Underground high-voltage cable fault accurate positioning method and system

Abstract

The invention provides an underground high-voltage cable fault accurate positioning method which comprises the following steps: when an underground high-voltage cable has a fault, measuring a three-phase sheath current signal from a sheath grounding terminal in real time and linearly superposing the three-phase sheath current signal into a composite signal; determining the arrival time of the coaxial wave as a base point, and carrying out time window sliding cutting on the synthesized signal to obtain a plurality of pieces of time sequence data which contain the arrival time of the coaxial wave and are fixed in length; after normalized wavelet decomposition is carried out on the multiple pieces of time sequence data, the multiple pieces of time sequence data are imported into a trained coaxial wave arrival recognition model, and multiple pieces of wavefront recognition time are obtained; and screening out two wavefront identification times meeting a predetermined condition from the plurality of wavefront identification times, and positioning the position from the fault point to the sheath grounding terminal in combination with the total length of the underground high-voltage cable and a predetermined coaxial wave propagation speed. According to the invention, the fault of the underground high-voltage cable can be quickly and accurately positioned on line based on the cable monitoring data with transient characteristics, so that the positioning accuracy is improved.

Description

Technical Field

[0001] This invention relates to the field of power cable technology, and in particular to a method and system for accurately locating faults in underground high-voltage cables. Background Technology

[0002] Cables are widely used in underground power transmission and distribution systems in urban and mining areas. To withstand high-voltage stress, cables are typically equipped with several layers (such as sheaths and armor layers) and use XLPE to cover the core conductor for good insulation. However, after many years of use, cables are prone to breakdown due to high voltage caused by insulation deterioration. Statistical observations show that when cables exceed 60% of their design life, the failure rate increases from 1% to 10%. Once a fault occurs in an underground cable, if it cannot be located and cleared in a timely manner, it will result in prolonged power outages and significant economic losses. Therefore, accurate and rapid fault location is of great significance and presents a significant challenge for underground high-voltage cables.

[0003] Currently, existing methods for locating faults in underground high-voltage cables are divided into offline and online methods. Offline methods involve injecting signals into the cable after a fault occurs to acquire data and locate the fault. However, this method requires manual operation on-site and segmented injection and analysis, often taking several hours, making it time-consuming, labor-intensive, and prone to operational errors, resulting in poor location accuracy. Online methods use sensors pre-installed on the cable to continuously monitor the cable and locate the fault. This method is more accurate in analyzing steady-state characteristic data; however, considering that the data obtained in actual operation is continuous transient signal data under permanent faults, analyzing transient characteristics leads to a decrease in location accuracy.

[0004] Therefore, in order to overcome the above-mentioned problems of existing online fault location methods for underground high-voltage cables, a new perspective is needed to study the transient characteristics of cables, so as to achieve the goal of accurate fault location for underground high-voltage cables. Summary of the Invention

[0005] The technical problem to be solved by the embodiments of the present invention is to provide a method and system for accurate location of underground high-voltage cable faults, which can quickly and accurately locate underground high-voltage cable faults online based on cable monitoring data with transient characteristics, thereby improving the location accuracy.

[0006] To address the aforementioned technical problems, this invention provides a method for accurately locating faults in underground high-voltage cables, the method comprising the following steps:

[0007] S1. When a fault occurs in an underground high-voltage cable, the three-phase sheath current signal is measured in real time from the sheath grounding terminal at one end of the underground high-voltage cable and linearly superimposed into a composite signal.

[0008] S2. Determine the arrival time of the coaxial wave, and construct multiple time windows of equal length and different time ranges based on the arrival time of the coaxial wave. Based on the constructed multiple time windows, perform feature extraction on the synthesized signal to obtain multiple time series data containing the arrival time of the coaxial wave and with fixed length; wherein, the arrival time of the coaxial wave is used to indicate the time difference between two consecutive acquisitions of the fault traveling wave signal at the sheath grounding terminal after it is generated at the fault point of the underground high-voltage cable.

[0009] S3. Normalize wavelet decomposition is performed on all the time series data, and the data are respectively imported into the pre-trained coaxial wave arrival recognition model to obtain the corresponding multiple wavefront recognition times.

