Use of Capacitive Probes to Detect Asymmetry and Earth Fault in a Medium-Voltage Power Network

Abstract

The detection of short circuits in a medium-voltage (MV) network is a complex issue due to the way the neutral point works. An additional difficulty is the relatively large load asymmetry. The methods used so far include complex equipment (e.g., a system of voltage transformers) for use mainly in power stations. The detection of short circuits deep in the network is therefore difficult, and this could facilitate the process of fault localization and limit the areas that should be disconnected for the time of fault removal. This article presents the new concept of using a system of capacitive probes as a simple and cheap tool that allows for the detection of a short circuit in an MV network based on the assessment of the zero-voltage component. This component is considered to be one of the basic starting criteria for various types of specialist earth-fault protections. Appropriately placed capacitive probes—through the existence of capacitive coupling with phase conductors—record the voltages of individual phases, including the total resultant voltage, which is the criterion for detecting a short circuit in the system. An important advantage of using such a solution is that capacitive probes allow for voltage measurement and assessment of line asymmetry in a non-contact and, therefore, safe manner. The presented concept has been tested in the laboratory and supported by simulation studies. The modeling of the system was based on the parameters of real structures used in overhead lines, recreated in laboratory conditions. Obtaining positive results of the simulation studies—primarily the appropriate sensitivity of short-circuit detection, confirmed in the laboratory—allows for the creation of a prototype of the device and the commencement of field tests, which will be the subject of further work conducted by the authors.

1. Introduction

Analyzing the scope of applications of electromagnetic probes in the power industry, several areas in which they have proven to be particularly useful can be distinguished. Apart from typical applications related to remote control and communication with high-voltage objects, several characteristic tasks that field probes perform can be listed. Currently, the basic areas of applications include the following: diagnostics of high-voltage insulation systems; monitoring the operation of high-voltage devices; recording events related to the power devices and measurements of electrical quantities [1,2,3].

Diagnostic tests, which are currently increasingly associated with the continuous monitoring of high-voltage devices, can be considered the most dynamically developing field due to their main concerns, which include the objects of great importance in a power system (switchgears, transformers, and generators). The monitored quantity is the energy or peak value of the signal from partial discharge (PD) pulses [4] and the frequency of recorded events in a given time interval [5], as well as the trends of changes (also in connection with other parameters, e.g., the phase angle) [6]. The occurrence of partial discharges always indicates a developing defect in the insulation system, and therefore constitutes important information about the approaching failure of the monitored device. Quantitative analysis of the partial discharge generation phenomenon using field probes aims to determine the dynamics of the development of the defect that has occurred. In addition, qualitative analysis allows us to determine the cause of partial discharges, i.e., to identify the type of defect and often to locate its source [7,8,9].

Registration of events related to the operation of power devices is primarily the detection and recording of overvoltage phenomena caused by external factors (lightning surges) or internal factors (slow- and fast-changing surges). Such systems in their simplest form enable counting events and recording the value of the surge current (e.g., flowing through a surge arrester or lightning protection system) [10]. In a more complex configuration, the recorder can be used to analyze the shape of the voltage waveform and assess the scale of the threat to devices operating in its vicinity. Using a sufficiently large number of probes distributed over a given area, it is also possible to determine storm activity (systems creating so-called discharge density maps during the year) [11].

The next discussed application has the longest history. Measurements of quantities such as current, voltage, electromagnetic field strength, and radio-frequency disturbance voltage are very common not only in relation to high-voltage devices but also in the entire field of electrical engineering. In particular, the measurement of electric- and magnetic-field intensities and disturbances has recently gained importance because it concerns the extremely important and widely publicized problem of the electromagnetic compatibility of electrical devices and the assessment of their impact on the environment [12,13].

Field probes are also applied to assess the quality of electric power and can be used as indicators of power network disturbances (e.g., they can be used to detect short circuits and earth faults) [14].

