Power communication network fault diagnosis method, system and device based on deep learning

Abstract

The present invention discloses a method, system and device for diagnosing faults in a power communication network based on deep learning, and belongs to the technical field of power systems; in the present invention, the operation information of the power communication network is obtained, and the operation information is pre-processed, the stable pass value is integrated to obtain an operation data matrix, and the operation data matrix is ​​calculated and processed to obtain an operation data set; the operation data set is feature extracted to obtain an operation data feature sequence, and the power communication network operation curve is obtained according to the operation data feature sequence. In the present invention, by further analyzing and calculating the data in normal operation in the power communication network, the change trend of various data of the power communication network behind the stable pass value is analyzed, so as to realize the prediction and diagnosis of power communication network faults; at the same time, according to the abnormal signal generated by the diagnosis result, a corresponding maintenance and processing plan is formulated, including regular inspection, equipment maintenance, etc., to ensure the stable operation of the power communication network.

Description

Technical Field

[0001] The present invention relates to the technical field of power systems, and more specifically, to a method, system and device for diagnosing power communication network faults based on deep learning. Background Art

[0002] Power communication was originally developed to meet the needs of power systems for remote monitoring and control. With the continuous development and expansion of power systems, there is a need for real-time monitoring and control of power equipment distributed in a large area, which has prompted the application and development of power communication technology. Power communication is mainly used in power systems, including power production, transmission, distribution and consumption. It is widely used in power equipment and systems such as power plants, substations, and distribution networks to achieve functions such as remote monitoring, fault diagnosis, data acquisition and remote control of power grids. The power communication network plays an important role in key areas such as monitoring and control, emergency communications, and data transmission of power communication systems. Therefore, the occurrence of faults may lead to widespread impacts, and timely diagnosis and prevention of faults are required to ensure the stable operation and rapid recovery of the power communication system, thereby improving the reliability and safety of the power system.

[0003] Existing power communication fault diagnosis will contain a large amount of operating information. Manual processing will occupy a large number of operation and maintenance personnel, which is inefficient and prone to data consistency problems. As a result, the accuracy of manual fault judgment, analysis, and positioning is low, and the fault handling efficiency is not high. Summary of the invention

[0004] In view of the shortcomings of the prior art, the purpose of the present invention is to provide a power communication network fault diagnosis method, system and device based on deep learning.

[0005] The purpose of the present invention can be achieved by the following technical solutions:

[0006] A method for fault diagnosis of a power communication network based on deep learning comprises the following steps:

[0007] Step 1: Obtaining the operation information of the power communication network, and preprocessing the operation information to obtain a stable pass value and a comparison result representing the operation status of the power communication network;

[0008] Step 2: Integrate the stable pass values ​​to obtain an operation data matrix, and obtain an operation data set by calculating and processing the operation data matrix;

[0009] Step 3: Extract features from the operation data set to obtain an operation data feature sequence, and obtain an operation curve of the power communication network according to the operation data feature sequence;

[0010] Step 4: Obtain historical stable pass values ​​and historical disturbance pass values, process the power communication network operation curve according to the historical stable pass values ​​and historical disturbance pass values, and obtain the secondary operation processing value for abnormality determination;

[0011] Step 5: Develop a corresponding maintenance plan based on the secondary operation processing value, and dispatch corresponding staff to conduct fault inspection and repair;

[0012] Step 6: Generate a report on the operation information, secondary operation processing value and maintenance plan of the power communication network and upload it to the cloud for recording.

[0013] As a further solution of the present invention, the operation information in step one includes network operation bandwidth, signal strength, network traffic throughput and terminal load.

[0014] The steps for preprocessing the operation information in step 1 include:

[0015] Using the formula The network operation bandwidth, signal strength, network traffic throughput and terminal load are normalized to obtain the first-level operation processing value , the first level operation processing value Compare with the operation alarm value to generate a comparison result. The operation alarm value is the pre-set warning value in the power communication network. If the operation alarm value is exceeded, the entire system will give an alarm prompt;

[0016] Wherein, DKi is the real-time network operation bandwidth, i=1, 2, 3...n, XHi is the real-time signal strength; LLi is the real-time network traffic throughput; FZi is the real-time terminal load; a1, a2, a3 and a4 are preset proportional coefficients and are not zero;

[0017] When the first-level operation processing value is less than the operation alarm value, it means that the power communication network is operating normally, a normal signal is generated, and the first-level operation processing value less than the operation alarm value is defined as a stable pass value;

[0018] When the first-level operation processing value is greater than or equal to the operation alarm value, it means that there is an abnormality in the power communication network, and an abnormal signal is generated. The first-level operation processing value greater than or equal to the operation alarm value is defined as the disturbance pass value.

