Multi-level feedback fault diagnosis method and system for power communication network

Abstract

The present application discloses a multi-level feedback power communication network fault diagnosis method and system, which relates to the technical field of power communication network, including collecting alarm data of the power communication network, extracting data from the alarm data, and obtaining the electric transformer information in the alarm data; establishing a multi-level feedback model, inputting the electric transformer information into the multi-level feedback model, and the multi-level feedback model analyzes the electric transformer information to obtain feedback results; according to the feedback results, extracting the fault logic relationship in the feedback results, and gradually decomposing the fault level according to the fault logic relationship to obtain component-level faults; according to the component-level faults, obtaining the faulty component, collecting the power frequency of the faulty component, and diagnosing the faulty component according to the power frequency and the feedback results to obtain the diagnosis results. The present application has the effect of improving the diagnostic accuracy and efficiency of the power communication network in diagnosing faults when a fault occurs.

Description

Technical Field

[0001] The present application relates to the technical field of power communication networks, and in particular to a method and system for diagnosing faults in power communication networks with multi-level feedback. Background Art

[0002] The power communication network is established to ensure the safe and stable operation of the power system. It covers all aspects of the power grid, from power generation, transmission, transformation to distribution. This network uses a variety of communication methods, such as satellites, microwaves, carriers, optical cables, etc., to build a three-dimensional cross-communication network.

[0003] During the operation of the power communication network, some faults are inevitable, resulting in communication interruption, unstable transmission, etc., causing power allocation to be untimely or even impossible to allocate.

[0004] As the topology of the power communication network becomes more and more complex, fault diagnosis of the power communication network is becoming more and more difficult. The traditional fault diagnosis of the power communication network is mainly carried out through fault analysis through the rule base, and the cause of the fault is determined by the rules in the rule base. However, it becomes increasingly difficult to inspect the increasingly complex power communication network in this way. At the same time, the rule base cannot accurately analyze the faults of the power communication network, resulting in a decrease in the efficiency of fault diagnosis of the power communication network. Summary of the invention

[0005] The purpose of the present invention is to provide a multi-level feedback power communication network fault diagnosis method and system to solve the problems raised by the above background technology.

[0006] In a first aspect, the present application provides a multi-level feedback power communication network fault diagnosis method, the method comprising the following steps:

[0007] Collecting alarm data from the power communication network, extracting data from the alarm data, and obtaining power transformer information in the alarm data;

[0008] Establishing a multi-level feedback model, inputting the electrical transformation information into the multi-level feedback model, and the multi-level feedback model analyzing the electrical transformation information to obtain a feedback result;

[0009] According to the feedback result, extract the fault logic relationship in the feedback result, and gradually decompose the fault level according to the fault logic relationship to obtain the component level fault;

[0010] According to the component-level fault, a faulty component is obtained, the power frequency of the faulty component is collected, and the faulty component is diagnosed according to the power frequency and the feedback result to obtain a diagnosis result.

[0011] Preferably, the step of collecting alarm data of the power communication network, extracting data from the alarm data, and obtaining the power transformer information in the alarm data is specifically as follows:

[0012] Collecting alarm data from the power communication network, performing data correlation analysis on the alarm data, and obtaining a fault correlation degree of each piece of data in the alarm data;

[0013] Determine whether the fault correlation is higher than a preset correlation threshold, and extract the alarm data corresponding to the fault correlation being higher than the correlation threshold to obtain an alarm data set;

[0014] The alarm data set is screened to obtain the electrical transformer information in the alarm data.

[0015] Preferably, a multi-level feedback model is established, the electrical transformation information is input into the multi-level feedback model, and the multi-level feedback model analyzes the electrical transformation information to obtain a feedback result, specifically:

[0016] Establishing a multi-level feedback model, the multi-level feedback model includes an identification and division layer, a low-level feedback loop layer, a high-level feedback loop layer, and a loop fusion layer;

[0017] The electric transformation information is input into the multi-level feedback model, and the identification and division layer identifies the electric transformation information to obtain a voltage change diagram, a current change diagram, and an electromagnetic change diagram of the electric transformation information;

[0018] Obtaining a surface characteristic value of the electric variable information based on the voltage change diagram, the current change diagram, and the electromagnetic change diagram, and comparing the surface characteristic value with a preset boundary characteristic value;

