Equipment and procedures for monitoring the condition of cables in an electrical supply network

Abstract

A method and system for detecting the degradation of cable joints in an electrical line is described, in which the current in the shielding of the line is determined at a distance from the cable joints, current pulses from partial discharges are summed using an analog circuit, and subsequently digital processing takes place at a frequency in the kHz range, whereby the phase angle and intensity during current half-waves are determined during the digital processing.

Description

[0001] The invention relates to a device and a method for monitoring the condition of cables in an electrical supply network, in particular a medium-voltage network. In particular, the condition of the cable joints is monitored.

[0002] Sections of underground cables, for example in a medium-voltage network, are joined by cable joints, which age due to environmental influences combined with electrical stress. This aging can occur at different rates depending on the conditions under which the joint was installed and can be initiated by both degradation of the insulation between the conductor and the shield and degradation of the electrical connection between the conductors in the area of ​​the cable joint.

[0003] Degradation initially manifests itself through a temperature increase or slight partial discharges. As it progresses, partial discharges of increasing intensity develop. Above a critical intensity threshold, degradation in medium-voltage networks becomes self-reinforcing, meaning it accelerates rapidly until the cable joint is destroyed. A destroyed cable joint results in outages in the medium-voltage network.

[0004] In the future, increasing demands on distribution networks due to the integration of renewable energy generation units and electric vehicle charging infrastructure are expected to lead to a shorter lifespan for cable joints. The resulting increase in outages in medium-voltage networks will lead to growing economic losses. Retrofitting medium-voltage distribution networks with cable joint monitoring systems is therefore desirable.

[0005] It is known to monitor cable joints using sensors to detect, for example, temperature increases. A disadvantage of this is that all joints must be equipped for the use of such or similar technology, which is costly.

[0006] The object of the invention is to provide a method and a device for monitoring the condition of cables in an electrical supply network, in particular a medium-voltage network, with which the aforementioned disadvantages are reduced.

[0007] This problem is solved by a method having the features specified in claim 1. Another solution is a device having the features of claim 10.

[0008] In the inventive method for monitoring the condition of cables in an electrical supply network, particularly a medium-voltage network, the current flow in the conductor's ground shield is measured. The measured current flow signal is then filtered using an (analog) active bandpass filter, wherein the lower cutoff frequency of the bandpass filter is at least 0.5 MHz and the upper frequency of the bandpass filter is below 100 MHz. Furthermore, the bandpass-filtered measurement signal is summed or integrated using an analog circuit. The resulting summed or integrated signal is digitized at a sampling rate between 0.2 kHz and 50 kHz. At least one partial discharge measure is then determined from the digital signal.

[0009] The device according to the invention for monitoring the condition of a line in an electrical supply network, in particular a medium-voltage network, comprises a current sensor for current flowing in the line's ground shield. It further comprises an analog bandpass filter whose lower cutoff frequency is at least 0.5 MHz. The bandpass filter is configured to filter the signal from the current sensor; in other words, it is connected to the current sensor so that it uses its signal as an input signal.

[0010] The device further includes an integrator circuit designed to integrate the filtered signal. The integrator circuit is connected to the output of the bandpass filter. The device also includes an analog-to-digital converter for digitizing the integrated signal at a sampling rate between 0.2 kHz and 50 kHz, and a digital processing unit designed to determine a partial discharge measure from the digitized signal.

[0011] In other words, according to the invention, a first part of the evaluation of the current signal is performed using an analog circuit that filters and integrates it. A second part of the evaluation then takes place digitally. The data rate of the recorded digital data is approximately in the kHz range. This does not allow for the analysis of the characteristics of individual current pulses from partial discharges, which have a duration in the nanosecond range. However, simple digital components can be advantageously used, and the data transmission rate for transmitting the digital results is unproblematic for all types of data networks. In particular, this allows the use of existing data acquisition systems in 50 Hz networks.

