sensors

Article
Research on a Non-Contact Multi-Electrode Voltage Sensor and
Signal Processing Algorithm
Wenbin Zhang 1, *, Yonglong Yang 1 , Jingjing Zhao 2 , Rujin Huang 2 , Kang Cheng 2 and Mingxing He 2

1 College of Mechanical and Electrical Engineering, Kunming University of Science and Technology,
Kunming 650504, China
2 College of Science, Kunming University of Science and Technology, Kunming 650504, China
* Correspondence: [email protected]; Tel.: +86-1530-885-9327

Abstract: Traditional contact voltage measurement requires a direct electrical connection to the
system, which is not easy to install and maintain. The voltage measurement based on the electric
field coupling plate capacitance structure does not need to be in contact with the measured object or
the ground, which can avoid the above problems. However, most of the existing flat-plate structure
voltage measurement sensors are not only expensive to manufacture, but also bulky, and when the
relative position between the wire under test and the sensor changes, it will bring great measurement
errors, making it difficult to meet actual needs. Aiming to address the above problems, this paper
proposes a multi-electrode array structure non-contact voltage sensor and signal processing algorithm.
The sensor is manufactured by the PCB process, which effectively reduces the manufacturing cost
and process difficulty. The experimental and simulation results show that, when the relative position
of the wire and the sensor is offset by 10 mm in the 45◦ direction, the relative error of the traditional
single-electrode voltage sensor is 17.62%, while the relative error of the multi-electrode voltage sensor
designed in this paper is only 0.38%. In addition, the ratio error of the sensor under the condition of
power frequency of 50 Hz is less than ±1% and the phase difference is less than 4◦ . The experimental
Citation: Zhang, W.; Yang, Y.; Zhao,
results show that the sensor has good accuracy and linearity.
J.; Huang, R.; Cheng, K.; He, M.
Research on a Non-Contact
Keywords: non-contact voltage measurement; voltage sensor; electrode array; PCB process; Kalman
Multi-Electrode Voltage Sensor and
Signal Processing Algorithm. Sensors
filtering
2022, 22, 8573. https://doi.org/
10.3390/s22218573

Academic Editors: Xue-Bo Jin and

1. Introduction
Yuan Gao
Transmission lines are one of the most critical forms of transmission equipment in
Received: 18 September 2022 the power system. The accurate measurement of transmission line voltage is of great
Accepted: 30 October 2022 significance to the stable operation and power security of the power system. At present,
Published: 7 November 2022 the mainstream voltage measurement methods mainly include traditional contact voltage
Publisher’s Note: MDPI stays neutral measurement and new non-contact voltage measurement [1].
with regard to jurisdictional claims in At present, the traditional electric energy information acquisition mainly adopts the
published maps and institutional affil- contact voltage measurement technology, which requires the metal part of the cable to
iations. be in contact with the measurement probe [2,3]. The probe is connected to the secondary
equipment for voltage and current measurement. Contact voltage measurement requires
that the copper core and the measuring probe are in direct contact, which not only damages
the cable insulation layer, but also must be installed without a power supply, affecting
Copyright: © 2022 by the authors. the power supply quality and making installation difficult. In the process of contact
Licensee MDPI, Basel, Switzerland. measurement, the insulation layer of cable is damaged and the operation safety of the
This article is an open access article
power cable and equipment faces severe challenges.
distributed under the terms and
The non-contact voltage measurement (NCVM) technique can measure the electric
conditions of the Creative Commons
potential without any electrical contact [4], which is a smart and safe method of voltage
Attribution (CC BY) license (https://
measurement [2]. Non-contact voltage measurement technology is of great significance
creativecommons.org/licenses/by/
in the internet of things, power electronics, and high voltage engineering [5]. At present,
4.0/).

Sensors 2022, 22, 8573. https://doi.org/10.3390/s22218573 https://www.mdpi.com/journal/sensors

Sensors 2022, 22, 8573 2 of 13

there are two main non-contact voltage sensing methods: photoelectric sensing and electric
field coupling sensing [5–12]. Photoelectric sensing technology is based on the optical
principle of voltage measurement [13–16]. This method has good measurement accuracy,
but it is difficult to achieve large-scale promotion of engineering applications because of
the complex structure of optical devices, high cost, complex daily maintenance and repair,
and the fact that it is easily affected by ambient light. Non-contact voltage measurement
based on electric field coupling is more convenient for industrial application [12,17–23]. Its
basic principle is that, by applying a certain voltage or electric field, electromotive force
will be generated on the induction electrode, and a resistance or capacitance connection
can be formed between the electrodes. Based on this principle, different studies have been
carried out on non-contact voltage sensors. The authors of [24] designed and implemented
a voltage monitoring technology for overhead transmission lines based on contactless
capacitive coupling and supplemented by magnetic field sensing. In the work of [25], a
voltage acquisition system with a open close coaxial structure is introduced, which allows
the measurement of power line voltage waveform without electrical contact. However,
the structure is large in size, complicated in process, inconvenient for installation and
rotation, and high in cost. The work of [26] analyzes the capacitive non-contact AC voltage
measurement technology and the feasibility of measuring arbitrary waveform signals.
Several errors related to this technology are analyzed to show the influence of different
design parameters on the final accuracy. In order to optimize the design, different sensor
structures are studied. In the work of [27], a three-electrode sensor is designed and the
electrode is placed in a ring to improve the electric field distribution, insulation performance,
and sensor sensitivity [28]. In addition, the exact spatial location of overhead transmission
lines is usually unknown and dynamic in practice. The authors of [28] designed a non-
contact household voltage measurement device based on the capacitive coupling principle,
but this device has the problem of inconvenient installation. Shenil P. S. of Madras Institute
of Technology in India and others proposed an instrument scheme suitable for non-contact
measurement of AC voltage of insulated conductors. The AC voltage of the insulated
conductor can be measured using a suitable probe, which forms a capacitance network
between the conductor and the ground, and indirectly calculates the wire voltage by
online measuring the voltage across the capacitor [29]. In the laboratory, the prototype of
the measurement system was tested at different frequencies and voltages, the data were
processed by fast Fourier transform (FFT), and the measurement error was less than 0.75%.
According to Helmholtz theory, the induced electric field of the power line is spatially
distributed and this distribution may affect the integration of the electric field, thus gener-
ating corresponding probe charges. The sensing accuracy largely depends on the relative
position of the power line and the probe.
In a multi-sensor voltage measurement system, the measurement results mainly
depend on the processing of the measurement signals of all sensors. The requirement
for the signal processing method is to calculate the voltage value to be measured from
the measurement result containing a certain measurement error. For the measurement of
AC voltage, according to the principle of electric field induction, it can be known that the
average value of the sensor output in the sensor ring around the wire is proportional to the
wire voltage. Therefore, the voltage value to be measured can be obtained by averaging
the output signals of the voltage sensor loop. In addition, according to the characteristics
of the spatial electric field distribution around the wire, the output signal of the voltage
sensor ring can be spatially Fourier transformed, and the nonlinear equation can be solved
according to the analysis result, so as to obtain the value of the voltage to be measured.
Kalman filter is a method to obtain the best filter value through signal estimation and
correction based on the known system model and noise characteristics [30].
In view of the above analysis, this paper proposes a non-contact voltage sensor struc-
ture based on a multi-electrode array and designs a new electrode array structure composed
of six uniformly distributed electrodes. The direct contact of the circuit solves the intrusive
and safety problems of the touch sensor to a large extent. At the same time, it effectively
Sensors 2022, 22, 8573 3 of 13