[0019] If it is determined that the surface characteristic value is lower than the boundary characteristic value, the electrical transformation information is input into the low-level feedback loop layer, and a low-level feedback result is output;

[0020] If it is determined that the surface characteristic value is higher than or equal to the boundary characteristic value, the electrical transformation information is input into the low-level feedback loop layer and the high-level feedback loop layer respectively, and a low-level feedback result and a high-level feedback result are output respectively;

[0021] The low-level feedback result and the high-level feedback result are input into the recurrent fusion layer, and the low-level feedback result and the high-level feedback result are fused to generate a feedback result.

[0022] Preferably, the step of inputting the electric transformation information into the low-level feedback loop layer and outputting the low-level feedback result is specifically:

[0023] Inputting the electric change information into the low-level feedback loop layer, the low-level feedback loop layer generating a simulated fault and a simulated communication signal according to the voltage change graph and the current change graph;

[0024] After the simulated communication signal passes through the simulated fault, a simulated voltage change graph and a simulated current change graph of the simulated communication signal are obtained;

[0025] Comparing the simulated voltage change graph with the voltage change graph to obtain a voltage difference, and comparing the simulated current change graph with the current change graph to obtain a current difference;

[0026] Substitute the voltage difference and the current difference into the simulated fault to generate a new simulated fault, and repeat the operation of simulating the communication signal passing through the new simulated fault and the difference comparison operation until the simulated voltage change graph and the simulated current change graph are the same as the voltage change graph and the current change graph, input the new simulated fault at this time, and output the new simulated fault as a low-level feedback result.

[0027] Preferably, the steps of inputting the electric transformation information into the low-level feedback loop layer and the high-level feedback loop layer respectively, and outputting the low-level feedback result and the high-level feedback result respectively, are specifically:

[0028] Inputting the electrical conversion information into the low-level feedback loop layer and the high-level feedback loop layer respectively, and the low-level feedback loop layer outputs a low-level feedback result according to the electrical conversion information;

[0029] After receiving the electrical change information, the advanced feedback loop layer converts the electrical change information into magnetic change information, and extracts an electromagnetic change map based on the magnetic change information;

[0030] Based on the electromagnetic change map, feature extraction is performed on the electromagnetic change map to obtain deep features of the electromagnetic change map;

[0031] According to the deep features, the current change graph, the voltage change graph and the electromagnetic change graph are vertically correlated to obtain a correlation change graph;

[0032] A multi-layer convolution cycle is performed on the association change graph to obtain an initial feedback result, and the initial feedback result is substituted into the association change graph for a secondary convolution cycle to obtain an advanced feedback result.

[0033] Preferably, according to the feedback result, extracting the fault logic relationship in the feedback result, and gradually decomposing the fault level according to the fault logic relationship to obtain the step of component level fault, specifically comprises:

[0034] According to the feedback result, extract multiple logical key points in the feedback result, and logically connect the multiple logical key points to obtain the fault logical relationship in the feedback result;

[0035] According to the alarm data, the outermost fault level of the electric power communication network fault is obtained, and a target logic point is generated according to the outermost fault level;

[0036] Matching the target logic point with the fault logic relationship to obtain a logical position of the target logic point in the fault logic relationship;

[0037] The outermost fault level is subjected to fault decomposition according to the logical position and the logical relationship to obtain a component level fault.

[0038] Preferably, the steps of obtaining a faulty component according to the component-level fault, collecting the power frequency of the faulty component, diagnosing the faulty component according to the power frequency and the feedback result, and obtaining a diagnosis result are specifically:

[0039] According to the component-level fault, a faulty component corresponding to the component-level fault is obtained, and the power frequency during the operation of the faulty component is extracted;

[0040] Comparing the power frequency with a preset standard power frequency to obtain a frequency difference, and obtaining a difference reason for the frequency difference according to the frequency difference;

[0041] According to the difference reason, the target component having an abnormality in the faulty component is obtained, and the abnormal state of the target component is obtained by combining the target component and the feedback result;

[0042] A fault diagnosis is performed on the abnormal state to obtain a diagnosis result that a fault has occurred in the power communication network.

