Intelligent Fault Diagnosis Method for Distribution Lines Based on Multi-Source Information Fusion

Abstract

The present invention discloses an intelligent fault diagnosis method for distribution lines based on multi-source information fusion, belonging to the technical field of power system fault detection and diagnosis. It includes obtaining distribution line data and meteorological data; generating a spatio-temporal fusion feature matrix based on the wire spatial displacement amount by superimposing the distribution line data and the meteorological data; dynamically adjusting a preset initial threshold using the spatio-temporal fusion feature matrix; performing multiple rounds of verification on the abnormal fluctuations of the electrical parameters of the line according to the fault discrimination threshold, generating a fault location instruction, and performing real-time matching with the line topological structure, correcting the fault section priority queue based on the switch action feedback data of the feeder terminal, and triggering a fault removal action matching the priority. The present invention adopts a spatio-temporal feature fusion and dynamic threshold adjustment mechanism for current, voltage, meteorological and spatial displacement data, can accurately identify the fault type and real-time locate the fault section, and significantly improve the diagnosis accuracy and response speed in complex environments.

Description

Technical Field

[0001] The present invention relates to the technical field of power system fault detection and diagnosis, and particularly to an intelligent diagnosis method for distribution line faults based on multi-source information fusion. Background Art

[0002] Traditional distribution line fault diagnosis methods mainly rely on the judgment of single electrical signal thresholds such as current and voltage, and there is a problem of insufficient data dimension. When the line is affected by environmental temperature and humidity changes or mechanical deformation interference, fixed thresholds are prone to misjudgment. For example, the fluctuation of wire resistance affected by temperature may cause overcurrent false alarms, and the swing of the wire caused by strong wind may be misidentified as a short circuit fault.

[0003] In the prior art, abnormal situations are usually processed by comparing regular inspections with a preset rule library. For example, offline fault recording data is used to match typical waveforms, or the status of fault indicators is verified manually. Such methods need to rely on historical experience data and have a large response delay, and cannot meet the requirements of rapid judgment of dynamic changes in line topology and instantaneous faults.

[0004] Traditional solutions have obvious defects when facing multi-factor coupled faults. Hidden defects such as mechanical deformation and insulation aging are difficult to be identified in time through a single electrical parameter, and sudden changes in meteorological conditions may cover up early fault characteristics. Manual inspection is inefficient and cannot cover real-time monitoring, resulting in a long fault troubleshooting cycle and affecting power supply reliability. Summary of the Invention

[0005] To solve the above problems, the present invention provides an intelligent diagnosis method for distribution line faults based on multi-source information fusion, which adopts a spatio-temporal feature fusion and dynamic threshold adjustment mechanism for current, voltage, meteorological and spatial displacement data, can accurately identify the fault type and locate the fault section in real time, and significantly improves the diagnosis accuracy and response speed in complex environments.

[0006] The above object can be achieved by the following solutions:

[0007] Intelligent fault diagnosis method for distribution lines based on multi-source information fusion, including acquiring current signals, voltage signals and temperature signals of distribution lines, and synchronously collecting the humidity gradient change rate in meteorological data and the wire spatial displacement in geographical coordinate data to generate a multi-dimensional heterogeneous data set; calculating the entropy value of the time-domain distortion characteristics of the current signal and the voltage signal, and superimposing the humidity gradient change rate to generate a dynamic entropy change parameter, and at the same time performing spatial morphology calibration on the dynamic entropy change parameter based on the wire spatial displacement to generate a spatio-temporal fusion feature matrix; using the spatio-temporal fusion feature matrix to dynamically adjust a preset initial threshold; wherein, the adjustment process includes calculating the wire resistance change rate according to the real-time temperature data in the meteorological data, and performing weighted compensation on the resistance change rate through a sliding window algorithm to generate an adaptive fault discrimination threshold; performing multi-round verification on the abnormal fluctuation of the electrical parameters of the line according to the fault discrimination threshold, and when the fluctuation amplitude in three consecutive sampling periods exceeds the fault discrimination threshold and the high-frequency harmonic ratio reaches a preset threshold, generating a fault location instruction with a confidence weight; matching the fault location instruction with the line topology structure in real time, and correcting the fault section priority queue based on the switch action feedback data of the feeder terminal, and triggering a fault removal action matching the priority.

[0008] Optionally, the generating the spatio-temporal fusion feature matrix includes: extracting the time-domain waveform distortion feature of the current signal and the third-harmonic mutation rate of the voltage signal, and calculating the composite distortion entropy value of the two; superimposing the logarithmic attenuation coefficient of the humidity gradient change rate on the composite distortion entropy value to generate a dynamic entropy change parameter; performing morphological erosion processing on the dynamic entropy change parameter based on the fluctuation trajectory of the wire spatial displacement, and outputting the spatio-temporal fusion feature matrix.

[0009] Optionally, the generating the adaptive fault discrimination threshold includes: acquiring the historical mean data of the wire resistance change rate, and fitting the resistance change curve based on the real-time temperature data; combining the spatio-temporal fusion feature matrix, and using a dynamic Bayesian network to identify abnormal segments of the resistance change curve to generate a compensation weight factor; performing a convolution operation on the compensation weight factor and the initial threshold to generate a fault discrimination threshold including the dynamic coupling relationship of environmental parameters.

[0010] Optionally, the multi-round verification of the abnormal fluctuation of the electrical parameters of the line according to the fault discrimination threshold includes: when the first fluctuation amplitude exceeds the fault discrimination threshold, collecting the hydrogen volume fraction data of the distribution transformer oil chromatogram; if the hydrogen volume fraction exceeds the preset fraction threshold and the high-frequency harmonic ratio reaches the threshold, triggering a partial discharge pre-judgment signal; adjusting the fault discrimination threshold for the next two sampling periods according to the pre-judgment signal.

[0011] Optionally, the generation of the fault location instruction with confidence weights includes: obtaining the switch status data of the feeder terminal, and inversely calculating the fault current propagation path based on the switch status data; performing an overlap analysis on the propagation path and the abnormal area of the spatio-temporal fusion feature matrix, and increasing the confidence weight when the proportion of the overlapping area is greater than a preset proportion threshold; if there are multiple candidate sections, calculating the correlation coefficient between the change rate of the dynamic entropy change parameter and the wire spatial displacement amount, and selecting the section according to the magnitude of the correlation coefficient.