solves the problem that the accuracy of single sensor non-contact sensor measurement
is easily affected by the wire position. The sensor is manufactured using a PCB process,
ensuring that the sensor is cheap, accurate, lightweight, and compact. At the same time, a
signal processing algorithm for multi-sensor AC voltage measurement based on Kalman
filter is also proposed in this paper. According to the electric field distribution around
the wire, the state equation of the system is established and the recursive method of the
Kalman steady-state filter is given. The Kalman filter algorithm is analyzed from the
perspective of algorithm complexity, measurement accuracy, and flexibility, as well as
algorithm application conditions and requirements.
The multi-electrode array non-contact voltage sensor designed in this paper has the
advantages of convenient use, isolation from the measurement circuit, low insulation
design requirements, and convenient installation. It can be used for voltage measurement
in low-voltage distribution networks and user terminals. The signal processing algorithm of
multi-sensor AC voltage measurement based on the Kalman filter is beneficial to reduce the
influence of interference voltage on the measurement and it is not sensitive to the influence
of random errors caused by the sensor itself, maintaining high measurement accuracy.
This paper first introduces the research status of non-contact voltage measurement
technology and, according to the problem that the measurement accuracy of the existing
sensor is easily affected by the relative position of the wire and the sensor, a non-contact
voltage sensor structure and data based on a multi-electrode array is proposed. Processing
algorithm; the second part mainly introduces the principle of voltage measurement and the
parameter design of the sensor; the third part mainly introduces the experimental platform,
data processing algorithm, and test results of the sensor test; then, the application scenarios
and shortcomings of the sensor are discussed.

2. Design of the Voltage Sensor

2.1. Principle of the Voltage Sensor
For a long straight wire with an alternating current, an electric field that changes
with the internal current of the wire will be generated around the electrified wire and a
changing potential will be generated on the conductor around the wire. In addition to the
size, shape, position, and other factors of the conductor, the potential on the conductor
has a direct relationship with the potential of the electrified wire that generates the electric
field. Therefore, using this principle, while fixing other influencing factors, the internal
potential of the electrified wire can be measured in a non-contact manner. The traditional
voltage contact measurement method is generally to connect the measurement loop into the
circuit to be measured and directly collect signals. Different from the traditional method,
the principle of the non-contact voltage sensor is to select a metal plate to be coupled with
the wire to be measured, form a coupling capacitance at the electrode plate, and collect the
induced voltage signal on the coupling electrode plate.
The measurement principle of the non-contact transmission line voltage sensor pro-
posed in this paper is shown in Figure 1. The sensor uses the stray capacitance Cl1 between
the overhead transmission line and the sensor induction plate as the high-voltage arm
capacitance. Figure 1 shows the physical process and equivalent circuit of the induced
voltage VI N between the internal induced electrode group and the transmission line voltage.
It is assumed that the resistor R and the capacitor C are connected between VI N and VR ,
which represents the stray capacitance between the power line and the internal probe
group, while Cl2 represents the stray capacitance between all conductors connected to VR
and the ground. The current I flowing through the measuring device is shown in Figure 1.
Therefore, the current I in the circuit can be expressed as follows:

VS
I= (1)
1 1 R
+ +
jωCl1 jωCl2 1 + jωCR
Sensors 2022, 22, 8573 4 of 13

If the expected stray capacitances Cl1 and Cl2 are at a maximum, and if the value of
capacitance C is much larger than 10 n f , then the impedance of RC is much smaller than
R
the impedance of stray capacitance, so can be ignored, then
1 + jωCR

jVS
Sensors 2022, 22, x FOR PEER REVIEW I= 4 of 15 (2)
1 1
+
wCl1 wCl2

Therefore,
Therefore,the
thecurrent
voltageI inisthe
as circuit
follows:can be expressed as follows:
𝑉
𝐼= R jωVS R
(VI N − VR ) = I 1 + 1 =+ 𝑅  (1) (3)
1 + jωCR
𝑗𝜔𝐶 𝑗𝜔𝐶 +1 𝑗𝜔𝐶𝑅
1ωC + ωC1 (1 + jωCR)
l1 l2
If the expected stray capacitances 𝐶 and 𝐶 are at a maximum, and if the value of
capacitance − VRlarger
Voltage𝐶VisI Nmuch is anthan 10 𝑛𝑓, then
attenuated the impedance
version of 𝑅𝐶distortion
of VS without is much smaller thanshift. The
or phase
the impedance of stray capacitance, so
attenuation factor γ is as follows: can be ignored, then

𝑗𝑉Cl1 Cl2
𝐼γ
== (4)
1C (Cl1 1+ Cl2 ) (2)
+
𝑤𝐶 𝑤𝐶
Its value depends
Therefore, onisstray
the voltage capacitance.
as follows:
Only when the stray capacitances Cl1 and Cl2 are known or can be evaluated, can the
𝑅 𝑗𝜔𝑉 𝑅
amplitude of VS be(𝑉accurately
− 𝑉 ) = 𝐼 predicted=from VI N − VR . However, for power factor and
1 1 (3)
1 + 𝑗𝜔𝐶𝑅 + know(1the+ 𝑗𝜔𝐶𝑅)
power quality measurement, it is not necessary𝜔𝐶 to 𝜔𝐶 absolute value of the voltage,
but it is important to know the phase and amplitude of the voltage.

Transmission line

Cl1

VIN

VS R C

VR
I
Cl 2

Ground

Figure 1. measurement principle.

