A method for cable soft fault frequency domain positioning and evaluation

Abstract

This invention discloses a frequency domain method for locating and evaluating cable soft faults, comprising: acquiring a swept-frequency signal of the cable with the soft fault; extracting the real part of the cable transfer function (CFD) from the swept-frequency signal; separating the reflected wave from the real part of the CFD using a variational mode decomposition method to obtain the eigenmode function (EMF) signal of the reflected wave; estimating the frequency and attenuation factor of the CFD using a sliding-window TLS-ESPRIT algorithm based on the EMF signal of the reflected wave, and then estimating the amplitude and phase of the CFD using a least-squares method; determining the location of the cable soft fault based on the frequency of the CFD to obtain a cable soft fault location spectrum. This invention transforms the random cable transfer function signal into a deterministic autocorrelation function array signal, thereby achieving high frequency resolution for soft fault location and evaluation, and improving the accuracy of cable soft fault location.

Description

Technical Field

[0001] This invention relates to the field of cable fault detection technology, specifically to a method for frequency domain location and evaluation of cable soft faults. Background Technology

[0002] As power cables age, insulation aging becomes a prominent issue. Soft faults such as partial discharge and copper shield corrosion are gradually developing into hard faults, placing the cables in unsafe operating conditions. These soft faults create impedance changes due to variations in the cable's local capacitance, leading to varying degrees of traveling wave reflections. For example, water entering the cable produces negative reflections, while corrosion or cracks in the copper shield result in positive reflections. Therefore, accurately locating soft faults in cable lines and estimating reflection coefficients are crucial for diagnosing and monitoring the health of flexible cables.

[0003] Cable fault assessment generally involves two steps: location and estimation of reflection coefficient. Based on signal analysis methods, domains, and injected signals, existing soft fault location methods can be summarized as Time Domain Reflectometry (TDR), Time-Frequency Domain Reflectometry (TFDR), and Frequency Domain Reflectometry (FDR).

[0004] TDR and its improved methods (such as sequential time-domain reflectometry (STDR) and spread-spectrum time-domain reflectometry (SSTDR)) can effectively locate hard faults such as open circuits and short circuits. However, due to insufficient high-frequency components in the test signal, they are rarely used to locate soft faults such as those caused by friction or cracks. TFDR, as an improvement on TDR, adds frequency domain analysis to the time domain analysis and uses a Gaussian envelope signal to enhance the high-frequency components of the injected signal. Therefore, TFDR is more sensitive to soft faults. However, multiple reflections can cause severe cross-terms in TFDR, thus interfering with the identification of soft faults. In addition, TDR and TFDR are difficult to accurately assess the characteristics of faults. The FDR method uses a linearly swept frequency signal as the cable injection signal. By adjusting the frequency band, a cable transfer function with rich information about soft faults can be obtained, and combined with Fast Fourier Transform (FFT) to locate soft faults. However, within the test frequency band, the cable transfer function is an incomplete random signal, and spectral leakage and picket-fence effects occur during the FFT process, affecting the accuracy of soft fault location. Although windowing the cable transfer function can improve positioning accuracy, different types and orders of window functions have different effects on the main blade width, side blade level, and side blade attenuation rate, leading to misjudgment or omission of soft faults; and the FDR method cannot identify the type of soft fault. Summary of the Invention

[0005] To address the aforementioned shortcomings in the prior art, this invention provides a method for frequency domain location and evaluation of cable soft faults.

[0006] To achieve the above-mentioned objectives, the technical solution adopted by this invention is as follows:

[0007] In a first aspect, the present invention proposes a frequency domain location method for cable soft faults, comprising the following steps:

[0008] S1. Obtain the sweep frequency signal of the soft fault cable;

[0009] S2. Extract the real part of the cable transfer function from the sweep frequency signal of the soft-fault cable;

[0010] S3. The variational mode decomposition method is used to separate the cable body reflected wave from the real part signal of the cable transfer function to obtain the eigenmode function signal of the cable body reflected wave.

[0011] S4. Based on the eigenmode function signal of the reflected wave of the cable body, the frequency and attenuation factor of the cable transfer function are estimated by the sliding window TLS-ESPRIT algorithm, and then the amplitude and phase of the cable transfer function are estimated by the least squares method.

[0012] S5. Determine the location of the cable soft fault based on the frequency of the cable transfer function, and obtain the cable soft fault location spectrum.

