System and method for detecting insulation defects in underground power cables

Abstract

A system (2) for detecting insulation defects in an underground power cable (12) having one or more single conductors (16, 18, 20) surrounded by a conductive shield (22). The system (2) includes one or more external clamp-on sensors (4, 4', 4'', 4''') fastened to the outside of the power cable (12) or positioned in close proximity to the power cable (12). The clamp-on sensors (4, 4', 4'', 4''') are configured to provide two or more current measurements from outside the power cable (12) without being electrically connected to any of the one or more conductors (16, 18, 20) of the power cable (12). The system (2) includes a signal processing unit (50). The sensors (4, 4', 4'', 4''') are configured to detect partial discharge events (6, 6', 6'', 6''', 6''''). The signal processing unit (50) is adapted to use a mathematical statistical model (52) to process the measurements obtained by the sensors (4, 4', 4'', 4''') to identify whether the current measurements are caused by a partial discharge event (6, 6', 6'', 6''', 6'''') caused in a leakage structure (46) of the power cable (12).

Description

Technical Field

[0001] The present invention relates to a system for analyzing partial discharges and a method for performing an analysis of partial discharges. More specifically, the present invention relates to a system and method for detecting insulation defects in underground power cables.

Background Art

[0002] According to the International Electrotechnical Commission (IEC), an international standard, partial discharge is defined as follows: "A local electrical discharge that only partially bridges the insulation between conductors, which may or may not occur adjacent to a conductor."

[0003] Partial discharge occurs when impurities or voids inside the insulation of a conductor, or protrusions outside the insulation of the conductor, create a stressed region. The stressed region can be formed by sharp edges or protrusions around the conductor.

[0004] Partial discharge in a power cable includes several types of discharge phenomena, such as surface discharge occurring at the boundary of different insulating materials and internal discharge occurring in gaps or voids within a solid or liquid dielectric.

[0005] The detection and measurement of discharge are based on the energy exchange that occurs during discharge. These exchanges are represented as electrical pulse currents.

[0006] However, prior art solutions are difficult to set up and costly because they need to be electrically connected to the power cable. Therefore, when applying these solutions, it is necessary to interrupt the power cable. This has been a major unsolved problem for decades.

[0007] Prior art partial discharge analysis tools are expensive because they require costly sensors, and these sensors must be set up by being electrically connected around a conductor or single conductor. Therefore, these analysis tools are not suitable for use.

[0008] Technical staff in underground power distribution systems more frequently face the decision of whether to replace older sections of cable. Measurements of partial discharge are used to determine whether the cable should be replaced with new cable or repaired as needed.

[0009] However, studies have shown that using partial discharge measurements in power cables is difficult to definitively determine whether they should be replaced. Removing these cables based solely on partial discharge measurements is not a valid reason, because external events (e.g., lightning or network switching) as well as leakage structures in the power cable insulation can cause partial discharge events. A challenge with prior art solutions is the inability to distinguish whether a discharge event was caused by an external event (e.g., lightning) or by a leakage structure in the power cable.

[0010] U.S. Patent Application Publication No. 20090177420 discloses a device for identifying, locating, and identifying partial discharges occurring in partial discharge areas along electrical equipment. However, this device is not suitable for determining partial discharges with sufficient accuracy.

[0011] Therefore, it is desirable to have systems and methods that mitigate or even eliminate the disadvantages of prior art solutions. It is also desirable to have improved apparatus and methods for detecting, locating, and elucidating partial discharges. [Prior art documents] [Patent Documents]

[0012] [Patent Document 1] U.S. Patent Application Publication No. 20090177420 [Overview of the project] [Problems that the invention aims to solve]

[0013] Therefore, an object of the present invention is to provide a system that mitigates or even eliminates the aforementioned disadvantages of the prior art.

[0014] The object of the present invention is to provide a system and method for detecting, locating, and identifying partial discharges in multi-core cables. [Means for solving the problem]

[0015] The object of the present invention is a power quality analysis system defined in claim 1, and Claim 16 This can be achieved by the method defined in [the relevant section]. Preferred embodiments are defined in the dependent claims, described below, and illustrated in the accompanying drawings.

[0016] The system according to the present invention is for detecting insulation defects in underground power cables containing one or more conductors surrounded by a conductive shield. The system comprises one or more external clamp-on sensors that are attached to or positioned close to the power cable. These clamp-on sensors are configured to provide one or more current measurements from outside the power cable without being electrically connected to any of the conductors of the power cable. The system comprises a processing unit. The sensors are configured to detect partial discharge events. The system comprises a signal processing unit adapted to use a mathematical statistical model. This mathematical statistical model processes the measurements obtained by the sensors to identify whether the current measurement was caused by a partial discharge event due to a leakage structure in the power cable. The mathematical statistical model is constructed to produce a linear projection of the current in the conductor and shield. The mathematical statistical model is defined as follows: Y t =F t (θ t )+ε t ε t ~δ t (V t ) θ t =g t (θ t-1 )+Θ t Θ t ~δ 2 (W t ) Here Y t is the sensor (S 1 、S 2 、...、S n A vector used to determine the observed process at time t, based on data observed from ). θ t This is a vector that determines the latent probabilistic processing at time t, which includes the latent processing data, i.e., the currents arising from the cable conductor and shield, respectively; F t This is a regression matrix that determines the linear relationship between the latent process and the process observed at time t; g t This is an evolutionary matrix used in the aforementioned latent processing to determine the transition from time t-1 to time t; δ 1 and δ 2 These are the stochastic noise vectors for the observed and latent processes, respectively; V t This is the observed variance-covariance matrix; and W t This is the evolutionary variance-covariance matrix.

[0017] This makes it possible to provide a system capable of measuring partial discharge in power cables in a way that enables predictive and proactive maintenance, i.e., reliably identifying whether a power cable should be replaced. The system is configured to detect insulation defects in underground power cables containing a single conductor surrounded by a conductive shield. Both the single conductor and the surrounding conductive shield constitute the conductor. Therefore, a single conductor surrounded by a conductive shield is a two-conductor cable.

[0018] Prior art solutions are not configured to detect insulation defects in underground power cables containing at least one single conductor surrounded by a conductive shield. The challenge is that the number of equations is less than the number of unknowns. Therefore, since there are an infinite number of solutions, it is impossible to provide a single solution, and thus it is impossible to use a magnetic field detected by a sensor placed outside the power cable to determine whether a current measurement resulted from a partial discharge caused by a leakage structure in the power cable or the conductive shield.

[0019] However, the present invention applies a mathematical statistical model to identify whether the current measurement resulted from a partial discharge event caused by a leakage structure in the power cable or conductive shield. This is done by providing a sufficiently large number of equations by placing a sufficiently large number of subsensors outside or in close proximity to the outer surface of the power cable. This makes it possible to provide multiple measurements detected by several different sensors (one main sensor and several subsensors) arranged such that the relative positions of the sensors to each other are known (this can be done by placing the sensors in a predetermined mounting structure).

