Method and system for detecting high resistance grounding fault of cable branch box

Abstract

The application discloses a kind of cable branch box high resistance ground fault detection method and system, the method includes: the zero sequence current of each return line is power frequency synchronous sampling and spectrum analysis, extracts fundamental wave effective value, third harmonic effective value, and synchronously obtains load current effective value;The ratio of zero sequence current third harmonic and fundamental wave is calculated, and harmonic ratio average and load current average are respectively obtained in preset time window;After a plurality of groups of data covering complete day load cycle are accumulated, linear regression is carried out with load current average as independent variable and harmonic ratio average as dependent variable, and the regression slope is used as harmonic ratio load sensitivity;It is compared with the normal range established in historical normal operating state, when the deviation degree exceeds alarm threshold, it is judged that the return line exists high resistance ground fault.This method can utilize the correlation characteristics of harmonic ratio with load change to identify weak fault, improve detection sensitivity and reliability.

Description

Technical Field

[0001] This application relates to the field of power distribution network fault detection technology, specifically to a method and system for detecting high-resistance grounding faults in cable branch boxes. Background Technology

[0002] Cable distribution boxes are widely used in medium-voltage power distribution networks for cable collection, branching, and transfer. Their internal busbars, insulation supports, cable terminals, and connections are subjected to the combined effects of electrical, thermal, and environmental factors over long periods. This can lead to high-resistance grounding paths due to moisture, contamination, aging, or deterioration of contact points. In low-current grounding systems, high-resistance grounding fault currents are typically small, easily comparable to the normal cable-to-ground capacitance current, insulation leakage current, and measurement noise. In the initial stages of a fault, they may not necessarily cause significant zero-sequence voltage or current exceedances. Existing detection technologies primarily use criteria such as zero-sequence current amplitude, direction, transient characteristics, harmonic content, or multi-feature identification. However, in actual operation, these are easily affected by load fluctuations, system imbalances, background harmonics, changes in line parameters, and measurement errors of zero-sequence current transformers. Fixed threshold values ​​cannot simultaneously ensure both sensitivity and reliability.

[0003] Some methods based on transient waveforms or intelligent models still rely on high sampling accuracy, typical fault samples, and complex parameter tuning, which limits their applicability in the field. Furthermore, existing methods often focus on fault line identification, lacking sufficient ability to distinguish faults between branch cables, terminal joints, and the common insulation area of ​​branch boxes, and lack effective integration between routine online monitoring and transient insulation assessment during power transmission. Therefore, there is a need to research a technical solution that can extract stable features from weak fault signals while also considering fault detection, location determination, and insulation degradation early warning. Summary of the Invention

[0004] To overcome the problems existing in related technologies, this manual provides a method and system for detecting high-resistance grounding faults in cable branch boxes.

[0005] According to a first aspect of the embodiments of this application, a method for detecting high-resistance grounding faults in cable branch boxes is provided, comprising: The zero-sequence current of each outgoing line of the cable branch box is sampled synchronously at power frequency. Spectrum analysis is performed on the sampled data to extract the fundamental effective value and the third harmonic effective value of the zero-sequence current for each power frequency cycle. The effective value of the load current of each outgoing line is obtained synchronously. The zero-sequence current harmonic ratio is obtained by the ratio of the effective value of the third harmonic of the zero-sequence current to the effective value of the fundamental wave of the zero-sequence current. Using a preset time window as the unit, calculate the mean value of the zero-sequence current harmonic ratio and the mean value of the load current effective value within each time window; After accumulating multiple sets of time window data covering the complete daily load cycle, linear regression calculation is performed with the mean of the effective value of the load current in each time window as the independent variable and the mean of the zero-sequence current harmonic ratio as the dependent variable to obtain the regression slope. The regression slope is used as the load sensitivity of the harmonic ratio. The harmonic ratio load sensitivity is compared with a normal range established based on the harmonic ratio load sensitivity under the historical normal operation of the circuit. When the harmonic ratio load sensitivity deviates from the normal range by more than a preset alarm threshold, it is determined that there is a high-resistance grounding fault on the outgoing line.

[0006] As an optional approach, the establishment of the normal range specifically includes: taking the harmonic ratio load sensitivity sequence obtained by the circuit operating continuously for a preset number of days under the confirmed normal insulation operating condition, and calculating the mean and standard deviation of the sequence; the alarm threshold is the standard deviation multiplied by a preset threshold coefficient.

[0007] As an optional solution, the determination of a high-resistance grounding fault on the outgoing line must also meet the following requirement: the harmonic ratio load sensitivity deviates from the normal range by more than the alarm threshold in a consecutive preset number of daily load cycles.

[0008] As an optional solution, a multi-circuit lateral comparison is also included, specifically: comparing the harmonic ratio load sensitivity obtained from each outgoing line of the cable branch box; when the harmonic ratio load sensitivity of only one outgoing line deviates from the normal range by more than the alarm threshold, the fault is determined to be located at the loop cable or its terminal joint; when the harmonic ratio load sensitivity of multiple outgoing lines simultaneously deviates from the normal range by more than the alarm threshold, and the relative difference between the harmonic ratio load sensitivity of each circuit is less than a preset ratio, the fault is determined to be located in the common insulation area of ​​the cable branch box.

