Analysis of The Significance of Changes in The Num PDF

materials
Article
Analysis of the Significance of Changes in the Number and
Energy Parameters of Acoustic Emission Signals on the
Assessment of the Strength of Fibre–Cement Boards
Anna Adamczak-Bugno 1, * , Grzegorz Świt 1 , Aleksandra Krampikowska 1 and Edoardo Proverbio 2
1 Faculty of Civil Engineering and Architecture, Kielce University of Technology, 25-314 Kielce, Poland
2 Department of Engineering, University of Messina, Contrada di Dio Sant’Agata, 98166 Messina, Italy
* Correspondence: [email protected]
Abstract: The article presents the results of three-point bending tests carried out for samples cut
from full-size fibre–cement boards subjected to typical and exceptional conditions. The tests were
carried out with the simultaneous acquisition of acoustic emission signals. It has been noted that
some factors significantly deteriorate the strength parameters of the samples as well as cause the
occurrence of differences in the number of acoustic emission signals of various classes and their
energy parameters. A statistical analysis was carried out in order to repeat the relationship between
the strength parameters of the samples and the acoustic emission parameters. Based on the research,
it was found that the MOR bending strength for specimens exposed to fire and high temperature is
more than 50% lower than for air-dried specimens and specimens exposed to water. The increased
number of freeze–thaw cycles also has an impact on the strength of the specimens. Components
exposed to more than 10 freeze–thaw cycles had a strength more than 30% smaller than the reference
specimens soaked in water and exposed to bath-drying cycles. A similar dependency was indicated
Citation: Adamczak-Bugno, A.; Świt, by the number of signals of the individual classes, their energy parameters and their frequencies.
G.; Krampikowska, A.; Proverbio, E. The number, strength, duration and frequency also decreased along with the increase in the test case
Analysis of the Significance of number. On this basis, conclusions were drawn concerning the suitability of acoustic emission for the
Changes in the Number and Energy evaluation of the strength of fibre–cement elements.
Parameters of Acoustic Emission
Signals on the Assessment of the Keywords: fibre–cement composites; acoustic emission method; statistical analysis
Strength of Fibre–Cement Boards.
Materials 2022, 15, 5757. https://
doi.org/10.3390/ma15165757
Academic Editor: Jacek Domski 1. Introduction

1.1. Description of Fibre–Cement Boards
Received: 28 July 2022
Accepted: 17 August 2022
Fibre cement (Lat. fibro–fibre and caementum–cement) is a construction material
Published: 20 August 2022
consisting of cement or calcium silicate (formed by the chemical reaction of silicate and
calcium materials) and mineral fillers, reinforced with fibres (randomly dispersed or contin-
Publisher’s Note: MDPI stays neutral
uous strands and tapes or meshes and fabrics). Fibre–cement boards have high mechanical
with regard to jurisdictional claims in
strength (the minimum bending strength of flat cement–fibre boards for wall cladding is
published maps and institutional affil-
4 MPa), flexibility and durability [1–5].
iations.
Contemporary fibre–cement products (sold legally on the EU market) are harmless to
people. In most cases, fibre–cement cladding is resistant to corrosion, rot and fungi as well
as UV radiation.
Copyright: © 2022 by the authors.
Fibre–cement boards are manufactured using pressing and autoclaves. The press-
Licensee MDPI, Basel, Switzerland. ing pressure is approximately 650 N/cm2 . After the pressing cycle, the boards pass the
This article is an open access article curing stage within 6–8 h and are then placed in autoclaves, where they finally harden
distributed under the terms and at a high temperature of 175 ◦ C and under the pressure of 10 atm. Thanks to this tech-
conditions of the Creative Commons nology, fibre–cement boards have high mechanical strength and bending strength. Some
Attribution (CC BY) license (https:// manufacturers of fibre–cement cladding use recycled materials during production [5–8].
creativecommons.org/licenses/by/ The appearance of fibre–cement cladding may vary owing to the virtually limitless
4.0/). options regarding colour, texture and size. This fact, combined with the technical and
Materials 2022, 15, 5757. https://doi.org/10.3390/ma15165757 https://www.mdpi.com/journal/materials

Materials 2022, 15, 5757 2 of 23
functional performance, is the reason the use of fibre–cement rainscreen cladding is not
limited by regional architectural traditions, climatic conditions or intended use of the
building. Fibre–cement cladding may be full-body coloured, made with an outer textured
layer (imitating wood or stone), painted after installation or even covered with render.
That is why rainscreen cladding using fibre–cement cladding can be found on new and
reconstructed buildings of any type and with any function (apartment buildings and
single-family houses, office buildings, industrial facilities, hospitals, etc.) [9–12].
End-of-life fibre–cement cladding can be recycled.
Initially, fibre–cement products were made using asbestos. The first prototypes of
fibre–cement boards were manufactured towards the end of the 19th century by Ludwig
Hatschek from Austria as a mixture of cement and asbestos. Around the end of the 19th
century, there was a need for materials that would be less flammable, less expensive,
more resistant to variable temperatures, stronger and more durable than the conventional
construction materials known at the time. The first Austro-Hungarian asbestos factory
(owned by Ludwig Hatschek) manufactured fibre–cement roof cladding. The production
of wall cladding followed soon after.
By 1976, fibre–cement cladding was mostly made using asbestos. After it was es-
tablished that asbestos was harmful to people, approx. 200 types of fibres (e.g., basalt,
cellulose) were started to be used to make fibre–cement cladding [7,10,13].
1.2. Application and Operating Conditions of Fibre–Cement Boards

Fibre–cement cladding is an essential component of rainscreen cladding, which is
growing increasingly popular both with designers and clients. The popularity of such
facades is owed to aesthetic considerations (a very wide range of exterior cladding that
enables the adaptation of building facades to any urban environment), ease of installa-
tion, competitive prices, thermal insulation performance and inexpensive maintenance.
Rainscreen cladding is used on all types of buildings with various functions, fit-out spec-
ifications and furnishings. Such buildings include large shopping centres, sports and
recreational facilities, railway stations, high-end high-rise office buildings, small single-
family houses, etc. [14,15].
Rainscreen cladding is increasingly popular with clients and contractors and is be-
coming more widespread. Unfortunately, this increases the number of observed failures of
such solutions.
Typical rainscreen cladding faults include the separation of the cladding from the
support framework, which is particularly dangerous to people and property if it occurs at
a large height.
Rainscreen cladding has very high durability and reliability, but it requires continu-
ous monitoring to correctly assess its condition. The absence of monitoring and correct
diagnostics of rainscreen cladding often causes it to wear out and suffer damage.
Rainscreen cladding, regardless of the materials it is made of, has varying resistance
to operating conditions. The most aggressive factors that adversely affect the safety of use
of wall-cladding systems include climatic and environmental factors and anthropogenic
(human-related) factors [16–19].
Climatic and environmental factors include:
• Temperature variations
During use, exterior rainscreen cladding is exposed to significant temperature vari-
ations (daily and seasonal), which cause stresses and linear strain in the cladding due to
thermal expansion. Temperature variations during the day may reach more than 40 ◦ C,
and they often cause the cladding to crack and separate from the support framework (in
the case of adhesive cladding).
• Thermal shock
Fibre–cement cladding used as rainscreen cladding is exposed to thermal shocks,
i.e., sudden temperature variations during abrupt weather changes. Dark-coloured cladding
Materials 2022, 15, 5757 3 of 23
may heat up to more than 90 ◦ C during the summer season. When torrential rains start (the
water temperature may be approx. 10 ◦ C), the exterior parts of the cladding are cooled im-
mediately. The high difference between the temperatures out-side and inside the cladding
often results in the appearance of cracks.
• Impact of moisture on rainscreen cladding (rain and snow on the outside and water
vapour migration on the inside)
Changes in moisture content affect the shape and dimensions of the cladding. The
unavoidable factors that increase water migration into the cladding system include wind.
During rainfall, the suction caused by the wind causes the individual parts of the cladding
to move and form corridors for the migration of water into the system. If the plastic vapour
barrier is damaged or missing, the thermal insulation may become permanently damp.
• Chemical air pollution and moisture
Serious damage to the cladding, grids and mechanical fasteners may be caused by
chemical reactions occurring between the chemical elements in the air (due to industrial and
transport pollution, from combined heat and power plants, etc.) and moisture (water). Such
reactions occur in situations where chemical elements in the air dissolve in the moisture
penetrating into the cladding or settling on the metallic support frames.
• Wind actions
Wind fluctuations cause the appearance of long-time, recurring pushing-pulling load,
resulting in progressive damage to “cladding-framework” joints.
The most dangerous anthropogenic actions include impacts. Fibre–cement cladding
has a fairly low impact strength. Impact with a hard object (e.g., a rock) or soft object (e.g., a
ball used by children to play) may cause the cladding to crack. Even microcracks may allow
moisture into the cladding, accelerating its degradation.
These factors have a significant impact on the safety of the use of wall-cladding
systems, which is why such systems should be systematically monitored to ensure long
and safe use [20–23].
1.3. Diagnostic Methods Used to Monitor Fibre–Cement Cladding