[0020] The mathematical model is used to estimate the transfer function as a linear projection and statistical noise element from the conductors and screen to the main sensor and sub-sensor. When it is necessary to detect insulation defects in an underground power cable containing three conductors (where the conductors are electrically insulated from each other) and a conductive shield, the currents flowing through the conductors of the power cable are represented by I1, I2, and I3, and the electric charge flowing through the conductive shield is represented by I4.

[0021] The signals B1, B2, B3, B4... measured by the main sensor and sub-sensors can be calculated as the superposition of magnetic fields generated by currents I1, I2, I3, I4, using the following superposition equation: (1)

number

[0022] Here, Bj is the j-th order signal (measured by the j-th order sensor), and d ij This can be calculated by simple projection. This calculation can be achieved by simple projection because six cylindrical coordinate systems (three distances from the center of each of the three conductors to the target sensor, and the angular position of the sensor) are known.

[0023] According to Ampere's law, the magnetic field B at a distance r from the center of a conductor carrying an electric current I is defined as follows: (2) B2πr = μ0I, or (3) B = μ0I / 2πr, where μ0 is the permeability of free space. [Example 1]

[0024] In an underground power cable with three conductors surrounded by a conductive shield, there are seven unknown parameters (given that the distances from the main and sub-sensors to the center of the power cable are known), which need to be estimated using a superposition equation from the main and sub-sensors.

[0025] If there is one main sensor and eleven sub-sensors, there are 12-7=5 degrees of freedom. In practice, the current has 3 degrees of freedom, and therefore in the set of 12 equations, there are 2 degrees of freedom. Thus, in this particular example, currents I1, I2, I3, and I4 can be estimated and distinguished by multiple linear regression with 2 degrees of freedom.

[0026] The number of sub-sensors required depends on the number of conductors.

[0027] In one embodiment, the main sensor and additional sub-sensors are equipped with coils. Thereafter, when current flows through the conductor or conductive shield of the power cable, current is induced in each of the coils.

[0028] In one embodiment, the system is configured to detect insulation defects in an underground power cable containing several conductors surrounded by a conductive shield.

[0029] The system comprises one or more external clamp-on sensors, which are either attached to the outside of the power cable or positioned close to it. In one embodiment, these sensors may be attached to the power cable using a mechanical mounting structure. In one embodiment, this mechanical mounting structure is a cable tie.

[0030] These clamp-on sensors are configured to provide one or more current measurements from outside the power cable without being electrically connected to any of the conductors of the power cable. This is a significant advantage because it allows for new improvements to sensors in existing power cables, both underground and above ground.

[0031] This system includes a signal processing unit adapted to use a mathematical and statistical model for processing sensor-obtained measurements, and identifies whether current measurements are caused by a partial discharge event triggered by a leakage current structure in the power cable. Thus, these sensors are capable of detecting partial discharge events. The system is configured to perform an analysis of partial discharge in the power cable, thereby detecting insulation defects in the power cable.

[0032] In one embodiment, the power cable comprises several single conductors.

[0033] In one embodiment, the system comprises several spaced sensors arranged along the shield of a power cable. This makes it possible to detect the location of leakage current structures in the power cable. The more sensors there are, the more accurately the location of the leakage current structures in the power cable can be detected. Leakage current structures in power cables result in partial discharge events, which can be detected by sensors within a certain distance of the leakage current structure. Sensors closest to the location of the leakage current structure will detect a higher signal than sensors located at a greater distance from the location of the leakage current structure. This makes it possible to indicate the location of the sensor located at the shortest distance from the location of the leakage current structure in the power cable. This allows for the determination of the location of the leakage current structure in the power cable.

[0034] In one embodiment, the system comprises several sensors positioned and configured to determine the location of a partial discharge event. Since the signal intensity and signal frequency / wavelength of a partial discharge event depend on the distance between the sensor and the location of the event, it is possible to determine which sensor is located closest to the event by comparing the signal intensities of the sensors.

[0035] In one embodiment, the system comprises a main sensor member and one or more additional sensor members arranged along the outer circumference of the shield of a power cable. These sensor members are spaced apart in the tangential direction.

[0036] This allows these sensors to detect partial discharge events caused by leakage current structures, even without knowing their internal location within the power cable conductor.

[0037] It is important to emphasize that the main sensor member and one or more additional sensor members may be positioned at a short distance from the power cable shield. In one embodiment, the main sensor member and one or more additional sensor members are mounted on the power cable shield. The mounting of the main sensor member and one or more additional sensor members to the power cable shield may be done by cable ties or other mechanical mounting structures.

[0038] In one embodiment, the system includes a calibration unit. This calibration unit is configured to perform calibration on one or more sensors so that the system can detect whether a detected signal originates from a partial discharge signal caused by a leakage current structure in a power cable, or from a similar signal caused by an external event such as lightning or grid switching (noise).

[0039] Calibration is performed by applying the difference in the signals between the main sensor and one or more additional sensors (sub-sensors) to distinguish the current flowing through one or more single conductors from the current flowing through the surrounding shield.

[0040] In one embodiment, the calibration procedure includes a step of training a mathematical model, where the difference between signals detected by different sensors is used as input. After training, the mathematical model can split any signal into the following: a) A first portion resulting from the current flowing through one or more conductors, and b) A second portion resulting from the current flowing through the surrounding shield.

[0041] In one embodiment, the mathematical model is a mathematical statistical model configured to be used to assume a portion of the sensor signal flowing through one or more signal conductors and a shield. To correct for the actual environment of the sensor S1, it is possible to perform a transformation of the time dependence in the measurements of the sensor S1.

[0042] Less computationally intensive approaches can be achieved by using mathematical and statistical models to estimate whether time-independent transformations can be used.

[0043] For continuous measurements from sensor S1, using a time-independent transformation may be advantageous over a complete mathematical statistical model.

[0044] Testing whether the accuracy of the conversion is acceptable and updating it using mathematical and statistical models as needed can sometimes be beneficial.

[0045] The required number of samples can be determined by estimating the measurement error of the conversion. In this way, the results can be withheld until the desired accuracy is achieved.

[0046] Below, we describe one preferred method for estimating latent stochastic processes using mathematical statistical models. Latent stochastic processes can be modeled, for example, by a situation-space model defined below: (4)Y t =F t θ t +ε t ε t ~N(0,V t ) (5)θ t =G t θ t-1 +Θ t Θt ~N(0,W t ) Here, Y t The sensors (S1, S2, ...S n A vector (e.g., explicitly or explicitly defined) that determines the observed process at time t, with data observed from ); θ t F is a vector that determines the latent probabilistic processing at time t, comprising latent processing data, i.e., currents arising from the main conductor, shield, or external electromagnetic radiation of the cable, respectively; t G is a regression matrix that determines the linear relationship between the latent process and the process observed at time t. t In latent processing, the evolutionary matrix;ε is used to determine the linear transition from time t-1 to time t. t and Θ t V is the zero-mean multivariate Gaussian distributed noise vector in the observed and latent processing, respectively. t This is the observed variance-covariance matrix; and W t This is the evolutionary variance-covariance matrix.