[0013] Optionally, step S4 specifically includes the following sub-steps:

[0014] S41. The eigenmode function signal of the reflected wave of the cable body is truncated into continuous sub-data segments using a rectangular window;

[0015] S42. Convert the time-domain signal of each sub-data segment into a spatial-domain signal;

[0016] S43. Construct the first spatial matrix of the first subarray and the second spatial matrix of the second subarray based on the spatial domain signal;

[0017] S44. Construct a combined array signal based on the first spatial matrix and the second spatial matrix;

[0018] S45. Calculate the covariance matrix of the combined array signal;

[0019] S46. Perform eigenvalue decomposition on the covariance matrix to obtain the eigenvector matrix;

[0020] S47. Construct the first matrix based on the eigenvector matrix, and perform singular value decomposition on the first matrix to obtain the right orthogonal matrix of the first matrix;

[0021] S48. Construct the second matrix based on the right orthogonal matrix of the first matrix, and perform singular value decomposition on the second matrix to obtain the diagonal matrix of the second matrix;

[0022] S49. Calculate the frequency and attenuation factor of each harmonic component based on the diagonal matrix of the second matrix;

[0023] S410. Construct a third spatial matrix based on the spatial domain signal, and use the overall least squares method to estimate the amplitude and phase of the cable transfer function.

[0024] Optionally, step S43 specifically includes:

[0025] Construct a first subarray consisting of M array elements based on the spatial domain signal s(n);

[0026] The first spatial matrix of the first subarray is constructed based on the first N snapshots, and is represented as follows:

[0027]

[0028] X∈R M×N It is composed of the first [0, N+M-1] sample values ​​of s(n).

[0029] Optionally, step S43 specifically includes:

[0030] Construct a second subarray with M elements based on the spatial domain signal s(n+1);

[0031] The second space matrix of the second subarray is constructed based on the first N snapshots of the second subarray, and is represented as follows:

[0032]

[0033] Y∈R M×N It is composed of the first [1, N+M] sample values ​​of s(n+1).

[0034] Optionally, the combined array signal specifically includes:

[0035]

[0036] Optionally, step S47, which involves constructing the first matrix based on the eigenvector matrix, specifically includes:

[0037] Based on the eigenvector matrix, construct the first submatrix and the second submatrix respectively, as follows:

[0038] U1 = U[1:M, 1:2q]

[0039] U2=U[2:M+1,1:2q]

[0040] Where U is the eigenvector matrix;

[0041] The first matrix is ​​constructed based on the first and second submatrices, and is represented as follows:

[0042] U′=[U1,U2。

[0043] Optionally, the formulas for calculating the frequency and attenuation factor of each harmonic component in step S49 are as follows:

[0044]

[0045] in, Let κ′ be the frequency of the i-th harmonic component. i Let λ be the attenuation rate of the i-th harmonic component. i Let be the i-th eigenvalue in the diagonal matrix of the second matrix, and Δf be the step size of the sweep frequency signal.

[0046] Optionally, in step S410, the spatial domain signal s(n) is expanded as follows:

[0047]

[0048] And order

[0049]

[0050] Therefore, a third spatial matrix X′ is constructed based on the spatial domain signal s(n), and is expressed as:

[0051] X′=Bμ+ε

[0052] in,

[0053]

[0054] Among them, b 2q For the 2qth guiding matrix, ω i Let be the angular frequency of the i-th reflected wave.

[0055] μ is calculated using the least squares method, i.e.:

[0056] μ=(B H B) -1 B H X′

[0057] Thus, the amplitude a of the i-th harmonic component is obtained. i and phase θ i :

[0058]

[0059] Secondly, this invention proposes a method for assessing cable soft faults. By using the above-mentioned frequency domain location method for cable soft faults, a cable soft fault location spectrum is obtained to determine the location of the soft fault. Then, the reflection coefficient of the cable soft fault is calculated at each soft fault location based on the cable transfer function, and the variation characteristics of the cable soft fault are assessed based on the reflection coefficient.

[0060] Optionally, the formula for calculating the reflection coefficient is:

[0061]

[0062] Among them, P j Let A' be the reflection coefficient of the j-th soft fault location. j Let be the amplitude of the reflection at the j-th soft fault.