The mechanical assessment of fibre–cement boards after exposure to environmen-
tal and accidental operating factors is an important issue both from the scientific and
practical perspectives.
So far, the vast majority of research on fibre–cement boards has focused only on
determining the normative physical and mechanical parameters. The studies determined
the impact of operating conditions on the material, including the cycles of heat and rain
exposure, freezing and thawing, water-soaking and drying and high and low temperatures,
and the composition was modified primarily with respect to the type of the fibres. The
mechanical parameters of the boards were determined primarily by means of bending-
strength tests [24].
The most well-covered topic in the literature is mainly the impact of high temperatures
on concrete components and related processes [25]. The significance of this accidental factor
was often tested using non-destructive methods (NDT), primarily using the ultrasonic
method and acoustic emission [22,23].
Acoustic methods are advanced enough to also track the frequencies where the recorded
acoustic emission (AE) signals appear by using the Fourier transform and wavelet analysis [24].
The acoustic emission (AE) method is one of the popular methods for monitoring
civil engineering structures such as bridges, reinforced concrete and steel structures and
members, steel and plastic pipelines, compressed gas tanks, internal combustion engines
and power transformers. This procedure belongs to the group of passive methods, meaning
that AE devices do not emit any signals and do not affect the physical condition of the
object under test, only recording the physical effects appearing spontaneously in the moni-
Materials 2022, 15, 5757 4 of 23
tored object. The sources of acoustic emission signals include appearing and propagating
microcracks, cracks and corrosion processes.
An indirect objective of the research described in this paper was to analyse the changes
in the mechanical parameters of fibre–cement boards exposed to the previously mentioned
typical and accidental operating conditions. An important part of the testing procedure
was to analyse the acoustic emission signals recorded during the loading of material
samples. The primary objective of the research, in turn, was to examine the correlation of
the mechanical parameters of fibre–cement objects with the parameters of the recorded
acoustic emission signals. This enabled the assessment of the suitability of the acoustic
emission method as a potential reliable diagnostic tool to assess the condition of rainscreen
cladding made of fibre–cement boards [25–28].
The novelty of the research consists mainly in the use of multicriteria analysis of acous-
tic emission signals to monitor the condition of fibre–cement panels. The use of dividing the
recorded acoustic emission signals into classes is a much more reliable approach than the
commonly used inference about the state of various materials based on single parameters.
Establishing a statistical correlation between the mechanical and acoustic parameters for
panels subjected to typical and exceptional conditions will potentially make it possible to
use the acoustic emission method to assess the condition of panels at low load levels.
2. Materials and Methods

The analysis of the significance of the correlation between changes in the number and
energy parameters of acoustic emission signals and the strength of fibre–cement boards
was performed for the following test cases:
• air-dried specimens (F1 )—reference samples stored under constant laboratory condi-
tions (+23 ◦ C, 60% humidity);
• specimens soaked in water for 1 h (F2 )—samples immersed in water at room tempera-
ture (about 23 ◦ C) for 1 h and subjected to a wet bend test;
• specimens soaked in water for 24 h (F3 )—samples immersed in water at room temper-
ature (about 23 ◦ C) for 24 h and subjected to a wet bend test;
• specimens after 25 bath-drying cycles (F4 )—samples immersed in water at an ambient
temperature above 5 ◦ C (approximately 23 ◦ C) for 18 h and dried in a ventilated oven
at 60 ◦ C (±5 ◦ C), with a relative humidity of less than 20% for 6 h for 25 cycles;
• specimens after 50 bath-drying cycles (F5 )—samples immersed in water at an ambient
temperature above 5 ◦ C (approximately 23 ◦ C) for 18 h and dried in a ventilated oven
at 60 ◦ C (±5 ◦ C), with a relative humidity of less than 20% for 6 h for 50 cycles;
• specimens after 10 freeze–thaw cycles (F6 )—samples frozen in a freezer at a tempera-
ture of −20 ◦ C (±2 ◦ C) for 2 h, kept at this temperature for another hour, thawed in
a water bath at 20 °C (±2 ◦ C) for two hours and kept at this temperature for another
hour; repeated for 10 cycles;
ture of −20 ◦ C (±2 ◦ C) for 2 h, kept at this temperature for another hour, thawed in
a water bath at 20 °C (±2 ◦ C) for two hours and kept at this temperature for another
ture of −20 ◦ C (±2 ◦ C) for 2 h, kept at this temperature for another hour, thawed in a
water bath at 20 ◦ C (±2 ◦ C) for two hours and kept at this temperature for another
ture of −20 ◦ C (±2 ◦ C) for 2 h, kept at this temperature for another hour, thawed in a
water bath at 20 ◦ C (±2 ◦ C) for two hours and kept at this temperature for another
• specimens exposed to direct flame (heating the material to 400 ◦ C) for 2.5 min (F10 );
• specimens exposed to direct flame (heating the material to 400 ◦ C) for 5 min (F11 );
• specimens exposed to direct flame (heating the material to 400 ◦ C) for 7.5 min (F12 );
Recipes of fibre–cement boards and their technological process are strictly protected
by producers, so information on specific ingredients, their quantities and suppliers as well
as production details are severely limited. The tested fibre–cement boards were made with
the use of basic components such as: CEM I 42.5N Portland cement, cellulose fibers and
Materials 2022, 15, 5757lime flour additives. The plates were produced using the Hatschek process. 5 of 23
The acoustic emission method was used to assess the development of acoustic emis-
sion descriptors depending on the change of the mechanical parameters of fibre–cement
boards during• aspecimens
three-point bending
exposed test for
to direct previously
flame (heatingprepared specimens.
the material Figure
to 400 ◦ C) for 101min (F13 );
shows a scheme
• and photograph of the testing station. Dimensions
◦ are given in millime-
specimens exposed to a temperature of 230 C for 3 h (F14 )—samples fired in a labora-
ters. Markings: ltory
bs-distance
furnacebetween
to degradethe cellulose
axes of sensors,
fibers. ls-distance between the axes of
support and le-total length of the sample.
Each group consisted of 10 samples.
Acoustic emission
Recipes of measurements
fibre–cement wereboardsperformed using two frequency
and their technological process are sensors:
strictly protected
VS30-SIC with a flat characteristic in the range of 25–80 kHz and VS150-RIC
by producers, so information on specific ingredients, their quantities and with a meas-
suppliers as well
uring rangeasofproduction
100–450 kHz andare
details a peak
severelyat 150 kHz.The
limited. Thetested
AEWin procesoorboards
fibre–cement of acoustic
were made with
emission (Physical
the use Acoustic Corporation,such
of basic components Westas:Windsor, NJ, USA)
CEM I 42.5N was used
Portland in the
cement, tests. fibers and
cellulose
The fibre–cement compositeThe
lime flour additives. was subject
plates wereto produced
bending strength
using thetests using process.
Hatschek the Zwick
Roell strength testing machine with a load range from 0 to 10 kN.
The acoustic emission method was used to assess the developmentThe tests of the
of fibre–
acoustic emission
cement specimens were performed with a constant traverse speed of 0.1 mm/min.
descriptors depending on the change of the mechanical parameters of fibre–cement The boards
digital signals were processed using Vallen VisualAE (Vallen GMBH, Wolfratshausen,
during a three-point bending test for previously prepared specimens. Figure 1 shows
Germany) and Vallenand
a scheme VisualClass (VisualAE
photograph of the (Vallen
testing GMBH,
station. Wolfratshausen,
Dimensions are Germany)
given in millimeters.
software forMarkings:
the analysis of AE signals.
lbs -distance between the axes of sensors, ls -distance between the axes of support
and le -total length of the sample.
Materials 2022, 15, x FOR PEER REVIEW 6 of 24
(a)
(b)
Figure 1. Test stand: (a) scheme; (b) photograph.
Figure 1. Test stand: (a) scheme; (b) photograph.
3. Results Acoustic emission measurements were performed using two frequency sensors: VS30-
The AESICsignals
withrecorded during three-point
a flat characteristic in the bending
range oftests were
25–80 kHzbroken down into with a measuring
and VS150-RIC
classes using the k-means
range algorithm.
of 100–450 kHz and Theagrouping identified
peak at 150 four AEWin
kHz. The classes that were as-of acoustic emission
procesoor
signed to the processes occurring in the reinforced cement material under
(Physical Acoustic Corporation, West Windsor, NJ, USA) was used in load, based onthe tests.
the authors own prior research and data in the literature [11]:
The fibre–cement composite was subject to bending strength tests using the Zwick Roell
Class 1—commencement of microcracks;
strength testing machine with a load range from 0 to 10 kN. The tests of the fibre–cement
Class 2—development of a mesh and increase or crack width;
Class 3—delamination of the material and debonding of fibres;
Class 4—breaking and material failure.
The analysis of the test results included the monitoring of the data concerning the
number of signals of the individual classes for successive test cases, their strength and
Materials 2022, 15, 5757 6 of 23
specimens were performed with a constant traverse speed of 0.1 mm/min. The digital
signals were processed using Vallen VisualAE (Vallen GMBH, Wolfratshausen, Germany)
and Vallen VisualClass (VisualAE (Vallen GMBH, Wolfratshausen, Germany) software for
the analysis of AE signals.
3. Results
The AE signals recorded during three-point bending tests were broken down into
classes using the k-means algorithm. The grouping identified four classes that were assigned
to the processes occurring in the reinforced cement material under load, based on the
authors own prior research and data in the literature [11]:
Class 1—commencement of microcracks;
Class 2—development of a mesh and increase or crack width;
Class 3—delamination of the material and debonding of fibres;
Class 4—breaking and material failure.
The analysis of the test results included the monitoring of the data concerning the
number of signals of the individual classes for successive test cases, their strength and
Materials 2022, 15, x FOR PEER REVIEW duration and average frequencies. The analysis also concerned the strengths achieved 7 of 25 for
the individual specimens.
The analysis was carried out using IBM SPSS Statistics 26 (IBM, New York, NY, USA).
The significance level was adopted as 0.05. The normality of the distributions was verified
homogeneity in most cases, the average distributions were compared using a group of
using the Shapiro–Wilk test, and variance homogeneity was tested using the Levene test.
nonparametric tests for the independent variables, including, in particular, the Kruskal–
Since there was no normal distribution for some of the data and there was no variance
Wallis test for multiple groups. The dependency between the data was examined using
homogeneity in most cases, the average distributions were compared using a group of non-
the Spearman’s correlation coefficient due to the absence of normal distribution in the
parametric tests for the independent variables, including, in particular, the Kruskal–Wallis
data.
test for multiple groups. The dependency between the data was examined using the Spear-
man’s correlation coefficient due to the absence of normal distribution in the data.
3.1. Distribution of the Number and Selected Energy Parameters of Acoustic Emission Signals
and3.1.
Strength for the Individual
Distribution Testand
of the Number Cases
Selected Energy Parameters of Acoustic Emission Signals and
Strength
3.1.1. Number for of
theClass-1
Individual Test Cases
Signals
3.1.1.
The Number
graphicalofrepresentation
Class-1 Signals
of Kruskal–Wallis test results for independent samples
The graphical
of the number representation
of class-1 of Kruskal–Wallis
signals (Figure test results
2) indicates that for independent
the highest number of samples
class-1 of
the number of class-1 signals (Figure 2) indicates that the highest number
signals was recorded for components from test case F7 (specimens exposed to 25 freeze– of class-1 signals
was recorded for components from test case F (specimens exposed to 25
thaw cycles). Moreover, for this case, the scatter of 7the results was the greatest. Test casefreeze–thaw
cycles). Moreover,
F8 (specimens exposed to for50this case, the scatter
freeze–thaw cycles) of the results
contains was the
individual greatest.
outliers. TheTest case F8
smallest
(specimens
number exposed
of class-1 signalstowas
50 freeze–thaw
recorded forcycles)
case F14contains
(specimensindividual
fired in outliers. The smallest
a furnace).
number of class-1 signals was recorded for case F14 (specimens fired in a furnace).
Figure 2. Graphic presentation of the Kruskal–Wallis test results for independent samples: number
of class-1 signals.
of class-1 signals.
3.1.2. Average Strength of Class-1 Signals [nVs]

