Distribution line fault detection method, apparatus and device, and storage medium

Abstract

The invention discloses a distribution line fault detection method and device, equipment and a storage medium, and relates to the technical field of electric power inspection, hidden danger features are extracted through LSTM time sequence analysis, time sequence modeling is performed on current and voltage waveforms, harmonic distortion rate, waveform distortion and other features are extracted, and early abnormality of the equipment can be accurately identified through quantitative analysis of the features, so that the fault detection accuracy is improved. Early warning before a fault is achieved, and the passive situation that an existing system can only conduct positioning after the fault occurs is changed. The fault type is subjected to weighted fusion judgment by using the random forest + SVM algorithm, and the advantages of the two algorithms are integrated, so that the possible misjudgment condition of a single algorithm is greatly reduced, and the accuracy of fault judgment is improved. Secondly, waveform correction is introduced into a traditional double-end positioning formula, local spectrum analysis of edge nodes is combined, traveling wave propagation speed errors are corrected, dependence on clock synchronization of equipment at the two ends is reduced, and the fault positioning precision is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power inspection, and in particular to a power distribution line fault detection method, device, equipment and storage medium. BACKGROUND

[0002] With the advancement of smart grid construction, the accuracy and timeliness of power distribution line fault detection are increasingly required. Currently, the industry mainly relies on traveling wave distance measurement method and single intelligent algorithm for fault positioning and diagnosis. The traveling wave method relies on double-end detection device to capture fault traveling wave, and calculates the fault distance through time difference. Intelligent algorithms such as SVM or random forest are used for waveform feature extraction and fault classification.

[0003] However, the existing scheme still has significant defects: first, the positioning accuracy of traveling wave is limited by clock synchronization error and uncertainty of traveling wave propagation speed, especially in long lines and complex electromagnetic environment, the error is amplified, which is difficult to meet the high-precision requirement; second, single algorithm has limited recognition ability for non-steady-state signals (such as intermittent faults and harmonic disturbances), with high false alarm rate, and lacks timing warning capability for early hidden dangers such as insulation aging; third, the existing system is mostly limited to "after-treatment", and cannot realize state prediction and maintenance guidance before fault.

[0004] Therefore, there is an urgent need for an intelligent fault detection method that can integrate multi-source data, combine time series analysis and hybrid algorithm, and support hidden danger warning and high-precision positioning, to comprehensively improve the operation reliability and fault response efficiency of the distribution network. SUMMARY

[0005] The present application provides a power distribution line fault detection method, device, equipment and storage medium to solve the above problems.

[0006] Firstly, the present application improves the algorithm by extracting hidden danger features through LSTM time series analysis, modeling the current and voltage waveforms in time series, extracting harmonic distortion rate and waveform distortion, etc. Through quantitative analysis of these features, early abnormalities of equipment can be accurately identified, early warning before fault can be realized, and the passive situation of the existing system that can only locate after fault occurs is changed. The random forest + SVM algorithm is used to judge the fault type by weighted fusion, which combines the advantages of the two algorithms, greatly reduces the misjudgment of single algorithm, and improves the accuracy of fault judgment. Secondly, the waveform correction is introduced in the traditional double-end positioning formula, combined with the local spectrum analysis of the edge node, the error of traveling wave propagation speed is corrected, the dependence on the clock synchronization of the two ends is reduced, and the accuracy of fault positioning is improved.

[0007] On the one hand, the present application provides a power distribution line fault detection method, which comprises the following steps: Step S1: Real-time acquisition of current and voltage waveform data through Hall sensors and capacitive dividers deployed on the power distribution line, and transmission of the data to the edge computing node; Step S2: Time series modeling of the current and voltage waveform data using an LSTM neural network to extract harmonic distortion rate and waveform distortion features for identifying device operating states; Step S3: Analysis of the features based on a fusion diagnosis algorithm of random forest and support vector machine to determine whether the line state is fault-free, has hidden dangers, or has occurred a fault; Step S4: If it is determined that there are hidden dangers, a warning signal is generated; if it is determined that a fault has occurred, the fault point position is calculated through double-end traveling wave time difference correction and propagation speed dynamic calibration, and the positioning result is output.

