Power distribution network fault detection and positioning method

Abstract

The invention provides a power distribution network fault detection and positioning method. The method comprises the following steps: acquiring electrical quantity data and environmental data of each section; performing feature extraction on the electrical quantity data and the environmental data to obtain a first feature and a second feature corresponding to the electrical quantity data and the environmental data respectively; constructing a feature vector according to the first feature and the second feature of each section at the same moment, and constructing a multi-dimensional dynamic feature vector according to a plurality of feature vectors at continuous moments; and inputting the multi-dimensional dynamic feature vector into a pre-trained fault detection model to obtain a fault identification result. According to the invention, the fault of the power distribution network can be detected and positioned efficiently and accurately.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of transfer learning, in particular to a power distribution network fault detection and positioning method. BACKGROUND

[0002] With the rapid development of smart grid and distributed energy, the scale of power distribution network is expanding and the structure is becoming increasingly complex, and the real-time and accuracy of fault detection and positioning have become the key to ensuring power supply reliability.

[0003] Traditional power distribution network fault detection mainly relies on manual inspection or threshold judgment based on single electrical quantity (such as current and voltage), which has the problems of slow response speed, high false alarm rate and fuzzy fault positioning. SUMMARY

[0004] The purpose of the present application is to provide a power distribution network fault detection and positioning method, which aims to solve the problems of slow response speed, high false alarm rate and fuzzy fault positioning of traditional technology due to the main reliance on manual inspection or threshold judgment based on single electrical quantity (such as current and voltage).

[0005] In a first aspect, the present application provides a power distribution network fault detection and positioning method, which comprises:

[0006] Obtaining electrical quantity data and environmental data of each section;

[0007] Performing feature extraction on the electrical quantity data and the environmental data respectively to obtain first features and second features corresponding to the electrical quantity data and the environmental data respectively;

[0008] Constructing a feature vector according to the first feature and the second feature of each section at the same time, and constructing a multi-dimensional dynamic feature vector according to a plurality of feature vectors at consecutive time points;

[0009] Inputting the multi-dimensional dynamic feature vector into a pre-trained fault detection model to obtain a fault recognition result.

[0010] Further, the step of obtaining electrical quantity data and environmental data of each section comprises:

[0011] Dividing the target power distribution network into a plurality of sections and numbering each section;

[0012] Collecting bus zero sequence current, bus zero sequence voltage, three-phase current of each section, voltage between two sections, line current, section temperature and section humidity every first preset time.

[0013] Further, the step of performing feature extraction on the electrical quantity data and the environmental data respectively to obtain first features and second features corresponding to the electrical quantity data and the environmental data respectively comprises:

[0014] The bus zero sequence current is extracted according to the following formula:

[0015]

[0016]

[0017] wherein, is the kth harmonic of the bus zero sequence current from the ith section at the tth moment, is the total number of sampling points, is the bus zero sequence current of the ith section at the tth moment, is the first energy feature extracted from the bus zero sequence current of the ith section at the tth moment, and M represents the total number of harmonics;

[0018] The three-phase current is extracted according to the following formula:

[0019]

[0020] wherein, is the unbalance degree corresponding to the three-phase current of the ith section at the tth moment, are each a phase current in the three-phase current;

[0021] The section temperature and the section humidity are extracted according to the following formula:

[0022]

[0023] wherein, is the temperature-humidity coupling feature of the ith section at the tth moment, is the temperature of the ith section at the tth moment, is the humidity of the ith section at the tth moment.

[0024] Further, the step of constructing a feature vector according to the first feature and the second feature of each section at the same moment, and constructing a multi-dimensional dynamic feature vector according to a plurality of feature vectors at continuous moments comprises:

[0025] The feature vector is constructed according to the following formula:

[0026]

[0027] wherein, are respectively corresponding to a phase feature corresponding to the zero sequence voltage, a frequency feature, respectively, a second energy feature, a third energy feature, and a fourth energy feature corresponding to the three-phase current of the i-th section at the t-th moment, is a two-terminal voltage difference of the i-th section at the t-th moment, is a fifth energy feature corresponding to the line current of the i-th section at the t-th moment, is a line current fluctuation rate of the i-th section at the t-th moment;

[0028] The total number of moments is taken as the number of rows, and a multi-dimensional feature vector is constructed according to the feature vector.