Figure 1. Measurement principle.
VoltageParameter
2.2. Sensor is an attenuated version of 𝑉 without distortion or phase shift.
𝑉 − 𝑉 Design
The attenuation factor 𝛾 is as follows:
In order to obtain reasonable structural parameters, it is necessary to verify the influ-
𝐶 𝐶
ence of wire position change on the 𝛾= measurement results when the number of(4)electrode
𝐶(𝐶 + 𝐶models)
plates is different. By establishing simulation to change the wire position under
Its value
different depends
number on stray capacitance.
of electrode plates, the simulation conditions are set with the number
Only when
of electrode the stray
plates of 1, capacitances 𝐶 and 𝐶 are
2, 4, and 6, respectively, known
and or can
the wire is be evaluated,
moved by 2 can
mm, 4 mm,
6the amplitude
mm, 8 mm, of and bemm
𝑉 10 accurately
in the predicted
direction from
of x, 𝑉y, and . However,
− 𝑉 45 ◦ with the forpositive
power factor
half axis of x.
and power quality
Considering measurement,
that the it is not voltage
actual measured necessaryvalues
to know the absolute
of the sensor arevalue
220ofVthe
and 380 V,
voltage, but it is important to know the phase and amplitude of the voltage.
the reference voltage of the wire is uniformly set to 220 V in this simulation. In order to
facilitate the installation and use of the sensor, the overall size of the sensor is kept as small
as possible during the size design. Therefore, the initial distance between the electrode
plate and the wire is 30 mm, which is roughly the same as the actual size.
 Sensors 2022, 22, 8573 5 of 13

When only one plate is simulated, the initial distance between the plate and the wire
is 40 mm and the floating potential sensed by the plate is 85.00 V. Then, the wire is shifted
by 10 mm in the positive direction of the y-axis. At this time, the starting distance between
the plate and the wire is 50 mm and the floating potential sensed by the plate is 70.22 V,
with an error of 17.39%. The error is 17.62% when the wire is offset by 10 mm in the x
and y directions. Under the same conditions, the number of plates is changed to six, the
distance between each plate and the conductor is 50 mm, the average value of the initial
floating potential sensed by the plate is 67.77 V, and the average value of the initial floating
potential sensed by the plate is 67.51 V when the conductor is shifted by 10 mm in the x
and y directions. The error is 0.38%. Through simulation analysis, it can be seen that the
structure of the six-electrode array can effectively reduce the error caused by the change in
wire position compared with the sensor with one electrode plate.
Figure 2 shows a schematic diagram of the structure of a single electrode of the sensor.
Layers 1, 3, 5, and 7 represent electrode sensing layers with a thickness of 0.035 mm. Layers
2 and 4 represent FR4 materials with a thickness of 0.2 mm. Layer 6 represents FR4 with a
thickness of 1.2 mm. The sensor unit is composed of two capacitors and such a structure
is conducive to reducing external interference and improving sensor performance. The
Sensors 2022, 22, x FOR PEER REVIEWthird and fourth layers are ground layers, which can shield external signal interference. 6 of 15
In addition, the ground layer needs to be connected to the reference ground of the signal
under test to provide an accurate reference potential.

Unit: mm
top floor
0.035 (sensing layer)
The second floor
0.2
(media layer)
Seventh floor Fifth floor
Third floor
Sensor probe 0.017
(ground plane) (sensing layer) (ground plane)

Fourth floor
1.065
(isolation layer)
Fifth floor
0.017 (ground plane)

Sixth floor
0.2
(media layer)
Seventh floor Bottom view
0.035
(sensing layer)

Figure
Figure 2. Structuraldiagram
2. Structural diagramof of
thethe sensing
sensing electrode
electrode unit. unit.

According to the electric field coupled contactless voltage theory and simulation
results, the overall structure of the sensor is designed. The non-contact voltage sensor with
a multi electrode array designed in this paper is shown in Figure 3. The sensor includes two
groups of capacitive voltage sensors installed on an electrically insulating support member.
The first group of sensors is positioned along the inner closed path and the second group
of sensors is positioned along the outer closed path, surrounding the inner closed path and
connected in parallel. The wire to be measured can be placed on the central sensor of the
sensor, so that the sensor surrounds the axis of the wire to be measured, each sensor has a
signal electrode connected to the corresponding signal conductor, which can measure the
voltage in the measured conductor and the voltage waveform on the reference conductor.
The size of the sensor coupling capacitance Cm is related to the sensing distance of the
line under test and the area of the line under test. Among them, the sensing distance is
determined by the manufacturing process of the PCB board. In order to shield the vertical
interference of the induction head and attenuate the interference of the horizontal direction,
the area
Figure 3. of the top ground
Schematic diagramlayer in three-dimensional
of the Figure 2 should bestructure
larger than thatsensor.
of the of the sensor head.
Figure 4 shows the relationship between the shielding degree and the area of the shielding
layer. This paper defines the interference degree as the percentage of the interference
The size of the sensor coupling capacitance 𝐶 is related to the sensing distance of
the line under test and the area of the line under test. Among them, the sensing distance
is determined by the manufacturing process of the PCB board. In order to shield the ver-
tical interference of the induction head and attenuate the interference of the horizontal
 Seventh floor Fifth floor
Third floor
Sensor probe 0.017
(ground plane) (sensing layer) (ground plane)

Fourth floor
1.065
(isolation layer)
Sensors 2022, 22, 8573 6 of 13
Fifth floor
0.017 (ground plane)

Sixth floor
0.2 coupling capacitance Cj , where Cj is the coupling capacitance between
capacitance Cj to the
(media layer)
the side interference line and the sensor head. When the interference is strengthened on the
side of the electric Seventh
field coupling sensor, the interference degreefloor Bottom view
of the surrounding lines to
0.035
(sensing layer)
the sensor decreases with the increase in the area of the shielding layer. In order to ensure
Figure 2. Structural diagram of the sensing electrode unit.
the good shielding effect and small size of the sensor at the same time, this paper selects
∆x = 3 mm.

rs 2022, 22, x FOR PEER REVIEW 7 of

Figure 3. Schematic diagram of the three-dimensional structure of the sensor.