[0012] Optionally, the method further includes: collecting the difference in line capacitance current before and after fault excision, and performing similarity matching with the standard difference in a preset fault feature database; generating a fault type label according to the matching result, associating the fault type label with the environmental parameters, and appending the result to a preset historical fault case database; updating the default value of the subsequent initial threshold based on the fault frequency distribution under the same geographical coordinates in the historical fault case database.

[0013] Optionally, the updating of the default value of the subsequent initial threshold includes: extracting the environmental characteristic parameters and action delay data corresponding to the fault type label to generate a four-dimensional spatio-temporal feature vector; using a time series analysis algorithm to perform trend prediction on the four-dimensional spatio-temporal feature vector to generate a dynamic threshold recommended value within a preset future time period; superimposing the dynamic threshold recommended value and the real-time meteorological warning data, and determining the final default value of the initial threshold through the gradient descent method.

[0014] Optionally, the method for obtaining the wire spatial displacement amount includes: real-time scanning of the spatial coordinates of the distribution line corridor by lidar to generate a three-dimensional displacement trajectory of the wire suspension point; performing time synchronization matching on the displacement mutation amount of the three-dimensional displacement trajectory and the fluctuation period of the spatio-temporal fusion feature matrix; when the correlation coefficient between the displacement mutation amount and the electrical parameter fluctuation is greater than a preset correlation threshold, triggering a high-risk warning signal for line mechanical deformation.

[0015] Optionally, the triggering of the high-risk warning signal for line mechanical deformation includes: inputting the spatio-temporal coordinates corresponding to the high-risk warning signal into a preset insulator aging prediction model to calculate the remaining mechanical strength attenuation rate; adjusting the inspection cycle of this section according to the attenuation rate, and synchronously increasing the threshold value.

[0016] Based on the same inventive concept, the present invention also provides an intelligent fault diagnosis system for distribution lines based on multi-source information fusion. The system includes: a data acquisition module, configured to obtain current signals, voltage signals, and temperature signals of the distribution line, and simultaneously collect the humidity gradient change rate in meteorological data and the wire spatial displacement amount in geographic coordinate data to generate a multi-dimensional heterogeneous data set; a feature construction module, configured to calculate the entropy value of the time-domain distortion features of the current signal and the voltage signal, and superimpose the humidity gradient change rate to generate a dynamic entropy change parameter. At the same time, based on the wire spatial displacement amount, spatial form calibration is performed on the dynamic entropy change parameter to generate a spatio-temporal fusion feature matrix; an initial threshold adjustment module, configured to dynamically adjust a preset initial threshold by using the spatio-temporal fusion feature matrix; wherein, the adjustment process includes calculating the wire resistance change rate according to the real-time temperature data in the meteorological data, and performing weighted compensation on the resistance change rate through a sliding window algorithm to generate an adaptive fault discrimination threshold; a fault judgment module, configured to perform multiple rounds of verification on the abnormal fluctuation of the electrical parameters of the line according to the fault discrimination threshold. When the fluctuation amplitude in three consecutive sampling periods exceeds the fault discrimination threshold and the high-frequency harmonic ratio reaches a preset threshold, a fault location instruction with a confidence weight is generated; a fault processing module, configured to perform real-time matching of the fault location instruction with the line topology structure, and correct the fault section priority queue based on the switch action feedback data of the feeder terminal, and trigger a fault removal action matching the priority.

[0017] Compared with the prior art, the present invention has the following advantages:

[0018] 1. The present invention realizes the improvement of the accuracy and efficiency of distribution line fault diagnosis; by integrating electrical parameter data such as current, voltage, and temperature and meteorological information, a dynamic entropy change parameter and a spatio-temporal fusion feature matrix are established, solving the limitations of a single data source, and being able to effectively distinguish environmental interference from real fault signals; combining the high-frequency harmonic detection and humidity compensation mechanisms of multi-dimensional data analysis, the anti-interference ability under complex meteorological conditions such as humidity and high temperature is improved, and the misjudgment rate is significantly reduced;

[0019] 2. The method of adaptive threshold adjustment combining the sliding window algorithm and the dynamic Bayesian network is adopted to realize the real-time compensation of the wire resistance change rate, overcoming the problem of sensitivity loss caused by temperature fluctuation of the traditional fixed threshold; through three sampling period verifications and confidence weight determination, the reliability and real-time performance of fault identification are ensured, and false actions caused by instantaneous disturbances are avoided;

[0020] 3. By matching the fault location instruction with the real-time line topology structure and combining the switch feedback data of the feeder terminal, the priority queue of the fault section is dynamically corrected; this method optimizes the fault removal logic, shortens the location time, especially in the complex scenario of branch lines, can accurately lock the fault point and preferentially remove the high-risk section, and improves the power supply restoration speed.

[0021] Other features and advantages of the present invention will be described in the following specification, and, in part, will be obvious from the specification, or will be understood by implementing the present invention. The objectives and other advantages of the present invention can be achieved and obtained by the structures pointed out in the specification, claims, and drawings. Brief Description of the Drawings

[0022] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0023] Figure 1 is a schematic flowchart of the intelligent diagnosis method for distribution line faults based on multi-source information fusion according to an embodiment of the present invention.

[0024] Figure 2 is a heat map of the spatio-temporal fusion feature matrix according to an embodiment of the present invention.

[0025] Figure 3 is a curve graph of the dynamic threshold adjustment process according to an embodiment of the present invention.

[0026] Figure 4 is a three-dimensional wire displacement trajectory diagram according to an embodiment of the present invention.

[0027] Figure 5 is a schematic structural diagram of the intelligent diagnosis system for distribution line faults based on multi-source information fusion according to an embodiment of the present invention. Detailed Embodiments

[0028] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0029] Refer to Figure 1, an embodiment of the present invention proposes an intelligent fault diagnosis method for distribution lines based on multi-source information fusion. By adopting a spatio-temporal feature fusion and dynamic threshold adjustment mechanism for current, voltage, meteorological, and spatial displacement data, it can accurately identify the fault type and real-time locate the fault section, significantly improving the diagnosis accuracy and response speed in complex environments.

[0030] The method of this embodiment specifically includes:

[0031] Obtain the current signal, voltage signal, and temperature signal of the distribution line, and simultaneously collect the humidity gradient change rate in meteorological data and the wire spatial displacement amount in geographical coordinate data to generate a multi-dimensional heterogeneous data set;

[0032] Specifically, the humidity gradient change rate is the change rate of environmental humidity per unit time, reflecting the risk of condensation on the surface of insulating materials. The wire spatial displacement amount is the amplitude of the position change of the wire in three-dimensional space, characterizing the mechanical stability of the line under external forces. The heterogeneous data set is a set of multi-physical field data with time series alignment, covering four dimensions of electricity, heat, humidity, and mechanics.