The graphical representation of Kruskal–Wallis test results for independent samples
of the average strength of class-1 signals (Figure 3) indicates that the highest strength of
class-1 signals was recorded for components from test case F4 (specimens exposed to 25
Materials 2022, 15, 5757 7 of 23

of the average strength of class-1 signals (Figure 3) indicates that the highest strength
of class-1 signals was recorded for components from test case F4 (specimens exposed to
Materials 2022, 15, x FOR PEER REVIEW 25 bath-drying cycles). Moreover, for this case, the scatter of the results was the8 greatest.
of 25
The smallest average strength of class-1 signals was recorded for case F12 (specimens ignited
for 7.5 min).
Figure 3. Graphic presentation of the Kruskal–Wallis test results for independent samples: average
strength of3.class-1
Figure signals.
Graphic presentation of the Kruskal–Wallis test results for independent samples: average
strength of class-1 signals.
3.1.3. Average Duration of Class-1 Signals [µs]
3.1.3. The graphical
Average representation
Duration of Kruskal–Wallis
of Class-1 Signals [µs] test results for independent samples
of the The graphical
average duration representation
of class-1 of Kruskal–Wallis
signals (Figure 4) test results
indicates that forlongest
independent samples
The
of the graphical
average representation
duration of of Kruskal–Wallis
class-1 signals (Figuretest4)results forthe
indicates independent
that the
average
longest
du-
samples
average
ration
of the of class-1 signals
averageofduration was recorded
of class-1 signals for components
(Figure from test case F 7 (specimens ex-
duration class-1 signals was recorded for4)components
indicates that the test
from longest
caseaverage du-
F7 (specimens
posed of
ration to class-1
25 freeze–thaw cycles). Moreover,components
for this case,fromthe scatter of the results was ex-
the
exposed to 25signals was
freeze–thaw recorded
cycles). for
Moreover, for this test
case, thecase F7 (specimens
scatter of the results was
greatest.
posed to Test
25 cases
freeze–thaw F 1 (air-dried specimens), F3 (specimens soaked in water for 24 h) and
cycles). Moreover, for this case, the scatter of the results was the
the greatest. Test cases F1 (air-dried specimens), F3 (specimens soaked in water for 24 h)
F11 (specimens
greatest. ignited for 5 min)specimens),
contain individual outliers. The shortest
in water average
for 24 h)dura-
and FTest cases F1 (air-dried F3 (specimens soaked
11 (specimens ignited for 5 min) contain individual outliers. The shortest average
and
tion of class-1
F11 (specimens signals
ignited was
for recorded
5 min) for
contain case F 10 (specimens ignited for 2.5 min).
individual outliers. The shortest average dura-
duration of class-1 signals was recorded for case F10 (specimens ignited for 2.5 min).
tion of class-1 signals was recorded for case F10 (specimens ignited for 2.5 min).
Figure
Figure 4. Graphic
4. Graphic presentation
presentation of the
of the Kruskal–Wallis
Kruskal–Wallis testtest results
results for for independent
independent samples:
samples: average
average
duration
duration of class-1
of class-1 signals.
signals.
duration
3.1.4.ofNumber
class-1 signals.
of Class-2 Signals
3.1.4. Number of Class-2 Signals
3.1.4. The graphical representation of Kruskal–Wallis test results for independent samples
ofNumber of Class-2
the number Signals
of class-2 signals (Figure 5) indicates that the highest number of class-2
of the number of class-2 signals (Figure 5) indicatestest
The graphical representation of Kruskal–Wallis that the highest
results number of
for independent class-2
samples
signals
of was recorded
the number forsignals
of class-2 components
(Figurefrom test case that
5) indicates F4 (specimens
the highestexposed
numbertoof25 bath-
class-2
drying cycles). The largest scatter of results occurred for group F 6 (specimens exposed to
signals was recorded for components from test case F4 (specimens exposed to 25 bath-
10 freeze–thaw
drying cycles).
cycles). The Test
largest case Fof
scatter 2 (specimens soaked in water for 1 h) contains individ-
results occurred for group F6 (specimens exposed to
ual outliers. The smallest number of class-2 signals was recorded for case F14 (specimens
Materials 2022, 15, 5757 8 of 23
signals was recorded for components from test case F4 (specimens exposed to 25 bath-
Materials 2022, 15, x FOR PEER REVIEWdrying cycles). The largest scatter of results occurred for group F6 (specimens exposed
9 of 25 to
Materials 2022, 15, x FOR PEER REVIEW10 freeze–thaw cycles). Test case F2 (specimens soaked in water for 1 h) contains individual
9 of 25
outliers. The smallest number of class-2 signals was recorded for case F14 (specimens fired
in a furnace).
of class-25.signals.
Figure Graphic presentation of the Kruskal–Wallis test results for independent samples: number
of class-2 signals.
of class-2 signals.
3.1.5. Average
3.1.5. Average Strength
Strengthof Class-2
of Class-2Signals [nVs]
Signals [nVs]
3.1.5.The
Average
graphicalStrength of Class-2
representation Signals
of [nVs]
Kruskal–Wallis
The graphical representation of Kruskal–Wallis test results
test resultsfor
forindependent
independentsamples
samples of
of the average
theThe strength
graphical
average of class-2
representation
strength signals (Figure
of Kruskal–Wallis
of class-2 signals 6) indicates
test results
(Figure 6) indicates that the
for highest
that the highest
independent average
samples
average strength
strength
of of of class-2
theclass-2
average signalssignals
strength
was of was recorded
class-2
recorded for components
signals
for (Figure 6)
components from from
test test
indicates case
that
case F1the (air-dried
F1 highest
(air-dried spec-
average
specimens).
imens).
strength Moreover,
Moreover,of class-2 for
for this this was
signals
case, case,
the the scatter
recorded
scatter of the results
forresults
of the components
was the was
from the greatest.
test
greatest. case Test
F4 cases
1 (air-dried
TestFcases F4
spec-
(specimens
(specimens
imens).
exposed exposed
Moreover, to 25 bath-drying
for this case,
to 25 bath-drying cycles)
the scatter
cycles) and F
and F8of(specimens (specimens
the resultsexposed
8 was thetoexposed
greatest. to 50 freeze–
Test cases
50 freeze–thaw F4
cycles)
thaw cycles)
(specimens
contain contain individual
exposed
individual to
outliers. outliers.
25 bath-drying
The Theaverage
cycles)
smallest smallest
and F 8average ofstrength
(specimens
strength of class-2
exposed
class-2 signals 50 signals
to was freeze–
recorded
was recorded
for
thaw for case individual
case F9 contain
cycles) (specimens F9 (specimens toexposed
exposedoutliers. Thetosmallest
100 freeze–thaw
100 freeze–thaw cycles). cycles). of class-2 signals
average strength
was recorded for case F9 (specimens exposed to 100 freeze–thaw cycles).
Figure 6. Graphic
strength presentation
of class-2 signals. of the Kruskal–Wallis test results for independent samples: average
3.1.6. of
strength Average Duration of Class-2 Signals [µs]
class-2 signals.
3.1.6. Average Duration of Class-2 Signals
The graphical representation [µs]
of Kruskal–Wallis test results for samples of the average
3.1.6. Average
duration
The graphicalDuration
of class-2 of Class-2
signals
representation ofSignals
(Figure [µs] that
7) indicates
Kruskal–Wallis theresults
test longestforaverage
samplesduration of class-2
of the average
signals
duration ofwas recorded
class-2 signals for components
(Figure 7) from
indicates test
thatcase
the F (specimens
longest average soaked in
duration
The graphical representation of Kruskal–Wallis test2 results for samples of the average water
of for 1 h).
class-2
The
signals
durationlargest
was scatter
of recorded of results occurred
for components
class-2 signals for group
from testthat
(Figure 7) indicates F
casethe (specimens
5F2 longest
(specimens exposed
soaked
average to 50 bath-drying
in water
duration for 1
of class-2
h). The largest
signals scatter of
was recorded forresults occurred
components for group
from F5 (specimens
test case F2 (specimens exposed
soakedtoin50water
bath-dry-
for 1
ing cycles). Test case F (specimens ignited for 10 min) contains individual
h). The largest scatter of results occurred for group F5 (specimens exposed to 50 bath-dry-
13 outliers. The
shortest average duration of class-2 signals was recorded for case F (specimens
ing cycles). Test case F13 (specimens ignited for 10 min) contains individual outliers. The
9 exposed
to 100 freeze–thaw
shortest cycles). of class-2 signals was recorded for case F9 (specimens exposed
average duration
Materials 2022, 15, 5757 9 of 23

