Method for evaluating development of cable insulation water tree aging based on time-frequency domain features

Abstract

The application relates to the technical field of insulation state detection, in particular to a cable insulation water tree aging development evaluation method based on time-frequency domain characteristics, which comprises the following steps: applying variable-frequency alternating excitation voltages of at least two different voltage grades to an insulation cable, collecting response current signals of the cable at each detection frequency point under each voltage grade, and decomposing the resistance current signals and the capacitive current signals; segmenting the capacitive current signals of each detection frequency point, calculating the harmonic energy ratios of each signal segment, determining the capacitive current diffusion ratios representing the energy static diffusion characteristics; determining the net diffusion ratios of each detection frequency point, obtaining the dielectric loss correction coefficients of each detection frequency point, correcting the capacitive current signals according to the dielectric loss correction coefficients, calculating the dielectric loss tangent values representing the real dielectric loss characteristics, and evaluating the water tree aging degree of the cable insulation. The application improves the evaluation accuracy of the water tree aging degree of the cable.

Description

Technical Field

[0001] This application relates to the field of insulation condition detection technology, specifically to a method for assessing the aging development of water tree in cable insulation based on time-frequency domain characteristics. Background Technology

[0002] Cross-linked polyethylene (XLPE) cables are widely used in power transmission and distribution systems due to their excellent electrical properties and thermal stability. During long-term operation, the insulation layer of XLPE cables is inevitably affected by the humid environment and electric field coupling, gradually developing water treeing aging, which may lead to insulation breakdown, short circuits, and other accidents. Therefore, accurately assessing the degree of water treeing aging in XLPE cables is of great significance for ensuring the safe operation of power systems.

[0003] Currently, frequency domain dielectric spectroscopy (FDS) is a non-destructive testing method used to assess the insulation condition of cables. It applies a variable-frequency AC voltage and decomposes the response current to obtain resistive and capacitive currents, and then calculates the dielectric loss tangent to assess the insulation condition of the cable. However, when water treeing occurs in cables, the local high field strength at the micro-defects causes nonlinear polarization of the insulation material, resulting in significant harmonic components from the cable itself being mixed into the response current. Traditional FDS methods cannot separate and eliminate nonlinear harmonic interference, and only perform simple linear resistive-capacitive decomposition on the response current. This incorrectly includes all harmonic interference terms generated by nonlinear polarization in the capacitive current, leading to serious deviations in the calculation results of the dielectric loss tangent and making it impossible to accurately assess the true degree of cable aging. Summary of the Invention

[0004] To address the aforementioned technical issues, a method for assessing the aging development of water tree in cable insulation based on time-frequency domain characteristics is provided to resolve existing problems.

[0005] The solution to the technical problem in this application is to provide a method for assessing the aging development of water tree in cable insulation based on time-frequency domain characteristics, including the following steps:

[0006] At least two different voltage levels of variable frequency AC excitation voltage are applied to the insulated cable, and the response current signal of the cable at each detection frequency point under each voltage level is collected. The resistive current signal and capacitive current signal are decomposed from the response current signal.

[0007] The capacitive current signal at each detection frequency point is segmented, the harmonic energy distribution characteristics of each signal segment in the frequency domain are analyzed, the harmonic energy ratio of each signal segment is calculated, the trend of the change of the harmonic energy ratio is analyzed, and the trend of the change is extracted from the harmonic energy ratio to determine the capacitive current diffusion ratio that characterizes the static diffusion characteristics of energy.

[0008] The baseline level of the harmonic energy ratio of all signal segments at each detection frequency is evaluated, and the baseline is removed from the capacitive current diffusion ratio to determine the net diffusion ratio of each detection frequency. Combined with the fundamental frequency energy distribution characteristics of the capacitive current signal in the frequency domain, the dielectric loss correction coefficient characterizing the proportion of the true capacitive response component at each detection frequency is obtained. Based on this, the capacitive current signal is corrected. Using the correction results and the resistive current signal, the dielectric loss tangent value characterizing the true dielectric loss characteristics is calculated. The deviation of the dielectric loss tangent value at the same detection frequency under different voltage levels is used to evaluate the degree of water tree aging of the cable insulation.

[0009] Preferably, the segmentation of the capacitive current signal at each detection frequency point includes: for each detection frequency point, calculating the period of its AC excitation voltage, defining it as a reference period, defining multiple consecutive reference periods as a time window, and having an overlap length of one reference period between two adjacent time windows; and extracting the signal segments contained in each time window from the capacitive current signal to obtain all signal segments.

