Professional Documents
Culture Documents
1 s2.0 S1537511020300064 Main
1 s2.0 S1537511020300064 Main
ScienceDirect
Research Paper
article info
A single passive acoustic emission sensor was used to collect signals coming from an
Article history: obstructed pipe in a water recirculation system. Four geometrically different obstructions
Received 13 July 2019 were investigated. The flow field of water around each obstruction was visualised with the
Received in revised form use of 2D particle image velocimetry (PIV) to identify the different flow features. In parallel,
13 December 2019 the acoustic emission signals were acquired by locating a piezoelectric sensor on the outer
Accepted 28 December 2019 wall of the pipe at the tip of the obstruction. The acoustic emission signals were then pre-
Published online 20 January 2020 processed and the frequency domain was extracted for 100 recordings in each case. Signals
were processed further by using principle component analysis and a matrix is created for
Keywords: supervised machine learning algorithms. This methodology was applied over a range of
Obstruction four flow rates, all in fully developed turbulent flow. Results showed that different ob-
Hydrology structions generated different acoustic signals and flow fields, which reflected the different
Pipe Blockage flow fields observed with PIV. The average velocity and amplitude of the acoustic signals
Machine Learning increased in magnitude with increasing flow rate. The machine-learning algorithm with
Particle Image Velocimetry highest prediction values was quadratic support-vector machine with predictions in the
Acoustics area of 95% accuracy or above. This makes the combination of machine learning and a
single passive acoustic sensor a viable option to predict pipe obstructions and the type of
obstruction. This may lead to a useful application for urban water supply or sewage sys-
tems as well as agricultural practice for field irrigation or the detection of nozzle blockages.
© 2020 IAgrE. Published by Elsevier Ltd. All rights reserved.
* Corresponding author.
E-mail address: f.alberini@bham.ac.uk (F. Alberini).
https://doi.org/10.1016/j.biosystemseng.2019.12.015
1537-5110/© 2020 IAgrE. Published by Elsevier Ltd. All rights reserved.
b i o s y s t e m s e n g i n e e r i n g 1 9 1 ( 2 0 2 0 ) 4 8 e5 9 49
do require additional computational power (Sause, Gribov, classification between data classes, each class will be plotted
Unwin, & Horn, 2012). Another advantage of the technology in another dimensional space, the so-called feature space
proposed here, compared to hydrophones, is that the sensor is (Guan et al., 2019; Masud et al., 2010). Figure 2 1 illustrates the
placed on the outer wall of the pipe. Hydrophones are in-pipe presentation of machine learning data.
solutions, which makes it difficult to maintain or replace such
equipment. As hydrophone is microphones they also have
been waterproofed which makes them suitable to be 2. Materials and method
immersed in fluids. By positioning a hydrophone close to a
suspected leak or by applying those inside a pipe system the 2.1. Water recirculation system and pipe obstructions
distinctive sound of water escaping can be detected. Hydro-
phones can also be used to listen for leaks in pressurised pipe A water recirculation system was fed by a 40 l water tank and
systems that run close to the wall (Khalifa, Chatzigeorgiou, powered by a centrifugal pump (Alfa Laval, Lund, Skåne,
Youcef-Toumi, Khulief, & Ben-Mansour, 2010) and work well Sweden) of the type I KA-5 132SSS1. The internal diameter ID
for plastic piping (Hunaidi & Chu, 1999). of the pipework is 25.4 mm with a 120 mm in length pipe
A key challenge making use of such data is the sheer vol- segment from stainless steel, which is used for the AE data
ume of incoming data points. To make use of such complex acquisition of passive acoustic signals. A schematic drawing
and high-volume data machine learning can be used. The of the test rig is shown in Fig. 2.
