sensors 2021, 21, 1927 2 of 27
modelsof
models ofthe
thestructure.
structure.The
Themethods
methodsof ofvalidation
validationof ofthese
thesedata
dataand
andthe
thepossibilities
possibilitiesof
of
their further statistical analysis are also shown. An example of their use
their further statistical analysis are also shown. An example of their use for fatigue cal- for fatigue cal-
has been proposed,
culations
culations hasbeen
has as well. Theas
beenproposed,
proposed, tool
as consisting
well.
well. toolofconsisting
Thetool
The the big SHM
consisting system
ofthe
of the and
SHMthe
bigSHM
big FEM and
system
system models
andthethe
shows
FEM how the
FEMmodels
models collected
shows
shows howdata
how the can be used
thecollected
collected data
data for simulations
can
can beused
be usedfor of simulations
for future results.
simulations ofMoreover,
of the
futureresults.
future results.
analysis
Moreover,performed
Moreover, the in this
theanalysis
analysis way can provide
performed
performed inthis
in thiswayvaluable
way information
canprovide
can provide on the
valuable
valuable design of further
information
information onthe
on the
monitoring
design of systems
further for cable-stayed
monitoring systems bridges.
for cable-stayed
design of further monitoring systems for cable-stayed bridges. bridges.
2.2.The R˛edziński Bridge

TheRędziński
Rędziński and
Bridgeand its
anditsitsSHM
SHMSystem
System
2. The Bridge SHM System
The R˛e dzi ński
TheRędziński Bridge
RędzińskiBridge was
Bridgewas built
wasbuilt in
builtin 2011
in2011 and
2011and ititisisthe
andit main
themain bridge
mainbridge within
bridgewithin the
withinthe Wrocław
theWrocław
Wrocław
The is the
motorway ring-road.
motorwayring-road.
ring-road.TheThe structure
Thestructure is a concrete
structureisisaaconcrete cable-stayed
concretecable-stayed bridge
cable-stayedbridge with
bridgewith the
withthe spans
spansof
thespans of
motorway of
49.00 m + 256.00 m + 256.00 m + 49.00 m. The H-shaped pylon is 122.00 m high (Figure 1)
49.00m
49.00 m++256.00
256.00m m++256.00
256.00mm++49.00
49.00m. m.The
TheH-shaped
H-shapedpylon pylonisis122.00
122.00mmhigh
high(Figure
(Figure1)1)
measured from the foundation top surface.
measuredfrom
measured fromthe thefoundation
foundationtop topsurface.
surface.
Figure1.
Figure
Figure 1.1.Aerial
Aerialview
Aerial viewof
view ofofthe
theR˛
the Rędziński
Rędziński Bridge(photo:
Bridge
edziński Bridge (photo:Władysław
(photo: WładysławKluczewski).
Władysław Kluczewski).
Kluczewski).
Thecharacteristic
The characteristicfeature
characteristic feature
feature of
ofof
the the
the bridge
bridge
bridge isistwo
is two two separate
separate
separate boxgirder
boxbox
girdergirder concrete
concrete
concrete decks. decks.
decks.
Due
Due
to
Due to
theto thePolish
Polish
the Polish
law at law
theatat
law timetheof
the time
theof
time of thedesign
design
the design andconstruction
construction
and construction
and of
of theof thebridge,
bridge,
the bridge, aaseparate
a separateseparate
deck
deckstructure
deck structure
structure had
had tohad tobe
beinstalled
be installed
to installed
underunderunder
each eachroadway.
roadway.
each roadway.
The decks Theare
The decks
18.57
decks are
are 18.57
m18.57
widem mandwide and
2.50and
wide m
2.50m
high,
2.50 m high,suspended
suspended
high, suspended
to one pylon
totoone
one pylon
with
pylon with
160with 160stay
stay160
cables stay cables2).
(Figure
cables (Figure
The stay
(Figure 2).The
2). Thestay
stay
cables are cables
divided
cables are
are
divided
into
divided into
fourinto
equalfour
four equalThe
planes.
equal planes. Thetheoretical
theoretical
planes. The theoretical
cable lengthscablerange
cable lengths
lengthsfromrange
67.39
range from
m to
from 67.39
267.34
67.39 m mto
to267.34
m. 267.34
Each
m.
cableEach
m. Each cable
consists
cable ofconsists of
seven-wire
consists seven-wire
strands, which
of seven-wire strands, which
arewhich
strands, are
installed installed
are from from
24 tofrom
installed 24 to
48 strands 48 strands
24 to 48instrands
a bundle in
in
aabundle
bundledepending
depending depending
on the stayonon thestay
cable.
the staycable.
cable.
Figure2.2.Main
Figure Main propertiesofofthe
theRędziński
RędzińskiBridge
Bridge[7].
[7].
Figure 2. Main properties
properties of the R˛
edziński Bridge [7].
Sensors 2021, 21, x FOR PEER REVIEW

Sensors 2021, 21, 1927 3 of 27
Detailed stay cable parameters are presented in Table 1. The size and
Detailed stay cable parameters are presented in Table 1. The size and outsta
Detailed stay cable parameters are presented in Table 1. The size and outstanding
structure of the Rędziński Bridge were the reasons why it was equipped wit
structure of the Rędziński Bridge were the reasons why it was equipped with 222
structure of thetural health monitoring (SHM) system sensors (Figure 3). A detailed description
R˛edzitural health monitoring (SHM) system sensors (Figure 3). A detailed descri
ński Bridge were the reasons why it was equipped with 222 structural
health monitoring (SHM) system sensors (Figure 3). A detailed description of the design
design and construction of the bridge can be found in the study [7].
design and construction of the bridge can be found in the study [7].
and construction of the bridge can be found in the study [7].
Table 1. Dimensions of stay cables and the number of strands installed in them.

Horizontal Vertical Total
Number of
Stay Cable Number Length Length Length
Strands Horizontal
Horizontal
Number of
Number of Lh [m] Lv Vertical
Vertical
[m] Total Total
L [m]
Stay Cable Number
Stay Cable Number Length
Length Length
Length
Length
Length

1 21 Strands
24
Strands 19.06 64.64 67.39
2 22 28 L
Lh [m] h [m]
30.51 Lv [m] Lv [m]
67.58 L [m]
L [m]
74.15
3 23 28 42.33 69.94 81.75
1 41 21 24 21 24 3024 19.06 19.06
54.29 64.64 64.64
72.03 67.39 67.39
90.20
2 52 22 25 22 28 3228 66.30
30.51 30.51 74.07
67.58 67.58 99.41
74.15 74.15
6 26 34 78.31 76.01 109.14
3 3 23 23 28 28 42.33 42.33 69.94 69.94 81.75 81.75
7 27 38 90.37 77.95 119.35
4 84 24 28 24 30 4030 54.29 54.29
102.41 72.03 72.03 129.86
79.85 90.20 90.20
5 95 25 29 25 32 4632 66.30
114.45
66.30 74.07
81.75
74.07 99.41 99.41
140.65
106
6 26 30 26 34 4834 126.51
78.31 78.31 83.62
76.01 76.01 151.64
109.14 109.14
11 31 48 138.54 85.52 162.81
7 127 27 32 27 38 4838 90.37 90.37
150.57 77.95 77.95 174.10
87.41 119.35 119.35
8 138 28 33 28 40 4840 102.41 102.41
162.61 79.85 79.85 185.51
89.29 129.86 129.86
9 149 29 34 29 46 4846 174.59
114.45 114.45 91.16
81.75 81.75 196.96
140.65 140.65
15 35 48 186.63 93.06 208.55
10 10
16 30 36 30 48 4848 126.51 126.51
198.66 83.62 83.62
94.97 151.64 151.64
220.20
11 11
17 31 37 31 48 4848 138.54 138.54
210.70 85.52 85.52 231.91
96.90 162.81 162.81
12 18
12 32 38 32 48 4848 222.75
150.57 150.57 98.84
87.41 87.41 243.70
174.10 174.10
19 39 48 234.79 100.77 255.50
13 13
20 33 40 33 48 40
48 162.61 162.61
246.82 89.29 89.29
102.71 185.51 185.51
267.34
14 14 34 34 48 48 174.59 174.59 91.16 91.16 196.96 196.96
15 15 35 35 48 48 186.63 186.63 93.06 93.06 208.55 208.55
The installed sensors are measuring the following values: stress, temperature, acceler-
16 displacements
16 and
ations 36 36 [6]. The 48 48
connection the198.66
198.66
with 94.97 94.97
system is allowed 220.20 220.20
by an Internet
17 17
browser 37 that37
application provides an 48 of each
48 overview 210.70 210.70 It generates
sensor. 96.90 96.90 notifi-231.91
alerts and 231.91
18 18
cations 38 are
if some sensors 38 out of order. 48
48 Furthermore, 222.75 222.75
the application98.84
makes98.84 to243.70
243.70
it possible
create
19 diagrams
19 of39
measured
39 values48 and 48
to export234.79
them234.79
as .csv files100.77
which are compatible
100.77 255.50 255.50
with
20 calculation
20 programs.
40 40 After 7 years,
40 a
40 database of the
246.82 measured
246.82 values
102.71 was
102.71 compiled,
267.34 267.34
and it was compared with calculations conducted based on the bridge FEM model.
In case of the The
R˛edzi ńskiinstalled
The Bridge
installed SHM are system,
sensors
sensors mostlythe vibrating
are measuring
measuring wire
the following
following strain
values: gages
values: stress,
stress, temp
temperatu
were used. Measurement of all static quantities is performed at the same time,
celerations and displacements [6]. The connection with the system is allowe
celerations and displacements [6]. The connection with the system is allowed by with the
possibility of changing the time interval between readings from 10 to 60 min. Dynamic
ternet browser application that provides an overview of each sensor. It gen
ternet browser application that provides an overview of each sensor. It generates
quantities (accelerations) and forces are measured with a frequency of 100 Hz. Their
and notifications if some sensors are out of order. Furthermore, the applicati
and notifications if some sensors are out of order. Furthermore, the application ma
signals are sent to six local servers (SAD). Then they are saved on hard drives and sent
possible to create diagrams of measured values and to export them as .csv file
possible to create diagrams of measured values and to export them as .csv files whi
to the data collectioncompatible with calculation programs. After 7 years, a database of the meas
centre (CGD) located in the motorway management centre. In the
compatible with calculation programs. After 7 years, a database of the measured v
concrete and steel elements of the bridge structure, strains are measured and converted
was compiled,
was compiled, and it and
was it compared
was compared with calculations
with calculations conducted
conducted based based onb
on the
into corresponding stress values. System measurements are additionally supported by the
FEM model.
FEM model.
continuous geodetic measurement of pylon foot settlements.
In of
In case case
the of the Rędziński
Rędziński Bridge
Bridge SHM SHM system,
system, mostly mostly vibrating
vibrating wire s
wire strain
Each sensor has a so-called address. It is a code that indicates the place of its instal-
were used. Measurement of all static quantities is performed at the same tim
were used. Measurement of all static quantities is performed at the same time, wi
lation in the structure, the type of measured quantity and the method of measurement.
Sensors 2021, 21, 1927 4 of 27
Sensors 2021, 21, x FOR PEER REVIEW 4 of 29
Table 2 shows the method of marking the sensors on the platform, pylon and cables. The
sensors measuring
signals are thelocal
sent to six values in the
servers cablesThen
(SAD). are marked
they arewith
saved the
onletter
hardW. The and
drives cables are
sent to
numbered from the pylon to the span direction. Additional markings with the letters L
the data collection centre (CGD) located in the motorway management centre. In the
and P indicate the left and right deck of the structure. The letters W and Z indicate the stay
concrete and steel elements of the bridge structure, strains are measured and converted
cable attachment to the inner or outer edge of the deck, respectively. Table 3 presents a list
into corresponding stress values. System measurements are additionally supported by
of sensors used in the bridge.
the continuous geodetic measurement of pylon foot settlements.
Figure3.3.The
Figure Thestructural
structuralhealth
healthmonitoring
monitoring(SHM)
(SHM)system
systemscheme
scheme[6].
[6].
Table 2. Examples of sensor markings.

