A method and system for locating broadband impedance defects in submarine cables based on dielectric loss correction

Abstract

The application discloses a kind of based on dielectric loss correction submarine cable broadband impedance defect positioning method and system, the method synchronously collects cable first end conductor loop current, after direction normalization, non-main insulation current component correction and amplitude correction, by first end current phasor addition reconstruction main insulation equivalent current, calculate dielectric loss factor;Loss factor of continuous sampling period is carried out credibility screening, temperature and load normalization and trend calculation, obtain dielectric loss trend correction quantity, again by the same type cable standardization electric conductance frequency mapping coefficient mapping to broadband detection frequency band, correct unit length equivalent conductance, admittance and propagation constant, obtain correction phase constant;Subsequently inject broadband signal, collect port voltage current response, calculate reflection coefficient and extract phase anomaly amount, use correction phase constant to invert and fuse multi-frequency point defect distance;The application has alleviated the fixed phase constant mismatch problem caused by dielectric loss change, and improved defect positioning precision.

Description

Technical Field

[0001] This invention belongs to the field of online monitoring and defect location technology for the insulation status of submarine power cables, and particularly relates to a method and system for locating broadband impedance defects in submarine cables based on dielectric loss correction. Background Technology

[0002] Submarine power cables are widely used in offshore wind power transmission, island power supply, cross-sea power grid interconnection, and marine energy development. Due to their long laying distances, complex operating environments, and extremely high maintenance costs, a failure in the main insulation often results in prolonged power outages and significant economic losses. Therefore, online monitoring of the insulation condition of submarine power cables and accurate location of local defects have become key technical requirements for ensuring the reliable operation of marine power transmission systems.

[0003] In existing technologies, online monitoring of the dielectric loss factor (tanδ) can effectively reflect the overall moisture and aging trend of the cable's main insulation. However, this parameter is a macroscopic statistical quantity for the entire cable and cannot pinpoint the specific location of local defects. Wideband impedance detection methods, on the other hand, inject a wideband signal into the cable and analyze the phase spectrum of the port reflection coefficient. Using the phase constant β(ω), frequency domain anomalies are mapped to the cable length, thus enabling spatial location of local defects. These two methods are often used in parallel in engineering, with tanδ used to assess the overall degree of degradation and wideband impedance used to identify local impedance discontinuities.

[0004] However, the main drawback of existing technologies is that the accuracy of broadband impedance positioning is highly dependent on the accuracy of the phase constant β(ω), which shifts with the overall moisture absorption, aging, or increased dielectric loss of the cable's main insulation. Traditional methods use a fixed phase constant βbase(ω) under healthy conditions for distance inversion throughout the cable's entire lifespan. When the overall dielectric loss changes during actual operation, this directly introduces systematic positioning errors. Furthermore, existing solutions only use tanδ as an auxiliary evaluation index for the overall degradation level, without using it for dynamic calibration and correction of the equivalent admittance per unit length and the phase constant in the broadband impedance positioning model. This results in a significant decrease in the accuracy of local defect positioning under changing overall dielectric conditions of submarine power cables. Summary of the Invention

[0005] Purpose of the invention: The purpose of this invention is to provide a broadband impedance defect location method and system for submarine cables based on dielectric loss correction. This aims to alleviate the problem that existing broadband impedance defect location methods suffer from decreased location accuracy due to the use of a fixed phase constant when the overall dielectric loss of the submarine power cable changes. By using the online monitored dielectric loss factor to dynamically calibrate and correct the phase constant, the accuracy of locating local defects under overall damp or aging conditions can be improved.

[0006] Technical Solution: The present invention describes a broadband impedance defect location method for submarine cables based on dielectric loss correction. This method is applied to submarine power cables laid in a seawater environment. The metallic shielding layer, metallic sheath, armor layer, outer sheath, or a combination thereof of the submarine power cable form a continuous or nearly continuous distributed grounding boundary through seawater, the outer sheath discharge path, the armor layer discharge path, the end grounding structure, the distributed conductivity path, or the distributed capacitance coupling path. The method includes the following steps:

[0007] S1. At either the beginning or end of the submarine power cable, the voltage phase of the phase conductor under test relative to the grounding reference point is collected, and the loop current phasor of the phase conductor under test is simultaneously collected at the beginning of the submarine power cable as the first current phasor, and the loop current phasor of the phase conductor under test is simultaneously collected at the end of the submarine power cable as the second current phasor.

[0008] S2. Perform direction normalization, non-main insulation current component correction and amplitude-phase error correction on the first current phasor and the second current phasor to obtain the first corrected current phasor and the second corrected current phasor.

[0009] S3. Add the first corrected current phasor and the second corrected current phasor to obtain the remaining current phasor, and use the remaining current phasor as the equivalent current phasor of the main insulation of the submarine power cable. Calculate the dielectric loss factor based on the equivalent current phasor of the main insulation and the voltage phase reference. The dielectric loss factor is used to characterize the overall dielectric loss change of the main insulation of the submarine power cable and to correct the phase propagation parameters in the broadband impedance positioning model.

[0010] S4. Perform reliability screening, temperature normalization, load normalization, and long-term trend calculation on the dielectric loss factor of multiple consecutive sampling periods to obtain the dielectric loss trend value, and obtain the dielectric loss trend correction amount based on the dielectric loss trend value and the healthy benchmark dielectric loss factor.

[0011] S5. Calculate the equivalent conductivity correction per unit length based on the dielectric loss trend correction.

[0012] S6. Based on the conductivity frequency mapping coefficient obtained by fitting the calibration data of the same type of submarine power cable, the equivalent conductivity correction per unit length is mapped to a preset broadband detection frequency band to obtain the corrected equivalent conductivity per unit length within the preset broadband detection frequency band, and the corrected equivalent admittance per unit length is constructed.

[0013] S7. Calculate the corrected propagation constant based on the corrected unit length equivalent admittance and the preset unit length series impedance, and extract the phase propagation parameter from the corrected propagation constant as the corrected phase constant;

[0014] S8. Inject a broadband detection signal into the submarine power cable within the preset broadband detection frequency band, and collect the broadband voltage response and broadband current response at the port of the submarine power cable. Calculate the input impedance based on the broadband voltage response and the broadband current response, and calculate the reflection coefficient based on the input impedance.

[0015] S9. Extract the phase anomaly of the reflection coefficient corresponding to the local impedance anomaly from the reflection coefficient, and filter the frequency points where the phase anomaly of the reflection coefficient is located to obtain multiple effective frequency points.

[0016] S10. For the multiple effective frequency points, the defect distance is inverted by using the corrected phase constant and the phase anomaly of the reflection coefficient at the corresponding frequency to obtain the local defect distances corresponding to the multiple frequency points. The local defect distances corresponding to the multiple frequency points are then fused to obtain the final local defect distance.