[0033] Calculate the entropy value of the time-domain distortion characteristics of the current signal and the voltage signal, and superimpose the humidity gradient change rate to generate a dynamic entropy change parameter. At the same time, calibrate the dynamic entropy change parameter based on the wire spatial displacement amount to generate a spatio-temporal fusion feature matrix;

[0034] Specifically, the time-domain distortion characteristic is the degree of waveform deviation from the standard sine wave, used to quantify the current distortion caused by overload or short circuit. Entropy value calculation reflects the signal complexity through information entropy, effectively distinguishing between steady-state load and fault transient. The dynamic entropy change parameter is a composite index that combines electrical anomalies and environmental humidity effects. As Figure 2 shown, the spatio-temporal fusion feature matrix shows the change of the dynamic entropy parameter with time and spatial displacement through color mapping. The light-colored area represents a high correlation between electrical anomalies and environmental factors.

[0035] Dynamically adjust the preset initial threshold using the spatio-temporal fusion feature matrix; among them, the adjustment process includes calculating the wire resistance change rate according to the real-time temperature data in the meteorological data, and performing weighted compensation on the resistance change rate through a sliding window algorithm to generate an adaptive fault discrimination threshold;

[0036] Specifically, the sliding window algorithm performs a moving average process on the resistance change in consecutive time periods to smooth the influence of instantaneous temperature fluctuations. The adaptive fault discrimination threshold is a dynamic threshold value that changes with temperature, avoiding the failure of a fixed threshold in extreme weather. As Figure 3 shown, after the initial threshold is adjusted by the sliding window algorithm, the fault criterion is accurately triggered in the 12-15 hour period, avoiding misjudgment of transient disturbances.

[0037] Perform multiple rounds of verification on the abnormal fluctuations of the electrical parameters of the line according to the fault discrimination threshold. When the fluctuation amplitude in three consecutive sampling periods exceeds the fault discrimination threshold and the high-frequency harmonic ratio reaches the preset threshold, generate a fault location instruction with confidence weight;

[0038] Match the fault location instruction with the line topology structure in real time, and correct the fault section priority queue based on the switch action feedback data of the feeder terminal, triggering a fault removal action matching the priority.

[0039] Specifically, perform the shortest path matching between the suspected fault nodes in the location instruction and the line topology diagram of the SCADA system, and preferentially screen the three candidate sections with the shortest distance. Receive the switch opening and closing states uploaded by the feeder terminal. If the adjacent switches of a section have moved, reduce its priority. Finally, select the section with the highest priority to trigger the circuit breaker to trip and send a location report to the operation and maintenance terminal.

[0040] Construct a fault diagnosis model through spatio-temporal multi-dimensional data fusion, and dynamically adjust the threshold to adapt to complex environmental changes. The coupling calibration of electrical parameters and mechanical displacement effectively suppresses interference signals, and the correction of humidity data improves the sensitivity to insulation moisture faults. After implementation, it can improve the accuracy of fault identification, shorten the location time, and reduce the risk of misoperation in extreme weather.

[0041] Optionally, the generation of the spatio-temporal fusion feature matrix includes:

[0042] Extract the time-domain waveform distortion feature of the current signal and the third-harmonic mutation rate of the voltage signal, and calculate the composite distortion entropy value of the two;

[0043] Specifically, first collect the original current signal of the distribution line through the current transformer on the high-voltage side, and input this signal into a band-pass filter to remove low-frequency interference outside 50Hz power frequency. Use a differential operator to perform time-domain waveform analysis on the filtered current signal, calculate the distortion amplitude at each sampling point and the change amount of the time interval between adjacent wave peaks and wave valleys to obtain the current time-domain distortion feature. At the same time, perform a fast Fourier transform on the voltage signal collected by the voltage transformer, extract the ratio of the amplitude of the third-harmonic component to the fundamental wave, and use the rising slope of this ratio within a unit time as the third-harmonic mutation rate. The composite distortion entropy value is calculated through the following formula: , where is the number of sampling points, is the deviation value between the instantaneous current amplitude at the th sampling point and the standard sine wave, is the current effective value, is the time-varying derivative of the third-harmonic voltage. This formula combines the energy distribution of the current waveform distortion with the dynamic characteristics of the voltage harmonic variation, and characterizes the abnormal degree of the power quality of the line.

[0044] Superimpose the composite distortion entropy value on the logarithmic decay coefficient of the humidity gradient change rate to generate a dynamic entropy change parameter;

[0045] Superimpose the composite distortion entropy value on the logarithmic decay coefficient of the humidity gradient change rate to generate a dynamic entropy change parameter;

[0046] Specifically, collect the time series data of the environmental humidity sensor, calculate the absolute value of the humidity change rate within the last 30 minutes, and generate a humidity gradient change curve through piecewise linear interpolation. Use the natural logarithm function to perform non-linear scaling processing on the humidity gradient to generate a logarithmic decay coefficient. For the decay coefficient , there is: , where is the humidity change rate. Superimpose this coefficient on the composite distortion entropy value in proportion to obtain a dynamic entropy change parameter. For the dynamic entropy change parameter , there is: , where is the adjustment coefficient of the humidity change rate. This operation adds dynamic correction of the influence of humidity change on the conductivity of insulating materials while maintaining the original entropy value reflecting the characteristics of electrical parameters, and effectively improves the sensitivity to surface leakage current in high-humidity scenarios.

[0047] Perform morphological erosion processing on the dynamic entropy change parameter based on the fluctuation trajectory of the wire spatial displacement amount, and output a spatio-temporal fusion feature matrix.

[0048] Specifically, use a lidar scanning device to record the three-dimensional coordinates of the wire suspension point every 0.1 seconds, calculate the spatial displacement components at adjacent time points through difference calculation, and generate a wire fluctuation trajectory. Corresponding the dynamic entropy change parameter to the displacement trajectory points according to the sampling time, and construct a time-domain and space-domain joint distribution matrix. Use a 3×3 structuring element in the morphological erosion algorithm to traverse the matrix. The specific operations include: for each matrix element, compare its value with the values in the adjacent eight positions, and replace the element value with the minimum value in the neighborhood. After erosion, isolated noise peaks (such as abnormal entropy values caused by wire swings due to short-term strong winds) are suppressed, and the feature regions consistent with the physical deformation trend of the line are retained, and finally a spatio-temporal fusion feature matrix is generated.