[0010] S4. From the multiple wavefront identification times obtained, select two wavefront identification times that meet the predetermined conditions, and combine them with the total length of the underground high-voltage cable and the predetermined coaxial wave propagation speed to locate the position of the fault point from the sheath grounding terminal.

[0011] Specifically, step S2 includes:

[0012] The arrival time of the coaxial wave is determined, and based on the arrival time of the coaxial wave, intercepts are made before and after the base point each time to form multiple time windows of length K; where K is a positive number.

[0013] Based on each time window of equal length K, feature extraction is performed on the synthesized signal to obtain N time series data containing the arrival time of the coaxial wave and with a fixed length of K; where N is a positive integer.

[0014] Specifically, step S3 includes:

[0015] Wavelet decomposition is performed on all N time series data to obtain N wavelet data of size J×K×1; where J is the level of the wavelet decomposition transformation and its value is greater than 6.

[0016] In the obtained N wavelet data, the detailed coefficients of layers 2 to 6 are extracted and normalized to obtain N wavelet data with fundamental frequency and high-frequency noise components removed.

[0017] The N wavelet data obtained after removing the fundamental frequency and high-frequency noise components are respectively imported into the pre-trained coaxial wave arrival recognition model to obtain the wavefront recognition time corresponding to each wavelet data after removing the fundamental frequency and high-frequency noise components.

[0018] Specifically, step S4 includes:

[0019] From the multiple wavefront identification times obtained, the two wavefront identification times that appear most frequently and the next most frequently are selected.

[0020] Based on the two selected wavefront identification times, combined with the total length of the underground high-voltage cable and the predetermined coaxial wave propagation speed, the location of the fault point from the sheath grounding terminal is calculated.

[0021] Among them, through the formula Calculate the position x of the fault point from the grounding terminal of the sheath; where,

[0022] t ω1 The wavefront identification time with the highest frequency; t ω2 The wavefront identification time is the second most frequent occurrence; l is the total length of the underground high-voltage cable; v c The predetermined coaxial wave propagation speed.

[0023] This invention also provides a system for accurately locating faults in underground high-voltage cables, comprising:

[0024] The signal measurement and superposition unit is used to measure the three-phase sheath current signal in real time from the sheath grounding terminal at one end of the underground high-voltage cable when a fault occurs, and to linearly superimpose it into a composite signal.

[0025] A signal feature extraction unit is used to determine the arrival time of the coaxial wave, and to construct multiple time windows of equal length and different time ranges based on the arrival time of the coaxial wave. Based on the constructed multiple time windows, feature extraction is performed on the synthesized signal to obtain multiple time series data containing the arrival time of the coaxial wave and of fixed length. The arrival time of the coaxial wave is used to indicate the time difference between two consecutive acquisitions of the fault traveling wave signal at the sheath grounding terminal after it is generated at the fault point of the underground high-voltage cable.

[0026] The coaxial wave arrival recognition unit is used to perform normalized wavelet decomposition on the multiple time series data and import them into the pre-trained coaxial wave arrival recognition model to obtain the corresponding multiple wavefront recognition times.

[0027] The fault location unit is used to select two wavefront identification times that meet predetermined conditions from multiple wavefront identification times obtained, and combine them with the total length of the underground high-voltage cable and the predetermined coaxial wave propagation speed to locate the position of the fault point from the sheath grounding terminal.

[0028] The signal feature extraction unit includes:

[0029] The time window sliding interception module is used to determine the arrival time of the coaxial wave, and to intercept the time before and after the arrival time of the coaxial wave as the base point each time, so as to form multiple time windows of length K; where K is a positive number;

[0030] The time window feature extraction module is used to extract features from the synthesized signal according to each time window of equal length K, so as to obtain N time series data containing the arrival time of the coaxial wave and with a fixed length of K; where N is a positive integer.

[0031] The coaxial wave arrival identification unit includes

[0032] The signal wavelet decomposition module is used to perform wavelet decomposition on all N time series data to obtain N wavelet data of size J×K×1; where J is the level of the wavelet decomposition transformation and its value is greater than 6.

[0033] The wavelet coefficient extraction module is used to extract detailed coefficients from layers 2 to 6 in the obtained N wavelet data and perform normalization processing to obtain N wavelet data with fundamental frequency and high-frequency noise components removed.