[0029] Further, the step of inputting the multi-dimensional dynamic feature vector into the pre-trained fault detection model to obtain a fault recognition result comprises:

[0030] Obtain power distribution network data at consecutive moments under known faults and normal operation, and construct a historical multi-dimensional feature vector according to the power distribution network data;

[0031] Use CNN to extract spatial features of the feature vector sequence to obtain a feature map, and input the feature map into LSTM for time sequence feature learning to obtain a time feature vector, and input the time feature vector into SVM for fault classification training to obtain a trained hybrid model.

[0032] Further, the step of inputting the multi-dimensional dynamic feature vector into the pre-trained fault detection model to obtain a fault recognition result further comprises:

[0033] According to the fault recognition result, obtain the fault section number of the fault, and according to the fault section number, call the operating parameters of each node in the fault section, the operating parameters including voltage amplitude, current harmonic distortion rate, temperature fluctuation value and power factor;

[0034] Calculate the dynamic coupling value between each operating parameter at the same node, and determine whether the dynamic coupling value is within a preset threshold range;

[0035] If the dynamic coupling value is not within the preset threshold range, the node is a fault node;

[0036] If the dynamic coupling value is within the preset threshold range, the node is a normal node.

[0037] Further, the step of calculating the dynamic coupling value between each operating parameter at the same node comprises:

[0038] According to the following formula:

[0039]

[0040]

[0041] wherein, is the coupling degree between the i-th operating parameter and the j-th operating parameter, , , are weight coefficients, is the Pearson coefficient between the i-th operating parameter and the j-th operating parameter, is the variance of the product of the i-th operating parameter and the j-th operating parameter, , are the variances of the i-th operating parameter and the j-th operating parameter, respectively, is the mutual information of the i-th operating parameter and the j-th operating parameter, is the dynamic coupling value, and L is the total number of operating parameters, is the weight of the coupling degree between the i-th operating parameter and the j-th operating parameter.

[0042] In a second aspect, the present application provides a power distribution network fault detection and positioning system, the system comprising:

[0043] an information collection module configured to acquire electrical quantity data and environmental data of each section;

[0044] a feature extraction module configured to perform feature extraction on the electrical quantity data and the environmental data, respectively, to obtain first features and second features corresponding to the electrical quantity data and the environmental data, respectively;

[0045] a feature vector construction module configured to construct a feature vector according to the first features and the second features of each section at the same time, and construct a multi-dimensional dynamic feature vector according to a plurality of feature vectors at consecutive time points;

[0046] a fault recognition module configured to input the multi-dimensional dynamic feature vector into a pre-trained fault detection model to obtain a fault recognition result.

[0047] In a third aspect, the present application provides a storage medium storing one or more programs, which, when executed by a processor, implement the power distribution network fault detection and positioning method described above.

[0048] In a fourth aspect, the present application provides an electronic device comprising a memory and a processor, wherein:

[0049] the memory is configured to store a computer program;

[0050] the processor is configured to execute the computer program stored on the memory to implement the power distribution network fault detection and positioning method described above.

[0051] In summary, according to the power distribution network fault detection and positioning method described above, first, by constructing a multi-dimensional dynamic feature vector based on the fusion of electrical quantities and environmental quantities, combined with high-frequency sampling and section division technology, the temporal and spatial resolution of data acquisition is improved, and the dynamic coupling relationship between multi-source data is revealed, significantly enhancing the robustness and accuracy of fault identification. Second, a CNN-LSTM-SVM hybrid model architecture is used, which learns spatial features and time series features jointly, overcoming the limitations of single models and achieving efficient fault classification under complex conditions. Finally, based on the fault section positioning results, the dynamic coupling value of the node-level operating parameters is further quantified, improving the fault positioning accuracy from the section level to the node level, and significantly shortening the fault troubleshooting time. BRIEF DESCRIPTION OF DRAWINGS

[0052] Figure 1 The flow chart of the power distribution network fault detection and positioning method according to an embodiment of the present application is shown in

[0053] Figure 2 The structural diagram of the power distribution network fault detection and positioning system according to an embodiment of the present application is shown in

[0054] The following specific embodiments will further illustrate the present application in conjunction with the above-mentioned drawings. DETAILED DESCRIPTION

[0055] In order to make the purpose, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely below. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application. Unless otherwise defined, the technical terms or scientific terms used herein should be understood as the usual meaning understood by those skilled in the art in the field to which the present application belongs. The words such as "include" and similar words used herein mean that the elements or objects before the words cover the elements or objects listed after the words and their equivalents, without excluding other elements or objects.

