Content uploaded by Ronny Francis Ribeiro Junior
Author content
All content in this area was uploaded by Ronny Francis Ribeiro Junior on Dec 03, 2021
Content may be subject to copyright.
Available via license: CC BY 4.0
Content may be subject to copyright.
sensors
Article
An Ultrasonic-Capacitive System for Online Characterization of
Fuel Oils in Thermal Power Plants
Mateus Mendes Campos 1,2, Luiz Eduardo Borges-da-Silva 2, Daniel de Almeida Arantes 2,
Carlos Eduardo Teixeira 1, Erik Leandro Bonaldi 1, Germano Lambert-Torres 1, Ronny Francis Ribeiro Junior 1,
Gabriel Pedro Krupa 1, Wilson Cesar Sant’Ana 1,* , Levy Ely Lacerda Oliveira 1and Renato Guth de Paiva 3
Citation: Campos, M.M.;
Borges-da-Silva, L.E.; Arantes, D.d.A.;
Teixeira, C.E.; Bonaldi, E.L.;
Lambert-Torres, G.; Ribeiro Junior,
R.F.; Krupa, G.P.; Sant’Ana, W.C.;
Oliveira, L.E.L.; et al. An
Ultrasonic-Capacitive System for
Online Characterization of Fuel Oils
in Thermal Power Plants. Sensors
2021,21, 7979.
https://doi.org/10.3390/s21237979
Academic Editor: Pedro M. Ramos,
Olfa Kanoun and Pasquale Arpaia
Received: 4 October 2021
Accepted: 24 November 2021
Published: 29 November 2021
Publisher’s Note: MDPI stays neutral
with regard to jurisdictional claims in
published maps and institutional affil-
iations.
Copyright: © 2021 by the authors.
Licensee MDPI, Basel, Switzerland.
This article is an open access article
distributed under the terms and
conditions of the Creative Commons
Attribution (CC BY) license (https://
creativecommons.org/licenses/by/
4.0/).
1
Gnarus Institute, R&D Department, Itajuba 37500-052, MG, Brazil; mateusmcampos@unifei.edu.br (M.M.C.);
carlos.teixeira@institutognarus.com.br (C.E.T.); erik@institutognarus.com.br (E.L.B.);
germano@institutognarus.com.br (G.L.-T.); ronny@institutognarus.com.br (R.F.R.J.);
gabriel@institutognarus.com.br (G.P.K.); levy@institutognarus.com.br (L.E.L.O.)
2Institute of Engineering Systems and Information Technology, Itajuba Federal University, Pro-Reitoria de
Pesquisa e Pos-Graduacao (PRPPG), Itajuba 37500-903, MG, Brazil; leborges@unifei.edu.br (L.E.B.-d.-S.);
daniel_arantes@unifei.edu.br (D.d.A.A.)
3Norte Energia S.A., Vitoria do Xingu 68383-000, PA, Brazil; renatopaiva@norteenergiasa.com.br
*Correspondence: wilson_santana@ieee.org
Abstract:
This paper presents a ultrasonic-capacitive system for online analysis of the quality of fuel
oils (FO), which are widely used to produce electric energy in Thermal Power Plants (TPP) due to
their elevated heating value. The heating value, in turn, is linked to the quality of the fuel (i.e., the
density and the amount of contaminants, such as water). Therefore, the analysis of the quality is
of great importance for TPPs, either in order to avoid a decrease in generated power or in order to
avoid damage to the TPP equipment. The proposed system is composed of two main strategies: a
capacitive system (in order to estimate the water content in the fuel) and an ultrasonic system (in
order to estimate the density). The conjunction of the two strategies is used in order to estimate the
heating value of the fuel, online, as it passes through the pipeline and is an important tool for the
TPP in order to detect counterfeit fuel. In addition, the ultrasonic system allows the estimation of
the flow rate through the pipeline, hence estimating the amount of oil transferred and obtaining the
total mass transferred as a feature of the system. Experimental results are provided for both sensors
installed in a TPP in Brazil.
Keywords:
heat value; heavy fuel oil; thermal power plant; ultrasonic-capacitive system; water content
1. Introduction
The Brazilian electrical energy mix, since the 1990s, has shown a strong growth
in Thermal Power Plants (TPP) based on fossil fuels [
1
]. According with the Brazilian
regulatory agency for energy (ANEEL) [
2
], the granted electrical power to fossil fuels is
approximately 35 GW (representing about 16% of the Brazilian energy mix). Among the
fossil fuels, fuel oils (FOs, which include heavy and light fuel oils) represent a share of 23%
(around 8 GW).
The efficiency of power generation in a TPP is linked with the quality of the FOs,
besides the fact that oils with contaminants may induce to long-term damage to the equip-
ment in the TPP. The most common contaminant is water. The process of contamination
might be unintentional (due the processes involved in extraction and storage)—however,
it might also be due to counterfeit. The counterfeit/adulteration of a high-quality fuel
with some other substance of less quality and cheaper economical value is a common (and
deceptive) practice around the world, as presented in scientific reports from Bangladesh
[
3
], Ghana [
4
], India [
5
], and Tanzania [
6
]. In these aforementioned references, gasoline
or diesel is mixed with subsidized kerosene or lubricants and affect automotive engines.
In the case of the Brazilian TPPs, the fuel oils are usually loaded from tank trucks. These
Sensors 2021,21, 7979. https://doi.org/10.3390/s21237979 https://www.mdpi.com/journal/sensors
Sensors 2021,21, 7979 2 of 25
trucks are filled with an agreed amount of oil at a distribution center or refinery. During
the transportation, it is reported [
7
] that some truck drivers remove a small quantity of oil
(in order to sell it in the black market) and replace this amount with water.
The literature presents some methods to detect damage on the internal combustion
engines that convert the chemical energy of the fuels into rotational mechanical energy at
their crankshafts (which are tied to synchronous generators in order to produce electrical
energy). Such methods may rely on the direct measurement of internal parameters of the
engine (such as pressure and temperature at different parts of the engine) [
8
,
9
] or even the
use of voltage and current signals of the synchronous generator [10] in order to indirectly
detect mechanical issues within the engine. However, these methods cannot test the fuel
quality, which may cause a problem that will only be manifested in the long term.
In order to check the quality of their fuels, the Brazilian TPPs perform laboratory
analyses (which follow the standards ASTM D95 [
11
] for water content, ASTM D1298 [
12
]
for density, and ASTM D4868 [
13
] for heat value) on samples. However, the samples taken
are only a small fraction of the total amount of fuel that is loaded to the TPP pipelines.
Moreover, in the case of TPPs located far from the large centers (which is a common
scenario in the North and Northeast regions of Brazil), the laboratory tests are not even
performed (due to lack of certified laboratories nearby). Hence, the use of online methods
for the estimation of fuel quality (especially heavy fuel oils—HFO) is a general demand
from the TPP shareholders.
1.1. Literature Review
1.1.1. Online Measurement of Water Content in Fuel Oils
The work of Han
et al. [14]
presents a method based on differential pressure. Its
advantage is that it this is a simple (and low cost) method, requiring only a differential
pressure transducer. However, the accuracy of this method for estimating water content is
strongly dependent on the difference in the specific mass of fuel oil and water.
The literature also presents methods based on absorption of electromagnetic waves,
such as infrared (as presented by Al-Saiyed
et al. [15]
) and gamma ray (as presented by
Scheers
[16]
and Johansen and Tjugum
[17]
). The main advantage of methods based on
electromagnetic waves is the frequency range that can be used, whose wavelengths reach
molecular levels, detecting the fraction of water in molecules as well as other contami-
nants, sulfur, vanadium, etc. However, these techniques are more common for laboratory
applications that use small samples of fuel oils to be analyzed and obtain their characteris-
tics. Considering online models, these techniques are limited to small ducts (from mm to
few cm) [
18
], which differs from the applications proposed in the work (where the ducts
are larger than 3”). Another drawback of these techniques is the cost of processing.
The works of Gregory and Clarke
[19]
and Makeev
et al. [20]
present a method based
on microwaves. The main advantage is the possibility to measure the whole range of water
fractions (from 0% to 100%, with an extra advantage at the precision at lower fractions).
However, this technology has a high initial cost in relation to the others, and it is also
sensitive to variations in salinity in the range of higher fractions of water.