[0043] In a second aspect, the present application provides a multi-level feedback power communication network fault diagnosis system, the system comprising:

[0044] Data initialization module: used to collect alarm data from the power communication network, extract data from the alarm data, and obtain the power transformer information in the alarm data;

[0045] Model analysis module: used to establish a multi-level feedback model, input the electric transformation information into the multi-level feedback model, and the multi-level feedback model analyzes the electric transformation information to obtain feedback results;

[0046] A hierarchical decomposition module is used to extract the fault logical relationship in the feedback result according to the feedback result, and gradually decompose the fault hierarchy according to the fault logical relationship to obtain the component level fault;

[0047] Diagnostic module: used to obtain a faulty component according to the component-level fault, collect the power frequency of the faulty component, diagnose the faulty component according to the power frequency and the feedback result, and obtain a diagnostic result.

[0048] In summary, the present application includes at least one of the following beneficial technical effects:

[0049] According to the alarm data collected on the power communication network, the electric transformer information in the alarm data is obtained, and the electric transformer information is input into the established multi-level feedback model. The electric transformer information is classified according to the surface characteristic value of the electric transformer information, and the electric transformer information is cyclically analyzed through the low-level feedback loop layer and the high-level feedback loop layer in the multi-level feedback model to obtain low-level feedback results and high-level feedback results, and the feedback results are output. The fault logic relationship is obtained according to the feedback result, and the fault level is decomposed according to the fault logic relationship to obtain the component-level fault. The fault component is obtained according to the component-level fault, and the power frequency in the fault component is collected. The fault component is comprehensively diagnosed according to the power frequency and the feedback result to obtain the diagnosis result. The diagnostic accuracy and efficiency of the power communication network in diagnosing faults when faults occur are improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 It is a flowchart of the steps of the multi-level feedback power communication network fault diagnosis method provided in the embodiment of the present application.

[0051] Figure 2 It is a module block diagram of a multi-level feedback power communication network fault diagnosis system provided in an embodiment of the present application.

[0052] Explanation of the accompanying drawings: 1. Data initialization module; 2. Model analysis module; 3. Hierarchical decomposition module; 4. Diagnosis module. DETAILED DESCRIPTION

[0053] The following combination Figure 1-Figure 2 The present application is further described in detail, but the embodiments of the present invention are not limited thereto.

[0054] In the embodiments of the present application, a multi-level feedback method and system for diagnosing faults in a power communication network are disclosed.

[0055] In this embodiment, a multi-level feedback power communication network fault diagnosis method comprises the following steps:

[0056] S100: Collecting alarm data from the power communication network, extracting the alarm data, and obtaining the power transformer information in the alarm data;

[0057] S200: establishing a multi-level feedback model, inputting the electric transformer information into the multi-level feedback model, and the multi-level feedback model analyzes the electric transformer information to obtain a feedback result;

[0058] S300: extracting fault logic relationships in the feedback results according to the feedback results, and gradually decomposing the fault levels according to the fault logic relationships to obtain component-level faults;

[0059] S400: According to the component-level fault, a faulty component is obtained, the power frequency of the faulty component is collected, and the faulty component is diagnosed according to the power frequency and the feedback result to obtain a diagnosis result.

[0060] It should be pointed out that the above steps are only a preferred implementation order. In the specific implementation process, some steps can be replaced without affecting the overall implementation effect.

[0061] The steps of collecting alarm data from the power communication network, extracting the alarm data, and obtaining the power transformer information in the alarm data are specifically as follows:

[0062] Collect alarm data from the power communication network, perform data correlation analysis on the alarm data, and obtain the fault correlation of each piece of data in the alarm data;

[0063] Determine whether the fault correlation is higher than a preset correlation threshold, and extract the alarm data corresponding to the fault correlation being higher than the correlation threshold to obtain an alarm data set;

[0064] The alarm data set is screened to obtain the electric transformer information in the alarm data.

[0065] In application, taking the electric power communication network in a certain area as an example, a device in the electric power communication network fails, and the electric power communication network collects the alarm data of the device. The alarm data contains a variety of data, such as the fault condition of the equipment, the signal data after the signal changes after passing through the device, etc., and the correlation analysis is performed on these data to obtain the fault correlation of each data, and the alarm data with a correlation higher than the correlation threshold is extracted, such as the fault condition, signal data, etc., and then these data are screened and converted into electric transformation information.