term machine learning (ML) applies to artificial systems that All four in-pipe obstructions were designed in AutoCad
can generate knowledge based on experience. Such systems 2018 (Autodesk Inc., San Rafael, CA USA), and extruded by a
learn on given, and to the problem representative, examples. FlashForge Dreamer 3D printer (Zhejiang Flashforge 3D
It is the system's task to create logical paths or identify logical Technology Co., Ltd., Jinhua, Zhejiang, China). Each obstruc-
patterns to the given problem. In general, ML can either solve tion was slotted into the pipe and had a length of 1.5 times the
classification or regression problems. In the first case ML aims inner diameter (1.5 25.4 mm ¼ 38.1 mm). The four geometric
to classify data into classes, such as fault classification. obstructions were wall-leaning cone, cross (four wedges
Regression problems are those where the system will be asked meeting in the central focal point), three triangular aligned
to provide future predictions of a continuous variable (e.g. holes and a semicircle. An IMO iDrive AC Inverter Drive con-
temperature trends) (Kubat, 2015; Moraru, Pesko, Porcius, trol unit (IMO Precision Controls Ltd, Hatfield, Hertfordshire,
Fortuna, & Mladenic, 2010). UK) was used to set the four flow rates of 1300, 3000, 4530 and
In the end of this process, the system will create a general 6350 l h1. The pressure drop was measured by using a TPI
algorithm/system that is capable to solve problems to previ- 665L Digital Manometer (Test Products International Europe
ously unseen datasets. The system/algorithm can also eval- Ltd., Crawley, West Sussex, UK). All four in-pipe obstructions
uate unknown datasets (i.e. learning transfer) or fail to learn used in the investigation are shown in Fig. 3.
unknown data (i.e. overfitting) (Koza, Bennett, Andre, & The free area of each obstruction was Cone: 1.287 104 m,
Keane, 1996; Reitmaier, 2015). Three holes: 2.83 105 m, Semicircle: 2.53 104 m, Cross:
Based on the methodology ML approaches a problem and 2.53 104 m. The geometries chosen represent either po-
makes use of datasets three types of ML learning algorithms tential deposit types (Cone for slight build-up, Semicircle for
can be distinguished; namely supervised, unsupervised and half blocked pipe and Three holes for a full blockage with the
reinforced learning (Fischer, 1999). The majority of ML perforations delivering partial flow) and the potential design
employed in industry (between 80 to 90%) are of supervised of spray nozzle (Cross).
and unsupervised nature with a dominance towards super-
vised ML (Ge, Song, Ding, & Huang, 2017) (see Fig. 1). 2.2. 2D Particle image velocimetry
The way a system learns is by giving it labelled examples,
thus the answer to the question posed is already known. The PIV is an established technology in fluid dynamics to visualise
system will try to build a logical path with the help of algo- flow fields by a combining laser technology, a camera and the
rithms (quadratic support-vector machine, decision trees fluid seeded with particles. PIV is a non-intrusive technology
etc.). This is done with only a selected proportion of dataset, that delivers time-resolved velocity information and flow
whereas the remaining parts are used to optimise the algo- maps on a nearly instantaneous basis, however, is still mainly
rithm and finally test it to an unknown dataset, with a com- exclusively used within academia (Adrian, 2005). A pulsed
mon split being 60% for training, 20% for optimisation and 20% laser beam is configured into a light slab that is fired into a
for testing (Chong, Abraham, & Paprzycki, 2004; Wuest, translucent chamber. The fluid within the chamber contains
Weimer, Irgens, & Thoben, 2016). Results are projected as seeding particles that are assumed to behave as the fluid
observations, which are dependent on the number of features. within the chamber. The laser pulse time delay or interval is
An observation is a specific number set that characterises a referred to as Dt. By the help of cross-correlation algorithms,
class (Kotsiantis, Zaharakis, & Pintelas, 2007). Hence, data can the particle displacement (X, Y direction) and velocity can be
be seen as points in an n-dimensional space. The number of determined.
attributes n is directly related to the class observed from The time delay Dt between each frame depends on the
experience or the experiments. In order to achieve a better maximum distance a particle could travel within the
b i o s y s t e m s e n g i n e e r i n g 1 9 1 ( 2 0 2 0 ) 4 8 e5 9 51
Fig. 1 e Schematic projections of machine learning data, adopted from (Liu & Wang, 2017).
interrogation window. The formula Dt for is given below and v dpipe rfluid
Re ¼ (3)
follows Adrian (1986): hfluid
S Lp with dpipe being the internal pipe diameter and hfluid being the
Dt < 0:25 (1) dynamic viscosity of water.