[0019] As a further solution of the present invention, the stable traffic value in step 2 is normalized by a linear function Processing to obtain standardized stable traffic value of telecommunication communication network , according to the standardized stable traffic value of the telecommunication network, the operation data matrix is ​​obtained; the operation data matrix is ​​expressed as: ;

[0020] in, is the stable traffic value of the telecommunication network, H is the operation data matrix, arrive is the normalized stable traffic value of the telecommunication network from the first moment to the nth moment;

[0021] The running data matrix is ​​calculated by principal component analysis to obtain the running data matrix for dimension reduction. The calculation formula is: ; Wherein, P is the dimension reduction operation data matrix with z rows and x columns, Vi is the principal component contribution rate, H is the operation data matrix, and the operation data set is obtained by clustering the dimension reduction operation data matrix;

[0022] As a further solution of the present invention, a stable operation feature sequence and a disturbed operation feature sequence are obtained by extracting features from historical stable passage values ​​and historical disturbed passage values ​​respectively;

[0023] According to the stable operation characteristic sequence and the disturbance operation characteristic sequence, the stable operation curve and the disturbance operation curve are obtained respectively. According to the power communication network operation curve and the stable operation curve, the first Euclidean distance is obtained by calculating the Euclidean distance. The calculation formula is: ;

[0024] The second Euclidean distance is obtained by calculating the Euclidean distance according to the operation curve of the power communication network and the disturbance operation curve. The calculation formula is: ;

[0025] The secondary operation processing value is obtained by calculation according to the first Euclidean distance and the second Euclidean distance through the fuzzy decision model.

[0026] As a further solution in the present invention, the calculation formula of the secondary operation processing value is: ;

[0027] Among them, YCZB is the secondary operation processing value, β and is a fuzzy operator and is not zero, DT1 is the first Euclidean distance, and DT2 is the second Euclidean distance;

[0028] The secondary operation processing value is compared with the abnormal threshold. When the secondary operation processing value is higher than the abnormal threshold, it is determined that the operation of the power communication network is unstable and a secondary abnormal signal is generated.

[0029] When the secondary operation processing value is lower than or equal to the abnormal threshold, the power communication network is judged to be operating stably and a stable operation signal is generated.

[0030] As a further solution of the present invention, the abnormal signal in the comparison result between the first-level operation processing value and the operation alarm value and the second-level abnormal signal in the comparison result between the second-level operation processing value and the abnormal threshold are obtained, the abnormal signal and the second-level abnormal signal are combined, and according to the time sequence of the abnormal signals, the staff are arranged to deal with the abnormalities with the longest time in priority, and a maintenance plan is generated based on this for inspection and maintenance.

[0031] As a further solution of the present invention, a power communication network fault diagnosis system based on deep learning includes a data acquisition module, an edge computing module, an abnormality diagnosis module, a preventive maintenance module and a cloud storage module;

[0032] The data acquisition module is used to extract the operation information of the power communication network, pre-process the operation information to obtain the stable pass value that needs further analysis, and send the stable pass value to the edge computing module;

[0033] The edge computing module is used to integrate the stable pass values ​​and perform preliminary processing on the algorithm to obtain the operation data matrix, and send the operation data matrix to the abnormal diagnosis module;

[0034] The abnormal diagnosis module extracts features from the operation data set to obtain an operation data feature sequence, and obtains the power communication network operation curve based on the operation data feature sequence; obtains historical stable pass values ​​and historical disturbance pass values, obtains an anomaly detection model through deep learning based on the historical stable pass values ​​and historical disturbance pass values, and processes the power communication network operation curve through the anomaly detection model to obtain a secondary operation processing value;

[0035] The preventive maintenance module is used to dispatch corresponding staff to perform fault inspection and repair according to the secondary operation processing value;

[0036] The cloud storage module is used to store the operation information, secondary operation processing values ​​and maintenance plans of the power communication network.