Materials 2022, 15, x FOR PEER REVIEW cycles).Test case F13 (specimens ignited for 10 min) contains individual outliers.
10 of 25The
shortest average duration of class-2 signals was recorded for case F9 (specimens exposed to
100 freeze–thaw cycles).
Figure
Figure 7. Graphic presentation
7. Graphic of Kruskal–Wallis
the Kruskal–Wallis
testtest results for independent samples: average
duration of class-2presentation
signals. of the results for independent samples: average
duration of class-2 signals.
3.1.7. Number
3.1.7. Numberof Class-3
of Class-3Signals
Signals
3.1.7.The
Number of
graphicalClass-3 Signals
representation of of
Kruskal–Wallis
The graphical representation Kruskal–Wallistest testresults
resultsfor
forindependent
independentsamples
samples of
of theThe graphical
number of
the number representation
of class-3
class-3signals of
signals(FigureKruskal–Wallis
(Figure8)8) indicates
indicates test
thatthatresults
thethe for independent
highest
highest number
number ofsamples
class-3
of class-3 signals
of the
wasnumber
signals recorded offor
was recordedclass-3 signals from
for components
components (Figure from
test 8)case
indicates
testFcase Fthat the
1 (air-dried
(air-dried highest number
specimens).
specimens). The of class-3
The
largest largest
scatter of
1
signals
scatter was
of recorded
results for
occurred components
for group F from test
(specimens case F 1 (air-dried
exposed to 10 specimens).
freeze–thaw
results occurred for group F6 (specimens exposed to 10 freeze–thaw cycles). Test case F10
6 The largest
cycles). Test
scatter
case F10of(specimens
results
(specimens occurred
ignitedignited for
for 2.5 group
for F6 (specimens
2.5 contains
min) min) contains exposed
individual
individual outliers.to 10 freeze–thaw
outliers.
The cycles).
The smallest
smallest number ofTest
numberclass-3
case F 10 (specimens
of class-3 signals wasignited for
recorded 2.5
formin)
case contains
F individual
(specimens
signals was recorded for case F14 (specimens fired in a furnace).
14 fired outliers.
in a The
furnace). smallest number
of class-3 signals was recorded for case F14 (specimens fired in a furnace).
Figure
Figure 8. Graphic
presentation of the
of the Kruskal–Wallis
Kruskal–Wallis testtest results
results for for independent
independent samples:
samples: number
number
of
Figure class-3 signals.
8. Graphic
of class-3 signals. presentation of the Kruskal–Wallis test results for independent samples: number
of class-3 signals.
3.1.8.The The graphical
Average Strength representation of Kruskal–Wallis
of Class-3 Signals [nVs] test results for independent samples of
graphical
the average representation
strength of Kruskal–Wallis
of class-3 signals test results
(Figure 9) indicates that for
the independent
highest average samples
strength
of theThe graphical
strength of class-3ofsignals
Kruskal–Wallis
(Figure 9)test results for
indicates thatindependent
the highest samples
average to
of class-3 signals was recorded for components from test case F4 (specimens exposed
of the average
strength strength
of class-3 signalsof class-3 signalsfor(Figure 9) indicatestestthat theFhighest average
25 bath-drying cycles).wasTherecorded
largest scattercomponents
of results from
occurred case 4 (specimens
for group ex-
F9 (specimens
strength
posed to of
25 class-3 signalscycles).
bath-drying was recorded
The for components
largest scatter of fromoccurred
results test caseforF4 group
(specimens
F ex-
9 (spec-
exposed to 100 freeze–thaw cycles). Test case F14 (specimens fired in a furnace) contains
posed
imens to 25 bath-drying
exposed cycles). The
to 100 freeze–thaw largestTest
cycles). scatter of
F14results occurred
firedfor
ingroup F9 (spec-
individual outliers. The smallest strength case
of class-3(specimens
signals was a furnace)
recorded con-F
for case 10
imens
tains exposed
individual to 100 freeze–thaw
outliers. The cycles).
smallest Test
strength case
of F
class-3
14 (specimens
signals fired
was in a
recorded furnace)
for case con-
F10
(specimens ignited for 2.5 min).
tains individual
(specimens outliers.
ignited for 2.5The smallest strength of class-3 signals was recorded for case F10
min).
(specimens ignited for 2.5 min).
Materials 2022,2022,
Materials 15, x15,
FOR PEER REVIEW
5757 11 of10
25of 23
Figure
strength9. of
Graphic
class-3presentation
signals. of the Kruskal–Wallis test results for independent samples: average
strength
strength of class-3
of class-3 signals.
signals.
3.1.9. Average
3.1.9. Average Duration
Duration of Class-3 Signals
of Class-3 Signals [µs][µs]
3.1.9.The
Average Duration
graphical of Class-3 of
representation Signals [µs]
Kruskal–Wallis testtest
results forfor
independent samples
The graphical representation of Kruskal–Wallis results independent samples
of theThe
of the graphical
average
duration
durationof class-3 of
of class-3 Kruskal–Wallis
signals
signals (Figure test
10)10)
(Figure results
indicates for
indicates independent
thatthat
thethelongest
longestsamples
average
average
of the
durationaverage
duration of duration
of class-3 signals
class-3 of was
class-3
signals signals
recorded
was recordedfor(Figure 10) indicates
components
for components from fromthatcase
test theFcase
test longest
4 (specimensaverage
F4 (specimens ex-
duration
posed to of
25 class-3 signals
bath-drying was
cycles).recorded
The largestfor components
scatter of from
results test
occurredcase
exposed to 25 bath-drying cycles). The largest scatter of results occurred for group F13 F
for (specimens
4group F 13 ex-
(spec-
posed
imens to 25 bath-drying
ignited
(specimens for 10 min).
ignited cycles).
for Test
10 TheTest
cases
min). largest scatter
F2 (specimens
cases of resultsinsoaked
soaked
F2 (specimens occurred
water forinfor group
1 h)
water forFF11310h)
and (spec-
(spec-
and F10
imens
imens ignited for 10
ignited ignited
(specimens min).
for 5 min) Test cases
for 5contain F (specimens
individual
min) contain 2 soaked
outliers.
individual in water
The shortest
outliers. for
The shortest 1 h)
average and
average F
duration
10 (spec-
duration of of
imens
class-3 ignited
signals
class-3 for 5recorded
waswas
signals min) contain
recorded forforindividual
case outliers.
F13 F(specimens
case 13 (specimens
The
ignited shortest
for for
ignited 10average
10 min). min). duration of
class-3 signals was recorded for case F13 (specimens ignited for 10 min).
duration
Figure 10.of of class-3 signals. of the Kruskal–Wallis test results for independent samples: average
Graphic
duration class-3 presentation
signals.
3.1.10.TheThe graphical
Number representation
of Class-4 Signals of Kruskal–Wallis test results for independent samples
of the graphical
number ofrepresentation
class-4 signalsof(Figure
Kruskal–Wallis test results
11) indicates that theforhighest
independent
number samples
of class-4
of theThe graphical
number of representation
class-4 signals of Kruskal–Wallis
(Figure 11) indicatestest results
that the for independent
signals was recorded for components from test case F4 (specimens exposed toclass-4
highest number samples
of 25 bath-
of the
signals number
dryingwas of class-4
recorded
cycles). Thefor signals (Figure
components
largest scatter 11) indicates
offrom
resultstest case that
occurred F4 forthe highest
(specimens
group numbertoof25
F6 exposed
(specimens class-4
bath- to
exposed
signals
drying was recorded
cycles).
10 freeze–thaw for components
Thecycles).
largest scatter
The from
of results
smallest test case
occurred
number F4 (specimens
for
of class-4 group exposed
F6 (specimens
signals was to 25
forbath-
exposed
recorded casetoF14
drying cycles).
10 freeze–thaw The largest
cycles). The scatter
(specimens fired in a furnace). of
smallest results
number occurred
of for
class-4 group
signals F 6 (specimens
was recorded exposed
for case to
F 14
10 freeze–thaw
(specimens firedcycles). The smallest number of class-4 signals was recorded for case F14
in a furnace).
(specimens fired in a furnace).