[0063] The present invention has the following beneficial effects:

[0064] This invention employs a sliding window TLS-ESPRIT to perform harmonic recovery on the cable transfer function, replacing the traditional FT with a covariance matrix. This eliminates the randomness of the cable transfer function, transforming the random cable transfer function signal into a deterministic autocorrelation function array signal, thereby enabling high-precision location and assessment of soft faults. Attached Figure Description

[0065] Figure 1 This is a flowchart illustrating a frequency domain method for locating cable soft faults in this embodiment.

[0066] Figure 2 This is a schematic diagram of the receiving array model in this embodiment.

[0067] Figure 3 This is a coaxial cable model.

[0068] Figure 4 The diagram shows the soft fault analysis spectrum in the cable model. (a) represents the intrinsic mode function (IMF) signals s(n) of different frequencies obtained by the VMD method, with labels 1, 2, and 3 representing S(n), the reflected wave at the cable head, and the reflected wave in the cable body, respectively. (b) represents the location result of the determined terminal A, with labels 2, 4, 6, 8, and 10 representing the 2nd, 4th, 6th, 8th, and 10th sliding windows, respectively. (c) represents the location result of the determined soft fault and terminal B, with labels 2, 4, 6, 8, and 10 representing the 2nd, 4th, 6th, 8th, and 10th sliding windows, respectively. Detailed Implementation

[0069] The specific embodiments of the present invention are described below to enable those skilled in the art to understand the present invention. However, it should be understood that the present invention is not limited to the scope of the specific embodiments. For those skilled in the art, various changes are obvious as long as they are within the spirit and scope of the present invention as defined and determined by the appended claims. All inventions utilizing the concept of the present invention are protected.

[0070] Example 1

[0071] like Figure 1 As shown, this embodiment of the invention provides a frequency domain location method for cable soft faults, including the following steps S1 to S6:

[0072] S1. Obtain the sweep frequency signal of the soft fault cable;

[0073] In an optional embodiment of the present invention, one end of the soft fault cable is selected as the signal input terminal, and a series of sinusoidal signals of different frequencies are injected into the terminal by frequency sweeping to obtain the frequency sweep signal of the soft fault cable.

[0074] S2. Extract the real part of the cable transfer function from the sweep frequency signal of the soft-fault cable;

[0075] In an optional embodiment of the present invention, the cable transfer function is extracted from the frequency sweep signal of the soft-fault cable, the real part of the obtained cable transfer function is analyzed, and it is rewritten as a discrete expression, denoted as S(n):

[0076]

[0077] Where q is the number of harmonics in the cable transfer function, and a i Let θ be the amplitude of the i-th reflected wave. i Let κ′ be the phase of the i-th reflected wave. i This is the equivalent attenuation factor at the soft fault location. Let be the frequency of the i-th harmonic component, n be the sample number, n = 1, 2, ..., N', N' be the sample size, and Δf be the step size of the sweep signal.

[0078] S3. The variational mode decomposition method is used to separate the cable body reflected wave from the real part signal of the cable transfer function to obtain the eigenmode function signal of the cable body reflected wave.

[0079] In an optional embodiment of the invention, since the cable's transfer function is a superposition of multiple harmonics of different frequencies, and due to the large reflection coefficient at the cable test end, only a small portion of the test signal energy is injected into the cable. Furthermore, the signal attenuates over long distances within the cable. Compared to the reflection from the test end, the reflected wave from the soft fault is very weak and may be annihilated by the former. Therefore, this embodiment considers extracting the reflected wave from the test end in the cable transfer function and analyzing the reflected wave from the test end and the reflected wave from the soft fault separately. This embodiment uses a variational mode decomposition method assuming that all harmonic components are narrowband signals concentrated around their respective center frequencies, decomposing the real part signal S(n) into intrinsic mode function (IMF) signals s(n) of different frequencies.

[0080] S4. Based on the eigenmode function signal of the reflected wave of the cable body, the frequency and attenuation factor of the cable transfer function are estimated by the sliding window TLS-ESPRIT algorithm, and then the amplitude and phase of the cable transfer function are estimated by the least squares method.

[0081] In an optional embodiment of the present invention, the sliding window TLS-ESPRIT algorithm used in this embodiment is to truncate the intrinsic mode function signal s(n) of the reflected wave of the cable body into multiple sub-intervals using a rectangular window, and then use TLS-ESPRIT to estimate the parameters of the signal in each sub-interval.