Each sensor has a so-called address. It is a code that indicates the place of its instal-
lation in the Quantity
Measured structure, the type of measured
Sample Mark quantity and the method of measurement.
Description
Table 2 shows the method of marking the sensors on the platform, pylon and cables. The
The sensor located in the stay cable No. 16
sensors measuring the values in the cables are(W16),
marked with
which the letter
supports the W.
left The
deck cables
(L) fromare
Force in the
numbered stay the
from cable
pylon to W16-LW/F
the span direction.
the inside (W), measuring the tensionletters
Additional markings with the force L
and P indicate the left and right deck of the structure. The letters W and Z
of the stay cable (F). indicate the
stay cable attachment to the inner or outer edge The
of the deck,
sensor respectively.
located Table
on the stay cable3 No.
presents
21
a list of sensors used in the bridge.
Temperature W21-LZ/Te (W21), on the left deck (L) from the outside
(Z), measuring the temperature (Te).
Table 2. Examples of sensor markings.
Sensors measuring the tension force are only installed on the stay cables reference
Measured
strand. Sample Mark
The isotension method used for the installation Description
and tensioning process attempts to
Quantity
create uniform stresses by applying
The sensor located in the stay force
certain sequence and cableapplications
No. 16 (W16),to which
individual
sup-
Force in The
strands. the uniformity of the strand forces was checked by an individual strand lift-off
W16-LW/F ports the left deck (L) from the inside (W), measuring the ten-
stayeach
after cableinstallation at the quality control step [8].
sion force of the stay cable (F).
The relationship between the measured frequency by the vibrating wire strain gages
Tempera- The sensor located on the stay cable No. 21 (W21), on the left
and the deformation (strain) is described by the Formula (1). The fundamental frequency
W21-LZ/Te
ture deck
(resonant frequency) of the vibration(L) from the outside
of a wire (Z),tomeasuring
is related its tension,the temperature
length, (Te).
and mass:
Sensors measuring the tension force are

1 only
F installed on the stay cables reference
r
f =the installation and tensioning process attempts
strand. The isotension method used for (1)
2L m
to create uniform stresses by applying certain sequence and force applications to indi-

1. to create uniform stresses by applying certain sequence and force applications to indi‐
temperature sensors: Geokon model 4000
Strain and temperature measure‐
Table 3. List and description of sensors.
vidual strands. The uniformity of the strand forces was checked by an individual strand

to create uniform stresses by ments in the pylon and decks
lift‐off after each installation at the quality control step [8].
applying certain sequence and force applications
No. Description Function Photo * to indi‐
Vibrating wire strain gages integrated with
vidual strands. The uniformity of the strand forces was checked by an individual strand
1. Table 3. List and description of sensors.
Sensors 2021,temperature sensors: Geokon model 4000
21, 1927
lift‐off after each installation at the quality control step [8]. 5 of 27

No. Description
Function Photo *
ments in the pylon and decks

Table 3. List and description of sensors.

Table 3. List and descriptionFunction
of sensors.
No. Description
1. Vibrating wire strain gages integrated with Photo *
2. No. temperature sensors: Geokon model
Description Function
Strain and temperature measure‐ Photo *
4100/4150
Vibrating wire strain gages integrated with ments in the pylon and decks
1.
Vibrating wire strain gages integrated

2. Vibrating wire strain gages integrated with
temperature sensors: Geokon model Strain and temperature measurements
1. 1. with temperature sensors: Geokon
4100/4150 in the pylon and
decks
model 4000



2. temperature sensors: Geokon model
4100/4150
Vibrating wire strain gages integrated with Measurement of linear displace‐

3. Vibrating wire strain gages integrated with ments in the pylon and decks
temperature sensors: Geokon model 4420 ments
2. 2. temperature sensors: Geokon model
with temperature sensors: Geokon
Strain and temperature measurements
decks
in the pylon and
model 4100/4150
4100/4150

2.
3. temperature sensors: Geokon model
temperature sensors: Geokon model 4420 Strain and temperature measure‐
ments
4100/4150

3. 3. with temperature sensors: Geokon Measurement of linear displacements
model 4420 ments

Accelerometers:
Vibrating wire strain gages integrated with Measurement of structure accelera‐
Measurement of linear displace‐
4.
3.
Haneywell model MA321
temperature sensors: Geokon model 4420 tions
ments


3.
Accelerometers:
Accelerometers:
Sensors 2021, 21, x FOR PEER REVIEW ments
Measurement of structure accelera‐
6 of 31
4. 4. Measurement of structure
accelerations
tions


Accelerometers: Measurement of structure accelera‐
4.
tions
Temperature sensors: Geokon model
Temperature measurement of
5. 5. Temperature sensors: Geokon model 3800
3800 individual elements of the structure
Temperature measurement of indi‐
Accelerometers: Measurement of structure accelera‐

4. vidual elements of the structure
Haneywell model MA321 tions

Accelerometers:
Vibrating wire tiltmeters: Geokon model Measurement of structure accelera‐

Measurement of angular displace‐
4.
5.
6. Temperature sensors: Geokon model 3800
6350 tions
ments

vidual elements of the structure

Vibrating wire tiltmeters: Geokon model
Vibrating wire tiltmeters: Geokon
Measurement of angular displace‐
6. 6. Measurement of angular displacements
model
6350 6350 ments




5. Force sensor:
Temperature sensors: Geokon model 3800
Force sensor:
7. 7. Advitam Permanent Monostrand Load
Advitam Permanent Monostrand Load Measurement of forces in stay cables
Measurement of forces in stay cables
Cell Model MLC C P
Cell Model MLC C P
Force sensor:
7. Advitam Permanent Monostrand Load
Measurement of forces in stay cables
Cell Model MLC C P * Source of photos No. 1–6: https://www.geokon.com.
* Source of photos No. 1–6: https://www.geokon.com.

* Source of photos No. 1–6: https://www.geokon.com.
(resonant frequency) of the vibration of a wire is related to its tension, length, and mass:
1 𝐹
𝑓 (1)
2𝐿 𝑚
(resonant frequency) of the vibration of a wire is related to its tension, length, and mass:
Sensors 2021, 21, 1927 6 of 27
After the transformations, the following equation may be obtained:

2 2 2 2
ε= 4 f = 4ρL
4mL 4 f (2)
= EA = E (2)
where:
where:
f —frequency (Hz),
f—frequency L—wire length
(Hz), L—wire length (m), F—wire tension
(m), F—wire tension force
force (N), m—wire mass
(N), m—wire mass per
per
unit length (kg/m), ε—strain (-), E—Young’s Modulus (Pa), A—wire
unit length (kg/m), ε—strain (-), E—Young’s Modulus (Pa), A—wire cross section areacross section area
(m22),), ρ—wire
(m material density
ρ—wire material density (kg/m
(kg/m3).3 ).
Equation (1)
Equation (1) can
can also
also be
be used
used for
for the
the determination
determination ofof tension
tension forces
forces in
in stay
stay cables.
cables.
This approach was described in [9] and its usefulness for future research is discussed in
This approach was described in [9] and its usefulness for future research is discussed in
Section 9.
Section 9.
3. The FEM Models

3. The FEM Models
The bridge FEM models were created in the SOFiSTiK software (Figure 4) by using
The bridge FEM models were created in the SOFiSTiK software (Figure 4) by using
the documentation provided by the bridge designer. For a general analysis, a whole bridge
the documentation provided by the bridge designer. For a general analysis, a whole
model was created. The pylon and deck elements were modelled as beam elements, but
bridge model was created. The pylon and deck elements were modelled as beam ele-
real cross-sections, not only the geometric characteristics, were used as the input for the
ments, but real cross-sections, not only the geometric characteristics, were used as the
model. The steel core in the upper parts of the pylon was modelled in the specified sections,
input for the model. The steel core in the upper parts of the pylon was modelled in the
because SOFiSTiK enables the calculation of the geometric characteristics just as in the case
specified sections,
of a substitute because SOFiSTiK enables the calculation of the geometric character-
cross-section.
istics just as in the case of a substitute cross-section.
(a)
(b)
Figure 4. The FEM models: (a) the general bridge model; (b) a detailed single strand model.
Sensors 2021, 21, x1927
FOR PEER REVIEW 77of
of 29
27
The
The bridge
bridge staystay cable
cablesystem
systemisismade made of of Freyssinet
Freyssinet cables
cables [8]. [8].
EachEach
cablecable is com-
is composed
posed of a different number of strands (from 24 to 48). For the sake
of a different number of strands (from 24 to 48). For the sake of simplification, the cables of simplification, the
cables were modelled
were modelled as circular
as circular cross-sections
cross-sections with an withareaan corresponding
area corresponding to thetonumber
the num- of
ber of strands.
those those strands.
In the In the general
general model,model,
the cablesthe cables are modelled
are modelled as a element,
as a single single element,
which
which
can only canbeonly be tensioned.
tensioned. In this Inmodel,
this model, all calculations
all calculations werewere conducted
conducted according
according to theto
the linear analysis.
linear analysis.
In
In order
order to to test
test the
the exact
exact workwork of of a single strand, a representative strand model model was
developed
developed for each each cable
cable(Figure
(Figure4b). 4b).InInthis
this case,
case, both
both a non-linear
a non-linear modelmodel of geometry
of geometry and
and material
material was was adopted.
adopted. The cross-section
The cross-section was described
was described by theby the parameters
parameters for a sev-
for a seven-wire
en-wire
strand. Thestrand. The loading
loading on such on such awas
a strand strand wasweight
its own its ownand weight and theforce
the tension tension force
obtained
obtained
from the from
general theFEMgeneral
model FEM ormodel
from the or SHM
from the SHMThe
system. system.
model Thewasmodel was discre-
discretized with
elements
tized withfrom a length
elements fromofa0.1 m. The
length of 0.1verification of such a model
m. The verification of suchwas the overhang,
a model was the
calculated calculated
overhang, on the basison ofthe
its own
basisweight and weight
of its own the average
and tension force,tension
the average which in the case
force, which of
thethe
in longest
case of stay
thewas 1.20 m.
longest stayThiswasvalue
1.20 was consistent
m. This value was withconsistent
the designwith assumptions.
the design Such
as-
a model will
sumptions. be used
Such in the
a model willfuture to assess
be used in thethe durability
future to assessof the
strand anchorages
durability where
of strand an-a
bending moment
chorages appears, moment
where a bending depending on the depending
appears, tension force. on the tension force.
To be
To besure
surethat
thatthethegeneral
general model
model was was created
created correctly,
correctly, it wasit was
loadedloaded
with with
the loadthe
load that was used during the static and dynamic load tests,
that was used during the static and dynamic load tests, described in [10], i.e., with thedescribed in [10], i.e., with
the load
load of twenty
of twenty 40-tonne
40-tonne lorries
lorries on each
on each deck. deck.
TheThe displacement
displacement of the
of the FEM FEM model
model is
is shown in Figure 5 and is very close to the load tests values.
shown in Figure 5 and is very close to the load tests values. To compare, during the static To compare, during
the static proof-load
proof-load test, a displacement
test, a displacement of 467.0 mmofwas 467.0 mm wasThe
measured. measured.
calculatedThe calculated
displacement
displacement
totalled 431.2 mm.totalled 431.2the
During mm. testDuring the testdifferent
loads, several loads, several differentwere
load schemes loadused
schemesand
compared with the model. This article presents a comparison with a representative re-a
were used and compared with the model. This article presents a comparison with
representative
sult. Furthermore, result.
theFurthermore, the first five
first five frequencies of thefrequencies of the natural
natural vibrations werevibrations
calculatedwere and
compared in Table 4. After those numerical tests, the FEM model was used forwas
calculated and compared in Table 4. After those numerical tests, the FEM model theused
cal-
for the calculations
culations based on the based
SHM ondata.
the SHM data.
(b)
(a)
(c)
Figure 5. (a) The bridge during loading test [10]; (b) FEM model loaded; (c) corresponding loading deck nodes displacements
Figure 5. (a) The bridge during loading test [10]; (b) FEM model loaded; (c) corresponding loading deck nodes dis-
in mm.
placements in mm.