[0012] Furthermore, the bandpass filter is advantageously designed to cover the spectrum of current pulses corresponding to partial discharges at a distance from the point of origin of the partial discharge. A typical distance in practice for cable joints from a cable branch, such as a substation, is between 0.5 km and 1.5 km. At such a distance, the spectrum of partial discharges is reduced by approximately two orders of magnitude compared to the spectrum found at the point of origin. Since the spectrum of partial discharges is found in the range between 100 MHz and several GHz, the bandpass filter is designed with a lower cutoff frequency of at least 0.5 MHz for measurements at a distance of approximately 1 km. This frequency is above typical interference from power electronic converters but below the expected spectrum of partial discharges.

[0013] Advantageous embodiments of the method and apparatus according to the invention are described in the dependent claims. The embodiment of the independent claims can be combined with the features of one of the dependent claims or, preferably, with those of several dependent claims. Accordingly, the following additional features can be provided: The upper cutoff frequency of the bandpass filter can be at least 1 MHz, and in particular at least 5 MHz. The resulting frequency band encompasses the typical frequencies of current pulses from partial discharges, which are attenuated by approximately two orders of magnitude compared to their intrinsic frequency at a distance of around 1 km. At the same time, the frequency band is no wider than absolutely necessary for the measurement and thus at least partially excludes interference from other sources in the radio frequency range.

[0014] The partial discharge measure can be determined for a period whose length corresponds to one half-cycle of the voltage applied in the line. For a 50 Hz network, this period is therefore 10 ms. The position of this period relative to the voltage half-cycle can be freely chosen. It is advantageous for the period to correspond exactly to the positive or negative half-cycle. However, shifted time periods can also be used, for example, a period between the maximum and minimum of the voltage.

[0015] Since partial discharges are known to exhibit a polarity dependency, partial discharge measures for half-waves of positive polarity (or increasing voltage) and half-waves of negative polarity (or decreasing voltage) are expediently determined separately.

[0016] It is advantageous to determine the partial discharge measure for a large number of consecutive such periods. Since the voltage profile is always the same over these periods, apart from the sign, the resulting partial discharge measures are comparable.

[0017] In this way, a large number of partial discharge measurements can be conveniently determined over a measurement period ranging from seconds to minutes or even hours. For example, a measurement period of 1 second yields 100 partial discharge measurements for the 100 half-waves. These can be processed separately according to the positive and negative half-waves, or together, to further reduce the amount of data and its rate before transmission.

[0018] For example, an average value for the partial discharge values ​​can be determined over the measurement period. It is advantageous to calculate an average value for each polarity. This average value provides a more robust result than a single value for each polarity.

[0019] Furthermore, a variation in the partial discharge values ​​can also be determined over the measurement period, for example, a standard deviation. This variation provides at least some indication of how reliable the average value is.

[0020] A partial discharge measure can be used to determine the phase of the digitized signal relative to the voltage half-cycle in the line. The phase describes the temporal behavior of the current pulses during the voltage half-cycle. Since the current pulses are capacitively coupled discharges, their occurrence correlates with the rate of change of the voltage, i.e., dU / dt. Therefore, fewer current pulses should occur at the voltage maximum and minimum than at other times if they are partial discharges of the type under consideration. If, however, the current pulses occur regardless of their position in the half-cycle, they are not the partial discharges under consideration.

[0021] Regarding phase angle, it is particularly advantageous to consider not only the average partial discharge value but also its variation. If the phase angle variation is large, the values ​​are more likely disturbances than actual partial discharges, even if the average phase angle value suggests otherwise, as if they were capacitively coupled partial discharges at a cable joint. With low variation, the phase angle is reproducible, and the values ​​represent actual charge transfer due to insulation degradation.

[0022] As a partial discharge measure, a total charge transfer for the period under consideration can be determined alternatively or additionally. Since the integrator sums the signal, this measure corresponds to the signal at the end of the period, for example, at the end of the half-wave under consideration. The total charge transfer is a measure of the severity of the insulation degradation.