The work of Teixeira
et al. [21]
presents a method based on ultrasound. The advantage
of this method is that it allows the gathering of information online, adaptable to conduits
with larger diameters. However, the disadvantages are a relatively high cost and also the
need for a cumbersome calibration process in order to classify water percentages.
Another technology for the estimation of the water content is the measurement of
the capacitance of the emulsion. The fluid being measured acts as the dielectric of the
sensor, making it possible to measure the electrical permittivity of the fluid and, conse-
quently, the capacitance of the sensor–fluid system. The percentage of water in oil can be
calculated based on a predictable relationship in the electrical properties of the emulsions.
Hammer et al. [
22
] proposed a non-intrusive capacitive sensor, capable of measuring water
contents in the range between 0% and 80%. Shuller et al. [
23
] proposed a measurement
principle of the dielectric properties of the emulsion based on a single electrode excited by
Sensors 2021,21, 7979 3 of 25
an oscillator (in the range of Mega Hertz). Libert et al. [
24
,
25
] presented a system based on
capacitive sensors installed at the internal wall of the pipes.
Systems such as EIT (Electrical Impedance Tomography) or ECT (Electrical Capac-
itance Tomography) are techniques widely researched and disseminated in the litera-
ture
[26–28]
with great advances in the field of flow measurement and physical– characteri-
zation of multiphase fluids. The advantages found in these types of technique are mainly
focused on the potential to be applied in multiphase fluids, to present the flow image in
the penstock cross section and the possibility to be used on flow conditions unfavorable
for other methods
[26–28]
. The disadvantages, however, are linked to the development of
more consolidated techniques for commercial application, in addition to the complexity of
electronics [26].
The capacitive sensor used in this current work was proposed by the authors of
[
29
] and is based on paralleled disks, which are immersed in the emulsion inside the
pipeline. The advantage of the system proposed in [
29
] is the electronic topology based
on a Howland current source, capable of eliminating parasitic effects caused by influences
from cables, connectors, and constructive characteristics of the capacitive sensor that affect
the capacitance measurement. In addition, as will be seen in Section 2.2, the sensor is
based on several disks in order to perform a paralleled association of the capacitances,
which increases the total capacitance of the sensor. The fluid to be inspected fills the spaces
between each pair of disks, forming small associated capacitors. Higher capacitance values
allow it to have greater precision in measurements, as their proportion to the parasitic
elements (such as those from cables, traces of plates and electronic parts) is greater, thus
reducing the magnitude of errors caused by them.
1.1.2. Online Measurement of Density in Fuel Oils
The literature presents several methods for measuring the density: many methods
are based on direct physical relationships and others are based on indirect parameter
determinations [
30
]. Most of the established methods have limitations inherent to the sys-
tem, which often result in application restrictions in the sensing implementation (physical
process limitations such as tube diameter, limitations in process deviations, limitation in
the flow velocity of the fluid to be monitored, and others) [
31
]. In other cases, the presence
of bubbles, particles, or incrustations may cause problems in specific mass measurements.
Methods involving ultrasonic waves appear as alternatives to replace the standard meth-
ods, where these cannot be applied [
32
]. The most widespread ultrasonic techniques for
measurement of density are based on three different principles: speed of sound in the
propagation medium, acoustic impedance, and propagation on waveguides.
Waveguide approaches generally use time variations in the propagation of ultrasonic
waves on a transmission line immersed in the measurement medium [33].
The acoustic impedance method approach is based on the determination of the reflec-
tion and transmission coefficient and the transmission coefficients of flat ultrasonic waves
that propagate through different media [32].
Considering the determination of the density of any medium just by measuring the
speed of sound, Swoboda et al. [
34
] used measurements of speed of sound and temperature
in a sulfuric acid solution and found that the assessment of density was possible over
a relative range greater than 1.10–1.30 and over a temperature range greater than 10
◦
C
to 50
◦
C. The density and speed of sound are quantities that vary with temperature and
pressure [
35
]. For determination of density in controlled systems, where there is little
variation in temperature, pressure, and knowledge of the compressibility of the medium to
be measured, the application of this concept is quite interesting, as it reduces the complexity
of electronics in measuring the speed of sound and, by consequence, in the calculation of
the density.
Sensors 2021,21, 7979 4 of 25
1.2. Aim and Objectives
The proposition of this work is to present a methodology for measurement of the
quality of fuel oils based on two different strategies (capacitive and ultrasonic), in order
to estimate the water content and heating values. The methodology is applied in an
ultrasonic-capacitive hybrid on-line measurement system, which analyzes the fuel oil
constantly throughout the entire fueling process. The advantage of this system is that it
estimates the fuel oil quality parameters in real time, unlike the methods currently applied
in Brazilian TPPs, which are based on laboratory analysis of small samples of fuel oil.
However, laboratory results are not instantaneous, and some TPPs in Brazil do not even
have certified laboratories nearby. Hence, the proposition of this paper allows for an
immediate estimation of the fuel quality, while the fuel oil is flowing through the pipelines.
The originality of this paper lies precisely in the application of the proposed system at
the loading bays of thermoelectric power plants for online monitoring of fuel oil used for
energy generation.
1.3. Outline of the Paper
The paper is organized as follows: Section 2presents the fundamental concepts about
the two main strategies (capacitive and ultrasonic) and its correlation in order to infer about
the fuel quality. Section 3presents the experimental results obtained with the prototype
installed in a Brazilian power plant. Finally, Section 4presents the main conclusions of the
work and some opportunities for future research.
2. Materials and Methods
Figure 1presents the conceptual drawing of the proposed system. This system must be
installed in a section of the pipeline between the TPP fuel storage and the point where the
tank truck discharges the oil fuel. Section 2.1 briefly discusses the main characteristics of an
emulsion of water-oil. As the emulsion passes through the pipe, the proposed capacitive
sensor (described in detail in Section 2.2) estimates the water content of the emulsion ,and
the proposed ultrasonic sensors (described in detail in Section 2.3) estimate the density and
the flow rate. The box indicated as “control boards” contains an FPGA board that controls
and receives data sent from the capacitive system control board, ultrasonic system control
board, and, also, from the temperature sensor. Through ethernet, the FPGA makes the
data available to a remote analysis software, where the information of the aforementioned
sensors is used to estimate the water content, the density and the heating value (hence, the
quality) of the oil fuel. Section 2.4 presents the proposed procedure to estimate the heating
value.
Figure 1. Conceptual 3D drawing of the proposed system.
Sensors 2021,21, 7979 5 of 25
2.1. Characteristics of Fuel Oils
Fuel oils (FO) can be categorized in relation to viscosity as Light Fuel Oils (LFO) and
Heavy Fuel Oils (HFO). FOs also have several other parameters that are used in their
characterization—three of them, particularly in this work, are of special interest:
•Water content
is probably the main contaminant of fuel oils, since contamination can
occur in all parts of the production process, from the moment of extraction of crude oil
to the moment of sending it to the TPPs [
21
]. The lower the water-in-oil content, the
better the oil quality for power generation. The Resolution number 3 of the Brazilian
National Agency of Petroleum, Natural Gas and Biofuels (ANP) stipulates a maximum
of 2% water-in-oil content [36].
•Density
is the ratio of the mass of oil to its volume. The density of the FO is an
important property, which directly influences its combustion power and in the energy
production, as well as in the removal of contaminants in centrifuges [
37
]. Density can
be used to define other intrinsic characteristics of FO, such as the heat of combustion.
•Heating Value
shows the amount of energy that could be obtained from fuel combus-
tion. It is, usually defined in terms of a Higher Heating Value (HHV, or Gross Heat
of Combustion (GHC)) and a Lower Heating Value (LHV, or Net Heat of Combus-
tion (NHC)). The difference between the two is that HHV includes the latent heat
of evaporation from the water vapor formed during combustion, while the LHV
considers the loss of the latent heat needed to evaporate the water formed during
combustion [
38
,
39
]. The LHV is the indicator used by Brazilian TPPs in order to
estimate the generation capacity [
36
] and is the one that is going to be estimated in
Section 2.4.
2.2. Estimation of Water Content through Capacitance
The estimation of water content in fuel oils is a welcomed measure in the power
generation industry, as the presence of this contaminant reduces the efficiency of energy
production. Moreover, the presence of water in fuel oil tends to be corrosive to system
equipment, also causing environmental problems. Still, from an economic point of view,
water has no economic value for energy generation.