Table 4. Comparison of the calculated and measured frequencies.
FEM Tests
Sensors 2021,Form Number
21, 1927 Frequency Number and Form
[Hz] 8 of[Hz]
27
FEM
Table 4. Comparison of the calculated and measured frequencies. Tests
Form Number Frequency Number and Form
FEM
[Hz] Tests
[Hz]
Form Number Frequency Number and Form
Table 4. Comparison of the calculated and measured frequencies. FEM Tests
Form Number Frequency Number and Form [Hz] [Hz]
Form Number Frequency Number and Form [Hz]
FEM [Hz] Tests [Hz][Hz]
1. 0.25 0.25
1. 0.25 0.25

1. 0.25 0.25
1. 1. 0.250.25 0.25 0.25

2. 0.32 0.31
2. 0.32 0.31

2. 2. 0.320.32 0.31 0.31
2. 0.32 0.31

3. 0.50 0.48
3. 3. 0.500.50 0.48 0.48

3. 0.50 0.48
3. 0.50 0.48


4. 4. 0.510.51 0.48 0.48
4. 0.51 0.48

4. 0.51 0.48
4. 0.51 0.48

5. 5. 0.610.61 0.62 0.62
4. Comparison of the Measured Data

4.1. Temperatures
The analysis of the temperature change’s influence on cable forces was started with
4. Comparison of the Measured Data
the comparison between annual summer and winter values. Diagrams from the SHM
4.1. Temperatures
The analysis of the temperature change’s influence on cable forces was started with
the comparison between annual summer and winter values. Diagrams from the SHM
application showing the force and temperature plots in one of the longest cables between
January 2017 and August 2017 are presented in Figure 6.
of the calculations are shown in Table 6. The measured force differences in strands be-
tween the coldest and warmest days in 2017 are described as ΔF (SHM). The calculated
force change is described as ΔF (FEM).
Comparing the calculated and measured values, it is visible how much they differ
Sensors 2021, 21, 1927 9 of 27
from each other. This indicates how complicated the temperature distribution problem is
concerning bridges [11] or other civil engineering structures [12,13]. The SHM bridge
system measures these values at a few selected points, which gives only a local view of
these changes,
application and does
showing the not characterise
force the entire
and temperature temperature
plots distribution
in one of the in the
longest cables whole
between
structure. Therefore, these measurements cannot be
January 2017 and August 2017 are presented in Figure 6. used for global or long-term struc-
ture analyses.
(a)
(b)
Figure 6. The diagrams in the longest cable LW-40: (a) force diagram; (b) temperature diagram.
Figure 6. The diagrams in the longest cable LW-40: (a) force diagram; (b) temperature diagram.
Table 5 shows the temperature differences in the bridge parts (pylon, deck and cables)
Table 5. Measured temperature differences in structure elements.
which were recorded by the SHM system. Those changes were applied at the FEM model
as a constant increase in temperature
Lowest Daily for individual
Average structure
Highest Dailyelements.
Average The results of the
ΔT
calculations are shown in
Element Table 6. The
Temperature in measured
January force differences
Temperature in strands between the
in August
(°C)
coldest and warmest days in 2017 2017
are described as ∆F (SHM).
(°C) 2017 (°C)
The calculated force change
is described as ∆F (FEM).
decks −2.40 22.90 +25.30
pylon −2.40 22.80 +25.20
Table 5. Measured
cables temperature differences
−8.35 in structure elements.
26.00 +34.35
Lowest Daily Average Highest Daily Average

∆T
Element Temperature in January 2017 Temperature in August 2017
(◦ C)
(◦ C) (◦ C)
decks −2.40 22.90 +25.30
pylon −2.40 22.80 +25.20
cables −8.35 26.00 +34.35
Comparing the calculated and measured values, it is visible how much they differ
from each other. This indicates how complicated the temperature distribution problem
is concerning bridges [11] or other civil engineering structures [12,13]. The SHM bridge
system measures these values at a few selected points, which gives only a local view
of these changes, and does not characterise the entire temperature distribution in the
whole structure. Therefore, these measurements cannot be used for global or long-term
structure analyses.
The Shortest The Shortest The Longest The Longest
Cable Cable Cable Cable
Position
(LZ-W21) (LW-W1) (LZ-W20) (LW-W20)
(kN) (kN) (kN) (kN)
Sensors 2021, 21, 1927 10 of 27
Highest daily average force in
January 2017 63.14 51.82 73.17 68.31
(06/01/17)
Table 6. Measured forces in strands
Lowestofdaily
selected (the shortest
average force inand longest) cables compared with the calculated tension
force change. August 2017 62.80 51.04 73.04 67.91
(18/08/17)
The Shortest Cable The Shortest Cable The Longest Cable The Longest Cable
Position
ΔF(kN)
(LZ-W21) (SHM) (LW-W1) (kN)−0.34 −0.78 (kN)
(LZ-W20) −0.13(LW-W20)−0.40
(kN)
Highest daily average force in ΔF (FEM) −2.60 −2.63 −4.25 −4.21
January 2017 63.14 51.82 73.17 68.31
(06/01/17) In order to obtain a wider description of the nature of the work of the bridge struc-
Lowest daily average force in ture, the changes in extreme annual temperatures in individual elements of the structure
August 2017 were additionally
62.80 checked (Figure
51.047). 73.04 67.91
(18/08/17) Characteristic points in the structure were adopted for the thermal analysis:
∆F (SHM) • −0.34 stay cable (symbol
the longest −0.78W-20), −0.13 −0.40
∆F (FEM) • the shortest
−2.60 stay cable (symbol
−2.63 W-01), −4.25 −4.21
• bottom of the deck section (symbol PD),
• top of the deck section (symbol PG),
In order to obtain a wider description of the nature of the work of the bridge structure,
• steel in the pylon’s upper cross-beam (RS symbol),
the changes in extreme annual temperatures in individual elements of the structure were
• concretechecked
additionally in the pylon’s
(Figureupper
7). cross-beam (symbol RB).
Annual extreme temperature chart at selected points in the structure

50
40
W01 temp. max.
30 W01 temp. min.
W20 temp. max.
W20 temp. min.
20 PD temp. max.
PD temp. min.
PG temp. max.
10 PG temp. min.
RS temp. max.
RS temp. min.
0 RB temp. max.
RB temp. min.
-10
-20
12
13
14
15
16
17
18
19
11
20
20
20
20
20
20
20
20
20
Year
Figure 7. Annual
Figure 7. Annual extreme
extreme temperature
temperature chart
chartat
atselected
selectedpoints
pointsin
inthe
thestructure.
structure.
Characteristic
However, the points in the
measured structure were
temperature adopted
values can befor the for
used thermal analysis:
the local calculation of
•structural elements.
the longest For this
stay cable purpose,
(symbol the measured temperature changes were used to
W-20),
•calculate
the shortest
a singlestay cable
strand (symbol
in the FEM W-01),
model. This model, as mentioned, is described by the
•nonlinear
bottom of the deck
material section
model (symbol PD),
and geometry characteristics as well. The crucial value in these
• top of the deck section (symbol PG),
• steel in the pylon’s upper cross-beam (RS symbol),
• concrete in the pylon’s upper cross-beam (symbol RB).
However, the measured temperature values can be used for the local calculation of
structural elements. For this purpose, the measured temperature changes were used to
calculate a single strand in the FEM model. This model, as mentioned, is described by
the nonlinear material model and geometry characteristics as well. The crucial value in
these calculations was the starting daily strand tension force. Then, a uniform temperature
change ∆T was applied to the strand. Additionally, displacements of the theoretical an-

calculations was the starting daily strand tension force. Then, a uniform temperature
Sensors 2021, 21, 1927 change ΔT was applied to the strand. Additionally, displacements of the theoretical an-
11 of 27
chorages, resulting from heating or cooling in the global bridge model, were imple-
calculations
mented was8).the
(Figure Thestarting daily were
calculations strandperformed
tension force. Then, a uniform
by automatic iterations,temperature
where the
change ΔT was applied to the strand. Additionally, displacements
temporary structure stiffness was saved after each step. The calculations of the theoretical
ended an-
when the
chorages,
chorages,and
stiffness resulting
resulting from
from
cable sag heating
heating
values or
hadorbeen cooling
cooling in the global bridge model, were
in the global bridge model, were implemented
stabilized. imple-
mented8).
(Figure (Figure 8). The calculations
The calculations were performed
were performed by automatic by automatic iterations,
iterations, where where the
the temporary
temporary
structure structure
stiffness wasstiffness waseach
saved after saved after
step. Theeach step. The ended
calculations calculations
when theended whenand
stiffness the
stiffness
cable sagand cable
values had sag values
been had been stabilized.
stabilized.
Figure 8. Strand loading scheme: ΔT—uniform temperature change; di,j—theoretical anchorages