[0047] Model parameter matrix F t and G t This can be estimated, for example, by a Kalman filter, using prior data from a modeled system and / or a similar system, including data provided by the user and / or experts in the art. Standard statistical methods can be used to perform estimations in the process (e.g., estimation of information). This information can include, for example, estimated signals (e.g., trends) and / or forecasts of the process (e.g., predictions), as well as the relevant distributions in estimates, variability, and / or confidence intervals. These types of estimations facilitate the issuance of warnings and / or alerts. For example, an alert can be selected if the probability of an observed deviation in the process is estimated to be less than 0.1%.

[0048] The above model framework is a special case of a more general model framework: (6)Yt =f t (θ t )+ε t ε t ~δ1(V t ) (7)θ t =g t (θ t-1 )+Θ t Θ t ~δ2(W t ) Here, f t and g t is a general function, δ1 and δ2 are general statistical distributions, and all other terms are as described above.

[0049] Inferences in this more general model framework can be made, for example, by an extended Kalman filter when there is a relationship between latent and observed processes, and by a Kalman-Bouchy filter when time is defined on a continuous scale (e.g., as described).

[0050] Other time series analysis methods, as well as / or multivariate data analysis methods such as analysis of variance (ANOVA), Markov models, generalized linear models (GLMs), and multivariate Gaussian models, may similarly be used to estimate the above latent probabilistic processing and to infer the above information.

[0051] Applying additional sensors can be advantageous because it can accelerate the training process. Switching events in one or more conductors of a power cable can be used to perform calibration (throughout the training procedure).

[0052] A major advantage of the sensors used in the system and method according to the present invention is that they do not need to avoid power cables.

[0053] In one embodiment, the system includes a calibration unit configured to perform calibration of one or more sensors in order to calibrate the system against the physical arrangement of cables and the environment. This makes it possible to calibrate the sensors initially (during sensor setup).

[0054] In one embodiment, the calibration unit is separated from the sensor.

[0055] In one embodiment, the system comprises a plurality of sensors and a single, centrally located calibration unit configured to receive and process data from these sensors.

[0056] In one embodiment, the calibration unit is integrated with each of the sensors.

[0057] In one embodiment, the calibration unit is configured to calibrate the sensor as the main sensor member and several additional sensor members of the sensor move along the perimeter of a power cable.

[0058] This can be achieved by detecting several locations where local maximum sensor signal amplitudes exist. At these locations, the main sensor element and one of the additional sensor elements (sub-sensors) should be positioned. This procedure can be performed manually or by using a calibration tool that maintains the position trajectory and corresponding sensor signals.

[0059] In one embodiment, one or more external clamp-on sensors include an energy harvester.

[0060] In one embodiment, one or more external clamp-on sensors are electrically connected to the energy harvester.

[0061] In one embodiment, the energy harvester comprises a thermoelectric generator or an electric field environment harvesting device.

[0062] In one embodiment, the system: -A communication unit extending from the sensor toward the ground surface, - An antenna configured to transmit wireless signals, Equipped with, In this system, the measured values ​​obtained from the sensor are configured to be transmitted wirelessly via an antenna.

[0063] This makes it possible to detect sensor signals and transmit them wirelessly to a receiving device.

[0064] In one embodiment, the communication unit is configured to communicate wirelessly with the antenna.

[0065] In one embodiment, the system includes a wired connection between the communication unit and the antenna.

[0066] In one embodiment, the antenna is powered by an energy harvester, which is preferably a thermoelectric generator.

[0067] In one embodiment, the antenna is integrated into an antenna assembly which includes an energy harvester, which is preferably a thermoelectric generator or a solar cell.

[0068] In one embodiment, the shielding structure surrounds the sensor and the entire outer perimeter of the portion of the sensor to which the power cable extends, and the shielding structure is an electromagnetic field shield.

[0069] In one embodiment, the shielding structure is formed as a magnetic conductor shielding structure, configured to surround the sensor and the entire outer periphery of the portion of the sensor to which the power cable extends.

[0070] This makes it possible to provide electromagnetic shielding.

[0071] This makes it possible to reduce the effects of electromagnetic radiation.

[0072] In one embodiment, the processing unit includes a peak detector configured to analyze current measurements and detect arbitrary current peaks.

[0073] In one embodiment, the processing unit includes a high-pass filter and is configured to apply the current measurement to the high-pass filter.

[0074] In one embodiment, the processing unit includes an algorithm configured to automatically identify whether a current measurement was caused by a partial discharge event triggered by a leakage structure in a power cable.

[0075] The present invention provides a method for detecting insulation defects in underground power cables, which include one or more single conductors surrounded by a conductive shield. The method includes the step of attaching one or more external clamp-on sensors to the outside of the power cable or placing them close to the power cable. The power conductors of these clamp-on sensors are configured to provide one or more current measurements from outside the power cable without being electrically connected to any of the conductors of the power cable. The method includes the step of applying a signal processing unit to process the data. These sensors are configured to detect partial discharge events. The step of applying a signal processing unit to process the data is performed using a mathematical statistical model that processes the measurements made by the sensors to identify whether the current measurements were caused by a partial discharge event triggered by a leakage structure in the power cable. This method applies a mathematical statistical model used to construct linear projections of currents in conductors and shields. This mathematical statistical model is defined as follows: Y t =F t (θ t )+ε t ε t ~δ 1 (V t ) θ t =g t (θ t-1 )+Θ t Θ t ~δ 2 (W t ) Here Y t is the sensor (S 1 、S 2 、...、S n A vector used to determine the observed process at time t, based on data observed from ). θ t This is a vector that determines the latent probabilistic processing at time t, which includes the latent processing data, i.e., the currents arising from the cable conductor and shield, respectively; F t This is a regression matrix that determines the linear relationship between the latent process and the process observed at time t; g t This is an evolutionary matrix used to determine the transition from time t-1 to time t in latent processing; δ 1 and δ 2 These are the stochastic noise vectors for the observed and latent processes, respectively; V t This is the observed variance-covariance matrix; and W t This is the evolutionary variance-covariance matrix.

[0076] This makes it possible to provide a method that allows for the measurement of partial discharge in a power cable, enabling reliable identification of whether the power cable should be replaced or repaired. Thus, based solely on the partial discharge measurements obtained using this method, it is possible to determine when the power cable should be removed.

[0077] In one embodiment, the power cable comprises several single conductors.