[0017] Preferably, the continuous or near-continuous distributed grounding boundary refers to the following: within a preset monitoring length range Lm, the metal shielding layer, metal sheath, armor layer, outer sheath, or combination thereof form a potential constraint with the seawater, end grounding structure, or distributed discharge path along the length direction, so that the main insulation distributed branch current flows continuously or near-continuously into the equivalent ground of the metal shielding layer, metal sheath, armor layer, outer sheath, and seawater along the length direction of the submarine power cable; the near-continuous refers to the distance di between adjacent equivalent discharge paths, the equivalent discharge impedance Geq(i), or the equivalent potential difference Δφi satisfying a preset threshold, so that the corresponding cable segment can be processed according to the distributed parameter model; the preset threshold is determined based on the submarine power cable design parameters, outer sheath conductivity, seawater conductivity, armor layer discharge impedance, end grounding structure parameters, or initial commissioning calibration data.

[0018] Preferably, the direction unification includes: taking the direction of the current flowing into the submarine power cable conductor from the equipment in the first terminal station as the positive direction of the first current and taking the direction of the current flowing into the submarine power cable conductor from the equipment in the last terminal station as the positive direction of the last current, and unifying the directions of the first current phasor and the second current phasor.

[0019] The equivalent current phasor of the main insulation is obtained by adding the first corrected current phasor and the second corrected current phasor to form the residual current phasor, so that the sum of the current phasor of the first conductor loop after direction normalization and the current phasor of the last conductor loop under the ideal condition of ignoring the current of the main insulation distributed branch and the line loss approaches zero; the equivalent current phasor of the main insulation is the phasor composite value of the current of the main insulation distributed branch of the entire submarine power cable, and is not the current of the cross-interconnection grounding wire, the current of the metal sheath grounding wire, or the grounding box grounding current;

[0020] The voltage phase reference is either the voltage phase φu at one end or the equivalent voltage phase φeq of the main insulation after correction by the distributed parameter model of the submarine power cable. Based on the voltage phase reference, the equivalent current phasor of the main insulation is decomposed into an active loss current component Ir and a capacitive current component Ic, and the dielectric loss factor tanδk of the current sampling period is calculated. The dielectric loss factor tanδk of the current sampling period is determined by the ratio of the amplitude of the active loss current component Ir to the amplitude of the capacitive current component Ic. Wherein, Ir is the active loss current component in phase with the voltage phase reference, and Ic is the capacitive current component that leads or lags the voltage phase reference by a quarter of a period.

[0021] Preferably, in step S4, the reliability of the dielectric loss factor for multiple consecutive sampling periods is evaluated, and the dielectric loss factor that meets the preset reliability conditions is taken as the high reliability dielectric loss factor; the high reliability dielectric loss factor is normalized by temperature and load to obtain the high reliability normalized dielectric loss factor; the dielectric loss trend value is the moving average, exponentially weighted average, median, or robust regression trend value of the high reliability normalized dielectric loss factor within a continuous time window; the healthy benchmark dielectric loss factor is determined by the dielectric loss factor data that meets the preset reliability conditions during the initial commissioning of the submarine power cable, after maintenance, during the period when the main insulation status is confirmed to be normal by offline testing, or during historical healthy operation; the dielectric loss trend correction amount is the increment of the dielectric loss trend value relative to the healthy benchmark dielectric loss factor.

[0022] Preferably, the conductance frequency mapping coefficient is determined by the following calibration method:

[0023] Select calibration submarine power cable samples with the same or similar model, insulation material, conductor cross-section, insulation thickness, metal sheath structure, armor structure, and outer sheath structure as the submarine power cable to be tested. Under different overall moisture levels, different aging levels, or different dielectric loss levels, obtain the dielectric loss trend correction Δtanδq, the known local defect location xq, and the measured phase φmeas,q(ω) of the broadband reflection coefficient for the calibration submarine power cable samples. Using the conductance frequency mapping coefficient mG(ω) as the fitting parameter, construct the corrected unit length equivalent admittance ycorr,q(ω) and the corrected phase constant βcorr,q(ω) corresponding to the calibration sample based on Δtanδq, and calculate the broadband reflection coefficient model phase φmodel,q(ω) corresponding to the known local defect location xq. Determine mG(ω) by minimizing the weighted error between the measured phase and the model phase, where:

[0024] ;

[0025] Where q is the calibration sample number, W(ω) is the frequency weighting coefficient, and unwrap indicates phase expansion processing; the mG(ω) is represented by a piecewise constant, polynomial, spline function, lookup table interpolation function, or function based on the dielectric dispersion model of submarine power cable insulation material.

[0026] Preferably, the corrected equivalent conductance per unit length is calculated as follows:

[0027] ΔGδ=ω0C0Δtanδ;

[0028] Gcorr(ω)=Gbase(ω)+mG(ω)ΔGδ;

[0029] When Δtanδ is lower than the preset capacitance correction threshold, Ccorr(ω)=Cbase(ω); when Δtanδ is not lower than the preset capacitance correction threshold, Ccorr(ω)=Cbase(ω)+mC(ω)C0Δtanδ;

[0030] ycorr(ω)=Gcorr(ω)+jωCcorr(ω);

[0031] ;

[0032] γcorr(ω)=αcorr(ω)+jβcorr(ω);

[0033] Wherein, ΔGδ is the correction amount for equivalent conductance per unit length, ω0 is the angular frequency corresponding to dielectric loss factor monitoring, C0 is the main insulation capacitance per unit length, Δtanδ is the correction amount for dielectric loss trend, Gcorr(ω) is the corrected equivalent conductance per unit length, Gbase(ω) is the equivalent conductance per unit length under healthy reference conditions, mG(ω) is the conductance frequency mapping coefficient, ω is the angular frequency of the broadband detection signal, Ccorr(ω) is the corrected equivalent capacitance per unit length, Cbase(ω) is the equivalent capacitance per unit length under healthy reference conditions, mC(ω) is the capacitance frequency mapping coefficient, ycorr(ω) is the corrected equivalent admittance per unit length, j is the imaginary unit, zcorr(ω) is the series impedance per unit length, γcorr(ω) is the corrected propagation constant, αcorr(ω) is the corrected attenuation constant, and βcorr(ω) is the corrected phase constant.

[0034] Preferably, the broadband detection signal is injected at the onshore converter station-side terminal, the offshore booster station-side terminal, or an intermediate test port of the submarine power cable; the broadband detection signal is one or more of the following: a swept-frequency sine wave signal, a multi-sine wave signal, a pseudo-random sequence signal, a linear frequency modulation signal, or a step signal; the broadband detection signal is injected into the submarine power cable through a cable terminal coupling capacitor, a bushing tap, or a conductor port coupling device; the input impedance is determined by the broadband voltage response and the broadband current response, and the reflection coefficient is determined by the input impedance and the reference impedance.