The authors proposed in [
29
] a capacitive sensor based on paralleled disk, as presented
in Figure 2. As the sensor is immersed into the emulsion, the fluid act as the dielectric of an
array of capacitors. The importance of having an array of paralleled capacitors is that higher
values of capacitance increase the precision of the measurements, as the parasitic elements
(from the cables, circuit boards, etc.) have less influence on the process. In addition, the
proposed circuit uses a Howland current source topology, with enhancements that allow
its continuous oscillating operation and easy adjustment of output current and frequency.
Figure 2. Proposed capacitive sensor and its equivalent circuit.
Sensors 2021,21, 7979 6 of 25
The sensor must be installed in a transversal section of the pipeline, as presented in
Figure
3
. Its sensitive region is limited to the metallic disks that will be filled with the fluid.
The length of this sensitive region is 8 cm.
Figure 3. Proposed capacitive sensor installed in a transversal section of the pipeline.
The permittivity of water is approximately 40 times greater than the permittivity of
the fuel oil [
40
,
41
]—hence, the greater the water content in the emulsion, the greater its
measured capacitance will be.
In order to estimate the water content from the measured capacitance, a calibration
procedure is performed, based on the capacitances of four emulsion samples with known
water percentages (measured with a Karl Fischer laboratory equipment). (The calibration
procedure must be performed for each different sensor, as there might occur some mechan-
ical differences (such as the diameter of the metallic disks and the spaces between them)
from one sensor to another.) Table 1presents the data obtained in the calibration of the
prototype of the capacitive sensor (used in the experimental results of Section 3). The left
column contains the water concentrations (obtained in a Karl Fischer laboratory test) of the
four fuel samples (prepared with pure HFO mixed with different amounts of water), and
the right column contains the respective capacitances measured with the sensor.
Table 1. Calibration example of the capacitive sensor.
Karl Fischer Water Content Measured Capacitance
0.19% 261.8 pF
0.40% 286.1 pF
1.01% 353.2 pF
2.20% 482.9 pF
Through a linear interpolation of the data of Table 1, Equation (1) is obtained. This
equation relates a given measurement of capacitance (in pF) to the water content (in %).
fwater%=0.0087 ·CpF −2.0712 , (1)
Sensors 2021,21, 7979 7 of 25
where
CpF
is the measured capacitance value (in pF) and
fwater%
is the estimated water
content (in%).
2.3. Estimation of Density and Flow through Ultrasonic System
2.3.1. Density
According to [
42
,
43
], ultrasonic methods can determine the density
ρu
of a fluid
through the use of the Newton–Laplace Equation (2) and measurement of the speed of
sound in that medium.
ρu=1
κs·c2=B
c2, (2)
where
κs
is the isentropic (adiabatic) compressibility of the fluid and
c
is the speed of sound
at this fluid. Usually, the compressibility is also expressed in terms of its reciprocal, i.e., the
bulk modulus B=1/κs[32,44].
The speed of sound can be obtained either through a pulse-echo system [
45
] or through
a system of two transducers (emitter–receiver) [
46
,
47
]. Both principles are based on the
measurement of the wave propagation time in the fluid, differing only in the number of
ultrasonic transducers used.
In case of a pulse-echo system, the basic principle involves the use of an ultrasonic
transducer that outputs a very short pulsed signal of the proper ultrasonic frequency.
The ultrasonic wave travels through the propagation medium to the other side, where a
reflector material is positioned. Hence, there is a reflection of this ultrasonic wave, causing
a large part of the energy of the ultrasonic wave to return through the medium to the
transducer that emitted it.
In case of an emitter–receiver system, there are two transducers positioned on separate
sides forming an acoustic trajectory between them. While one ultrasonic transducer emits,
the other receives the ultrasonic signal that has propagated through the measurement
medium.
In the proposed system, multiple acoustic trajectories are used, as illustrated in Figure 4.
In the figure, two transducers are opposed at the conduit and form an acoustic trajectory
of length
L
and angle
ϕ
with the conduit wall. During transmission, the ultrasonic pulse
that travels in favor of the fluid flow through the conduit (in downstream direction) travels
the distance
L
in a time period
tdown
smaller than the pulse that travels in the opposite
direction (against the fluid flow, in upstream direction) in a time tup .
Sensors 2021,21, 7979 8 of 25
Figure 4.
Measurement of time-of-flight in a conduit using multiple acoustic trajectories of ultra-
sonic transducers.
Considering that, in Figure 4,
~
vax
is the axial velocity of the fluid flowing through the
conduit (whose amplitude has a value of
¯
vax
), its projection on the trajectory
L
(at an angle
ϕ
with the flow direction) is given by
¯
vax ·cos ϕ
. Hence, whenever there is a fluid flowing
through the conduit, the speed of the ultrasonic pulses downstream (
vdown
) and upstream
(vup ) can be calculated as (3) and (4), respectively.
vdown =c+¯
vax ·cos ϕ, (3)
vup =c−¯
vax ·cos ϕ. (4)
In addition, the propagation times of the ultrasonic pulses downstream (
tdown
) and
upstream (tup ) can be calculated as (5) and (6), respectively.
tdown =L
c+¯
vax ·cos ϕ, (5)
tup =L
c−¯
vax ·cos ϕ. (6)
Combining Equations (5) and (6), it can be shown that the speed of ultrasonic wave
c
can be obtained as a function of the downstream and upstream travel times as per Equation
(7).
c=L
2·tup +tdown
tup ·tdown
. (7)
As presented in Figure 1, the ultrasonic system has three layers of planes equivalent
to Figure 4, and each layer has two crossed ultrasonic trajectories. Hence, Equation (7)
Sensors 2021,21, 7979 9 of 25
can be modified to (8), in order to obtain an average velocity
¯
c
considering each of the six
trajectories
Li
and each of the six pairs of downstream and upstream times (
tdown,i
and
tup,i, respectively).
¯
c=1
6
6
∑
i=1
Li
2·tup,i+tdown,i
tup,i·tdown,i
, (8)
where iis an index corresponding to one of the six trajectories.
The technique used to measure the ultrasonic wave propagation times is based on the
zero-crossing detection [
48
] of the pulses at the receivers. Figure 5presents the detection of
the zero-crossing of an ultrasonic wave. After the transmission of the ultrasonic pulse, the
FPGA starts counting the time of several zero-crossings at the receiver. In order to avoid any
initial transient, the first cycles are skipped. This is performed by detecting a pre-defined
threshold (indicated by (1) in the figure). After the threshold, more three zero-crossings are
skipped (until the first valid zero-crossing, indicated by (2) at the figure). Then the time
from the starting of the pulse until next
M
number of zero-crossings (
tToF1
until
tToFM
) are
computed. The average time of the valid zero crossings is calculated (recursively) using
Equation (9).
Figure 5. Measurement of the transit time through multiple zero crossings.
¯
tToF =1
M· tToF1+
M
∑
j=2tToFj −(j−1)·¯
tc/2!, (9)
where
¯
tToF
is the average of the zero crossing times. It is important to note that
¯
tToF
can be
any of the
tup,i
or
tdown,i
times in Equation (8). Furthermore,
¯
tc/2
is the average time of the
half cycle of oscillation of the signal, and is calculated with (10).
¯
tc/2 =1
M−1·
M−1
∑
j=1tToFj+1−tToF j . (10)
Finally, the average speed of the ultrasound waves (calculated as (8)) at the fluid is
used in Equation (2) in order to obtain the density ρuof the fluid.
Sensors 2021,21, 7979 10 of 25
2.3.2. Flow Rate
According with [
49
,
50
], the flow rate
Qv
using multiple acoustic trajectories arranged
at different heights in the cross section of the conduit can be calculated as (11). This
calculation performs a numerical integration of
N
trajectories arranged in distinct heights
in the cross section of the penstock. The most used integration methods [
51
] are: Gauss–
Legendre (for rectangular sections) and Gauss–Jacobi (for circular sections). The OWICS
(Optimal Weighted Integration, for circular sections) and the OWIRS (Optimal Weighted
Integration for rectangular sections) [
48
] can also be highlighted. What differentiates the
methods are the speed profiles [
49
,
52
], whereas the Gauss–Legendre and the Gauss–Jacobi
propose uniform speed profiles [51].
Qv=D
2·
N
∑
i=1
wi·¯
vax,i(zi)·Li·sin(ϕi), (11)
where
zi
is the position of the trajectory in relation to the cross section of the conduit.