[0078] In one embodiment, the method includes the following steps: - The step of exposing a portion of the power cable at one or more locations along the extension of the power cable; - The step of securing one or more external clamp-on sensors to the outside of the power cable or placing them close to the power cable at each location; - A step of establishing a connection between each of the external clamp-on sensors and the signal processing unit.

[0079] Thus, by using sensors that enable the detection of insulation defects in underground power cables using this method, the method can be used to implement new improvements to existing power cables.

[0080] In one embodiment, the method includes the step of calibrating a sensor.

[0081] In one embodiment, the process includes a step of calibrating the sensor using a calibration unit. This calibration unit is configured to calibrate the sensor by moving the main sensor component and several additional sensor components of the sensor along the perimeter of a power cable.

[0082] By detecting several locations where local maximum sensor signal amplitudes exist, it is possible to identify the most likely locations for positioning the main sensor component and one or more additional sensor components. Therefore, the main sensor component, or one of the additional sensor components (sub-sensors), should be positioned at these locations. This procedure can be performed manually or by using calibration tools that maintain the position trajectory and corresponding sensor signals.

[0083] In one embodiment, each sensor comprises a main sensor member and one or more additional sensor members. The method includes the step of arranging the main sensor member and the additional sensor members tangentially separated along the outer circumference of the shield of a power cable.

[0084] In one embodiment, the method includes the step of applying a sensor configured to measure both the magnetic and electric fields of a power cable.

[0085] In one embodiment, these sensors are configured to communicate wirelessly with one or more external devices.

[0086] In one embodiment, at least some of the sensors are powered by at least one energy harvesting device.

[0087] The present invention will be better understood from the detailed description given below in this specification. The accompanying drawings are provided for illustrative purposes only and do not limit the present invention. [Brief explanation of the drawing]

[0088] [Figure 1] This is a schematic diagram of the system according to the present invention, which includes several external clamp-on sensors attached to the outside of an underground power cable. [Figure 2A] This figure shows a curve representing the current as a function of time, indicating a partial discharge signal. [Figure 2B] This figure shows a curve representing the current as a function of time, where no partial discharge signal is present. [Figure 3] This is a schematic diagram of the system according to the present invention, including an external clamp-on sensor that is attached to the outside of an underground power cable. [Figure 4A] This figure shows the sensor according to the present invention. [Figure 4B] This is a diagram showing power cables. [Figure 4C] This figure shows the sensor shown in Figure 4A attached to the power cable shown in Figure 4B. [Figure 5A] This diagram shows a power cable that includes a leakage current structure. [Figure 5B] This figure shows a power cable, as indicated in Figure 5A, where the leakage current structure has been replaced by a melted area. [Figure 6A]This figure shows the sensor according to the present invention. [Figure 6B] This is a diagram showing power cables. [Figure 6C] This figure shows the sensor shown in Figure 6A, which is attached to a power cable. [Figure 7A] This diagram shows the partial discharge currents flowing through the conductor and shield, respectively. [Figure 7B] This figure shows the sensor according to the present invention. [Figure 7C] This figure shows another sensor according to the present invention. [Figure 8] This is a flowchart illustrating the method according to the present invention. [Figure 9A] This figure shows the sensor according to the present invention. [Figure 9B] This figure shows the sensor shown in Figure 9A, which is attached to a power cable having three conductors and a conductive shield. [Figure 9C] This figure shows a sensor according to the present invention, attached to the outside of a power cable. [Figure 9D] This figure shows another sensor according to the present invention, attached to the outside of a power cable. [Figure 10A] This figure shows a sensor according to the present invention, positioned on the outer surface of a power cable. [Figure 10B] This figure shows another sensor according to the present invention, positioned on the outer surface of a power cable. [Figure 10C] This figure shows a sensor according to the present invention, arranged on the outer surface of a multi-core power cable. [Figure 10D] This figure shows how the method according to the present invention can be used so that a conductor is treated as a single conductor. [Modes for carrying out the invention]

[0089] Next, with detailed reference to drawings intended to illustrate preferred embodiments of the present invention, a schematic of the system according to the present invention is shown in Figure 1. The system comprises several external clamp-on sensors 4, 4', 4'', 4'''', 4'''' which are attached to the outside of the underground power cable 12.

[0090] Sensors 4, 4', 4'', 4'''', and 4'''' are spaced apart from each other. Therefore, sensors 4, 4', 4'', 4'''', and 4'''' will capture signals at different locations along the power cable 12.

[0091] Below each sensor 4, 4', 4'', 4''', 4'''', the detected sensor signals 6, 6', 6'', 6''', 6'''', processed using a signal processing unit, are plotted against time. The sensor signals 6, 6', 6'', 6''', 6'''' are processed through a processing procedure performed using the signal processing unit to remove the main signal (e.g., a 50 Hz AC current) and partial discharge events caused by switching events. This processing procedure may include one or more filtering steps.

[0092] Sensors 4, 4', 4'', 4'''', and 4'''' are configured to detect current. Therefore, these plots represent current versus time. It can be seen that sensors 4'' and 4'''', located at the shortest distance from the partial discharge event 10 (represented by the thick arrows), detect a larger signal than the remaining sensors 4, 4', and 4'''', which are located at a longer distance from the partial discharge event 10.

[0093] In this way, the location of the partial discharge event 10 can be identified using sensors 4, 4', 4'', 4''', and 4''''. The location of the partial discharge event 10 can also be identified by comparing the amplitudes of the processed sensor signals 6, 6', 6'', 6''', and 6''''.

[0094] Figure 2A illustrates a curve 54 representing current as a function of time. A partial discharge signal 6 caused by a leakage current structure in an underground power cable is shown. The partial discharge signal 6 is measured by a sensor schematically illustrated in Figure 1. The frequency of the partial discharge signal 6 is significantly higher than the “main signal” (e.g., a 50 Hz AC current). Thus, the partial discharge signal 6 is shown as a peak.

[0095] Figure 2B illustrates curve 54', which represents the current as a function of time, in the absence of a partial discharge signal. Thus, curve 54' corresponds to the modified version of curve 54, where the partial discharge signal 6 has been removed.

[0096] Figure 3 illustrates a schematic of System 2 according to the present invention. System 2 includes an external clamp-on sensor 4 that is attached to the outside of the underground power cable 12. System 2 is configured to detect insulation defects in the power cable 12.

[0097] The power cable 12 comprises three conductors 16, 18, and 20 surrounded by a conductive shield 22. Although not shown, system 2 also comprises several external clamp-on sensors 4, which may be attached to the outside of the power cable 12 or positioned very close to it.

[0098] The clamp-on sensor 4 is configured to provide one or more current measurements from outside the power cable 12 without being electrically connected to any of the conductors 16, 18, 20 of the power cable 12. The sensor 4 is configured to detect partial discharge events. This is done by measuring the magnetic field generated by the current flowing through the conductors 16, 18, 20 and the conductive shield 22. Since each sensor 4 can only measure the superimposed magnetic field (the sum of the magnetic fields generated by the currents flowing through the conductors 16, 18, 20 and the conductive shield 22), additional information needs to be provided.