[0034] The coaxial wave arrival recognition module is used to import the N wavelet data obtained after removing the fundamental frequency and high-frequency noise components into the pre-trained coaxial wave arrival recognition model to obtain the wavefront recognition time corresponding to each wavelet data after removing the fundamental frequency and high-frequency noise components.

[0035] The fault location unit includes:

[0036] The wavefront identification time filtering module is used to filter out the two wavefront identification times that appear most frequently and the next most frequently from the multiple wavefront identification times obtained.

[0037] The fault location module is used to calculate the distance of the fault point from the sheath grounding terminal by combining the two selected wavefront identification times with the total length of the underground high-voltage cable and the predetermined coaxial wave propagation speed.

[0038] Among them, through the formula Calculate the position x of the fault point from the grounding terminal of the sheath; where,

[0039] t ω1 The wavefront identification time with the highest frequency; t ω2 The wavefront identification time is the second most frequent occurrence; l is the total length of the underground high-voltage cable; v c The predetermined coaxial wave propagation speed.

[0040] Implementing the embodiments of the present invention has the following beneficial effects:

[0041] This invention is based on the real-time measurement of transient three-phase sheath current signals at the sheath grounding terminal at one end of an underground high-voltage cable. The sum of the three-phase sheath current signals (i.e., linear superposition) is used to replace the measurement of the iron core conductor, and a coaxial wave arrival recognition model (CNN neural network) is used to process the recognition of the coaxial wave arrival time, so as to quickly locate the location of the fault point (such as the distance of the fault point from the sheath grounding terminal). Thus, based on the transient cable monitoring data, it is possible to quickly and accurately locate underground high-voltage cable faults online, which greatly improves the positioning accuracy. Attached Figure Description

[0042] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, obtaining other drawings based on these drawings without creative effort still falls within the scope of the present invention.

[0043] Figure 1 This is a diagram of the cable mutual coupling distributed parameter model in a method for accurate fault location of underground high-voltage cables provided in an embodiment of the present invention;

[0044] Figure 2 A schematic diagram of the modal traveling wave propagation of a cross-bonded cable in a method for precise fault location of an underground high-voltage cable provided in an embodiment of the present invention;

[0045] Figure 3 A flowchart illustrating a method for accurately locating faults in underground high-voltage cables, provided as an embodiment of the present invention;

[0046] Figure 4 This is a time series data distribution map obtained by feature extraction of synthetic signals through a time window in a method for accurate fault location of underground high-voltage cables provided in an embodiment of the present invention.

[0047] Figure 5 for Figure 4 Distribution of wavelet data obtained after normalized wavelet decomposition of time series data;

[0048] Figure 6 A logic diagram of the coaxial wave arrival identification model in a method for accurate fault location of underground high-voltage cables provided in an embodiment of the present invention;

[0049] Figure 7 This is a structural schematic diagram of an underground high-voltage cable fault accurate location system provided in an embodiment of the present invention. Detailed Implementation

[0050] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings.

[0051] The inventors discovered that the modal traveling wave propagation under underground high-voltage cable faults can be derived based on a multi-conductor coupling model to obtain the traveling wave propagation equations for the cable's modal voltage and current. Based on these equations, the propagation characteristics of the traveling wave in the sheath layer, the main conductor, and at cross-connection joints, as well as the characteristics of the sheath current, are analyzed. This allows the sum of the three-phase sheath currents measured at the sheath grounding terminal at one end of the underground high-voltage cable to replace the core conductor measurements, which is then used for fault location in underground high-voltage cables. The specific process is as follows:

[0052] First, according to Figure 1 The distributed parameter model of cable mutual coupling is constructed, and the model equations are shown in equations (1) to (2):

[0053]

[0054]

[0055] in, V i (x,t), I i (x,t) represent the voltage and current at position x, respectively; the subscript i indicates the layer; R ii ,L ii G ii and C ii Represent the resistance, inductance, conductance, and capacitance of the i-th layer, respectively; R ij ,L ij G ij and C ij These represent the mutual parameters of resistance, inductance, conductance, and capacitance between the i-th and j-th layers, respectively.

[0056] Decoupling equations (1) to (2) into modal modes, we obtain equations (3) to (4):

[0057]

[0058] Where Λ = S -1 ZYS; V M =S -1 V;I M =Q -1 I;V M ,I M Let Q represent the voltage and current vectors in the modal modes, respectively. -1 =S T .