displacements.
Figure
Figure 8.
TableStrand
8. Strand loading
loading
7 shows scheme:
thescheme: ∆T—uniform
ΔT—uniform
described temperature
temperature
calculation change;
results change;
for i,jd
the dthreei,j —theoretical
—theoretical anchorages
longestanchorages
stay cables.
displacements.
displacements.
Based on the graphs in Figures 9 and 10, the range of the maximum and minimum force
as well as the temperature changes were determined. Due to disturbances in the meas-
Table
Table 77shows
shows the described
the describedcalculation results for the
forthree longest stay cables. Based
urement caused by vehicle traffic, it calculation
was decidedresults
to present the
the three
extremelongest
valuesstay
as acables.
value
on the on
Based graphsgraphs
in Figures 9 and 10, the range ofrange
the maximum and minimum force as force
well
range. It isthe
worth noting in Figures 9 and
that between 10, the
10:00 and 11:00ofa.m.,
the maximum
there was aand minimum
rapid temperature
as
as the
welltemperature
as the changeschanges
temperature were determined.
were Due to disturbances
determined. Due to in the measurement
disturbances in the meas-
change by about 4 °C, which resulted with a force peak at the same time. As for the value,
caused
urement bycaused
vehiclebytraffic, it was
vehicle decided
traffic, it wastodecided
present the
to extremethe
present values as a values
extreme value range. It is
as a value
aworth
similar forcethat
noting change occurred
between 10:00 between
and 11:000:00 andthere
a.m., 6:00 a.m.aby
was the
rapid temperature
temperature difference
change by
range.
of It is worth noting that between 10:00 and 11:00 a.m., there was a rapid temperature
3.7 °C.
about 4 ◦ C, which resulted with a force peak at the same time. As for the value, a similar
change by about 4 °C, which resulted with a force peak at the same time. As for the value,
force change
a similar occurred
force change between
occurred 0:00 and 6:00
between 0:00 a.m.
and by
6:00the temperature
a.m. of 3.7 ◦ C.
differencedifference
by the temperature
Table 7. Comparison of forces changes between measured and calculated values on the basis of the
of 3.7 strand
single °C. FEM model.
Table 7. Comparison of forces changes between measured and calculated values on the basis of the single strand FEM model.
Cable7. Comparison
Table Initial Maximum
of Minimum
forces changes between measured Calculated
and calculated values onCalculated
the basis of the
Maximum Minimum −ΔT +ΔT Calculated Calculated
Cable Initial Force Num-
single Force
strand FEM Force
model. Range Force Range
◦ ◦ Maximum Minimum
Force Range Force Range −∆T ( C) +∆T ( C) Maximum
(°C) (°C) Minimum
Number (kN) ber (kN) (kN) (kN) (kN) Force (kN) Force
(kN) Force (kN) Force(kN)
(kN)
Cable Initial Maximum Minimum Calculated Calculated
LZ-W16 104.5 104.7–105.3 102.9–103.4 −3.7 +15.3
−ΔT+15.3
+ΔT 104.91 102.78
LZ-W16 104.5 104.7–105.3
Num- 102.9–103.4
Force Force Range Force−Range
3.7 104.91
Maximum 102.78
Minimum
LZ-W18 103.5 103.4–104.4 100.9–101.3 (°C) +15.3
−3.7 (°C) 103.41 101.28
LZ-W18 103.5 103.4–104.4
ber (kN) 100.9–101.3
(kN) −3.7
(kN) +15.3 103.41
Force (kN) 101.28
Force (kN)
LZ-W20 68.5 69.3–69.9 66.5–67.2 −3.7 +15.3 69.31 67.18
LZ-W20 68.5 LZ-W16 104.5
69.3–69.9 104.7–105.3
66.5–67.2 102.9–103.4
−3.7 −3.7+15.3
+15.3 104.91
69.31 102.78
67.18
LZ-W18 103.5 103.4–104.4 100.9–101.3 −3.7 +15.3 103.41 101.28
LZ-W20 68.5 69.3–69.9 66.5–67.2 −3.7 +15.3 69.31 67.18
Figure 9. Daily strands temperature change (1 August 2017).
Figure 9. Daily strands temperature change (1 August 2017).

Sensors 2021,
Sensors 21, x1927
2021, 21, FOR PEER REVIEW 1212of
of 29
27
Force in strand [kN]
0:00
1:00
2:00
3:00
4:00
5:00
6:00
7:00
8:00
9:00
10:00
12:00
13:00
14:00
15:00
16:00
17:00
18:00
19:00
20:00
21:00
22:00
23:00
24:00
11:00
0:00
1:00
2:00
3:00
4:00
5:00
6:00
7:00
8:00
9:00
10:00
12:00
13:00
14:00
15:00
16:00
17:00
18:00
19:00
20:00
21:00
22:00
23:00
24:00
11:00
0:00
1:00
2:00
3:00
4:00
5:00
6:00
7:00
8:00
9:00
10:00
12:00
13:00
14:00
15:00
16:00
17:00
18:00
19:00
20:00
21:00
22:00
23:00
24:00
11:00
Figure
Figure 10. Daily force
10. Daily force course
course in
in single
single strands
strands of
of cables
cables LZ-W16,
LZ-W16, LZ-W18 and LZ-W20 (1 August 2017).
The presented
The presented analysis
analysis and
and results
results comparison
comparison show show that
that aa detailed
detailed calculation
calculation ap-ap-
proach accurately
proach accurately reflects
reflects the
the work
work of of the
the structure
structure in in the
the context
context of of the
the temperature
temperature
changes of
changes of cable
cable elements.
elements. The The temperature
temperature changes
changes are are significant
significant in in relation
relation toto the
the
displacement of the cable anchorages relative to each other. Applying
displacement of the cable anchorages relative to each other. Applying them in the cables them in the cables
detailed in
detailed in the
the FEM
FEM model
model allows
allows for
for aa more
more accurate
accurate estimation
estimation of of force
force changes
changes andand
stresses in particular cable sections. However, the analysis of the temperature
stresses in particular cable sections. However, the analysis of the temperature distribu- distribution
in whole
tion in wholesuspended
suspended structures composed
structures composed of various
of variousmaterials
materialsis aiscomplex
a complex issue and
guess-
requires
work and further and more
requires furtheraccurate measurements
and more accurate than just SHM-based
measurements than measurements
just SHM-based to
be able to create a representative load model. A complex description
measurements to be able to create a representative load model. A complex description of of temperature distri-
bution in bridges
temperature is presented,
distribution e.g., is
in bridges in presented,
[11]. It is also
e.g.,worth adding
in [11]. thatworth
It is also the differences
adding that in
temperature distribution, that should be applied to the calculation models,
the differences in temperature distribution, that should be applied to the calculation are included in
the design standards. The analysis performed according to SHM measurements, supported
models, are included in the design standards. The analysis performed according to SHM
by FEM calculations, allows to check the plausibility of the standard assumptions.
measurements, supported by FEM calculations, allows to check the plausibility of the
standard
4.2. Trafficassumptions.
Loads
Figure 11 is a diagram for the cable force in the year 2017 in one of the medium-length
4.2. Traffic Loads
cables. Major force peaks occurred on the diagram in February, May, July, October and
Figure 11
December. Thisis was
a diagram for the
the reason cable force
for creating in the lines
influence year for
2017
theincable
one tension
of the medi-
force
um-length cables. Major force peaks occurred on the diagram in February,
(Figure 12). Table 8 shows calculations that were helpful in the assessment of what May, July,
kind of
October and December. This was the reason for creating influence lines for the
vehicles may have crossed the bridge. A force sensor installed on the cable only measures cable ten-
sion force in
the force (Figure 12). Table
one single 8 shows
strand. calculations
For the that wereithelpful
force assessment, in the assessment
is necessary to know how of
what kind of vehicles may have crossed the bridge. A force sensor installed
many strands are installed in each cable. Multiplying the force results from the SHM on the cable
only measures the force in one single strand. For the force assessment, it is necessary to
know
knowhowhowmanymanystrands
strandsareareinstalled
installedinineach
eachcable.
cable.Multiplying
Multiplyingthe theforce
forceresults
resultsfrom
from
the SHM measurement by the number of strands is allowed due
the SHM measurement by the number of strands is allowed due to the cable assembling to the cable assembling
technology
technologyused.
used.ItItprovides
providesthe thesame
samestress
stressinineach
eachstrand
strand[8].[8].
Some
Some of the peaks on the force diagram may be consideredtotobe
of the peaks on the force diagram may be considered bemeasuring
measuringerrors.
errors.
Sensors 2021, 21, 1927 13 of 27
InInorder
order to exclude them, it is necessary to check whether the neighbouringcables
to exclude them, it is necessary to check whether the neighbouring cableshave
have
similar
similarpeaks
peaksatatthethesame
sametime.
time.Table
Table88presents
presentsestimations
estimationsfor forthe
theproved
provedpeaks
peaksonly
only
(also
(also marked on Figure 11). The bigger peaks cannot be caused by a single vehicle,sosoitit
marked on Figure 11). The bigger peaks cannot be caused by a single vehicle,
was
wasnecessary
measurement
necessarybytotoconsider
the number
consider aasituation
of strands
situation with aatraffic
is allowed
with jam
due
traffic jam oror
to the
the accumulation
cable
the assembling
accumulation ofofheavy
heavyve-
technologyve-
hicles.
used. The
It situation
provides the is presented
same stress in
in the
eachlast column
strand [8].
hicles. The situation is presented in the last column of Table 8.of Table 8.
Figure
Figure11.
Figure 11.Force
11. Forcediagram
Force diagramin
diagram inthe
in themiddle-length
the middle-lengthcables
middle-length cablesLZ-W12
cables LZ-W12and
LZ-W12 andLW-W12
and LW-W12with
LW-W12 withselected
with selectedmaximum
selected maximumforce
maximum forcepeaks
force peaksin
peaks inin2017.
2017.
2017.
Figure
Figure12.
Figure 12.Influence
12. Influenceline
Influence lineof
line oftension
of tensionforce
tension forcefor
force forcables
for cablesLZ-W12
cables LZ-W12and
LZ-W12 andLW-W12.
and LW-W12.
LW-W12.
Table Selected
8. the
After and estimated
assessment force peaks
of crossing cars, ainsimulation
the year 2017.
of a moving vehicle on one bridge
After the assessment of crossing cars, a simulation of a moving vehicle on one bridge
deck
deck was
wascarried
carriedout
outwith
withthe
theFEM
FEM model.
model. The
Theaim
aimofofEstimated
that
thatcalculation
calculationwaswastotoassess
assessthethe
Monthly Force Force Difference
Peak Number Peak biggest deck displacement and accelerations during the crossing and to check the
biggest Strand
Average deck displacement
Difference and
in accelerations
in Cable during Loading
the crossing
on and to check
Example theme-
Vehicles me-
(Month) Value dium force
diumForceforcepeaks
peaksappearing
Strand on
appearing onthe
thecable
(48 force
forcespectrum.
Strands)
cable Whole
spectrum. The calculation
calculationwas
Deck
The wascarried
carriedoutout
for a 40-tonne truck that was moving with the speed of 80 km/h
for a 40-tonne truck that was moving with the speed of 80 km/h (Figure (Figure 13).
13).Figure
Figure
3 loaded concrete
14a
14a
shows
showsthe thevertical
verticaldeck
deckdisplacements
displacementsforforthree
threepoints
pointsofofthe
thedeck
deckstructure
structure (description
(description
mixers
1 (February) 101.8 kN 97.3 kN 4.5 kN 216.0 kN 360 t
under the figure).
under the figure). (3 × 40 t) + 10 lorries
(10 × 15 t) and cars
Military transport
(e.g., Stanag MLC
2 (May) 103.6 kN 97.9 kN 5.7 kN 273.6 kN 460 t
150) + lorries (15 t)
and cars
Military transport
(e.g., Stanag MLC
3 (July) 105.0 kN 98.0 kN 7.0 kN 336.0 kN 570 t
150) + lorries (15 t)
and cars
Table 8. Selected and estimated force peaks in the year 2017.
Sensors 2021, 21, 1927 Force 14 of 27
Force Estimated
Peak Number Peak Monthly Average Difference Example
Difference in Loading on
(Month) Value Strand Force in Cable Vehicles
Strand Whole Deck
Some of the peaks on the force diagram (48 may
Strands)
be considered to be measuring errors.
In order to exclude them, it is necessary to check whether the neighbouring 3cables loadedhaveconcrete m
similar
1 (February) 101.8 kN peaks at the
97.3 kN same time. Table
4.5 kN 8 presents estimations
216.0 kN for the proved
360 t peaks
(3 × 40 t) (also
only + 10 lorries (
marked on Figure 11). The bigger peaks cannot be caused by a single vehicle, so it was
t) and cars
necessary to consider a situation with a traffic jam or the accumulation of heavy vehicles.
The situation is presented in the last column of Table 8. Military transpo
2 (May) 103.6 kNAfter the
97.9 kN
assessment 5.7 kNcars, a simulation
of crossing 273.6 kN of a moving460 tvehicle on(e.g., Stanag MLC 1
one bridge
lorries
deck was carried out with the FEM model. The aim of that calculation was to assess the (15 t) and c
biggest deck displacement and accelerations during the crossing and to check the Military
medium transpo
3 (July) force
105.0 kN peaks appearing
98.0 kN on the cable
7.0 kNforce spectrum.
336.0 kNThe calculation
570 t was carried out for a MLC 1
(e.g., Stanag
40-tonne truck that was moving with the speed of 80 km/h (Figure 13). Figure 14a shows
lorries (15 t) and c
the vertical deck displacements for three points of the deck structure (description under
the figure).
Figure 13. Screenshots