Materials 2022, 15, 5757 11 of 23
11. Graphic presentation of the Kruskal–Wallis test results for independent samples: number
Figuresignals.
of class-4
of class-4 signals.
of class-4 signals.
The
3.1.11. graphical
Strength of Class-4ofSignals
Kruskal–Wallis
[nVs] test results for samples of the average
The graphical representation of Kruskal–Wallis test results for samples of the average
strength of class-4
The graphical signals (Figure
representation 12) indicates
of Kruskal–Wallis that the
test highest
results average
for samplesstrength
of the of class-
average
strength of class-4 signals (Figure 12) indicates that the highest average strength of class-4
4 signals of
strength was recorded
class-4 for (Figure
signals components from testthat
12) indicates casethe
F1 highest
(air-dried specimens).
average strengthTheoflargest
class-
signals was recorded for components from test case F1 (air-dried specimens). The largest
scatter
4 signalsof results occurred for group F 8 (specimens exposed to 50 freeze–thaw cycles). Test
scatterwas of recorded for components
results occurred for group from
F8 test case F1 (air-dried
(specimens exposed specimens). The largest
to 50 freeze–thaw cycles).
cases F6of(specimens
scatter results exposed fortoexposed
10 freeze–thaw cycles), F7 (specimens exposed to 25 freeze–
Test cases F6 occurred
(specimens group F8 to
(specimens exposed
10 freeze–thaw to 50 freeze–thaw
cycles), F7 (specimens cycles). Test to
exposed
thaw
cases cycles),
F F
(specimens9 (specimens exposed to 100 freeze–thaw cycles) and F11 (specimens ignited
exposed to 10 freeze–thaw cycles), F (specimens exposed to 25 freeze–
25 freeze–thaw cycles), F9 (specimens exposed to 100 freeze–thaw cycles) and F11 (speci-
6 7
for 5 min)
thaw cycles), contain outliers. The smallest
F9 (specimens average strength of class-4 Fsignals was recorded
mens ignited for 5 min) exposed to 100
contain outliers. freeze–thaw
The smallest cycles) andstrength
average 11 (specimens ignited
of class-4 signals
for case
for was
5 min) F 13 (specimens ignited for 10 min).
containfor
recorded outliers.
case F The smallest average
(specimens strength
ignited for 10 min).of class-4 signals was recorded
13
for case F13 (specimens ignited for 10 min).
Figure
Figure 12. Graphic
presentation of Kruskal–Wallis
of the the Kruskal–Wallis
test test results
results for independent
for independent samples:
samples: average
average
strength
strength of class-4
of class-4 signals.
signals.
3.1.12.
strength of Average Duration of Class-4 Signals [µs]
class-4 signals.
3.1.12.The graphical
Duration of Class-4 ofSignals
Kruskal–Wallis
[µs] test results for independent samples
of the average duration of class-4 signals (Figure 13) indicates that the longest average
of the average
The graphical
duration duration of class-4
representation
of class-4 signals
signals wasofrecorded (Figure
Kruskal–Wallis 13) indicates
test
for components thatindependent
resultsfrom
for the case
test longest average
F2 samples
(specimens
duration
of the
soaked of
average class-4
in water signals
duration
for 1 of was recorded
h).class-4 signals
The largest for components
(Figure
scatter from
13) indicates
of results occurred test
thatfor case
thegroupF 2 (specimens
longest average
F4 (specimens
soaked
duration inof
exposed water 25for
class-4
to 1 h). The
signals
bath-drying waslargest scatter
recorded
cycles). Testforofcomponents
case results occurred
from for
F11 (specimens testgroup
ignitedcaseforFF245 (specimens
(specimens
min) contains
exposed
soaked intowater
individual 25 outliers.
bath-drying
The
for 1 h). cycles).
shortest
The Test caseduration
largestaverage
scatter ofF11results
(specimens ignited
ofoccurred
class-4 signals for
was
for group 5 min)
recorded
F contains
for case
4 (specimens
individual
F
exposed outliers.
(specimens The
ignitedshortest
for 10 average
min). duration of class-4 signals was
to 25 bath-drying cycles). Test case F11 (specimens ignited for 5 min) contains recorded for case
13
F 13 (specimens ignited for 10 min).
individual outliers. The shortest average duration of class-4 signals was recorded for case
F13 (specimens ignited for 10 min).
Materials 2022, 15, 5757 12 of 23
Figure
duration
Figure of 13. Graphic
13.class-4
Graphic presentation
signals.
presentation of Kruskal–Wallis
of the the Kruskal–Wallis
testtest results
results for independent
samples: average
average
3.1.13. Average Frequency of AE Events before Reaching Fmax [kHz]
3.1.13.
The Average
graphicalFrequency
representationof AEofEvents before Reaching
Kruskal–Wallis Fmax [kHz]
test results for independent samples
of the average frequency of AE events before reachingtest Fmax (Figure for14) indicates that the
ofThe
the graphical representation
average frequency of AEof Kruskal–Wallis
events before reachingresults
Fmax (Figure independent samples
14) indicates that the
longest
of the average
average frequencies
frequency were
of AE recorded
events for components
before reaching Fmaxfrom test 14)
(Figure caseindicates
F3 (specimens
that the
longest average frequencies were recorded for components from test case F3 (specimens
soaked
longestin average
water forfrequencies
24 h). Moreover,were for this case,
recorded forthe scatter of the
components results
from test was the
case F greatest.
3 (specimens
soaked in water for 24 h). Moreover, for this case, the scatter of the results was the
Test case
soaked F (specimens
in12water
greatest. Testfor Fignited
24 h).
case for 7.5for
Moreover, min)thiscontains
case, theindividual outliers.
scatter of the resultsThe
was smallest av-
the greatest.
12 (specimens ignited for 7.5 min) contains individual outliers. The
erage frequencies
Testsmallest (specimens
case F12averageof AE events before
ignited of
frequencies forAE reaching
7.5events F
min) contains
max were recorded
individual
before reaching for case
outliers.
Fmax F (specimens
The smallest
14
were recorded av-F
for case 14
fired in
erage a furnace).
frequencies of AE events
(specimens fired in a furnace). before reaching Fmax were recorded for case F14 (specimens
fired in a furnace).
FigureFigure 14. Graphic