A specific application of electromagnetic probes, in the areas of current and voltage measurement, may be the in-depth detection of short circuits in medium-voltage distribution networks. Such a solution using inductive coupling was proposed by, for example, the SOFTIN company [15,16]. The idea of the system is based on the following principle. The network system is divided into sections, each of which is monitored by one signaling device (Figure 1). The signaling device set is the combination of a magnetic field probe with a measuring system and a signaling system (communication between the system elements is carried out by radio). The probe detecting sudden changes in the magnetic field intensity—resulting from an earth fault or phase-to-phase fault—is installed several meters under the line and connected with a wire to the measuring system (Figure 1b). An important element of the installation process is to position the probe so that it is as symmetrical as possible in relation to the individual phases because the recorded signal is a result of the field intensity from the current flowing in the individual wires (incorrect positioning will result in a lack of symmetry during normal network operation and, in extreme cases, even start-up of the signaling device).

An alternative to inductive probes may be capacitive probes, although they are based on a different starting criterion (in this case, the zero-component voltage and not the current). Such a solution is certainly cheaper (initial estimates indicate a cost of the system of around USD 100) than the typical devices used in the power industry (voltage transformers) and has a number of interesting properties, the most important of which is operation without contact with high voltage, which limits the impact of various types of dangerous overvoltage or overcurrent interferences. In addition, that kind of system will react to the zero-voltage component and will be sensitive to changes in the symmetry of the three-phase voltage of the network, which, in addition to detecting earth faults, can be used to monitor the quality of transmitted electrical energy. A prototype system of this type proposed by the authors will be presented in this article.

This article is divided into the four following parts: Section 2 will present the concept of a new diagnostic method for detecting ground faults based on the readings of capacitive probes; Section 3 describes the probes used and the simulation environment, along with the proposed model corresponding to the actual operating conditions of the system; Section 4, in turn, includes the presentation and discussion of the results; Section 5 is a summary, indicating the authors’ further plans for continuing the research.

2. The Concept of Using Capacitive Probes to Detect Asymmetries and Earth Faults in Medium-Voltage Power Networks

The issue of detecting voltage asymmetry in a three-phase power network is important for two reasons. First, asymmetry is one of the parameters indicating a decrease in the quality of electric power, which is currently—according to the regulation [17] and the guidelines of the standard [18]—an unacceptable situation in the power industry. Second, such a situation may indicate the occurrence of a serious disturbance in the operation of the power system, such as an earth fault, which in turn is particularly important from the point of view of the operation of the medium-voltage network. Both the first and the second event can be detected using the capacitive probes described in this article, which will be the subject of further detailed considerations.

The issue of detecting earth faults is a complicated problem, analyzed for many years. The team of Professor J. Lorenc from the Institute of Electrical Power Engineering of the Poznań University of Technology has special achievements in this area, he is the creator of admittance criteria for detecting earth faults, which revolutionized this field of technology [19,20,21].

Currently, scientific work focuses on the development of methods for effective detection, identification, and location of short-circuit locations and the selection of grounding element parameters (resistors, the Petersen coil) in order to limit short-circuit currents. The detection and location methods include methods based on traveling waves, described, e.g., in [22], or a very reliable method based on effective impedance calculation [19]. More and more often, advanced artificial intelligence approaches are used in analytical processes of short-circuit detection, such as the Recurrent Neural Network (RNN) model created by Deep Learning (DL) [23] or wavelet continuous transform described, e.g., in [24]. This fault-detection and -location algorithm can also be established on the optimally allocated phasor measurement units [25]. Regardless of the use of advanced algorithms and methods for analyzing ground-fault events, it is important to use appropriate criteria that will enable the start-up and activation of appropriate protection algorithms that will cause the short-circuited line to be switched off.

Currently, several criteria are used to detect such events in medium-voltage networks. Table 1 presents them collectively, taking into account the way the neutral point works [26]. The table shows that one of the basic and universal criteria for detecting earth faults is to determine the value of the zero-sequence voltage component, which can be written as follows:

$$U_0={\displaystyle \frac{1}{3}}\left(U_{L1}+U_{L2}+U_{L3}\right),$$

(1)

where

• U₀—zero sequence voltage;
• U_L1, U_L2, U_L3—voltage of individual phases.

Table 1. Selection of protection against earth faults in line fields depending on the neutral point operation method [21].