[0023] The effects of load and environmental factors such as temperature and humidity on the partial discharge coefficient can be factored out. Machine learning can be used for this purpose.

[0024] A long-term increase in the intensity of partial discharges, adjusted for load and environmental influences, is usefully used as an indicator of insulation degradation. An increase that is disproportionate to the cumulative load indicates critical degradation and triggers a warning signal, which is sent, for example, to a control center or mobile monitoring devices.

[0025] The current sensor is preferably a potential-free current sensor, in particular a high-frequency current transformer or a Rogowski coil. These measure with a high bandwidth and simplify the design, especially in a medium-voltage network.

[0026] In one embodiment of the invention, a second (analog) active bandpass filter, a second integrator circuit, and a second analog-to-digital converter are provided. The cutoff frequencies of the second bandpass filter are configured such that the filtered frequency band has a non-overlapping range with that of the first bandpass filter; in particular, the two frequency bands are disjoint, and especially the lower cutoff frequency of the second bandpass filter is equal to the upper cutoff frequency of the first bandpass filter. In other words, two parallel devices are created that share the current sensor but use differently configured bandpass filters.If the signals from the two devices created in this way are compared, for example by calculating the ratio, the distance of partial discharges can be determined, since the distance causes a frequency shift and thus a different signal strength in the two devices. If the current sensor is located at a cable branch, the determined distance can be used to identify one of several affected cable joints.

[0027] The invention will now be described and explained in more detail with reference to the exemplary embodiments shown in the figures. The figures show: Fig. 1. a current pulse representing a partial discharge, in detail; Fig. 2 several current pulses representing partial discharges; Fig. 3 a section of a medium-voltage network with a cable joint and a system for detecting partial discharges; Fig. 4 a section of a medium voltage network with two cable joints and a system for detecting partial discharges.

[0028] The degradation of cable joints 111 in an exemplary medium-voltage network 200 initially manifests itself through a temperature increase or slight partial discharges. Subsequently, partial discharges of increasing intensity develop.

[0029] Fig. Figure 1 shows the course of a current pulse 100a in a time-current diagram. This pulse represents a partial discharge in a conductor 110 of the medium-voltage network 200, i.e., a current flow from the conductor 110 through its insulation to the surrounding shield 112. The current pulse 100a is only about 20 ns long. In a frequency spectrum, such short current pulses therefore manifest themselves at frequencies from 50 MHz upwards.

[0030] Over a longer period, in this case 50 µs, such partial discharges occur sporadically, as can be seen in the example shown in Fig. Figure 2 shows that, in addition to current pulse 100a, further current pulses 100b...d occur here. Also, the one in Fig. The depicted period of 50 µs is still small compared to the duration of a voltage half-cycle, which is 10 ms at a typical mains frequency of 50 Hz. Within a single voltage half-cycle, a large number of such current pulses (100 a...d) can therefore occur.

[0031] For the detection of partial discharges, i.e., current pulses 100 a...d, a [missing information] is used. Fig. The device 10 shown in Figure 3 is used. The device 10 comprises a high-frequency current transformer 11. This is arranged so that it measures the current flowing in the ground shield of line 110 without any potential difference. The signal from the current transformer 11, which is terminated to ground by a resistor 113, is used at a first node 123 as the input signal to an active bandpass filter 12 consisting of a high-pass filter 120, an inverting operational amplifier 133, and a low-pass filter 130. The high-pass filter 12 consists of a resistor 121 and a capacitor 122 connected in series between the first node 123 and the inverting input of the operational amplifier 133. The non-inverting input of the operational amplifier is connected to ground.The output of the operational amplifier is connected to the inverting input via the parallel connection of the feedback resistor 131 and the feedback capacitor 132, which thus act as a low-pass filter 130. The high-pass filter 120 and the low-pass filter 130 each define the lower and upper cutoff frequencies of the active band-pass filter, respectively.