[0049] For example, monitoring equipment is installed on 10kV distribution lines in coastal typhoon-prone areas. When the instantaneous wind speed reaches 15m / s, the conductors produce spatial displacement due to violent swings, and the strong wind carries salt mist, causing a sudden increase in humidity. The current transformer detects the current distortion caused by the sudden increase in load, and the voltage transformer records that the amplitude of the third harmonic increases from 1% to 5% within 2 seconds. The current waveform distortion feature extraction shows that the distortion rate is 12%, the voltage third harmonic mutation rate is 4% / s, the composite distortion entropy value is calculated to be 2.3, the ambient humidity increases from 60% to 95% within 5 minutes, the logarithmic attenuation coefficient of the humidity gradient change rate is 0.8, and the corrected dynamic entropy change parameter increases to 2.3×1.24≈2.85. The laser radar measures that the horizontal displacement trajectory of the conductor fluctuates periodically in the same period. After morphological corrosion treatment eliminates accidental displacement spikes, only the displacement area synchronized with the current harmonic surge is retained in the space-time matrix. This method effectively distinguishes the conductor swing interference caused by typhoons from actual insulation degradation faults. Traditional methods would misjudge current distortion caused by strong winds as short circuit faults, but through humidity correction and spatial corrosion processing, the system filters out the instantaneous disturbance effect of salt fog humidity on electrical parameters, and confirms the real fault based on the time synchronization of displacement trajectory and harmonic changes. Operators can accurately locate a section of line with reduced insulation due to salt fog corrosion, avoiding the waste of resources caused by power outages for maintenance.

[0050] Optionally, generating an adaptive fault discrimination threshold comprises:

[0051] Obtain historical average data of the wire resistance change rate, and fit the resistance change curve based on the real-time temperature data;

[0052] Specifically, retrieve the wire resistance value and corresponding ambient temperature records of the target line in the last 30 days from the cloud storage module, calculate the correlation coefficient between the daily average temperature and the resistance fluctuation, and generate the historical resistance change mean sequence. ,in, is the number of historical resistance change means, is the temperature range (each 5°C is a segment), In the same temperature range The offset of the wire resistance relative to the nominal value. Real-time data from the wire surface temperature sensor is collected, and the resistance change rate is calculated in real time according to the resistance temperature coefficient formula of the wire material. Resistance change rate at time ,have:

[0053] ,

[0054] in, The wire is at the reference temperature The nominal resistance value under is the temperature coefficient of resistance, is the temperature of the wire surface at a moment. Using the least squares method, the calculated in real time is fitted with the historical mean value to construct a cubic polynomial curve of temperature-resistance change for predicting the theoretical change range of the wire resistance under the current environment.

[0055] Combined with the spatio-temporal fusion feature matrix, a dynamic Bayesian network is used to identify the abnormal segments of the resistance change curve and generate a compensation weight factor;

[0056] Specifically, the spatio-temporal fusion feature matrix is input into the dynamic Bayesian network in a time-slice manner. The network contains hidden state nodes (indicating normal / abnormal wire state) and observation nodes (composed of parameters such as the entropy value of electrical parameters and displacement correlation in the spatio-temporal matrix). The state transition probability is defined as:

[0057] where, is the first probability value, is the second probability value. The observation probability is calculated through the deviation degree between the data of each dimension of the spatio-temporal matrix and the current resistance change curve. For the deviation degree between the data of each dimension of the spatio-temporal matrix and the current resistance change curve at time

[0058] , there is:

[0059] where, is the expected value of dimension for calculating the deviation degree between the observed value and the expected value, is the dynamic entropy change parameter at time, is the cubic polynomial curve function of temperature-resistance change.

[0060] If is abnormal, then:

[0061] ,

[0062] where, is the sensitivity coefficient, is the natural constant. According to the Viterbi algorithm, the maximum probability state sequence is inferred. When 3 consecutive time slices are abnormal, mark this segment of the resistance change curve as an abnormal segment and generate a compensation weight factor :

[0063] ,

[0064] In the formula, is the initial time of the abnormal segment, is the end time of the abnormal segment, is for the moment from to definite integral.

[0065] Convolve the compensation weight factor with the initial threshold to generate a fault discrimination threshold including the dynamic coupling relationship of environmental parameters.

[0066] Specifically, input the initial threshold (determined by 120% of the rated current of the line), convolve it with the compensation weight factor. For the fault discrimination threshold at the moment , there is:

[0067] ,

[0068] In the formula, represents the offset relative to the current time slice , is the Gaussian kernel function, is the sliding window width. This operation weights and fuses the thresholds within the current moment and its front and rear time windows with the weight factor, enabling the threshold to dynamically reflect the coupling effects of abnormal wire resistance and environmental parameters (temperature, humidity).

[0069] Exemplarily, when a cold snap suddenly strikes in winter in the north, the temperature of a 110 kV line drops from -5°C to -20°C within 1 hour. Through historical data, it is found that the average resistance of this line increases by 8% compared to the standard value at -20°C, and the fitting curve predicts 7.5%. However, the actual sensor measurement is 14%, deviating from the predicted value. The spatio-temporal fusion matrix shows that the increase in the current entropy value is abnormal and the wire displacement is stable (no sign of icing). The dynamic Bayesian network determines that this change in resistance does not conform to the normal expansion caused by low temperature, triggering the generation of the abnormal compensation weight factor . The initial threshold is adjusted to through convolution operation. 0.95 is the value of the Gaussian kernel function, significantly improving the sensitivity of overcurrent detection in this section. The real reason for the abnormal increase in wire resistance during the cold snap is that the contact resistance increases due to oxidation of a certain joint, rather than simply a temperature drop. The traditional method only relies on temperature compensation and sets the threshold to 2300 A, which may not be able to detect this hidden danger (the actual fault current is 2500 A). After using the dynamic Bayesian network to identify the irreversible abnormality of the resistance change, the system accurately adjusts the threshold to 2593 A, sensitively triggering an alarm and locating the hidden danger point. Thus, it avoids the fusing accident caused by overheating of the joint and effectively distinguishes different influence modes of environmental factors and equipment deterioration.