¯
vax,i(zi)
is the average axial speed along trajectory
i
, calculated according to Equation (12).
wi
are
weighting coefficients (which depend on the number of trajectories and on the integration
technique in use, according to Equation (13)).
D
is the dimension of the parallel conduit for
the intersection of two acoustic planes (i.e., the diameter of a circular section conduit).
N
is
the total number of acoustic trajectories in a measurement plane.
Li
is the distance (length)
of the acoustic trajectory i, and ϕiis the angle between the acoustic trajectories.
¯
vax,i(zi) = Li
2·cos(ϕi)·tup −tdown
tup ·tdown
, (12)
where
tup
is the measured propagation time upstream and
tdown
is the measured propaga-
tion time downstream. Their difference is known as ∆t=tup −tdown .
According to [49,51], the weighting coefficients can be calculated as (13).
wi=1
1−4·z2
i
D2κ·2
D·ZD/2
−D/2 1−4·z2
i
D2!κ
·
L
i.dz , (13)
where the parameter
κ
is dependent on the type of integration used (
κ=
0.5 for the Gauss–
Legendre method and
κ=
0.6 for the OWICS method, considering circular cross-section
conduits in both cases). The variable
L
i
is the Lagrange polynomial integration, calculated
as (14).
L
i(z)=
N
∏
j=1
j6=i
z−zj
zi−zj
. (14)
The developed system was based on a measurement configuration using six acoustic
trajectories introduced in a circular spool, as presented in Figure 6. The acoustic trajectories
are distributed in two crossed measurement planes (
A
and
B
), containing three trajectories
in each plane. Therefore, the flow rate for a measurement plan (
A
or
B
) is determined from
Equation (11) and results in (15). Furthermore, each sensor forms an angle of
ϕi=
45
◦
with
the conduit wall.
Sensors 2021,21, 7979 11 of 25
Figure 6. Ultrasonic sensors forming two planes of three trajectories.
QA|B=D
2·
3
∑
i=1
wi·¯
vax,i(zi)·Li·sin(ϕi). (15)
Using (15) for both planes
A
and
B
, the resultant flow rate can be obtained as the
average (16).
Qv=QA+QB
2. (16)
Each of the two planes has a trajectory positioned at the center of the circular spool and
two other trajectories equidistant from the center, forming the positions
z1=−
0.707107
·D/
2,
z2=
0 and
z3= +
0.707107
·D/
2. For each trajectory, based on Equations (13) and (14), the
respective weights
wi
are presented in Table 2, either considering Gauss–Jacobi or OWICS
integration methods.
Table 2. Weighting coefficients for a circular cross section conduit.
Trajectory iPosition zi/(D/2)wifor Gauss–Jacobi wifor OWICS
1−0.707107 0.555360 0.546150
2 0 0.785398 0.792715
3 +0.707107 0.555360 0.546150
Table 3presents the measured lengths (
Li
) of the trajectories for either planes
A
and
B
.
Ideally, the lengths
L1
and
L3
of both planes should have the same value, as well as the
lengths
L2
of both planes. However, due to a not so precise drilling, there are these 1 mm
discrepancies.
Table 3. Lengths of the acoustic trajectories.
Trajectory iLength Lifor Plane A Length Lifor Plane B
1 162.26 mm 163.54 mm
2 203.00 mm 202.83 mm
3 163.05 mm 162.93 mm
2.3.3. Evaluation Setup
In order to evaluate the ultrasonic system, a reduced scale flow setup has been assem-
bled, as presented in Figure
7
. This setup is formed by a pump driven by an electric motor
(whose speed is controlled by an inverter) and a closed loop hydraulic system.
Sensors 2021,21, 7979 12 of 25
Figure 7. Evaluation setup for the ultrasonic sensors.
With the setup of Figure
7
, it was possible to vary the flow of the hydraulic circuit
by controlling the rotation of the motor that drives the pump. In total, five motor speeds
were used (0 RPM, 450 RPM, 500 RPM, 600 RPM, and 700 RPM) in order to achieve a more
stable flow in the system. The motor speeds are reduced to the pump using a gear train (of
ratio 2.54).
Water has been used as the circulating fluid, as it is it is cheaper than oil and its
physical–chemical parameters are well known in the literature. For each motor speed,
100 complete transit time measurements were acquired for each of the 12 trajectories.
Means and standard deviations were calculated from the measurements taken, as well as
the uncertainties associated with the measurement.
Evaluation of Flow Measurements
Table
4
presents the obtained flows (measured with the ultrasonic system) for each
motor rotation (as well as the respective pump rotations).
Sensors 2021,21, 7979 13 of 25
Table 4. Relation between flow and motor/pump angular speed.
Motor Rotation Pump Rotation Flow
0 RPM 0.00 RPM 0.00 m3/h
450 RPM 117.17 RPM 11.47 m3/h
500 RPM 196.85 RPM 13.19 m3/h
600 RPM 236.22 RPM 16.07 m3/h
700 RPM 275.59 RPM 19.29 m3/h
Ideally, the flow measurements obtained with the proposed sensors should be vali-
dated using a certified flow measurement system. As this was not possible, the solution
found was to compare the measured flows against the data sheet of the pump
[53]
, which
relates flow with pump rotation. From the data of Table
4
, a linear regression is performed,
resulting in the Equation (17).
Qproto ty pe =6.91275 ×10−2×RPM −2.43041 ×10−1, (17)
where
RPM
is the motor rotation (in RPM) and
Qproto ty pe
is the interpolated flow (in m
3
/h)
measured with the prototype.
The manufacturer curve of the pump can be obtained in
[53]
and results in Equation (
18
).
Qpum p =6.88505 ×10−2×RP M −2.32743 ×10−1, (18)
where Qpump is the reference flow (in m3/h) for a motor rotation RPM.
Based on Equations (
17
) and (
18
), for some values of motor rotation, the flows of
the proposed system can be compared against the theoretical flows at the pump. This
comparison is presented in Table 5.
Table 5. Comparison between the theoretical flows against the interpolated flows of the prototype.
Motor Rotation Qpum p Qprototype Deviation
400 RPM 27.41 m3/h 27.31 m3/h 0.37%
500 RPM 34.32 m3/h 34.19 m3/h 0.37%
600 RPM 41.23 m3/h 41.08 m3/h 0.38%
700 RPM 48.15 m3/h 47.96 m3/h 0.38%
Evaluation of Speed-of-Sound and Density Measurements
As the fluid used in the evaluation setup was water, and the fluid temperature at
the moment of the tests was 19.40
◦
C, and the value of speed of sound can be obtained
as 1480.36 m/s
[54]
and the value of density can be obtained as 998.34 kg/m
3[55]
. These
theoretical values can be used in order to validate the measurements of speed of sound
and density performed with the proposed prototype. These comparisons are presented in
Tables 6and 7.
Table 6.
Validation of measurements of speed of sound of water circulating in the evaluation setup at
19.40 ◦C, whose theoretical value is 1480.36 m/s [54].
Motor Rotation Speed of Sound Deviation
0 RPM 1479.01 m3/h 0.091%
450 RPM 1480.52 m3/h 0.011%
500 RPM 1480.31 m3/h 0.004%
600 RPM 1480.73 m3/h 0.025%
700 RPM 1481.10 m3/h 0.034%
Sensors 2021,21, 7979 14 of 25
Table 7.
Validation of measurements of density of water circulating in the evaluation setup at 19.40
◦
C,
whose theoretical value is 998.34 kg/m3[55].
Motor Rotation Density Deviation
0 RPM 1000.16 kg/m30.183%
450 RPM 998.13 kg/m30.021%
500 RPM 998.41 kg/m30.007%
600 RPM 997.84 kg/m30.050%
700 RPM 997.35 kg/m30.069%
As observed in Tables 6and 7, the deviations are below 0.2%.
2.4. Estimation of Heating Value
The heating values of any organic compound are associated with the bonding energies
between the atoms that form the chemical structure of the compound and, therefore,
the character of the bonds [
56
]. However, the possibility of calculation of the heating
value for petroleum-derived fuels with reasonable accuracy from elementary composition
alone took many researchers to establish empirical correlations from their commonly
measured characteristics [
57
,
58
]. These correlations are often expressed in the form of
linear combinations of the percentages by weight of the elements of the atoms of carbon
(C), hydrogen (H), and oxygen (O), and sometimes expanded to sulfur (S) and nitrogen
(N). The reason is that the main elements in the chemical composition of fossil fuels are, in
fact, limited to C, H, O, N, and S in their organic part [
58
]. Furthermore, the contaminants
present in the oils have a great influence on the efficiency of their combustion. Water is the
contaminant that has the greatest influence on HHV and LHV levels, where the greater the
amount of water, the lower the values of the two factors will be. Furthermore, sulfur and
ashes are two other important contaminants.