[0032] The output of operational amplifier 133 is connected via a charging resistor 141 to one of the AC inputs of a diode bridge rectifier 143. The diode bridge rectifier 143, together with a charging resistor 141, a discharging resistor 144, and a capacitor 142, forms an integrator 14. Depending on the polarity, the charging resistor 141, two of the diodes of the bridge rectifier 143, and the capacitor 142 are connected in series to ground. The other two diodes of the bridge rectifier are reverse-biased for the respective polarity. The series connection of the charging resistor 141 and the capacitor 142 determines the time constant of the integration for sharp pulses. It is expediently set to match the lower cutoff frequency of the active bandpass filter 12.The discharge resistor 144, connected in parallel to capacitor 142, together with the capacitance of capacitor 142, determines the discharge time constant and thus the time that the voltage signal at the output of the integrator 14 is available for further processing. The output voltage of the integrator 14 is the voltage applied across the parallel circuit consisting of capacitor 142 and discharge resistor 144.

[0033] This is connected to the input of a digital processing unit 15, which includes an analog-to-digital converter 151. The analog-to-digital converter 151 is configured to sample and convert the incoming analog voltage signal at a data rate of 5 kHz. In other embodiments, other data rates can be used, expediently between 1 kHz and 20 kHz.

[0034] The analog processing circuitry, including the active bandpass filter 12 and the integrator 14, is located locally within the cable. However, it can be situated at a considerable distance from a cable joint 111. The processing unit 15, on the other hand, can be located partially far away, for example, as a cloud service or as a program on a specific PC. At least the analog-to-digital converter 151 is located locally within the analog circuitry.

[0035] The processing unit 15 is designed to perform multiple evaluations on the data received from the analog-to-digital converter 151. The evaluation is carried out separately for each individual half-wave of the voltage in line 110, i.e., at a mains frequency of 50 Hz, in blocks of 10 ms duration, the position of which corresponds to the temporal position of the voltage half-waves.

[0036] For each block, a phase angle for the partial discharges, i.e., current pulses, is determined. The phase angle describes the temporal behavior of the current pulses' occurrence during the voltage half-cycle. Since the current pulses are capacitively coupled discharges, their occurrence correlates with the rate of change of the voltage, i.e., dU / dt. Therefore, fewer current pulses should occur at the voltage maximum and minimum than at other times if they are partial discharges of the type under consideration. If, however, the current pulses occur regardless of their position in the half-cycle, they are not the partial discharges under consideration.

[0037] Furthermore, the intensity of the partial discharges, i.e., the current pulses, is determined for each block. This intensity corresponds to the summed charge transfer within the voltage half-wave and to the summed and digitized voltage signal of the analog circuit at the end of a block.

[0038] The phase relationship indicates whether the current is truly a capacitively coupled partial discharge, an incipient ground fault with resistive coupling and high charge transfer, or simply interference. With interference, the phase relationship fluctuates significantly and is therefore not reproducible, and the charge transfer also varies considerably.

[0039] Specifically, in processing unit 15, a variation in the values, for example a standard deviation, for the phase angle is determined over a multiple of current half-waves. Only if this falls below a threshold value are the signals not considered interference. A variation in the values ​​can also be determined for the intensity.

[0040] The values ​​determined in this way for the phase position and intensity are averaged over several seconds or minutes for each positive and negative voltage half-wave and stored with a resolution of seconds or minutes.

[0041] The degradation determination process is therefore twofold. An analog evaluation circuit combines the current pulses 100a...d of the partial discharges into an integrated signal. This signal is digitized at a sampling rate in the kHz range. The subsequent digital processing therefore only needs to process the values ​​in the kHz range, which simplifies the digital processing. However, a detailed spectral analysis of the current pulses 100a...d is not required.

[0042] Correlations with electrical load flow and environmental conditions such as humidity and temperature are determined using machine learning and used to isolate the degradation of cable joint 111 from load and environmental influences. The long-term increase in partial discharge intensity, adjusted for load and environmental influences, is used as an indicator of degradation.