[0009] As an optional approach, the spectrum analysis specifically includes: performing a discrete Fourier transform on the sampled data for each power frequency cycle, taking the magnitude of the Fourier coefficients corresponding to the fundamental frequency point and the third harmonic frequency point respectively, and calculating the effective value of the fundamental frequency of the zero-sequence current and the effective value of the third harmonic frequency of the zero-sequence current based on the magnitude values.

[0010] As an optional solution, it also includes power transmission transient insulation assessment, specifically: when the switch of the feeder where the cable branch box is located is closed, the high-speed waveform recording mode is entered with the closing signal as the trigger condition, and the zero-sequence current transient waveform and three-phase voltage waveform of each outgoing line within a preset time period from the closing time are recorded, and the ambient temperature at the closing time is also recorded.

[0011] As an optional approach, the method also includes deducting the capacitor charging component, specifically: determining the closing phase angle based on the three-phase voltage waveform, establishing a transient circuit model by combining the ground capacitance parameters of the cable line and the system impedance parameters, and calculating the theoretical capacitor charging current waveform; subtracting the theoretical capacitor charging current waveform from the zero-sequence current transient waveform to obtain the residual current.

[0012] As an optional approach, the method also includes the extraction of dielectric absorption current features, specifically: taking the data segment of the residual current within the time interval from the end of the capacitor charging inrush current decay after the circuit breaker is closed to the end of the preset time period, fitting the data segment with an exponential decay model, and extracting the initial amplitude and decay time constant of the dielectric absorption current.

[0013] As an optional solution, temperature correction is also included, specifically: based on the ambient temperature at the closing time and the preset reference temperature, the initial amplitude of the dielectric absorption current and the decay time constant are respectively corrected by temperature to obtain the initial amplitude of the dielectric absorption current and the decay time constant at the reference temperature; the temperature correction coefficient used for temperature correction is obtained by regression calibration of multiple power supply data of the same circuit at different ambient temperatures.

[0014] According to a second aspect of the embodiments of this application, a high-resistance grounding fault detection system for cable branch boxes is also provided, comprising: The sampling and spectrum analysis module is used to synchronously sample the zero-sequence current of each outgoing line of the cable branch box at the power frequency, perform spectrum analysis on the sampled data, and extract the effective value of the fundamental wave and the effective value of the third harmonic of the zero-sequence current for each power frequency cycle; and synchronously acquire the effective value of the load current of each outgoing line. The harmonic ratio calculation module is used to obtain the zero-sequence current harmonic ratio based on the ratio of the effective value of the third harmonic of the zero-sequence current to the effective value of the fundamental wave of the zero-sequence current. The time window statistics module is used to calculate the mean value of the zero-sequence current harmonic ratio and the mean value of the load current effective value in each preset time window. The load sensitivity calculation module is used to accumulate multiple sets of time window data covering the complete daily load cycle, and then perform linear regression calculation with the mean of the effective value of the load current in each time window as the independent variable and the mean of the zero-sequence current harmonic ratio as the dependent variable to obtain the regression slope. The regression slope is used as the harmonic ratio load sensitivity. The fault determination module is used to compare the harmonic ratio load sensitivity with a normal range established based on the harmonic ratio load sensitivity under the historical normal operation of the circuit. When the harmonic ratio load sensitivity deviates from the normal range by more than a preset alarm threshold, it is determined that there is a high-resistance grounding fault on the outgoing line.

[0015] This application uses the ratio of the third harmonic of the zero-sequence current to the fundamental current as the analysis object. The regression slope of this ratio as a function of load current is defined as the harmonic ratio load sensitivity. Utilizing the difference between the linear capacitance characteristics of normal insulation and the nonlinear volt-ampere characteristics of high-resistance grounding channels, stable features related to faults are extracted from normal capacitive current, load fluctuations, and background harmonics, reducing the risk of missed detection when weak fault currents do not reach the traditional amplitude threshold. By establishing the historical normal range for each circuit and introducing continuous periodic judgment, false alarms caused by occasional interference can be suppressed. Combined with the lateral comparison of sensitivity across multiple circuits, faults in circuit cables or terminal joints can be distinguished from faults in the common insulation area of ​​branch boxes. During feeder energization, the theoretical capacitance charging component is further subtracted, and the initial amplitude and decay time constant of the dielectric absorption current are extracted and temperature-corrected. This allows daily online fault detection and transient insulation assessment during energization to mutually verify each other, thereby improving the reliability of high-resistance grounding fault identification, fault location, and insulation degradation early warning. Attached Figure Description

[0016] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this specification and, together with the description, serve to explain the principles of this specification.

[0017] Figure 1 This is a flowchart of a high-resistance grounding fault detection method for a cable branch box according to one embodiment of this application.