[0082] Step S4 in this embodiment specifically includes the following sub-steps S41 to S410:

[0083] S41. The eigenmode function signal of the reflected wave of the cable body is truncated into continuous sub-data segments using a rectangular window;

[0084] In this embodiment, the addition of a symmetrical negative frequency component doubles the number of harmonics, resulting in 2q harmonics. Assume the sample size of the cable transfer function s(n) in each subinterval is L, denoted as {s(n)}. n＝1，2，...，L ,

[0085] S42. Convert the time-domain signal of each sub-data segment into a spatial-domain signal, taking into account the noise in the environment, as follows:

[0086]

[0087] in, w(n) is a Gaussian white noise signal, i is the index of the reflected wave, i = 1, 2, ..., 2q, and q is the number of reflected waves. is the equivalent propagation coefficient of the i-th reflected wave.

[0088] S43. Constructing the first spatial matrix of the first subarray and the second spatial matrix of the second subarray based on the spatial domain signal specifically includes:

[0089] Define the two processes as follows:

[0090]

[0091] Here, x(n) and y(n) are two different subarrays containing M (L>M>2q) array elements, namely the first subarray and the second subarray;

[0092] Then, based on the first N snapshots of the first subarray x(n), the first spatial matrix X of the first subarray is constructed as follows:

[0093]

[0094] Where, x k =[s(k-1),s(k),...,s(k+N-1)] T (k = 1, 2, ..., N, L > N > 2q) represents the data vector received in the k-th snapshot in the first spatial matrix X, further expressed as:

[0095] x k =Cs k +w(k)

[0096] in,

[0097]

[0098] Among them, a k (n) represents the amplitude of the reflected wave from the k-th snapshot, c(ω) 2q ) is the 2qth steering vector, and w(k) is the white noise of the kth snapshot.

[0099] The second spatial matrix Y of the second subarray y(n) is constructed based on the first N snapshots of the second subarray, and is expressed as:

[0100]

[0101] Among them, y k =[s(k),x(k+1),...,x(k+N)] T (k = 1, 2, ..., N) represents the data vector received in the k-th snapshot in the second spatial matrix Y, further expressed as:

[0102] y k =Cφs k +w(k+1)

[0103] in,

[0104]

[0105] φ is a unitary matrix that satisfies φ H φ=φφ H =I, where I is the identity matrix.

[0106] Based on the spatial domain signal s(n) and the data vector x in the first spatial matrix X k The data vector y in the second space matrix Y k A receiver array model can be constructed, such as Figure 2 As shown. Figure 2 d m Let o be the distance between the m-th (0≤m≤M) array element and the first array element. i It is the i-th reflected wave, which can be regarded as a far-field signal.

[0107] S44. Construct a combined array signal based on the first spatial matrix and the second spatial matrix;

[0108] In this embodiment, a combined array signal Z is constructed based on the first spatial matrix X and the second spatial matrix Y, and is represented as follows:

[0109]

[0110] Where S = [s1 s2 … s N W is the Gaussian white noise matrix.

[0111] S45. Calculate the covariance matrix of the combined array signal;

[0112] This embodiment calculates the covariance matrix R of the combined array signal Z. z , is represented as:

[0113]

[0114] Where E[·] is the expectation symbol, R S Let S be the covariance matrix of S, and σ be the standard deviation of the white noise.

[0115] S46. Perform eigenvalue decomposition on the covariance matrix to obtain the eigenvector matrix;

[0116] This embodiment focuses on the covariance matrix R. z Eigenvalue decomposition is performed, and it is expressed as:

[0117]

[0118] Where U is R z The eigenvector matrix Σ is the eigenvector matrix of R. z eigenvalue matrix, Σ s It is an eigenvalue matrix containing the first 2q largest eigenvalues ​​of Σ, U sIt is the signal subspace, U W It is the noise subspace, Σ W It is a diagonal matrix formed by the eigenvalues ​​of the noise.

[0119] S47. Construct the first matrix based on the eigenvector matrix, and perform singular value decomposition on the first matrix to obtain the right orthogonal matrix of the first matrix;

[0120] This embodiment of constructing the first matrix U' based on the eigenvector matrix U specifically includes:

[0121] Based on the eigenvector matrix U, construct the first submatrix U1 and the second submatrix U2 respectively, as follows:

[0122] U1 = U[1:M, 1:2q]

[0123] U2=U[2:M+1,1:2q]

[0124] Wherein, U1 takes the first row to the Mth row (1:M) and the first column to the 2qth column (1:2q) of U; U2 takes the second row to the M+1th row (2:M+1) and the first column to the 2qth column (1:2q) of U.