presentation of the of the Kruskal–Wallis
Kruskal–Wallis test results
test results for independent
samples: average
average
frequency of Graphic
frequency
Figure 14. AEofevents beforebefore
reaching
AEpresentation
events max. Fmax .
the FKruskal–Wallis
ofreaching test results for independent samples: average
frequency of AE events before reaching Fmax.
3.1.14. MORMOR
3.1.14. BendingBending Strength
Strength [MPa][MPa]
3.1.14.
The MOR The graphical
Bending
graphical representation
Strength
representation [MPa] of Kruskal–Wallis
of Kruskal–Wallis test results
test results for independent
for independent samplessamples
of
of the MOR the MOR bending
bending strength strength (Figure
(Figureof15) 15) indicates that
indicates that test the
the highest highest strength
strength was recorded
was recorded
The graphical representation Kruskal–Wallis results for independent samples
for components
forofcomponents fromcase
from test test 3case F3 (specimens
soaked soaked infor
water for 24 h). Moreover,
this for
the MOR bending strengthF(Figure
(specimens
15) indicates in water
that the highest24 h). Moreover,
strength was for
recorded
case, this case, the scatter of the results was the greatest. Test cases F 1 (air-dried specimens),
for the scatter of the
components from results wasFthe
test case greatest. Test
3 (specimens cases
soaked inFwater
1 (air-dried
for 24specimens),
h). F5 (spec-
Moreover, for this
F5 (specimens exposed to 50 bath-drying cycles) and F12 (specimens ignited for 7.5 min)
imens
case,exposed toof
the scatter 50the
bath-drying
results wascycles) and F12Test
the greatest. (specimens ignited for
cases F1 (air-dried 7.5 min) contain
specimens), F5 (spec-
contain individual outliers. The smallest MOR bending strength was recorded for case F14
individual
imens exposed outliers.
to The smallest MOR
50 bath-drying bending
cycles) and strength was recorded
F12 (specimens ignited for case F14 (speci-
7.5 min) contain
(specimens fired in a furnace).
mens fired inoutliers.
individual a furnace). The smallest MOR bending strength was recorded for case F14 (speci-
mens fired in a furnace).
Materials 2022, 15, 5757 13 of 23
Figure 15. Graphic presentation of the Kruskal–Wallis test results for independent samples: MOR
bending strength.
Figure15.
Figure Graphicpresentation
15.Graphic presentationofofthe
theKruskal–Wallis
Kruskal–Wallistest
testresults
resultsfor
forindependent
independentsamples: MOR
samples:MOR
3.2. Breakdown
bending of
strength.
bending strength. the Number of Signals and Selected Energy Parameters into Groups
Depending on the Significance of Changes
3.2. Breakdown of the Number of Signals and Selected Energy Parameters into Groups Depending
3.2. Classification
Breakdown
on the of theoftrees
Significance using
Number
Changes the CHAID
of Signals algorithm
and Selected were
Energy used to into
Parameters identify
Groupsthe divisions
into signaloncharacteristics.
Depending the Significance of Changes
Classification trees using the CHAID algorithm were used to identify the divisions
intoClassification trees using the CHAID algorithm were used to identify the divisions
signal characteristics.
into signal characteristics. Signals
3.2.1. Number of Class-1
3.2.1.Four groups
Number were identified.
of Class-1 Signals For every successive group, there is a significant drop
in theNumber
3.2.1. number
Four ofofClass-1
groups class-1 signals (Figure
Signals
were identified. 16): successive group, there is a significant drop in
For every
•the number
Groups
Four 1,class-1
groups
of 2, were
3, 4 and 6; (Figure
identified.
signals For 16):
every successive group, there is a significant drop
in•• theGroups
Groups
number1,5,of2,
7,class-1
8, 9 and 10;
signals
3, 4 and 6; (Figure 16):
• Groups
•• Groups 11 and 12;
Groups1,5,2,7,3,8,49and
and6;10;
••• Groups
Groups5,11
Groups 137,and
and 14.
8, 912;
and 10;
•• Groups11
Groups 13and
and12;
14.
• Groups 13 and 14.
Figure 16. Classification tree for the number of class-1 signals.

Figure 16. Classification
3.2.2. Average Strengthtree
of for the number
Class-1 Signalsof[nVs]
class-1 signals.
Three groups were identified. For every successive group, there is a significant drop
3.2.2. 16.
Figure Average Strength
Classification of Class-1
tree Signals class-1
[nVs] signals.
in the average strength offor the number
class-1 signalsof(Figure 17):
• Groups 1, 2, 4, 5, 6 and 7;
• Groups 3, 8 and 9;
• Groups 10, 11, 12, 13 and 14.
Three groups were identified. For every successive group, there is a significant drop
in the average strength of class-1 signals (Figure 17):
• Groups 1, 2, 4, 5, 6 and 7;
Materials 2022, 15, 5757 14 of 23
• Groups 10, 11, 12, 13 and 14.
Figure 17. Classification tree for the average strength of class-1 signals.
3.2.3.Three
Figure Average Duration
17. Classification
groups weretreeoffor
Class-1 Signals
the average
identified. [µs]
strength
For every of class-1 signals.
successive group, there is a significant drop
in theThree groups
average wereof
duration identified. For every
class-1 signals successive
(Figure 18): group, there is a significant drop
•in the average
Groups 1, duration
2, 5 and 7;of class-1 signals (Figure 18):
•• Groups 4,
Groups 1, 6,
2, 85 and
and 9;
7;
•• Groups 3,
Groups 4, 10,
6, 811,
and
12,9;13 and 14.
• Groups 3, 10, 11, 12, 13 and 14.
Figure 18. Classification tree for the average duration of class-1 signals.
Six groups were identified. For every successive group, there is a significant drop in
Six groups
the number were identified.
of class-2 For every
signals (Figure 19): successive group, there is a significant drop in
the number of class-2 signals (Figure 19):
• Groups 2 and 4;
•• Groups
Groups 21 and
and 4;
3;
•• Groups
Groups 15,and 3; 7;
6 and
Materials 2022, 15, 5757 • Groups 2 and 4; 15 of 23
• Groups 1 and 3;
•• Groups8,8,9 9and
Groups and10;10;
•• Groups13
Groups 13and and14.14.

3.2.5. 19.
Figure Average Strength
Classification treeoffor
Class-2 Signals
the number [nVs] signals.
of class-2
Four groups were identified. For every successive group, there is a significant drop
Four groups were identified. For every successive group, there is a significant drop in
the average strength of class-2 signals (Figure 20):
•• Groups
Groups 1 and
1 and 4; 4;
•• Groups 2, 3, 5, 76,and
Groups 2, 3, 5, 6, 7 and
8; 8;
•• Groups
Groups 9, 9,
1010andand14; 14;
•• Groups
Groups 11,11,
12 12
and and
13. 13.
Figure
3.2.6. 20. Classification
Average Duration of tree for the
Class-2 average
Signals strength of class-2 signals.
[µs]
3.2.6.
the Average
average Duration
duration of Class-2
of class-2 Signals21):
signals (Figure [µs]
• Four groups
Groups were
1, 2, 3 and 4; identified. For every successive group, there is a significant drop
•in the
Groups 5 and 10;
average duration of class-2 signals (Figure 21):
• Groups 6, 7, 8 and 11;
• Groups 1, 2, 3 and 4;
• Groups 9, 12, 13 and 14.
• Groups 5 and 10;
• Groups 6, 7, 8 and 11;
• Groups 9, 12, 13 and 14.
in the average duration of class-2 signals (Figure 21):
• Groups 1, 2, 3 and 4;
Materials 2022, 15, 5757 16 of 23
• Groups 6, 7, 8 and 11;
• Groups 9, 12, 13 and 14.
FigureNumber
Number of Class-3treeSignals
for the average duration of class-2 signals.
Four groups were were identified.

identified.For Forevery
everysuccessive
successivegroup,
group,there
thereis is a significant
a significant drop
drop in
in
thethe number
number of of class-3
class-3 signals
signals (Figure
(Figure 22):22):
•• Groups 1, 2, 4 and 5;
•• Groups 3, 6 and 7;
•• Groups 8, 9, 10, 11 and 12;
•• Groups 13 and 14.

3.2.8.Four groups
Average were identified.
Strength of Class-3 For every
Signals successive group, there is a significant drop in
[nVs]
the average strength of class-3 signals (Figure 23):
• theGroups
in average1,strength
2, 4 and 5;
of class-3 signals (Figure 23):
• Groups 3 and 6;
•• Groups
Groups 1,
7, 2, 4 and
8, 9, 5; 14;
13 and
Groups 10, 11 and 12.
• Groups 7, 8, 9, 13 and 14;
• Groups 10, 11 and 12.
• Groups 1, 2, 4 and 5;
• Groups 3 and 6;
Materials 2022, 15, 5757 • Groups 7, 8, 9, 13 and 14; 17 of 23
• Groups 10, 11 and 12.
Figure
3.2.9.Average
Average Duration
Duration treeoffor
of the average
Class-3
Class-3 strength
Signals
Signals [µs] of class-3 signals.
[µs]
Fivegroups
Five groupswere
wereidentified.
identified.For
Forevery
everysuccessive
successive group,
group, there
there is aissignificant
a significant drop
drop in
in the
the average
average duration
duration of class-3
of class-3 signals
signals (Figure
(Figure 24):24):
•• Groups
Groups11and
and2;2;
•• Groups
Groups33and
and4;4;
•• Groups
Groups55and
and6;6;
•• Groups
Groups7,7,88and
and9;9;
•• Groups
Groups 10, 11,12,
10, 11, 12,1313and
and14.
14.
SixNumber
3.2.10. groups were identified.
of Class-4 SignalsFor every successive group, there is a significant drop in
•the number
Groups 1ofand 3; signals (Figure 25):
class-4
• Groups 2 and 4;
• Groups 1 and 3;
• Groups 2 and 4;
• Groups 1 and 3;
• Groups 2 and 4;
25.Classification
Figure 25.
Figure Classificationtree for for
tree the the
number of class-4
number signals.
of class-4 signals.
3.2.11. Average
Four groups Strength of Class-4
were identified. Signals
For every [nVs] group, there is a significant drop in
successive
Materials 2022, 15, x FOR PEER REVIEW
the Four groups
average strengthwere identified.
of class-4 signals For every
(Figure 20 of 25
26):successive group, there is a significant drop
in
• the average
Groups 1, 2,strength
3 and 4; of class-4 signals (Figure 26):
•• Groups
Groups5,1,6,2,
7, 38,and
9 and
4; 11;
•• Groups 10 and
Groups 10 and 12;12;
• Groups 5, 6, 7, 8, 9 and 11;
•• Groups
Groups 13
13 and
and 14.
14.
Four groups
the average wereofidentified.
duration For every
class-4 signals successive
(Figure 27): group, there is a significant drop
• Groups 1, 2, 3, 4, 5 and 6;
•• Groups
Groups 1,
7, 2, 3, 4, 9;
8 and 5 and 6;
•• Groups
Groups 7,
10,8 11
and 9; 12;
and
•• Groups
Groups 10, 11 and
13 and 14 12;
• Groups 13 and 14
• Groups 1, 2, 3, 4, 5 and 6;
Materials 2022, 15, 5757 • Groups 10, 11 and 12; 19 of 23
• Groups 13 and 14
Figure 27.
Figure Classification tree
27. Classification tree for
for the
the average
average duration
duration of
of class-4
class-4 signals.
signals.