[0010] Preferably, the calculation of the harmonic energy ratio of each signal segment includes: performing frequency domain analysis on each signal segment for each detection frequency point to obtain a spectrum diagram; summing the energy of all frequency components in the spectrum diagram to obtain the total energy; using each detection frequency point as the fundamental frequency, defining the signal component corresponding to the fundamental frequency in the spectrum diagram as the fundamental frequency component; calculating the proportion of the energy corresponding to the fundamental frequency component in the spectrum diagram in the total energy, performing a negative mapping on it, and using it as the harmonic energy ratio of each signal segment.

[0011] Preferably, the calculation process of the capacitive current diffusion ratio is as follows: the harmonic energy ratios of all signal segments at each detection frequency point are used to form a harmonic energy sequence; the harmonic energy sequence is decomposed into a trend sequence to extract the trend sequence; the trend sequence is nonlinearly fitted to obtain the fitting value corresponding to each signal segment; the corresponding fitting value is removed from the harmonic energy ratio of each signal segment to obtain the capacitive current diffusion ratio of each signal segment.

[0012] Preferably, determining the net diffusion ratio at each detection frequency point includes: extracting a baseline from the harmonic energy sequence, taking the constant value corresponding to the baseline as the baseline level value; calculating the average value of the capacitive current diffusion ratio of all signal segments at each detection frequency point, and taking the difference between the average value and the baseline level value as the net diffusion ratio.

[0013] Preferably, the dielectric loss correction coefficient is calculated as follows: frequency domain analysis is performed on the capacitive current signal decomposed at each detection frequency point, and the proportion of the energy corresponding to the fundamental frequency component in the sum of the energies corresponding to all frequency components is extracted from the obtained spectrum as the fundamental frequency energy ratio; the dielectric loss correction coefficient is the result of positive fusion of the fundamental frequency energy ratio and the net diffusion ratio.

[0014] Preferably, the specific process of the forward fusion is as follows: the sum of the fundamental frequency energy ratio and the net diffusion ratio is used as the dielectric loss correction coefficient for each detection frequency point.

[0015] Preferably, the correction of the capacitive current signal includes: calculating the effective value of the decomposed capacitive current signal for each detection frequency point, and multiplying it by the dielectric loss correction coefficient as the corrected effective value of the capacitive current.

[0016] Preferably, the dielectric loss tangent is the ratio of the effective value of the resistive current signal to the corrected effective value of the capacitive current.

[0017] Preferably, the assessment of the water tree aging degree of the cable insulation includes: two voltage levels respectively and 0.25 The voltage level is With a voltage level of 0.25 The ratio of the dielectric loss tangent at the same detection frequency point is denoted as the relative ratio; the average of the relative ratios at all detection frequencies is taken as the nonlinearity; if the nonlinearity is less than or equal to a preset first value, the cable is not in a water tree aging state; if the nonlinearity is greater than the preset first value and less than or equal to a preset second value, the cable is in a mild water tree aging state; if the nonlinearity is greater than the preset second value and less than or equal to a preset third value, the cable is in a moderate water tree aging state; otherwise, the cable is in a severe water tree aging state.

[0018] This application has at least the following beneficial effects:

[0019] This application, by applying a variable-frequency AC excitation voltage, can obtain the dielectric response characteristics of cable insulation over a wide frequency range, and initially decouple the response current into resistive and capacitive currents, providing basic data for subsequent removal of nonlinear harmonic errors. The capacitive current signal is segmented, and the harmonic energy distribution characteristics of each signal segment in the frequency domain are analyzed, and the harmonic energy ratio of each signal segment is calculated. Its beneficial effect lies in separating the interference information originally mixed in the time domain in the frequency domain through time-series segmentation of the capacitive current and extraction of the harmonic energy ratio, directly quantifying the harmonic interference. The proportion of energy to total energy provides an objective and quantifiable indicator of interference level for subsequent calibration. Determining the capacitive current diffusion ratio, which characterizes the static diffusion features of energy, has the beneficial effect of dynamically capturing the nonlinear distortion evolution during polarization. By eliminating the time-varying trend of the harmonic energy ratio, the influence of the dynamic evolution of nonlinear polarization on capacitive current measurement during the test is eliminated, allowing the capacitive current diffusion ratio to reflect the static characteristics of energy diffusion more stably and purely. Removing the baseline from the capacitive current diffusion ratio to determine the net diffusion ratio at each detection frequency point has the beneficial effect of… The advantages of this method are: by eliminating the baseline of the harmonic energy ratio, systematic errors such as spectral leakage caused by segmented processing are eliminated, making the net diffusion ratio more realistically reflect the energy diffusion characteristics of the cable insulation itself due to water tree aging; a dielectric loss correction coefficient is obtained, the beneficial effect of which is that by superimposing the fundamental frequency energy ratio, which has not undergone energy diffusion, with the net diffusion ratio, which reflects the degree of energy diffusion, the proportion of the measured capacitive current occupied by the true capacitive response is quantified, and the true capacitive response component is restored; the capacitive current signal is corrected, and the dielectric loss tangent value, which characterizes the true dielectric loss characteristics at each detection frequency, is calculated using the correction result and the resistive current signal to evaluate the degree of water tree aging of the cable insulation. The beneficial effect of this method is that by using the dielectric loss correction coefficient to correct the capacitive current, the true capacitive current response is restored from the measurement value exaggerated by harmonics, fundamentally solving the measurement deviation caused by harmonic interference, and making the dielectric loss tangent value free from the false influence of nonlinear harmonics, which can more realistically reflect the actual loss characteristics of the cable insulation, and significantly improving the accuracy and reliability of the assessment of the degree of water tree aging of cables. Attached Figure Description