[0043] A disproportionate increase in load compared to the cumulative load is used as an indicator of critical degradation and triggers a warning or alarm, for example, in the control room or on a mobile monitoring device. The AI-based or classical / empirical model-based evaluation algorithm can be implemented, at least partially, on the actual monitoring device.

[0044] Device 10 is connected to line 110 at a specific location via current transformer 11. When measuring the current pulses 100a...d at their point of origin, i.e., at the cable joint 111 itself, their frequencies are roughly in the range between 3 MHz and 100 MHz. The bandpass filter formed by the high-pass filter 120 and the low-pass filter 130 can therefore be designed, at close range, such that its lower cutoff frequency is 3 MHz and its upper cutoff frequency is 100 MHz.

[0045] However, the device 10 is advantageously located not at a cable joint 111, but away from it at a cable branch, for example at a substation or a local distribution transformer. It is then typically located a considerable distance from the cable joint 111, for example at a distance of 500 m to approximately 1.5 km. At such distances, the current pulses 100a...d of the partial discharges are lengthened by the influence of cable capacitance and cable inductance and are thus shifted towards lower frequencies in the frequency spectrum. The frequency shift is approximately two orders of magnitude for distances on the order of 1 km.

[0046] The bandpass filter formed by the high-pass filter 120 and the low-pass filter 130 is therefore designed at such a distance such that its lower cutoff frequency is 0.5 MHz and its upper cutoff frequency is 5 MHz. The frequencies are determined taking into account the shift caused by the distance from the location of the partial discharges, such that the lower cutoff frequency lies below the frequency components typical for partial discharges but above interference from power electronic converters, and that the upper cutoff frequency lies in the range of the frequency characteristic of partial discharges in cables.

[0047] In a second embodiment, which is described in Fig. As shown in Figure 4, two devices 41, 42 are connected in parallel to each other to the same current transformer 11. The devices 41, 42 essentially correspond to the device 10 with the difference that they share the current transformer 11 and the further difference that the cutoff frequencies for the respective bandpass filter are different.

[0048] In this embodiment, the first device 41 uses cutoff frequencies of 1 MHz and 10 MHz for the bandpass filter, while the second device 42 uses cutoff frequencies of 10 MHz and 100 MHz. During operation, the resulting signals can be used individually, as already described in the first embodiment. Furthermore, a comparison can be made between the signals resulting from devices 41 and 42; for example, a ratio can be calculated. From this ratio, it can be deduced how far away a source of current pulses, i.e., partial discharges, is from the current transformer 11.

[0049] If the first cable joint 111a, which is closer to the current transformer 11, is a source of partial discharges, then these will manifest themselves due to the proximity in a higher frequency spectrum, i.e. with a stronger signal in the frequency band between 10 MHz and 100 MHz than in the frequency band between 1 MHz and 10 MHz.

[0050] If, however, the second cable joint 111b, located further away from the current transformer 11, is a source of partial discharges, these will manifest themselves due to the greater distance in a lower frequency spectrum, i.e., with a weaker signal in the frequency band between 10 MHz and 100 MHz than in the frequency band between 1 MHz and 10 MHz. The ratio of the signal strengths of the second to the first device 41, 42 would therefore be lower in this case than the same ratio in the case of the closer cable joint 111a mentioned above. Thus, it is advantageous to detect from a central location which of the cable joints 111a, b is the source of the partial discharges.

[0051] In another embodiment, a difference or correlation measurement is performed. For this purpose, a setup is used that is similar to the setup consisting of Fig.This corresponds to 4. The second bandpass filter is designed for a narrow frequency band that is contained within the frequency band of the first bandpass filter. Improved signal quality is achieved in this narrow frequency band.

[0052] The correlation of the amplified signals in the same frequency range can be used as an evaluation criterion. This simultaneously provides a normalization measure for the amplified signals, counteracting the disproportionate effect of the amplified first signal. A change in the correlation value between the narrowband and the broadband bandpass filters can be used as an additional output value. This advantageously increases the robustness of the evaluation algorithm.