[0125] Construct the first matrix U' based on the first submatrix U1 and the second submatrix U2, and represent it as follows:

[0126] U′=[U1,U2。

[0127] Then, singular value decomposition is performed on the first matrix U', which is represented as:

[0128]

[0129] in, and These are the left and right orthogonal matrices of U', V is a diagonal matrix composed of the singular values ​​of U'. 11 V 12 V 21 and V 22 yes The submatrix.

[0130] S48. Construct the second matrix based on the right orthogonal matrix of the first matrix, and perform singular value decomposition on the second matrix to obtain the diagonal matrix of the second matrix;

[0131] This embodiment is based on the right orthogonal matrix of the first matrix U'. Construct the second matrix ψ, denoted as:

[0132]

[0133] in, V represents 22 The inverse matrix, U ψ Σ is the eigenvector matrix of ψ. ψ =diag(λ1 λ2 … λ) 2q ) is a diagonal matrix composed of the eigenvalues ​​of ψ.

[0134] S49. Calculate the frequency and attenuation factor of each harmonic component based on the diagonal matrix of the second matrix;

[0135] In this embodiment, because ψ and φ have the same eigenvalues Therefore, the frequency of the i-th harmonic component can be obtained. and attenuation rate κ′ i The calculation formulas are respectively

[0136]

[0137] in, Let κ′ be the frequency of the i-th harmonic component. i Let λ be the attenuation rate of the i-th harmonic component. i Let be the i-th eigenvalue in the diagonal matrix of the second matrix, and Δf be the step size of the sweep frequency signal.

[0138] S410. Construct a third spatial matrix based on the spatial domain signal, and use the least squares method to estimate the amplitude and phase of the cable transfer function.

[0139] In this embodiment, the spatial domain signal s(n) is expanded as follows:

[0140]

[0141] And order

[0142]

[0143] Therefore, a third spatial matrix X′ is constructed based on the spatial domain signal s(n), and is expressed as:

[0144] X′=Bμ+ε

[0145] in,

[0146]

[0147] Among them, b 2q For the 2qth guiding matrix, ω i Let be the angular frequency of the i-th reflected wave.

[0148] Since ε is an error vector, in order to minimize the error, this embodiment uses the least squares method to calculate μ, that is:

[0149] μ=(B H B)-1 B H X′

[0150] Thus, the amplitude a of the i-th harmonic component is obtained. i and phase θ i :

[0151]

[0152] Where i = 1, 2, ..., q.

[0153] S5. Determine the location of the cable soft fault based on the frequency of the cable transfer function, and obtain the cable soft fault location spectrum.

[0154] In an optional embodiment of the present invention, when at a distance x from the terminal i When a soft fault occurs at a distance, a new resonant frequency will appear in the cable, the value of which is related to the x-axis of the soft fault. i Location-dependent, therefore the location of a cable soft fault can be determined based on the frequency of the cable transfer function, i.e.:

[0155]

[0156] Where v is the propagation speed of the injected signal in the cable.

[0157] Example 2

[0158] Based on the frequency domain localization method for cable soft faults described in Example 1, this embodiment further calculates the reflection coefficient of the cable soft fault according to the amplitude and phase of the cable transfer function, and evaluates the changing characteristics of the cable soft fault based on the reflection coefficient.

[0159] In an optional embodiment of the present invention, the coefficient and frequency of the reflected wave respectively reflect the characteristics and location of the cable soft fault. The reflection coefficient directly reflects the changing characteristics of the soft fault. Therefore, this embodiment also includes decoupling the amplitude of the cable transfer function and calculating the reflection coefficient at the soft fault, thereby evaluating the cable soft fault.

[0160] The amplitude of the reflected wave at a soft fault is the result of the coupling of multiple reflection coefficients. Therefore, assessing the state of a soft fault requires decoupling the reflection coefficients. Since the amplitude of the reflection at a soft fault already includes the coupling relationship of the reflection coefficients, only the amplitude of the reflection at the soft fault needs to be considered, while multiple reflections are ignored. Furthermore, the amplitudes of multiple reflections are very small, so the reflection at the soft fault location can be retrieved by setting a threshold.