[0035] Preferably, the reflection coefficient phase anomaly is obtained by subtracting the current reflection coefficient phase from the healthy baseline reflection coefficient phase, or by subtracting the current reflection coefficient phase from the reflection coefficient phase of the defect-free model. When extracting the reflection coefficient phase anomaly Δφ(ω), the reflection coefficient phase undergoes phase expansion, end reflection background subtraction, and phase smoothing. Only frequency points with a signal-to-noise ratio higher than a preset threshold, phase continuity meeting preset conditions, and repeatability consistency meeting preset conditions are selected as effective frequency points. For each effective frequency point, the defect distance is calculated using the following formula:

[0036]

[0037] The method calculates the final local defect distance x by weighted averaging, weighted median, robust regression, or consistent clustering of the local defect distances x(ω) corresponding to multiple effective frequency points. The method also calculates the positioning reliability Q, which is determined by one or more of the following: reliability of dielectric loss factor monitoring, stability of dielectric loss trend, repeatability of reflection coefficient phase anomaly, broadband impedance signal-to-noise ratio, spectral stability, end equipment status, seawater temperature fluctuation, conductor load fluctuation, or broadband signal coupling stability.

[0038] Preferably, the method further includes a time-based verification positioning step: calculating the verification defect distance based on the corrected equivalent propagation velocity determined by the corrected propagation constant and the time difference between the broadband detection signal propagating from the detection port to the local defect and reflecting back to the detection port; when the deviation between the verification defect distance and the final local defect distance is less than a preset distance threshold, the positioning reliability is improved; when the positioning reliability is lower than the preset threshold, the deterministic local defect distance is not output, but a prompt indicating that repeated detection is required or the positioning reliability is insufficient is output.

[0039] When the dielectric loss trend value tanδc continuously increases relative to the healthy baseline dielectric loss factor tanδbase and no stable local phase anomaly appears in the reflection coefficient phase, an overall insulation degradation or overall moisture in the submarine power cable is indicated. When the dielectric loss trend value tanδc does not continuously increase but a stable local phase anomaly appears in the reflection coefficient phase, a local structural defect, joint anomaly, or local anomaly in the sheath or shielding layer is indicated. When the dielectric loss trend value tanδc continuously increases and a stable local phase anomaly appears in the reflection coefficient phase, a local insulation defect in the submarine power cable and its location result are indicated.

[0040] The submarine cable broadband impedance defect location system based on dielectric loss correction described in this invention includes a dielectric loss online acquisition unit, a dielectric loss trend analysis unit, a mapping coefficient calibration unit, a parameter correction unit, a broadband signal injection unit, a broadband response acquisition unit, a reflection coefficient phase anomaly extraction unit, a defect distance inversion unit, and a diagnostic output unit.

[0041] The online dielectric loss acquisition unit includes a voltage phase acquisition device, a first-end current acquisition device, an end-end current acquisition device, and a synchronous sampling device. It is used to acquire the voltage phase of the phase conductor under test relative to the grounding reference point at one end of the submarine power cable, and to synchronously acquire the conductor loop current phasor at the first and end ends. After direction normalization, non-main insulation current component correction and amplitude and phase error correction, the main insulation equivalent current phasor is reconstructed from the small margin of the first and end conductor loop current phasors and the dielectric loss factor is calculated.

[0042] The dielectric loss trend analysis unit is used to perform reliability screening, temperature normalization, load normalization, and long-term trend calculation on the dielectric loss factor of multiple consecutive sampling periods to obtain the dielectric loss trend correction amount.

[0043] The mapping coefficient calibration unit is used to fit the conductivity frequency mapping coefficient mG(ω) based on the calibration data of the same type of submarine power cable, with the goal of minimizing the error between the measured reflection coefficient phase and the model reflection coefficient phase.

[0044] The parameter correction unit is used to calculate the corrected equivalent conductance per unit length based on the dielectric loss trend correction amount and the conductance frequency mapping coefficient mG(ω), construct the corrected equivalent admittance per unit length, and calculate the corrected propagation constant and the corrected phase constant.

[0045] The broadband signal injection unit includes a cable terminal coupling capacitor, a bushing tap, or a conductor port coupling device, used to inject broadband detection signals into submarine power cables.

[0046] The broadband response acquisition unit is used to acquire the broadband voltage response and broadband current response of the submarine power cable port;

[0047] The reflection coefficient phase anomaly extraction unit is used to calculate the input impedance and reflection coefficient based on the wideband voltage response and wideband current response, and to extract the reflection coefficient phase anomaly quantity corresponding to the local impedance anomaly.

[0048] The defect distance inversion unit is used to invert the defect distance by using a corrected phase constant to assess the phase anomaly of the reflection coefficient.

[0049] The diagnostic output unit is used to output local defect distance, overall insulation degradation indication, local structural defect indication, or location reliability.

[0050] Beneficial Effects: Compared with the prior art, the present invention has the following significant advantages: 1. The present invention synchronously collects the current in the conductor loops at the beginning and end of the submarine cable and normalizes its direction. It then uses the small margin from the sum of the two to reconstruct the equivalent current of the entire main insulation section and obtains the online dielectric loss factor. This provides a dynamic correction basis for the phase propagation parameters in the broadband impedance positioning model, effectively alleviating the problem of decreased positioning accuracy caused by the mismatch of the fixed phase constant due to changes in overall dielectric loss. 2. The present invention introduces a conductivity frequency mapping coefficient to map the low-frequency equivalent conductivity change obtained based on the dielectric loss trend correction to a preset broadband detection frequency band. This constructs a physically consistent corrected unit length equivalent admittance and propagation constant, improving the phase propagation parameters within the broadband band. 3. After extracting the phase anomaly of the reflection coefficient, this invention filters the effective frequency points and uses the corrected phase constant at the corresponding frequency to perform defect distance inversion. Then, by fusing the positioning results of multiple frequency points, the final distance is output, which effectively suppresses the interference of single-frequency noise and abnormal disturbances on the positioning results, and enhances the anti-interference ability and robustness of local defect positioning. 4. By simultaneously analyzing the long-term trend of dielectric loss and the phase anomaly pattern of the reflection coefficient, this invention can effectively distinguish between the overall insulation degradation and local structural defects of submarine cables. Combined with the quantitative index of positioning reliability, the output strategy is judged, avoiding the output of unreliable positioning results when the data quality is insufficient, and improving the comprehensive condition diagnosis capability of cables. Attached Figure Description

[0051] Figure 1 This is a schematic diagram of the method flow of the present invention;

[0052] Figure 2 This is a structural block diagram of the submarine power cable defect location system of the present invention;

[0053] Figure 3 This is a schematic diagram of the continuous or near-continuous distributed grounding boundary of the submarine power cable of the present invention;

[0054] Figure 4 This is a schematic diagram illustrating the principle of online acquisition of dielectric loss factor in submarine power cable conductor circuits according to the present invention.

[0055] Figure 5 This is a schematic diagram of the calibration process for the conductivity frequency mapping coefficient mG(ω) of this invention;

[0056] Figure 6 This is a schematic diagram illustrating the principle of phase anomaly localization based on the modified phase constant βcorr(ω) of the reflection coefficient in this invention.