Figure 3. Schematic diagram of the three-dimensional structure of the sensor.
7

The
6 size of the sensorcoupling
x capacitance
side interference line 𝐶 is related to the sensing distan
the line under test and the area of the line under test. Among them, the sensing dis
is determined
5
by the manufacturing process of the PCB board. In order to shield th
Cj/Cm×100% (%)

tical interference
4 of the induction head and attenuate the interference of the horiz
direction, the area of the top ground shield
Intermediate layer in Figure 2 should be larger than that
3
sensor head. Figure 4 shows the relationship between the shielding degree and the
of the2 shielding layer. This paper defines the interference degree as the percentage
interference capacitance 𝐶 to the coupling capacitance 𝐶 , where 𝐶 is the couplin
1
pacitance between the Interference
side interference
level is less line
than 1%and the sensor head. When the interfere
strengthened
0 on the side of the electric field coupling sensor, the interference deg
the surrounding
-1
lines to the sensor decreases with the increase in the area of the shie
layer. In order to ensure
-1 0 1 2 the
3 good
4 5shielding
6 7 effect
8 and small size of the sensor at the
△x(mm)
time, this paper selects ∆𝑥 = 3 mm.
Figure 4. The relationship between anti-interference degree and shielding layer area.

Figure 3.4. Research

The relationship between
on Experiment andanti-interference degree and shielding layer area.
Data Processing Algorithm
3.1. Experimental Platform Construction and Data Collection
3. Research onschematic
The Experiment andofData
diagram Processing
the non-contact Algorithm
voltage measurement system based on
the multi-capacitor array is shown in Figure 5. A 50 Hz power frequency voltage is
3.1. Experimental Platform Construction and Data Collection
generated by a voltage source and the experimental comparison test is carried out using
The schematic
an Agilent diagram
16-digit of the non-contact
digital multimeter. voltage
The pico 5443D measurement
16-bit PC oscilloscopesystem based
is utilized to on th
multi-capacitor array is shown in Figure 5. A 50 Hz power frequency voltage is generate
simultaneously record the waveforms of the voltage sensor and digital multimeter. This
sensor mainly measures the rated voltage of 220 V, so it is measured within the range of
by a voltage source and the experimental comparison test is carried out using an Agilen
80–120% of the rated voltage specified by the IEC60044-7 standard. The power frequency
16-digit digital multimeter.
experimental The pico
platform is shown 5443D
in Figure 6. 16-bit PC oscilloscope is utilized to simultan
ously record the waveforms of the voltage sensor and digital multimeter. This senso
mainly measures the rated voltage of 220 V, so it is measured within the range of 80–120
of the rated voltage specified by the IEC60044-7 standard. The power frequency exper
mental platform is shown in Figure 6.

16-bit digital multimeter

 multi-capacitor array is shown in Figure 5. A 50 Hz power frequency voltage is generated
by a voltage source and the experimental comparison test is carried out using an Agilent
16-digit digital multimeter. The pico 5443D 16-bit PC oscilloscope is utilized to simultane-
ously record the waveforms of the voltage sensor and digital multimeter. This sensor
mainly measures the rated voltage of 220 V, so it is measured within the range of 80–120%
Sensors 2022, 22, 8573 7 of 13
of the rated voltage specified by the IEC60044-7 standard. The power frequency experi-
mental platform is shown in Figure 6.

16-bit digital multimeter

Sensors

Transmission
line
High voltage
source
Power supply

Computer
16-bit data
Sensors 2022, 22, x FOR PEER REVIEW collector 8 of 15

Figure 5.
Figure 5. Schematic
Schematic diagram
diagram of
of the
the measurement
measurement system.
system.

Figure6.6.Power
Figure Powerfrequency
frequencyexperimental
experimentalplatform.
platform.(1)—high
(1)—highvoltage
voltagesource,
source,(2)—PC
(2)—PCterminal
terminaldata
data
acquisition, (3)—pico 5443D 16-bit PC oscilloscope, (4)—Agilent 16-bit digital multimeter, (5)—
acquisition, (3)—pico 5443D 16-bit PC oscilloscope, (4)—Agilent 16-bit digital multimeter, (5)—shield
shield cover, (6)—wire to be tested, (7)—sensor, and (8)—9V dry battery.
cover, (6)—wire to be tested, (7)—sensor, and (8)—9 V dry battery.

Inorder
In ordertotoobtain
obtainthe
thelinearity
linearity ofof
thethe voltage
voltage sensor,
sensor, thethe output
output voltage
voltage of DC-5
of the the DC-5
kV
kV programmable AC source is adjusted while the distance
programmable AC source is adjusted while the distance between the sensor and between the sensor andthe
the
transmission line remains unchanged and the output voltage of the digital
transmission line remains unchanged and the output voltage of the digital multimeter and multimeter and
thevoltage
the voltagesensor
sensorisisrecorded
recorded at at
thethe same
same time.
time. TheThe fitting
fitting curve
curve of experimental
of the the experimental re-
results
issults is obtained,
obtained, but before
but before this,
this, the the experimental
experimental datato
data need need to be processed
be processed to obtain
to obtain the
the most
most accurate
accurate measurement
measurement data. data.

3.2.Signal
3.2. SignalProcessing
ProcessingAlgorithms
Algorithms
Tomeet
To meetthe
theperformance
performance requirements
requirements of of a multi-sensor
a multi-sensor voltage
voltage measurement
measurement sys-
system,
tem, signal
signal processing
processing algorithms
algorithms are key.are key.
The The processing
signal signal processing
algorithmalgorithm
must notmust
onlynot only
ensure
ensure
the the accuracy
accuracy of the measurement,
of the measurement, but also be but also
easy to be easy to implement,
implement, flexible, andflexible, and
adaptable.
adaptable. In a multi-sensor voltage measurement system, the measurementdepend
In a multi-sensor voltage measurement system, the measurement results mainly results
mainly
on depend onof
the processing thethe
processing of thesignals
measurement measurement signals of
of all sensors. all output
The sensors.ofThe
theoutput
voltageof
the voltage
sensor sensor
represents represents
the theof
distribution distribution of the space
the space electric electricthe
field around field around
wire, the wire,
including the
including the
interference interference
field field signal
signal introduced introduced
by other by other
space electric space
fields and electric fields
the noise and the
caused by
noise caused by the uncertainty of the position of the voltage sensor itself. At present, the
the uncertainty of the position of the voltage sensor itself. At present, the commonly used
commonly used signal processing algorithms mainly include the mean value method, the
spatial Fourier transform method, and the Kalman filter method.
The average value algorithm approximates the voltage value to be measured by sum-
ming the discrete measurement values, so the approximation error is closely related to the
number of measurement values, that is, the number of sensors in the sensor ring. It can be
seen that, only when the number of sensors is sufficient, the influence of the approxima-
Sensors 2022, 22, 8573 8 of 13