[0161] Assuming that based on the amplitude distribution of the cable transfer function obtained in Example 1, p soft faults are identified, and the reflection coefficient of the j-th (j = 1, 2, ..., p) soft fault is defined as P. j(P1 represents the head-end reflection coefficient), the reflection amplitude is A' j Then establish P j With A' j The coupling relationship between them is as follows:

[0162]

[0163] Based on the above formula, the amplitude relationship can be obtained as follows:

[0164]

[0165] Therefore, the formula for calculating the reflection coefficient is:

[0166]

[0167] Among them, P j Let A' be the reflection coefficient at the j-th soft fault. j Let be the amplitude of the reflection at the j-th soft fault.

[0168] Application examples

[0169] The coaxial cable model involved in this application example is as follows: Figure 3 As shown, the cable length is l = 500m; Z0 is the characteristic impedance of the cable; terminal A is the test terminal, l' is the length of the test lead (l' = 0.5m), and the capacity of the test lead is set to 0.8 times that of the cable; terminal B is set as an open circuit, Z... L Load impedance (Z) L =∞). A soft fault is set at a distance of 300m from terminal A (i.e., x1 = 300m). The length of the soft fault is d1 = 0.5m. The capacity value at the soft fault location is increased to 1.2 times the capacity value of the cable itself.

[0170] The sampling step size is Δf = 2000Hz; a series of sinusoidal signals from 0.15 to 150MHz are injected into terminal A by frequency sweeping. The result of decomposing the cable transmission function using the Variational Mode Decomposition (VMD) method is as follows: Figure 4 As shown in (a). Then, the cable soft fault frequency domain localization method provided in Example 1 was used to locate the soft fault, and the result is as follows. Figure 4 As shown in (b) and 4(c). In this application example, L = 200, M = 0.4L, and N = 0.6L.

[0171] Figure 4 (a) shows that the VMD method can effectively separate the reflected wave at terminal A. The waveform indicates that it is a cosine function within the interval (0-2π) with a negative sign. The method proposed in this invention can accurately locate the positions of terminal A, the soft fault, and terminal B, such as... Figure 4 (b) and Figure 4 As shown in (c).

[0172] The position, amplitude, and initial phase parameters of terminal A, soft fault, and terminal B estimated by the proposed method are shown in Table 1.

[0173] Table 1 shows the location, amplitude, and initial phase parameters of the soft fault estimated by the proposed method.

[0174]

[0175] As shown in Table 1, the relative positioning errors of the proposed method for terminal A, soft fault, and terminal B are 2.24%, 0.0128%, and 0.004%, respectively.

[0176] When the capacitance of the test line is set to 0.8 times that of the cable, the amplitude of the reflection coefficient of terminal A is 0.0557 and the phase is 3.1416 rad. Table 1 shows that the relative errors in amplitude and phase of the estimation of the reflection coefficient of terminal A using the method proposed in Example 2 are 0.53% and 0.61%, respectively. When the capacitance of the soft fault is set to 1.2 times that of the cable, the amplitude of the reflection coefficient of the soft fault should be 0.0455 and the phase should be 3.1416 rad. Using the method proposed in Example 2 to decouple the amplitude of the soft fault in Table 1, the estimated amplitude and phase of the soft fault reflection coefficient are 0.0432 and 3.1254 rad, respectively, with relative errors in amplitude and phase of 5.05% and 0.52%, respectively. Using the method proposed in Example 2, the amplitude of terminal B is decoupled, and its reflection coefficient is estimated to be 0.9255, the phase is 0.1449 rad, and the relative amplitude error is 7.45%. With the end open-circuited and fully reversible, the absolute phase error is 0.1449 rad.

[0177] Specific embodiments have been used to illustrate the principles and implementation methods of this invention. The descriptions of the embodiments above are only for the purpose of helping to understand the method and core ideas of this invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this invention. Therefore, the content of this specification should not be construed as a limitation of this invention.

[0178] Those skilled in the art will recognize that the embodiments described herein are intended to help the reader understand the principles of the invention, and should be understood that the scope of protection of the invention is not limited to such specific statements and embodiments. Those skilled in the art can make various other specific modifications and combinations based on the technical teachings disclosed in this invention without departing from the spirit of the invention, and these modifications and combinations are still within the scope of protection of this invention.