[0037] As a further solution of the present invention, a power communication network fault diagnosis device based on deep learning includes a memory and a processor, the memory stores a computer program, and the processor implements a power communication network fault diagnosis method when executing the computer program.

[0038] Compared with the existing solutions, the present invention has the following beneficial effects:

[0039] In the present invention, by further analyzing and calculating the data in normal operation of the power communication network, the change trend of various data of the power communication network behind the stable pass value is analyzed to achieve the prediction and diagnosis of power communication network faults; at the same time, according to the abnormal signal generated by the diagnosis result, a corresponding maintenance processing plan is formulated, including regular inspection, equipment maintenance, etc., to ensure the stable operation of the power communication network;

[0040] The present invention first obtains the operation information of the electric power communication network, pre-processes the operation information, obtains the first-level operation processing value for preliminary judgment, further judges the stable pass value in the first-level operation processing value, and improves the reliability of fault diagnosis in the electric power communication network; uses the stable pass value in the first-level operation processing value for further analysis to obtain the operation curve of the electric power communication network, and then processes the operation curve of the electric power communication network through an abnormal detection model to obtain the second-level operation processing value, realizes secondary judgment of the data, analyzes the change trend of various data of the electric power communication network behind the stable pass value, and improves the accuracy of fault diagnosis. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 This is a flow chart of a power communication network fault diagnosis method based on deep learning in the present invention.

[0042] Figure 2 This is a system block diagram of a power communication network fault diagnosis system based on deep learning in the present invention. DETAILED DESCRIPTION

[0043] The present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. The embodiments of the present invention are provided for the purpose of illustration and description, and are not intended to be exhaustive or to limit the present invention to the disclosed forms. Many modifications and variations will be apparent to those of ordinary skill in the art. The embodiments are selected and described in order to better illustrate the principles and practical applications of the present invention, and to enable those of ordinary skill in the art to understand the present invention and thereby design various embodiments with various modifications suitable for specific uses. Example

[0044] Reference Figure 1 As shown, the present invention discloses a power communication network fault diagnosis method based on deep learning, comprising the following steps:

[0045] Step 1: Acquire the operation information of the power communication network, and pre-process the operation information to obtain the primary processing value, the stable pass value in the primary processing value, and the comparison result representing the operation status of the power communication network;

[0046] Step 2: Integrate the stable pass values ​​in the pre-processed operation information to obtain an operation data matrix, and process the operation data matrix through a fusion algorithm to obtain an operation data set;

[0047] Step 3: Extract features from the operation data set to obtain an operation data feature sequence, and obtain an operation curve of the power communication network according to the operation data feature sequence;

[0048] Step 4: Obtain historical stable pass values ​​and historical disturbance pass values, obtain an anomaly detection model through deep learning based on the historical stable pass values ​​and historical disturbance pass values, and process the power communication network operation curve through the anomaly detection model to obtain a secondary operation processing value;

[0049] Step 5: Develop a corresponding maintenance plan based on the secondary operation processing value, and dispatch corresponding staff to conduct fault inspection and repair;

[0050] Step 6: Generate a report on the operation information, secondary operation processing value and maintenance plan of the power communication network and upload it to the cloud for recording.

[0051] The operation information includes network operation bandwidth, signal strength, network traffic throughput and terminal load; the operation steps for preprocessing the operation information include:

[0052] Using the formula The network operation bandwidth, signal strength, network traffic throughput and terminal load are normalized to obtain the first-level operation processing value The first-level operation processing value is the normalization processing of the operation information in the power communication network, and is used to preliminarily determine the operation status of the power communication network. Compare with the operation alarm value to generate a comparison result. The operation alarm value is the pre-set warning value in the power communication network. If the operation alarm value is exceeded, it means that the operation of the power communication network is abnormal, and the whole system will issue an alarm prompt;

[0053] Wherein, DKi is the real-time network operation bandwidth, i=1, 2, 3...n, XHi is the real-time signal strength; LLi is the real-time network traffic throughput; FZi is the real-time terminal load; a1, a2, a3 and a4 are preset proportional coefficients and are not zero;

[0054] When the first-level operation processing value is less than the operation alarm value, it means that the operation state of the power communication network is normal operation, a normal signal is generated, and the first-level operation processing value less than the operation alarm value is defined as a stable pass value;

[0055] When the first-level operation processing value is greater than or equal to the operation alarm value, it means that the operation state of the power communication network is abnormal, an abnormal signal is generated and an alarm is issued, and the first-level operation processing value greater than or equal to the operation alarm value is defined as a disturbance pass value; the disturbance pass value is recorded, and the disturbance pass value is eliminated; the stable pass value in the first-level operation processing value is further processed;

[0056] The disturbance pass value in the first-level operation processing value is saved, and the stable pass value in the first-level operation processing value is further analyzed to improve the reliability of fault diagnosis in the power communication network.