[0057] Figure 7 This is a schematic diagram of the segmented calibration results of the conductivity frequency mapping coefficient mG(ω) of the present invention. Detailed Implementation

[0058] The technical solution of the present invention will be further described below with reference to the accompanying drawings.

[0059] like Figure 3 As shown, the object under test is a submarine power cable laid in a seawater environment. The cable includes a conductor, main insulation, metallic shield, metallic sheath, armor layer, outer sheath, or semi-conductive layer. The metallic shield, metallic sheath, armor layer, outer sheath, or combinations thereof can form a continuous or nearly continuous distributed grounding boundary through seawater, the outer sheath discharge path, the armor layer discharge path, the end grounding structure, the distributed conductivity path, or the distributed capacitance coupling path. Within a preset monitoring length range Lm, when the distance di of adjacent equivalent discharge paths, the equivalent discharge impedance Geq(i), or the adjacent equivalent potential difference Δφi meets the threshold, this section of cable can be treated according to the distributed parameter model. Figure 3 In this context, Ceq represents the equivalent coupling capacitance, Geq represents the equivalent discharge conductance, di represents the distance between adjacent equivalent discharge paths, Δφi represents the difference between adjacent equivalent potentials, and dth and Δφth represent the corresponding threshold values, respectively.

[0060] like Figure 1 and Figure 4 As shown, the voltage phase φu of the conductor under test relative to the grounding reference point is first acquired at either the beginning or end of the submarine power cable. Simultaneously, the loop current phasors I1 and I2 of the conductor under test are acquired at both the beginning and end. The acquisition locations for the loop current phasors are located in the conductor loop or the terminal outgoing loop of the submarine power cable. The positive direction of the current at the beginning is defined as the flow from the equipment within the beginning station into the cable conductor, and the positive direction of the current at the end is defined as the flow from the equipment within the end station into the cable conductor; I1 and I2 are then normalized according to this definition.

[0061] After directional normalization, I1 and I2 are corrected for non-main insulation current components, including grounding wire current, sheath current, grounding box current, or other components unrelated to the main insulation distribution branches. Subsequently, amplitude and phase error correction is performed to eliminate current acquisition channel gain error, phase deviation, and synchronization sampling error, yielding I1′ and I2′. Under ideal conditions where the main insulation distribution branch current and line losses are ignored, the sum of the phasors of the first and last conductor loop currents after directional normalization should approach zero. Therefore, their small margin can be used to characterize the combined value of the main insulation distribution branch currents.

[0062] The equivalent current phasor of the main insulation is reconstructed by the following formula:

[0063] Iins = I1′ + I2′

[0064] Wherein, Iins is the equivalent current phasor of the main insulation of the entire submarine power cable, not the cross-interconnection grounding current, sheath grounding current, or grounding box current. Using the voltage phase φu at one end or the equivalent active voltage phase φeq of the main insulation corrected by the distributed parameter model as the voltage phase reference, Iins is decomposed into an active power loss current component Ir and a capacitive current component Ic, and the dielectric loss factor for the current sampling period is calculated using the following formula:

[0065] tanδk=|Ir| / |Ic|

[0066] Where Ir is used to characterize the active power loss current of the main insulation, and Ic is used to characterize the capacitance current of the main insulation. For example... Figure 4 As shown in the lower section, trend processing is performed on tanδk for continuous sampling periods, including confidence screening, temperature normalization, load normalization, outlier removal, and long-term trend calculation. Confidence screening can be based on signal-to-noise ratio, amplitude stability, phase consistency, channel status, and synchronization status; temperature normalization is used to compensate for the influence of seawater temperature, cable conductor temperature, or ambient temperature on tanδ; load normalization is used to compensate for the influence of conductor load and heating state changes on tanδ; outlier removal is used to eliminate isolated points and abrupt changes. After the above processing, the dielectric loss trend value tanδc is obtained.

[0067] Δtanδ=tanδc-tanδbase

[0068] Wherein, tanδbase is the health baseline dielectric loss factor, which can be determined from high-reliability data during the initial operation of the cable, after maintenance, during offline testing confirming the normal condition of the main insulation, or during historical healthy operation. Δtanδ is the dielectric loss trend correction, used to describe the change in the current overall dielectric loss level relative to the health baseline.

[0069] exist Figure 1 S5 and Figure 6In the correction process, Δtanδ is first converted into a correction amount for the equivalent conductance per unit length:

[0070] ΔGδ=ω0C0Δtanδ

[0071] Where ω0 is the angular frequency corresponding to dielectric loss factor monitoring, for example, ω0 = 2π × 50 rad / s for power frequency monitoring; C0 is the main insulation capacitance per unit length. Since the online dielectric loss monitoring frequency and the broadband impedance detection frequency band are different, it is necessary to map ΔGδ to the broadband detection frequency band through the conductivity frequency mapping coefficient mG(ω).

[0072] like Figure 5 As shown, the calibration of mG(ω) can be carried out according to the following procedure: Select calibration samples that are the same as or similar to the submarine power cable under test in terms of model, insulation material, conductor cross-section, insulation thickness, metal sheath structure, armor structure, and outer sheath structure; construct sample working conditions by different overall moisture levels, different aging levels, or different dielectric loss levels; obtain the dielectric loss trend correction Δtanδq, the known local defect location xq, and the measured phase φmeas,q(ω) of the reflection coefficient obtained after injecting broadband detection signals for each calibration sample; set the representation method of mG(ω), such as piecewise constant, polynomial, spline function, lookup table interpolation function, or function based on the dielectric dispersion model of insulation material; then construct Gcorr,q(ω), ycorr,q(ω), and γcorr,q(ω) based on the candidate mG(ω), calculate the model phase φmodel,q(ω) corresponding to the known local defect location xq, and fit it with the goal of minimizing the weighted phase error:

[0073] mG(ω)=argmin ΣqΣω W(ω)|unwrap[φmeas,q(ω)-φmodel,q(ω)]|²

[0074] Where q is the calibration sample number, W(ω) is the frequency weighting coefficient, and unwrap indicates phase expansion processing. For frequency bands with high signal-to-noise ratio and good repeatability, W(ω) can be taken as a larger value; for frequency bands with high noise or phase discontinuity, W(ω) can be taken as a smaller value or set to zero. After calibration, mG(ω) can be stored as a frequency mapping table, a piecewise constant table, or a continuous curve.

[0075] In a segmented constant calibration embodiment, the calibration frequency range is 10kHz to 3MHz, divided into four frequency bands: F1 is 10kHz ≤ f < 50kHz, F2 is 50kHz ≤ f < 200kHz, F3 is 200kHz ≤ f < 1MHz, and F4 is 1MHz ≤ f ≤ 3MHz. Within each frequency band, mG(ω) takes values ​​of m1, m2, m3, and m4, respectively. Figure 7 As shown in Table 1, a set of typical calibration results are presented.