[0099] System 2 includes a signal processing unit 50, which is configured to provide any additional information as needed. The signal processing unit 50 is adapted to use a mathematical statistical model 52. The mathematical statistical model 52 is configured to process the measurements obtained by the sensor 4 of System 2 and to identify the origin of the magnetic field measured by the sensor 4. This means that the signal processing unit 50 can divide the measured signal into the following: a) Currents flowing through conductors 16, 18, and 20, and b) Current flowing through shield 22.

[0100] Thus, the signal processing unit 50 can identify whether the current measurement was caused by a partial discharge event triggered by the leakage structure of the power cable 12.

[0101] The power cable is connected to a connection assembly located at the power plant 8. The shield 22 is confirmed to be grounded. When lightning 14 strikes the power plant 8, it causes a partial discharge event. Thus, the current from the partial discharge event flows through conductors 16, 18, and 20. Since the shield 22 is conductive and surrounds conductors 16, 18, and 20, the current is induced in the shield 22. However, due to capacitive coupling between conductors 16, 18, and 20 and the surrounding shield 22, the induced current generated by the partial discharge current (caused by lightning) is delayed by 90° (or 1 / 4 wavelength). Thus, by comparing the current from the partial discharge event in the shield 22 with the current from the partial discharge event in conductors 16, 18, and 20, it is possible to determine whether the current from the partial discharge event was caused by a leakage current structure in the power cable 12.

[0102] System 2 can identify whether the current measurement was caused by a partial discharge event triggered by the leakage structure of the power cable 12. The system comprises an antenna 28 and a communication unit 24 positioned to transmit signals detected by sensor 4 to antenna 28. The communication unit 24 may be configured to transmit signals to antenna 28 via a wired connection or a wireless connection.

[0103] Antenna 28 transmits the wireless signal 30 received by the signal processing unit 50. The wireless signal 30 is transmitted via the internet 26.

[0104] In one embodiment, the sensor 4 and / or communication unit 24 includes, or is electrically connected to, an energy harvester, which is positioned and configured to collect energy, thereby supplying electrical energy to the sensor 4 and / or communication unit 24. In one embodiment, the energy harvester includes a thermoelectric generator or an electric field harvesting device. In one embodiment, the energy harvester includes a solar panel.

[0105] Figure 4A illustrates a sensor 4 according to the present invention. Sensor 4 comprises a main sensor member 34 and several additional sensor members (subsensors) 36, 36', 36'', 36'''. The main sensor member 34 is electrically connected to the additional sensor members 36, 36', 36'', 36''' by an electrical connector 38.

[0106] Each of the main sensor member 34 and the additional sensor members 36, 36', 36'', 36''' is configured to detect a magnetic field generated by the current flowing through the causative structure. Having several sensor members 34, 36, 36', 36'', 36''' makes it possible to position the sensor members 34, 36, 36', 36'', 36''' at different tangential positions around the power cable 12, as shown in Figure 4C. This allows the data from the sensor members 34, 36, 36', 36'', 36''' to be processed (by a signal processing unit 50, for example, as described with reference to Figure 3) so that the measured signal can be separated into the current flowing through one or more conductors of the power cable and the current flowing through the shield of the power cable.

[0107] In one embodiment, the main sensor member 34 includes an integrated communication unit (not shown). In one embodiment, the main sensor member 34 is configured to receive signals from additional sensor members 36, 36', 36'', 36''' and to transmit the signals measured by the main sensor member 34, as well as the signals from the additional sensor members 36, 36', 36'', 36'''', to a receiving device via either a wired or wireless connection.

[0108] Figure 4B illustrates a power cable 12 according to the present invention. The power cable 12 corresponds to the type of power cable 12 shown and described with reference to Figure 5A.

[0109] Figure 4C shows the sensor 4 shown in Figure 4A mounted on the power cable 12 shown in Figure 4B. The sensor 4 can be mounted on the power cable 12 using any preferred mounting structure. In one embodiment, the sensor 4 is mounted on the power cable 12 by one or more cable ties (not shown). The shielding structure 60 surrounds the sensor 4 and the entire outer circumference of the portion of the power cable 12 on which the sensor 4 extends. The shielding structure 60 is an electric field shield.

[0110] The shielding structure 60 is applied to electrically isolate the sensor 4 and the power cable 12 from the environment in which the cable 12 extends. In one embodiment, the shielding structure 60 is a conductive housing used to block electrostatic fields.

[0111] Figure 5A illustrates an underground power cable 12 located below the ground surface 32. The power cable 12 comprises three single conductors 16, 18, and 20. Each of the three single conductors 16, 18, and 20 is surrounded by an insulator 42. The insulators 42 of the three single conductors 16, 18, and 20 are surrounded by an insulating structure 44. The insulating structure 44 is surrounded by a filler 40. The filler 40 is surrounded by a metal lead shield 22.

[0112] It can be seen that the power cable 12 is damaged and includes a leakage structure 46 (enlarged for illustrative purposes). The leakage structure 46 extends through the shield 22. Thus, water and moisture can enter the cable 12 through the leakage structure 46. The leakage structure 46 extends through the packing 40, the insulating structure 44, and the insulating material 42. Thus, when water or moisture enters the leakage structure 46, the electrical impedance along the path of the leakage structure 46 is reduced to the extent that a short circuit (current traveling along the unexpected path of the leakage structure 46) is created.

[0113] During a short circuit, the current flow is at its maximum, causing the temperature inside the leakage structure 46 to rise. Thus, a short circuit generally results in the melting of the leakage structure 46, thereby creating a molten region 48 as shown in Figure 5B. The molten region 48 can seal the damaged structure (leakage structure 46) of the power cable for at least a short time. Thus, the power cable 12 may undergo several short circuits, as illustrated with reference to Figure 5A, before it becomes necessary to replace the power cable 12.

[0114] Figure 6A illustrates a sensor 4 according to the present invention. The sensor 4 comprises a main sensor member 34 and a single additional sensor member (sub-sensor) 36. The main sensor member 34 is electrically connected to the additional sensor member 36 via an electrical connector 38.

[0115] The main sensor member 34 and the additional sensor member 36 are configured to detect a magnetic field generated by the current flowing through the causative structure. Having several sensor members 34, 36 makes it possible to position the sensor members 34, 36 at different tangential positions around the power cable 12, as shown in Figure 4C. This makes it possible to process the data from the sensor members 34, 36 (for example, by the signal processing unit 50 as described with reference to Figure 3), thereby separating the measured signal into the current flowing through one or more conductors of the power cable and the current flowing through the shield of the power cable.