signal processing algorithms mainly include the mean value method, the spatial Fourier
transform method, and the Kalman filter method.
The average value algorithm approximates the voltage value to be measured by
summing the discrete measurement values, so the approximation error is closely related
to the number of measurement values, that is, the number of sensors in the sensor ring.
It can be seen that, only when the number of sensors is sufficient, the influence of the
approximation error on the measurement results is small enough. Further, when the
number of sensors is small, the measurement error will be very large. For the spatial
Fourier transform, as the measurement error decreases exponentially with the number of
sensors, high accuracy can be obtained without a large number of sensors. Even when the
relative position of the interference voltage and the sensor ring is special, the influence
of the interference voltage on the measurement accuracy can be completely eliminated
in theory. Therefore, when the number of sensors is limited, the accuracy of the spatial
Fourier transform is far better than the average algorithm. Kalman filter is a method to
obtain the best filter value through signal estimation and correction based on the known
system model and noise characteristics. The algorithm based on the Kalman filter can be
regarded as an improvement of the average algorithm. As the object of averaging is the
optimally filtered sensor output sequence, it can greatly reduce the measurement error of
the averaging algorithm when the number of sensors is limited.
In summary, this paper uses the Kalman filter algorithm to process the signal. Kalman
filter is a method used to obtain the best filter value by estimating and correcting the
signal for the known system model and noise characteristics. This paper proposes a signal
processing algorithm for a multi-sensor AC voltage measurement system based on Kalman
filtering. According to the electric field distribution around the wire, the state equation
of the system is established and the recursive method of Kalman steady-state filtering
is provided. The algorithm based on Kalman filter can be regarded as an improvement
to the traditional average algorithm. As the object of averaging is the optimally filtered
sensor output sequence, the measurement error of the averaging algorithm when the
number of sensors is limited can be greatly reduced. According to the analysis of the
characteristics of the interference voltage, determining the appropriate filter parameters is
the key to the success or failure of the Kalman filter algorithm. When the characteristics of
the interference voltage are unknown, it is a necessary part of the measurement system to
correct the Kalman filter by means of parameter estimation.
In a variety of applications, the structure of multiple parallel conductors is most
common. In the modeling of multi-sensor AC voltage measurement, the double parallel
straight wire model is often used. By studying the electric field distribution characteristics
of double conductors, it is not difficult to generalize to more parallel straight conductors.
As shown in Figure 7, there are two parallel long straight wires, where the distance is D, the
magnitude of the voltage on the two wires is U, and the direction of the current is opposite.
The distances between point P and the two wires are r1 and r2 , respectively. When the
distance between the wires and the radius of the wires is much smaller than the length
of the wires, it can be approximated as the parallel plane electric field generated by two
infinitely long straight wires.
Assuming that the linear charge density of a uniformly charged infinitely long straight
wire is τ, the dielectric constant in a vacuum is ε 0 , the point P0 is selected as the reference
point, the vertical distance between the point P0 and the wire is r0 , and ϕ(r0 ) = 0. In
Figure 7, the potential difference of two infinitely long straight wires at point P is as follows:
τ r
ϕ (r1 ) = − ln 1 (5)
2πε 0 r0
τ r2
ϕ (r2 ) = ln (6)
2πε 0 r0
 Sensors 2022, 22, 8573 9 of 13

According to Formula (6) and Formula (7), the potential difference at point P can be
obtained as follows:
 
τ r2 r
ϕ ( P ) = ϕ (r1 ) + ϕ (r2 ) = ln − ln 1 (7)
2πε 0 r0 r0

Then,
τ r2
ϕ( P) = ln (8)
2πε 0 r1
The multi-sensor system in this paper adopts the sensor ring structure. For the
convenience of calculation, Formula (9) is transformed into the expression in the cylindrical
coordinate system:
q
τ ( D − r cos ϕ)2 + (r sin ϕ)2
ϕ( P) = ln (9)
2πε 0 r

Take the negative gradient of ϕ P as the following:

3Dr cos ϕ − 2D2 − 3r2

 
∂ϕ τ
EP = − = 1+ (10)
∂r 2πε 0 D2 + r2 − 2Dr cos ϕ

Assuming D = βr, then:

3β cos ϕ − 2β2 − 1
 
τ
EP = 1+ 2 (11)
2πε 0 β − 2β cos ϕ + 1

Among them, 3β cos ϕ − 2β2 − 1 /( β2 − 2β cos ϕ + 1) represents the influence of the

interference electric field generated by the parallel wires. When the interference voltage
value is not equal to the voltage value to be measured, and is K times the voltage value
to be measured, the influence of the interference electric field is K 3β cos ϕ − 2β2 − 1 /

( β2 − 2β cos ϕ + 1).
The sensor structure designed in this paper is a ring structure and a voltage measure-
Sensors 2022, 22, x FOR PEER REVIEW 10 o
ment system realized by multiple sensors is used, which can greatly reduce the difficulty of
analysis and calculation. The sensor arrangement of the multi-sensor system is shown in
Figure 8. A plurality of voltage sensors is evenly distributed on a ring of radius r.

D
U1 r1
P
r2
y
U2

Figure 7.Schematic
Figure 7. Schematic diagram
diagram of double
of double parallel
parallel AC conductors.
AC conductors.

The multi-sensor system in this paper adopts the sensor ring structure. For the c
venience of calculation, Formula (9) is transformed into the expression in the cylindr
coordinate system:

𝜏 (𝐷 − 𝑟 cos 𝜑) + (𝑟 sin 𝜑)
 1)/(𝛽 − 2𝛽 cos 𝜑 + 1).
The sensor structure designed in this paper is a ring structure and a voltage meas-
urement system realized by multiple sensors is used, which can greatly reduce the diffi-
culty of analysis and calculation. The sensor arrangement of the multi-sensor system is
Sensors 2022, 22, 8573 shown in Figure 8. A plurality of voltage sensors is evenly distributed on a ring of 10
radius
of 13
𝑟.

Voltage sensor

r
D

Voltage to be Interference
measured U1 voltage U2

Figure8.8. Schematic
Figure Schematicdiagram
diagramof
ofthe
thesensor
sensorarrangement
arrangementof
ofthe
themulti-sensor
multi-sensorsystem.
system.