Protection Type	Neutral Point Connection
	Insulated	Impedance Earthed		Resonant Earthed	Inductor and Resistor Parallel Circuit
		Without ACFA *	With ACFA *
U0>	+++	+++	+++	+++	+++
directional watt metric protection	-	-	+++	+++	+++
directional varmetric protection	+++	-	-	-	-
I0>	++	+	+	+++	+++
Y0>	++	+	+	+++	+++
G0> unidirectional	-	-	+++	+++	+++
G0> directional	-	-	+	+	+
B0> undirectional	+++	-	-	-	-
RYY0>	-	-	+++	-	-

Explication: +++ recommended, ++ application possible in most cases, + application possible in exceptional cases, - application impossible * Active Current Forcing Automation.

Table 1. Selection of protection against earth faults in line fields depending on the neutral point operation method [21].

Protection Type	Neutral Point Connection
	Insulated	Impedance Earthed		Resonant Earthed	Inductor and Resistor Parallel Circuit
		Without ACFA *	With ACFA *
U0>	+++	+++	+++	+++	+++
directional watt metric protection	-	-	+++	+++	+++
directional varmetric protection	+++	-	-	-	-
I0>	++	+	+	+++	+++
Y0>	++	+	+	+++	+++
G0> unidirectional	-	-	+++	+++	+++
G0> directional	-	-	+	+	+
B0> undirectional	+++	-	-	-	-
RYY0>	-	-	+++	-	-

Explication: +++ recommended, ++ application possible in most cases, + application possible in exceptional cases, - application impossible * Active Current Forcing Automation.

Formula (1), in accordance with the definition contained in the standard [27], describes the zero component in a three-phase system of sinusoidal voltages, which appears as a result of a ground fault or ground asymmetry in fault conditions covered by this article.

Zero-voltage protection (U₀>) does not generally occur on its own. Its common use in protection equipment results from the fact that it is frequently a starting component for many other criteria, such as the following: directional protections and the admittance group. It is also often a criterion that starts algorithms’ advanced and complex analyses aimed at identifying and localizing the incident as presented in [28,29,30]. In some cases, when it is not possible to meet the conditions for selecting earth-fault protection, it can even be a basic protection but acting as a signaling device. In typical solutions, special voltage filters are used to measure the zero sequence (i.e., voltage transformers connected in an open delta configuration). Using appropriately located capacitive probes (from one to three probes, depending on the configuration of the power pole), a similar measurement effect can be obtained but with the use of simpler and much cheaper equipment than, e.g., HV transformers.

The possibilities of using capacitive probes for high-voltage measurement, mentioned in the Introduction Section and described in detail in [15], have so far been considered in the context of single-phase systems, where the influence of neighboring phases was a type of disturbance that had to be eliminated. In a situation where we are interested in the resultant total voltage of all three phases, which is reflected in the zero-sequence voltage, the existence of capacitive couplings between the measuring probe and individual phases turns out to be an advantage. With the proficient placement of individual probes, it can be used in a practical way. Figure 2 shows two typical MV support structure arrangements, in which the insulators operate in a triangular and flat configuration. Against this background, capacitive probes were placed in several variants to present the measurement idea in an illustrative way. For a triangular system, the zero-sequence component—due to the geometric symmetry of the system (and, thus, the symmetrical couplings between individual phases and the probe)—can be determined using a single probe. In a flat system, however, it is necessary to use two or three probes, depending on whether the recording includes measurement of the voltage of individual phases or only the zero-sequence component. In further analyses and simulations, variants with one (Figure 2a) and two probes (Figure 2b) were considered the most optimal.