The standard ASTM D4868 [
13
] presents an empirical method to estimate the HHV
and the LHV based on the fuel density and the amounts of water, sulfur, and ashes. It is
described by Equations (19) and (20).
HHVASTM D4868 =51.916 −8.792 ·10−6·ρ2·h1−fwater +fsul f ur +fashesi
+9.420 ·fsul f ur , (19)
where
ρ
is the density of the fuel at 15
◦
C (in kg/m
3
), obtained from a laboratory test.
fwater
is the water content (in
V/V
),
fsul f ur
is the sulfur content (in
V/V
), and
fash
is the ash
content (in V/V).
LHVASTM D4868 =46.423 −8.792 ·10−6·ρ2+3.170 ·10−3·ρ·h1−fwater +fsul f u r +fashesi
+9.420 ·fsul f ur −2.449 ·f wa ter . (20)
The amount of sulfur can be obtained through laboratory analysis using technolo-
gies such as ultraviolet fluorescence, non-dispersive infrared, or X-ray fluorescence spec-
troscopy [
59
]. These technologies measure the sulfur content at the molecular level and are
very specific and difficult to use online. The same can be said for the ash content. Hence,
this work considers only the water content.
The water content is obtained online (with the sensor and method presented in Section
2.2). Furthermore, the density is obtained online (with the sensor and method presented in
Section 2.3). These two are applied in simplified versions of Equations (19) and (20), without
considering the sulfur and ashes contents, resulting in Equations (21) and (22).
HHV =51.916 −8.792 ·10−6·ρ2
0·(1−f wat er), (21)
Sensors 2021,21, 7979 15 of 25
LHV =46.423 −8.792 ·10−6·ρ2
0+3.170 ·10−3·ρ0·(1−f wat er)−2.449 ·f water , (22)
where
ρ0
is the density of the fuel compensated in relation to the amount of water using
Equation (23).
ρ0=ρu−ρw·fwater
1−fwater , (23)
where
ρu
is the density of the fuel, as obtained with the ultrasonic sensor of Section 2.3.
ρw
is the density of water.
3. Results and Discussion
The proposed system has been installed in a Brazilian TPP. Figure 8presents a pho-
tograph of the setup, installed in a section of the pipeline between the TPP fuel storage
and the point where the tank truck unloads the fuel oil. The system is composed of one
capacitive sensor and six pairs of ultrasonic sensors, as described in Section 2.
Figure 8. Proposed system installed at the fuel input of a Brazilian TPP.
Figure 9presents a photograph of the proposed capacitive sensor for determination of
water content, as described in Section 2.2.
Figure 9. Proposed capacitive sensor for determination of water content.
Figure 10 presents a photograph of the proposed ultrasonic sensor for determination
of density and flow, as described in Section 2.3.
Sensors 2021,21, 7979 16 of 25
Figure 10. Proposed ultrasonic sensor for determination of density and flow.
The following results were obtained at three distinct operations of loading, with three
different tank trucks. These tests were carried out at different periods due to the require-
ment of the plant itself for energy generation. During these periods, the measurement
system was turned on and followed the supply process carried out by TPP. General loading
information is presented in Table 8.
Table 8. Recorded loading operations at the TPP.
Loading Date Time Duration
1 9 June 2021 10:50–11:35 4502000
2 30 June 2021 10:52–11:36 4400000
3 30 June 2021 12:05–12:50 4502000
The prototype was configured to generate information in a 20-s update cycle at the time
of loading. In these 20 s, acquisitions of signals and data processing were performed. In
case of the ultrasonic system (which is responsible for measurement of flow measurement,
sound velocity and, consequently, density), 600 UP and DOWN transit time samples were
acquired for each acoustic trajectory, and from these measurements, the corresponding
calculations were performed. Therefore, each 20 s update presents data for an average of
600 samples of the ultrasonic system. In case of the capacitive system, 5000 capacitance
measurement samples were acquired per update cycle that are taken into account to
calculate the water content of the emulsion.
The uncertainties presented in the graphs that follows are defined according to the
guidelines given in references
[60]
. In the case of the ultrasonic system process, the
uncertainties are correlated to the measurement of transit times and take into account
their mean, standard deviation, and number of samples for each update cycle. The other
uncertainties (path length, angle, etc.) related to uncertainty propagation are static, as
they involve measurements based on calibrated equipment. The same principle is used
in the capacitive system, where the capacitance uncertainty is estimated by the mean,
standard deviation, and number of samples of measurements performed in each update
cycle. Uncertainty propagation for the other quantities, density, and LHV use the concept
of uncertainty propagation [61].
3.1. Flow
Figures
11
and
12
present the results of the flow measurements performed for the
loading operations of Table
8
. Figure
11
presents the general perspective of the three
loading operations. It can be noted that the flow measurements of the three operations are
very close over time.
Sensors 2021,21, 7979 17 of 25
Figure 11.
Measurement of flow using the proposed ultrasonic sensor of Section 2.3 for the three
loading operations of Table 8.
Figure
12
presents a zoom in the region of Figure
11
where the loading is actually
happening, in order to visualize more clearly the calculated uncertainties. It is possible to
observe that some samples present high levels of uncertainty. For the first loading (blue
color) it is observed that the uncertainty levels at the beginning of the loading (from the
start of the procedure until near the 15 min mark) had some samples with large levels
of uncertainty, while the remaining samples had lower levels. These variations in the
uncertainties are related to the dispersion of transit time differences (Equation (
12
)) with
high standard deviations and, consequently, higher levels of uncertainty in the flow.
Figure 12.
Zoomed view on the measurements of flow, in order to better visualize the uncertainties.
Table
9
presents the summary of the flow measurements of the three loading operations
of Table
8
. As noted, the average values of the means are close to each other and have
similar standard deviations from the mean. Concerning the uncertainties, it can be seen
that the maximum uncertainty observed between loading operations was 6.74%. Variations
in uncertainties are related to the standard deviation of the difference in transit time
(∆t=tup −tdown ) of the trajectories.
Sensors 2021,21, 7979 18 of 25
Table 9. Summary of the flow measurements of the three loading operations of Table 8.
Loading Average of the Means Standard Deviation of the Means Minimum Uncertainty Maximum Uncertainty
161.3 m3/h 0.3 m3/h 0.23% 6.74%
260.8 m3/h 0.3 m3/h 0.23% 5.93%
360.7 m3/h 0.3 m3/h 0.23% 1.31%
The average fluid temperatures during the loading operations were 17.75
◦
C, 19.58
◦
C,
and 21.55 ◦C, for loading operations 1, 2, and 3, respectively.
3.2. Capacitance and Water Content
Figure
13
presents the results of the flow measurements performed for the loading
operations of Table 8. Table 10 presents the summary of the capacitance measurements of
the three loading operations.
Figure 13.
Measurement of capacitance using the proposed capacitive system of Section 2.2 for the
three loading operations of Table 8.
Table 10. Summary of the capacitance measurements of the three loading operations of Table 8.
Loading Average of the Means Standard Deviation of the Means Minimum Uncertainty Maximum Uncertainty
1 290.5 pF 0.3 pF 0.032% 0.054%
2 297.4 pF 0.5 pF 0.028% 0.152%
3 295.9 pF 0.3 pF 0.024% 0.048%
Based on the measured capacitances, applying Equation (
1
), the water content for
the loading operations of Table
8
can be obtained. Figure
14
presents the estimated water
contents. Table
11
presents the summary of the water content measurements of the three
loading operations. It can be noted that the average water contents were 0.456%, 0.503%,
and 0.514%, for loading operations 1, 2, and 3, respectively.
Sensors 2021,21, 7979 19 of 25
Figure 14.
Measurement of water content using the proposed capacitive system of Section 2.2 for the
three loading operations of Table 8.
Table 11. Summary of the water content measurements of the three loading operations of Table 8.