vs
For the flow field visualisation a TSI 2D PIV system (TSI Inc.,
with S being the magnification factor, Lp is the length of the Missoula, MT, USA) was used. A dual head green 532 nm
interrogation area in pixels vs for the superficial velocity in the Nd:Yag laser (Litron Nano PIV, Rugby, Warwickshire, UK)
pipe. pulsing at 7 Hz has been formed to light sheet. The laser was
The superficial velocity vs , a hypothetical flow velocity mounted and levelled 500 mm above the acrylic chamber,
calculated which does not take local differences into account, with the light sheet following the length of the chamber. The
within the pipe was calculated by dividing the volumetric flow system was synchronised to a single TSI Power View 4 meg-
rate V_ by the area Acircle of the pipe: apixels (2048 2048 pixels) 12-bit CCD camera that could be
controlled in its X-Y-Z direction by a remotely controllable
V_
vs ¼ (2) frame construction. The CCD camera is connected to another
Acircle
52 b i o s y s t e m s e n g i n e e r i n g 1 9 1 ( 2 0 2 0 ) 4 8 e5 9
synchroniser (TSI 610035) which was attached to a desktop PC relative variance the mean and the variance for each column
(DELL Inc., Round Rock, Texas, USA). The 2D PIV system is (frequency) has to be found by using the following formulae:
controlled by TSI Insight 4G software. Since fully developed
1 XN 2
turbulent flow is apparent, 2 500 images are taken. The area Vj ¼ Aij mj (4)
N 1 i¼1
is set to 32 32 pixels under the application of a recursive
Nyquist grid. The resolution was adjusted for each experiment
1 XN
as changeover between obstructions made readjustments mj ¼ Aij (5)
N i¼1
necessary. Captured images were combined, Fourier trans-
formed, averaged and debugged in order to determine the PN 2
i¼1 Aij mj
1
final averaged flow field. The generated flow field was VRelative; j ¼ N1 1
P N
(6)
N i¼1 Aij
visualised with Tecplot 360 software (Constellation Software,
Toronto, ON, Canada). The seeding particles used are silver with Aij being the magnitude of the FFT transform of the fre-
coated, hollow glass spheres (D ¼ 10 mm, r ¼ 1400 kg m3) quency j for the ith row which corresponds to the ith buffer.
(Dantec Dynamics, Skovlunde, Ballerup, Denmark). The relative variance VRelative was chosen instead of the vari-
ance as this way absolute values can be weighted by the
2.3. Passive acoustic emission arithmetic mean values. Thus small values with a relative
small absolute changes but large relative changes can be
The acoustic emissions were detected with a piezoelectric considered.
VS375-M sensor (Vallen Systeme GmbH, Icking, Bavaria,
Germany). The sensor is linked to an AEP5H preamplifier
2.4. Supervised machine learning
(Vallen Systeme GmbH, Germany) along with a DCPL2
decoupling unit (Vallen Systeme GmbH, Icking, Bavaria,
Support-vector machines (SVM) are a classifier within ML. The
Germany), a PicoScope 5000 Series oscilloscope (Pico Tech-
basic concept of SVM is the establishment of hyperplanes,
nology Ltd, UK) and a personal computer using PicoScope
seeking to create separation between classes. A hyperplane
version 6.13.15 software (Pico Technology Ltd, UK). One
(for problems higher than 2-dimensional or binary classifica-
hundred recordings, so-called buffers, each of a length of
tion) is the n-dimensional space that creates the boundary of
500 ms, a resolution of 16-bit and an amplitude of maximum
data points for different clusters that belong to different
±1 V were taken. The sampling number was set to 600 kS to
classes.
ensure that the sampling frequency of 1.2 MHz is at least
The general equation for a hyperplane for the case of linear
twice the number of the resonance frequency to meet
separability is given in Eq. (7):
Shannon's theorem (Shannon, 1948). The choice of 500 ms
was justified as a 10 s buffer and was divided as long as the WT X þ b ¼ 0 (7)
lowest recording time was reached where fast Fourier T
transform (FFT) spectra were still visually similar. The sensor where W ¼ ½w1 w2 $$$ wm and b is a scalar. W represents
was placed on the outer side of the pipe wall and it sat on top the normal vector to the hyperplane (W, b) and X being the
of a circular shaped pinhead, located on the tip of the position vector of the points lying on the plane. The classes
obstruction. This meant that sensor and fluid were never in are separated by their value being either positive (þ) or
direct contact, which brings hygienic as well as maintenance negative () measured based upon their distance from the
benefits (see patent application GB1909291.5). hyperplane (Dagher, 2008).