[0056] Please refer to Figure 1 , which shows a power distribution network fault detection and positioning method according to an embodiment of the present application. The method includes steps S101 to S104, wherein:

[0057] Step S101: Obtain electrical quantity data and environmental data for each section;

[0058] It should be noted that in order to accurately realize fault location of the power distribution network and improve the efficiency in the process of fault location, the target power distribution network is first divided into multiple sections, and each section is numbered, and then each section is checked for faults. Specifically, the bus zero sequence current, bus zero sequence voltage, three-phase current of each section, section voltage, line current, section temperature, section humidity of each section are collected every first preset time, which constitutes the electrical quantity data and environmental data.

[0059] It should be noted that the first preset time is set to continuously and real-timely collect data related to the power distribution network, so as to comprehensively reflect the running state of each section of the power distribution network. The first preset time can be 1s, 3s, etc.

[0060] Step S102: feature extraction is performed on the electrical quantity data and the environmental data respectively, to obtain first features and second features corresponding to the electrical quantity data and the environmental data respectively;

[0061] It should be noted that the first features include first energy features, phase features, frequency features, second energy features, third energy features, fourth energy features, unbalance degree, voltage difference between two ends, fifth energy features, fluctuation rate; the second features include temperature-humidity coupling features.

[0062] In some embodiments, the bus zero sequence current is extracted according to the following formula:

[0063]

[0064]

[0065] wherein, is the kth harmonic of the bus zero sequence current of the ith section at the tth moment, is the total number of sampling points, is the bus zero sequence current of the ith section at the tth moment, is the first energy feature extracted from the bus zero sequence current of the ith section at the tth moment, and M represents the total number of harmonics; in addition, the second energy feature to the fifth energy feature are obtained in the same way as the first energy feature, except that the processing objects are different, so they are not repeated in this embodiment.

[0066] In addition, the phase feature extraction and the frequency feature extraction are conventional technologies, and the zero sequence voltage is determined, i.e. can be directly obtained. The fluctuation rate of the line current is the difference between the line current at the current moment and the line current at the previous moment, and then divided by the line current at the previous moment.

[0067] In addition, the three-phase current is extracted according to the following formula:

[0068]

[0069] wherein, is the unbalance degree corresponding to the three-phase current of the i-th section at the t-th moment, are one of the three-phase currents;

[0070] The section temperature and the section humidity are characterized according to the following formula:

[0071]

[0072] wherein, is the temperature-humidity coupling feature of the i-th section at the t-th moment, is the temperature of the i-th section at the t-th moment, is the humidity of the i-th section at the t-th moment.

[0073] wherein, are the phase feature and the frequency feature corresponding to the zero sequence voltage respectively, are the second energy feature, the third energy feature and the fourth energy feature corresponding to the three-phase current of the i-th section at the t-th moment respectively, is the voltage difference across the ends of the i-th section at the t-th moment, is the fifth energy feature corresponding to the line current of the i-th section at the t-th moment, is the line current fluctuation rate of the i-th section at the t-th moment;

[0074] In summary, specific feature extraction formulas are designed for different types of electrical quantity data and environmental data such as bus zero sequence current, three-phase current, section temperature, humidity, etc. Through these formulas, features closely related to various types of data and representative features can be accurately extracted, such as the harmonic energy features of the bus zero sequence current, the unbalance degree features of the three-phase current, and the temperature-humidity coupling features, etc. This provides high-quality feature information for subsequent fault detection and positioning, and helps to improve the accuracy and reliability of fault detection.

[0075] Step S103: constructing a feature vector according to the first feature and the second feature of each section at the same moment, and constructing a multi-dimensional dynamic feature vector according to a plurality of feature vectors at consecutive moments;

[0076] In some embodiments, the feature vector is constructed according to the following formula:

[0077]

[0078] wherein,​ respectively, are the phase characteristics and frequency characteristics corresponding to the zero sequence voltage, respectively, are the second energy characteristic, the third energy characteristic, and the fourth energy characteristic corresponding to the three-phase current of the i-th section at the t-th moment, is the voltage difference across the i-th section at the t-th moment, is the fifth energy characteristic corresponding to the line current of the i-th section at the t-th moment, is the line current fluctuation rate of the i-th section at the t-th moment;

[0079] The total number of moments is taken as the number of rows, and a multi-dimensional feature vector is constructed according to the feature vectors.