According to the electric field distribution of the transmission wire, the output se-
quence U(n) of the voltage sensor can be obtained as shown in Formula (12). Taking the
voltage U1 to be measured as the system state quantity and the voltage sensor output
sequence U(n) as the measurement value, the state equation and the measurement equation
are established respectively. In the state equation shown in Formula (13), as the value of the
voltage to be measured remains relatively unchanged, the system state remains unchanged.
In the measurement equation shown in Formula (14), the measurement noise z(n) is the
influence of the interference electric field on the output of the voltage sensor.
!
n
− 2β2 − 1


τ 3β cos 2π N
U(n) = V1 (n) + V2 (n) 2 n

(12)
2πε 0 β − 2β cos 2π N +1

V1 (n + 1) = V1 (n) (13)
λ
V(n) = V (n) + z(n) (14)
2πε 0 1
Then,
n
− 2β2 − 1


λ 3β cos 2π N
z(n) = ·V2 (n) 2 n

(15)
2πε 0 β − 2β cos 2π N +1
The corresponding Kalman steady-state filter equation is as follows:
  
λ
V̂1 (n|n ) = V̂1 (n|n − 1) + M U (n) − · V̂1 (n|n − 1)

2πε 0 (16)
V̂1 (n|n + 1) = V̂1 (n|n )


Among them, V̂1 (n|n − 1) is the estimated value of the state quantity V1 based on the
first n − 1 measurement values; V̂1 (n|n ) is the state variable based on the nth measurement
value. For the estimated value of the quantitative update, M is the steady-state gain that
minimizes the covariance of the estimated error for the system noise Z (n). The properties
of the system noise z(n) determine the value of M.
By synthesizing the two Kalman steady-state filter equations, we can obtain the
following:  
λ
V̂1 (n + 1|n ) = 1 − M V̂1 (n|n − 1) + MÛ (n) (17)
2πε 0
 
λ λ λ
Û (n|n ) = 1−M V̂1 (n|n − 1) + M Û (n) (18)
2πε 0 2πε 0 2πε 0
where Û (n|n ) is the optimal filtering result of the measured value. The filtered sensor
output sequence is averaged and the result is used as the voltage value to be measured.
Linear fitting is performed on the processed experimental data, as shown in Figure 9;
the expression of the fitting curve is y = 3.0640x − 2.0298 (x is the voltage of the transmis-
 Sensors 2022, 22, 8573 11 of 13

Sensors 2022, 22, x FOR PEER REVIEW 12 of 15

sion line, kV; y is the output voltage of the voltage sensor, mV). The results show that the
voltage sensor has good linearity.

850
Measured Voltage
% Error 0.010
800
Measured Voltage[V]
750 0.005

% Error
700
0.000
650

600 -0.005

550
-0.010

500
160 180 200 220 240 260 280
Actual Voltage[V]

Figure
Figure9.9.The
Themeasured
measuredvoltage
voltageusing
usingthe
theprototype
prototypeand
andthe
theratio
ratioerror
errorcharacteristics.
characteristics.

According
Accordingtotothethemeasured
measuredvoltage
voltageandandactual
actualvoltage
voltageininFigure
Figure9,9,ititcan
canbebeobtained
obtained
that
thatthe
thetheoretical
theoreticalvoltage
voltagedivider
dividerratio
ratio𝐾Knofofthe
thevoltage
voltagesensor
sensorisisabout
about325.47
325.47and
andthe
the
voltage
voltagereduction
reductionvalue
value𝑈Uc can
canbe
beexpressed
expressed byby the
the following
following formula:
formula:

𝑈c =
U =K𝐾n 𝑈
Ud (19)
(19)
In order to evaluate the accuracy of the sensor more intuitively, it is necessary to
In order to evaluate the accuracy of the sensor more intuitively, it is necessary to
calculate the relative error (δ%) [31] and phase error 𝜑 of the sensor:
calculate the relative error (δ%) [31] and phase error ϕu of the sensor:
𝑈 −𝑈
δ%= Uc − Ul × 100% (20)
δ% = 𝑈 × 100% (20)
Ul
𝜑 =𝜑 −𝜑 (21)
ϕu = ϕd − ϕt (21)
where 𝑈 is the voltage of the wire, 𝜑 is the phase measured by the voltage probe, and
is theUoutput
𝜑where phase ofofthe
l is the voltage theTektronix
wire, ϕd ishigh-voltage probe. Itby
the phase measured can
thebevoltage
seen from
probe, Table
and 1ϕt
that, within the 80–120% rated voltage range specified by the standard, the measurement
is the output phase of the Tektronix high-voltage probe. It can be seen from Table 1 that,
error of the sensor is controlled within ±1% and the phase error is controlled withinerror
within the 80–120% rated voltage range specified by the standard, the measurement 4°.
◦ . U in
𝑈of intheTable
sensor1 isisthe rated voltage
controlled withinand
±1%theand
rated
thevoltage measured
phase error in this paper
is controlled is 4220
within V.n
Table 1 is the rated voltage and the rated voltage measured in this paper is 220 V.
Table 1. Calculation results of the sensor measurement value ratio difference and angular differ-
ence.
Table 1. Calculation results of the sensor measurement value ratio difference and angular difference.

Measuring Point
Measuring Point 𝑼𝒍 /V Ul /V 𝑼𝒔 /mVUs /mV 𝜺/% ε/% 𝝋𝒖ϕ/(°) ◦
u /( )
80%𝑈80%Un 175.82175.82 535 535 −0.97%
−0.97% 3.12
3.12
85%𝑈85%Un 186.85186.85 578 578 0.65%
0.65% 2.79
2.79
90%𝑈90%Un 197.87197.87 613 613 0.86%
0.86% 3.41
3.41
95%Un 208.76 640 −0.21% 3.16
95%𝑈 208.76 640 −0.21% 3.16
100%Un 219.84 669 −0.98% 2.84
100%𝑈
105%Un 219.84230.76 669 702 −0.98%
−0.94% 2.84
2.61
105%𝑈
110%Un 230.76241.78 702 742 −0.94%
−0.16% 2.61
2.53
110%𝑈 n
115%U 241.78252.77 742 782 −0.16%
− 0.94% 2.53
3.18
120%Un 263.95 813 0.27% 3.12
115%𝑈 252.77 782 −0.94% 3.18
120%𝑈 263.95 813 0.27% 3.12
The instantaneous value waveforms of the input and sensor output voltages are shown
in Figure 10. It can be seen that their amplitudes and phase angles are basically the same
The instantaneous value waveforms of the input and sensor output voltages are
and the sensor can highly recover the voltage to be measured.
shown in Figure 10. It can be seen that their amplitudes and phase angles are basically the
same and the sensor can highly recover the voltage to be measured.
Sensors 2022, 22, x FOR PEER REVIEW 13 of 15
Sensors 2022, 22, 8573 12 of 13

Output voltage/V

Figure
Figure10.
10.Momentary
Momentaryvalues
valuesofofinput
input(red)
(red)and
andsensor
sensoroutput
output(blue)
(blue)voltage.
voltage.