[0070] Optionally, the multi-round verification of the abnormal fluctuation of the electrical parameters of the line according to the fault discrimination threshold includes:

[0071] When the amplitude of the first fluctuation exceeds the fault discrimination threshold, collect the hydrogen gas volume fraction data of the distribution transformer oil chromatogram;

[0072] Specifically, when the electrical parameter fluctuation first breaks through the fault discrimination threshold after dynamic adjustment, the system immediately triggers the oil chromatogram on-line monitoring device to start gas sampling. The hydrogen gas dissolved in the transformer oil is separated by an oil-gas separation membrane, and its volume fraction in the mixed gas is measured by a gas sensor. The data of three consecutive cycles are continuously collected in this process, and the median is taken as the current hydrogen gas volume fraction. If at least two of the hydrogen gas volume fractions in three consecutive samplings are greater than the fraction threshold, it is determined as an abnormal state.

[0073] If the hydrogen gas volume fraction exceeds the preset fraction threshold and the high-frequency harmonic proportion reaches the preset threshold, trigger the partial discharge pre-judgment signal;

[0074] Specifically, compare the current hydrogen gas volume fraction with the preset threshold , and at the same time calculate the total energy proportion of high-frequency harmonics above 2 kHz in the voltage signal through Fourier transform . Assuming the threshold is 5%, if it satisfies and , generate the partial discharge pre-judgment signal. The intensity of the pre-judgment signal is graded according to the following formula:

[0075] ,

[0076] where is the minimum value function. The high-frequency harmonic proportion is the ratio of the high-frequency component to the total harmonic distortion, reflecting the change of the spectrum characteristics caused by the discharge pulse. The partial discharge pre-judgment signal is a risk level index characterizing the early discharge activity caused by insulation defects.

[0077] Adjust the fault discrimination threshold for the subsequent two sampling periods according to the pre-judgment signal.

[0078] Specifically, use the exponential decay model to dynamically correct the subsequent threshold. For the corrected fault discrimination threshold at time , there is: , where is an adjustment coefficient, which can be 0.15. This adjustment reduces the threshold when the system detects the risk of partial discharge and improves the sensitivity to weak fault characteristics. The exponential decay model is a mathematical correction method that scales down the threshold proportionally based on the predicted signal strength, ensuring that the system enters the high-sensitivity monitoring mode in high-risk scenarios.

[0079] Exemplarily, the oil chromatogram monitoring of a transformer in a suburban substation in a certain city shows that the volume fraction of hydrogen gas continuously remains at 0.6% - 0.7%, and the proportion of high-frequency harmonics in the voltage reaches 7%. Judging by the median of 0.65% exceeds the standard; the proportion of high-frequency harmonics of 7% exceeds the threshold of 5%, triggering a pre-judgment signal of partial discharge ; the original dynamic threshold of 300A is adjusted to 300×(1 - 0.15×1) = 255A; abnormal currents with an amplitude of 265A are captured in the subsequent two samplings, and it is confirmed as a creeping discharge fault on the surface of the bushing. Early identification of hidden creeping discharge faults that cannot be detected by traditional current thresholds, avoiding the all-station power outage accident caused by insulation breakdown.

[0080] 5 Optionally, the generation of the fault location instruction with confidence weight includes:

[0081] Obtain the switch status data of the feeder terminal, and reverse calculate the fault current propagation path based on the switch status data;

[0082] Specifically, obtain the current opening and closing status and action timestamps of each sectional switch through the FTU (Feeder Terminal Unit), and establish an adjacency matrix in combination with the line topology structure. Using the reverse path tracing algorithm:

[0083] ,

[0084] wherein, is the adjacency matrix, is the fault current vector. The algorithm recursively searches upstream from the end node where the fault current is detected to generate a set of possible propagation paths . The reverse path tracing algorithm is a graph theory method for analyzing the possible sources of fault current through the switch status. The adjacency matrix is a two-dimensional boolean array describing the connection relationship of line nodes.

[0085] Perform an overlap degree analysis on the propagation path and the abnormal area of the spatio-temporal fusion feature matrix, and increase the confidence weight when the proportion of the overlapping area is greater than the preset proportion threshold;

[0086] Specifically, extract the region coordinates in the spatio-temporal fusion feature matrix whose entropy value exceeds the safety critical value, and map them to the line topology coordinate system to form a polygon abnormal area. Calculate the spatial overlap degree between each propagation path and the polygon abnormal area. For the th propagation path, the spatial overlap degree , there is:

[0087] In the formula, the th propagation path, is a polygonal abnormal area. For the proportion threshold , it can be 70%. If , then increase the confidence weight of this path. For the confidence weight of the th propagation path, there is:

[0088] ,

[0089] In the formula, is the adjustment coefficient of the propagation path. The overlapping area ratio is a quantitative index of the geometric overlapping degree between the path and the abnormal area, representing the spatial correlation of the fault location. The confidence weight is a dynamic weighting coefficient based on the overlapping degree, used for priority sorting.

[0090] If there are multiple candidate sections, calculate the correlation coefficient between the change rate of the dynamic entropy change parameter and the wire spatial displacement amount, and select the section according to the magnitude of the correlation coefficient.

[0091] Specifically, when there are multiple candidate sections, that is, there are candidate paths with the same weight. Calculate the correlation coefficient between the wire spatial displacement amount corresponding to each path section and the change rate of the dynamic entropy change parameter of the electrical parameter. For the correlation coefficient of the th candidate path, there is:

[0092] ,

[0093] In the formula, is the covariance function, is the wire spatial displacement amount of the section corresponding to the th candidate path, is the change rate of the dynamic entropy change parameter of the electrical parameter, is the standard deviation of the wire spatial displacement amount, is the standard deviation of the change rate of the dynamic entropy change parameter of the electrical parameter. Select the section with the largest value as the final positioning target, and promote its priority to the head of the queue. The correlation coefficient is a statistical index to measure the synchronization of wire mechanical deformation and electrical abnormality, preventing misselection of pseudo-fault paths caused only by switch misoperation. Avoiding positioning deviation caused by a single criterion, locking the real fault source by comprehensively considering spatial and physical characteristics, and reducing ineffective inspections.

[0094] Optionally, the method further includes:

[0095] Collect the difference in line capacitance current before and after the fault removal, and perform a similarity match with the standard difference in the preset fault feature database;

[0096] Specifically, after the line is powered off, use a capacitive current transformer to record the capacitance current waveforms in the 5 cycles before and 10 cycles after the fault removal respectively, calculate the difference through integration, and for the capacitance current difference , there is:

[0097] ,

[0098] Among them, is the standard time window, such as a 0.1-second time window, to is the time period before the fault removal, to is the time period after the removal, is the capacitance current waveform before the fault removal, is the capacitance current waveform after the fault removal, and is calculated with the nominal values of various fault types in the database, such as short circuit, grounding, and disconnection for the cosine similarity. For the similarity , there is:

[0099] ,

[0100] If the similarity is greater than the preset threshold, it is determined to be the same type of fault mode. The capacitance current difference is the energy change amount of the capacitive current before and after the fault, reflecting the change in the line charge distribution state. The cosine similarity is used to quantify the matching degree between the current difference and the historical pattern.