[0008] In an implementation manner of the present application, in the step S2, the LSTM neural network extracts the harmonic distortion rate THD and the waveform distortion coefficient K by time series modeling of the current and voltage waveforms, and the calculation method is as follows:

[0009] wherein, is the effective value of the fundamental voltage, is the effective value of the hth harmonic voltage, and H is the highest harmonic number considered; The waveform distortion coefficient K is calculated by the following formula:

[0010] wherein, is the actual sampling value, is the ideal sinusoidal wave reference value, is the standard deviation, and N is the number of sampling points.

[0011] In an implementation manner of the present application, in the step S3, the random forest algorithm is used for importance evaluation and preliminary classification of multi-dimensional features, and the probability distribution of each category is output; the SVM algorithm receives the probability distribution as an input feature vector, and performs final classification decision through kernel function mapping to a high-dimensional space, and the decision function is:

[0012] wherein, is the Lagrange multiplier, is the sample label, is the kernel function, b is the bias term, and the weighted fusion mechanism improves the ability to distinguish intermittent faults and non-fault disturbances.

[0013] In an implementation form of the present application, in the step S4, the double-ended traveling wave time difference correction compensates the clock synchronization error by using the following formula:

[0014] wherein, is the delay compensation amount obtained by fitting the historical communication delay data; The traveling wave propagation speed v is dynamically calibrated according to the ambient temperature T and the line load L:

[0015] wherein, is the reference speed, and are the reference temperature and load, and are the material and environmental coefficients.

[0016] In an implementation form of the present application, the hidden danger early warning further comprises trend analysis based on the LSTM output features, if the harmonic distortion rate THD continuously rises in a plurality of continuous time windows and exceeds the set threshold, it is determined that there is an insulation aging hidden danger, a warning signal is generated and pushed to the operation and maintenance terminal.

[0017] In an implementation form of the present application, the fault point position is calculated based on the following formula:

[0018] wherein, L is the total length of the line, v is the calibrated traveling wave propagation speed, is the corrected time difference.

[0019] In an implementation form of the present application, the fusion diagnosis algorithm of the random forest and the support vector machine adopts a weighted voting mechanism to obtain the final state judgment result; the final classification result is determined by the following formula:

[0020] wherein, represents the category of the fault, and the category includes: no fault, hidden danger, fault, is the prediction probability of the random forest model for the category , is the hard decision output of the SVM model, is the Kronecker delta function, which is 1 when , otherwise 0, and are the weights given to the random forest and the SVM model respectively, and .

[0021] The application also provides a power distribution line fault detection device, the device comprising: A data acquisition unit is configured to acquire current and voltage waveform data in real time through a Hall sensor and a capacitor voltage divider deployed on the power distribution line, and transmit the data to an edge computing node; A distortion extraction unit is configured to use an LSTM neural network to perform time series modeling on the current and voltage waveform data, extract harmonic distortion rate and waveform distortion features, and identify equipment operating states; A fault analysis unit is configured to analyze the features based on a fusion diagnosis algorithm of random forest and support vector machine, and determine whether the line state is fault-free, has hidden dangers, or has faults; A result calculation unit is configured to generate a warning signal when it is determined that there are hidden dangers, and calculate a fault point position through double-end traveling wave time difference correction and propagation speed dynamic calibration, and output a positioning result if it is determined that there is a fault.

[0022] The application also provides a power distribution line fault detection device, the device comprising: at least one processor; and, a memory in communication with the at least one processor; wherein, the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to complete the foregoing power distribution line fault detection method.

[0023] The application also provides a non-volatile computer storage medium for power distribution line fault detection, which stores computer executable instructions, and the computer executable instructions are executed by a processor to implement the foregoing power distribution line fault detection method.

[0024] The power distribution line fault detection method, device, equipment and storage medium provided by the application have the following beneficial effects: 1. The application realizes high-precision time series feature extraction and state classification of power distribution line current and voltage waveforms by fusing LSTM neural network and random forest-support vector machine hybrid algorithm. LSTM effectively captures harmonic distortion and waveform abnormal trend, and the hybrid algorithm improves the ability to distinguish intermittent faults and interference signals, significantly reduces the false alarm rate, and enhances the accuracy and robustness of fault identification.