Five groups were identified. For every successive group, there is a significant drop in
Five groups
duration AE eventsFor every
(Figure successive group, there is a significant drop
28):
in the average duration of AE events (Figure 28):
• Groups 1, 2, 3 and 4;
•• Groups
Groups 1, 2, 3 6;
5 and and 4;
Groups 7, 8 and 9;
•• Groups
Groups 7,
108and
and11;9;
•• Groups
Groups 10
12,and 11; 14.
13 and
• Groups 12, 13 and 14.
Figure 28.
Figure Classification tree
28. Classification tree for
for the
the average
average frequency
frequency of
of AE
AE events
events before
before reaching
reaching F
Fmax
max.
.
3.2.14. MOR Bending Strength [MPa]

3.2.14. MOR Bending Strength [MPa]
Five groups were identified. For every successive group, there is a significant drop in
Five groups
duration AE eventsFor every
(Figure successive group, there is a significant drop
29):
• Groups 1, 2, 3 and 4;
•• Groups
Groups 1, 2, 3 6;
5 and and 4;
Groups 7, 8 and 9;
•• Groups
Groups 7,
108and
and11;9;
Groups 12, 13 and 14.
• Groups 12, 13 and 14.
• Groups 1, 2, 3 and 4;
• Groups 5 and 6;
• Groups 12, 13 and 14.
Figure 29. Classification tree for MOR bending strength.