[0020] The following section provides a more detailed description of the cable insulation water tree aging development assessment method based on time-frequency domain characteristics, in conjunction with the accompanying drawings.

[0021] Figure 1 A flowchart illustrating the steps of the cable insulation water tree aging development assessment method based on time-frequency domain characteristics provided in this application embodiment;

[0022] Figure 2 A flowchart illustrating the steps of the method for obtaining the dielectric loss correction coefficient provided in this application embodiment. Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description of the cable insulation water tree aging development assessment method based on time-frequency domain characteristics, in conjunction with the accompanying drawings and implementation examples, is provided. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0024] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains.

[0025] Please see Figure 1 The diagram illustrates a flowchart of a method for assessing the aging of cable insulation water tree based on time-frequency domain characteristics, according to an embodiment of this application. The method includes the following steps:

[0026] Step 1: Apply at least two different voltage levels of variable frequency AC excitation voltage to the insulated cable, collect the response current signal of the cable at each detection frequency point under each voltage level, and decompose the resistive current signal and capacitive current signal from the response current signal.

[0027] The stable operation of cross-linked polyethylene (XLPE) cables is mainly affected by operating conditions, external environment, and insulation aging. During long-term operation, the cable insulation layer is inevitably affected by the humid environment and electric field coupling, gradually developing water treeing aging. The dielectric properties of XLPE cable insulation are an important reference for characterizing its aging degree. Among them, frequency domain dielectric spectroscopy (FDS) technology is a broadband AC dielectric spectrum testing method. Its basic principle is to apply a variable frequency AC voltage to the XLPE cable and obtain the polarization characteristics and loss characteristics of the insulation material by measuring the complex dielectric response of the insulation material at different frequencies. When an excitation is applied to the XLPE cable, the response current generated by the excitation can be regarded as the dielectric response of the excitation. Therefore, under the action of a uniform alternating electric field applied to the XLPE cable, the cable insulation can be equivalent to a parallel resistance and capacitance model. At this time, the response current can be decoupled into a resistive current reflecting the AC conductivity characteristics and a capacitive current reflecting the polarization characteristics. Then, the degree of insulation aging can be quantitatively characterized by the dielectric loss tangent.

[0028] The specific process of traditional FDS technology is as follows: a sinusoidal AC excitation voltage of a certain frequency is applied to the cable through a high-voltage AC testing system. The response current signal is obtained through a current detection device. Under the condition that the insulating material exhibits linear dielectric properties, the cable can be equivalently represented as a parallel model of a resistor and a capacitor, and the response current signal... It can be decomposed into resistive current. and capacitive current The linear superposition, i.e. Then, the dielectric loss tangent is calculated. ,in, This is the effective value of the resistive current. This is the effective value of the capacitive current; therefore, the smaller the dielectric loss tangent, the less aged the cable is, and vice versa.

[0029] However, in practice, as the cable ages due to water treeing, the micron-sized dendritic channels formed inside the insulation material under the combined action of electric field and moisture can have a tip radius of curvature as small as less than 100 nanometers. This results in a local electric field strength far exceeding the applied electric field, which in turn induces nonlinear orientation polarization in the polar groups (such as -OH, -COOH) in the XLPE molecular chain and in the interfacial water molecules. Therefore, in addition to resistive and capacitive currents, the response current also contains harmonic current signals generated by nonlinear effects within the cable itself. ,Right now Because the traditional FDS method can only decompose the response current into two items, namely resistive current and capacitive current, it cannot independently extract the harmonic components, resulting in harmonic current signals. It was incorrectly classified as capacitive current, meaning the capacitive current actually used in the calculation became... , This represents the effective value of the harmonic current, from which the dielectric loss tangent is calculated as: The dielectric loss tangent value deviates from the true value, thus interfering with the accuracy of the test on the cable insulation performance.