This idea of the hyperplane dates back to the mid-1930s
2.3.1. Data processing (Fisher, 1936), however, a first appearance within the back-
Recorded acoustic emission was converted to make them ground of artificial networks was not done until 1958
available for MATLAB R2018a (MathWorks Inc, Natick, MA, (Rosenblatt, 1958). These initial concepts were further devel-
USA). All data are processed on a computing system with 8-GB oped by Vapnik and Chervonenkis (1974) and this lead to
memory. As a first step, each buffer was assigned with a SVMs. Noble and Street (2006) stated four key concepts behind
corresponding label, describing the experimental condition every SVM. Those are.
(e.g. flow rate 1, obstacle “cross”). Once assigned the data was
abstracted from environmental noise. This was regarded to be The separation of the hyperplane,
frequencies <4 kHz. Literature reports environmental noise to The maximum margin hyperplane,
be in the range 2e6 kHz or below (Chang et al., 2011; Forrest, The soft margin, and
1994). The kernel function.
Discrete Fourier transformation was applied to transform
the time-bound signals into the frequency domain (FFT To make the data available for the MATLAB R2018a Clas-
spectrum). When converted into MATLAB, positive and sification Learner Application (CLA) (MathWorks Inc, Natick,
negative infinitive values appeared, representing over range MA, USA) feature scaling and mean normalisation were
values. These values were filtered and replaced by the applied to the reduced FFT spectrum. The necessity of this
value ± 1 or for those values that are below the detection limit operation was to have features with comparable ranges of
of the sensor the value zero was inserted. values.
The last reduction step was the selection of only the 5,000 Principal component analysis (PCA) (15 components) was
FFT values with the largest relative variance. To get the enabled for the CLA. The goal of the PCA was to project the
b i o s y s t e m s e n g i n e e r i n g 1 9 1 ( 2 0 2 0 ) 4 8 e5 9 53
Fig. 5 e Example contour plots of 2D PIV results for one obstruction (Semicircle) over four flow rates. The Reynolds numbers
for the highest point of restriction are (a) Re ≈ 36,251, (b) Re ≈ 83,701, (c) Re ≈ 126,195 and (d) Re ≈ 176,182. Blanked sections
show the in-pipe obstruction.
54 b i o s y s t e m s e n g i n e e r i n g 1 9 1 ( 2 0 2 0 ) 4 8 e5 9
Fig. 6 e Example contour plots of 2D PIV results for one flow rate (4,530 l h¡1) for all obstruction types. The Reynolds
numbers for the highest point of restriction are (a) Re ≈ 126,195, (b) Re ≈ 126,195, (c) Re ≈ 267,489 and (d) Re ≈ 252,390.
With increasing flow rates the increases in velocity of the Semicircle (Fig. 5 (a)e(d)) (reverse flow from the bottom
magnitude can be observed (Figs. 5 and 6 (a)e(d)). For the semi- to the open area above the obstruction). For the Cone
circular obstruction, a drag up effect is apparent along the tip obstruction (Fig. 6 (d)), preferential flow is shown following the
Fig. 7 e Pressure drop Dp of different obstacles against Reynold number (Re) for the unobstructed pipe.
b i o s y s t e m s e n g i n e e r i n g 1 9 1 ( 2 0 2 0 ) 4 8 e5 9 55
Fig. 8 e Cropped and environmental noise cleared FFT spectra for different obstructions (Cone (red) and Semicircle (black))
but with the same pressure drop Dp (1 kPa).
open space above the cone. When inserting the Cross type 3.4. Acoustic emission signals
obstruction (Fig. 6 (b)) a butterfly wing-like structure in the
flow appears, provoked by vortices immediately downstream The signal was collected from a piezoelectric passive acoustic
of the obstruction. In Fig. 6 (c) jetting occurs for the Three emission sensor. After the removal of environmental noise, a
holes obstruction. The jet increased in width with increasing discrete Fourier transformation was applied to convert the
flowrate. signal from the time-domain into the frequency-domain.