[0018] Figure 2 This is a schematic diagram comparing the load sensitivity of a normal circuit and a high-resistance ground fault circuit in one embodiment of this application.

[0019] Figure 3 This is a schematic diagram of fault location based on load sensitivity of multi-loop harmonic ratio in one embodiment of this application.

[0020] Figure 4 This is a schematic block diagram of a high-resistance grounding fault detection system for a cable branch box according to this application.

[0021] Figure 5 This is a structural diagram of an electronic device according to one embodiment of this application. Detailed Implementation

[0022] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0023] This embodiment applies to cable branch boxes operating in a low-current grounding system mode in, for example, a 10kV distribution network. The branch box houses multiple cable outgoing lines, each equipped with a zero-sequence current transformer. The distribution automation terminal has power frequency synchronous sampling and event-triggered waveform recording functions. During normal system operation, the three-phase voltage contains third harmonic components originating from the transformer core magnetization nonlinearity and nonlinear loads, with amplitudes ranging from 1% to 3% of the fundamental voltage. The feeder where the branch box is located experiences periodic energizing events due to maintenance or power transfer operations.

[0024] The zero-sequence current amplitude of a high-resistance grounding fault is typically only a few tenths of an ampere to one or two amperes, on the same order of magnitude as the leakage current of a normal cable to ground capacitance. Traditional detection methods rely on zero-sequence current amplitude exceeding limits or power direction criteria, which are insufficiently sensitive to high-resistance faults. This method is based on the following physical mechanism: normal cable insulation is a linear capacitor, and each frequency component of the zero-sequence current maintains a strictly linear proportional relationship with its corresponding voltage component. This proportional relationship does not change due to small fluctuations in bus voltage with the load. The high-resistance grounding fault channel has nonlinear volt-ampere characteristics, and its impedance changes with the applied voltage amplitude. In the nonlinear characteristic, the higher-order terms are more sensitive to changes in voltage amplitude than the lower-order terms, causing the third harmonic current and the fundamental current in the fault channel to no longer respond proportionally to voltage fluctuations. Therefore, by quantifying the sensitivity of the zero-sequence current harmonic ratio to load changes, the nonlinear fault signal can be separated from the linear leakage current background.

[0025] The implementation process of the method described in this application will be explained in detail below with reference to specific embodiments.

[0026] Please see Figure 1 , Figure 1 This is a flowchart of a high-resistance grounding fault detection method for a cable branch box according to one embodiment of this application, such as... Figure 1 As shown, the method includes steps 101-105: Step 101: Zero-sequence current power frequency synchronous sampling and spectrum analysis. Power frequency synchronous sampling is performed on the zero-sequence current of each outgoing line in the cable branch box. Spectrum analysis is performed on the sampled data to extract the fundamental RMS value and the third harmonic RMS value of the zero-sequence current for each power frequency cycle; the RMS value of the load current of each outgoing line is obtained synchronously.

[0027] According to an embodiment of this disclosure, the power distribution terminal performs power frequency synchronous sampling of the zero-sequence current of each outgoing line of the cable branch box. The sampling frequency can be configured to, for example, 1.6 kHz, i.e., 32 sampling points are collected per power frequency cycle. The effective value of the load current of each outgoing line is obtained synchronously. This effective value of the load current is obtained by measuring the effective value of the current of each phase using a three-phase current transformer and then summing the three phases.

[0028] Spectral analysis is performed on the sampled data to extract the fundamental RMS value and the third harmonic RMS value of the zero-sequence current for each power frequency cycle. Specifically, a Discrete Fourier Transform (DFT) is performed on the sequence of N sampling points for each power frequency cycle. The Fourier coefficient magnitudes at the fundamental frequency point (the first harmonic corresponding to 50Hz) and the third harmonic frequency point (the third harmonic corresponding to 150Hz) are taken respectively, and the RMS values ​​of the corresponding frequency components are calculated accordingly. The Discrete Fourier Transform is a standard frequency domain analysis method that decomposes a discrete-time signal into complex amplitudes of each frequency component; for the N-point sequence... , No. The Fourier coefficients corresponding to the second harmonic are: ; in For the first The instantaneous value of zero-sequence current at each sampling point, in A; Number of sampling points per cycle; For harmonic order; The imaginary unit is used. For the fundamental frequency, take... For the third harmonic, take The effective values ​​of each harmonic are obtained by converting the coefficient modulus: the peak amplitude is... Valid value The unit is A. Based on this, the effective value of the fundamental frequency of the zero-sequence current is obtained. and the effective value of the third harmonic of the zero-sequence current .

[0029] Step 102: Obtain the zero-sequence current harmonic ratio based on the ratio of the effective value of the third harmonic of the zero-sequence current to the effective value of the fundamental wave of the zero-sequence current.

[0030] The effective value of the third harmonic of the zero-sequence current extracted in the aforementioned steps and the fundamental effective value of zero-sequence current Take the ratio to obtain the zero-sequence current harmonic ratio. : ; in and All are RMS current values, in amperes (A), and ratios. It reflects the proportion of the third harmonic component in the zero-sequence current relative to the fundamental component. For linear insulating capacitors, its value is approximately equal to the amplitude ratio of the third harmonic to the fundamental voltage in the system voltage, and it does not change with load variations.