[0025] 2. The introduction of double-end traveling wave time difference dynamic correction and propagation speed environment adaptive calibration mechanism effectively overcomes the problems of clock synchronization error and speed uncertainty in traditional traveling wave positioning. Through historical delay fitting and temperature load dynamic compensation, the fault point positioning accuracy is greatly improved, which is especially suitable for high reliability fault ranging requirements in long lines and complex environments.

[0026] 3. The system possesses proactive risk warning capabilities. Based on LSTM output characteristics, it performs continuous trend analysis to identify potential risks such as insulation aging at an early stage. An early warning is issued once the harmonic distortion rate continues to exceed the standard, supporting early intervention by maintenance personnel. This achieves a shift from "post-event handling" to "pre-event prevention," improving the operational safety and maintenance efficiency of the distribution network. Attached Figure Description

[0027] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings: Figure 1 A flowchart of a power distribution line fault detection method provided in this application embodiment; Figure 2 A diagram illustrating the composition of a power distribution line fault detection device provided in this application embodiment; Figure 3 This is a schematic diagram of a power distribution line fault detection device provided in an embodiment of this application. Detailed Implementation

[0028] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0029] This application provides a method, apparatus, device, and storage medium for detecting power distribution line faults. The technical solutions proposed in this application will be described in detail below with reference to the accompanying drawings.

[0030] Figure 1 This is a flowchart illustrating a power distribution line fault detection method provided in an embodiment of this application. Figure 1 As shown, the method mainly includes the following steps: Step S1: Real-time acquisition of current and voltage waveform data is achieved by using Hall sensors and capacitive voltage dividers deployed on the power distribution line, and the data is transmitted to the edge computing node; Step S2: Use an LSTM neural network to perform time-series modeling on the current and voltage waveform data, extract harmonic distortion rate and waveform distortion features to identify the operating status of the equipment; Step S3: Based on the fusion diagnostic algorithm of random forest and support vector machine, analyze the features to determine whether the line status is fault-free, has hidden dangers, or has failed. Step S4: If it is judged that there is a hidden danger, a warning signal is generated; if it is judged that a fault occurs, the fault point position is calculated through double-end traveling wave time difference correction and dynamic calibration of propagation speed, and the positioning result is output.

[0031] In the present application, in the step S2, the LSTM neural network models the time sequence of the current and voltage waveforms, and the input sequence length (time window size) of the LSTM neural network is dynamically determined based on the power distribution line signal characteristics and the sampling frequency: the initial window obtained by offline cross-validation (recommended 50 ms, corresponding to 2500 data points under 50 kHz sampling) is preferentially adopted, and is self-adaptively adjusted according to the real-time waveform complexity - expanded to 80 ms to smooth the noise under normal state, and reduced to 20 ms to capture the transient characteristics under abnormal state. The data in the window is input into the LSTM after normalization, and the time sequence characteristics such as harmonic distortion rate (THD) and waveform distortion coefficient (K) are extracted through the gating mechanism.

[0032] The harmonic distortion rate THD and the waveform distortion coefficient K are extracted, and the calculation method is as follows:

[0033] wherein, is the effective value of the fundamental voltage, is the effective value of the hth harmonic voltage, and H is the highest harmonic number considered; The waveform distortion coefficient K is calculated by the following formula:

[0034] wherein, is the actual sampling value, is the ideal sinusoidal wave reference value, is the standard deviation, and N is the number of sampling points.

[0035] In the present application, in the step S3, the random forest algorithm is used to evaluate the importance of multi-dimensional features and preliminarily classify, and the probability distribution of each class is output; the SVM algorithm receives the probability distribution as the input feature vector, and makes the final classification decision through the kernel function mapping to the high-dimensional space, and the decision function is:

[0036] wherein, is the Lagrange multiplier, is the sample label, is the kernel function, and b is the bias term, which improves the ability to distinguish between intermittent faults and non-fault disturbances through a weighted fusion mechanism.