Figure from the moving
13. Screenshots from thevehicle
moving on vehicle
the FEMon
bridge model.
the FEM bridge model.
Sensors 2021, 21, 1927 15 of 27
(a) (b)
Figure
Figure 14.
14. (a)
(a) Node
Node displacements
displacements for
for 33 measuring
measuring points
points on
on one
one span,
span, magenta
magenta line:
line: deck
deck point
point under
under the
the longest
longest cables
cables
W40, dark blue line: deck point under the medium–long cables W32, brown line: deck point under the shortest
W40, dark blue line: deck point under the medium–long cables W32, brown line: deck point under the shortest cables W21; cables
W21; (b) tension
(b) tension force
force in thein the whole
whole cablescables W12,
W12, red redouter
line: line: outer deck cable,
deck cable, green green line: inner
line: inner deck cable.
deck cable.
The
The SHM
SHM system
system is is not
not equipped
equipped with with vertical
vertical displacement
displacement sensors.
sensors. WithWith aa proper
proper
FEM model, the bridge supervisor is able to assess that important
FEM model, the bridge supervisor is able to assess that important value. The biggest value. The biggest
displacement
displacement was was noticed
noticed in in the
the middle
middle part part of the span (26.05 mm). Figure 14b shows
the force change
the force change that appeared in the medium-length cables W12. ThereThere can be noticed a
difference
difference between the inner and outer cables. The The outer
outer cables
cables are
are aa little
little longer,
longer, because
of
of the
the deck
deckrotation
rotationtotoprovide
providethe thetransverse
transverse slope
slopeononthethe
road surface.
road surface. It causes an un-
It causes an
symmetrical
unsymmetrical load distribution
load distribution between
between cables.
cables.Such
Sucha adifference
differenceisisalsoalsonoticed
noticed by by the
SHM
SHM system
system (Figure
(Figure 11) in the corresponding cable pair.
Table 9 shows
Table shows the thecalculation
calculationofofforceforce change
change forfor
thethe LZ-W12
LZ-W12 andand LW-W12
LW-W12 cables.
cables. The
The
values values
werewere calculated
calculated for a for a single
single strand strand
(these(these particular
particular stays stays
consistconsist of 48 strands
of 48 strands each).
It was Itassumed
each). was assumed that onethatvehicle weighing
one vehicle 40 tonnes
weighing is moving
40 tonnes on the deck.
is moving on theThe results
deck. The
show that
results show forthat
a simple
for aand quick
simple andestimation of the tension
quick estimation of theforce changes,
tension forceitchanges,
is enough it to
is
use the tension
enough to use the forcetension
influence line.
force For a more
influence line.detailed description,
For a more detailed a complete
description, analysis
a com- of
the vehicle’s
plete analysiscrossing should be
of the vehicle’s confuted,
crossing because
should then, the because
be confuted, work of then,
the stay thecable
worksystem
of the
is more accurately included. As a result, the difference between
stay cable system is more accurately included. As a result, the difference between the force change in the
external and internal cable is noticeable. This relationship is also visible
force change in the external and internal cable is noticeable. This relationship is also vis- in Figure 11, where
the in
ible measured
Figure 11, force alsothe
where takes into account
measured force alsothe takes
self-weight of the the
into account structure.
self-weightFor larger
of the
vehicles passing,
structure. For larger different
vehiclespeaks are visible.
passing, different peaks are visible.
The presented
The presented and and verified
verifiedapproaches
approachesfor forestimating
estimatingchanges
changes inintension
tension forces cancan
forces be
used to characterize the work of cables that are not equipped
be used to characterize the work of cables that are not equipped with sensors. with sensors.
Another important issue is the estimation of the vibration acceleration of the structure
on the basis of vehicle passage. Although the SHM system is equipped with vibration
sensors, these do not give accurate results as they record the measured values every 5 min.
Sensors 2021, 21, 1927

Table 9. Comparison of tension force changes in one strand of LZ-W12 and LW-W12 cables calcu-16 of 27
lated according to the influence line and with the modelled dynamic passage of a 40 t vehicle.
Force Change Read from the

Force Change Calculated
Stay Cable Number FEM Model during the
For dynamic values, such a timefrom interval is completely
Influence Lines inefficient. It only gives information
about the range of changes in the vibration acceleration of the structure.DynamicBoth Crossing
phenomena
LZ-W12 1.07 kN 1.20
are presented in the daily measurement of the acceleration of the vertical vibrations kN of the
LW-W12
deck at the anchorage of the W-32 cables. 1.05Note
kN that the system acceleration
0.94 kNsensor requires
calibration; changes oscillate around a value other than zero (Figure 15a).
Another important issue is the estimation of the vibration acceleration of the
structure on the basis
Table 9. Comparison of vehicle
of tension force passage.
changes inAlthough the
one strand of SMH and
LZ-W12 system
LW-W12is equipped with
cables calculated
vibration
accordingsensors, these do
to the influence linenot
andgive
withaccurate results
the modelled as they
dynamic record
passage of athe
40 tmeasured
vehicle. values
every 5 min. For dynamic values, such a time interval is completely inefficient. It only
Force Change Calculated Force Change Read from the FEM
gives information
Stay Cable Number about the range of changes in the vibration acceleration of the struc-
from Influence Lines Model during the Dynamic Crossing
ture. Both phenomena are presented in the daily measurement of the acceleration of the
LZ-W12 of the deck at the
vertical vibrations 1.07anchorage
kN of the W-32 cables.1.20 kNthat the system
Note
acceleration sensor requires calibration;
LW-W12 1.05 kNchanges oscillate around a0.94valuekN other than zero
(Figure 15a).
(a)
Acceleration [m/s 2 ]
10:00
12:00
13:00
14:00
15:00
16:00
17:00
18:00
19:00
20:00
21:00
22:00
23:00
24:00
0:00
1:00
2:00
3:00
4:00
5:00
6:00
7:00
8:00
9:00
11:00
(b)
Figure 15. (a) Deck accelerations near the cable W-32 anchorages registered by the SHM system; (b) vertical acceleration in
Figure 15. (a) Deck accelerations near the cable W-32 anchorages registered by the SHM system; (b) vertical acceleration
inthe
themiddle
middlespan
spandeck
deckpoint
point(below the
(below cables
the W-32)
cables during
W-32) the
during vehicle
the crossing.
vehicle crossing.
Figure 15b shows the results of the vertical accelerations in the middle span deck point.
The maximum accelerations are between 0.06 m/s2 and 0.08 m/s2 and are the effect of a
heavy vehicle crossing. Lower calculated values (0.02 m/s2 –0.04 m/s2 ), before the crossing,
Sensors 2021, 21, 1927 17 of 27
correspond to the values change range measured with the SHM system. Accelerations of
the vibrations at this level are negligible. The calculation model assumes no damping. The
damping factor for such a structure as the R˛edziński Bridge should be tested separately,
which is a subject of ongoing investigation. The presented approach aims to show only the
maximum possible values of vibration accelerations, which supplement the data measured
by the SHM system. Further advanced static and dynamic analyses should be carried out
in accordance with [14,15].
The comparison shows that due to the required sampling frequency in measurements
of the vibration acceleration, long-term monitoring (several years) would be too burden-
some. For this purpose, high-capacity, high-speed storage drives would be needed, which
from a maintenance point of view, could prove to be an expensive undertaking.
The peaks appearing on the force diagrams in SHM data often approach the upper
safety limit set in the SHM system, therefore the cause of their appearance should be
determined each time. Often, these can be just a measurement error. However, when they
are a real representation of the load and occur too often, they can cause fatigue damage to
cables. Therefore, their source should be determined, and the effects of this load checked
in FEM models. It should also be remembered that exceptional vehicles often pass on the
bridges and need the appropriate permits. The manager equipped with the SHM system
of the facility and its FEM model can, using such tools, easily determine the conditions
and possibilities of such a passage and what impact it may have on the durability of the
structure. In addition, we are constantly dealing with an increase in vehicle traffic on the
roads. The ongoing updating of calculations compared with the results from the SHM
system will allow for the accurate determination of the current load capacity of the object.
When the calculated and measured values start to differ significantly, this will indicate a
possible failure of the facility and a quick decision may prevent a catastrophe.
5. Cable Forces Analysis

From the design and analytical point of view, it is important to determine the impact
of individual types of loads on the fatigue strength of stay cables. The works [6,16]
discuss various approaches to considering the moving loads on bridges. It is assumed that
continuous vehicle traffic has a completely different impact on the forces and displacements
(vibrations) of cables than other short-term intense weather changes such as windstorms or
big amplitudes of temperature. The amount of collected data allows one to create statistics
on the occurrence of particular types of bridge loads. An additional advantage of the
monitoring system measurements is the fact that the data are collected continuously in
real time. Parallel simulations based on the already collected data can be verified on an
ongoing basis. The measured tension force in the strands is the effect of various impacts on
the bridge, which include dead load, thermal impact, vehicle crossing and wind pressure
on the structures. The whole impact can be described with the general formula:
F(t) = g(t) + T(t) + q(t) + w(t), (3)
where:
F(t)—total force in the cable strand; g(t)—component of force from dead load; T(t)—
component from thermal interactions; q(t)—random function that illustrates the impact of
passing vehicles; w(t)—random function representing the influence of wind; t—time.
Comparing the diagrams in Figures 9 and 10, the influence of temperature on the
change of the tension force is clearly visible. The temperature drop leads to increased
tension in the stay cable and vice versa. Analysing the fluctuations of forces during the
day, one can ignore the influence of rheological changes in the concrete of the bridge and
the pylon. This change, in the form of function g(t) in Equation (3), is negligibly small, and
only noticeable in the long-term analysis.
For this reason, in the first phase of the analysis presented, Equation (3) will be
simplified to:
F(t) = G(t) + Q(t), (4)
Sensors 2021, 21, 1927 18 of 27
where:
G(t)—total tension force from own weight and temperature changes; Q(t)—changes in
the tension force from variable loads. Function G(t) was approximated by successive series
of the Fourier series:
n
G (t) = a0 + ∑ ai cos(iωt) + bi sin(iωt) (5)
i =1
The correctness of fitting the G(t) function to F(t) was measured with four parameters.
Sum of Squares Due to Error. This statistic measures the total deviation of the response
values from the fit to the response values. It is also known as the summed square of
residuals and is usually labelled as SSE. The closer the value is to 0, the better the match.
R-Square. This statistic measures how successful the fit is in explaining the variation
of the data. In other words, R-square is the square of the correlation between the response
values and the predicted response values. It is also known as the square of the multiple
correlation coefficient and the coefficient of multiple determination. The closer the value is
to 1, the better the match.
Root Mean Squared Error. This statistic is also known as the fit standard error and the
standard error of the regression. It is an estimate of the standard deviation of the random
component in the data. The closer the value is to 0, the better the match. Table 10 presents
an example of the scope of the particular fit parameters for subsequent stays on 17 August
2018. Figure 16 presents the result of such an operation for one whole month (May 2013)
for the W16 stay. The first graph is a pure data plot from the SHM system, F(t), the next is
a graph of successive approximated daily functions, G(t), and the last one shows filtered
force changes—amplitudes, Q(t).
Table 10. Sample parameters from 17 August 2018.
Cable SSE R-Sqr RMSE