[0031] Step 103: Calculate the mean value of the zero-sequence current harmonic ratio and the mean value of the load current effective value within a preset time window.

[0032] In one embodiment, a preset time window is set, the window length of which can be configured, for example, 5 minutes. Using this time window as a unit, the average value of the zero-sequence current harmonic ratio obtained cycle by cycle in the aforementioned steps and the average value of the effective load current obtained cycle by cycle are calculated respectively. Taking a 5-minute window as an example, each window contains approximately 15,000 power frequency cycles of data. Taking the arithmetic mean can suppress random measurement noise in a single cycle, thus obtaining the average value of the zero-sequence current harmonic ratio for that window. and the mean value of the effective value of the load current .

[0033] The selection principle for the time window length is as follows: if the window is too short, noise suppression will be insufficient; if the window is too long, details of the load change process cannot be captured. Based on the typical load fluctuation rate of a 10kV distribution network, 3 to 10 minutes is a suitable range.

[0034] Step 104: Perform linear regression calculation on accumulated time window data to obtain the harmonic ratio load sensitivity. After accumulating multiple sets of time window data covering the complete daily load cycle, perform linear regression calculation with the mean of the effective value of the load current in each time window as the independent variable and the mean of the zero-sequence current harmonic ratio as the dependent variable to obtain the regression slope. Use the regression slope as the harmonic ratio load sensitivity.

[0035] After accumulating data from multiple time windows covering the entire daily load cycle, the average of the effective load current values ​​for each time window is used. The independent variable is the mean of the zero-sequence current harmonic ratio. Linear regression was performed with the variable as the dependent variable. The data accumulation period can be configured, for example, from 3 to 7 days, to cover the complete load variation pattern of weekdays and rest days. Using a 5-minute window and a cumulative 5-day period as an example, approximately 1440 data pairs were obtained. ,in This is the window number.

[0036] In one embodiment, linear regression employs the least squares method, which involves fitting a linear model once on all pairs of data: Minimize the sum of squared residuals. Least squares is the standard method for solving linear regression problems. It obtains the analytical solution for the parameters by taking the partial derivative of the sum of squared residuals with respect to the parameters to be estimated and setting it to zero. Regression slope. The calculation method is as follows: ; in For the total number of data pairs; The arithmetic mean of the effective values ​​of all window load currents, in A; It is the arithmetic mean of the mean values ​​of the zero-sequence current harmonic ratios for all windows.

[0037] The obtained regression slope Harmonic ratio load sensitivity refers to the deviation of the zero-sequence current harmonic ratio caused by a unit change in load current. For normal linear insulation, the theoretical value of harmonic ratio load sensitivity is 0. When a nonlinear fault path exists, the harmonic ratio load sensitivity deviates from 0, and the larger the absolute value, the more significant the nonlinearity of the fault path.

[0038] As an example, the harmonic ratio load sensitivity calculated by linear regression after accumulating 5 days of data for a normal circuit is: The value is close to 0; the harmonic ratio and load sensitivity of a circuit with a high-impedance fault obtained during the same period are... The difference between the two is about an order of magnitude, indicating that the zero-sequence current harmonic ratio of the latter has a significant response to load changes.

[0039] Figure 2 This is a schematic diagram comparing the load sensitivity of a normal circuit and a high-resistance ground fault circuit in one embodiment of this application, as shown below. Figure 2 As shown, the horizontal axis represents the average effective value of the load current within each time window, and the vertical axis represents the average ratio of the third harmonic effective value of the zero-sequence current to the fundamental effective value within the corresponding time window. Hollow circles represent data points in normal circuits, where the zero-sequence current harmonic ratio fluctuates only slightly during load current changes. The fitted solid line remains essentially horizontal, and the regression slope K is close to zero, indicating that the current components of each frequency maintain a relatively stable proportional relationship in normal linear insulation. Solid triangles represent data points in circuits with high-resistance grounding faults, where the harmonic ratio shows a significant trend with increasing load current. The dashed line fitting result has a significant slope, meaning K for faults deviates significantly from zero. Therefore, the regression slope can be used as the load sensitivity of the harmonic ratio, and the degree of deviation from the normal range can be used to identify high-resistance grounding faults with nonlinear volt-ampere characteristics.

[0040] Step 105: Comparison of harmonic ratio load sensitivity with normal range and determination of high-resistance grounding fault. The harmonic ratio load sensitivity is compared with the normal range established based on the harmonic ratio load sensitivity under the historical normal operation of the circuit. When the harmonic ratio load sensitivity deviates from the normal range by more than a preset alarm threshold, it is determined that there is a high-resistance grounding fault in the outgoing line.