[0037] In this application, in step S4, the double-ended traveling wave time difference correction is compensated for clock synchronization error using the following formula:

[0038] in, This is the delay compensation amount obtained by fitting historical communication delay data; The traveling wave propagation speed v is dynamically calibrated based on the ambient temperature T and the line load L.

[0039] in, As the reference speed, and For reference temperature and load, and For materials and environmental factors.

[0040] In this application, the hazard warning also includes trend analysis based on LSTM output characteristics. If the harmonic distortion rate (THD) continues to rise and exceeds the set threshold within multiple consecutive time windows, it is determined to be a hazard of insulation aging, and a warning signal is generated and pushed to the operation and maintenance terminal.

[0041] In this application, the location of the fault point is calculated based on the following formula:

[0042] Where L is the total length of the line, and v is the calibrated traveling wave propagation speed. This is the corrected time difference.

[0043] In this application, the fusion diagnostic algorithm of random forest and support vector machine adopts a weighted voting mechanism to obtain the final state judgment result; the final classification result Determined by the following formula:

[0044] in, This indicates the category of the fault, including: no fault, potential problem, and fault. It is a random forest model for categories The predicted probability, It is the hard decision output of the SVM model. It is the Kroneckerdelta function, when The value is 1 if it is true, and 0 otherwise. and These are the weights assigned to the random forest and SVM models, respectively. .

[0045] Weights in Random Forest and SVM Models , ) Combination setting by "offline calibration + online fine-tuning". In the offline calibration stage: based on historical samples (including three states of no fault, hidden danger and fault), cross-validation method is used to traverse the weight combination, and the initial weight is determined by taking the weighted F1 value (taking into account accuracy and recall rate) as the optimization target (recommended = 0.4, = 0.6, which can be adjusted according to the line type); in the online fine-tuning stage: dynamic adaptation according to real-time running state - hidden danger warning stage is improved to 0.6, fault diagnosis stage is improved to 0.7, and in strong interference environment, it is improved to 0.8; the weight adjustment needs to meet , to ensure that the fusion algorithm adapts to the dynamic scene requirements of the distribution line.

[0046] The above is a distribution line fault detection method provided by the embodiment of the application, based on the same inventive concept, the embodiment of the application also provides a distribution line fault detection device, Figure 2 The composition diagram of the distribution line fault detection device provided by the embodiment of the application is shown in Figure 2 , which mainly includes: A data acquisition unit 201 is configured to acquire current and voltage waveform data in real time through a Hall sensor and a capacitor voltage divider deployed on a distribution line, and transmit the data to an edge computing node; A distortion extraction unit 202 is configured to use an LSTM neural network to perform time series modeling on the current and voltage waveform data, extract harmonic distortion rate and waveform distortion features, and identify device operating state; A fault analysis unit 203 is configured to analyze the features based on a fusion diagnosis algorithm of random forest and support vector machine, and determine whether the line state is no fault, hidden danger or fault; A result calculation unit 204 is configured to generate a warning signal when it is determined that there is a hidden danger, and calculate the fault point position by double-end traveling wave time difference correction and propagation speed dynamic calibration, and output the positioning result if it is determined that there is a fault.

[0047] The above is a distribution line fault detection device provided by the embodiment of the application, based on the same inventive concept, the embodiment of the application also provides a distribution line fault detection device, Figure 3 The schematic diagram of the distribution line fault detection device provided by the embodiment of the application is shown in Figure 3As shown, the device mainly comprises: at least one processor 301; and a memory 302 connected in communication with the at least one processor; wherein the memory 302 stores instructions executable by the at least one processor 301, and the instructions are executed by the at least one processor 301 to enable the at least one processor 301 to complete the power distribution line fault detection method described above.

[0048] In addition, the embodiment of the present application also provides a non-volatile computer storage medium for power distribution line fault detection, which stores computer executable instructions, and the computer executable instructions are executed by a processor to implement the power distribution line fault detection method described above.

[0049] The following is an application example of the present application in a specific scenario.