[0101] Generate a fault type label according to the matching result, and append the fault type label associated with the environmental parameters to the preset historical fault case library;

[0102] Specifically, if the capacitance current difference matches the short-circuit nominal value, generate a "short circuit between lines" label, and combine this label with the environmental temperature, humidity, and wind speed data at the time of the fault into a tuple. Store this tuple in the NoSQL database of the historical case library according to the timestamp, and adopt a sharding storage strategy to improve the retrieval efficiency. The fault type label is a classification code, identifying the specific physical cause of the fault. The environmental parameter association is to bind and store the meteorological conditions and the fault type for subsequent pattern analysis.

[0103] Update the default value of the subsequent initial threshold based on the fault frequency distribution under the same geographical coordinates in the historical fault case library.

[0104] Specifically, the line area is divided according to the geographical grid, and the average monthly occurrence times of the same type of faults in each grid are counted. The exponential weighted moving average model is used to adjust the initial threshold For the updated initial threshold There is:

[0105] ,

[0106] In the formula, is the decay coefficient of the average monthly occurrence times, is the average monthly occurrence times, is the natural constant.

[0107] Exemplarily, 6 "wet flash grounding faults" occurred in a grid of a coastal industrial zone in the past three months. Calculate the difference in capacitance current and the similarity with the nominal value of wet flash Similarity , the matching is successful; when adding tags, the data of humidity 92% and temperature 28°C are associated, and the initial threshold is updated according to the average monthly occurrence times times / month to obtain . When a similar fault occurs in the next month, the alarm is triggered 200 ms in advance due to the decrease of the threshold. Significantly improve the fault warning speed in high-incidence areas and avoid the expansion of faults caused by long-term exposure of equipment to corrosive and humid environments.

[0108] Optionally, the default value of the initial threshold after the update includes:

[0109] Extract the environmental characteristic parameters and action delay data corresponding to the fault type label to generate a four-dimensional spatio-temporal feature vector;

[0110] Specifically, extract the environmental parameters and action delay corresponding to the fault label from the historical case library, construct a vector, the environmental parameters include temperature, humidity, and wind speed, and the action delay is the time from detection to excision; after normalization, sort according to the time series to generate a spatio-temporal feature matrix. The action delay is an index of the system response efficiency, reflecting the action speed of the protection device. The four-dimensional spatio-temporal feature vector is a multi-dimensional data expression that integrates time, space, environment, and performance.

[0111] Use the time series analysis algorithm to perform trend prediction on the four-dimensional spatio-temporal feature vector to generate a recommended value of the dynamic threshold within a future preset time period;

[0112] Specifically, based on the ARIMA model, perform individual prediction on each dimension, set the autoregressive order to 3, the difference order to 1, and the moving average order to 2. After predicting the parameter values for the next 24 hours, generate a recommended value of the dynamic threshold by weighted fusion of temperature, humidity, wind speed, and action delay. For the recommended value of the dynamic threshold There is:

[0113] ,

[0114] In the formula, is the weight coefficient, which is determined by ridge regression training, is the temperature factor, is the humidity factor, is the wind speed factor, is the action time delay factor. The ARIMA model is an autoregressive integrated moving average model, which is suitable for non-stationary time series prediction. Ridge regression is a linear regression method with L2 regularization to prevent overfitting of the weight coefficient.

[0115] After superimposing the dynamic threshold recommended value and the real-time meteorological warning data, the final initial threshold default value is determined by the gradient descent method.

[0116] Specifically, receive the rainfall probability and lightning warning level for the next 6 hours issued by the meteorological station, construct a correction term. For the correction term , there is:

[0117] ,

[0118] In the formula, is the weight coefficient, is the rainfall probability, is the lightning warning level. The gradient descent method is used to iteratively solve the optimal threshold. For the final initial threshold default value , there is:

[0119] ,

[0120] Among them, is the theoretical safety threshold calculated according to the line carrying capacity, is the gradient descent method iterative function of the initial threshold default value.

[0121] Exemplarily, for a certain mountainous area line, it is predicted that the humidity will rise to 85% and the lightning level will be 3 in the next 24 hours. The ARIMA predicts that the humidity factor is 86% and the wind speed factor is 8 m / s. Combining with the lightning warning calculation

[0122] , the dynamic threshold recommended value , after gradient descent optimization, it is calculated that , which is 3% higher than the original threshold of 1000 A. During the actual thunderstorm, the line inrush current reaches 1015 A. Due to the threshold adjustment, the normal load is prevented from being cut off by mistake. Through the dynamic optimization of the threshold driven by meteorological warnings, the real faults and the instantaneous overcurrents caused by lightning can be effectively distinguished, and the power supply continuity can be maintained.

[0123] Optionally, the acquisition method of the wire spatial displacement amount includes:

[0124] The spatial coordinates of the power distribution line corridor are scanned in real time by lidar to generate the three-dimensional displacement trajectory of the wire suspension points;

[0125] Specifically, rotary lidar devices are installed on both sides of the corridor where the wire is located. A laser pulse is emitted every 0.05 seconds. After receiving the reflected signal, the distance between the wire suspension point and the radar station is calculated by the time-of-flight method. Combining the observation data of multiple-station radars, the real-time spatial coordinates of the suspension points are reconstructed based on the three-dimensional point cloud triangulation algorithm. The coordinate sequences of 20 consecutive cycles (within 1 second) form the displacement trajectory curve. The lidar time-of-flight method is a precise measurement technique for calculating distance based on the time delay between the emission and reception of light pulses. Three-dimensional point cloud triangulation is a geometric method for solving three-dimensional spatial coordinates using multi-station observation data. The displacement trajectory curve is a description of the continuous movement path of the wire in space. As Figure 4 shown, the marked points show the key displacement nodes of the wire, reflecting the spatio-temporal synchronization of mechanical vibration and electrical signals.