[0059] By solving equations (3) to (4), the modal voltage and current are obtained and expressed in traveling wave form as follows:

[0060]

[0061] Among them, F V and B V These are the positive and negative wave vectors of the voltage, respectively; F I and B I These are the positive and negative wave vectors of the current, respectively; γ m Let be the propagation constant of mode m (m = 1, 2, ..., M), which is calculated from the square root of the Mth eigenvalue in the diagonal matrix Λ.

[0062] Secondly, according to Figure 2 This diagram illustrates the modal traveling wave propagation in a cross-bonded cable. Assuming a fault occurs in the third section, the modal current wave propagates to both terminals. At this point, the coaxial mode wave propagates fastest, followed by the intersheath mode wave, and finally the ground mode wave. It should be noted that the waves in each modal mode are decoupled and propagate independently, but they interfere with each other at the cross-bonding joint due to phase transposition.

[0063] With coaxial mode wave I cf For example, when they propagate from the third part to the second part, a new wave T is generated. c- s I cf The interlaminar mode wave received at time (xl / 3) / Vc + l / 3Vs is not a wave directly generated by the fault location, resulting in the loss of some fault location information. In contrast, the first coaxial wave received at the terminal always originates directly from the fault location because its propagation is not affected by the crosslink. Therefore, the interlaminar mode wave is not the most ideal electrical quantity for fault location because it is difficult to identify whether the wave originates from the fault location.

[0064] from Figure 2 It can be seen that the first two waves reaching terminal 1 are coaxial mode waves I. cf The intersheath mode wave T c-s I cf It is generated between the first and second parts. The terminal current of the mode can be expressed by the Laplace transform as shown in equations (10) to (11):

[0065]

[0066] Where Vc and Vs represent the velocities of the coaxial mode and the intersheath mode, respectively; β ct β strepresents the reflection coefficient matrix of the coaxial mode and the intersheath mode, respectively; x and l represent the fault distance and length of the cable, respectively.

[0067] By substituting equations (10) to (11), the current in the core conductor and the current in the sheath are derived as shown in equations (12) to (13):

[0068]

[0069] Among them, I ck and I sk (k=a,b,c) represents the core conductor current wave and sheath current wave of phase k at terminal 1 (for simplicity, they will be referred to as conductor current and sheath current below); I cfi (i = 1, 2, 3) represents a coaxial current wave; Z cci (i = 1, 2, 3) represents the characteristic impedance of coaxial mode i; Z0 represents the equivalent impedance connected at terminal 1.

[0070] At this time, equations (12) to (13) represent the conductor current I, respectively. ck (k=a,b,c) initially consists only of coaxial mode waves, while the sheath current I sk It consists of coaxial and sheath-intermodal wave modes.

[0071] As the previous analysis shows, the coaxial mode wave is the preferred localized charge. However, according to equation (13), the sheath current I... sk Still contains conductive current I ck All information. Therefore, it is necessary to eliminate intersheath mode waves, mainly because intersheath mode waves have a potential similar to conductor currents.

[0072] Finally, the intersheath mode wave is eliminated by adding the three-phase sheath currents given in equation (13) to obtain the synthesized signal, as shown in equation (14):

[0073]

[0074] In summary, from equation (14), it can be determined that the sheath-exit mode wave is eliminated, and the sum is left only with the coaxial mode wave, similar to the conductor current given in (12). Therefore, the sum of the three-phase sheath currents can be measured at the sheath grounding terminal at one end of the underground high-voltage cable to replace the core conductor measurement for fault location.

[0075] At this time, as Figure 3 As shown in the figure, this is an embodiment of the present invention, in which the inventors propose a method for accurately locating faults in underground high-voltage cables based on the above conclusions. The method includes the following steps:

[0076] Step S1: When a fault occurs in the underground high-voltage cable, the three-phase sheath current signal is measured in real time from the sheath grounding terminal at one end of the underground high-voltage cable and linearly superimposed into a composite signal.

[0077] The specific process is as follows: First, when a fault occurs in the underground high-voltage cable, based on... Figure 2 First, select or specify the sheath grounding terminal at one end of the underground high-voltage cable for measurement to obtain the three-phase sheath current signal in real time. Second, according to equation (14), perform linear superposition calculation on the measured three-phase sheath current signals, and use the sum as the composite signal I. sum It should be noted that the synthesized signal I sum It contains only the coaxial mode wave corresponding to the core conductor current, and other interference components are effectively eliminated.