[0050] An intelligent fault detection system of the present application is applied to a 10kV power distribution line. The line is 8km long, with multiple distribution transformers and branch lines distributed along the way, and the surrounding environment is complex, with various influencing factors such as industrial interference and weather changes. During system operation, the current and voltage data of the line are continuously collected by Hall sensors and capacitor dividers and transmitted to edge computing nodes. The LSTM time series modeling algorithm analyzes the current and voltage waveform data for a continuous week and finds that the harmonic distortion rate of a certain section of the line is gradually increasing from the initial 2% to 5%, exceeding the normal range (normal range is 0-3%), and the system determines that there is a risk of insulation aging in this section of the line, and immediately issues a hidden danger warning signal. After receiving the warning, the staff timely inspected the line and found that the insulation layer at a certain point of the line was slightly damaged and was replaced in advance, avoiding the occurrence of a fault.

[0051] When a short circuit fault occurs in the line, the system immediately starts the fault handling process. The time difference Δt of the traveling wave arriving at both ends of the line is calculated to be 0.002s, and the 0.0003s error caused by communication delay is eliminated by double-end time difference Δt correction, and the traveling wave propagation speed v is dynamically calibrated in combination with the environmental temperature, line load and other factors at that time, from the original 2.9x10 8 m / s to 2.85x10 8 m / s. Finally, according to the corrected and calibrated parameters, the positioning formula is used to calculate the distance of the fault point from one end of the line as 3.2km, and the staff quickly find the fault point and perform repair according to the positioning, and the fault positioning and repair time is shortened by about 50% compared with the traditional system. Among them, the fault point positioning time is shortened from 40 minutes to 15 minutes, and the number of ineffective repair operations is reduced by 82% compared with last month, fully verifying the quantitative improvement effect of fault handling efficiency.

[0052] It can be seen from the specific embodiment that the application can effectively realize hidden danger early warning and high-precision fault positioning in actual application, and significantly improves the operation reliability and fault processing efficiency of the distribution line.

[0053] The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable data processing apparatus to produce a computer-implemented process such that the instructions executed by the computer or other programmable data processing apparatus provide steps for implementing the functions specified in the flowchart block or blocks. Figure 1 one or more flow or blocks Figure 1 means for functionally implementing the steps listed in the flowchart block or blocks.

[0054] These computer program instructions can also be stored in a computer readable memory capable of directing the computer or other programmable data processing apparatus to work in a specific manner, so that the instructions stored in the computer readable memory produce a manufactured product including instruction means, which implements the functions specified in the flowchart block or blocks. Figure 1 one or more flow or blocks Figure 1 the functions specified in the flowchart block or blocks.

[0055] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus, so that a series of operational steps are performed on the computer or other programmable data processing apparatus to produce a computer-implemented process, so that the instructions executed on the computer or other programmable data processing apparatus provide steps for implementing the functions specified in the flowchart block or blocks. Figure 1 one or more flow or blocks Figure 1 the functions specified in the flowchart block or blocks.

[0056] In a typical configuration, the computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memories.

[0057] Each of the embodiments in the present application is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other. Each embodiment focuses on the difference from other embodiments. In particular, for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the part of the method embodiment.

[0058] It is also to be noted that the terms "comprising", "including", and any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can include other elements not expressly listed or inherent to such process, method, article, or apparatus. An element proceeded by "comprises a... " does not, without more constraints, exclude the existence of additional identical elements in the process, method, article, or apparatus that comprises the element.

[0059] The above description is merely illustrative of the application, and not restrictive. Various modifications and changes can become apparent to those skilled in the art. Incorporating any modification, equivalent substitution, improvement, etc. within the spirit and principle of the application, shall be included in the scope of the claims of the application.