Figure 29. Classification
3.3. Correlation betweentree for MOR bending strength.
Parameters
In order to test the normality of the data distribution, the Shapiro–Wilk test was per-
3.3. Correlation between Parameters
formed. In most cases, the data are not normally distributed. Therefore, the Spearman cor-
In order
relation to testwas
coefficient the used
normality of the data
to investigate thedistribution,
relationshipthe Shapiro–Wilk
between test was
the indicated per-
changes.
formed. In most cases, the data are not normally distributed.
It was found that a significant correlation exists between: Therefore, the Spearman
correlation coefficient was used to investigate the relationship between the indicated
• The number of class-1 signals and the bending strength MOR (ρ = 0.874, p = 0.000):
changes.
the more signals there are, the higher the MOR bending strength.
It was found that a significant correlation exists between:
• Average strength of class-1 signals and bending strength MOR (ρ = 0.581, p = 0.000):
• The number
the higher ofaverage
the class-1 signals and the is,
signal strength bending strength
the higher MORbending
the MOR (𝜌 = 0.874, p = 0.000):
strength.
• the Themore signals
number there are,
of class-2 the higher
signals and thethe MOR bending
bending strengthstrength.
MOR (ρ = 0.844, p = 0.000):
• Average
the morestrength of class-1
signals there signals
are, the higherand
thebending strengthstrength.
MOR bending MOR (𝜌 = 0.581, p = 0.000):
• the higher the average signal strength is, the higher
Average strength of class-2 signals and bending strength MOR the MOR bending strength.
(ρ = 0.483, p = 0.000):
the higher the average signal strength is, the higher the MOR bending strength.
• Average duration of class-2 signals and bending strength MOR (ρ = 0.548, p = 0.000):
the greater the mean duration of the signals is, the higher the MOR bending strength.
• Average frequency of AE events before reaching Fmax and the bending strength MOR
(ρ = 0.875, p = 0.000): the higher the frequency of AE events is, the higher the MOR
bending strength.
4. Discussion
The test results shown above indicate that the operating conditions affect the me-
chanical parameters of fibre–cement boards exposed to loads. The impact of moisture and
low and high temperature significantly reduces the strength of fibre–cement boards in
comparison with air-dried objects under optimum humidity conditions. In most cases, the
strength parameters decrease with the increase in the test case number. The highest strength
decrease was observed in specimens exposed to flame and high temperature, which are
destroyed by the action of high temperature due to the nature of the fibres.
Materials 2022, 15, 5757 21 of 23
The behaviour of fibre–cement boards under load is connected with the presence
of the acoustic emission phenomenon. The decrease in the strength parameters in the
objects results in a change in the acoustic characteristics and the number of signals in the
individual classes. The graphical representation of the Kruskal–Wallis test results indicates
that the MOR bending strength for specimens exposed to fire and high temperature is
more than 50% lower than for air-dried specimens and specimens exposed to water. The
increased number of freeze–thaw cycles also has an impact on the strength of the specimens.
Components exposed to more than 10 freeze–thaw cycles had a strength more than 30%
smaller than the reference specimens soaked in water and exposed to bath-drying cycles. A
similar dependency was indicated by the number of signals of the individual classes, their
energy parameters and frequencies. The number, strength, duration and frequency also
decrease along with the increase in the test case number.
The acoustic emission parameters that are most sensitive to the reduction in the
mechanical parameters of the boards are the average strength and the average duration of
class 3 signals, the average strength and the average duration of the class 4 signals and the
average frequency of events before reaching the maximum load. The occurrence of class 3
signals with an average strength of signals in the range from 200 to 400 nVs and a duration
in the range of 1000–1500 µs and of class 4 signals with an average strength in the range of
100 to 300 nVs and a duration in the range of 100–150 µs at average event frequencies above
200 kHz, but not higher than 300 kHz, is associated with a reduction in the strength of the
boards from 30% to 50%. The occurrence of class 3 signals with an average strength of
signals below 200 nVs and duration less than 1000 µs and the occurrence of class 4 signals
with an average strength of less than 100 nVs and with a duration of less than 100 µs with
an average frequency of events below 200 kHz is associated with a reduction in the strength
of the boards by more than 50%.
5. Conclusions
The research of fibre–cement boards, so far, has focused mainly on the influence of
operational factors and high temperatures on the boards, determined by the examination
of the physicochemical parameters-mainly on the flexural strength (MOR). The literature
describes only a few cases of testing fibre–cement boards using non-destructive methods,
including the acoustic emission method. There are results available that have tried to
apply the acoustic emission method to determine the effect of cellulose fibers on the
strength of fibre–cement boards and distinguish the EA events emitted by the fibers from
those emitted by the cement matrix. Research has shown that this method is suitable for
testing fibre–cement boards. Attempts were also made to use the acoustic emission method
to study the effect of fire and high temperatures on fibre–cement boards, where it was
proposed to combine the acoustic emission method with artificial intelligence, including
artificial neural networks (ANN). These studies confirmed the effectiveness of using the
acoustic emission method to monitor the condition of fibre–cement boards, however, the
issue of the statistical correlation of mechanical and acoustic parameters was not discussed.
The following conclusions can be drawn based on the results shown above:
• During the three-point bending of fibre–cement specimens, various material damage
mechanisms occur.
• Some conditions that may appear during the use of fibre–cement components have a
significant adverse effect on their strength parameters.
• The reduction in strength parameters is strictly connected with a change of acoustic
parameters recorded for a material under load.
• The highest decrease in strength is connected with the most significant reduction in
the number of signals of the individual classes, their strength, duration and frequency.
• The highest decrease in the strength, number and parameters of AE signals was
observed for the specimens exposed to fire and high temperature.
Materials 2022, 15, 5757 22 of 23
• After analysing the results of the research, it was found that the acoustic emission
method was suitable to be used for the monitoring and diagnostics of fibre–cement
boards under load.
Due to the existence of a significant correlation between the mechanical parameters
of fibre–cement boards and the number and parameters of acoustic emission signals of
individual classes, according to the authors, further research should be undertaken in
order to describe the specific ranges of the parameters of acoustic emission signals of
individual classes, which would enable conclusions about the condition of fibre–cement
boards under a low load value. The authors also plan to include the ultrasound method in
the applied methodology, which would enable the detection of possible material voids and
discontinuities in the distribution of reinforcing fibers.
Author Contributions: Conceptualization, A.A.-B. and G.Ś.; methodology, G.Ś.; software, A.A.-B.;
validation, A.A.-B., A.K. and E.P.; formal analysis, A.K.; investigation, A.A.-B.; resources, E.P.; data
curation, G.Ś.; writing—original draft preparation, A.A.-B.; writing—review and editing, A.K.;
visualization, E.P.; supervision, G.Ś.; project administration, E.P.; funding acquisition, G.Ś. All authors
have read and agreed to the published version of the manuscript.
Funding: The project entitled “Sildig AE-system for identifying and locating gas infrastructure
defects us-ing the acoustic emission method” is co-financed by the European Regional Development
Fund under Sub-measure 1.1.1 of the Intelligent Development Operational Program 2014–2020,
contract no. POIR.01.01.01-00-1019/19-00 of June 10, 2020 The project is carried out in a consortium
by Polska Spółka Gazownictwa sp.z o.o. (Polish Gas Distribution Group-Consortium Leader) and
Kielce University of Technology (Consortium Member).
Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.
Data Availability Statement: Not applicable.
Conflicts of Interest: The authors declare no conflict of interest.
References
1. Schabowicz, K.; Gorzelańczyk, T. Fabrication of fibre cement boards. In The Fabrication, Testing and Application of Fibre Cement
Boards, 1st ed.; Ranachowski, Z., Schabowicz, K., Eds.; Cambridge Scholars Publishing: Newcastle Upon Tyne, UK, 2018; pp. 7–39.
ISBN 978-1-5276-6.
2. Faruk, O.; Bledzki, A.K.; Fink, H.P.; Sain, M. Biocomposites reinforced with natural fibers. Prog. Polym. Sci. 2012, 37, 1552–1596.
[CrossRef]
3. Tonoli, G.H.D.; Santos, S.F.; Savastano, H.; Delvasto, S.; de Gutiérrez, R.M.; de Murphy, M.D.M.L. Effects of natural weathering
on microstructure and mineral composition of cementitious roofing tiles reinforced with fique fibre. Cem. Concr. Compos. 2011, 33,
225–232. [CrossRef]
4. Khorami, M.; Ganjian, E. The effect of limestone powder, silica fume and fibre content on flexural behaviour of cement composite
reinforced by waste Kraft pulp. Constr. Build. Mater. 2013, 46, 142–149. [CrossRef]
5. Claramunt, J.; Ardanuy, M.; García-Hortal, J.A.; Filho, R.D.T. The hornification of vegetable fibers to improve the durability of
ement mortar composites. Cem. Concr. Compos. 2011, 33, 586–595. [CrossRef]
6. Dzioba, I.; Lipiec, S.; Pała, R.; Furmańczyk, P. On Characteristics of Ferritic Steel Determined during the Uniaxial Tensile Test.
Materials 2021, 14, 3117. [CrossRef] [PubMed]
7. Schabowicz, K.; Józwiak-Niedzwiedzka, D.; Ranachowski, Z.; Kudela, S.; Dvorak, T. Microstructural characterization of cellulose
fibres in reinforced cement boards. Arch. Civ. Mech. Eng. 2018, 4, 1068–1078. [CrossRef]
8. Domaneschi, M.; Niccolini, G.; Lacidogna, G.; Cimellaro, G.P. Nondestructive Monitoring Techniques for Crack Detection and
Localization in RC Elements. Appl. Sci. 2020, 10, 3248. [CrossRef]
9. Schabowicz, K. Modern acoustic techniques for testing concrete structures accessible from one side only. Arch. Civ. Mech. Eng.
2015, 15, 1149–1159. [CrossRef]
10. Dzioba, I.; Zvirko, O.; Lipiec, S. Assessment of Operational Degradation of Pipeline Steel Based on True Stress–Strain Diagrams.
In Degradation Assessment and Failure Prevention of Pipeline Systems; Springer: Berlin/Heidelberg, Germany, 2021; Volume 102, pp.
175–186.
11. Adamczak-Bugno, A.; Świt, G.; Krampikowska, A. Assessment of Destruction Processes in Fibre-Cement Composites Using the
Acoustic Emission Method and Wavelet Analysis. IOP Conf. Ser. Mater. Sci. Eng. 2019, 471, 032042. [CrossRef]
Materials 2022, 15, 5757 23 of 23
12. Park, W.H.; Packo, P.; Kundu, T. Acoustic source localization in an anisotropic plate without knowing its material properties—A
new approach. Ultrasonics 2017, 79, 9–17. [CrossRef]
13. Habib, A.; Kim, C.H.; Kim, J.-M. A Crack Characterization Method for Reinforced Concrete Beams Using an Acoustic Emission
Technique. Appl. Sci. 2020, 10, 7918. [CrossRef]
14. Calabrese, L.; Khaskhoussi, A.; Patanè, S.; Proverbio, E. Assessment of Super-Hydrophobic Textured Coatings on AA6082
Aluminum Alloy. Coatings 2019, 9, 352. [CrossRef]
15. Tsangouri, E.; Aggelis, D.G. A review of acoustic emission as indicator of reinforcement effectiveness in concrete and ce-mentitious
composites. Constr. Build. Mater. 2019, 224, 198–205. [CrossRef]
16. Świt, G.; Adamczak, A.; Krampikowska, A. Wavelet Analysis of Acoustic Emissions during Tensile Test of Carbon Fibre Reinforced
Polymer Composites. IOP Conf. Ser. Mater. Sci. Eng. 2017, 245, 22–31. [CrossRef]
17. Afroughsabet, V.; Ozbakkaloglu, T. Mechanical and durability properties of high-strength concrete containing steel and polypropy-
lene fibers. Constr. Build. Mater. 2015, 94, 73–82. [CrossRef]
18. Buratti, N.; Mazzotti, C.; Savoia, M. Post-cracking behavior of steel and macro-synthetic fibre-reinforced concretes. Constr. Build.
Mater. 2011, 25, 2713–2722. [CrossRef]
19. Calabrese, L.; Bonaccorsi, L.; Proverbio, E. Corrosion protection of aluminum 6061 in NaCl solution by silane-zeolite composite
coatings. J. Coat. Technol. Res. 2012, 9, 597–607. [CrossRef]
20. Goszczyńska, B.; Świt, G.; Trampczy
˛ ński, W.; Krampikowska, A. Application of the acoustic emission method of identification
and location of destructive processes to the monitoring concrete bridges. In Proceedings of the 7th International Conference of
Bridge Maintenance, Safety and Management, IABMAS 2014, Shanghai, China, 7–11 July 2014; pp. 688–694.
21. Adamczak-Bugno, A.; Świt, G.; Krampikowska, A. Fibre-Cement Panel Ventilated Façade Smart Control System. Materials 2021,
14, 5076. [CrossRef]
22. Drelich, R.; Gorzelanczyk, T.; Pakuła, M.; Schabowicz, K. Automated control of cellulose fibre cement boards with a non-contact
ultrasound scanner. Autom. Constr. 2015, 57, 55–63. [CrossRef]
23. Landis, E.N.; Kravchuk, R.; Loshkov, D. Experimental investigations of internal energy dissipation during fracture of fiber-
reinforced ultra-high-performance concrete. Front. Struct. Civil Eng. 2019, 13, 190–200. [CrossRef]
24. De Sutter, S.; Verbruggen, S.; Tysmans, T.; Aggelis, D. Fracture monitoring of lightweight composite-concrete beams. Compos.
Struct. 2017, 167, 11–19. [CrossRef]
25. Schabowicz, K.; Gorzelanczyk, T.; Szymków, M. Identification of the degree of degradation of fibre-cement boards exposed to fire
by means of the acoustic emission method and artificial neural networks. Materials 2019, 12, 656. [CrossRef] [PubMed]
26. Zangana, S.; Epaarachchi, J.; Ferdous, W.; Leng, J.; Schubel, P. Behaviour of continuous fibre composite sandwich core under
low-velocity impact. Thin-Walled Struct. 2021, 158, 107157. [CrossRef]
27. Szewczak, E.; Winkler-Skalna, A.; Czarnecki, L. Sustainable Test Methods for Construction Materials and Elements. Materials
2020, 13, 606. [CrossRef] [PubMed]
28. EN 12467:2012; Fiber-Cement Flat Sheets—Product Specification and Test Methods. European Committee for Standardization:
Brussels, Belgium, 2012.

Analysis of The Significance of Changes in The Num PDF

Uploaded by

Copyright:

Available Formats

Analysis of The Significance of Changes in The Num PDF

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

Analysis of The Significance of Changes in The Num PDF

Uploaded by

Copyright:

Available Formats

materials

Academic Editor: Jacek Domski 1. Introduction

Materials 2022, 15, 5757. https://doi.org/10.3390/ma15165757 https://www.mdpi.com/journal/materials

1.2. Application and Operating Conditions of Fibre–Cement Boards

1.3. Diagnostic Methods Used to Monitor Fibre–Cement Cladding

2. Materials and Methods

Materials 2022, 15, x FOR PEER REVIEW 6 of 24

3.1.2. Average Strength of Class-1 Signals [nVs]

3.1.2. Average Strength of Class-1 Signals [nVs]

Materials 2022, 15, x FOR PEER REVIEW 10 of 25

Materials 2022, 15, x FOR PEER REVIEW 12 of 25

FigureFigure 14. Graphic

Materials 2022, 15, x FOR PEER REVIEW 14 of 25

Materials 2022, 15, 5757 13 of 23

Figure 16. Classification tree for the number of class-1 signals.

Materials 2022, 15, x FOR PEER REVIEW 16 of 25

Materials 2022, 15, x FOR PEER REVIEW 17 of 2

Figure 19. Classification tree for the number of class-2 signals.

Materials 2022, 15, x FOR PEER REVIEW 18 of 25

Four groups were were identified.

Figure 22. Classification tree for the number of class-3 signals.

Materials 2022, 15, x FOR PEER REVIEW 19 of 25

3.2.13. Average Frequency of AE Events before Reaching Fmax [kHz]

3.2.14. MOR Bending Strength [MPa]

Figure 29. Classification tree for MOR bending strength.

You might also like