[0116] In one embodiment, the main sensor member 34 includes an integrated communication unit (not shown). In one embodiment, the main sensor member 34 is configured to receive signals from an additional sensor member 36 and to transmit signals measured by the main sensor member 34, as well as signals from the additional sensor member 36, to a receiving device via either a wired or wireless connection.

[0117] Figure 6B illustrates a power cable 12 according to the present invention. The power cable 12 comprises a single conductor 16 surrounded by an insulator 42. The insulator 42 is surrounded by an insulating structure 44. A filler 40 is provided between the insulating structure 44 and the surrounding metal lead shield 22.

[0118] Figure 6C shows the sensor 4 shown in Figure 6A mounted on the power cable 12 shown in Figure 6B. The sensor 4 can be mounted on the power cable 12 using any preferred mounting structure. In one embodiment, the sensor 4 is mounted on the power cable 12 by one or more cable ties (not shown).

[0119] Figure 7A illustrates a first curve 56 representing the current flowing through the conductor of a power cable during a partial discharge event versus time, as shown in Figure 5A, and a sensor corresponding to that shown in Figure 4C or Figure 6C is used to detect the current. Figure 7 further illustrates a dotted second curve 58 representing the current flowing through the shield of a power cable during a partial discharge event versus time.

[0120] The first curve 56 has an amplitude of A2, which is greater than the amplitude A1 of the second curve 58. Furthermore, the second curve is delayed.

[0121] Since the shield is conductive and surrounds the conductor of the power cable (see Figure 5A), current is induced in the shield as it flows through the conductor. Due to capacitive coupling between the conductor and the surrounding shield, the current induced in the shield is delayed by 90° (or 1 / 4 of the wavelength λ). Thus, the second curve 58 (showing the induced current in the shield caused by the current of an external partial discharge event) is delayed by 90° compared to the first curve 56. Therefore, by comparing the first solid curve 56 and the second dotted curve 58, the state of the current in the partial discharge event reveals whether the current in the partial discharge event was caused by the leakage structure of the power cable. Curves 56 and 58 shown in Figure 7A clearly show that the partial discharge event occurred in the conductor of the power cable.

[0122] Figure 7B illustrates a sensor 4 according to the present invention. Sensor 4 is essentially equivalent to the sensor shown in Figure 6A. However, the main sensor member 34 includes a signal processing unit 50. The signal processing unit 50 applies an algorithm that uses a mathematical model 52 according to the present invention. In another embodiment, the signal processing unit 50 and the mathematical model 52 may be located in an additional sensor member 36, or in a separate device (not shown) of sensor 4.

[0123] Figure 7C shows another sensor 4' according to the present invention. Sensor 4 is essentially equivalent to the sensor shown in Figure 4A. However, the main sensor member 34 includes a signal processing unit 50. The signal processing unit 50 applies an algorithm that uses a mathematical model 52 according to the present invention. In another embodiment, the signal processing unit 50 and the mathematical model 52 may be located in additional sensor members 36, 36', 36'', or in a separate device (not shown) of sensor 4'.

[0124] Figure 8 is a flowchart illustrating a method according to the present invention. The first step of this method is to detect signals measured by sensors. These sensors will provide different data because they are located at different positions.

[0125] The data detected by the sensor is processed. The processing procedure includes filtering using a high-pass filter. The processing procedure also includes signal amplification.

[0126] In the next step, it is determined whether the detected signal was caused by a partial discharge triggered by a damaged (leakage-prone) power cable. This can be done using a processing unit, as illustrated with reference to Figure 3.

[0127] In one embodiment, these steps are repeated multiple times to continuously monitor the power cable.

[0128] In one embodiment, an additional step is performed, which includes determining the location of a partial discharge event. This makes it possible to apply the method according to the present invention to indicate the location of a damaged portion of the power cable.

[0129] Figure 9A illustrates a sensor 4 according to the present invention. Sensor 4 comprises a main sensor member 34 and a plurality of additional sensor members 35, 35', 35'', 35''', 36, 36', 36'', 36''', 37, 37'. The main sensor member 34 and the additional sensor members 35, 35', 35'', 35''', 36, 36', 36'', 36''', 37, 37' are attached to a mounting structure 62. Therefore, the relative positions of the main sensor member 34 and the additional sensor members 35, 35', 35'', 35''', 36, 36', 36'', 36''', 37, 37' are clearly defined and known. A mathematical model is used to estimate the transfer function, as a linear projection and statistical noise element, from the conductor and screen to the main sensor 34 and sub-sensors 35, 35', 35'', 35''', 36, 36', 36'', 36''', 37, 37'. The shape of the mounting structure 62 may be arc-shaped as shown in Figure 9A. However, the shape of the mounting structure 62 may be different, as long as the relative positions of the main sensor member 34 and the additional sensor members 35, 35', 35'', 35''', 36, 36', 36'', 36''', 37, 37' are clearly defined and known. The shape of the mounting structure 62 may be straight, for example, so that the main sensor member 34 and the additional sensor members 35, 35', 35'', 35''', 36, 36', 36'', 36''', 37, 37' are arranged along a straight line.

[0130] Figure 9B illustrates the sensor 4 shown in Figure 9A, mounted on a power cable 12 having three conductors 16, 18, and 20 and a conductive shield 22. The conductors 16, 18, and 20 are electrically insulated from each other and surrounded by the conductive shield 22, which is spaced apart from the conductors 16, 18, and 20. A shield structure 60, formed as an electric field shield, surrounds the power cable 12 and the sensor 4.

[0131] Since the main sensor member 34 and the additional sensor members 35, 35', 35'', 35''', 36, 36', 36'', 36''', 37, and 37' extend along the outer surface of the power cable 12, the distances from the center of the power cable 12 to the main sensor member 34 and the additional sensor members 35, 35', 35'', 35''', 36, 36', 36'', 36''', 37, and 37' are known. To identify whether the current measurement resulted from a partial discharge event caused by a leakage structure in the power cable 12 or the conductive shield 22, the present invention applies the measurements obtained by the main sensor member 34 and the additional sensor members 35, 35', 35'', 35''', 36, 36', 36'', 36''', 37, 37', as well as known parameters (the relative positions of the main sensor member 34 and the additional sensor members 35, 35', 35'', 35''', 36, 36', 36'', 36''', 37, 37', as well as the diameter of the power cable 12). By arranging the main sensor member 34 and the additional sensor members 35, 35', 35'', 35''', 36, 36', 36'', 36''', 37, 37' in a predetermined mounting structure 62, the relative positions of the main sensor member 34 and the additional sensor members 35, 35', 35'', 35''', 36, 36', 36'', 36''', 37, 37' in relative positions become known.

[0132] By providing a sufficiently large number of additional sensor members 35, 35', 35'', 35''', 36, 36', 36'', 36''', 37, 37', it is possible to provide a sufficient number of measurements and therefore a sufficient number of equations, solve simultaneous equations, and thereby determine the current flowing through the conductor and the current flowing through the conductive shield 22.