Loading Average of the Means Standard Deviation of the Means Minimum Uncertainty Maximum Uncertainty
1 0.456% 0.002% 0.18% 0.30%
2 0.503% 0.003% 0.14% 0.77%
3 0.514% 0.003% 0.12% 0.25%
The results obtained using the capacitive sensor should have, ideally, been compared
against a laboratory test. However, the only laboratory results available had been obtained
almost three months before the installation of the prototype, but they are used here, in
order to give an idea of how close the proposed system compares to a laboratory analysis.
The complete laboratory results are presented in the Appendix
A
. The results obtained
with the prototype indicate around 0.5% of water content. The laboratory results indicated
0.05%. This a (somehow) large discrepancy—although both values are well below the limit
of 2% determined by the Brazilian National Agency of Petroleum, Natural Gas and Biofuels
(ANP) [36].
3.3. Speed of Sound and Density
Figure
15
presents the results of the online measurements of the speed of sound at the
emulsion, performed for the loading operations of Table
8
. Table
12
presents the summary
of the speed of sound measurements of the three loading operations.
Sensors 2021,21, 7979 20 of 25
Figure 15.
Measurement of speed of sound using the proposed ultrasonic system of Section 2.3 for
the three loading operations of Table 8.
Table 12. Summary of the speed of sound measurements of the three loading operations of Table 8.
Loading Average of the Means Standard deviation of the Means Minimum Uncertainty Maximum Uncertainty
11352.4 m/s 0.5 m/s 0.122% 0.122%
21367.6 m/s 1.2 m/s 0.121% 0.142%
31363.5 m/s 1.3 m/s 0.121% 0.122%
Based on the measured speeds of sound, applying Equation (
23
), the densities for the
loading operations of Table
8
can be obtained. Figure
16
presents the estimated densities.
Table
13
presents the summary of the density measurements of the three loading operations.
Figure 16.
Measurement of density using the proposed ultrasonic system of Section 2.3 for the three
loading operations of Table 8.
Table 13. Summary of the density measurements of the three loading operations of Table 8.
Loading Average of the Means Standard Deviation of the Means Minimum Uncertainty Maximum Uncertainty
1949.2 kg/m30.6 kg/m30.556% 0.556%
2928.2 kg/m31.6 kg/m30.556% 0.575%
3933.8 kg/m31.7 kg/m30.556% 0.566%
The results obtained with the prototype indicate average density values from 928.2 to
949.2 kg/m
3
. The laboratory results of the Appendix
A
indicated 927.7 kg/m
3
. This indi-
Sensors 2021,21, 7979 21 of 25
cates a maximum discrepancy of 2.3%, although the laboratory tests have been performed
three months before the installation of the prototype.
3.4. Heating Value
Using the measured values of the water content (data from Figure 14) and the density
(data from Figure 16), the lower heating value can be estimated online, as presented in
Figure 17.
Figure 17.
Measurement of Lower Heating Value using both the proposed capacitive system of
Section
2.2
and the proposed ultrasonic system of Section 2.3 for the three loading operations of Table
8.
Table
14
presents the summary of the LHV measurements of the three loading opera-
tions. It can be observed that the calculated uncertainties had values below 0.25%.
Table 14. Summary of the Lower Heating Values measurements of the three loading operations of Table 8.
Loading Average of the Means Standard Deviation of the Means Minimum Uncertainty Maximum Uncertainty
141.31 MJ/kg 0.01 MJ/kg 0.192% 0.222%
241.57 MJ/kg 0.02 MJ/kg 0.181% 0.210%
341.49 MJ/kg 0.02 MJ/kg 0.180% 0.211%
The results obtained with prototype indicate average LHV values from 41.31 to 41.57
MJ/kg. The laboratory results of the Appendix
A
indicate 41.59 MJ/kg. This indicates a
discrepancy below 1%, although the laboratory tests have been performed three months
before the installation of the prototype.
4. Conclusions
This paper presented a hybrid online monitoring system for the determination of fuel
oil quality. It is proposed the association of two different measurement techniques (i.e.,
capacitive and ultrasonic techniques) and the correlation of their information in order to
assess the quality parameters. The capacitive technique aims to measure the water content
in the fuel oil, while the ultrasonic technique directly measures the fuel’s density and flow.
The combination of information from both techniques serves to infer about other quality
indicators and carry out calibration processes.
The estimation of water content through the capacitive technique is based on the fact
that the dielectric properties of the fuel oil changes according to its water content. Hence, a
capacitive sensor (composed of multiple metallic disks) is proposed in order to measure
the capacitance of the emulsion (fuel oil with water contamination) that fills the spaces
between the disks.
Sensors 2021,21, 7979 22 of 25
The estimation of density and flow through the ultrasonic technique is based on the
propagation times of the ultrasonic pulses on the fuel oil. A system composed of an array
of sensors is proposed in order to estimate the velocity of the pulses, even with the fuel in
motion through the pipeline where the sensors are installed.
With the information of both the water content and the density, the heating value of
the fuel oil can be estimated. The heating value is an important parameter that shows the
quality of the fuel, which implies in the efficiency of the combustion cycle of the power
generators. Usually, combustion-based power plants have to perform laboratory tests in
order to determine the quality of their fuels. In these cases, samples are collected and sent
to a laboratory for the analysis. With the proposed system, the analysis can be performed
instantaneously, while the tank truck is loading the plant.
The proposed system, at the current stage of research, is an invasive procedure. As
a future work, it is being studied a non invasive procedure, for both the capacitive and
ultrasonic systems. Furthermore, as future work, it is planned to analyze the influence of
flow conditions on the measurements of capacitance and the effects of the temperature on
the density of the fuel oil. To date, due to intermittent loading operations at the TPP (few
operations per month due to low demand), some long-term measurements, such as the
evaluation of drift in the sensors, cannot be performed.
Author Contributions:
Conceptualization, M.M.C., L.E.B.-d.-S., D.d.A.A., C.E.T., and E.L.B.; data
curation, L.E.B.-d.-S, C.E.T., E.L.B., G.L.-T., and L.E.L.O.; formal analysis, M.M.C., D.d.A.A., C.E.T.,
R.F.R.J., and G.P.K.; investigation, M.M.C., D.d.A.A., C.E.T., L.E.B.-d.-S., R.F.R.J., and G.P.K.; method-
ology and software, M.M.C., D.d.A.A., C.E.T., G.P.K., L.E.L.O., and E.L.B.; project administration,
L.E.B.-d.-S., C.E.T., E.L.B., G.L.-T., L.E.L.O., and R.G.d.P.; validation, M.M.C., D.d.A.A., C.E.T., and
R.F.R.J.; supervision, L.E.B.-d.-S., C.E.T., E.L.B., G.L.-T., L.E.L.O., and R.G.d.P.; writing, M.M.C. and
W.C.S. All authors have read and agreed to the published version of the manuscript.
Funding: This research received no external funding.
Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.
Data Availability Statement: Not applicable.
Acknowledgments:
The authors would like to thank Norte Energia S.A. and the Brazilian Research
Agencies CNPq, CAPES, and ANEEL R&D for their support of this project.
Conflicts of Interest: The authors declare no conflict of interest.
Abbreviations
The following abbreviations are used in this manuscript:
FO Fuel Oil
GHC Gross Heat of Combustion
HFO Heavy Fuel Oil
HHV Higher Heating Value
LFO Light Fuel Oil
LHV Lower Heating Value
NHC Net Heat of Combustion
TPP Thermal Power Plant
Appendix A. Laboratory Tests Performed on a Sample of HFO
Table
A1
presents an excerpt of the results sheet of a laboratory analysis performed
on a sample of HFO of the same TPP where the prototype has been installed. The test,
however, has been performed some months before the installation of the system.
Sensors 2021,21, 7979 23 of 25
Table A1.
Results of laboratory tests performed on a sample of HFO (from the results sheet provided by the Brazilian
laboratory Intertek).
Methods Tests Results Units
ASTM D1298 density (20/4◦C) 927.7 kg/m3
ASTM D95 water by distillation 0.05 % of volume
ASTM D4868 lower heating value 41.59 kJ/kg
References
1.
Santos, A.Q.O.; da Silva, A.R.; Ledesma, J.J.G.; de Almeida, A.B.; Cavallari, M.R.; Junior, O.H.A. Electricity Market in Brazil: A
Critical Review on the Ongoing Reform. Energies 2021,14, 2873. doi:10.3390/en14102873.
2.