[0053] In other embodiments, several additional devices can be used. Their respective bandpass filters are then designed, like the second bandpass filter, for a narrower frequency band compared to the first bandpass filter, whereby the various narrow frequency bands do not overlap or at most only partially overlap; in other words, several narrow frequency bands are covered. Reference sign 10, 41, 42 Facility 11 Current transformer 113 Termination resistor 12 Active (analog) bandpass filter 120 high-pass filters 121 Resistance 122 Capacitor 123 knots 130 low-pass filters 131 Feedback resistor 132 Feedback capacitor 133 operational amplifiers 14 integrators 141 Charging resistor 142 Capacitor 143 Bridge rectifiers 144 Discharge resistor 15 digital processing equipment 151 Analog-to-Digital Converters 100 a...d current pulse 110 Line 111, 111a, b cable sleeve 112 Shielding

Claims

[1] Method for monitoring the condition of a line (110) of an electrical supply network, in particular a medium-voltage network, comprising the steps: - Measurement of the current flow in the ground shield (112) of the line (110), - filtering the measurement signal with an analogue bandpass filter, using at least 0.5 MHz as the lower limit frequency of the bandpass filter, - Integrating the filtered measurement signal with an analog integrator circuit (14), - Digitizing the integrated signal with a sampling rate between 0.2 kHz and 50 kHz, - Determining a partial discharge measure from the digitized signal. [2] Method according to claim 1, wherein at least 5 MHz is used as the upper limit frequency of the bandpass filter. [3] Method according to claim 1 or 2, wherein the partial discharge index is determined for a period of time whose length corresponds to a half-wave of the voltage applied in the line (110). [4] Method according to one of the preceding claims, in which a phase position of the digitized signal in comparison to the half-wave of the voltage in the line (110) is determined as the partial discharge measure. [5] Method according to one of the preceding claims, in which a total charge transfer for a period of time is determined as the partial discharge measure. [6] Method according to one of the preceding claims, in which an average value for the partial discharge index is determined over a period of time which is at least 1 s, in particular at least 1 min. [7] Method according to one of the preceding claims, in which influences of the load and environmental influences on the partial discharge index are eliminated, in particular by means of machine learning. [8] Method according to one of the preceding claims, in which a warning signal is triggered if the partial discharge index increases disproportionately compared to the cumulative load. [9] Device (10, 41, 42) for monitoring the condition of a line (110) of an electrical supply network, in particular a medium-voltage network, comprising - a current sensor (11) for current flowing in the ground shield (112) of the line (110), - an analogue bandpass filter whose lower limit frequency is at least 0.5 MHz, the bandpass filter being arranged to filter the signal from the current sensor (11), - an integrator circuit (14) arranged to integrate the filtered signal, - an analog-to-digital converter (151) for digitizing the integrated signal with a sampling rate of between 0.2 kHz and 50 kHz, - a digital processing device (15) designed to determine a partial discharge measure from the digitized signal. [10] Device (10, 41, 42) according to claim 9, wherein the current sensor (11) is a potential-free measuring current sensor (11), in particular a high-frequency current transformer (11) or a Rogowski coil. [11] Device (10, 41, 42) according to claim 9 or 10, comprising a second analog bandpass filter, a second integrator circuit and a second analog-to-digital converter, wherein the cutoff frequencies of the second bandpass filter are arranged such that the filtered frequency band has a non-overlapping range with that of the first bandpass filter, in particular the two frequency bands are disjoint, in particular the lower cutoff frequency of the second bandpass filter is equal to the upper cutoff frequency of the bandpass filter. [12] Device (10, 41, 42) according to claim 11, wherein the digital processing device (15) is arranged to determine the distance of a source of partial discharges from the current sensor (11) from the signals of the analog-digital converter (151) and the second analog-digital converter, in particular by forming a ratio of the signals of the two analog-digital converters (151).

Images