[0133] Figure 9C illustrates a sensor 4 according to the present invention, attached to the outside of a power cable comprising three conductors 16, 18, and 20 and a conductive shield 22. The conductors 16, 18, and 20 are electrically insulated from each other and surrounded by a conductive shield 22 that is spaced apart from the conductors 16, 18, and 20.

[0134] Sensor 4 comprises a main sensor member 34 and several additional sensor members 35, 35', 35'', 35'''', 36, 36', 36'', 36'''', 37, and 37', which are arranged relative to each other in a predetermined manner and attached to the power cable. Thus, the distances from the center of the power cable to the main sensor member 34 and the additional sensor members 35, 35', 35'', 35''', 36, 36', 36'', 36'''', 37, and 37' are known. In this way, using the same principle described with reference to Figure 9B, it is possible to detect whether the current measurement resulted from a partial discharge event caused by a leakage structure in the power cable or the conductive shield.

[0135] Figure 9D illustrates another sensor 4 according to the present invention, which is attached to the outside of a power cable and corresponds to the one shown in Figure 9C. Sensor 4 comprises a main sensor member 34 and a number of additional sensor members 35, 35', 35'', 35'''', 36, 36', 36'', 36'''', 37, 37', 37'', arranged along the outer surface of the power cable. Thus, the main sensor member 34 and the additional sensor members 35, 35', 35'', 35''', 36, 36', 36'', 36'''', 37, 37', 37'' are positioned relative to each other in a predetermined manner. As a result, the distances from the center of the power cable to the main sensor member 34 and the additional sensor members 35, 35', 35'', 35''', 36, 36', 36'', 36'''', 37, 37'' are known. Thus, using the same principle described with reference to Figures 9B and 9C, it is possible to detect whether the current measurement originated from a partial discharge event caused by a leakage structure in the power cable or conductive shield.

[0136] Figure 10A illustrates a sensor 4 according to the present invention, positioned on the outer surface of a power cable comprising three conductors and a conductive shield. These conductors are electrically insulated from each other and surrounded by a conductive shield that is spaced apart from the conductors.

[0137] Sensor 4 comprises a main sensor member 34 and several additional sensor members 35, 35', 35'', 35'''', 36, 36', 36'', which are mounted on the power cable in a predetermined manner (along the cylindrical surface of the power cable). The main sensor member 34 and the additional sensor members 35, 35', 35'', 36, 36', 36'' are positioned along a plane extending perpendicular to the longitudinal axis X of the power cable. Therefore, the distances from the center of the power cable to the main sensor member 34 and the additional sensor members 35, 35', 35'', 35'''', 36, 36', 36'' are known.

[0138] Therefore, using the same principle described with reference to Figures 9B and 9D, it is possible to detect whether the current measurement originated from a partial discharge event caused by a leakage structure in the power cable or conductive shield.

[0139] Figure 10B illustrates another sensor 4 according to the present invention, positioned on the outer surface of a power cable having three conductors and a conductive shield. This power cable corresponds to the one shown in Figure 10A.

[0140] Sensor 4 comprises a main sensor member 34 and several additional sensor members 35, 35', 35'', 35'''', 36, 36', 36'', which are arranged in a predetermined manner along the surface of the power cable and attached to the power cable. The main sensor member 34 and the additional sensor members 35, 35', 35'', 35''', 36, 36', 36'' are arranged along a plane angled at an angle α (other than 90°) with respect to the longitudinal axis X of the power cable. If the angle α is unknown, the method according to the present invention can be used to identify whether the current measurement resulted from a partial discharge event caused by a leakage structure in the power cable or the conductive shield. The distance from the center of the power cable to the main sensor member 34 and the additional sensor members 35, 35', 35''', 36, 36', 36'' is known. Therefore, using the same principle described with reference to Figures 9B and 9D, it is possible to detect whether the current measurement originated from a partial discharge event caused by a leakage structure in the power cable or conductive shield.

[0141] Figure 10C illustrates another sensor 4 according to the present invention, which is located on the outer surface of a multi-core power cable, corresponding to those shown in Figures 9C and 9D. Sensor 4 is essentially corresponding to those shown in Figure 9D.

[0142] Figure 10D illustrates how the method according to the present invention is used so that the conductors of a multi-core cable, as shown in Figure 10C, can be treated as a single estimated conductor 21. This is possible because there is zero current flow through the conductors of the multi-core cable. [Explanation of Symbols]

[0143] 2 (Partial Discharge Analysis) System 4, 4', 4'', 4'''' (partial discharge) sensors 6, 6', 6'', 6''', 6'''' partial discharge signal 8 Power Plants 10. Partial discharge events 12 Power cables (e.g., multi-core) 14 Lightning 16, 18, 20 Single conductor of power cable 21 Estimated Conductors 22 Metal lead shields 24 Communication Units 26. Cloud (Servers accessed via the Internet) 28 Antennas 30 Transmitted signal 32 Ground surface 34 Main sensor component 35, 35', 35'', 35'''' Additional sensor components 36, 36', 36'', 36'''' Additional sensor components 37, 37', 37'' Additional sensor components 38 connectors 40 fillings 42 Insulators 44 Insulation structure 46 Earth leakage structure 48 Dissolution area 50 processing units 52 Mathematical Models 54, 54' curve 56, 58 curve 60 Shield structure 62 Mounting structure A1, A2 amplitude α angle λ wavelength B Magnetic field line X Longitudinal axis

Claims

1. A system (2) for detecting insulation defects in an underground power cable (12) comprising one or more single conductors (16, 18, 20) surrounded by a conductive shield (22), wherein the system (2) comprises two or more external clamp-on sensors (subsensors) (4, 4', 4'', 4''') that are attached to or positioned close to the underground power cable (12), and the external clamp-on sensors (4, 4', 4'', 4''') are not electrically connected to any of the one or more single conductors (16, 18, 20) of the underground power cable (12) and can be detected from outside the underground power cable (12) or configured to provide multiple current measurements, the system (2) comprises a signal processing unit (50), the external clamp-on sensors (4, 4', 4'', 4'''') are configured to detect partial discharge events (6, 6', 6'', 6'''', 6''''), the signal processing unit (50) is adapted to use a mathematical statistical model (52), the mathematical statistical model (52) processes the measurements obtained by the external clamp-on sensors (4, 4', 4'', 4'''') to identify whether the current measurements were caused by partial discharge events (6, 6', 6'', 6'''', 6'''') triggered in the leakage structure (46) in the underground power cable (12), Here, the system (2) is configured such that the mathematical statistical model (52) produces a linear projection of the current in the single conductors (16, 18, 20) and the conductive shield (22), and the mathematical statistical model (52) is Yt=Ft(θt)+εt εt~δ1(Vt) θt=gt(θt-1)+Θt Θt~δ2(Wt) It is characterized by being defined as follows: Here, Yt is a vector that determines the observed processing at time t, comprising data observed from the external clamp-on sensors (S1, S2, ..., Sn). θt is a vector for determining the latent probabilistic processing at time t, which comprises latent processing data, i.e., currents generated from the single conductors (16, 18, 20) and the conductive shield (22) of the underground power cable (12). Ft is a regression matrix that determines the linear relationship between the latent probabilistic processing and the processing observed at time t. gt is an evolutionary matrix used in the aforementioned latent probability processing to determine the transition from time t-1 to time t. δ1 and δ2 are the stochastic noise vectors for the observed and latent processes, respectively. Vt is the observational variance-covariance matrix, Wt is the evolutionary variance-covariance matrix. System (2).