These acoustic “fingerprints” are presented in Figs. 9 and 10.
3.2. Pressure drop measurements Increases in flowrate came with increases in the peak in-
tensity Y. In addition, an increase in flowrate produced wider
The pressure drop was measured 5 mm upstream and 120 mm peaks and more defined shapes in the spectrum. A peak shift
downstream of the obstruction. Plotting the pressure drop Dp towards the lower frequencies is characteristic of increases in
against the Reynolds number for the unobstructed pipe as per flowrate (Fig. 9). At higher frequencies a periodic pattern of
Eq. (3) the following figure can be obtained (Fig. 7). spikes and valleys can be observed and this remains
The obstruction type with the three holes shows a very throughout flowrate increases.
different pressure drop to the other three obstructions. This Figure 10 shows that for each obstruction type at a fixed
will be explained with the high degree of restriction along flow rate of 4,530 l h1 unique acoustic “fingerprints” can be
each hole (D ¼ 2 mm). Overall, all curves are in a parallel retrieved. By building these into a database this sensing
arrangement. For Re z 18,125 the semicircle and the cone technique can be used not only be used to detect blockages in
share a value at 1 kPa. Also for Re z 88,406 the data points are pipe systems, but also determine the type or degree of
nearly congruent as they differ only by 0.7 kPa. This shows obstruction. This information could be used to determine
that it is challenging to distinguish between different runs service points or the cleaning of the system.
purely based on pressure drop measurements.
3.5. Supervised machine learning
3.3. Pressure drop measurements versus acoustics
A total of 22 different ML algorithms were investigated tried,
As per section 3.2, it is difficult to distinguish between different with quadratic SVM delivering best prediction accuracies.
experimental runs when looking at the pressure drop values. Quadratic SVM was also shown to be capable to give highly
However, when linking the pressure drop to acoustic response, accurate predictions with only a limited number of input data
distinguishable responses can be retrieved. Even for the case points. The figure below represents a selection of a few of the
where obstructions share the same pressure drop of 0.1 kPa, it algorithms tested and their performance against the number
is possible to get different obstructions (Fig. 8). of features (Fig. 11).
This indicates that the sensing technique is more sensitive Among all 22 algorithms, sub space KNN performed the
than conventional methods. In addition, the detection of poorest in terms of prediction accuracy and consistency of
these pressure drops can subsequently be used to predict prediction. The prediction accuracies for the ML on unseen
leakages in a very early stage. datasets are presented in Table 1:
56 b i o s y s t e m s e n g i n e e r i n g 1 9 1 ( 2 0 2 0 ) 4 8 e5 9
Fig. 11 e Performance plot for quadratic SVM learner against the number of features (PCA 15 components enabled).
range of the poorest performing prediction and the best pre- the pressure drop is marginal. This makes it a highly sensitive
diction in its average deviation by only 2.5%. Within runs for the technology that might be applied in the future to the early
same cases (e.g. obstruction Cross against Run 1 to 4) the range stage detection of upstream blockages.
is in the 3%. Hence, data is repeatable which is of high impor- Highest prediction accuracy levels in supervised ML are
tance when aiming for an industrial solution to build, for attained when applying the quadratic SVM learner. For all
example a surveillance system for an irrigation system. cases under investigation accuracy levels were in the region of
95% or above with all experiments been repeated four times
and a hundred buffers used for each case.
The combination of the sensor, pipe device and ML can be
4. Conclusions used as an in situ measuring tool for pipe blockages and may
be implemented as a prediction tool for water supply system,
A single passive acoustic emission sensor has been placed on spray systems etc.
a water channel. This channel has been obstructed by Future research will focus on the refining of the algorithms
different geometries, aiming to represent blockages through and create a user interface. Further, scale-up must be
debris build up and a spray nozzle. The different geometries considered. Other work will focus on branched systems and
lead to different flow patterns as supported by 2D PIV. When the maximum detection distance of the sensor system. In-
analysing the acoustic signals, higher flowrates produced field applications for irrigation and spraying systems are to
higher peak values and different obstructions could be be investigated and application to other materials than water
distinguished through different FFT spectra. Even marginally must also be studied.
different or similar pressure drops still deliver distinguishably
different acoustic responses. This technology also makes it
possible for the first time (in reported literature) to measure
such features in a single phased, fully flooded channel. In Declaration of interest
addition this work shows that the proposed technology is also
capable to deliver very different “acoustic fingerprints” when None.