[0078] Step S2: Determine the arrival time of the coaxial wave, and construct multiple time windows of equal length and different time ranges based on the arrival time of the coaxial wave. Based on the constructed multiple time windows, perform feature extraction on the synthesized signal to obtain multiple time series data containing the arrival time of the coaxial wave and with fixed length. The arrival time of the coaxial wave is used to indicate the time difference between two consecutive acquisitions of the fault traveling wave signal at the sheath grounding terminal after it is generated at the fault point of the underground high-voltage cable.

[0079] The specific process is as follows: First, the arrival time of the coaxial wave is determined. Then, using this arrival time as a base point, interceptions are made before and after this base point each time to form multiple time windows of length K; where K is a positive number. It should be noted that the optimal coaxial wave arrival time is the time difference between the first and second acquisitions of the fault traveling wave signal.

[0080] Secondly, based on each time window of equal length K, the synthesized signal I is... sum Feature extraction is performed to obtain N time series data points containing the arrival times of coaxial waves and with a fixed length of K, such as... Figure 4 As shown; where N is a positive integer.

[0081] It should be noted that the arrival times of the coaxial waves preceding and following each of the above time series data are, for example... Axial wave arrival time; K is the data length, also the time window length; f s The sampling rate is 50MHz; the range of α is [0.5,1]; the number of datasets selected is N = 0.45 × K.

[0083] Step S3: Normalize wavelet decomposition is performed on all the multiple time series data, and the results are imported into the pre-trained coaxial wave arrival recognition model to obtain the corresponding multiple wavefront recognition times.

[0084] The specific process is as follows: First, wavelet decomposition (e.g., using the bior1.1 mother wavelet) is performed on all N time series data to obtain N wavelet data of size J×K×1, such as... Figure 5 As shown; where J is the level of the wavelet decomposition transform, and its value is greater than 6.

[0085] Secondly, from the obtained N wavelet data, detailed coefficients from layers 2 to 6 are extracted and normalized to obtain N wavelet data with fundamental frequency and high-frequency noise components removed. It should be noted that the wavelet decomposition coefficients are normalized to ensure that the amplitudes of the wavelet data N are consistent under different fault conditions.

[0086] It should be noted that the wavelet decomposition formula is introduced. Extract wavelet data for wave arrival characteristics; where x j,L [n] and x j,H [n] are the approximation coefficients and detail coefficients of level j, respectively; g[m] and h[m] are filter coefficients of length m; the spectrum decomposed at level j [f s / 2 j+1 ,f s / 2 j ] represents the detailed coefficient, [0,f s / 2 j+1 [ ] are approximate coefficients. Therefore, in order to exclude the fundamental frequency component and the noise that is more dominant in the high frequency, the detailed coefficients of levels 2 to 6 after min-max normalization are extracted as direct inputs to the coaxial wave arrival identification model.

[0087] Finally, the N wavelet data obtained after removing the fundamental frequency and high-frequency noise components are imported into the pre-trained coaxial wave arrival recognition model to obtain the wavefront recognition time corresponding to each wavelet data after removing the fundamental frequency and high-frequency noise components.

[0088] It should be noted that a CNN neural network is introduced to construct a coaxial wave arrival recognition model, which is used to identify coaxial wave patterns and specifically handles grid-like topological data. This model matches wavelet data of size J×K×1 in wavelet decomposition. It should also be noted that the training and testing of the coaxial wave arrival recognition model are implemented using common techniques in this field, which will not be elaborated upon here.

[0089] At this point, the model input data is processed through convolutional layers to create feature maps, then downsampled through pooling layers to extract coaxial wave features, and the features obtained from wavelet decomposition are refined. With these improved features, the coaxial wave arrival time can be more directly identified by the subsequent fully connected layers, thus ensuring positioning accuracy.

[0090] Therefore, N wavelet data points, after removing the fundamental frequency and high-frequency noise components, are sequentially input into the coaxial wave arrival recognition model to derive N corresponding wavefront recognition times. The results are then encoded using one-hot encoding, as shown below. Figure 6 As shown.