[0041] According to an embodiment of this disclosure, the harmonic ratio load sensitivity obtained in the aforementioned steps is compared with a normal range established based on the harmonic ratio load sensitivity under the historical normal operating conditions of the circuit. The normal range is established by taking a data sequence composed of harmonic ratio load sensitivity calculated for each daily load cycle within a preset number of days (e.g., 30 days) of continuous operation under confirmed normal insulation conditions, and calculating the mean of this sequence. and standard deviation .

[0042] The alarm threshold is the standard deviation. Multiply by a preset threshold coefficient The threshold coefficient is determined as follows: Several historical cases of confirmed high-impedance faults in similar branch boxes are selected; the standard deviation of the load sensitivity of the harmonic ratio in each case from the normal mean is calculated; and 80% of the smallest deviation is taken as the threshold coefficient. The value of ; when historical failure case data is lacking, It can be initially set to, for example, 2.5, with a value range of 2 to 3, and then gradually adjusted as more fault cases are accumulated. The adjustment method is the same as the initial calibration method, that is, the minimum deviation multiple is recalculated to 80% by incorporating newly acquired fault case data.

[0043] In one embodiment, the fault determination condition is that when the harmonic ratio load sensitivity deviates from the normal range by more than a preset alarm threshold, the following condition is met: ; in The current calculated harmonic ratio represents the load sensitivity, in units of... ; The mean value is within the normal range, in units of ; The standard deviation of the normal range, in units ; This is a threshold coefficient. When this condition is met, it is determined that there is a high-resistance grounding fault on the outgoing line.

[0044] In some embodiments, the determination also needs to meet a time-duration condition: the harmonic ratio load sensitivity deviates from the normal range and exceeds the alarm threshold for a consecutive preset number of daily load cycles. The preset number can be configured as, for example, 3 daily load cycles, meaning that the harmonic ratio load sensitivity meets the alarm condition for 3 consecutive days before a final determination is made. This time-duration condition is used to exclude single anomalies caused by occasional interference.

[0045] Continuing with the previous example, suppose the mean value of the harmonic ratio load sensitivity sequence during the 30-day normal operation of a certain circuit is... Standard deviation Threshold coefficient The alarm threshold is The harmonic ratio and load sensitivity of the aforementioned faulty circuit. Deviation is The value far exceeds the alarm threshold, thus meeting the judgment criteria.

[0046] In some embodiments, optionally, after obtaining the harmonic ratio load sensitivity of each outgoing line connected to the cable branch box, a multi-circuit lateral comparison can be further performed to assist in fault location. Specifically, the comparison logic includes: when the harmonic ratio load sensitivity of only one outgoing line deviates from the normal range by more than the alarm threshold while the remaining outgoing lines are within the normal range, the fault is determined to be located at that cable or its terminal joint; when the harmonic ratio load sensitivity of multiple outgoing lines simultaneously deviates from the normal range by more than the alarm threshold, and the relative difference between the harmonic ratio load sensitivities of each circuit is less than a preset proportion, the fault is determined to be located in the common insulation area of ​​the cable branch box, such as the deterioration of the busbar insulation to ground. The preset proportion can be configured, for example, 20%, and is determined by calculating the percentage of the difference between the maximum and minimum values ​​of the harmonic ratio load sensitivity of each abnormal circuit to the maximum value. The reason for multiple circuits being abnormal at the same time and having similar values ​​is that each circuit shares the same nonlinear fault path through a common point of common insulation deterioration.

[0047] Figure 3 This is a schematic diagram illustrating fault location based on load sensitivity of multi-loop harmonic ratios in one embodiment of this application. Figure 3 As shown, a common insulation area is set up in the cable branch box and multiple outgoing lines are connected. The harmonic ratio load sensitivity of each circuit is calculated and compared laterally with its deviation from the corresponding normal range. Figure 3 In (a), only the harmonic ratio load sensitivity of loop 2 exceeds the alarm threshold, while the other loops are within the normal range. This indicates that the abnormality only occurs along this loop, and it can be determined that the fault is located at the cable or terminal joint corresponding to loop 2. Figure 3 In (b), circuits 1 through 3 simultaneously exhibit abnormalities, and the harmonic ratio and load sensitivity values ​​of each abnormal circuit are close and relatively small, indicating that multiple circuits are affected by the same nonlinear fault channel. This suggests that the fault is located in the common insulation area of ​​the cable branch box. Through the above multi-circuit lateral comparison, branch circuit faults and common insulation faults can be further distinguished based on the completed high-resistance grounding fault identification.

[0048] Optionally, the degree to which the harmonic ratio load sensitivity deviates from the normal range can be graded for assessment. A first-level threshold and a second-level threshold are set, with the second-level threshold being greater than the first-level threshold; for example, the first-level threshold is set to... The second-level threshold is taken as follows: When the deviation is between the first and second level thresholds, it is marked as a minor anomaly, and it is recommended to increase the monitoring frequency; when the deviation exceeds the second level threshold, it is marked as a serious anomaly, and it is recommended to arrange maintenance.