[0080] By extracting features according to different types of data, a feature vector containing various feature information is constructed, such as the phase and frequency characteristics of the zero sequence voltage, various energy characteristics of the three-phase current, the fluctuation rate of the line current, etc. This construction method effectively integrates multi-source features, so that the feature vector can comprehensively describe the operation state of the distribution network section at a certain moment, lays a foundation for subsequent construction of a multi-dimensional dynamic feature vector, and helps to improve the recognition ability of the fault detection model for complex fault modes. In addition, by constructing a multi-dimensional feature vector with the total number of moments as the number of rows, the operation characteristics of the distribution network section at different moments are comprehensively represented, forming a multi-dimensional dynamic feature vector that can reflect the dynamic operation process of the distribution network. This dynamic feature representation method can capture the change law of the operation state of the distribution network over time, so that the fault detection model can better learn the feature differences before and after the fault occurs, improve the real-time performance and accuracy of fault detection, and timely discover potential fault hazards.

[0081] Step S104: inputting the multi-dimensional dynamic feature vector into a pre-trained fault detection model to obtain a fault recognition result.

[0082] ​In this step, with respect to the construction process of the pre-trained fault detection model, power distribution network data under known faults and normal operation at continuous time points is acquired, and historical multi-dimensional feature vectors are constructed according to the power distribution network data; a CNN is used to extract spatial features of the feature vector sequence, to obtain a feature map, and the feature map is input into an LSTM for time sequence feature learning, to obtain a time feature vector, and the time feature vector is input into an SVM for fault classification training, to obtain a trained hybrid model. The CNN-LSTM-SVM hybrid model is used for fault classification training, which fully gives play to the advantages of each model. The CNN can effectively extract spatial features of the feature vector sequence and capture spatial correlation information between features; the LSTM is good at processing time sequence data and can learn the change law of the power distribution network operation state over time; and the SVM, as a powerful classification model, can accurately classify the learned time feature vector. This hybrid model architecture improves the overall performance of the fault detection model and can more accurately identify fault types in the power distribution network.

[0083] In addition, in some embodiments, in order to accurately and efficiently locate the fault node, the fault section number of the fault occurrence is also acquired according to the fault identification result, and the operation parameters of each node in the fault section are recalled according to the fault section number, including the voltage amplitude, the current harmonic distortion rate, the temperature fluctuation value and the power factor; the dynamic coupling value between each operation parameter under the same node is calculated, and it is judged whether the dynamic coupling value is within a preset threshold range; if the dynamic coupling value is not within the preset threshold range, the node is a fault node; if the dynamic coupling value is within the preset threshold range, the node is a normal node.

[0084] The dynamic coupling value is calculated according to the following formula:

[0085]

[0086]

[0087] wherein, is the coupling degree between the i th operation parameter and the j th operation parameter, , , are all weight coefficients, is the Pearson coefficient between the i th operation parameter and the j th operation parameter, is the variance of the product of the i th operation parameter and the j th operation parameter, , are respectively the variance of the i th operation parameter and the j th operation parameter, is the mutual information of the i th operation parameter and the j th operation parameter, is a dynamic coupling value, L is the total number of operating parameters, is the weight of the coupling degree between the i-th operating parameter and the j-th operating parameter.

[0088] By analyzing the dynamic coupling value of the node operating parameters, not only can the fault node be determined, but also a certain reference basis can be provided for the analysis of the fault cause. When the dynamic coupling value is not within the preset threshold range, it indicates that there is an abnormal relationship between the operating parameters of the node, which may be caused by equipment failure, line aging, environmental influence and other factors. Maintenance personnel can further analyze the fault cause according to these abnormal parameters, take targeted maintenance measures, improve the quality and effect of maintenance, and ensure the long-term stable operation of the power distribution network.

[0089] In summary, according to the above power distribution network fault detection and positioning method, first, by constructing a multi-dimensional dynamic feature vector based on the fusion of electrical quantities and environmental quantities, combined with high-frequency sampling and section division technology, not only the spatio-temporal resolution of data acquisition is improved, but also the dynamic coupling relationship between multi-source data is revealed, which significantly enhances the robustness and accuracy of fault identification. Secondly, the CNN-LSTM-SVM hybrid model architecture is adopted, and the joint learning of spatial features and time series features is realized, which overcomes the limitations of single model and realizes efficient fault classification under complex working conditions. Finally, based on the fault section positioning result, the dynamic coupling value of the node-level operating parameters is further quantified, the fault positioning accuracy is improved from the section level to the node level, and the fault troubleshooting time is greatly shortened.