The idea of measuring the zero-sequence voltage using capacitive probes results directly from the principle of measuring high voltage with a field probe, which was described in great detail in the publication [31]. Each live phase is connected to the probe via capacitive coupling, creating a voltage divider system. This is illustrated in Figure 3. Therefore, the voltage recorded using the probe is the sum of the phase voltages, reduced to a low level in accordance with the measurement system ratio Formula (2). If the condition of the appropriate location of the probe or probe system in relation to the individual phases is met, the value of this voltage is equal to zero for full-voltage symmetry or different from zero in the case of any asymmetry. This situation can be described with the simplified formula below:

$$U=\nu ·\left(U_{L1}+U_{L2}+U_{L3}\right),$$

(2)

where

• U_L1, U_L2, U_L3—voltage of individual phases;
• ν—ratio of the measuring system, calculated from the following formula:

$$\nu ={\displaystyle \frac{C+C_p}{C}}-j{\displaystyle \frac{1}{\omega C_p{R}_p}}$$

(3)

where

• C_p—capacitance of the measuring system (probes, wires, and meter);
• R_p—input resistance of the voltage meter;
• C—coupling capacitance between any phase wire and the probe for a system with one probe or equal to the following:

$$C=C_{11}+C_{21}=C_{12}+C_{22}=C_{13}+C_{23},$$

(4)

for the system with two probes.

The value of capacitive coupling, designated in Formulas (3) and (4) as C, C₁₁, C₂₁, C₂₂, and C₂₃, depends on the distance of the probe from the live object and is of the order of tenths of pF (this will be discussed in the Section 3). The C_p value is the self-capacitance of the probe together with the measuring lead (in the tests it was 420–470 pF, depending on the configuration—see

Table 2. Parameters of models used in simulations.

Model Parameter	Type of System
	Flat	Triangular
Coupling capacitance C between probe and power line wire [pF]	0.147	0.088
Capacitance of measurement system CP [pF]	470	420
Neutral point grounding resistance RN [Ω]	56

). The R_p resistor has a resistance of 1 MΩ and represents the input resistance of the measuring device (in this case a digital oscilloscope).

Figure 3. Capacitive coupling system between (a) three phases and one probe; (b) three phases and two probes.

Figure 3. Capacitive coupling system between (a) three phases and one probe; (b) three phases and two probes.

The concept presented above obviously requires deeper analysis and testing. Laboratory tests and computer simulations were used to verify this concept, which will be presented in detail in the following paragraphs.

3. Materials and Methods

A unipolar capacitive probe of our own design was used to implement the prototype solution. It is shown in Figure 4, and in more detail in the previously mentioned article [31]. Due to its intended use in outdoor conditions, it was appropriately encapsulated so that the tight construction would provide it with proper resistance to the influence of external environmental factors. The test provided in the climatic chamber showed that an increase in humidity of up to 99% (with a simultaneous increase in temperature) does not cause any changes in the probe’s operation, i.e., the value of the signal recorded by it does not change.

The most important features of the probe include its directivity and distance characteristics, which show how the probe’s sensitivity (the amplitude of the signal it records) changes depending on its position relative to the source of electric field propagation. The characteristics are shown in Figure 5 and Figure 6, respectively.

During the tests, the probe was connected to a DSO 3032 digital oscilloscope (Tektronix, Beaverton, OR, USA) with standard parameters (bandwidth 300 MHz, sampling rate 2.5 GS/s). The initial tests aimed at creating a simulation model that would reflect the probe’s field operating conditions as realistically as possible; therefore, the values of the individual parameters of the equivalent circuit were determined based on measurements of the actual system simulated in laboratory conditions. For the purposes of their determination, a fragment of a medium-voltage overhead line was prepared—a pole with a metal crossbar, insulators, and phase conductors—a close environment to which the capacitive probe will operate under operating conditions.

Considering the fact that, in the future, the system will operate under various types of higher-frequency interference, it was additionally equipped with a three-stage sequential filter, thanks to which it is possible to measure the voltage with a capacitive probe even in the presence of strong electromagnetic interference, such as corona discharges, lightning discharges, and substation switching operations. The filtering algorithm consists of such modules as the accumulation filter, wavelet filter, and Butterworth low-pass filter, which—as shown by detailed studies presented in article [31]—gives very good results.

The simulation studies were conducted using the Orcad environment ver. 16.6 (Cadence Design Systems, San Jose, CA, USA). The model implemented in this software is shown in Figure 7. The key elements in the diagram are outlined below:

-

AC voltage sources L1, L2, and L3;

-

Transformer neutral point grounding resistor R4;

-

Capacitive couplings C1, C2, and C3 of individual phases with the field probe;

-

Probe capacitance C5;

-

Input resistance R1 of the measuring device.