[0030] Based on the above analysis, the influence of nonlinear harmonic interference on dielectric loss measurement is eliminated by introducing time-frequency domain analysis. Therefore, FDS technology is used to test the dielectric response of XLPE cables. The specific test process is as follows:

[0031] Dielectric response testing experiments were conducted using an AC testing system. This system included a signal generator, a high-voltage amplifier, a micro-current sensor, a current amplifier, and a shielded box. The signal generator produced a sinusoidal voltage signal at a set frequency. The high-voltage amplifier amplified the voltage amplitude output by the signal generator to the required test level. The micro-current sensor acquired the weak response current signal from the cable. The current amplifier conditioned and amplified the signal output by the sensor. The shielded box isolated external electromagnetic interference. The selection and connection methods of the above equipment are well-known technologies and will not be elaborated here.

[0032] Using an AC testing system, variable frequency AC excitation voltages of different voltage levels are applied to the cable. After the excitation voltage is applied, the response current signal of the cable at each detection frequency point under different voltage levels is continuously collected by a micro current sensor.

[0033] The applied voltage levels are respectively and 0.25 , for Furthermore, the frequency conversion AC excitation voltage is a sinusoidal AC voltage, the detection frequency range is 0.001Hz~1000Hz, and the test temperature is required to be controlled at 30℃.

[0034] It should be noted that, in order to avoid polarization residual crosstalk between different detection frequencies, the cable was cut into multiple independent short cable samples. Each sample was only used to apply an AC excitation voltage at a single specific frequency and was discarded after the test. No reuse test was performed at other frequencies. Therefore, an excitation voltage at a single specific frequency was applied to each sample, and the response current signal at that detection frequency was collected 10 minutes after the AC electric field was applied. Thus, excitation at different specific frequencies was applied to all samples, and the response current signal at each detection frequency was obtained.

[0035] In addition, the response current signal at each detection frequency point contains at least 30 complete sine cycles. As another implementation method, the implementer can set it according to the actual situation.

[0036] Furthermore, the response current signal is decomposed as follows:

[0037] The response current signal is then decomposed into resistive current signal and capacitive current signal;

[0038] In this embodiment, the resistive current signal and the capacitive current signal are decomposed using the current phase shift addition and subtraction method. The current phase shift addition and subtraction method is a well-known technology and will not be described in detail here. As other implementation methods, implementers may also use other methods of the prior art, such as the voltage phase shift addition and subtraction method, the zero-crossing detection method, etc. This embodiment does not impose any special restrictions on this.

[0039] Thus, the response current signals of the cable at various detection frequencies under different voltage levels are obtained, and the resistive current signal and capacitive current signal are decomposed.

[0040] Step 2: Segment the capacitive current signal at each detection frequency point, analyze the harmonic energy distribution characteristics of each signal segment in the frequency domain, calculate the harmonic energy ratio of each signal segment, analyze the trend of the change of the harmonic energy ratio, and extract the trend of the change from the harmonic energy ratio to determine the capacitive current diffusion ratio that characterizes the static diffusion characteristics of energy.

[0041] In standard frequency domain dielectric spectrum testing, the excitation voltage applied to XLPE cables is a standard sinusoidal signal. In the equivalent parallel resistance-capacitance model of the insulated cable, based on the principle of "parallel current division," the total response current theoretically contains only two components: resistive current and capacitive current. Since the resistor is a purely linear element, changing only its amplitude without altering its frequency or phase, the resistive current remains in phase and frequency with the excitation voltage. The capacitor, however, causes the phase of the capacitive current to lead the phase of the voltage. Ideally, the frequency components remain the same as the excitation voltage. Therefore, when the cable insulation exhibits linear dielectric characteristics, the total response current is the vector sum of resistive and capacitive currents, neither of which generates additional frequency components.

[0042] However, when water treeing occurs in the cable, it leads to interference from harmonic currents inherent to the cable itself in the response current. Since the decomposed resistive current is in phase and frequency with the excitation voltage and has no additional frequency components, the influence of the harmonic current is entirely attributed to the capacitive current component, causing the measured value of the capacitive current to deviate from its true value. To eliminate this deviation, a quantitative analysis of the harmonic interference in the capacitive current is required, specifically:

[0043] For each detection frequency point, the period of its AC excitation voltage is calculated and defined as the reference period. Multiple consecutive reference periods are defined as a time window, and there is an overlap length of one reference period between two adjacent time windows.