[0076] Table 1

[0077] frequency band Frequency range Values ​​of mG(ω) Calibration of root mean square residual phase error F1 10kHz≤f<50kHz m1＝0.82 4.8° F2 50kHz≤f<200kHz m2＝0.67 3.9° F3 200kHz≤f<1MHz m3＝0.41 4.5° F4 1MHz≤f≤3MHz m4＝0.18 6.2°

[0078] The above results indicate that in lower bandwidth detection bands, the effect of dielectric loss variation on equivalent conductance and phase constant is more pronounced at operating frequencies. As frequency increases, mG(ω) gradually decreases, suggesting that the equivalent effect of power frequency dielectric loss variation on high-frequency propagation phase is partially weakened by frequency dispersion, liner coupling, and high-frequency loss mechanisms. In another embodiment, mG(ω) can be expressed as a second-order polynomial with respect to the logarithm of frequency:

[0079] mG(ω)=a0+a1log10(f / f0)+a2[log10(f / f0)]²

[0080] Where f is the frequency, f0 is the reference frequency, and a0, a1, and a2 are coefficients obtained by fitting calibration data. This method is suitable for wideband sweep detection scenarios that require continuous interpolation of frequency points.

[0081] like Figure 6 As shown in the correction process, in the broadband impedance positioning model of submarine power cables, the series impedance per unit length and the parallel admittance per unit length can be expressed as:

[0082] z(ω)=R(ω)+jωL(ω)

[0083] y(ω)=G(ω)+jωC(ω)

[0084] At the dielectric loss factor monitoring frequency ω0, the equivalent conductance per unit length and the dielectric loss factor can be expressed as:

[0085] Gδ=ω0C0tanδ

[0086] When the dielectric loss factor changes relative to the healthy baseline value, the equivalent conductance per unit length correction is:

[0087] ΔGδ=ω0C0(tanδc-tanδbase)

[0088] Using the calibrated mG(ω), the corrected equivalent conductance per unit length in the wideband detection band is:

[0089] Gcorr(ω)=Gbase(ω)+mG(ω)ΔGδ

[0090] When Δtanδ is small and the overall dielectric constant change is negligible, Ccorr(ω) = Cbase(ω). When Δtanδ is large and the dielectric constant change caused by moisture, aging, or water tree development cannot be ignored, the equivalent capacitance per unit length can be further modified.

[0091] Ccorr(ω)=Cbase(ω)+mC(ω)C0Δtanδ

[0092] Where Cbase(ω) is the equivalent capacitance per unit length under healthy baseline conditions, and mC(ω) is the capacitance frequency mapping coefficient. Therefore, the corrected equivalent admittance per unit length is constructed:

[0093] ycorr(ω)=Gcorr(ω)+jωCcorr(ω)

[0094] Calculate the corrected propagation constant by considering the series impedance per unit length zcorr(ω):

[0095] γcorr(ω)=√[zcorr(ω)·ycorr(ω)]

[0096] γcorr(ω)=αcorr(ω)+jβcorr(ω)

[0097] Where αcorr(ω) is the corrected attenuation constant and βcorr(ω) is the corrected phase constant. Subsequent defect distance inversion uses βcorr(ω) instead of the fixed phase constant βbase(ω) under healthy conditions.

[0098] like Figure 1 S8 Figure 2 Wideband signal injection unit and wideband response acquisition unit and Figure 6 As shown in the detection port diagram, the system can inject a broadband detection signal at the onshore converter station terminal, the offshore booster station terminal, or an intermediate test port on the submarine power cable. The broadband detection signal can be a swept-frequency sine wave, multiple sine waves, a pseudo-random sequence signal, a linear frequency modulated signal, or a step signal, and can be injected into the cable through cable terminal coupling capacitors, bushing taps, or conductor port coupling devices. The broadband response acquisition unit acquires the broadband voltage response U(ω) and broadband current response I(ω) at the port and calculates the input impedance.

[0099] Zin(ω)=U(ω) / I(ω)

[0100] Further calculation of the reflection coefficient:

[0101] Γ(ω)=[Zin(ω)-Zref] / [Zin(ω)+Zref]

[0102] Where Zref is the reference impedance. For implementations using incident and reflected wave decomposition, the reflection coefficient phase can also be obtained based on the ratio of the reflected signal Vr(ω) to the injected signal U(ω). The system compares the current reflection coefficient phase with the healthy reference reflection coefficient phase, or with the reflection coefficient phase of a defect-free model, to obtain the reflection coefficient phase anomaly Δφ(ω) corresponding to the local impedance anomaly.

[0103] like Figure 6 As shown in the phase anomaly curve in the lower left, when extracting Δφ(ω), the phase of the reflection coefficient is subjected to phase expansion processing to eliminate discontinuities caused by ±π jumps; end reflections, coupling device reflections, and background reflections are subtracted; and the phase curve is smoothed or robustly filtered to reduce the impact of noise. To improve positioning stability, only frequency points with a signal-to-noise ratio higher than a preset threshold, phase continuity meeting preset conditions, and repeatability consistency meeting preset conditions are selected as effective frequency points.

[0104] For each effective frequency point, the defect distance is calculated using the following formula:

[0105] x(ω)=Δφ(ω) / [2βcorr(ω)]

[0106] In this calculation, the 2 in the denominator represents the round-trip propagation of the broadband detection signal from the detection port to the local defect and then back to the detection port. When defect distance results are obtained at multiple frequency points, the final defect distance x can be obtained by weighted averaging, weighted median, robust regression, or consistent clustering. Figure 6 The lower right section shows a comparison of the feedback distance using a fixed βbase(ω) and a modified βcorr(ω): using βcorr(ω) can reduce the deviation between the inverted location and the actual defect location.

[0107] In an alternative implementation, the verification location can also be based on the reflection time. The corrected equivalent propagation velocity vcorr is determined according to the corrected propagation constant, and the verification defect distance is calculated using the following formula:

[0108] xv=vcorrΔt / 2

[0109] Where Δt is the time difference between the broadband detection signal propagating from the detection port to the local defect and reflecting back to the detection port. When the deviation between xv and x is less than a preset distance threshold, the positioning reliability Q is increased. The positioning reliability Q can be determined by one or more of the following: dielectric loss factor monitoring reliability, dielectric loss trend stability, reflection coefficient phase anomaly repeatability, broadband impedance signal-to-noise ratio, spectral stability, end device status, seawater temperature fluctuation, conductor load fluctuation, or broadband signal coupling stability. When Q is lower than the preset threshold, the deterministic local defect distance is not output, but a prompt indicating that repeated detection is required or the positioning reliability is insufficient is output.