In order for a meter input not to reduce the signal amplitude and change the registered signal shape (standard resistance 1 MΩ—resistor R1 in the diagram in Figure 7), a so-called voltage follower implemented by an operational amplifier with high internal impedance (marked as OPAMP in Figure 7) was installed between the A/D converter card and the probe. The results of this system function are shown in Figure 8. Comparing the amplitudes of both waveforms, it can be stated that the use of a properly selected voltage follower allows the signal to be amplified more than six times, which is of significant importance in the context of zero-sequence voltage field measurements.

As mentioned in the introduction to this section, the possibility of using capacitive probes as an asymmetry and earth-fault detector will be considered based on a computer simulation in the Orcad—PSpice program ver. 16.6. Initial tests show that properly selected model parameters (Figure 7) allow us to obtain results that are adequate to reality (Figure 9). Comparing the obtained waveforms from the measurement and from the simulation (point by point), an average relative error of 2% was obtained, which is a satisfactory result. The above allows us to assume that the results obtained as a result of the simulation will have an appropriate practical dimension.

The basic parameters of the equivalent circuit are the values of the coupling capacitances between the phase conductors and the probe. For this purpose, an experiment was carried out using elements of a real medium-voltage power pole. A support insulator was installed on the crossarm in a flat configuration, through which a live wire was led. At an appropriate distance below the insulator, a capacitive probe was placed on a tripod and then moved in a direction parallel to the crossarm, determining the value of the coupling capacitance with a step of 1 cm (Figure 10). The dependence of the coupling capacitance on the position of the probe was obtained (Figure 11). In the case of a system with one probe, the value of the coupling capacitance entered into the model was read from the characteristic (Figure 12) for the distance resulting from the geometric dimensions of the crossarm, included in the documentation provided by the manufacturer. Determining this parameter for a system of two probes was slightly more difficult because it required finding the position of the probes at which the assumption described in Formula (4) would be met. The results of the analysis of this problem are shown in Figure 12. The optimal distance l in relation to the axis of the extreme insulators—at which, with full symmetry of the voltages of all three phases, the measuring system indicates a voltage close to zero—was 61 cm.

4. Results and Discussion

Computer simulations were performed in the model of the electrical system shown in Figure 13. The model parameters were adopted in accordance with the data in Table 2. Due to the fact that the work was of a conceptual nature, it was decided to conduct analyses only for the system with a neutral point grounded by a resistor. The simulations included two groups of issues. First, it was assessed how the natural asymmetry in a three-phase system affects the signal recorded by capacitive probes. In the second part, single-phase earth-fault cases were analyzed with different values of the transition resistance at the fault location. The asymmetry was obtained by changing the voltage amplitude of one of the phases so as to obtain the maximum values of the asymmetry coefficient allowed by various standards [32]:

$$K_{2U}={\displaystyle \frac{∆U_{\mathrm{m}\mathrm{a}\mathrm{x}}}{U_{\mathrm{a}\mathrm{v}}}}·100\%,$$

(5)

where

• ΔU_max—maximum deviation of any of the three-phase voltages from the average value;
• U_av—average value of phase voltages [18].

Figure 13. Diagram of the measurement system model used in the simulation. C—coupling capacitance of the probe with the phase wire of the power line; C_P, R_P—measurement system parameters; R_N—neutral point grounding resistance; R_E—earth-fault resistance.

Figure 13. Diagram of the measurement system model used in the simulation. C—coupling capacitance of the probe with the phase wire of the power line; C_P, R_P—measurement system parameters; R_N—neutral point grounding resistance; R_E—earth-fault resistance.

The earth-fault resistance R_E varied in the range of 0 Ω to 5000 Ω. The obtained results aimed at determining the minimum starting threshold for earth-fault protection.

For each of the selected models of the crossarm system (flat and triangular), many simulations were performed in which the parameters were changed, as described in

Table 3. The values of the voltage amplitude recorded at the output of the capacitive probe and the relative values referred to the maximum voltage value that can occur in the measuring system; simulations were performed for different parameters of the equivalent circuit.