[0044] In this embodiment, for each detection frequency point Calculate the period of its AC excitation voltage, and define it as the reference period. ,but Using the aforementioned reference period as the basic unit, the length of each time window is set to... The overlap length between two adjacent time windows is one reference period. To ensure the continuity of the time window and avoid information loss due to window boundary truncation, as another implementation method, the implementer can set it according to the actual situation.

[0045] The signal segments contained in each time window are extracted from the capacitive current signals at each detection frequency point; frequency domain analysis is performed on the signal segments to obtain the spectrum.

[0046] In this embodiment, Fourier transform is used for frequency domain analysis and to obtain the spectrum. The Fourier transform is a well-known technique and will not be described in detail here.

[0047] The sum of the energies of all frequency components in the statistical spectrum is the total energy.

[0048] Using each detection frequency as the fundamental frequency, the signal component corresponding to the fundamental frequency in the spectrum is defined as the fundamental frequency component; the proportion of the energy corresponding to the fundamental frequency component in the total energy in the spectrum is statistically analyzed, and a negative mapping is performed on it to obtain the harmonic energy ratio of the signal segment;

[0049] In this embodiment, the negative mapping process is as follows: the difference between the value 1 and the percentage is taken as the harmonic energy ratio.

[0050] It should be noted that the more energy the harmonic components occupy in the spectrum, that is, the greater the harmonic energy ratio, the greater the interference of the harmonic current in the cable body.

[0051] Furthermore, in conventional signal processing, the arithmetic mean of the harmonic energy ratios of all signal segments can be directly calculated as a measure of the overall harmonic interference level, which is then used to correct the dielectric loss tangent. However, this simple averaging method has significant limitations: it assumes that harmonic interference remains constant throughout the entire test cycle. In reality, as the aging of cable water treeing deepens and the charge accumulation effect changes during the test, the harmonic energy ratios in different signal segments are not constant but exhibit a certain time-varying pattern. Simple averaging will obscure the dynamic evolution information of harmonic interference during nonlinear polarization, leading to distortion in the estimation of harmonic components in capacitive current, and thus affecting the accuracy of dielectric loss tangent correction.

[0052] In the initial stage of the test, the water and polar groups within the water tree channel are in thermal equilibrium, exhibiting disordered Brownian motion. The orientation of the dipoles is random. At this point, the applied electric field has just been applied, and the dipoles begin to align in response to the electric field. However, a large number of hydrogen bonds in the water have not yet fully broken, resulting in a weak nonlinear polarization effect. Therefore, the harmonic current of the cable body is not fully excited, and the degree of harmonic interference exhibits some random fluctuations. As the electric field continues to act, space charge accumulates within the water tree channel, further enhancing the local electric field strength at the tip of the water tree. Simultaneously, the local temperature gradually increases. Under these conditions, hydrogen bonds gradually dissociate, and the nonlinear orientation polarization effect of polar groups and water molecules under the influence of a strong electric field is significantly enhanced, leading to a gradual increase in the degree of interference from the harmonic current of the cable body. When the test continues to a certain stage, the process of hydrogen bond breaking and recombination gradually approaches dynamic equilibrium, and the nonlinear polarization effect reaches a relatively stable state. At this point, the overall degree of interference from the harmonic current of the cable body also reaches its maximum value and remains relatively stable. In summary, throughout the entire testing process, the degree of harmonic interference exhibited a progressive pattern of "initial fluctuation → gradual increase → tendency to stabilize," which reflects the nonlinear polarization dynamic evolution process of water-tree aged cables under the action of AC electric field.

[0053] Based on the above analysis, the interference of harmonics exhibits a certain degree of variation. Removing this variation allows us to obtain the most accurate capacitive current, specifically:

[0054] The harmonic energy ratios of all signal segments at each detection frequency point are used to form a harmonic energy sequence. The harmonic energy sequence is then decomposed to extract the trend sequence.

[0055] In this embodiment, the STL (Seasonal-Trend decomposition using Loess) algorithm is used to decompose the trend and extract the trend sequence. The STL algorithm is a well-known technology and will not be described in detail here.

[0056] Nonlinear fitting is performed on the trend sequence to obtain the fitted value corresponding to each signal segment;

[0057] In this embodiment, a polynomial fitting method is used for nonlinear fitting. The polynomial fitting method is a well-known technique and will not be described in detail here.

[0058] The capacitive current diffusion ratio of each signal segment is obtained by removing the corresponding fitted value from the harmonic energy ratio of each signal segment.

[0059] In this embodiment, the difference between the harmonic energy ratio of each signal segment and the fitted value is used as the capacitive current diffusion ratio.