LZ-W20 1.2 0.984 0.067
Sensors 2021, 21, x FOR PEER REVIEW LZ-W16 2.1 0.981 0.089 19 of 29
LZ-W01 5.4 0.976 0.141

Stress [MPa]
Stress [MPa]
Amplitude [MPa]
Figure 16. Graphs

Figure 16. Graphs showing
showing split
split data
data according
according to
to the
the described
described algorithm.
algorithm.
6. Histogram Interpretation
Figure 17a shows these histograms for the discussed cables. Ranges from −5 to 5
MPa were defined in order to improve their readability. By examining the entire stress
Sensors 2021, 21, 1927 19 of 27
On the basis of such mathematical operations, monthly and annual histograms de-
scribing the range of stresses and their changes were made for the separated graphs. Their
statistical interpretation will be presented in the next chapter. The presented algorithm
can also be an alternative or supplement to the commonly used rain-flow cycle counting
method. The basics of such probabilistic calculations are described, e.g., in [17,18].
The presented mathematical procedure allows for a better understanding of the charac-
teristics of the structure’s work. According to the authors, subsequent SHM systems could
be equipped with such, or even more complex, algorithms for an immediate separation of
the current measured signal.
6. Histogram Interpretation
Figure 17a shows these histograms for the discussed cables. Ranges from −5 to 5 MPa
were defined in order to improve their readability. By examining the entire stress spectra,
stress changes of up to 60 MPa appear in them. Figure 17b presents the histograms of
stresses in the analysed cables without separating thermal influences and dead weight. The
dissipation of stress values over a year is visibly larger due to the influence of temperature
on the tension force in the cable.
According to the Palmgren–Miner rule, each cycle ni (two amplitudes are a whole
cycle) causes a small damage in the steel structure n/N. N is the amount of particular
destructive cycles due to the Wöhler curve and n is the cycles measured. More extensive
descriptions of the advanced fatigue calculations, considered in this article, can be found
in [19,20]. If D describes the sum of all damages caused by different cycles, a simple
equation can be formulated, namely:
q
n
D= ∑ Nii , (6)
i =1
In order to compare this classical method of counting cycles, comparative calculations

were made to the approach proposed in this paper. Two comparison parameters were
introduced: Da and Ds .
n
Da = ∑ σa,i ni (7)
i
1 n
2∑
Ds = | σs,i | ni (8)
i
At this point, Formulas (7) and (8) should be interpreted. Based on the Formula (6),
parameter ni was omitted. This procedure is allowed due to the fact that ni is a constant
value resulting from the i-th interval directly from the Wöhler curve. The magnitude of
σa,i is the amplitude of one complete symmetric cycle of change in stress, calculated with
the rain-flow procedure. These are values greater than zero. Coefficient Ds contains stress
| σs,i | oscillations around the average value. Due to the method adopted here, this value is
0, and therefore the fluctuations are negative and positive. They can be treated as separate
upper and lower half cycles. Therefore, value 0.5 appears before the sum sign in Formula
(8) to enable a comparison with parameter Da . Table 11 compares parameters Da and Ds
for the analysed stay cables from 2018.
Table 11. Parameters Da and Ds in year 2018.
Stay Cable Ds Da Ds /Da

LZ-W01 16,705.2 13,983.9 1.19
LZ-W16 34,320.3 28,593.4 1.20
LZ-W20 52,831.9 44,848.1 1.18
Sensors 2021, 21, 1927 20 of 27
(a) (b)
Figure 17.(a)
Figure17. (a)Histograms
Histogramsof
ofstress
stresschanges.;
changes.;(b)
(b)amplitude
amplitudehistograms
histogramswithout
withoutseparated
separatedtemperature
temperatureinfluences.
influences.
As can be seen, ratio Ds /Da is greater than zero. On this basis, it can be concluded
At this point, Formulas (7) and (8) should be interpreted. Based on the Formula (6),
that the changes in the forces in the stays are not symmetrical. Tension is greater than
parameter ni was omitted. This procedure is allowed due to the fact that ni is a constant
compression.
value resulting from the i-th interval directly from the Wöhler curve. The magnitude of
The histograms created are an introduction to performing accurate fatigue calculations.
is the amplitude of one complete symmetric cycle of change in stress, calculated with
The, methods described in the literature [20,21] are based on the constant mean force, the
the rain-flow procedure. These are values greater than zero. Coefficient Ds contains stress
accompanying amplitudes and the Wöhler curve. The above analyses show that with long-
| , | oscillations around the average value. Due to the method adopted here, this value is
term observations, the mean tension force changes. Therefore, it is important to separate
theand
0, therefore
long-term theshort-term
and fluctuations are negative
effects, and theand positive. of
calculation They
thecan be treated
failure as sepa-
accumulation
Sensors 2021,21,
Sensors2021, 21,1927
x FOR PEER REVIEW 21 of
22 of 27
29
Table 13. Values

parameter of the failure
D should accumulation
be performed parameter
for smaller D. ranges, e.g., by weeks or months.
data
Even if the parameter D will not increase quickly, it can be used in further calculations
Year Cable LW-20 Cable LW-16 Cable LW-01
as an index reducing the load capacity of the material due to microscopic damage to the
2011 0.00 0.00 0.00
material structure. Moreover, the above analysis of the stay cable statics shows that the
classical fatigue strength calculation methods for cables, based on the Wöhler0.00
2012 0.00 0.00 curve, may be
2013 1.71 × 10
inaccurate. Therefore, the so-called fatigue surfaces or Haigh diagrams should×be
7.57 × 10 1.74 10used
−6 −7 −9
as it
2014
is indicated 3.48 × Their
in [18,21,22]. 10 −6
methodology is 1.46 × 10on a separate influence
based
−6 3.67 ×of10the
−9
mean
force2015
and amplitudes, 5.62which
× 10−6solves the problem 1.97of× the
10−6differences between4.73the
× 10presented
−9
2016
parameters Da and 8.04
Ds . × 10 −6 2.58 × 10 −6 5.88 × 10−9
The general idea
2017 1.08of×calculating
10−5 the D parameter
3.33 × 10on −6 the basis of the 7.32
fatigue
× 10surface
−9 is
shown2018 in Figure 18. In order
1.43 × 10 to
−5 create such a surface,
4.15 × 10two−6 data distributions should
8.85 × 10 −9be used:
p(σ2025
a )—the
* probability 8.75distribution
× 10−5 function of2.15
amplitude
× 10−5 stress changes, and
4.24 p(σ
× 10 −8m )—the
probability
2050 * distribution 5.02 ×function
10−3 of the average5.80tension
× 10−4 stress. The functions describing
1.26 × 10 −6
these distributions
2075 * are presented
3.89 × 10 −2 in Table 12. They
2.83 × 10 were
−3 calculated on6.30 × 10−6 of the
the basis
separated diagrams
* for simulated data. shown in Figures 16 and 17.
Figure18.
Figure 18.Calculation
Calculationscheme
schemeofofthe
thefatigue
fatigue parameter
parameter D,D, based
based on on
thethe fatigue
fatigue surfaces
surfaces [18,21,22].
[18,21,22]. Description
Description of mark-
of markings
ings in the
in the text. text.
Table Applied, exemplary types of probability distribution function of the analysed quantities.
12. Simulations
7. Data
A frequent problem of large Distribution
Probability monitoring Function
systems are their temporary
(PDF) failures
Distribution related
Parameters
to electricity supply, etc. This may result in longer or shorter periods related to the lack of
Average stress Normal σ—average value
measured data. To do this, in order 1to keep −( the
σm −overview, the data should deviation
be properly

µ )2
σm p(σm ) = √ exp 2σ2
µ—standard
simulated. The advantage of long-term σ 2π SHM measurements is that their database is con-
stantly growing. The histograms and
Stress amplitudes graphs
Weibull created above showλ—scale how data can be di-
parameter
σa (k −1) −(σa /λ)k
vided and σa ) = λk form,
σa sorted. In such ap(prepared λ they
e can be used to simulate
k—shape consecutive
parameter or
missing data. Figure 19 shows the idea of such a simulation scheme.
The fatigue surface N(σa , σm ) is calculated and described individually for each steel,
just like the Wöhler curves [18]. The stress probability density area p(σa , σm ) is the product
of the single probability distribution functions (PDFs), described above. R is the yield
strength of steel; n1 is the number of cycles in the measured period t. The whole idea is
derived directly from Formula (6).
Sensors 2021, 21, 1927 22 of 27
Table 13 shows the calculated parameters for measurements from the available years
and for future data simulations, in accordance with the projected durability of the particular
bridge. The method of data simulation is described in Section 7.
Table 13. Values of the failure accumulation parameter D.
Year Cable LW-20 Cable LW-16 Cable LW-01