[0057] It should be further explained that the stable pass value represents that the power communication network can continue to operate stably. However, since the basic values ​​in the operation information of the power communication network change at any time and the change range is too large, it is necessary to further analyze the stable pass value;

[0058] The stable pass value is normalized by a linear function Processing to obtain standardized stable traffic value of telecommunication communication network , according to the standardized stable traffic value of the telecommunication network, the operation data matrix is ​​obtained; the operation data matrix is ​​expressed as: ;

[0059] in, is the stable pass value in the first-level operation processing value, H is the operation data matrix, arrive is the normalized stable traffic value of the telecommunication network from the first moment to the nth moment;

[0060] Principal component analysis is a statistical method used to simplify data sets, reduce data dimensions, and retain the main information of the data as much as possible. It transforms data from the original coordinate system to a new coordinate system through linear transformation, so that the axes in the new coordinate system, i.e., the principal components, are mutually orthogonal and arranged according to the size of the variance, thereby capturing the most important variation information in the data;

[0061] The running data matrix is ​​calculated by principal component analysis to obtain the running data matrix for dimension reduction. The calculation formula is: ; Wherein, P is the running data matrix for dimensionality reduction with z rows and x columns, Vi is the principal component contribution rate, H is the running data matrix, and the running data set is obtained by clustering processing based on the running data matrix for dimensionality reduction.

[0062] The operating data set is subjected to feature extraction to obtain an operating data feature sequence, and the power communication network operating curve is obtained according to the operating data feature sequence.

[0063] By extracting features from historical stable passage values ​​and historical disturbance passage values, stable operation feature sequences and disturbance operation feature sequences are obtained;

[0064] According to the stable operation characteristic sequence and the disturbance operation characteristic sequence, the stable operation curve and the disturbance operation curve are obtained respectively. According to the power communication network operation curve and the stable operation curve, the first Euclidean distance is obtained by calculating the Euclidean distance. The calculation formula is: ;

[0065] The second Euclidean distance is obtained by calculating the Euclidean distance according to the operation curve of the power communication network and the disturbance operation curve. The calculation formula is: ;

[0066] The secondary operation processing value is obtained by calculation based on the fuzzy decision model and the first Euclidean distance and the second Euclidean distance. The secondary operation processing value is a further data anomaly determination indicator of the stable pass value in the primary operation processing value. It is used to analyze the change trend of various data in the power communication network behind the stable pass value, and improve the fault diagnosis accuracy. The calculation formula is: ;

[0067] Among them, YCZB is the secondary operation processing value, β and is a fuzzy operator and is not zero, DT1 is the first Euclidean distance, and DT2 is the second Euclidean distance;

[0068] The secondary operation processing value is compared with the abnormal threshold. When the secondary operation processing value is higher than the abnormal threshold, the power communication network is judged to be unstable, a secondary abnormal signal is generated, and inspection and maintenance are carried out;

[0069] When the secondary operation processing value is lower than or equal to the abnormal threshold, the power communication network is determined to be in stable operation, and a stable operation signal is generated;

[0070] By further analyzing and calculating the stable pass values ​​in the normal operation of the power communication network, and analyzing the changing trends of various data of the power communication network behind the stable pass values, the prediction and diagnosis of power communication network faults can be achieved.

[0071] The abnormal signal and the secondary abnormal signal are combined, and according to the time sequence of the abnormal signals, the staff are arranged to deal with the abnormalities with the longest duration first, and a maintenance plan is generated for inspection and maintenance.

[0072] Generate a report on the operation information, secondary operation processing values ​​and maintenance plan of the power communication network and upload it to the cloud for recording.