58 b i o s y s t e m s e n g i n e e r i n g 1 9 1 ( 2 0 2 0 ) 4 8 e5 9
Finfer, D., Parker, T. R., Mahue, V., Amir, M., Farhadiroushan, M.,
Acknowledgment & Shatalin, S. (2015). Non-intrusive multiple zone distributed
acoustic sensor flow metering. SPE Annual Technical Conference
Funding Note. and Exhibition, 1e9. https://doi.org/10.2118/174916-MS.
This work was supported by the School of Chemical Fischer, P. (1999). Algorithmisches lernen (1st ed.). https://doi.org/
10.1007/978-3-663-11956-2.
Engineering, University of Birmingham [EPSRC grant number
Fisher, R. A. (1936). The use of multiple measurements in
EP/R513167/1]. taxonomic problems. Annals of Eugenics, 7(2), 179e188. https://
The authors wish to express their gratitude for the support doi.org/10.1111/j.1469-1809.1936.tb02137.x.
received from Philip W Harris and Robert W Sharpe in the Forrest, T. G. (1994). From sender to receiver: Propagation and
engineering workshop at the University of Birmingham. environmental effects on acoustic signals. American Zoologist,
Further thanks goes to Dr Thomas Mills providing necessary 34(6), 644e654. https://doi.org/10.1093/icb/34.6.644.
Ge, Z., Song, Z., Ding, S. X., & Huang, B. (2017). Data mining and
training and enabling access to the 3D printing system. For
analytics in the process industry: The role of machine
technical support, the authors wish to acknowledge Matteo
learning. IEEE Access, 5, 20590e20616. https://doi.org/10.1109/
Antonelli and Giuseppe Forte. ACCESS.2017.2756872.
Gleick, P. H. (2014). Water, drought, climate change, and conflict
in Syria. Weather, Climate, and Society, 6(3), 331e340. https://
references doi.org/10.1175/WCAS-D-13-00059.1.
Guan, R., Wang, X., Marchese, M., Yang, M. Q., Liang, Y., &
Yang, C. (2019). Feature space learning model. Journal of
Adrian, R. J. (1986). Image shifting technique to resolve directional Ambient Intelligence and Humanized Computing, 10(5), 2029e2040.
ambiguity in double-pulsed velocity. Applied Optics, 25(21), https://doi.org/10.1007/s12652-018-0805-4.
3855e3858. Retrieved from http://www.ncbi.nlm.nih.gov/ Hauptmann, P., Hoppe, N., & Püttmer, A. (2002). Application of
pubmed/18235708. ultrasonic sensors in the process industry. Measurement Science
Adrian, R. J. (2005). Twenty years of particle image velocimetry. and Technology, 13(8), 201. https://doi.org/10.1088/0957-0233/
Experiments in Fluids, 39(2), 159e169. https://doi.org/10.1007/ 13/8/201.
s00348-005-0991-7. Hou, R., Hunt, A., & Williams, R. (1999). Acoustic monitoring of
Borenstein, J., & Koren, Y. (1988). Obstacle avoidance with pipeline flows: Particulate slurries. Powder Technology, 106(1e2),
ultrasonic sensors. IEEE Journal of Robotics and Automation, 4(2), 30e36. https://doi.org/10.1016/S0032-5910(99)00051-0.
213e218. https://doi.org/10.1109/56.2085. Hunaidi, O., & Chu, W. T. (1999). Acoustical characteristics of leak
Borodina, I. A., Zaitsev, B. D., & Teplykh, A. A. (2018). The signals in plastic water distribution pipes. Applied Acoustics,
influence of viscous and conducting liquid on the 58(3), 235e254. https://doi.org/10.1016/S0003-682X(99)00013-4.
characteristics of the slot acoustic wave. Ultrasonics, 82, 39e43. Khalifa, A. E., Chatzigeorgiou, D. M., Youcef-Toumi, K.,
https://doi.org/10.1016/j.ultras.2017.07.011. Khulief, Y. A., & Ben-Mansour, R. (2010). Quantifying acoustic
Boyd, J. W. R., & Varley, J. (2001). The uses of passive and pressure sensing for in-pipe leak detection. Sound,
measurement of acoustic emissions from chemical Vibration and Design, 13, 489e495. https://doi.org/10.1115/
engineering processes. Chemical Engineering Science, 56(5), IMECE2010-40056.