2011 0.00 0.00 0.00
2012 0.00 0.00 0.00
2013 1.71 × 10−6 7.57 × 10−7 1.74 × 10−9
2014 3.48 × 10−6 1.46 × 10−6 3.67 × 10−9
2015 5.62 × 10−6 1.97 × 10−6 4.73 × 10−9
2016 8.04 × 10−6 2.58 × 10−6 5.88 × 10−9
2017 1.08 × 10−5 3.33 × 10−6 7.32 × 10−9
2018 1.43 × 10−5 4.15 × 10−6 8.85 × 10−9
2025 * 8.75 × 10−5 2.15 × 10−5 4.24 × 10−8
2050 * 5.02 × 10−3 5.80 × 10−4 1.26 × 10−6
2075 * 3.89 × 10−2 2.83 × 10−3 6.30 × 10−6
* for simulated data.
7. Data Simulations
A frequent problem of large monitoring systems are their temporary failures related
to electricity supply, etc. This may result in longer or shorter periods related to the lack
of measured data. To do this, in order to keep the overview, the data should be properly
simulated. The advantage of long-term SHM measurements is that their database is
constantly growing. The histograms and graphs created above show how data can be
divided and sorted. In such a prepared form, they can be used to simulate consecutive or
missing data. Figure 19 shows the idea of such a simulation scheme.
Figure
Figure 19. Data simulation 19. Data
scheme. simulation
Step scheme.
1: no data; step 2:Step 1: no data;
simulating thestep 2: simulating
average course ofthe average
forces course
on the of forces
following on
days
the following
on the basis of a randomly selected days on the
Fourier basis
series from of athe
randomly
created selected
database; Fourier
step 3:series
linearfrom
datathe
fitcreated database;
correction; step 4:step 3:
sim-
ulation of amplitudes linear datainverse
using the fit correction; step 4:function.
distribution simulation of amplitudes using the inverse distribution function.
UsingUsing
data from the relevant
data from months
the relevant in the previous
months years, an
in the previous empirical
years, distribution
an empirical of
distribution
the stress change distribution can be created (Figure 20). Additionally, a matrix containing
of the stress change distribution can be created (Figure 20). Additionally, a matrix con-
taining all approximated functions (Fourier series) is described by Formula (5). The sim-
ulation began with a random selection of successive polynomials describing the mean
force value. This fit was next adjusted linearly to obtain signal continuity. The last step
was to add random amplitudes to which the inverse cumulative distribution method was
Figure 19. Data simulation scheme. Step 1: no data; step 2: simulating the average course of forces on the following days
on the basis of a randomly selected Fourier series from the created database; step 3: linear data fit correction; step 4: sim-
ulation of amplitudes using the inverse distribution function.
Sensors 2021, 21, 1927 23 of 27
Using data from the relevant months in the previous years, an empirical distribution
of the stress change distribution can be created (Figure 20). Additionally, a matrix con-
taining all approximated functions (Fourier series) is described by Formula (5). The sim-
all approximated
ulation functions
began with a random (Fourier series)
selection is described
of successive by Formuladescribing
polynomials (5). The simulation
the mean
began with a random selection of successive polynomials describing the
force value. This fit was next adjusted linearly to obtain signal continuity. meanThe
force
lastvalue.
step
This fit was next adjusted linearly to obtain signal continuity. The last
was to add random amplitudes to which the inverse cumulative distribution method step was to add
was
random
applied, amplitudes to which cumulative
using the empirical the inverse distribution
cumulative distribution
function justmethod
created.was
All applied,
calcula-
using the empirical
tions were cumulative
made in the distribution
MatLab software function
according just created.
to original All calculations were
scripts.
made in the Matlab software according to original scripts.
Figure 20.
Figure Empirical cumulative
20. Empirical cumulative distributions
distributions for
for stress
stress changes
changes in
in the
the shortest
shortest (LW-01)
(LW-01) and
and longest
longest cable
cable(LW-20).
(LW-20).
Figure
Figure 2121shows
showsananexample
exampleof of
a graph with
a graph missing
with and and
missing simulated data parts
simulated obtained
data parts ob-
by 24 of 29
tained by using the method specified in Figures 19 and 20. The chart obtained in way
using the method specified in Figures 19 and 20. The chart obtained in this can
this way
be used for further theoretical analyses. In a similar way, assuming different levels of
can be used for further theoretical analyses. In a similar way, assuming different levels of
the average tension of the cables, a series of data simulations for subsequent years were
the average tension of the cables, a series of data simulations for subsequent years were
performed.
performed. In Inthis way,DDparameters
thisway, parameterswere
werecalculated
calculated according
according to the formula
to the in Figure
formula 18.
in Figure
Table 13 summarises the calculated annual D parameters for measured and simulated
18. Table 13 summarises the calculated annual D parameters for measured and simulated (years
marked with *) data,
(years marked with respectively. First, monthly
*) data, respectively. First,values werevalues
monthly calculated
wereand summedand
calculated up
over the following years.
summed up over the following years.
Force diagram with missing data

111
110
109
108
107
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
Day
Force diagram with simulated data
111
110
109
108
107
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
Day
Figure 21.
Figure 21. LW-16
LW-16 force
force chart
chart in
in July
July 2016.
2016. Upper
Upper chart
chart with
with missing
missing data
dataand
andlower
lowerwith
withsimulated
simulatedone.
one.
From the results in Table 13, it is clear that the bridge was designed correctly, and
the stay cables are not exposed to fatigue damage from changes in tensioning forces.
Another sensitive point may be their anchors, where the influence of cable bending
should also be considered [3–5]. This is a separate research topic that is to be the subject
of future analyses. However, the data show that the greatest D parameter increment oc-
Sensors 2021, 21, 1927 24 of 27
From the results in Table 13, it is clear that the bridge was designed correctly, and the
stay cables are not exposed to fatigue damage from changes in tensioning forces. Another
sensitive point may be their anchors, where the influence of cable bending should also be
considered [3–5]. This is a separate research topic that is to be the subject of future analyses.
However, the data show that the greatest D parameter increment occurs in the LW-20 cable,
with the largest force/stress amplitude, and in the LW-16 stay with the highest average
tension force.
8. Characteristics of the Influence of Individual Loads on the Cable Tension Force