ANEEL. Sistemas de Geracao da ANEEL—SIGA. [S.1.] 2021. Available online: http://www.aneel.gov.br/siga (accessed on 17
September 2021).
3. Ehsan, M.; Rahman, M.M.; Saadi, H. Effect of fuel adulteration on engine crankcase dilution. J. Mech. Eng. 2010,41, 114–120.
4.
Boadu, K. Effects of adulteration on diesel oil with kerosene fuel in Ghana. J. Appl. Sci. Environ. Manag.
2019
,23, 1195–1200.
doi:10.4314/jasem.v23i7.1.
5. Gawande, A.P.; Kaware, J.P. Fuel adulteration consequences in India: A review. Sci. Rev. Chem. Commun. 2013,3, 161–171.
6.
Massawe, E.A.; Kilavo, H.; Sam, A. Fuel Adulteration in Tanzania and its Consequencies: An Overview. Res. J. Eng. Appl. Sci.
2013,4, 8.
7.
Mazivila, S.J.; Gontijo, L.C.; Santana, F.B.d.; Mitsutake, H.; Santos, D.Q.; Borges Neto, W. Fast Detection of Adulter-
ants/Contaminants in Biodiesel/Diesel Blend (B5) Employing Mid-Infrared Spectroscopy and PLS-DA. Energy Fuels
2015
,
29, 227–232. doi:10.1021/ef502122w.
8.
Vilas Boas, F.M.; Borges-da Silva, L.E.; Villa-Nova, H.F.; Bonaldi, E.L.; Oliveira, L.E.L.; Lambert-Torres, G.; Assuncao, F.d.O.;
Costa, C.I.d.A.; Campos, M.M.; Sant’Ana, W.C.; et al. Condition Monitoring of Internal Combustion Engines in Thermal Power
Plants Based on Control Charts and Adapted Nelson Rules. Energies 2021,14, 4924. doi:10.3390/en14164924.
9.
Assuncao, F.d.O.; Borges-da Silva, L.E.; Villa-Nova, H.F.; Bonaldi, E.L.; Oliveira, L.E.L.; Lambert-Torres, G.; Teixeira, C.E.;
Sant’Ana, W.C.; Lacerda, J.; Silva Junior, J.L.M.d.; et al. Reduced Scale Laboratory for Training and Research in Condition-Based
Maintenance Strategies for Combustion Engine Power Plants and a Novel Method for Monitoring of Inlet and Exhaust Valves.
Energies 2021,14, 6298. doi:10.3390/en14196298.
10.
Mendonca, P.; Bonaldi, E.; de Oliveira, L.; Lambert-Torres, G.; da Silva, J.B.; da Silva, L.B.; Salomon, C.; Santana, W.; Shinohara, A.
Detection and modelling of incipient failures in internal combustion engine driven generators using Electrical Signature Analysis.
Electr. Power Syst. Res. 2017,149, 30–45. doi:10.1016/j.epsr.2017.04.007.
11.
American Society for Testing and Materials. ASTM D95-13 (2018)—Standard Test Method for Water in Petroleum Products and
Bituminous Materials by Distillation; ASTM International, West Conshohocken, USA; 2018.
12.
American Society for Testing and Materials. ASTM D1298-12b (2017)—8 ASTM. ASTM D1298-12b(2017)—Standard Test Method for
Density, Relative Density, or API Gravity of Crude Petroleum and Liquid Petroleum Products by Hydrometer Method; ASTM International,
West Conshohocken, USA; 2017.
13.
American Society for Testing and Materials. ASTM D4868-00 (2010)—Standard Test Method for Estimation of Net and Gross Heat of
Combustion of Burner and Diesel Fuels; ASTM International, West Conshohocken, USA; 2010.
14.
Han, Y.; Zhao, A.; Zhang, H.; Ren, Y.; Liu, W.; Jin, N. Differential pressure method for measuring water holdup of oil–water two-
phase flow with low velocity and high water-cut. Exp. Therm. Fluid Sci.
2016
,72, 197–209. doi:10.1016/j.expthermflusci.2015.11.008.
15.
Al-Saiyed, M.R.; Warsi, S.; Phillips, J.; Gilani, S.; Zareef, M.; Lievois, J. Infrared Water Cut Meter for Cost Effective Individual
Monitoring of Complex Wells. In Proceedings of the 26th International North Sea Flow Measurement Workshop, St. Andrews,
Scotland, 21–24 October 2008.
16.
Scheers, A. An Oil/Water/Gas Composition Meter Based on Multiple Energy Gamma Ray Absorption (MEGRA) Measurement; Interna-
tional Atomic Energy Agency (IAEA), Vienna, Austria, 2000; pp. 105–122.
17.
Johansen, G.A.; Tjugum, S.A. Fluid composition analysis by multiple gamma-ray beam and modality measurements. In Pro-
ceedings of the 25th International North Sea Flow Measurement Workshop; NFSOGM - Norwegian Society for Oil and Gas
Measurement, Oslo, Norway; 2007.
18.
Rigaku Applied Technologies, Inc. NEX OL Process Elemental Analyzer for Liquid Streams—On-line, Real-Time Analysis
by Energy Dispersive X-ray Fluorescence (EDXRF)—Online Catalog. 2021. Available online: https://rigakuedxrf.com/
nex-ol.php?gclid=Cj0KCQjw7pKFBhDUARIsAFUoMDZ9rZZZnKoF9lLcjd8FPYUlaNiSO7OXMWHKJ57xc1WCqM1RNT3W4
RgaAopxEALw_wcB (accessed on 17 September 2021).
19.
Gregory, A.; Clarke, R. A review of RF and microwave techniques for dielectric measurements on polar liquids. IEEE Trans.
Dielectr. Electr. Insul. 2006,13, 727–743. doi:10.1109/TDEI.2006.1667730.
20.
Makeev, Y.V.; Lifanov, A.; Sovlukov, A.S. Microwave measurement of water content in flowing crude oil. Autom. Remote Control
2013,74, 157–169.
Sensors 2021,21, 7979 24 of 25
21.
Teixeira, C.E.; da Silva, L.E.B.; Veloso, G.F.; Lambert-Torres, G.; Campos, M.M.; Noronha, I.; Bonaldi, E.L.; de Oliveira, L.E.L. An
ultrasound-based water-cut meter for heavy fuel oil. Measurement 2019,148, 106907. doi:10.1016/j.measurement.2019.106907.
22.
Hammer, E.; Tollefsen, J.; Olsvik, K. Capacitance transducers for non-intrusive measurement of water in crude oil. Flow Meas.
Instrum. 1989,1, 51–58. doi:10.1016/0955-5986(89)90010-1.
23.
Schuller, R.; Gundersen, T.; Halleraker, M.; Engebretsen, B. Measurement of water concentration in oil/water dispersions with a
circular single-electrode capacitance probe. IEEE Trans. Instrum. Meas. 2004,53, 1378–1383. doi:10.1109/TIM.2004.830786.
24.
Libert, N.; Morales, R.E.; da Silva, M.J. Capacitive measuring system for two-phase flow monitoring. Part 1: Hardware design
and evaluation. Flow Meas. Instrum. 2016,47, 90–99. doi:10.1016/j.flowmeasinst.2015.12.009.
25.
Libert, N.; Morales, R.E.; da Silva, M.J. Capacitive measuring system for two-phase flow monitoring. Part 2: Simulation-based
calibration. Flow Meas. Instrum. 2016,50, 102–111. doi:10.1016/j.flowmeasinst.2016.05.011.
26.
Hansen, L.S.; Pedersen, S.; Durdevic, P. Multi-Phase Flow Metering in Offshore Oil and Gas Transportation Pipelines: Trends and
Perspectives. Sensors 2019,19, 2184. doi:10.3390/s19092184.
27.
Wang, F.; Marashdeh, Q.; Fan, L.S.; Warsito, W. Electrical Capacitance Volume Tomography: Design and Applications. Sensors
2010,10, 1890–1917. doi:10.3390/s100301890.
28.
Mosorov, V.; Zych, M.; Hanus, R.; Sankowski, D.; Saoud, A. Improvement of Flow Velocity Measurement Algorithms Based on
Correlation Function and Twin Plane Electrical Capacitance Tomography. Sensors 2020,20, 306. doi:10.3390/s20010306.