4.4.Conclusions
Conclusions
Aimingat
Aiming ataddressing
addressing thethe problems
problemsof ofthe
theexisting
existingvoltage
voltage sensors
sensorssuch
suchas as
large volume,
large vol-
difficult installation, complex manufacturing process, high cost, and the fact that the relative
ume, difficult installation, complex manufacturing process, high cost, and the fact that the
position of the wire and the sensor has a great influence on the measurement accuracy, this
relative position of the wire and the sensor has a great influence on the measurement ac-
paper proposes a multi-electrode array non-contact voltage sensor manufactured using
curacy, this paper proposes a multi-electrode array non-contact voltage sensor manufac-
PCB technology for the structure and a signal processing algorithm for a multi-sensor
tured using PCB technology for the structure and a signal processing algorithm for a
voltage measurement system. The experimental measurement of the developed sensor
multi-sensor voltage measurement system. The experimental measurement of the devel-
is carried out in the laboratory. The experimental and simulation results show that the
oped sensor is carried out in the laboratory. The experimental and simulation results show
ratio error is less than ±1% and the phase difference is less than 4◦ under the condition
that the ratio error is less than ±1% and the phase difference is less than 4° under the con-
of power frequency 50 Hz. When the relative position of the wire and the sensor is in the
dition of power◦ frequency 50 Hz. When the relative position of the wire and the sensor is
direction of 45 and the offset is 10 mm, the relative error of the voltage sensor with single-
in the direction of 45° and the offset is 10 mm, the relative error of the voltage sensor with
electrode structure is 17.62%, while the relative error of the multi-electrode voltage sensor
single-electrode structure is 17.62%, while the relative error of the multi-electrode voltage
designed in this paper is only 0.38%. Compared with the single-electrode structure, the
sensor
sensordesigned
structurein this paper
proposed is only
in this 0.38%.
paper Compared
is effective. with
Errors the single-electrode
caused by changes in thestructure,
relative
the sensor of
positions structure
wires andproposed
sensorsinare this paper isThe
reduced. effective.
sensor Errors caused
meets the by changesneeds
development in theof
relative positions of wires and sensors are reduced. The sensor meets the
smart grid measurement sensor intelligence, miniaturization, and convenience. The sensor development
needs
can beofused
smartingrid measurement
low-voltage sensor networks;
distribution intelligence,forminiaturization,
example, for the and convenience.of
measurement
The
220 V single-phase line voltage. However, the sensor designed in this paper onlythe
sensor can be used in low-voltage distribution networks; for example, for meas-
measures
urement
the low voltage of the power frequency of 50 Hz. In the future, in-depth research paper
of 220 V single-phase line voltage. However, the sensor designed in this can be
only measures
carried the lowthe
out to reduce voltage
error of the sensor,
of the power frequency of 50 Hz.itsInbandwidth
as well as increase the future,andin-depth
range,
research can
among others. be carried out to reduce the error of the sensor, as well as increase its band-
width and range, among others.
Author Contributions: Conceptualization, W.Z.; Data curation, Y.Y. and M.H.; Investigation, Y.Y.;
Author Contributions:
Methodology, Y.Y. andConceptualization,
J.Z.; Software, K.C.;W.Z.; Data curation,
Validation, Y.Y.
Y.Y., R.H. and
and M.H.;
M.H.; Investigation,
Visualization, Y.Y.;
Y.Y. and
Methodology, Y.Y. and J.Z.;
K.C.; Writing—original Software,
draft, K.C.; Validation,
Y.Y.; Writing—review Y.Y., R.H.
& editing, Y.Y.and
andM.H.; Visualization,
J.Z. All authors haveY.Y.
readand
and
K.C.; Writing—original
agreed to the publisheddraft, Y.Y.;
version of Writing—review
the manuscript. & editing, Y.Y. and J.Z. All authors have read
and agreed to the published version of the manuscript.
Funding: This research was supported by the “Research and development of new smart sensor
Funding: This
technology research the
to promote wasdevelopment
supported by of the “Research
green and development funding.
energy” (202104BN050011) of new smart sensor
technology to promote the development of green energy” (202104BN050011) funding.
Conflicts of Interest: The authors declare no conflict of interest.
Conflicts of Interest: The authors declare no conflict of interest.
References
References
1. Yang, P.; Wen, X.; Lv, Y.; Chu, Z.; Peng, C. A non-intrusive voltage measurement scheme based on MEMS electric field sensors:
1. Yang, P.; Wen,
Theoretical X.; Lv,and
analysis Y.; Chu, Z.; Peng, verification
experimental C. A non-intrusive voltage
of AC power measurement
lines. scheme2021,
Rev. Sci. Instrum. based92,
on065002.
MEMS [CrossRef]
electric field sensors:
[PubMed]
2. Theoretical analysisH.A.
Reza, M.; Rahman, and experimental
Non-Invasive verification
Voltage of AC
Measurement power for
Technique lines. Rev. Sci.
Low Voltage AC Instruments 2021, 92,of 065002,
Lines. In Proceedings the 2021
https://doi.org/10.1063/5.0052678.
IEEE 4th International Conference on Electronics Technology (ICET), Chengdu, China, 7–10 May 2021; pp. 143–148.
2.3. Reza,
Zhang,M.;Y.;Rahman, H.A.
Zhang, C.; Li,Non-Invasive Voltage
H.; Chen, Q. An onlineMeasurement Technique
detection method for Low
for capacitor Voltage
voltage AC Lines.with
transformer In Proceedings of the 2021
excessive measurement
IEEE
error4th International
based Conference
on multi-source on Electronics
heterogeneous Technology
data fusion. (ICET), Chengdu,
Measurement 2021, 187, China,
110262.7–10 May 2021; pp. 143–148.
[CrossRef]
4. Femine, A.D.; Gallo, D.; Landi, C.; Schiavo, A.L.; Luiso, M. Low Power Contactless Voltage Sensor for Low Voltage Power
Systems. Sensors 2019, 19, 3513. [CrossRef] [PubMed]
Sensors 2022, 22, 8573 13 of 13