[0126] Perform time synchronization matching between the displacement mutation amount of the three-dimensional displacement trajectory and the fluctuation period of the spatio-temporal fusion feature matrix;

[0127] Specifically, extract the maximum absolute value of the coordinate change amount between adjacent cycles on the displacement trajectory. If the change amount of the x-axis coordinate is the largest, then the displacement mutation amount is equal to the absolute value of the x-axis coordinate change amount; when the maximum absolute value of the coordinate change amount is greater than a preset value such as 0.5 m, it is marked as the displacement mutation amount period. Synchronously read the mutation timestamps in the spatio-temporal fusion feature matrix where the entropy value fluctuation exceeds the safety benchmark. If there is an overlapping area where the difference between the displacement mutation amount period and the mutation timestamp is less than 0.1 s, it is determined as a spatio-temporal correlation event. The spatio-temporal fusion feature matrix is a data structure containing electrical parameters, environmental, and displacement characteristics. Time synchronization matching is a comparison process between the mechanical displacement and the occurrence time of electrical anomalies, with an error window of 0.1 second.

[0128] When the correlation coefficient between the displacement mutation amount and the fluctuation of electrical parameters is greater than a preset correlation threshold, a high-risk warning signal for line mechanical deformation is triggered.

[0129] Specifically, calculate the Pearson correlation coefficient between the displacement mutation amount sequence and the electrical parameter entropy change parameter sequence :

[0130]

[0131] where, is the th displacement mutation amount, is the th electrical parameter entropy change parameter, is the mean value of the displacement mutation amount, is the mean value of the entropy change parameter of the electrical parameter. If the Pearson correlation coefficient

[0132] number and the duration exceeds 5 seconds, a level III high-risk warning signal is triggered and the corresponding line section is located.

[0133] Exemplarily, in a mountainous area line, conductor galloping occurs under strong wind weather. The laser radar measures that the sudden change in displacement is 6m and lasts for 8 seconds. The spatio-temporal fusion matrix shows a sharp increase in entropy value during the same period, and the calculated Pearson correlation coefficient . A high-risk warning is triggered and the section between tower No. 3 and tower No. 4 is locked. The hidden danger of loosening of the suspension clamp caused by conductor galloping is discovered 12 hours in advance, avoiding the accident of tower collapse due to fracture. Traditional line patrols require visual inspections and it is difficult to detect such gradual defects.

[0134] Optionally, the high-risk warning signal for triggering the mechanical deformation of the line includes:

[0135] Input the spatio-temporal coordinates corresponding to the high-risk warning signal into a preset insulator aging prediction model to calculate the remaining mechanical strength attenuation rate;

[0136] Specifically, the insulator aging prediction model adopts an LSTM neural network structure. The input parameters include the historical environmental temperature and humidity data corresponding to the spatio-temporal coordinates, the average value of the sudden change in displacement within 30 days before the warning occurs, and the cumulative number of power frequency flashovers of the insulators at the same location. The output is the mechanical strength attenuation rate for the next 30 days. For the mechanical strength attenuation rate , there is:

[0137] ,

[0138] where, is the mechanical strength attenuation rate adjustment coefficient, is the average value adjustment coefficient of the sudden change in displacement, is the cumulative power frequency flashover times adjustment coefficient, is the average value of the sudden change in displacement within 30 days before the warning occurs, is the cumulative number of power frequency flashovers of the insulators at the same location, is the historical environmental temperature and humidity data corresponding to the spatio-temporal coordinates, is the natural constant.

[0139] Adjust the inspection cycle of this section according to the attenuation rate, and synchronously increase the threshold value.

[0140] Specifically, the inspection cycle adjustment formula is:

[0141] ,

[0142] In the formula, is the adjusted inspection cycle, is the reference attenuation coefficient, which can be taken as 0.3. At the same time, the high-frequency harmonic threshold of partial discharge detection is increased from 5% to 3% to improve the sensitivity.

[0143] Exemplarily, after a certain coastal salt spray section triggers an early warning, the input to the model gives (standard value 0.1). The inspection cycle is adjusted from 28 days to days, and the threshold is reduced to 3%. During the subsequent inspection on the 7th day, it is found that the salt density on the surface of the insulator exceeds the standard and is cleaned in time. To avoid the brittle fracture accident of the insulator caused by salt spray corrosion, at the same time, the early signs of corona discharge are captured by the adaptive adjustment of the high-frequency threshold. Sub-meter-level deformation monitoring is realized through lidar and spatially and temporally aligned with electrical characteristics, enabling the system to determine whether mechanical disturbances such as conductor galloping and icing pose potential safety hazards. Combining with the LSTM model to predict the life attenuation path of the insulator, realizing hierarchical dynamic adjustment of detection parameters, and accurately allocating operation and maintenance resources to high-risk points. After implementation, the ability to prevent hidden mechanical defects is significantly improved, and the service life of key equipment is extended.

[0144] Based on the same inventive concept, as Figure 5 shown, the present invention also provides an intelligent fault diagnosis system for distribution lines based on multi-source information fusion, and the system includes:

[0145] A data acquisition module, configured to acquire current signals, voltage signals, and temperature signals of the distribution line, and synchronously collect the humidity gradient change rate in the meteorological data and the conductor spatial displacement amount in the geographic coordinate data to generate a multi-dimensional heterogeneous data set;

[0146] A feature construction module, configured to calculate the entropy value of the time-domain distortion features of the current signal and the voltage signal, and superimpose the humidity gradient change rate to generate a dynamic entropy change parameter. At the same time, based on the conductor spatial displacement amount, spatial form calibration is performed on the dynamic entropy change parameter to generate a spatio-temporal fusion feature matrix;

[0147] An initial threshold adjustment module, configured to dynamically adjust a preset initial threshold by using the spatio-temporal fusion feature matrix; wherein, the adjustment process includes calculating the conductor resistance change rate according to the real-time temperature data in the meteorological data, and performing weighted compensation on the resistance change rate through a sliding window algorithm to generate an adaptive fault discrimination threshold;

[0148] A fault judgment module, configured to perform multiple rounds of verification on the abnormal fluctuation of the electrical parameters of the line according to the fault discrimination threshold. When the fluctuation amplitude in three consecutive sampling periods exceeds the fault discrimination threshold and the high-frequency harmonic ratio reaches a preset threshold, a fault location instruction with a confidence weight is generated;

[0149] The fault handling module is used to perform real-time matching between the fault location instruction and the line topology structure, and correct the fault section priority queue based on the switch action feedback data of the feeder terminal, triggering a fault removal action that matches the priority.