Case Description	Value of the Voltage Amplitude		Relative Value of Measurement Signal
	Flat Formation	Triangular Formation
	[V]		[%]
Phase voltage L1 increased by 2%	0.140	0.095	1.0
Phase voltage L1 decreased by 2%	0.140	0.095	1.0
Earth-fault resistance RE = 0 Ω	14.050	9.457	100.0
Earth-fault resistance RE = 25 Ω	9.714	6.538	69.1
Earth-fault resistance RE = 250 Ω	2.571	1.731	18.3
Earth-fault resistance RE = 2500 Ω	0.307	0.207	2.2

. As a result of the simulation, the values of the voltage amplitude recorded at the output of the capacitive probe were determined. The table also lists the relative values of the measurement signal, i.e., the voltage amplitude values obtained in a given case; they are referred to the maximum values that may appear in the measuring system, i.e., in the case of a three-phase short circuit.

Comparing the voltage values for the flat and triangular formations, it can be seen that they are proportional to each other. This is due to the fact that, from an electrical point of view, they differ only in the value of the transformation ratio (coupling capacitance), which is about 33% lower in the triangular system.

Simulations 1 and 2 show that a change in the voltage amplitude of one of the phases by 2%, i.e., within the limits permitted by the standard [18], does not cause a significant increase in the signal recorded by the probe. Signal amplitude values below 150 mV in the flat system and 100 mV in the triangular system should, therefore, not be treated as an emergency condition. On the other hand, the situation of grounding of one of the phases causes significant asymmetry. This is manifested by an increase in the signal recorded by the probe to a level of at least 307 mV (flat formation) and 207 mV (triangular formation), in the most unfavorable situation among those analyzed, i.e., for an earth-fault resistance of R_E = 2500 Ω. When the earth-fault resistance decreases, the measurement signal increases exponentially (Figure 14).

To sum up, in the tested model system, the threshold voltage value that should trigger the short-circuit signaling can be 300 mV for a system with two probes and 200 mV for a single probe, which is approximately 2% of the maximum signal value that can be recorded in these systems.

For typical earth-fault protections used in networks with a neutral point grounded by a resistor, the starting voltage value is assumed to be 3–10 V [26]. Assuming a maximum measurement signal of 100 V, this gives a relative value of 3–10%, and is therefore comparable to that obtained using capacitive probes.

The results presented above confirm the validity of the concept presented in this section. Of course, due to the complexity of the issue and its importance for the correct operation of the power grid, this concept requires a prototype solution and field tests. A particularly important problem, in a situation where we are dealing with flexible conductors exposed to wind, is the assessment of the impact of vibrations of these conductors (e.g., Aeolian vibrations or “galloping conductors”) [33] because they can temporarily change the coupling capacitance between the overhead line conductors and the probes. These issues will be the subject of further development work conducted by the authors.

5. Conclusions

This article presents a development method for detecting asymmetries and short circuits in a medium-voltage power network using a system of several capacitive probes (one, two, or three). The research conducted in laboratory conditions and simulation analyses was to answer the question of whether this type of probe system—suitably distributed in space and coupled with phase conductors—is a sufficiently sensitive tool for the unambiguous detection of a short circuit in a power network based on the criterion of assessing the zero-voltage component. The research results confirmed the proper sensitivity of the proposed method, which, taking into account the appropriate scale of the recorded signals, is comparable to the parameters obtained in standard solutions (using voltage transformers). Devices installed deep in the network that support quick assessment of disturbance states, i.e., fit into the strategy of the so-called intelligent networks (Smart Grid), are certainly the future of modern energy. Nowadays, technologies using non-contact solutions are increasingly used because they provide—thanks to galvanic separation—greater safety of the detection device operation, which is why the proposed method also fits into these contemporary trends. Obtaining positive results from the laboratory tests became the basis for conducting further tests, this time on prototype devices installed in the field. Therefore, the next stage of the work conducted by the authors will be to create a prototype device. This will allow for tests and the collection of data on the behavior of the system in a typical power network, the assessment of the impact of interference on the system operation, and the determination of criteria for the correct detection of short circuits in the network.