Claims

1. A frequency domain method for locating soft faults in cables, characterized in that, Includes the following steps: S1. Obtain the sweep frequency signal of the soft fault cable; S2. Extract the real part of the cable transfer function from the sweep frequency signal of the soft-fault cable; S3. The variational mode decomposition method is used to separate the cable body reflected wave from the real part signal of the cable transfer function to obtain the eigenmode function signal of the cable body reflected wave. S4. Based on the eigenmode function signal of the reflected wave from the cable body, the frequency and attenuation factor of the cable transfer function are estimated using the sliding window TLS-ESPRIT algorithm, and then the amplitude and phase of the cable transfer function are estimated using the least squares method; specifically... It includes the following steps: S41. The eigenmode function signal of the reflected wave from the cable body is truncated into continuous sub-data segments using a rectangular window; S42. Convert the time-domain signal of each sub-data segment into a spatial-domain signal; S43. Construct the first spatial matrix of the first subarray and the second spatial matrix of the second subarray based on the spatial domain signal; S44. Construct a combined array signal based on the first spatial matrix and the second spatial matrix; S45. Calculate the covariance matrix of the combined array signal; S46. Perform eigenvalue decomposition on the covariance matrix to obtain the eigenvector matrix; S47. Construct the first matrix based on the eigenvector matrix, and perform singular value decomposition on the first matrix to obtain the right orthogonal matrix of the first matrix; S48. Construct the second matrix based on the right orthogonal matrix of the first matrix, and perform singular value decomposition on the second matrix to obtain the diagonal matrix of the second matrix; S49. Calculate the frequency and attenuation factor of each harmonic component based on the diagonal matrix of the second matrix; S410. Construct a third spatial matrix based on the spatial domain signal, and use the overall least squares method to estimate the amplitude and phase of the cable transfer function; S5. Determine the location of the cable soft fault based on the frequency of the cable transfer function, and obtain the cable soft fault location spectrum.

2. The cable soft fault frequency domain location method according to claim 1, characterized in that, Step S43 specifically includes: According to spatial domain signals Construct the first subarray consisting of M array elements; The first spatial matrix of the first subarray is constructed based on the first N snapshots, and is represented as follows: 。 3. The cable soft fault frequency domain location method according to claim 2, characterized in that, Step S43 specifically includes: According to spatial domain signals Construct a second subarray comprising M array elements; The second space matrix of the second subarray is constructed based on the first N snapshots of the second subarray, and is represented as follows: 。 4. The cable soft fault frequency domain location method according to claim 3, characterized in that, The combined array signal is specifically: 。 5. The cable soft fault frequency domain location method according to claim 4, characterized in that, Step S47, which involves constructing the first matrix based on the eigenvector matrix, specifically includes: Based on the eigenvector matrix, construct the first submatrix and the second submatrix respectively, as follows: ； ； Where U is the eigenvector matrix and q is the number of reflected waves; The first matrix is ​​constructed based on the first and second submatrices, and is represented as follows: 。 6. The cable soft fault frequency domain location method according to claim 5, characterized in that, The formulas for calculating the frequency and attenuation factor of each harmonic component in step S49 are as follows: ； in, Let i be the frequency of the i-th harmonic component. Let be the attenuation rate of the i-th harmonic component. Let i be the i-th eigenvalue in the diagonal matrix of the second matrix. The step size of the sweep frequency signal.

7. The cable soft fault frequency domain location method according to claim 6, characterized in that, In step S410, the spatial domain signal is... Expanded to: ； And order ； Therefore, based on spatial domain signals Constructing the third space matrix , is represented as: ； in, ； in, This is the 2qth orientation matrix. Let be the angular frequency of the i-th reflected wave. Let L be the error vector, and let L be the sample size of the cable transfer function s(n) in each sub-interval; Calculate using the least squares method ,Right now: ； Thus, the amplitude a of the i-th harmonic component is obtained. i and phase θ i : 。 8. A method for assessing soft faults in cables, characterized in that, The cable soft fault location spectrum is obtained by the frequency domain localization method of any one of claims 1 to 7, the location of the soft fault is determined, and then the reflection coefficient of the cable soft fault is calculated according to the amplitude of the cable transfer function at each soft fault location. The variation characteristics of the cable soft fault are evaluated based on the reflection coefficient.

9. The cable soft fault assessment method according to claim 8, characterized in that, The formula for calculating the reflection coefficient is: ； Among them, P j Let A' be the reflection coefficient at the j-th soft fault. j Let be the amplitude of the reflection at the j-th soft fault.

Images