[0066] The steps of establishing a multi-level feedback model, inputting the electric transformer information into the multi-level feedback model, and analyzing the electric transformer information by the multi-level feedback model to obtain the feedback result are as follows:

[0067] Establish a multi-level feedback model, which includes a recognition and division layer, a low-level feedback loop layer, a high-level feedback loop layer, and a loop fusion layer;

[0068] The electric transformation information is input into the multi-level feedback model, and the identification and division layer identifies the electric transformation information to obtain the voltage change diagram, current change diagram and electromagnetic change diagram of the electric transformation information;

[0069] Based on the voltage change diagram, the current change diagram and the electromagnetic change diagram, the surface characteristic value of the electric variable information is obtained, and the surface characteristic value is compared with the preset boundary characteristic value;

[0070] If it is determined that the surface characteristic value is lower than the boundary characteristic value, the electrical transformation information is input into the low-level feedback loop layer, and the low-level feedback result is output;

[0071] If it is determined that the surface characteristic value is higher than or equal to the boundary characteristic value, the electrical transformation information is input into the low-level feedback loop layer and the high-level feedback loop layer respectively, and the low-level feedback result and the high-level feedback result are output respectively;

[0072] The low-level feedback result and the high-level feedback result are input into the recurrent fusion layer, and the low-level feedback result and the high-level feedback result are fused to generate the feedback result.

[0073] In application, taking the power communication network in a certain area as an example, a multi-level feedback model is established. After the electric transformer information is input into the multi-level feedback model, the identification and division layer will identify the electric transformer information and generate the voltage change diagram, current change diagram and electromagnetic change diagram in the electric transformer information. According to these three diagrams, the surface characteristic value of the electric transformer information, namely, the voltage change characteristic, the current change characteristic and the electromagnetic change characteristic, is obtained, and these three characteristics are fused. The surface characteristic value is compared with the demarcation characteristic value. If the surface characteristic value is lower than the demarcation characteristic value, it indicates that this fault is a common and simple fault. Then the electric transformer information is input into the low-level feedback loop layer, the low-level feedback result is output, and the low-level feedback result is output as the feedback result. If the surface characteristic value is higher than or equal to the demarcation characteristic value, it indicates that this fault is an uncommon and more complex fault. Then the electric transformer information is input into the low-level feedback loop layer and the high-level feedback loop layer at the same time, and the low-level feedback result and the high-level feedback result are output respectively, and the two are fused to generate the feedback result.

[0074] The steps of inputting the electric transformation information into the low-level feedback loop layer and outputting the low-level feedback result are as follows:

[0075] The electric transformation information is input into the low-level feedback loop layer, and the low-level feedback loop layer generates simulated faults and simulated communication signals according to the voltage change diagram and the current change diagram;

[0076] After the simulated communication signal is subjected to a simulated fault, a simulated voltage change graph and a simulated current change graph of the simulated communication signal are obtained;

[0077] The simulated voltage change graph is compared with the voltage change graph to obtain the voltage difference, and the simulated current change graph is compared with the current change graph to obtain the current difference;

[0078] Substitute the voltage difference and the current difference into the simulated fault to generate a new simulated fault, and repeat the operation of simulating the communication signal passing through the new simulated fault and the difference comparison operation until the simulated voltage change graph and the simulated current change graph are the same as the voltage change graph and the current change graph, input the new simulated fault at this time, and output the new simulated fault as a low-level feedback result.

[0079] In application, taking the power communication network in a certain area as an example, after the electric transformer information is input into the low-level feedback loop layer, the low-level feedback loop layer generates simulated faults and simulated communication signals according to the voltage change graph A and the current change graph B provided by the identification and division layer, and obtains the simulated voltage change graph C and the simulated current change graph D by passing the simulated communication signal through the simulated fault, and then compares A with C, B with D to obtain the voltage difference A / C and the current difference B / D, and substitutes the voltage difference A / C and the current difference B / D into the simulated fault to optimize the simulated fault, and then passes the simulated communication signal and repeats the above comparison process and the optimization process of the simulated fault until the voltage difference A / C and the current difference B / D are both 1, that is, there is no difference, and then the simulated fault is output as the low-level feedback result.