[0110] like Figure 1 End output box and Figure 2 As shown in the diagnostic output unit, the system outputs diagnostic results based on tanδc, reflection coefficient phase anomaly characteristics, and location reliability. When tanδc continuously increases relative to tanδbase, but no stable local anomaly appears in the reflection coefficient phase, it outputs a warning of overall insulation degradation or overall moisture. When tanδc does not continuously increase, but a stable local anomaly appears in the reflection coefficient phase, it outputs a warning of local structural defects, joint anomalies, or local anomalies in the sheath or shielding layer. When tanδc continuously increases and a stable local anomaly appears in the reflection coefficient phase, it outputs a warning of local insulation defects and their locations. When tanδc shows a single anomaly but the dielectric loss reliability is low, and no stable local anomaly appears in the reflection coefficient phase, it outputs a warning of measurement disturbance, changes in the status of end equipment, or the need for repeated testing.

[0111] like Figure 2 As shown, the system includes an online dielectric loss acquisition unit, a dielectric loss trend analysis unit, a mapping coefficient calibration unit, a parameter correction unit, a broadband signal injection unit, a broadband response acquisition unit, a reflection coefficient phase anomaly extraction unit, a defect distance inversion unit, and a diagnostic output unit. The online dielectric loss acquisition unit receives health benchmark / historical database and end-to-end acquisition data; the mapping coefficient calibration unit receives calibration data from the same type of submarine power cable and outputs mG(ω); the parameter correction unit converts Δtanδ to ΔGδ and generates Gcorr(ω), ycorr(ω), and βcorr(ω); the broadband signal injection unit injects a broadband detection signal into the cable; the broadband response acquisition unit acquires U(ω) and I(ω); the reflection coefficient phase anomaly extraction unit outputs Δφ(ω); the defect distance inversion unit obtains the defect distance based on x(ω)=Δφ(ω) / [2βcorr(ω)]; and the diagnostic output unit outputs local defect distance, overall insulation degradation indication, local structural defect indication, and location reliability.

[0112] In an embodiment of a 220kV cross-linked polyethylene insulated submarine power cable, the cable length is 35km and the rated frequency is 50Hz. This submarine power cable is laid in a seawater environment. The metal sheath, metal shield, armor layer, outer conductive layer, or a combination thereof form a continuous or nearly continuous distributed grounding boundary through seawater, the outer sheath discharge path, the armor layer discharge path, the end grounding structure, the distributed conductivity path, or the distributed capacitance coupling path, and no cross-interconnection grounding structure is used. During the initial commissioning phase, the health baseline dielectric loss factor tanδbase = 2.6 × 10⁻⁴ was determined based on the ledger and type test report. After a period of operation, through small margin reconstruction of the current in the first and last conductor loops and long-term trend calculation, tanδc = 1.5 × 10⁻³ was obtained within the current trend window; therefore, Δtanδ = 1.24 × 10⁻³. The unit length main insulation capacitance C0 = 0.18μF / km, ω0 = 2π × 50rad / s, then:

[0113] ΔGδ=ω0C0Δtanδ=2π×50×0.18×10^-6×1.24×10^-3S / km=7.01×10^-8S / km

[0114] Based on the calibration data of the same type of submarine power cable, mG(ω) = 0.67 in the 50kHz to 200kHz frequency band. The corrected equivalent conductance increment per unit length is mG(ω)ΔGδ = 0.67 × 7.01 × 10^-8 S / km = 4.70 × 10^-8 S / km. The system then constructs ycorr(ω) and calculates γcorr(ω) and βcorr(ω). Subsequently, a 10kHz to 3MHz sweep frequency signal is injected into the cable, and the port broadband voltage response and broadband current response are collected and Γ(ω) is calculated. Stable phase anomaly characteristics are detected in the Γ(ω) phase curve. When calculated using the fixed phase constant βbase(ω) under healthy conditions, the defect location is 18.6km; when calculated using the corrected βcorr(ω), the defect location is 17.9km. Combining construction data and subsequent investigation, a local mechanical stress concentration area exists near 17.8km. The location errors of the two methods are shown in Table 2.

[0115] Table 2

[0116] method Location results Real location relative error Fixed phase constant 18.6km 17.8km 4.49% Corrected phase constant 17.9km 17.8km 0.56%

[0117] Location errors can also be compared under different dielectric loss levels. A sample of the same type of submarine power cable was selected, with the actual location of the local defect at 20.00 km. The wideband detection frequency range was 10 kHz to 3 MHz. Figure 7 The results of the piecewise constant calibration of mG(ω) are shown in Table 3.

[0118] Table 3

[0119] Operating conditions tanδc Δtanδ Uncorrected location results Uncorrected error Corrected positioning results Corrected error Health benchmarks 2.3×10^-4 0 20.08km 0.40% 20.05km 0.25% Mild aging 4.6×10^-4 2.3×10^-4 20.65km 3.25% 20.22km 1.10% Moderate aging 1.0×10^-3 7.7×10^-4 21.20km 6.00% 20.48km 2.40% Severe aging 2.0×10^-3 1.77×10^-3 21.85km 9.25% 20.72km 3.60%

[0120] As can be seen from the above results, as the dielectric loss factor increases, the positioning error using a fixed phase constant under healthy conditions gradually increases; after using a modified phase constant, the positioning error is significantly reduced, indicating that the scheme can compensate for the impact of changes in the overall dielectric loss of submarine power cables on broadband impedance phase propagation parameters.

[0121] In the embodiment where the number of calibration samples affects mG(ω), the true location of the local defect is 15.01 km, the dielectric loss trend value of the submarine power cable under test is tanδc = 1.3 × 10^-3, and the health baseline is tanδbase = 2.4 × 10^-4. The corrected location results under different calibration sample numbers and representation methods are shown in Table 4.

[0122] Table 4

[0123] Standardized sample size mG(ω) representation Corrected positioning results relative error 2 groups Piecewise constant 15.42km 2.80% 4 groups Piecewise constant 15.20km 1.33% 6 groups Piecewise constant 15.12km 0.80% 6 groups Second-order polynomial 15.15km 1.00%

[0124] Therefore, as the number of calibration samples increases, the fitting result of mG(ω) becomes more stable, and the corrected positioning error decreases. When the number of samples meets the requirements of the field or experiment, piecewise constants, lookup table interpolation, or polynomial functions can all effectively correct the broadband phase propagation parameters.