29.
de Almeida Arantes, D.; Borges da Silva, L.E.; Teixeira, C.E.; Campos, M.M.; Lambert-Torres, G.; Bonaldi, E.L.; de Lacerda de
Oliveira, L.E.; da Costa, G.A. Relative Permittivity Meter Using a Capacitive Sensor and an Oscillating Current Source. In
Proceedings of the IECON 2019—45th Annual Conference of the IEEE Industrial Electronics Society, Lisbon, Portugal, 14–17
October 2019; Volume 1, pp. 806–811. doi:10.1109/IECON.2019.8927762.
30.
Bjørndal, E. Acoustic Measurement of Liquid Density with Applications for Mass Measurement of Oil. Ph.D. Thesis, University
of Bergen, Bergen, Norway, 2007.
31.
Bjorndal, E.; Froysa, K.E. Acoustic methods for obtaining the pressure reflection coefficient from a buffer rod based measurement
cell. IEEE Trans. Ultrason. Ferroelectr. Freq. Control 2008,55, 1781–1793. doi:10.1109/TUFFC.2008.862.
32.
Hoche, S.; Hussein, M.; Becker, T. Ultrasound-based density determination via buffer rod techniques: A review. J. Sens. Sens.
Syst. 2013,2, 103–125. doi:10.5194/jsss-2-103-2013.
33.
Kim, J.O.; Bau, H.H. Instrument for simultaneous measurement of density and viscosity. Rev. Sci. Instrum.
1989
,60, 1111–1115.
doi:10.1063/1.1140325.
34.
Swoboda, C.A.; Cannon, P.H.; Fredrickson, D.R.; Gabelnick, S.D.; Hornstra, F.; Phan, K.A.; Yao, N.P. Development of an ultrasonic
technique to measure specfic gravity in lead-acid battery electrolyte. IEEE Trans. Sonics Ultrason. (USA)
1983
,Volume 30, pp.69–77.
doi:10.1109/T-SU.1983.31389.
35. Davis, L.A.; Gordon, R.B. Compression of mercury at high pressure. J. Chem. Phys. 1967,46, 2650–2660.
36.
Agencia Nacional do Petroleo, Gas Natural e Biocombustiveis. Resolucao ANP n
º
3 DE 27/01/2016. 2016. Available online:
https://www.gov.br/anp/pt-br/assuntos/producao-de-derivados-de-petroleo-e-processamento-de-gas-natural/producao-
de-derivados-de-petroleo-e-processamento-de-gas-natural/oleo-combustivel (accessed on 28 September 2021).
37. Speight, J.G. The Chemistry and Technology of Petroleum; CRC Press: Boca Raton, FL, USA, 2014.
38.
Al-Ghouti, M.; Al-Degs, Y.; Mustafa, F. Determination of hydrogen content, gross heat of combustion, and net heat of combustion
of diesel fuel using FTIR spectroscopy and multivariate calibration. Fuel 2010,89, 193–201.
39.
Demirbas, A.; Ak, N.; Aslan, A.; Sen, N. Calculation of higher heating values of hydrocarbon compounds and fatty acids. Pet. Sci.
Technol. 2018,36, 712–717. doi:10.1080/10916466.2018.1443126.
40. Malmberg, C.; Maryott, A. Dielectric constant of water from 0 to 100 C. J. Res. Natl. Bur. Stand. 1956,56, 1–8.
41. Warren, W. BS and W Measurement Principles and Practices. J. Pet. Technol. 1962,14, 1207–1212.
42.
Kang, S.; You, Y.; Feng, M. Study of high-pressure physical properties of marine heavy oil. In Key Engineering Materials; Trans
Tech Publ.: Bach, Switzerland; 2016; Volume 693, pp. 411–418.
43.
Akahama, Y.; Kawaguchi, S.; Hirao, N.; Ohishi, Y. Elastic softening of bulk modulus of monoclinic HfO2 under high pressure.
Appl. Phys. Lett. 2020,117, 182903. doi:10.1063/5.0029148.
44.
Boehman, A.L.; Morris, D.; Szybist, J.; Esen, E. The impact of the bulk modulus of diesel fuels on fuel injection timing. Energy
Fuels 2004,18, 1877–1882. doi:10.1021/ef049880j.
45. McClements, D.; Fairley, P. Ultrasonic pulse echo reflectometer. Ultrasonics 1991,29, 58–62. doi:10.1016/0041-624X(91)90174-7.
46. Hale, J. Ultrasonic density measurement for process control. Ultrasonics 1988,26, 356–357. doi:10.1016/0041-624X(88)90036-4.
47.
Henning, B.; Daur, P.C.; Prange, S.; Dierks, K.; Hauptmann, P. In-line concentration measurement in complex liquids using
ultrasonic sensors. Ultrasonics 2000,38, 799–803. doi:10.1016/S0041-624X(99)00190-0.
48.
Campos, M.M. Medidor de Vazao de Multi-Trajetorias (in Portuguese). Master’s Thesis, Universidade Federal de Itajuba, Itajuba,
Brazil, 2017.
49.
Gruber, P.; Wermelinger, F.; Hug, S. Experience with the ATT method for discharge measurement with adaptive weighting of the
path velocities. In Procedings of the HYDRO 2013 Conference, Innsbruck, Austria, 7 – 9 October 2013.
50.
Gruber, P.; Marushchenko, S. Comparative study of 4(8)-path and 5(10)-path configurations for ATT flow measurements in
circular conduits. In Procedings of the IGHEM 2014 Conference, Itajuba, Brazil, 16–19 September 2014.
Sensors 2021,21, 7979 25 of 25
51.
Tresch, T.; Lüscher, B.; Staubli, T.; Gruber, P. Presentation of optimized integration methods and weighting corrections for the
acoustic discharge measurement. In Procedings of the IGHEM 2008 Conference, Milano, Italy, 3–6 September 2008.
52.
Hug, S.; Staubli, T.; Gruber, P. Comparison of measured path velocities with numerical simulations for heavily disturbed velocity
distributions. In Procedings of the IGHEM 2012 Conference, Trondheim, Norway, 27–30, June 2012.
53.
RZR Bombas Positivas. Bombas de Engrenagens Internas Modelo RZR-30000. 2014. Available online: https://rzrbombas.com.
br/wp-content/themes/rzr/catalogos/RZR-30.000-SVA.pdf (accessed on 31 October 2021).
54.
The Engineering ToolBox. Water—Speed of Sound. 2021. Available online: https://www.engineeringtoolbox.com/sound-speed-
water-d_598.html (accessed on 31 October 2021).
55.
Jones, Frank E and Harris, Georgia L. ITS-90 density of water formulation for volumetric standards calibration. J. Res. Natl. Inst.
Stand. Technol. 1992,97, 335. doi:10.6028/jres.097.013.
56.
Buck, W.; Rudtsch, S., Thermal Properties. In Springer Handbook of Materials Measurement Methods; Czichos, H., Saito, T., Smith, L.,
Eds.; Springer Berlin Heidelberg: Berlin/Heidelberg, Germany, 2006; pp. 399–429. doi:10.1007/978-3-540-30300-8_8.
57.
Marlair, G.; Cwiklinski, C.; Tewarson, A. An analysis of some practical methods for estimating heats of combustion in fire safety
studies. In Proceedings of the Interflam 99 Conference, Edinburgh, United Kingdom, 29 June – 1 July 1999.
58.
Ponomarev, D.A.; Mettee, H.D.; Miller, J. Empirically estimated heats of combustion of oxygenated hydrocarbon bio-type oils.
BioResources 2015,10, 3638–3656.
59.
Corilo, Y.E.; Rowland, S.M.; Rodgers, R.P. Calculation of the total sulfur content in crude oils by positive-ion atmospheric
pressure photoionization Fourier transform ion cyclotron resonance mass spectrometry. Energy Fuels
2016
,30, 3962–3966.
doi:10.1021/acs.energyfuels.6b00497.
60.
BIPM, IEC, IFCC, ILAC, ISO, IUPAC, IUPAP and OIML. Evaluation of Measurement Data—Guide to the Expression of Uncertainty
in Measurement. 2008. Available online: https://www.bipm.org/utils/common/documents/jcgm/JCGM_100_2008_E.pdf
(accessed on 30 October 2021).
61.
BRITISH STANDARD. BS ISO 5168:2005—Measurement of Fluid Flow—Procedures for the Evaluation of Uncertainties; International
Organization for Standardization (ISO), Geneva, Switzerland; 2005.