[0080] The steps of inputting the electric transformation information into the low-level feedback loop layer and the high-level feedback loop layer, and outputting the low-level feedback result and the high-level feedback result respectively, are specifically:

[0081] The electric transformation information is input into the low-level feedback loop layer and the high-level feedback loop layer respectively, and the low-level feedback loop layer outputs a low-level feedback result according to the electric transformation information;

[0082] After receiving the electrical change information, the advanced feedback loop layer converts the electrical change information into magnetic change information, and extracts the electromagnetic change map based on the magnetic change information;

[0083] Based on the electromagnetic change map, feature extraction is performed on the electromagnetic change map to obtain deep features of the electromagnetic change map;

[0084] According to the deep characteristics, the current change graph, the voltage change graph and the electromagnetic change graph are vertically correlated to obtain a correlation change graph;

[0085] A multi-layer convolution cycle is performed on the association change graph to obtain an initial feedback result, and the initial feedback result is substituted into the association change graph for a secondary convolution cycle to obtain an advanced feedback result.

[0086] In application, taking the power communication network in a certain area as an example, after the electric transformation information is input into the advanced feedback loop layer, the advanced feedback loop layer converts the electric transformation information into magnetic transformation information, that is, converts the current change signal in the electric transformation information into a magnetic field change signal, and generates an electromagnetic change map according to the converted magnetic field change signal. According to the electromagnetic change map, the electromagnetic change map is feature extracted, including the shape change feature of the magnetic field change, the magnetic field intensity change feature, etc., and then the above-obtained features are fused to obtain the deep features of the electromagnetic change map. According to the deep features, the current change map, the voltage change map and the electromagnetic change map are vertically associated, that is, the horizontal axes of the above three change maps are fused, and the three change maps are established on the same horizontal axis, and then the associated change map is obtained, and then the associated change map is subjected to multi-layer convolution cycles, that is, the associated change map is subjected to multiple convolution cycles from left to right along the direction of the horizontal axis to generate an initial feedback result, and then the initial feedback result is substituted into the associated change map and subjected to a second multi-layer convolution cycle to obtain an advanced feedback result.

[0087] According to the feedback results, the fault logic relationship in the feedback results is extracted, and the fault level is gradually decomposed according to the fault logic relationship to obtain the steps of component-level faults, which are specifically:

[0088] According to the feedback results, multiple logical key points in the feedback results are extracted, and the multiple logical key points are logically connected to obtain the fault logical relationship in the feedback results;

[0089] According to the alarm data, the outermost fault level of the power communication network fault is obtained, and the target logic point is generated according to the outermost fault level;

[0090] Matching the target logic point with the fault logic relationship to obtain the logical position of the target logic point in the fault logic relationship;

[0091] The outermost fault level is decomposed according to the logical position and logical relationship to obtain the component level fault.

[0092] In application, taking the electric power communication network in a certain area as an example, according to the feedback results, multiple logical keywords in the feedback results are extracted, and the logical key points are connected to generate a fault logical relationship. The fault logical relationship is presented in a tree shape. Then, according to the alarm data, the outermost fault level of the power fault network is a system-level fault. The system-level fault also includes unit-level faults, which include component-level faults, and component-level faults include component-level faults. The logical keywords in the crown part of the tree-like fault logical relationship correspond to system-level faults. The root position is found along the tree-like fault logical relationship. It is found that the logical keyword at the root position corresponds to a component-level fault. The system-level fault is then decomposed to obtain the component-level fault.

[0093] According to the component-level fault, the faulty component is obtained, the power frequency of the faulty component is collected, and the faulty component is diagnosed according to the power frequency and the feedback result to obtain the diagnosis result, specifically, the steps are as follows:

[0094] According to the component-level fault, the faulty component corresponding to the component-level fault is obtained, and the power frequency during the operation of the faulty component is extracted;

[0095] Compare the power frequency with the preset standard power frequency to obtain a frequency difference, and obtain the cause of the frequency difference according to the frequency difference;

[0096] According to the difference reason, the target component with abnormality in the faulty component is obtained, and the abnormal state of the target component is obtained by combining the target component and the feedback result;

[0097] Perform fault diagnosis on the abnormal state and obtain the diagnosis result of the fault in the power communication network.