[0073] It should be further explained that the existing power communication fault diagnosis will contain a large amount of operation information, and manual processing will occupy a large number of operation and maintenance personnel, which is inefficient and prone to data consistency problems, resulting in low accuracy of manual fault judgment, analysis, and positioning, and low fault handling efficiency;

[0074] The present invention further analyzes and calculates the data in normal operation in the power communication network, analyzes the changing trends of various data in the power communication network behind the stable pass value, and realizes the prediction and diagnosis of power communication network faults; at the same time, according to the abnormal signal generated by the diagnosis result, a corresponding maintenance and processing plan is formulated, including regular inspection, equipment maintenance, etc., to ensure the stable operation of the power communication network and reduce the problems of low efficiency and low accuracy caused by manual processing; and the present invention provides high decision-making accuracy, which can restore the power communication network to a normal state in time and ensure the stable operation of the power communication network.

[0075] The present invention also includes a power communication network fault diagnosis device based on deep learning, including a memory and a processor, the memory stores a computer program, the processor is used to execute the computer program, and the computer program includes steps one to six in the above-mentioned power communication network fault diagnosis method. Example

[0076] Reference Figure 2 As shown, based on Example 1, the present invention discloses a power communication network fault diagnosis system based on deep learning, including a data acquisition module, an edge computing module, an abnormality diagnosis module, a preventive maintenance module and a cloud storage module;

[0077] The data acquisition module is used to extract the operation information of the power communication network, pre-process the operation information to obtain the stable pass value that needs further analysis, and send the stable pass value to the edge computing module for further analysis to achieve accurate judgment and early warning of faults; the operation information includes network operation bandwidth, signal strength, network traffic throughput and terminal load;

[0078] The edge computing module is used to integrate the stable pass values ​​and perform preliminary processing on the algorithm to obtain the operation data matrix, and send the operation data matrix to the abnormal diagnosis module;

[0079] The abnormal diagnosis module extracts features from the operation data set to obtain an operation data feature sequence, and obtains the power communication network operation curve based on the operation data feature sequence; obtains historical stable pass values ​​and historical disturbance pass values, obtains an anomaly detection model through deep learning based on the historical stable pass values ​​and historical disturbance pass values, and processes the power communication network operation curve through the anomaly detection model to obtain a secondary operation processing value;

[0080] The preventive maintenance module is used to dispatch corresponding staff to perform fault inspection and repair according to the secondary operation processing value;

[0081] The cloud storage module is used to store the operation information, secondary operation processing values ​​and maintenance plans of the power communication network, providing data support for the operation of the system.