[0091] Step S4: From the multiple wavefront identification times obtained, select two wavefront identification times that meet the predetermined conditions, and combine them with the total length of the underground high-voltage cable and the predetermined coaxial wave propagation speed to locate the position of the fault point from the sheath grounding terminal.

[0092] The specific process is as follows: First, from the multiple wavefront identification times obtained, the two wavefront identification times that appear most frequently and those that appear next most frequently are selected. For example, by using a voting statistical method for multiple identification results, the wavefront identification error is generally less than 5 sampling points, corresponding to the time... Select the wavefront arrival time t that occurs most frequently. ω1 and t ω2 As the final screening result.

[0093] Secondly, based on the two selected wavefront identification times, combined with the total length of the underground high-voltage cable and the predetermined coaxial wave propagation speed, the distance from the fault point to the sheath grounding terminal is calculated. This is achieved using the formula... Calculate the distance x from the fault point to the sheath grounding terminal; t ω1 The wavefront identification time with the highest frequency; t ω2 The wavefront identification time is the second most frequent occurrence; l is the total length of the underground high-voltage cable; v c This is the predetermined coaxial wave propagation speed.

[0094] In one example, the object of measurement is a 110kV cross-shaped buried cable with a total length of 1.5km, divided into 3 sections. The three-phase sheath current is measured at the sheath grounding terminal at one end of the cable, with a data acquisition sampling rate of 50MHz, thus obtaining a composite signal I obtained by linear superposition of the three-phase sheath current signals. sum .

[0095] At this time, a ground fault is applied at a distance of 300m and 600m. The calculated location of the fault point from the sheath grounding terminal is shown in Table 1 below, which makes the fault location accuracy rate as high as 99.3% or more.

[0096] Table 1

[0097]

[0098] like Figure 7 As shown in the figure, an underground high-voltage cable fault precise location system is provided in an embodiment of the present invention, comprising:

[0099] The signal measurement and superposition unit 110 is used to measure the three-phase sheath current signal in real time from the sheath grounding terminal at one end of the underground high-voltage cable when a fault occurs, and to linearly superimpose it into a composite signal.

[0100] The signal feature extraction unit 120 is used to determine the arrival time of the coaxial wave, and construct multiple time windows of equal length and different time ranges based on the arrival time of the coaxial wave. Based on the constructed multiple time windows, the synthesized signal is subjected to feature extraction to obtain multiple time series data containing the arrival time of the coaxial wave and with fixed length. The arrival time of the coaxial wave is used to indicate the time difference between two consecutive acquisitions of the fault traveling wave signal at the sheath grounding terminal after it is generated at the fault point of the underground high-voltage cable.

[0101] The coaxial wave arrival recognition unit 130 is used to perform normalized wavelet decomposition on the multiple time series data and import them into the pre-trained coaxial wave arrival recognition model to obtain the corresponding multiple wavefront recognition times.

[0102] The fault location unit 140 is used to select two wavefront identification times that meet predetermined conditions from the multiple wavefront identification times obtained, and combine them with the total length of the underground high-voltage cable and the predetermined coaxial wave propagation speed to locate the position of the fault point from the sheath grounding terminal.

[0103] The signal feature extraction unit 120 includes:

[0104] The time window sliding interception module is used to determine the arrival time of the coaxial wave, and to intercept the time before and after the arrival time of the coaxial wave as the base point each time, so as to form multiple time windows of length K; where K is a positive number;

[0105] The time window feature extraction module is used to extract features from the synthesized signal according to each time window of equal length K, so as to obtain N time series data containing the arrival time of the coaxial wave and with a fixed length of K; where N is a positive integer.

[0106] The coaxial wave arrival identification unit 130 includes

[0107] The signal wavelet decomposition module is used to perform wavelet decomposition on all N time series data to obtain N wavelet data of size J×K×1; where J is the level of the wavelet decomposition transformation and its value is greater than 6.

[0108] The wavelet coefficient extraction module is used to extract detailed coefficients from layers 2 to 6 in the obtained N wavelet data and perform normalization processing to obtain N wavelet data with fundamental frequency and high-frequency noise components removed.

[0109] The coaxial wave arrival recognition module is used to import the N wavelet data obtained after removing the fundamental frequency and high-frequency noise components into the pre-trained coaxial wave arrival recognition model to obtain the wavefront recognition time corresponding to each wavelet data after removing the fundamental frequency and high-frequency noise components.