[0098] In application, taking the electric power communication network in a certain area as an example, the faulty component corresponding to the component-level fault can be obtained according to the component-level fault, and the power frequency F of the faulty component during operation is extracted. The power frequency is compared with the standard power frequency V to obtain the frequency difference F / V. According to F / V and the feedback result, the cause of the frequency difference is that a certain component in the component has a virtual connection, resulting in an unstable connection. The component is then determined as the target component, and the virtual connection is regarded as the abnormal state of the target component. The virtual connection is diagnosed for the fault and the diagnosis result of the electric power communication network is that the component in the faulty component has a virtual connection.

[0099] An embodiment of the present invention provides a multi-level feedback power communication network fault diagnosis system. The system is based on any one of the multi-level feedback power communication network fault diagnosis methods described above. The system includes:

[0100] Data initialization module 1: used to collect alarm data from the power communication network, extract the alarm data, and obtain the power transformer information in the alarm data;

[0101] Model analysis module 2: used to establish a multi-level feedback model, input the electric transformer information into the multi-level feedback model, and the multi-level feedback model analyzes the electric transformer information to obtain feedback results;

[0102] Hierarchical decomposition module 3: used to extract the fault logical relationship in the feedback result according to the feedback result, and gradually decompose the fault hierarchy according to the fault logical relationship to obtain the component level fault;

[0103] Diagnostic module 4: used to obtain the faulty component according to the component-level fault, collect the power frequency of the faulty component, diagnose the faulty component according to the power frequency and the feedback result, and obtain the diagnosis result.

[0104] The above are all preferred embodiments of the present application, and the protection scope of the present application is not limited thereto. Therefore, any equivalent changes made according to the structure, shape, and principle of the present application should be included in the protection scope of the present application.

Claims

1. A multi-level feedback power communication network fault diagnosis method, characterized in that: The following steps are involved: Collecting alarm data from the power communication network, extracting data from the alarm data, and obtaining power transformer information in the alarm data, specifically: Collecting alarm data from the power communication network, performing data correlation analysis on the alarm data, and obtaining a fault correlation degree of each piece of data in the alarm data; Determine whether the fault correlation is higher than a preset correlation threshold, and extract the alarm data corresponding to the fault correlation being higher than the correlation threshold to obtain an alarm data set; Screening the alarm data set to obtain the electric transformer information in the alarm data; Establishing a multi-level feedback model, inputting the electrical transformation information into the multi-level feedback model, and the multi-level feedback model analyzing the electrical transformation information to obtain feedback results, specifically: Establishing a multi-level feedback model, the multi-level feedback model includes an identification and division layer, a low-level feedback loop layer, a high-level feedback loop layer, and a loop fusion layer; The electric transformation information is input into the multi-level feedback model, and the identification and division layer identifies the electric transformation information to obtain a voltage change diagram, a current change diagram, and an electromagnetic change diagram of the electric transformation information; Obtaining a surface characteristic value of the electric variable information based on the voltage change diagram, the current change diagram, and the electromagnetic change diagram, and comparing the surface characteristic value with a preset boundary characteristic value; If it is determined that the surface characteristic value is lower than the boundary characteristic value, the electrical transformation information is input into the low-level feedback loop layer, and a low-level feedback result is output; If it is determined that the surface characteristic value is higher than or equal to the boundary characteristic value, the electrical transformation information is input into the low-level feedback loop layer and the high-level feedback loop layer respectively, and a low-level feedback result and a high-level feedback result are output respectively; Inputting the low-level feedback result and the high-level feedback result into the loop fusion layer, fusing the low-level feedback result with the high-level feedback result to generate a feedback result; According to the feedback result, extract the fault logic relationship in the feedback result, and gradually decompose the fault level according to the fault logic relationship to obtain the component level fault; According to the component-level fault, a faulty component is obtained, the power frequency of the faulty component is collected, and the faulty component is diagnosed according to the power frequency and the feedback result to obtain a diagnosis result.