[0090] As shown in Figure 2 , an embodiment of the present application proposes a power distribution network fault detection and positioning system, the system comprises:

[0091] An information acquisition module 10 is configured to acquire electrical quantity data and environmental data of each section;

[0092] A feature extraction module 20 is configured to extract features from the electrical quantity data and the environmental data respectively, to obtain first features and second features corresponding to the electrical quantity data and the environmental data respectively;

[0093] A feature vector construction module 30 is configured to construct a feature vector according to the first features and the second features of each section at the same time, and to construct a multi-dimensional dynamic feature vector according to a plurality of feature vectors at consecutive time points;

[0094] A fault identification module 40 is configured to input the multi-dimensional dynamic feature vector into a pre-trained fault detection model to obtain a fault identification result.

[0095] In another aspect, the present application also proposes a storage medium having one or more programs stored thereon, which programs are executed by a processor to implement the above-mentioned power distribution network fault detection and positioning method.

[0096] Another aspect of the present application also provides an electronic device comprising a memory and a processor, wherein the memory is configured to store a computer program, and the processor is configured to execute the computer program stored in the memory to implement the power distribution network fault detection and location method described above.

[0097] Those skilled in the art can understand that the logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a list of executable instructions for implementing the logic function, which can be embodied in any computer readable medium for use by or in connection with an instruction execution system, apparatus or device, such as a computer-based system, a system including a processor or other system that can fetch the instructions from the instruction execution system, apparatus or device and execute the instructions, or in conjunction with these instruction execution systems, apparatus or devices. For the purpose of this specification, the "computer readable medium" can be any device that can contain, store, communicate, propagate or transport the program for use by or in connection with the instruction execution system, apparatus or device, or in conjunction with these instruction execution systems, apparatus or devices.

[0098] More specific examples (a non-exhaustive list) of the computer readable medium include the following: an electrical connection having one or more wires (electrical devices), a portable computer diskette (magnetic devices), a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). In addition, the computer readable medium can even be paper or another suitable medium upon which the program is printed, as the program can be electronically captured, for example, by optically scanning the paper or other suitable medium, then electronically converted into a form that is suitable for use in a computer system, stored, and / or otherwise processed.

[0099] It should be understood that parts of the present application can be implemented in hardware, software, firmware or a combination thereof. In the above described embodiments, a plurality of steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented in hardware, and as in another embodiment, any of the following technologies known in the art or a combination thereof can be used: discrete logic circuitry having logic gates for implementing logic functions on data signals, application specific integrated circuits having appropriate combinational logic gates, programmable gate arrays (PGA), field programmable gate arrays (FPGA), etc.

[0100] While the embodiments of the application have been illustrated and described in detail, it will be readily apparent to those skilled in the art that various modifications and changes can be made to the embodiments without departing from the scope and spirit of the application, as described in the claims. Moreover, the application described is not limited in its application to the details set forth in the description or illustrated in the drawings. The application is capable of other embodiments and of being practiced or carried out in various ways.

Claims

1. A method for detecting and locating faults in a distribution network, characterized in that: The method comprises: Obtain electrical quantity data and environmental data for each section; performing feature extraction on the electrical quantity data and the environmental data to obtain first features and second features corresponding to the electrical quantity data and the environmental data, respectively; Construct a feature vector based on the first feature and the second feature of each segment at the same time, and construct a multi-dimensional dynamic feature vector based on multiple feature vectors at consecutive times; The multi-dimensional dynamic feature vector is input into a pre-trained fault detection model to obtain a fault identification result.

2. The method for detecting and locating distribution network faults according to claim 1, wherein: The step of obtaining electrical quantity data and environmental data of each section includes: Divide the target distribution network into multiple sections and number each section; The bus zero-sequence current, bus zero-sequence voltage, three-phase current of each section, voltage at both ends of the section, line current, section temperature, and section humidity of each section are collected at first preset time intervals.