1749e1767. https://doi.org/10.1016/S0009-2509(00)00540-6. Khulief, Y. A., & Khalifa, A. (2012). On the in-pipe measurements
Butterfield, J. D., Collins, R. P., Krynkin, A., & Beck, S. B. M. (2017). of acoustic signature of leaks in water pipelines. Vibration,
Experimental investigation into the influence of backfill types Acoustics and Wave Propagation, 12, 429. https://doi.org/10.1115/
on the vibro-acoustic characteristics of leaks in MDPE pipe. IMECE2012-86731.
Procedia Engineering, 186, 311e318. https://doi.org/10.1016/ Kotsiantis, S. B., Zaharakis, I. D., & Pintelas, P. E. (2007).
j.proeng.2017.03.252. Supervised machine learning: A review of classification
Butterfield, J. D., Krynkin, A., Collins, R. P., & Beck, S. B. M. (2017). techniques general issues of supervised learning algorithms.
Experimental investigation into vibro-acoustic emission 31, 501e520.
signal processing techniques to quantify leak flow rate in Kotze , R., Ricci, S., Birkhofer, B., & Wiklund, J. (2016). Performance
plastic water distribution pipes. Applied Acoustics, 119, tests of a new non-invasive sensor unit and ultrasound
146e155. https://doi.org/10.1016/j.apacoust.2017.01.002. electronics. Flow Measurement and Instrumentation, 48, 104e111.
Chang, T.-Y., Liu, C.-S., Huang, K.-H., Chen, R.-Y., Lai, J.-S., & https://doi.org/10.1016/j.flowmeasinst.2015.08.013.
Bao, B.-Y. (2011). High-frequency hearing loss, occupational Koza, J. R., Bennett, F. H., Andre, D., & Keane, M. A. (1996).
noise exposure and hypertension: A cross-sectional study in Automated design of both the topology and sizing of analog
male workers. Environmental Health, 10(1), 35. https://doi.org/ electrical circuits using genetic programming. In , Vol. 96.
10.1186/1476-069X-10-35. Artificial intelligence in design (pp. 151e170). https://doi.org/
Chatzigeorgiou, D., Youcef-Toumi, K., & Ben-Mansour, R. (2015). 10.1007/978-94-009-0279-4_9.
Design of a novel in-pipe reliable leak detector. IEEE, 20(2), Kubat, M. (2015). An introduction to machine learning. https://
824e833. https://doi.org/10.1109/TMECH.2014.2308145. doi.org/10.1007/978-3-319-20010-1.
Chong, M., Abraham, A., & Paprzycki, M. (2004). Traffic accident Li, S., Song, Y., & Zhou, G. (2018). Leak detection of water
data mining using machine learning paradigms. In Fourth distribution pipeline subject to failure of socket joint based on
International Conference on Intelligent Systems Design and acoustic emission and pattern recognition. Measurement, 115,
Applications (ISDA’04), Hungary (pp. 415e420). 39e44. https://doi.org/10.1016/j.measurement.2017.10.021.
Dagher, I. (2008). Quadratic kernel-free non-linear support vector Liu, C., & Wang, W. (2017). Contextual regression: An accurate
machine. Journal of Global Optimization, 41(1), 15e30. https:// and conveniently interpretable nonlinear model for mining
doi.org/10.1007/s10898-007-9162-0. discovery from scientific data. BioRxiv. https://doi.org/10.1101/
Das, S., Das, R., & Mazumdar, A. (2013). Circulation characteristics 210997.
of horseshoe vortex in scour region around circular piers. Lohmar, B., Wang, J., Rozelle, S., Huang, J., & Dawe, D. (2003).