2. The multi-level feedback power communication network fault diagnosis method according to claim 1 is characterized in that: The step of inputting the electric transformation information into the low-level feedback loop layer and outputting the low-level feedback result is specifically: Inputting the electric change information into the low-level feedback loop layer, the low-level feedback loop layer generating a simulated fault and a simulated communication signal according to the voltage change graph and the current change graph; After the simulated communication signal passes through the simulated fault, a simulated voltage change graph and a simulated current change graph of the simulated communication signal are obtained; Comparing the simulated voltage change graph with the voltage change graph to obtain a voltage difference, and comparing the simulated current change graph with the current change graph to obtain a current difference; Substitute the voltage difference and the current difference into the simulated fault to generate a new simulated fault, and repeat the operation of simulating the communication signal passing through the new simulated fault and the difference comparison operation until the simulated voltage change graph and the simulated current change graph are the same as the voltage change graph and the current change graph, input the new simulated fault at this time, and output the new simulated fault as a low-level feedback result.

3. The multi-level feedback power communication network fault diagnosis method according to claim 2 is characterized in that: The steps of inputting the electric conversion information into the low-level feedback loop layer and the high-level feedback loop layer respectively, and outputting the low-level feedback result and the high-level feedback result respectively, are specifically: Inputting the electrical conversion information into the low-level feedback loop layer and the high-level feedback loop layer respectively, and the low-level feedback loop layer outputs a low-level feedback result according to the electrical conversion information; After receiving the electrical change information, the advanced feedback loop layer converts the electrical change information into magnetic change information, and extracts an electromagnetic change map based on the magnetic change information; Based on the electromagnetic change map, feature extraction is performed on the electromagnetic change map to obtain deep features of the electromagnetic change map; According to the deep features, the current change graph, the voltage change graph and the electromagnetic change graph are vertically correlated to obtain a correlation change graph; A multi-layer convolution cycle is performed on the association change graph to obtain an initial feedback result, and the initial feedback result is substituted into the association change graph for a secondary convolution cycle to obtain an advanced feedback result.

4. The multi-level feedback power communication network fault diagnosis method according to claim 1, characterized in that: According to the feedback result, extract the fault logic relationship in the feedback result, and gradually decompose the fault level according to the fault logic relationship to obtain the steps of component level fault, specifically: According to the feedback result, extract multiple logical key points in the feedback result, and logically connect the multiple logical key points to obtain the fault logical relationship in the feedback result; According to the alarm data, the outermost fault level of the electric power communication network fault is obtained, and a target logic point is generated according to the outermost fault level; Matching the target logic point with the fault logic relationship to obtain a logical position of the target logic point in the fault logic relationship; The outermost fault level is subjected to fault decomposition according to the logical position and the logical relationship to obtain a component level fault.

5. The multi-level feedback power communication network fault diagnosis method according to claim 4 is characterized in that: The steps of obtaining a faulty component according to the component-level fault, collecting the power frequency of the faulty component, diagnosing the faulty component according to the power frequency and the feedback result, and obtaining a diagnosis result are specifically: According to the component-level fault, a faulty component corresponding to the component-level fault is obtained, and the power frequency during the operation of the faulty component is extracted; Comparing the power frequency with a preset standard power frequency to obtain a frequency difference, and obtaining a difference reason for the frequency difference according to the frequency difference; According to the difference reason, the target component having an abnormality in the faulty component is obtained, and the abnormal state of the target component is obtained by combining the target component and the feedback result; A fault diagnosis is performed on the abnormal state to obtain a diagnosis result that a fault has occurred in the power communication network.

6. A multi-level feedback power communication network fault diagnosis system, the system using the multi-level feedback power communication network fault diagnosis method according to any one of claims 1 to 5, characterized in that: The system comprises: Data initialization module: used to collect alarm data from the power communication network, extract data from the alarm data, and obtain the power transformer information in the alarm data; Model analysis module: used to establish a multi-level feedback model, input the electric transformation information into the multi-level feedback model, and the multi-level feedback model analyzes the electric transformation information to obtain feedback results; A hierarchical decomposition module is used to extract the fault logical relationship in the feedback result according to the feedback result, and gradually decompose the fault hierarchy according to the fault logical relationship to obtain the component level fault; Diagnostic module: used to obtain a faulty component according to the component-level fault, collect the power frequency of the faulty component, diagnose the faulty component according to the power frequency and the feedback result, and obtain a diagnostic result.

Images