[0049] In some embodiments, in addition to the above-mentioned online continuous monitoring based on daily load changes, optionally, the method also includes extracting insulation status information using the transient process generated when the feeder resumes power supply. After voltage is applied, the cable insulation material exhibits dielectric absorption current in addition to the instantaneous capacitance charging current and steady-state leakage current. This dielectric absorption current is generated by dipole orientation polarization and interface polarization in the insulation material, decays exponentially, and typically has a decay time constant of several seconds to tens of seconds. Moisture-affected or aged insulation contains highly polar substances such as water molecules, resulting in an increased initial amplitude of the dielectric absorption current and a longer decay time constant.

[0050] When the switch of the feeder where the cable branch box is located is closed, the distribution terminal enters the high-speed waveform recording mode with the closing signal as the trigger condition. The sampling rate is not less than 5kHz. The zero-sequence current transient waveform and three-phase voltage waveform of each outgoing line are recorded within a preset time (e.g., 60s) from the closing time. At the same time, the ambient temperature at the closing time is also recorded.

[0051] Subsequently, the capacitor charging component is subtracted. Specifically, the closing phase angle (i.e., the phase of the voltage waveform at the moment of closing) is determined based on the three-phase voltage waveform. A transient circuit model is established by combining the cable line's capacitance to ground parameters (obtained from the distribution network line parameter database) and system impedance parameters (system equivalent resistance and equivalent inductance). This model is a standard RLC second-order transient circuit. The theoretical capacitor charging current waveform is calculated based on the closing phase angle and circuit parameters. The residual current is obtained by subtracting the theoretical capacitor charging current waveform from the zero-sequence current transient waveform. In practical applications, the system impedance parameters can be fine-tuned near their nominal values ​​using the least squares method to ensure that the inrush current envelope calculated by the model matches the measured inrush current envelope as closely as possible within the first 200ms.

[0052] In one embodiment, the dielectric absorption current characteristic parameter is extracted from the residual current. Specifically, a data segment of the residual current is taken from the time interval between the completion of capacitor charging inrush current decay after closing (e.g., 0.5s) and the end of a preset time period (e.g., 60s). This data segment is then fitted using an exponential decay model. ; in Residual current, in amperes (A); This represents the initial amplitude of the dielectric absorption current, in amperes (A). The decay time constant is expressed in seconds. The current is the steady-state leakage current, in amperes (A). The time after closing the circuit breaker is in seconds. The fitting process employs the Levenberg-Marquardt iterative method, aiming to minimize the sum of squared residuals. This method is a standard algorithm for solving nonlinear least squares problems, achieving convergence by introducing a damping factor to control the iteration step size based on the Gauss-Newton method. Initial values ​​are set as follows: Pick , Pick , Take 10 seconds.

[0053] As an example, the residual current of a certain outgoing line is 0.15A at 0.5s, 0.08A at 10s, and 0.038A at 60s; after fitting, it is found that... A, s, A. Verification The model value at time s is A is basically consistent with the actual measurement.

[0054] In some embodiments, temperature correction is performed after extracting the initial amplitude and decay time constant of the dielectric absorption current, because the dielectric absorption current is affected by ambient temperature. This is based on the ambient temperature at the time of closing the circuit. and preset reference temperature (e.g. 20) C), the initial amplitude and decay time constant of the dielectric absorption current are corrected to their equivalent values ​​at the reference temperature: ; ; in This is the temperature correction factor for the initial amplitude of the dielectric absorption current, in units of... ; Temperature correction factor for decay time constant, in units The temperature correction coefficient was obtained through regression calibration using multiple power transmission data from the same circuit under different ambient temperatures, meaning the data collection spanned a temperature range of at least 15°C. C's multiple sets of data, respectively for and , and Performing linear regression yields the slope. and In the initial stage where sufficient calibration data is lacking, and The initial setting can be, for example, 0.03, based on the typical temperature characteristics of cross-linked polyethylene insulation materials. It will be updated as more data is accumulated.

[0055] In some embodiments, the insulation condition assessment method involves comparing the initial amplitude of the dielectric absorption current and the decay time constant at the reference temperature with a baseline established from the corresponding parameters during the initial normal operation of the circuit. The baseline is taken from parameters extracted and temperature-corrected during several power-on events in the initial stage of the circuit's commissioning, and the mean and standard deviation are calculated as the baseline reference. When the initial amplitude of the dielectric absorption current at the reference temperature exceeds the preset deviation range of the corresponding baseline, or the decay time constant at the reference temperature exceeds the preset deviation range of the corresponding baseline, it is determined that the insulation of the outgoing line of the circuit has deteriorated. The method for determining the preset deviation range is similar to the method for determining the alarm threshold of the harmonic ratio load sensitivity mentioned above, taking the baseline mean plus the baseline standard deviation multiplied by the alarm coefficient. The alarm coefficient can be initially set to, for example, 2.0, with a value range of 1.5 to 3.0, and is calibrated and updated as operating data accumulates. An increase in the initial amplitude of the dielectric absorption current reflects an increase in the content of polar substances in the insulation, and an increase in the decay time constant reflects enhanced interface polarization, both of which are typical characteristics of moisture-induced deterioration.