Water Science and Engineering, 6(1), 59e77. https://doi.org/ China's agricultural water policy reforms: Increasing investment,
10.3882/j.issn.1674-2370.2013.01.005. resolving conflicts, and revising incentives.
b i o s y s t e m s e n g i n e e r i n g 1 9 1 ( 2 0 2 0 ) 4 8 e5 9 59
Martini, A., Rivola, A., & Troncossi, M. (2018). Autocorrelation Technology, 45, 34e42. https://doi.org/10.1016/
analysis of vibro-acoustic signals measured in a test field for j.tust.2014.09.003.
water leak detection. Applied Sciences, 8(12), 2450. https:// Reitmaier, T. (2015). Aktives Lernen für Klassifikationsprobleme unter
doi.org/10.3390/app8122450. der Nutzung von Strukturinformationen (Band 5 von). Kassel:
Martini, A., Troncossi, M., & Rivola, A. (2016). Leak detection in Kassel University Press GmbH.
water-filled small-diameter polyethylene pipes by means of Rosenblatt, F. (1958). The perceptron: A probabilistic model for
acoustic emission measurements. Applied Sciences, 7(1), 2. information storage and organization in the brain. Psychological
https://doi.org/10.3390/app7010002. Review, 65(6), 386e408. https://doi.org/10.1037/h0042519.
Masud, M. M., Chen, Q., Gao, J., Khan, L., Han, J., & Sause, M. G. R., Gribov, A., Unwin, A. R., & Horn, S. (2012). Pattern
Thuraisingham, B. (2010). Classification and novel class recognition approach to identify natural clusters of acoustic
detection of data streams in a dynamic feature space. In Joint emission signals. Pattern Recognition Letters, 33(1), 17e23.
European conference on machine learning and knowledge discovery https://doi.org/10.1016/j.patrec.2011.09.018.
in databases (6322nd ed., pp. 337e352). https://doi.org/10.1007/ Shannon, C. E. (1948). A mathematical theory of communication.
978-3-642-15883-4_22. Bell System Technical Journal, 27(3), 379e423. https://doi.org/
Moraru, A., Pesko, M., Porcius, M., Fortuna, C., & Mladenic, D. 10.1002/j.1538-7305.1948.tb01338.x.
(2010). Using machine learning on sensor data. Journal of The Economist Newspaper Limited. (2016). Precision agriculture:
Computing and Information Technology, 18(4), 341. https:// TV dinners. Economist, 3.
doi.org/10.2498/cit.1001913. Thenkabail, P. S., Hanjra, M. A., Dheeravath, V., & Gumma, M.
Noble, W. S., & Street, P. (2006). What is a support vector (2017). Global croplands and their water use from remote
machine? Nature Biotechnology, 24(12), 1565e1567. sensing and nonremote sensing perspectives. In Q. Weng (Ed.),
O'Keefe, C. V., Maron, R., Felix, J., van der Spek, A. M., & Advances in environmental remote sensing: Sensors, algorithms,
Rothman, P. (2010). Non-invasive passive array technology for and applications (1st ed., pp. 383e421). https://doi.org/10.1201/
improved flow measurements of slurries and entrained air. In b10599.
C. D. R. A. Holdings (Ed.), The 4th International platinum (2018). General overview of the consequences of
Treszkai, A.
conference: Platinum in transition ‘Boom or Bust’ (pp. 21e30). african water conflicts. UKH Journal of Social Sciences, 2(2),
Retrieved from https://www.saimm.co.za/Conferences/ 11e18. https://doi.org/10.25079/ukhjss.v2n2y2018.pp11-18.
Pt2010/021-030_OKeefe.pdf. Vapnik, V. N., & Chervonenkis, A. Y. (1974). Teoriya raspoznavaniya
, G. (1992). Vibroacoustical energy flow through straight
Pavic obrazov: Statisticheskie problemy obucheniya [Theory of pattern
pipes. Journal of Sound and Vibration, 154(3), 411e429. https:// recognition: Statistical problems of learning]. Moscow.
doi.org/10.1016/0022-460X(92)90776-T. Wuest, T., Weimer, D., Irgens, C., & Thoben, K.-D. (2016). Machine
Rahman, M., Håkansson, U., & Wiklund, J. (2015). In-line learning in manufacturing: Advantages, challenges, and
rheological measurements of cement grouts: Effects of water/ applications. Production & Manufacturing Research, 4(1), 23e45.
cement ratio and hydration. Tunnelling and Underground Space https://doi.org/10.1080/21693277.2016.1192517.