[0082] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0083] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", etc. should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be an indirect connection through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0084] The above is a detailed description of an embodiment of the present invention, but the content is only a preferred embodiment of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. A power communication network fault diagnosis method based on deep learning, characterized in that: The steps include: Step 1: Obtaining the operation information of the power communication network, and preprocessing the operation information to obtain a stable pass value and a comparison result representing the operation status of the power communication network; The steps for preprocessing the operation information include: Using the formula The network operation bandwidth, signal strength, network traffic throughput and terminal load are normalized to obtain the first-level operation processing value , the first level operation processing value Compare with the operation alarm value to generate a comparison result. The operation alarm value is the pre-set warning value in the power communication network. If the operation alarm value is exceeded, the entire system will give an alarm prompt; Wherein, DKi is the real-time network operation bandwidth, i=1, 2, 3...n, XHi is the real-time signal strength; LLi is the real-time network traffic throughput; FZi is the real-time terminal load; a1, a2, a3 and a4 are preset proportional coefficients and are not zero; When the first-level operation processing value is less than the operation alarm value, it means that the power communication network is operating normally, a normal signal is generated, and the first-level operation processing value less than the operation alarm value is defined as a stable pass value; When the first-level operation processing value is greater than or equal to the operation alarm value, it means that there is an abnormality in the power communication network, and an abnormal signal is generated. The first-level operation processing value greater than or equal to the operation alarm value is defined as the disturbance pass value; Step 2: Integrate the stable pass values ​​to obtain an operation data matrix, and obtain an operation data set by calculating and processing the operation data matrix; The stable pass values ​​are integrated to obtain the operation data matrix, and the operation data matrix is ​​processed by calculation to obtain the operation data set: the stable pass value is normalized by the linear function Processing to obtain standardized stable traffic value of telecommunication communication network , the operation data matrix is ​​obtained according to the normalized stable traffic value of the telecommunication network; the operation data matrix is ​​expressed as ; in, is the stable traffic value of the telecommunication network, H is the operation data matrix, arrive is the normalized stable traffic value of the telecommunication network from the first moment to the nth moment; The operating data matrix is ​​calculated by principal component analysis to obtain the operating data matrix for dimension reduction. The calculation formula is: ; Wherein, P is the dimension reduction operation data matrix with z rows and x columns, Vi is the principal component contribution rate, H is the operation data matrix, and the operation data set is obtained by clustering the dimension reduction operation data matrix; Step 3: Extract features from the operation data set to obtain an operation data feature sequence, and obtain an operation curve of the power communication network according to the operation data feature sequence; Step 4: Obtain historical stable pass values ​​and historical disturbance pass values, process the power communication network operation curve according to the historical stable pass values ​​and historical disturbance pass values, and obtain the secondary operation processing value for abnormality determination; The method of processing the power communication network operation curve according to the historical stable pass value and the historical disturbance pass value to obtain the secondary operation processing value for abnormal judgment is as follows: According to the stable operation characteristic sequence and the disturbance operation characteristic sequence, the stable operation curve and the disturbance operation curve are obtained respectively. According to the power communication network operation curve and the stable operation curve, the first Euclidean distance is obtained by calculating the Euclidean distance. The calculation formula is: ; The second Euclidean distance is obtained by calculating the Euclidean distance according to the operation curve of the power communication network and the disturbance operation curve. The calculation formula is: ; Obtaining a secondary operation processing value by calculating according to the first Euclidean distance and the second Euclidean distance through a fuzzy decision model; The calculation formula for the secondary operation processing value is: ; Among them, YCZB is the secondary operation processing value, β and is a fuzzy operator and is not zero, DT1 is the first Euclidean distance, and DT2 is the second Euclidean distance; The secondary operation processing value is compared with the abnormal threshold. When the secondary operation processing value is higher than the abnormal threshold, it is determined that the operation of the power communication network is unstable and a secondary abnormal signal is generated. When the secondary operation processing value is lower than or equal to the abnormal threshold, the power communication network is determined to be in stable operation, and a stable operation signal is generated; Step 5: Develop a corresponding maintenance plan based on the secondary operation processing value, and dispatch corresponding staff to conduct fault inspection and repair; Step 6: Generate a report on the operation information, secondary operation processing value and maintenance plan of the power communication network and upload it to the cloud for recording.

2. A method for fault diagnosis of a power communication network based on deep learning according to claim 1, characterized in that: The operation information in step 1 includes network operation bandwidth, signal strength, network traffic throughput and terminal load.

3. A method for fault diagnosis of a power communication network based on deep learning according to claim 2, characterized in that: Obtain the abnormal signal from the comparison result between the first-level operation processing value and the operation alarm value, and the second-level abnormal signal from the comparison result between the second-level operation processing value and the abnormal threshold, combine the abnormal signal and the second-level abnormal signal, and arrange staff to deal with the longest-lasting abnormality in chronological order according to the abnormal signals, and generate a maintenance plan for inspection and maintenance.

4. A power communication network fault diagnosis system based on deep learning, used to implement the diagnosis method according to any one of claims 1 to 3, characterized in that: It includes data acquisition module, edge computing module, abnormal diagnosis module, preventive maintenance module and cloud storage module; The data acquisition module is used to extract the operation information of the power communication network, pre-process the operation information to obtain the stable pass value that needs further analysis, and send the stable pass value to the edge computing module; The edge computing module is used to integrate the stable pass values ​​and perform preliminary processing on the algorithm to obtain the operation data matrix, and send the operation data matrix to the abnormal diagnosis module; The abnormal diagnosis module extracts features from the operation data set to obtain an operation data feature sequence, and obtains the power communication network operation curve based on the operation data feature sequence; obtains historical stable pass values ​​and historical disturbance pass values, obtains an anomaly detection model through deep learning based on the historical stable pass values ​​and historical disturbance pass values, and processes the power communication network operation curve through the anomaly detection model to obtain a secondary operation processing value; The preventive maintenance module is used to dispatch corresponding staff to perform fault inspection and repair according to the secondary operation processing value; The cloud storage module is used to store the operation information, secondary operation processing values ​​and maintenance plans of the power communication network.

5. A power communication network fault diagnosis device based on deep learning, comprising a memory and a processor, characterized in that: The memory stores a computer program, and the processor implements the steps of any one of claims 1 to 3 when executing the computer program.

Images