[0060] It should be noted that the fitted value characterizes the macroscopic trend of polar group orientation polarization evolving over time during the test, while the capacitive current diffusion ratio strips away this evolution trend and more accurately captures the nonlinear harmonic energy diffusion characteristics caused by water tree aging. This process eliminates the influence of nonlinear polarization of the cable body harmonic current on the capacitive current during the test. The larger the value, the higher the degree to which the capacitive response energy that originally belonged to the fundamental frequency diffuses or "escapes" to other frequency components.

[0061] Thus, the capacitive current diffusion ratio of each signal segment at each detection frequency point is obtained.

[0062] Step 3: Evaluate the baseline level of the harmonic energy ratio of all signal segments at each detection frequency, remove the baseline from the capacitive current diffusion ratio, determine the net diffusion ratio of each detection frequency, and combine the fundamental frequency energy distribution characteristics of the capacitive current signal in the frequency domain to obtain the dielectric loss correction coefficient that characterizes the proportion of the true capacitive response component at each detection frequency.

[0063] The flowchart of the method for obtaining the dielectric loss correction coefficient provided in this application embodiment is as follows: Figure 2 As shown.

[0064] Furthermore, when segmenting the capacitive current signal using a time window, directly truncating the time window to obtain a specific length will cause spectral leakage during the frequency domain transformation due to the non-periodic truncation effect of the time domain signal. This will cause the signal energy that should be concentrated in the fundamental frequency component to spread to surrounding frequencies, generating sidelobe interference in the spectrum. Consequently, false harmonic energy unrelated to the actual polarization of the cable will be introduced into each frequency component. Therefore, the harmonic energy ratio calculated after truncation will have a systematic deviation from the frequency domain response of the original true signal.

[0065] Meanwhile, given that the truncation length is the same for all time windows, the truncation process has the same effect on the frequency domain response of the signal in different time windows. Therefore, the false harmonic energy is a relatively stable background constant in the frequency domain response of different time windows, that is, it is an approximately constant constant term in the harmonic energy sequence. Therefore, by extracting the baseline of the harmonic energy ratio sequence and removing it, the algorithm error introduced by the time domain truncation can be eliminated, and the true nonlinear polarization response of the cable can be restored.

[0066] The baseline is extracted from the harmonic energy sequence, and the constant value corresponding to the baseline is used as the baseline level value.

[0067] In this embodiment, the least squares method is used to fit the harmonic energy sequence with a low-order polynomial, and the constant term of the fitting function is used as the baseline level value. The least squares method is a well-known technique and will not be described in detail here. As another implementation method, the implementer may also use other methods to extract the baseline value, specifically: extract all the minimum values ​​in the harmonic energy sequence and calculate the median of all the minimum values ​​as the baseline level value. The baseline extraction is a well-known technique and will not be described in detail here.

[0068] Calculate the average value of the capacitive current diffusion ratio for all signal segments at each detection frequency point, and take the difference between it and the baseline level value as the net diffusion ratio;

[0069] Frequency domain analysis is performed on the capacitive current signal decomposed at each detection frequency point, and the proportion of the energy corresponding to the fundamental frequency component in the sum of the energies of all frequency components is extracted from the obtained spectrum as the fundamental frequency energy ratio.

[0070] In this embodiment, Fourier transform is used for frequency domain analysis and to obtain the spectrum. The Fourier transform is a well-known technique and will not be described in detail here.

[0071] The sum of the fundamental frequency energy ratio and the net diffusion ratio is used as the dielectric loss correction coefficient for each detection frequency point at each voltage level.

[0072] It should be noted that the baseline level reflects the degree of systematic error introduced by the time-domain truncation process; the net diffusion ratio reflects the proportion of effective capacitive response energy diffused from the fundamental frequency to other frequency components due to water tree aging of the cable body after removing errors. The larger the value, the more severe the degree of capacitive current energy diffusion; the fundamental frequency energy ratio reflects the proportion of capacitive current that has not undergone energy diffusion under ideal conditions, that is, the proportion of the ideal linear response to the total response. By superimposing the fundamental frequency energy ratio reflecting the absence of energy diffusion with the net diffusion ratio reflecting the fundamental frequency energy diffusion, the dielectric loss correction coefficient is obtained, which characterizes the proportion of the actual capacitive current response to the measured capacitive current. That is, this correction coefficient reflects the proportion of the decomposed capacitive current signal occupied by the real capacitive current, and is used to recover the real capacitive current from the overestimated capacitive current signal.

[0073] Thus, the dielectric loss correction coefficients for each detection frequency point at each voltage level are obtained.