[0056] The two methods described above can operate in tandem. Specifically, the harmonic ratio load sensitivity method continuously monitors each outgoing line during daily operation; the dielectric absorption current feature extraction method is triggered and executed during each power-on operation. When the same circuit exhibits anomalies in both methods, the fault or degradation determination result for that circuit has a higher confidence level. Through load response characteristic analysis of the harmonic ratio load sensitivity, this scheme transforms high-resistance grounding fault detection from relying on amplitude thresholds to relying on the correlation characteristics between signals and operating conditions, achieving the identification of weak nonlinear fault signals without lowering the setting threshold. The power-on transient insulation assessment transforms transient processes neglected in routine operation and maintenance into a source of diagnostic information for insulation status. The combination of these two methods provides cable branch boxes with comprehensive online monitoring capabilities for coverage fault detection and degradation early warning.

[0057] Please see Figure 4 , Figure 4 This is a schematic block diagram of a high-resistance grounding fault detection system for a cable branch box according to this application. Figure 4 As shown, the system includes: The sampling and spectrum analysis module 401 is used to perform power frequency synchronous sampling of the zero-sequence current of each outgoing line of the cable branch box, perform spectrum analysis on the sampled data, extract the fundamental effective value and the third harmonic effective value of the zero-sequence current for each power frequency cycle, and synchronously acquire the effective value of the load current of each outgoing line. The harmonic ratio calculation module 402 is used to obtain the zero-sequence current harmonic ratio based on the ratio of the effective value of the third harmonic of the zero-sequence current to the effective value of the fundamental wave of the zero-sequence current. The time window statistics module 403 is used to calculate the average value of the zero-sequence current harmonic ratio and the average value of the load current in each time window, with a preset time window as the unit. The load sensitivity calculation module 404 is used to accumulate multiple sets of time window data covering the complete daily load cycle, and then perform linear regression calculation with the mean of the effective value of the load current in each time window as the independent variable and the mean of the zero-sequence current harmonic ratio as the dependent variable to obtain the regression slope. The regression slope is used as the harmonic ratio load sensitivity. The fault determination module 405 is used to compare the harmonic ratio load sensitivity with the normal range established based on the harmonic ratio load sensitivity under the historical normal operation state of the circuit. When the harmonic ratio load sensitivity deviates from the normal range by more than a preset alarm threshold, it is determined that there is a high-resistance grounding fault in the outgoing line.

[0058] The implementation process of the functions and roles of each module in the above system is detailed in the implementation process of the corresponding steps in the above method, and will not be repeated here.

[0059] Based on the same application concept as the above method, this application proposes an electronic device, see [link to application]. Figure 5 As shown, it includes: a processor 5001 and a machine-readable storage medium 5002, the machine-readable storage medium 5002 storing machine-executable instructions that can be executed by the processor 5001; the processor 5001 is used to execute the machine-executable instructions to implement the cable branch box high-resistance grounding fault detection method disclosed in the above example of this application.

[0060] Based on the same concept as the above-described method, this application also provides a machine-readable storage medium storing a plurality of computer instructions. When executed by a processor, these computer instructions can implement the cable branch box high-resistance grounding fault detection method disclosed in the above examples of this application. The aforementioned machine-readable storage medium can be any electronic, magnetic, optical, or other physical storage device, and can contain or store information such as executable instructions, data, etc.

[0061] Based on the same application concept as the above method, this application embodiment also provides a computer program product, which includes a computer program that, when executed by a processor, implements the cable branch box high-resistance grounding fault detection method disclosed in the above examples of this application.

[0062] In the above embodiments, the descriptions of each embodiment have different focuses. Parts not described in detail in a certain embodiment can be referred to in the relevant descriptions of other embodiments. The above descriptions are merely preferred embodiments of this application and explanations of the technical principles used. Those skilled in the art should understand that the scope of the invention involved in this application is not limited to the technical solutions formed by specific combinations of the above technical features, but should also cover other technical solutions formed by arbitrary combinations of the above technical features or their equivalent features without departing from the inventive concept.

Claims

1. A method for detecting high-resistance grounding faults in cable branch boxes, characterized in that, include: The zero-sequence current of each outgoing line of the cable branch box is sampled synchronously at power frequency. Spectrum analysis is performed on the sampled data to extract the fundamental effective value and the third harmonic effective value of the zero-sequence current for each power frequency cycle. The effective value of the load current of each outgoing line is obtained synchronously. The zero-sequence current harmonic ratio is obtained by the ratio of the effective value of the third harmonic of the zero-sequence current to the effective value of the fundamental wave of the zero-sequence current. Using a preset time window as the unit, calculate the mean value of the zero-sequence current harmonic ratio and the mean value of the load current effective value within each time window; After accumulating multiple sets of time window data covering the complete daily load cycle, linear regression calculation is performed with the mean of the effective value of the load current in each time window as the independent variable and the mean of the zero-sequence current harmonic ratio as the dependent variable to obtain the regression slope. The regression slope is used as the load sensitivity of the harmonic ratio. The harmonic ratio load sensitivity is compared with a normal range established based on the harmonic ratio load sensitivity under the historical normal operation of the circuit. When the harmonic ratio load sensitivity deviates from the normal range by more than a preset alarm threshold, it is determined that there is a high-resistance grounding fault on the outgoing line.