3. The method for detecting and locating distribution network faults according to claim 2, wherein: The step of extracting features from the electrical quantity data and the environmental data to obtain first features and second features corresponding to the electrical quantity data and the environmental data, respectively, comprises: The busbar zero-sequence current is characterized by the following formula: in, is the kth harmonic of the busbar zero-sequence current from the i-th section at the t-th moment, is the total number of sampling points, is the busbar zero-sequence current of the ith section at the tth moment, is the first energy feature extracted from the bus zero-sequence current of the i-th section at the t-th time, M represents the total number of harmonics; The three-phase current is characterized by the following formula: in, is the unbalance degree corresponding to the three-phase current of the i-th section at the t-th moment, Each is one phase current among the three-phase currents; The segment temperature and segment humidity are extracted according to the following formula: in, is the temperature-humidity coupling characteristic of the i-th segment at the t-th moment, is the temperature of the i-th section at the t-th moment, is the humidity of the i-th segment at the t-th moment.

4. The method for detecting and locating distribution network faults according to claim 3, wherein: The steps of constructing a feature vector based on the first feature and the second feature of each segment at the same time, and constructing a multi-dimensional dynamic feature vector based on multiple feature vectors at consecutive times include: The feature vector is constructed according to the following formula: in, Respectively The phase characteristics and frequency characteristics corresponding to the zero-sequence voltage, are the second energy characteristic, the third energy characteristic, and the fourth energy characteristic corresponding to the three-phase current of the i-th section at the t-th moment, is the voltage difference between the two ends of the i-th segment at the t-th moment, is the fifth energy characteristic corresponding to the line current of the i-th section at the t-th time, is the line current fluctuation rate of the i-th section at the t-th moment; The total number of moments is used as the number of rows, and a multidimensional feature vector is constructed according to the feature vector.

5. The method for detecting and locating distribution network faults according to claim 4, wherein: The step of inputting the multi-dimensional dynamic feature vector into a pre-trained fault detection model to obtain a fault identification result includes: Acquire distribution network data at consecutive moments of known faults and normal operation, and construct a historical multidimensional feature vector based on the distribution network data; CNN is used to extract spatial features from the feature vector sequence to obtain a feature map, which is then input into LSTM for time series feature learning to obtain a time feature vector, which is then input into SVM for fault classification training to obtain a trained hybrid model.

6. The method for detecting and locating distribution network faults according to claim 5, characterized in that: After the step of inputting the multi-dimensional dynamic feature vector into a pre-trained fault detection model to obtain a fault identification result, the method further includes: Obtain the fault section number where the fault occurred based on the fault identification result, and retrieve the operating parameters of each node in the fault section based on the fault section number, the operating parameters including voltage amplitude, current harmonic distortion rate, temperature fluctuation value and power factor; Calculating the dynamic coupling value between various operating parameters under the same node, and determining whether the dynamic coupling value is within a preset threshold range; If the dynamic coupling value is not within the preset threshold range, the node is a faulty node; If the dynamic coupling value is within a preset threshold range, the node is a normal node.

7. The method for detecting and locating distribution network faults according to claim 6, wherein: The step of calculating the dynamic coupling value between various operating parameters under the same node includes: Calculated according to the following formula: in, is the coupling degree between the i-th operating parameter and the j-th operating parameter, 、 、 are weight coefficients, is the Pearson coefficient between the i-th operating parameter and the j-th operating parameter, is the variance of the product of the i-th operating parameter and the j-th operating parameter, 、 are the variances of the i-th and j-th operating parameters, respectively. is the mutual information between the i-th operating parameter and the j-th operating parameter, is the dynamic coupling value, L is the total number of operating parameters, is the weight of the coupling degree between the i-th operating parameter and the j-th operating parameter.

8. A distribution network fault detection and location system, characterized in that: The system comprises: Information acquisition module, used to obtain electrical quantity data and environmental data of each section; a feature extraction module, configured to extract features from the electrical quantity data and the environmental data, respectively, to obtain first features and second features corresponding to the electrical quantity data and the environmental data, respectively; A feature vector construction module is used to construct a feature vector based on the first feature and the second feature of each segment at the same time, and to construct a multi-dimensional dynamic feature vector based on multiple feature vectors at consecutive times; The fault identification module is used to input the multi-dimensional dynamic feature vector into a pre-trained fault detection model to obtain a fault identification result.

9. A storage medium, characterized in that: The storage medium stores one or more programs, which, when executed by a processor, implement the distribution network fault detection and location method according to any one of claims 1 to 7.

10. An electronic device comprising a memory and a processor, wherein: The memory is used to store computer programs; When the processor is used to execute the computer program stored in the memory, it implements the distribution network fault detection and location method according to any one of claims 1 to 7.

Images