Claims

1. A power distribution line fault detection method, characterized by, The method comprises the following steps: Step S1: Real-time acquisition of current and voltage waveform data by Hall sensors and capacitive voltage dividers deployed on the power distribution line, and transmission of the data to an edge computing node; Step S2: Time series modeling of the current and voltage waveform data using an LSTM neural network, extraction of harmonic distortion rate and waveform distortion features, and identification of device operating state; Step S3: Analysis of the features based on a fusion diagnosis algorithm of random forest and support vector machine to determine the line state as fault-free, with hidden danger or fault; Step S4: If it is determined that there is hidden danger, a warning signal is generated; if it is determined that there is a fault, the fault point position is calculated by double-end traveling wave time difference correction and dynamic calibration of propagation speed, and the positioning result is output.

2. The method of claim 1, wherein, In step S2, the LSTM neural network extracts the harmonic distortion rate THD and the waveform distortion coefficient K by time series modeling of the current and voltage waveforms, and the calculation method is as follows: wherein Veff is the effective value of the fundamental voltage, Veh is the effective value of the hth harmonic voltage, H being the highest harmonic number considered; The waveform distortion coefficient K is calculated by the following formula: wherein, is the actual sampled value, is the ideal sinusoidal reference value, is the standard deviation, N is the number of sampling points.

3. The method of claim 1, wherein, In step S3, the random forest algorithm is used for importance evaluation and preliminary classification of multi-dimensional features, and the probability distribution of each category is output; the SVM algorithm receives the probability distribution output by the random forest as an input feature vector, and makes a final classification decision by kernel function mapping to a high-dimensional space, and the decision function is: wherein, is a Lagrange multiplier, is a sample label, is a kernel function, and b is a bias term, which improves the ability to distinguish intermittent faults and non-fault disturbances through a weighted fusion mechanism.

4. The method of claim 1, wherein, In step S4, the double-end traveling wave time difference correction compensates for the clock synchronization error using the following formula: wherein, is a delay compensation quantity obtained by fitting historical communication delay data; The traveling wave propagation speed v is dynamically calibrated according to the environmental temperature T and the line load L: wherein, is the reference speed, and is the reference temperature and load, and are material and environmental coefficients.

5. The method of claim 1, wherein, Hidden danger warning also includes trend analysis based on the LSTM output features. If the harmonic distortion rate THD continuously rises in multiple consecutive time windows and exceeds the set threshold, it is determined that there is insulation aging hidden danger, a warning signal is generated and pushed to the operation and maintenance terminal.

6. The method of claim 1, wherein, The fault point position is calculated based on the following formula: Wherein, L is the total length of the line, v is the calibrated wave propagation speed, is the corrected time difference.

7. The method of claim 3, wherein, The fusion diagnosis algorithm of the random forest and the support vector machine adopts a weighted voting mechanism to obtain a final state judgment result; the final classification result is determined by the following formula: wherein, represents the class of the fault, the classes including: no fault, hidden danger, fault, is the predicted probability of the class by the random forest model, is the hard decision output of the SVM model, is the Kronecker delta function, which is 1 when and 0 otherwise, and are the weights assigned to the random forest and SVM model, respectively, and .

8. A power distribution line fault detection apparatus, characterized by, The device comprises: A data acquisition unit for real-time acquisition of current and voltage waveform data by Hall sensors and capacitive voltage dividers deployed on the power distribution line, and transmission of the data to an edge computing node; A distortion extraction unit for time series modeling of the current and voltage waveform data using an LSTM neural network, extraction of harmonic distortion rate and waveform distortion features, and identification of device operating state; A fault analysis unit for analysis of the features based on a fusion diagnosis algorithm of random forest and support vector machine to determine the line state as fault-free, with hidden danger or fault; A result calculation unit for generating a warning signal if it is determined that there is hidden danger; if it is determined that there is a fault, the fault point position is calculated by double-end traveling wave time difference correction and dynamic calibration of propagation speed, and the positioning result is output.

9. A power distribution line fault detection apparatus, characterized by, The device comprises: At least one processor; and A memory in communication connection with the at least one processor; wherein The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to complete the power distribution line fault detection method of any one of claims 1-7. 10.A non-transitory computer storage medium storing computer-executable instructions for power distribution line fault detection, the computer-executable instructions comprising: The computer executable instructions are executed by the processor to implement the power distribution line fault detection method according to any one of claims 1-7.

Images