2. The method according to claim 1, characterized in that, The establishment of the normal range specifically includes: taking the harmonic ratio load sensitivity sequence obtained by the circuit operating continuously for a preset number of days under the confirmed normal insulation operating condition, and calculating the mean and standard deviation of the sequence; the alarm threshold is the standard deviation multiplied by a preset threshold coefficient.

3. The method according to claim 1, characterized in that, The determination that there is a high-resistance grounding fault on the outgoing line also requires that the harmonic ratio load sensitivity deviates from the normal range and exceeds the alarm threshold in a continuous preset number of daily load cycles.

4. The method according to claim 1, characterized in that, It also includes multi-circuit lateral comparison, specifically: comparing the harmonic ratio load sensitivity obtained from each outgoing line of the cable branch box; when the harmonic ratio load sensitivity of only one outgoing line deviates from the normal range by more than the alarm threshold, it is determined that the fault is located at the loop cable or its terminal joint; when the harmonic ratio load sensitivity of multiple outgoing lines deviates from the normal range by more than the alarm threshold at the same time, and the relative difference between the harmonic ratio load sensitivity of each circuit is less than a preset ratio, it is determined that the fault is located in the common insulation area of ​​the cable branch box.

5. The method according to claim 1, characterized in that, The spectrum analysis specifically includes: performing a discrete Fourier transform on the sampled data for each power frequency cycle, taking the magnitude of the Fourier coefficients corresponding to the fundamental frequency point and the third harmonic frequency point respectively, and calculating the effective value of the fundamental frequency of the zero-sequence current and the effective value of the third harmonic frequency of the zero-sequence current based on the magnitude values.

6. The method according to claim 1, characterized in that, It also includes transient insulation assessment of power transmission, specifically: when the switch of the feeder where the cable branch box is located is closed, the high-speed waveform recording mode is entered with the closing signal as the trigger condition, and the zero-sequence current transient waveform and three-phase voltage waveform of each outgoing line are recorded within a preset time period from the closing time, while the ambient temperature at the closing time is also recorded.

7. The method according to claim 6, characterized in that, It also includes the deduction of the capacitor charging component, specifically: determining the closing phase angle based on the three-phase voltage waveform, establishing a transient circuit model in combination with the ground capacitance parameters of the cable line and the system impedance parameters, and calculating the theoretical capacitor charging current waveform; Subtract the theoretical capacitor charging current waveform from the zero-sequence current transient waveform to obtain the residual current.

8. The method according to claim 7, characterized in that, It also includes the extraction of dielectric absorption current features, specifically: taking the data segment of the residual current within the time interval from the end of the capacitor charging inrush current decay after the circuit breaker is closed to the end of the preset time period, fitting the data segment with an exponential decay model, and extracting the initial amplitude and decay time constant of the dielectric absorption current.

9. The method according to claim 8, characterized in that, It also includes temperature correction, specifically: based on the ambient temperature at the closing time and the preset reference temperature, the initial amplitude of the dielectric absorption current and the decay time constant are respectively corrected by temperature to obtain the initial amplitude of the dielectric absorption current and the decay time constant at the reference temperature; the temperature correction coefficient used for temperature correction is obtained by regression calibration through multiple power supply data of the same circuit at different ambient temperatures.

10. A high-resistance grounding fault detection system for cable branch boxes, characterized in that, include: The sampling and spectrum analysis module is used to synchronously sample the zero-sequence current of each outgoing line of the cable branch box at the power frequency, perform spectrum analysis on the sampled data, and extract the effective value of the fundamental wave and the effective value of the third harmonic of the zero-sequence current for each power frequency cycle; and synchronously acquire the effective value of the load current of each outgoing line. The harmonic ratio calculation module is used to obtain the zero-sequence current harmonic ratio based on the ratio of the effective value of the third harmonic of the zero-sequence current to the effective value of the fundamental wave of the zero-sequence current. The time window statistics module is used to calculate the mean value of the zero-sequence current harmonic ratio and the mean value of the load current effective value in each preset time window. The load sensitivity calculation module is used to accumulate multiple sets of time window data covering the complete daily load cycle, and then perform linear regression calculation with the mean of the effective value of the load current in each time window as the independent variable and the mean of the zero-sequence current harmonic ratio as the dependent variable to obtain the regression slope. The regression slope is used as the harmonic ratio load sensitivity. The fault determination module is used to compare the harmonic ratio load sensitivity with a normal range established based on the harmonic ratio load sensitivity under the historical normal operation of the circuit. When the harmonic ratio load sensitivity deviates from the normal range by more than a preset alarm threshold, it is determined that there is a high-resistance grounding fault on the outgoing line.

Images