Claims

1. A broadband impedance defect location method for submarine cables based on dielectric loss correction, characterized in that, The method is applied to submarine power cables laid in a seawater environment. The metallic shielding layer, metallic sheath, armor layer, outer sheath, or a combination thereof of the submarine power cable form a continuous or nearly continuous distributed grounding boundary through seawater, the outer sheath discharge path, the armor layer discharge path, the end grounding structure, the distributed conductivity path, or the distributed capacitance coupling path. The method includes the following steps: S1. At either the beginning or end of the submarine power cable, the voltage phase of the phase conductor under test relative to the grounding reference point is collected, and the loop current phasor of the phase conductor under test is simultaneously collected at the beginning of the submarine power cable as the first current phasor, and the loop current phasor of the phase conductor under test is simultaneously collected at the end of the submarine power cable as the second current phasor. S2. Perform direction normalization, non-main insulation current component correction and amplitude-phase error correction on the first current phasor and the second current phasor to obtain the first corrected current phasor and the second corrected current phasor. S3. Add the first corrected current phasor and the second corrected current phasor to obtain the remaining current phasor, and use the remaining current phasor as the equivalent current phasor of the main insulation of the submarine power cable. Calculate the dielectric loss factor based on the equivalent current phasor of the main insulation and the voltage phase reference. The dielectric loss factor is used to characterize the overall dielectric loss change of the main insulation of the submarine power cable and to correct the phase propagation parameters in the broadband impedance positioning model. S4. Perform reliability screening, temperature normalization, load normalization, and long-term trend calculation on the dielectric loss factor of multiple consecutive sampling periods to obtain the dielectric loss trend value, and obtain the dielectric loss trend correction amount based on the dielectric loss trend value and the healthy benchmark dielectric loss factor. S5. Calculate the equivalent conductivity correction per unit length based on the dielectric loss trend correction. S6. Based on the conductivity frequency mapping coefficient obtained by fitting the calibration data of the same type of submarine power cable, the equivalent conductivity correction per unit length is mapped to a preset broadband detection frequency band to obtain the corrected equivalent conductivity per unit length within the preset broadband detection frequency band, and the corrected equivalent admittance per unit length is constructed. S7. Calculate the corrected propagation constant based on the corrected unit length equivalent admittance and the preset unit length series impedance, and extract the phase propagation parameter from the corrected propagation constant as the corrected phase constant; S8. Inject a broadband detection signal into the submarine power cable within the preset broadband detection frequency band, and collect the broadband voltage response and broadband current response at the port of the submarine power cable. Calculate the input impedance based on the broadband voltage response and the broadband current response, and calculate the reflection coefficient based on the input impedance. S9. Extract the phase anomaly of the reflection coefficient corresponding to the local impedance anomaly from the reflection coefficient, and filter the frequency points where the phase anomaly of the reflection coefficient is located to obtain multiple effective frequency points. S10. For the multiple effective frequency points, the defect distance is inverted by using the corrected phase constant and the phase anomaly of the reflection coefficient at the corresponding frequency to obtain the local defect distances corresponding to the multiple frequency points. The local defect distances corresponding to the multiple frequency points are then fused to obtain the final local defect distance.

2. The method according to claim 1, characterized in that, The continuous or near-continuous distributed grounding boundary refers to the following: within a preset monitoring length range Lm, the metal shielding layer, metal sheath, armor layer, outer sheath, or combination thereof form a potential constraint with seawater, end grounding structure, or distributed discharge path along the length direction, so that the main insulation distributed branch current flows continuously or near-continuously into the equivalent ground of the metal shielding layer, metal sheath, armor layer, outer sheath, and seawater along the length direction of the seabed power cable; the near-continuous refers to the distance di between adjacent equivalent discharge paths, the equivalent discharge impedance Geq(i), or the equivalent potential difference Δφi satisfying a preset threshold, so that the corresponding cable segment can be processed according to the distributed parameter model; The preset threshold is determined based on the submarine power cable design parameters, outer sheath conductivity, seawater conductivity, armor layer discharge impedance, end grounding structure parameters, or initial commissioning calibration data.

3. The method according to claim 1, characterized in that, The direction unification includes: taking the direction of the current flowing into the submarine power cable conductor from the equipment in the first terminal station as the positive direction of the current at the first terminal and taking the direction of the current flowing into the submarine power cable conductor from the equipment in the last terminal station as the positive direction of the current at the last terminal, and unifying the directions of the first current phasor and the second current phasor. The equivalent current phasor of the main insulation is obtained by adding the first corrected current phasor and the second corrected current phasor to form the residual current phasor, so that the sum of the current phasor of the first conductor loop after direction normalization and the current phasor of the last conductor loop under the ideal condition of ignoring the current of the main insulation distributed branch and the line loss approaches zero; the equivalent current phasor of the main insulation is the phasor composite value of the current of the main insulation distributed branch of the entire submarine power cable, and is not the current of the cross-interconnection grounding wire, the current of the metal sheath grounding wire, or the grounding box grounding current; The voltage phase reference is either the voltage phase φu at one end or the equivalent voltage phase φeq of the main insulation after correction by the distributed parameter model of the submarine power cable. Based on the voltage phase reference, the equivalent current phasor of the main insulation is decomposed into an active loss current component Ir and a capacitive current component Ic, and the dielectric loss factor tanδk of the current sampling period is calculated. The dielectric loss factor tanδk of the current sampling period is determined by the ratio of the amplitude of the active loss current component Ir to the amplitude of the capacitive current component Ic. Wherein, Ir is the active loss current component in phase with the voltage phase reference, and Ic is the capacitive current component that leads or lags the voltage phase reference by a quarter of a period.

4. The method according to claim 1, characterized in that, In step S4, the reliability of the dielectric loss factor for multiple consecutive sampling periods is evaluated, and the dielectric loss factor that meets the preset reliability conditions is taken as the high reliability dielectric loss factor. The high-confidence dielectric loss factor is normalized by temperature and load to obtain the high-confidence normalized dielectric loss factor. The media loss trend value is the moving average, exponentially weighted average, median or robust regression trend value of the high-confidence normalized media loss factor within a continuous time window. The health baseline dielectric loss factor is determined by dielectric loss factor data that meet the preset confidence conditions during the initial commissioning of submarine power cables, after maintenance, during offline testing confirming normal main insulation status, or during historical healthy operation. The dielectric loss trend correction is the increment of the dielectric loss trend value relative to the healthy baseline dielectric loss factor.

5. The method according to claim 1, characterized in that, The conductivity frequency mapping coefficient is determined by the following calibration method: Select calibration submarine power cable samples with the same or similar model, insulation material, conductor cross-section, insulation thickness, metal sheath structure, armor structure, and outer sheath structure as the submarine power cable to be tested. Under different overall moisture levels, different aging levels, or different dielectric loss levels, obtain the dielectric loss trend correction Δtanδq, the known local defect location xq, and the measured phase φmeas,q(ω) of the broadband reflection coefficient for the calibration submarine power cable samples. Using the conductance frequency mapping coefficient mG(ω) as the fitting parameter, construct the corrected unit length equivalent admittance ycorr,q(ω) and the corrected phase constant βcorr,q(ω) corresponding to the calibration sample based on Δtanδq, and calculate the broadband reflection coefficient model phase φmodel,q(ω) corresponding to the known local defect location xq. Determine mG(ω) by minimizing the weighted error between the measured phase and the model phase, where: ； Where q is the calibration sample number, W(ω) is the frequency weighting coefficient, and unwrap indicates phase expansion processing; the mG(ω) is represented by a piecewise constant, polynomial, spline function, lookup table interpolation function, or function based on the dielectric dispersion model of submarine power cable insulation material.