5. Guo, Z.; Shi, W.; Lei, T.; He, Z.; Kang, P. Transient voltage measurement for UHV AC substation using non-contact electro-optic
field sensor. In Proceedings of the 18th International Conference on AC and DC Power Transmission (ACDC 2022), online,
2–3 July 2022; pp. 695–699. [CrossRef]
6. Fuqing, Y.; Lu, J. Non-contact sensing and modelling on the voltage and electric field for powered transmission line. J. Phys. Conf.
Ser. 2020, 1453, 012115. [CrossRef]
7. Wang, J.; Gao, L. Research on the Algorithm and Test of Transmission Line Voltage Measurement Based on Electric Field Integral
Method. IEEE Access 2018, 6, 72766–72773. [CrossRef]
8. Xu, Q.; Dong, C.; Huang, Y. An Electronic Voltage Transducer Based on Electrostatic Charge Escape. IEEE Trans. Instrum. Meas.
2022, 71, 6006808. [CrossRef]
9. Wetula, A.; Bień, A.; Parekh, M. New Sensor for Medium- and High-Voltage Measurement. Energies 2021, 14, 4654. [CrossRef]
10. Fan, J.; Ai, C.; Guo, A.; Yan, X.; Wang, J. Evaluation of Electric Field Integral Voltage Measurement Method of Transmission Line
Based on Error Transmission and Uncertainty Analysis. Sensors 2021, 21, 4340. [CrossRef] [PubMed]
11. Wang, J.; Zhao, Y.; Li, W.; Zeng, X.; Tang, J.; Wang, Y.; Deng, X. Research on Transmission Line Voltage Measurement Method of
D-Dot Sensor Based on Gaussian Integral. Sensors 2018, 18, 2455. [CrossRef]
12. Li, J.; Liu, J.; Peng, C.; Liu, X.; Wu, Z.; Zheng, F. Design and Testing of a Non-Contact MEMS Voltage Sensor Based on Single-Crystal
Silicon Piezore-sistive Effect. Micromachines 2022, 13, 619. [CrossRef]
13. Leng, Y.; Shu, L.; Qian, Q.; Chen, C.; Lin, Y. Contactless Voltage Measurement Considering Spatially Dependent Voltage
Compensation. Sens. Mater. 2022, 34, 1057. [CrossRef]
14. Peng, J.; Bian, J.; Jia, S.; Kang, X.; Yu, H.; Yang, Y. Design of an electro-optical sensor with high sensitivity for AC electric field
measurement. Sens. Rev. 2022, 42, 187–194. [CrossRef]
15. Liu, S.; Chen, Y.; Jiang, J.; Wu, Y.; Guo, J.; Chen, L. Quantum enhanced electro-optic sensor for E-field measurement. Opt. Express
2021, 29, 32865–32874. [CrossRef] [PubMed]
16. Li, J.; Lin, Y.; Xie, N. Design of optical voltage sensor based on electric field regulation and rotating isomerism electrode. arXiv
2022, arXiv:2206.06822.
17. Edmunds, J.L.; Sonmezoglu, S.; Martens, J.; Von Meier, A.; Maharbiz, M.M. Optical voltage sensor based on a piezoelectric thin
film for grid applications. Opt. Express 2021, 29, 33716. [CrossRef]
18. Yang, P.; Wen, X.; Chu, Z.; Ni, X.; Peng, C. Non-intrusive DC voltage measurement based on resonant electric field microsensors.
J. Micromech. Microeng. 2021, 31, 064001. [CrossRef]
19. Shenil, P.; George, B. An auto-balancing scheme for non-contact ac voltage measurement. In Proceedings of the 2018 IEEE 9th
International Workshop on Applied Measurements for Power Systems (AMPS), Bologna, Italy, 26–28 September 2018; pp. 1–5.
20. Gorla, D.P.M.; Janus, P.; Edin, H. Non-contact Voltage Measurement Technique for On-Line Monitoring of Transient Overvolt-ages.
In Proceedings of the Nordic Insulation Symposium, Trondheim, Norway, 13–15 June 2022; Volume 27.
21. Yan, X.; Wang, L.; Zhou, Y. Design of Reciprocal Optical Path for Voltage Measurement Based on Pockels Effect. In Proceedings of
the 2021 International Conference on Digital Society and Intelligent Systems (DSInS), Chengdu, China, 19–21 November 2021;
pp. 271–274.
22. Jiang, S.; Jia, R.; Tao, K.; Wu, Y.; Liu, S. High-resistance voltage dividers fabricated by thin polysilicon films in silicon drift
detectors. J. Mater. Sci. Mater. Electron. 2019, 30, 6617–6627. [CrossRef]
23. Martins, A.V.; Bacurau, R.M.; dos Santos, A.D.; Ferreira, E.C. Nonintrusive energy meter for nontechnical losses identification.
IEEE Trans. Instrum. Meas. 2019, 69, 1140–1147. [CrossRef]
24. Hrbac, R.; Kolar, V.; Bartlomiejczyk, M.; Mlcak, T.; Orsag, P.; Vanc, J. A Development of a Capacitive Voltage Divider for High
Voltage Measurement as Part of a Combined Current and Voltage Sensor. Elektron. Elektrotech. 2020, 26, 25–31. [CrossRef]
25. Haberman, M.A.; Spinelli, E.M. A noncontact voltage measurement system for power-line voltage waveforms. IEEE Trans.
Instrum. Meas. 2019, 69, 2790–2797. [CrossRef]
26. Haberman, M.A.; Spinelli, E.M. Noncontact ac voltage measurements: Error and noise analysis. IEEE Trans. Instrum. Meas. 2018,
67, 1946–1953. [CrossRef]
27. Thomas, A.J.; Iyyappan, C.; Reddy, C.C. On the measurement of surface voltage of insulators and bushings. IEEE Trans. Power
Deliv. 2021, 37, 464–471. [CrossRef]
28. Zhou, Z.; Shu, G.; Li, D. Research on non-contact household voltage measuring device based on capacitive coupling principle.
Electr. Autom. 2020, 42, 98–100.
29. Shenil, P.; George, B. Development of a nonintrusive true-rms ac voltage measurement probe. IEEE Trans. Instrum. Meas. 2019, 68,
3899–3906. [CrossRef]
30. Kong, H.; Shan, M.; Sukkarieh, S.; Chen, T.; Zheng, W.X. Kalman filtering under unknown inputs and norm constraints. Automatica
2021, 133, 109871. [CrossRef]
31. Kozlov, V.K.; Ivanov, D.A.; Kirzhatskikh, E.R. Voltage Measuring Sensor Based on Capacitive Voltage Divider. In Proceedings of
the 2022 International Ural Conference on Electrical Power Engineering (UralCon), Magnitogorsk, Russia, 23–25 September 2022;
pp. 250–254.