[0110] The fault location unit 140 includes:

[0111] The wavefront identification time filtering module is used to filter out the two wavefront identification times that appear most frequently and the next most frequently from the multiple wavefront identification times obtained.

[0112] The fault location module is used to calculate the distance of the fault point from the sheath grounding terminal by combining the two selected wavefront identification times with the total length of the underground high-voltage cable and the predetermined coaxial wave propagation speed.

[0113] Among them, through the formula Calculate the position x of the fault point from the grounding terminal of the sheath; where,

[0114] t ω1 The wavefront identification time with the highest frequency; t ω2 The wavefront identification time is the second most frequent occurrence; l is the total length of the underground high-voltage cable; v c The predetermined coaxial wave propagation speed.

[0115] Implementing the embodiments of the present invention has the following beneficial effects:

[0116] This invention is based on the real-time measurement of transient three-phase sheath current signals at the sheath grounding terminal at one end of an underground high-voltage cable. The sum of the three-phase sheath current signals (i.e., linear superposition) is used to replace the measurement of the iron core conductor, and a coaxial wave arrival recognition model (CNN neural network) is used to process the recognition of the coaxial wave arrival time, so as to quickly locate the location of the fault point (such as the distance of the fault point from the sheath grounding terminal). Thus, based on the transient cable monitoring data, it is possible to quickly and accurately locate underground high-voltage cable faults online, which greatly improves the positioning accuracy.

[0117] It is worth noting that the system modules included in the above system embodiments are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of each functional module are only for easy differentiation and are not used to limit the scope of protection of the present invention.

[0118] Those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, such as ROM / RAM, disk, optical disk, etc.

[0119] The above description discloses only preferred embodiments of the present invention and should not be construed as limiting the scope of the present invention. Therefore, equivalent variations made in accordance with the claims of the present invention are still within the scope of the present invention.

Claims

1. A method for accurately locating faults in underground high-voltage cables, characterized in that, The method includes the following steps: S1. When a fault occurs in an underground high-voltage cable, the three-phase sheath current signal is measured in real time from the sheath grounding terminal at one end of the underground high-voltage cable and linearly superimposed into a composite signal. S2. Determine the arrival time of the coaxial wave, and construct multiple time windows of equal length and different time ranges based on the arrival time of the coaxial wave. Based on the constructed multiple time windows, perform feature extraction on the synthesized signal to obtain multiple time series data containing the arrival time of the coaxial wave and with fixed length; wherein, the arrival time of the coaxial wave is used to indicate the time difference between two consecutive acquisitions of the fault traveling wave signal at the sheath grounding terminal after it is generated at the fault point of the underground high-voltage cable. S3. Normalize wavelet decomposition is performed on all the time series data, and the data are respectively imported into the pre-trained coaxial wave arrival recognition model to obtain the corresponding multiple wavefront recognition times. S4. From the multiple wavefront identification times obtained, select two wavefront identification times that meet the predetermined conditions, and combine them with the total length of the underground high-voltage cable and the predetermined coaxial wave propagation speed to locate the position of the fault point from the sheath grounding terminal.

2. The method for precise fault location of underground high-voltage cables as described in claim 1, characterized in that, Step S2 specifically includes: The arrival time of the coaxial wave is determined, and based on the arrival time of the coaxial wave, intercepts are made before and after the base point each time to form multiple time windows of length K; where K is a positive number. Based on each time window of equal length K, feature extraction is performed on the synthesized signal to obtain N time series data containing the arrival time of the coaxial wave and with a fixed length of K; where N is a positive integer.

3. The method for precise fault location of underground high-voltage cables as described in claim 2, characterized in that, Step S3 specifically includes: Wavelet decomposition is performed on all N time series data to obtain N wavelet data of size J×K×1; where J is the level of the wavelet decomposition transformation and its value is greater than 6. In the obtained N wavelet data, the detailed coefficients of layers 2 to 6 are extracted and normalized to obtain N wavelet data with fundamental frequency and high-frequency noise components removed. The N wavelet data obtained after removing the fundamental frequency and high-frequency noise components are respectively imported into the pre-trained coaxial wave arrival recognition model to obtain the wavefront recognition time corresponding to each wavelet data after removing the fundamental frequency and high-frequency noise components.