The analysis presented above shows how differentiated the influence of individual
loads is on the total tension force in stay cables. For the concrete cable-stayed structure,
which is the R˛edziński Bridge, its main component comes from its dead load.
Based on the observations, a cyclic change in the tension force in the cables was found.
It increases in the summer and decreases in the winter. Similar fluctuations occur in the
cycle of daily changes, with the difference being that during the day, there is a decrease
in strength and an increase at night. This phenomenon is a result of the deformation of
the pylon and the deck due to temperature changes. In daily cycles, the tension force is
affected by the deformation of the superstructure itself, while the difference in the mean
tension force between the summer and winter results from the mutual displacement of
the anchorages on the pylon and decks. Local and rapid changes in the tension force are
caused by the passage of heavy vehicles or by the overlapping effects of individual moving
loads. These types of effects are noticeable within the entire group of stay cables, located in
the load zone of the decks.
The authors realize that these are not all of the factors influencing the long-term stress
of cables in cable-stayed bridges. The influence of rheology on force redistribution, pylon
settlement [23] or the complicated issue of aerodynamics or seismic issues [24,25] remains
a problem. However, in this paper, the aim was to initially decode the values measured by
a specific SHM system, which in the coming years will provide a lot of information that
could complement the analysis that has already been started.
9. Discussion
Monitoring and continuous observation of bridge structures has become an important
part of the work of designers and scientists over the last few decades [1,7,15,17,26–30]. Both
new structures and old ones require supervision for maintenance purposes [31]. Structural
health monitoring (SHM) allows one to monitor the structure according to its purpose: old
bridges should be monitored in the places of identified damage [32], and the examination
of new bridges enables a real assessment of the structure’s behaviour. SHM system saves
a lot of information from a big number of sensors. The analysis of the results should be
approached with caution. First of all, on the basis of the measurements, the behaviour of
the whole structure should be assessed, and then it should be ensured that the results are
reliable. For this purpose, it is worth creating an auxiliary FEM model of the tested object to
verify the measured data, which has been shown in this paper. Measurements only carried
out with SHM systems, or theoretical calculations alone, do not allow a good verification of
the bridge behaviour and the assessment of its condition. Furthermore, the load test results
of the R˛edziński Bridge provided additional information to verify the FEM model. Such a
set of available research tools and methods can be useful for assessing the durability of the
bridge structure and its elements. However, continuous monitoring with the support of
SHM systems and data verification is a reliable source of information about the behaviour
of cable-stayed bridges, but a detailed analysis of data from SHM also allows one to create
an extensive statistical database. By adding calculations based on realisable FEM models,
unmeasured SHM values can be defined. The calculation of the D parameter based on the
available statistics and the implementation of advanced simulations ultimately allows one
to develop a trend of the force changes influence on the stay cable system.
Sensors 2021, 21, 1927 25 of 27
It can be stated that the creation of an efficient tool in the form of a coupled monitoring
system with a numerical model can be easily used in practice. The analysis of the collected
data presented in the article shows how complex the problem of determining the impact of
individual loads on the effort of the structure is. In the case of structures with a complex,
hyper static scheme, such as the described bridge, intuitive determination of the influences
may turn out to be wrong.
There are, however, exceptional situations that cannot be detected on the basis of
SHM measurements. Figure 22 shows the deformed sheath of the LZ-W11 stay cable.
This situation occurred as a result of a passenger vehicle fire near its bottom anchorage
zone. Unfortunately, no force sensor is installed in this stay cable. However, checking
the force values from neighbouring cables and their implementation into the numerical
model allowed for the estimation of the force after the accident in the damaged stay cable.
Ultimately, it could be concluded that, apart from the deformation of the cover, the internal
strands were not damaged.
Figure22.
Figure 22.LZ-W11
LZ-W11cable
cablesheath
sheathdeformed
deformedasasa aresult
resultofofa avehicle
vehiclefire.
fire.
AnAnalternative
alternativemethod
methodthatthatcan
canbe beused
usedfor
forthe
thedetermination
determinationofofa astay
staycable
cableforce
force
aredynamic
are dynamicmeasurements
measurementsby byusing,
using,e.g.,
e.g.,external
externalaccelerometers.
accelerometers.TheThepotential
potentialofofthis
this
methodisisa asubject
method subjectofofthethePolish–Portuguese
Polish–PortugueseNAWA-FCT
NAWA-FCTresearchresearchproject
project“Dynamic
“Dynamic
monitoringofofbridge
monitoring bridgestructures”,
structures”,supported
supportedby bythe
thePolish
PolishNational
NationalAgency
Agencyfor forAcademic
Academic
Exchangeand
Exchange andFundação
Fundaçãopara paraaaCiência
CiênciaeeaaTecnologia
TecnologiaininPortugal,
Portugal,that
thatisiscarried
carriedout,
out,
among others, on the Rędziński Bridge (Figure 23). Dynamic measurements
among others, on the R˛edziński Bridge (Figure 23). Dynamic measurements indicate the indicate the
wideapplicability,
wide applicability,easiness
easinessand
andpracticality
practicalitytotovalidate
validatethe
thedesign
designassumptions
assumptionsofofbridges
bridges
and
andassess
assessthethecondition
conditionofoftheir
theirstay
staycable
cablesystems
systems[9,14].
[9,14].This
Thismethod
methodcan cansupport
supportthethe
SHM
SHMsystem
system and
andthethe
FEM FEMmodel-based
model-based force analysis.
force It will
analysis. It be used
will be as the as
used future research
the future re-
direction on the R˛on
search direction edzi ński
the Bridge. Bridge.
Rędziński
It is worth adding that the analysis of SHM data from the first 10 years of the bridge’s
existence shows that it was designed correctly. Continuous observation of data from the
monitoring system ensures the safety of the structure and its users, despite minor accidents
such as a vehicle fire. Furthermore, the road administration (General Directorate for National
Roads and Motorways) receives annual reports based on measurements from the SHM
system supported by a constant FEM analysis.
Exchange and Fundação para a Ciência e a Tecnologia in Portugal, that is carried out,
among others, on the Rędziński Bridge (Figure 23). Dynamic measurements indicate the
wide applicability, easiness and practicality to validate the design assumptions of bridges
and assess the condition of their stay cable systems [9,14]. This method can support the
Sensors 2021, 21, 1927 SHM system and the FEM model-based force analysis. It will be used as the future 26 ofre-
27
search direction on the Rędziński Bridge.
Figure 23.
Figure Accelerometers used
23. Accelerometers used for
for determination
determination of
of aa stay
stay cable
cable force.
force.
It is
Author worth adding
Contributions: that the analysis
Conceptualization, P.H.of
andSHMM.T.;data from theM.T.;
methodology, firstsoftware,
10 years of vali-
M.T.; the
dation, M.T.;
bridge’s formalshows
existence analysis, M.T.;
that investigation,
it was designedJ.B., P.H. andContinuous
correctly. M.T.; resources, J.B., P.H. and
observation M.T.;
of data
data curation,
from M.T.; writing—original
the monitoring system ensures draft the
preparation,
safety ofM.T.;
thewriting—review
structure andand its editing, P.H. and
users, despite
M.T.; visualization, P.H. and M.T.; supervision, J.B. and P.H.; project administration,
minor accidents such as a vehicle fire. Furthermore, the road administration (General J.B. and P.H.;
funding acquisition, J.B. and P.H. All authors have read and agreed to the published version of the
Directorate for National Roads and Motorways) receives annual reports based on meas-
manuscript.
urements from the SHM system supported by a constant FEM analysis.
Funding: The research project “Dynamic monitoring of bridge structures”, mentioned in Section 9, is
co-financed by the Polish National Agency for Academic Exchange (grant number: PPN/BIL/2018/1/
00235/U/00001).
Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.
Data Availability Statement: The data presented in this study are available on request from the
corresponding authors.
Acknowledgments: The authors are indebted to the General Directorate for National Roads and
Motorways in Wrocław, Poland for providing access to data collected by the SHM system of the
R˛edziński Bridge since 2011. The authors would like also to thank the Mosty-Wrocław (design office),
Freyssinet Poland (installation of stay cable system) and NeoStrain (SHM system supplier) for their
good cooperation and valuable technical support. This joint collaboration enabled this research study,
the practical experience and the scientific development.
Conflicts of Interest: The authors declare no conflict of interest.
References
1. Wenzel, H. Health Monitoring of Bridges; John Wiley & Sons Ltd.: Chichester, UK, 2009.
2. Maljaars, J.; Vrouwenvelder, T. Fatigue failure analysis of stay cables with initial defects: Ewijk bridge case study. Struct. Saf.
2014, 51, 47–56. [CrossRef]
3. Tokujama, S.; Ishihara, S.; Taniyama, S. Full Scale Fatigue Tests for Stay Cable Systems; ETH Zürich: Zürich, Switzerland, 1995.
[CrossRef]
4. Winkler, J.; Georgias, C.; Fischer, G.; Wood, S.; Ghannoum, W. Structural response of a multi-strand stay cable to cyclic bending
load. Struct. Eng. Int. 2015, 2, 141–150. [CrossRef]
5. Winkler, J. Parallel Monostrand Stay Cable Bending Fatigue. Ph.D. Thesis, Technical University of Denmark, Copenhagen,
Denmark, 2014.
6. Biliszczuk, J.; Hawryszków, P.; Teichgraeber, M. Structural health monitoring system of a concrete cable-stayed bridge. Archit.
Civ. Eng. Environ. 2018, 11, 69–77. [CrossRef]
7. Biliszczuk, J.; Barcik, W.; Onysyk, J.; Toczkiewicz, R.; Tukendorf, A.; Tukendorf, K. R˛edziński Bridge in Wrocław—The largest
concrete cable-stayed bridge in Poland. Struct. Eng. Int. 2014, 24, 285–292. [CrossRef]
Sensors 2021, 21, 1927 27 of 27
8. Hawryszków, P.; Hildebrand, M. Installation of the largest stay cable system in Poland—the R˛edziński bridge in Wrocław.
In Proceedings of the 18th IABSE Congress On Innovative Infrastructures—Toward Human Urbanism, Seoul, Korea,
19–21 September 2012. [CrossRef]
9. Hawryszków, P. Investigation of tension forces in a stay cable systems of a road bridge using vibration methods. In Proceedings
of the EVACES’15, 6th International Conference on Experimental Vibration Analysis for Civil Engineering Structures, MATEC
Web of Conferences, Zurich, Switzerland, 19–21 October 2015. [CrossRef]
10. Bień, J.; Kużawa, M.; Kamiński, T. Validation of numerical models of concrete box bridges based on load test results. Arch. Civ.
Mech. Eng. 2015, 15, 1046–1060. [CrossRef]
11. Hildebrand, M.; Nowak, Ł. Measurement of temperature distribution within steel box girder of Vistula River bridge in Central
Europe. Balt. J. Road Bridge Eng. 2020, 15, 71–95. [CrossRef]
12. Kang, F.; Li, J.; Zhao, S.; Wang, Y. Structural health monitoring of concrete dams using long-term air temperature for thermal
effect simulation. Eng. Struct. 2019, 180, 642–653. [CrossRef]
13. Kang, F.; Li, J.; Dai, J. Prediction of long-term temperature effect in structural health monitoring of concrete dams using support
vector machines with Jaya optimizer and salp swarm algorithms. Adv. Eng. Softw. 2019, 131, 60–76. [CrossRef]
14. Caetano, E. Cable Vibrations in Cable-Stayed Bridges, 9th ed.; IABSE International Association for Bridge and Structural Engineering:
Switzerland, Zürich, 2007.
15. Gimsing, N.J.; Georgakis, C.T. Cable Supported Bridges: Concept and Design; John Wiley & Sons, Ltd.: Chichester, UK, 2011.
16. Greco, F.; Lonetti, P.; Pascuzzo, A. Dynamic Analysis of Cable-Stayed Bridges Affected by Accidental Failure Mechanisms under
Moving Loads. Math. Probl. Eng. 2013, 2013, 302706. [CrossRef]
17. Mankar, A.; Bayane, I.; Sørensen, J.D.; Brühwiler, E. Probabilistic reliability framework for assessment of concrete fatigue of
existing RC bridge deck slabs using data from monitoring. Eng. Struct. 2019, 201. [CrossRef]
18. Szala, G. Application of two-parameter fatigue characteristics in fatigue persistence calculations of structural components under
conditions of a broad spectrum of loads. Pol. Marit. Res. 2016, 4, 138–145. [CrossRef]
19. Heywood, R.B. Designing Against Fatigue; Chapman Hall: London, UK, 1962.
20. Schijve, J. Fatigue of Structures and Materials; Springer: Dordrecht, The Netherlands, 2009.
21. Panontin, T.; Sheppard, S. (Eds.) Fatigue and Fracture Mechanics; ASTM International: West Conshohocken, PA, USA, 1999;
Volume 29. [CrossRef]
22. Serensen, S.V.; Kogayev, V.P.; Shnejderovich, R.M. Permissible Loading and Strength Calculations of Machine Components; Maschinos-
troenie: Mosow, Russia, 1975.
23. Winkelmann, K.; Żyliński, K.; Górski, J. Probabilistic analysis of settlements under a pile foundation of a road bridge pylon. Soils
Found. 2020, 61, 80–94. [CrossRef]
24. Jang, M.; Lee, Y.; Won, D.; Kang, Y.-J.; Kim, S. Static Behaviors of a Long-span Cable-Stayed Bridge with a Floating Tower under
Dead Loads. J. Mar. Sci. Eng. 2020, 8, 816. [CrossRef]
25. Seeram, M.; Manohar, Y. Two-Dimensional Analysis of Cable Stayed Bridge under Wave Loading. J. Inst. Eng. India Ser. A 2018,
99, 351–357. [CrossRef]
26. Bayane, I.; Brühwiler, E. Structural condition assessment of reinforced-concrete bridges based on acoustic emission and strain
measurements. J. Civ. Struct. Health Monit. 2020, 10, 1037–1055. [CrossRef]
27. Song, G.; Wang, C.; Wang, B. Structural Health Monitoring (SHM) of Civil Structures. Appl. Sci. 2017, 7, 789. [CrossRef]
28. Svensson, H. Cable-Stayed Bridges, 40 Years of Experience Worldwide; Ernst und Sohn: Berlin, Germany, 2012.
29. Chakraborty, J.; Katunin, A.; Klikowicz, P.; Salamak, M. Embedded ultrasonic transmission sensors and signal processing
techniques for structural change detection in the Gliwice bridge. Procedia Struct. Integr. 2019, 17, 387–394. [CrossRef]
30. Wang, X.; Chakraborty, J.; Niederleithinger, E. Noise Reduction for Improvement of Ultrasonic Monitoring Using Coda Wave
Interferometry on a Real Bridge. J. Nondestruct. Eval. 2021, 40, 14. [CrossRef]
31. Sheer, J. Failed Bridges: Case Studies, Causes and Consequences; Ernst und Sohn: Berlin, Germany, 2010.
32. Hui, L.; Shunlong, L.; Jinping, O.; Hongwei, L. Reliability assessment of cable-stayed bridges based on structural health
monitoring techniques. Struct. Infrastruct. Eng. 2012, 8, 829–845.

Sensors: SHM System and A FEM Model-Based Force Analysis Assessment in Stay Cables

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

Sensors: SHM System and A FEM Model-Based Force Analysis Assessment in Stay Cables

Uploaded by

Copyright:

Available Formats

sensors

Keywords: bridges; stay cables; durability; maintenance; monitoring; SHM; FEM

Citation: Biliszczuk, J.; Hawryszków,

Sensors 2021, 21, 1927. https://doi.org/10.3390/s21061927 https://www.mdpi.com/journal/sensors

Sensors 2021, 21, 1927 2 of 27

2.2.The R˛edziński Bridge

Sensors 2021, 21, x FOR PEER REVIEW 4 of 29

Table 2. Examples of sensor markings.

Sensors measuring the tension force are

Sensors 2021, 21, x FOR PEER REVIEW 6 of 29

After the transformations, the following equation may be obtained:

3. The FEM Models

1. 0.25 0.25

2. 0.32 0.31

2. 0.32 0.31

3. 0.50 0.48

3. 3. 0.500.50 0.48 0.48

Sensors 2021, 21, x FOR PEER REVIEW 10 of 31

4. 0.51 0.48

4. Comparison of the Measured Data

Lowest Daily Average Highest Daily Average

Annual extreme temperature chart at selected points in the structure

Sensors 2021, 21, x FOR PEER REVIEW 11 of 29

Figure 8. Strand loading scheme: ΔT—uniform temperature change; di,j—theoretical anchorages

Figure 9. Daily strands temperature change (1 August 2017).

Figure 9. Daily strands temperature change (1 August 2017).

Force in strand [kN]

Figure 13. Screenshots

Sensors 2021, 21, 1927

Force Change Read from the

5. Cable Forces Analysis

F(t) = g(t) + T(t) + q(t) + w(t), (3)

Table 10. Sample parameters from 17 August 2018.

Cable SSE R-Sqr RMSE

LZ-W01 5.4 0.976 0.141

Figure 16. Graphs

In order to compare this classical method of counting cycles, comparative calculations

Table 11. Parameters Da and Ds in year 2018.

Stay Cable Ds Da Ds /Da

Table 13. Values

Table 13. Values of the failure accumulation parameter D.

Year Cable LW-20 Cable LW-16 Cable LW-01

Force diagram with missing data

8. Characteristics of the Influence of Individual Loads on the Cable Tension Force

You might also like