[0150] It should be noted that the electrical connections between the above-mentioned various units do not necessarily represent direct connections of the lines. Indirect connection methods, as long as they achieve the purpose of the present invention, can be applied to the embodiments of the present invention. The above are only exemplary embodiments of the present invention, and the scope of the present invention cannot be limited thereby.

[0151] That is, any equivalent changes and modifications made in accordance with the teachings of the present invention still fall within the scope covered by the present invention. After considering the specification and the disclosure of the practical truth, those skilled in the art will easily think of other implementation schemes of the present invention. This application aims to cover any variations, uses, or adaptive changes of the present invention, which follow the general principles of the present invention and include the common general knowledge or conventional technical means in the technical field not recorded in the present invention.

Claims

1. An intelligent fault diagnosis method for distribution lines based on multi-source information fusion, characterized in that, The method includes: Obtaining the current signal, voltage signal and temperature signal of the distribution line, and synchronously collecting the humidity gradient change rate in the meteorological data and the wire spatial displacement amount in the geographic coordinate data to generate a multi-dimensional heterogeneous data set; Calculating the entropy value of the time-domain distortion characteristics of the current signal and the voltage signal, and superimposing the humidity gradient change rate to generate a dynamic entropy change parameter. At the same time, based on the wire spatial displacement amount, spatial morphology calibration is performed on the dynamic entropy change parameter to generate a spatio-temporal fusion feature matrix; Dynamically adjusting a preset initial threshold by using the spatio-temporal fusion feature matrix; wherein, the adjustment process includes calculating the wire resistance change rate according to the real-time temperature data in the meteorological data, and performing weighted compensation on the resistance change rate through a sliding window algorithm to generate an adaptive fault discrimination threshold; Performing multi-round verification on the abnormal fluctuation of the electrical parameters of the line according to the fault discrimination threshold. When the fluctuation amplitude in three consecutive sampling periods exceeds the fault discrimination threshold and the high-frequency harmonic ratio reaches a preset threshold, generating a fault location instruction with a confidence weight; Real-time matching the fault location instruction with the line topology structure, and correcting the fault section priority queue based on the switch action feedback data of the feeder terminal to trigger a fault removal action matching the priority; Among them, the generating of the spatio-temporal fusion feature matrix includes: Extracting the time-domain waveform distortion feature of the current signal and the third-harmonic mutation rate of the voltage signal, and calculating the composite distortion entropy value of the two; Superimposing the logarithmic decay coefficient of the humidity gradient change rate on the composite distortion entropy value to generate a dynamic entropy change parameter; Performing morphological erosion processing on the dynamic entropy change parameter based on the fluctuation trajectory of the wire spatial displacement amount, and outputting a spatio-temporal fusion feature matrix.

2. The intelligent fault diagnosis method for distribution lines based on multi-source information fusion according to claim 1, characterized in that The generating of the adaptive fault discrimination threshold includes: Obtaining the historical mean data of the wire resistance change rate, and fitting the resistance change curve based on the real-time temperature data; Combining the spatio-temporal fusion feature matrix, using a dynamic Bayesian network to identify abnormal segments of the resistance change curve, and generating a compensation weight factor; Performing a convolution operation on the compensation weight factor and the initial threshold to generate a fault discrimination threshold including the dynamic coupling relationship of environmental parameters.

3. The intelligent fault diagnosis method for distribution lines based on multi-source information fusion according to claim 2, characterized in that, The multi-round verification of the abnormal fluctuation of the electrical parameters of the line according to the fault discrimination threshold includes: When the first fluctuation amplitude exceeds the fault discrimination threshold, collecting the hydrogen volume fraction data of the distribution transformer oil chromatogram; If the hydrogen volume fraction exceeds the preset fraction threshold and the high-frequency harmonic ratio reaches the preset threshold, triggering a partial discharge pre-judgment signal; Adjusting the fault discrimination threshold for the subsequent two sampling periods according to the pre-judgment signal.

4. The intelligent fault diagnosis method for distribution lines based on multi-source information fusion according to claim 3, characterized in that, The generating of the fault location instruction with a confidence weight includes: Obtaining the switch state data of the feeder terminal, and inversely calculating the fault current propagation path based on the switch state data; Performing an overlap analysis on the propagation path and the abnormal area of the spatio-temporal fusion feature matrix, and increasing the confidence weight when the overlapping area ratio is greater than the preset ratio threshold; If there are multiple candidate sections, calculate the correlation coefficient between the change rate of the dynamic entropy change parameter and the wire spatial displacement, and select the section according to the magnitude of the correlation coefficient.

5. The intelligent fault diagnosis method for distribution lines based on multi-source information fusion according to claim 4, wherein The method further includes: Collect the difference in line capacitance current before and after fault removal, and perform similarity matching with the standard difference in the preset fault feature database; Generate a fault type label according to the matching result, and append the fault type label associated with the environmental parameters to the preset historical fault case database; Update the default value of the subsequent initial threshold based on the fault frequency distribution under the same geographical coordinates in the historical fault case database.

6. The intelligent fault diagnosis method for distribution lines based on multi-source information fusion according to claim 5, characterized in that The update of the default value of the subsequent initial threshold includes: Extract the environmental characteristic parameters and action delay data corresponding to the fault type label to generate a four-dimensional spatio-temporal feature vector; Use a time series analysis algorithm to perform trend prediction on the four-dimensional spatio-temporal feature vector to generate a dynamic threshold recommended value within a preset future time period; After superimposing the dynamic threshold recommended value and the real-time meteorological warning data, determine the final default value of the initial threshold by the gradient descent method.

7. The intelligent fault diagnosis method for distribution lines based on multi-source information fusion according to claim 3, characterized in that The acquisition method of the wire spatial displacement includes: Real-time scan the spatial coordinates of the distribution line corridor by lidar to generate the three-dimensional displacement trajectory of the wire suspension point; Perform time synchronization matching between the displacement mutation amount of the three-dimensional displacement trajectory and the fluctuation period of the spatio-temporal fusion feature matrix; When the correlation coefficient between the displacement mutation amount and the electrical parameter fluctuation is greater than the preset correlation threshold, trigger a high-risk warning signal for line mechanical deformation.

8. The intelligent fault diagnosis method for distribution lines based on multi-source information fusion according to claim 7, characterized in that The triggering of the high-risk warning signal for line mechanical deformation includes: Input the spatio-temporal coordinates corresponding to the high-risk warning signal into the preset insulator aging prediction model to calculate the remaining mechanical strength attenuation rate; Adjust the inspection cycle of this section according to the attenuation rate and synchronously increase the threshold.

Images