4. The method for precise location of underground high-voltage cable faults as described in claim 1, characterized in that, Step S4 specifically includes: From the multiple wavefront identification times obtained, the two wavefront identification times that appear most frequently and the next most frequently are selected. Based on the two selected wavefront identification times, combined with the total length of the underground high-voltage cable and the predetermined coaxial wave propagation speed, the location of the fault point from the sheath grounding terminal is calculated.

5. The method for precise fault location of underground high-voltage cables as described in claim 4, characterized in that, Through formula Calculate the position x of the fault point from the grounding terminal of the sheath; where, t ω1 The wavefront identification time with the highest frequency; t ω2 The wavefront identification time is the second most frequent occurrence; l is the total length of the underground high-voltage cable; v c The predetermined coaxial wave propagation speed.

6. A precise fault location system for underground high-voltage cables, characterized in that, include: The signal measurement and superposition unit is used to measure the three-phase sheath current signal in real time from the sheath grounding terminal at one end of the underground high-voltage cable when a fault occurs, and to linearly superimpose it into a composite signal. A signal feature extraction unit is used to determine the arrival time of the coaxial wave, and to construct multiple time windows of equal length and different time ranges based on the arrival time of the coaxial wave. Based on the constructed multiple time windows, feature extraction is performed on the synthesized signal to obtain multiple time series data containing the arrival time of the coaxial wave and of fixed length. The arrival time of the coaxial wave is used to indicate the time difference between two consecutive acquisitions of the fault traveling wave signal at the sheath grounding terminal after it is generated at the fault point of the underground high-voltage cable. The coaxial wave arrival recognition unit is used to perform normalized wavelet decomposition on the multiple time series data and import them into the pre-trained coaxial wave arrival recognition model to obtain the corresponding multiple wavefront recognition times. The fault location unit is used to select two wavefront identification times that meet predetermined conditions from multiple wavefront identification times obtained, and combine them with the total length of the underground high-voltage cable and the predetermined coaxial wave propagation speed to locate the position of the fault point from the sheath grounding terminal.

7. The underground high-voltage cable fault precise location system as described in claim 6, characterized in that, The signal feature extraction unit includes: The time window sliding interception module is used to determine the arrival time of the coaxial wave, and to intercept the time before and after the arrival time of the coaxial wave as the base point each time, so as to form multiple time windows of length K; where K is a positive number; The time window feature extraction module is used to extract features from the synthesized signal according to each time window of equal length K, so as to obtain N time series data containing the arrival time of the coaxial wave and with a fixed length of K; where N is a positive integer.

8. The underground high-voltage cable fault precise location system as described in claim 7, characterized in that, The coaxial wave arrival identification unit includes The signal wavelet decomposition module is used to perform wavelet decomposition on all N time series data to obtain N wavelet data of size J×K×1; where J is the level of the wavelet decomposition transformation and its value is greater than 6. The wavelet coefficient extraction module is used to extract detailed coefficients from layers 2 to 6 in the obtained N wavelet data and perform normalization processing to obtain N wavelet data with fundamental frequency and high-frequency noise components removed. The coaxial wave arrival recognition module is used to import the N wavelet data obtained after removing the fundamental frequency and high-frequency noise components into the pre-trained coaxial wave arrival recognition model to obtain the wavefront recognition time corresponding to each wavelet data after removing the fundamental frequency and high-frequency noise components.

9. The underground high-voltage cable fault precise location system as described in claim 8, characterized in that, The fault location unit includes: The wavefront identification time filtering module is used to filter out the two wavefront identification times that appear most frequently and the next most frequently from the multiple wavefront identification times obtained. The fault location module is used to calculate the distance of the fault point from the sheath grounding terminal by combining the two selected wavefront identification times with the total length of the underground high-voltage cable and the predetermined coaxial wave propagation speed.

10. The underground high-voltage cable fault precise location system as described in claim 9, characterized in that, Through formula Calculate the position x of the fault point from the grounding terminal of the sheath; where, t ω1 The wavefront identification time with the highest frequency; t ω2 The wavefront identification time is the second most frequent occurrence; l is the total length of the underground high-voltage cable; v c The predetermined coaxial wave propagation speed.

Images