2. The system (2) according to claim 1, characterized in that the underground power cable (12) comprises several single conductors (16, 18, 20).

3. The system (2) according to claim 1 or 2, further comprising several spaced sensors (4, 4', 4'', 4'''') arranged along the conductive shield (22) of the underground power cable (12).

4. The system (2) according to any one of claims 1 to 3, comprising one main sensor member (34) and one or more additional sensor members (36, 36', 36'', 36''') arranged along the outer circumference of the conductive shield (22) of the underground power cable (12), wherein the main sensor member and the additional sensor members (34, 36, 36', 36'', 36''') are spaced apart in the tangential direction.

5. The system (2) according to any one of claims 1 to 4, further comprising a calibration unit (54) configured to perform calibration of one or more of the external clamp-on sensors (4, 4', 4'', 4'''') in order to calibrate the system (2) with respect to the physical arrangement of cables and the environment.

6. The system (2) according to claim 5, characterized in that the calibration unit (54) is separated from the external clamp-on sensors (4, 4', 4'', 4'''').

7. The system (2) according to claim 5, characterized in that the calibration unit (54) is integrated with each of the external clamp-on sensors (4, 4', 4'', 4'''').

8. The system (2) according to claim 5, characterized in that the calibration unit (54) is configured to calibrate the external clamp-on sensors (4, 4', 4'', 4''') as the main sensor member (34) and several additional sensor members (36, 36', 36'', 36''') of the external clamp-on sensors (4, 4', 4'', 4''') move along the perimeter of the underground power cable (12).

9. The system (2) according to any one of claims 1 to 8, characterized in that one or more of the external clamp-on sensors (4, 4', 4'', 4''') are equipped with an energy harvester.

10. The system (2) according to claim 9, characterized in that the environment generator comprises a thermoelectric generator or an electric field environment power generation device.

11. A communication unit (24) extends from the sensor (4) toward the ground surface, The system (2) according to any one of claims 1 to 10, comprising an antenna (28) configured to transmit a wireless signal (30), wherein the measurement values ​​obtained by the sensor (4) are wirelessly transmitted by the antenna (30).

12. The system (2) according to any one of claims 1 to 11, characterized in that the shield structure (60) surrounds the sensors (4, 4', 4'', 4''') and the entire outer circumference of the portion of the underground power cable (12) to which the sensors (4, 4', 4'', 4'''') extend, and the shield structure (60) is an electromagnetic field shield (60).

13. The system (2) according to any one of claims 1 to 12, wherein the signal processing unit (50) is equipped with a peak detector, and the peak detector is configured to analyze the current measurement value and to detect an arbitrary current peak.

14. The system (2) according to any one of claims 1 to 13, characterized in that the signal processing unit (50) is equipped with a high-pass filter and is configured to apply the current measurement value to the high-pass filter.

15. The system (2) according to any one of claims 1 to 14, wherein the signal processing unit (50) comprises an algorithm configured to automatically identify whether the current measurement was caused by a partial discharge event (6, 6', 6'', 6'''', 6'''') triggered in the leakage structure (46) of the power cable (12).

16. A method for detecting an insulation defect in an underground power cable (12) having one or more single conductors (16, 18, 20) surrounded by a conductive shield (22), the method comprising the steps of securing two or more external clamp-on sensors (4, 4', 4'', 4'''') outside the underground power cable (12) or positioning them close to the underground power cable (12), wherein the power conductors of the external clamp-on sensors are configured to provide one or more current measurements from outside the underground power cable (12) without being electrically connected to any of the one or more conductors (16, 18, 20) in the underground power cable (12), and the method is The step of providing a signal processing unit (50) for processing the data is performed using a mathematical statistical model (52), which processes the measurements obtained from the external clamp-on sensors (4, 4', 4'', 4''') to identify whether the current measurements were caused by partial discharge events (6, 6', 6'', 6'''', 6'''') triggered in the leakage structure (46) of the underground power cable (12), Here, the method is configured such that the mathematical statistical model (52) produces a linear projection of the current in the conductors (16, 18, 20) and the conductive shield (22), and the mathematical statistical model (52) is Yt=Ft(θt)+εt εt~δ1(Vt) θt=gt(θt-1)+Θt Θt~δ2(Wt) It is characterized by being defined as follows: Here, Yt is a vector that determines the observed processing at time t, comprising data observed from the external clamp-on sensors (S1, S2, ..., Sn). θt is a vector that determines the latent probabilistic processing at time t, which includes latent processing data, i.e., currents generated from the conductor (12) and shield (22) of the cable (12), respectively. Ft is a regression matrix that determines the linear relationship between the latent probabilistic processing and the processing observed at time t. gt is an evolutionary matrix used in the aforementioned latent probability processing to determine the transition from time t-1 to time t. δ1 and δ2 are the stochastic noise vectors for the observed and latent processes, respectively. Vt is the observational variance-covariance matrix, and Wt is the evolutionary variance-covariance matrix. method.

17. The method according to claim 16, characterized in that the underground power cable (12) comprises several single conductors (16, 18, 20).

18. The steps of exposing a portion of the underground power cable (12) at one or more locations along the extension of the underground power cable (12), The steps include: securing one or more external clamp-on sensors (4, 4', 4'', 4''') at each location outside the underground power cable (12) or positioning them close to the underground power cable (12); The steps include establishing a connection between each of the external clamp-on sensors (4, 4', 4'', 4'''') and the signal processing unit (50), The method according to claim 16 or 17, characterized by including the following:

19. The method according to any one of claims 16 to 18, characterized by comprising the step of calibrating the external clamp-on sensors (4, 4', 4'', 4'''').

20. The method according to any one of claims 16 to 19, wherein each sensor (4, 4', 4'', 4'''') comprises a main sensor member (34) and one or more additional sensor members (36, 36', 36'', 36''''), and the method includes the step of arranging the main sensor member (34) and the additional sensor members (36, 36', 36'', 36'''') tangentially separated along the outer circumference of the shield (22) of the underground power cable (12).

Images