[0074] Step 4: Based on the dielectric loss correction coefficient, the capacitive current signal is corrected. Using the correction result and the resistive current signal, the dielectric loss tangent value, which characterizes the actual dielectric loss characteristics, is calculated. The deviation of the dielectric loss tangent value at the same detection frequency under different voltage levels is used to evaluate the degree of water tree aging of the cable insulation.

[0075] Furthermore, based on the resistive and capacitive current signals decomposed at each detection frequency, and the dielectric loss correction coefficient, the dielectric loss tangent is calculated, specifically as follows:

[0076] For each detection frequency point, the effective value of the decomposed capacitive current signal is calculated, and its product with the dielectric loss correction coefficient is used as the corrected effective value of the capacitive current.

[0077] The ratio of the effective value of the decomposed resistive current signal to the effective value of the corrected capacitive current is calculated and used as the dielectric loss tangent value at each detection frequency.

[0078] It should be noted that the calculation of the effective value of the current is a well-known technique and will not be elaborated here.

[0079] It should be noted that the decomposed capacitive current signal contains a large number of harmonics, thus being significantly "exaggerated," resulting in an underestimation of the uncorrected dielectric loss tangent, which masks the true aging condition. By using the dielectric loss correction coefficient, the capacitive current signal is corrected. The dielectric loss correction coefficient reflects the proportion of the true capacitive current in the decomposed capacitive current signal, and the corrected effective value of the capacitive current reflects the true capacitive current response after eliminating harmonic interference, characterizing the true dielectric loss characteristics. This makes the dielectric loss tangent more accurately reflect the aging state of the cable insulation, thereby improving the accuracy of water tree aging assessment.

[0080] Furthermore, the degree of cable aging is assessed based on the dielectric loss tangent values ​​at all detection frequencies for different voltage levels, specifically:

[0081] With voltage level With a voltage level of 0.25 The ratio of the dielectric loss tangent values ​​at the same detection frequency between two points is denoted as the relative ratio.

[0082] The mean of the relative ratios corresponding to all detection frequency points is taken as the nonlinearity;

[0083] If the nonlinearity is less than or equal to the preset first value, the cable is not in a water tree aging state; if the nonlinearity is greater than the preset first value and less than or equal to the preset second value, the cable is in a mild water tree aging state; if the nonlinearity is greater than the preset second value and less than or equal to the preset third value, the cable is in a moderate water tree aging state; if the nonlinearity is greater than the preset third value, the cable is in a severe water tree aging state.

[0084] In this embodiment, for a 10kV cable, the first preset value is set to 1, the second preset value is set to 2, and the third preset value is set to 6. In other implementations, the implementer can set these values ​​according to the actual cable specifications used.

[0085] It should be noted that the dielectric response of a healthy cable is linear, and the dielectric loss tangent is basically the same at different voltage levels. However, water treeing aging can lead to severe nonlinearity, and the higher the voltage level, the greater the dielectric loss tangent.

[0086] It should be understood that, although Figure 1 The steps in the flowchart are shown sequentially as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order in which these steps are executed, and they can be performed in other orders. Figure 1 At least some of the steps in the process may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least some of the sub-steps or stages of other steps.

[0087] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0088] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application. Therefore, any simple modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of this application, without departing from the content of the technical solution of this application, shall fall within the protection scope of the technical solution of this application.

Claims

1. A method for assessing the aging development of water tree in cable insulation based on time-frequency domain characteristics, characterized in that, The method includes the following steps: At least two different voltage levels of variable frequency AC excitation voltage are applied to the insulated cable, and the response current signal of the cable at each detection frequency point under each voltage level is collected. The resistive current signal and capacitive current signal are decomposed from the response current signal. The capacitive current signal at each detection frequency point is segmented, the harmonic energy distribution characteristics of each signal segment in the frequency domain are analyzed, the harmonic energy ratio of each signal segment is calculated, the trend of the change of the harmonic energy ratio is analyzed, and the trend of the change is extracted from the harmonic energy ratio to determine the capacitive current diffusion ratio that characterizes the static diffusion characteristics of energy. The baseline level of the harmonic energy ratio of all signal segments at each detection frequency is evaluated, and the baseline is removed from the capacitive current diffusion ratio to determine the net diffusion ratio of each detection frequency. Combined with the fundamental frequency energy distribution characteristics of the capacitive current signal in the frequency domain, the dielectric loss correction coefficient characterizing the proportion of the true capacitive response component at each detection frequency is obtained. Based on this, the capacitive current signal is corrected. Using the correction results and the resistive current signal, the dielectric loss tangent value characterizing the true dielectric loss characteristics is calculated. The deviation of the dielectric loss tangent value at the same detection frequency under different voltage levels is used to evaluate the degree of water tree aging of the cable insulation.