6. The method according to claim 1, characterized in that, The corrected equivalent conductance per unit length is calculated as follows: ΔGδ=ω0C0Δtanδ; Gcorr(ω)=Gbase(ω)+mG(ω)ΔGδ; When Δtanδ is lower than the preset capacitance correction threshold, Ccorr(ω)=Cbase(ω); when Δtanδ is not lower than the preset capacitance correction threshold, Ccorr(ω)=Cbase(ω)+mC(ω)C0Δtanδ; ycorr(ω)=Gcorr(ω)+jωCcorr(ω); ； γcorr(ω)=αcorr(ω)+jβcorr(ω); Wherein, ΔGδ is the correction amount for equivalent conductance per unit length, ω0 is the angular frequency corresponding to dielectric loss factor monitoring, C0 is the main insulation capacitance per unit length, Δtanδ is the correction amount for dielectric loss trend, Gcorr(ω) is the corrected equivalent conductance per unit length, Gbase(ω) is the equivalent conductance per unit length under healthy reference conditions, mG(ω) is the conductance frequency mapping coefficient, ω is the angular frequency of the broadband detection signal, Ccorr(ω) is the corrected equivalent capacitance per unit length, Cbase(ω) is the equivalent capacitance per unit length under healthy reference conditions, mC(ω) is the capacitance frequency mapping coefficient, ycorr(ω) is the corrected equivalent admittance per unit length, j is the imaginary unit, zcorr(ω) is the series impedance per unit length, γcorr(ω) is the corrected propagation constant, αcorr(ω) is the corrected attenuation constant, and βcorr(ω) is the corrected phase constant.

7. The method according to claim 1, characterized in that, The broadband detection signal is injected at the onshore converter station-side terminal, the offshore booster station-side terminal, or an intermediate test port of the submarine power cable; the broadband detection signal is one or more of the following: swept-frequency sine wave signal, multiple sine wave signal, pseudo-random sequence signal, linear frequency modulation signal, or step signal; the broadband detection signal is injected into the submarine power cable through a cable terminal coupling capacitor, bushing tap, or conductor port coupling device; the input impedance is determined by the broadband voltage response and the broadband current response, and the reflection coefficient is determined by the input impedance and the reference impedance.

8. The method according to claim 1, characterized in that, The reflection coefficient phase anomaly is obtained by subtracting the current reflection coefficient phase from the healthy baseline reflection coefficient phase, or by subtracting the current reflection coefficient phase from the reflection coefficient phase of the defect-free model. When extracting the reflection coefficient phase anomaly Δφ(ω), the reflection coefficient phase undergoes phase expansion, end reflection background subtraction, and phase smoothing. Only frequency points with a signal-to-noise ratio higher than a preset threshold, phase continuity meeting preset conditions, and repeatability consistency meeting preset conditions are selected as effective frequency points. For each effective frequency point, the defect distance is calculated using the following formula: The method calculates the final local defect distance x by weighted averaging, weighted median, robust regression, or consistent clustering of the local defect distances x(ω) corresponding to multiple effective frequency points. The method also calculates the positioning reliability Q, which is determined by one or more of the following: reliability of dielectric loss factor monitoring, stability of dielectric loss trend, repeatability of reflection coefficient phase anomaly, broadband impedance signal-to-noise ratio, spectral stability, end equipment status, seawater temperature fluctuation, conductor load fluctuation, or broadband signal coupling stability.

9. The method according to claim 1, characterized in that, It also includes a time-based verification and positioning step: calculating the verification defect distance based on the corrected equivalent propagation velocity determined by the corrected propagation constant and the time difference between the broadband detection signal propagating from the detection port to the local defect and reflecting back to the detection port; when the deviation between the verification defect distance and the final local defect distance is less than a preset distance threshold, the positioning reliability is improved; when the positioning reliability is lower than the preset threshold, the deterministic local defect distance is not output, but a prompt indicating that repeated detection is required or the positioning reliability is insufficient is output. When the dielectric loss trend value tanδc continuously increases relative to the healthy baseline dielectric loss factor tanδbase and no stable local phase anomaly appears in the reflection coefficient phase, an overall insulation degradation or overall moisture in the submarine power cable is indicated. When the dielectric loss trend value tanδc does not continuously increase but a stable local phase anomaly appears in the reflection coefficient phase, a local structural defect, joint anomaly, or local anomaly in the sheath or shielding layer is indicated. When the dielectric loss trend value tanδc continuously increases and a stable local phase anomaly appears in the reflection coefficient phase, a local insulation defect in the submarine power cable and its location result are indicated.

10. A broadband impedance defect location system for submarine cables based on dielectric loss correction, characterized in that, The system for implementing the method according to any one of claims 1-9 includes an online dielectric loss acquisition unit, a dielectric loss trend analysis unit, a mapping coefficient calibration unit, a parameter correction unit, a broadband signal injection unit, a broadband response acquisition unit, a reflection coefficient phase anomaly extraction unit, a defect distance inversion unit, and a diagnostic output unit. The online dielectric loss acquisition unit includes a voltage phase acquisition device, a first-end current acquisition device, an end-end current acquisition device, and a synchronous sampling device. It is used to acquire the voltage phase of the phase conductor under test relative to the grounding reference point at one end of the submarine power cable, and to synchronously acquire the conductor loop current phasor at the first and end ends. After direction normalization, non-main insulation current component correction and amplitude and phase error correction, the main insulation equivalent current phasor is reconstructed from the small margin of the first and end conductor loop current phasors and the dielectric loss factor is calculated. The dielectric loss trend analysis unit is used to perform reliability screening, temperature normalization, load normalization, and long-term trend calculation on the dielectric loss factor of multiple consecutive sampling periods to obtain the dielectric loss trend correction amount. The mapping coefficient calibration unit is used to fit the conductivity frequency mapping coefficient mG(ω) based on the calibration data of the same type of submarine power cable, with the goal of minimizing the error between the measured reflection coefficient phase and the model reflection coefficient phase. The parameter correction unit is used to calculate the corrected equivalent conductance per unit length based on the dielectric loss trend correction amount and the conductance frequency mapping coefficient mG(ω), construct the corrected equivalent admittance per unit length, and calculate the corrected propagation constant and the corrected phase constant. The broadband signal injection unit includes a cable terminal coupling capacitor, a bushing tap, or a conductor port coupling device, used to inject broadband detection signals into submarine power cables. The broadband response acquisition unit is used to acquire the broadband voltage response and broadband current response of the submarine power cable port; The reflection coefficient phase anomaly extraction unit is used to calculate the input impedance and reflection coefficient based on the wideband voltage response and wideband current response, and to extract the reflection coefficient phase anomaly quantity corresponding to the local impedance anomaly. The defect distance inversion unit is used to invert the defect distance by using a corrected phase constant to assess the phase anomaly of the reflection coefficient. The diagnostic output unit is used to output local defect distance, overall insulation degradation indication, local structural defect indication, or location reliability.

Images