2. The method for assessing the aging development of water tree in cable insulation based on time-frequency domain characteristics as described in claim 1, characterized in that, The segmentation of the capacitive current signal at each detection frequency point includes: for each detection frequency point, calculating the period of its AC excitation voltage, defining it as a reference period, defining multiple consecutive reference periods as a time window, and having an overlap length of one reference period between two adjacent time windows; and extracting the signal segments contained in each time window from the capacitive current signal to obtain all signal segments.

3. The method for assessing the aging development of water tree in cable insulation based on time-frequency domain characteristics as described in claim 1, characterized in that, The calculation of the harmonic energy ratio of each signal segment includes: performing frequency domain analysis on each signal segment for each detection frequency point to obtain a spectrum diagram; summing the energy of all frequency components in the spectrum diagram to obtain the total energy; using each detection frequency point as the fundamental frequency, defining the signal component corresponding to the fundamental frequency in the spectrum diagram as the fundamental frequency component; calculating the proportion of the energy corresponding to the fundamental frequency component in the spectrum diagram in the total energy, performing a negative mapping on it, and using it as the harmonic energy ratio of each signal segment.

4. The method for assessing the aging development of water tree in cable insulation based on time-frequency domain characteristics as described in claim 1, characterized in that, The calculation process of the capacitive current diffusion ratio is as follows: the harmonic energy ratios of all signal segments at each detection frequency point are used to form a harmonic energy sequence, and the harmonic energy sequence is decomposed into a trend sequence to extract the trend sequence. The trend sequence is nonlinearly fitted to obtain the fitted value corresponding to each signal segment; the corresponding fitted value is removed from the harmonic energy ratio of each signal segment to obtain the capacitive current diffusion ratio of each signal segment.

5. The method for assessing the aging development of water tree in cable insulation based on time-frequency domain characteristics as described in claim 4, characterized in that, Determining the net diffusion ratio at each detection frequency point includes: extracting a baseline from the harmonic energy sequence and using the constant value corresponding to the baseline as the baseline level value; calculating the average value of the capacitive current diffusion ratio of all signal segments at each detection frequency point, and using the difference between the average value and the baseline level value as the net diffusion ratio.

6. The method for assessing the aging development of water tree in cable insulation based on time-frequency domain characteristics as described in claim 3, characterized in that, The calculation process of the dielectric loss correction coefficient is as follows: frequency domain analysis is performed on the capacitive current signal decomposed at each detection frequency point, and the proportion of the energy corresponding to the fundamental frequency component in the sum of the energies corresponding to all frequency components is extracted from the obtained spectrum as the fundamental frequency energy ratio; the dielectric loss correction coefficient is the result of positive fusion of the fundamental frequency energy ratio and the net diffusion ratio.

7. The method for assessing the aging development of water tree in cable insulation based on time-frequency domain characteristics as described in claim 6, characterized in that, The specific process of the forward fusion is as follows: the sum of the fundamental frequency energy ratio and the net diffusion ratio is used as the dielectric loss correction coefficient for each detection frequency point.

8. The method for assessing the aging development of water tree in cable insulation based on time-frequency domain characteristics as described in claim 1, characterized in that, The correction of the capacitive current signal includes: calculating the effective value of the decomposed capacitive current signal for each detection frequency point, and multiplying it by the dielectric loss correction coefficient to obtain the corrected effective value of the capacitive current.

9. The method for assessing the aging development of water tree in cable insulation based on time-frequency domain characteristics as described in claim 8, characterized in that, The dielectric loss tangent is the ratio of the effective value of the resistive current signal to the corrected effective value of the capacitive current.

10. The method for assessing the aging development of water tree in cable insulation based on time-frequency domain characteristics as described in claim 1, characterized in that, The assessment of the water tree aging degree of the cable insulation includes: two voltage levels respectively and 0.25 The voltage level is With a voltage level of 0.25 The ratio of the dielectric loss tangent at the same detection frequency point is denoted as the relative ratio; the average of the relative ratios at all detection frequencies is taken as the nonlinearity; if the nonlinearity is less than or equal to a preset first value, the cable is not in a water tree aging state; if the nonlinearity is greater than the preset first value and less than or equal to a preset second value, the cable is in a mild water tree aging state; if the nonlinearity is greater than the preset second value and less than or equal to a preset third value, the cable is in a moderate water tree aging state; otherwise, the cable is in a severe water tree aging state.

Images