Intelligent GIS device fault location and detection method and system based on cloud-edge collaboration

Abstract

An intelligent GIS device fault location and detection method based on cloud-edge collaboration, relating to the technical field of computers. The idea of cloud-edge collaboration is utilized, and comprehensive device monitoring, fault prediction, and accurate location are realized by means of a digital twin model in combination with an abnormal change mode recognition method; the interaction relationship between the modules of a GIS device is simulated by means of the digital twin model; a large cloud interaction detection model constructed by using a graph neural network algorithm can monitor and analyze interaction data of the GIS device in real time, and can improve the precision and recall of interaction anomaly detection; and an intelligence algorithm and automatic repair technology are utilized, so that device faults can be automatically repaired when fault location is carried out, thereby improving the reliability and stability of the device.

Description

A method and system for fault location detection of intelligent GIS equipment based on cloud-edge collaboration Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a method and system for locating and detecting faults of intelligent GIS equipment based on cloud-edge collaboration. Background Art

[0002] Traditional GIS equipment troubleshooting typically relies on manual inspection and judgment to locate and address faults. However, this approach has several limitations and drawbacks. First, due to the complex and diverse structure and functions of GIS equipment, traditional methods may not accurately identify hidden fault signals, resulting in failures not being discovered and addressed in a timely manner. Second, traditional methods often require significant manpower and time, resulting in inefficient troubleshooting and potentially long downtime and repair delays. Finally, existing technologies lack comprehensive equipment monitoring and fault prediction capabilities, making it impossible to detect potential failure signs in advance and, consequently, preventing preventive maintenance measures.

[0003] Summary of the Invention

[0004] The purpose of this section is to summarize some aspects of the embodiments of the present invention and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section and the abstract and title of this application to avoid obscuring the purpose of this section, the abstract and the title of the invention, and such simplifications or omissions should not be used to limit the scope of the present invention.

[0005] In view of the limitations and defects in the existing GIS equipment fault handling, the present invention is proposed.

[0006] Therefore, the problem to be solved by the present invention is how to solve the limitations of traditional methods in GIS equipment fault location and processing by applying digital twin models, cloud-edge collaboration ideas and abnormal change pattern recognition methods, so as to improve the ability of equipment monitoring, fault prediction and precise positioning.

[0007] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0008] In the first aspect, an embodiment of the present invention provides a method for fault location detection of intelligent GIS equipment based on cloud-edge collaboration, which includes establishing a digital twin model containing all modules of the GIS equipment and their interaction relationships in the cloud, generating a large amount of fault simulation data by simulating the changes in interaction relationships under different fault conditions; extracting the interaction data between components, building a cloud-based large interaction detection model based on graph neural networks, and optimizing the model hyperparameters, and deploying the optimized large interaction detection model to the cloud platform; deploying sensors locally on the GIS equipment, collecting component interaction data, and building an edge small interaction location model to analyze the interaction data; when the edge small interaction location model detects an interaction anomaly, requesting the cloud The large-scale interactive detection model performs a global analysis, and the cloud-based large-scale interactive detection model comprehensively determines the fault source location and cause, and feeds back the results to the edge-based small-scale interactive positioning model; the edge-based small-scale interactive positioning model assists in making detailed local judgments based on the feedback to achieve accurate fault positioning, and sends the positioning results to the cloud-based large-scale interactive detection model, requesting correlation analysis; the cloud-based large-scale interactive detection model gives the fault type and root cause based on the positioning results, and the edge-based small-scale interactive positioning model verifies the root cause to complete the overall fault positioning; based on the fault results output by multiple models, the repair knowledge base is queried to determine the optimization plan, and control instructions are issued to carry out the automated maintenance process. The system performs a secondary judgment to ensure that the fault is resolved.

[0009] As a preferred solution of the intelligent GIS equipment fault location and detection method based on cloud-edge collaboration described in the present invention, requesting the cloud-side large interaction detection model to perform a global analysis includes the following steps: the edge small interaction location model continuously detects the interaction data between components. When the interaction data meets the judgment criteria, the edge small interaction location model determines that an interaction anomaly occurs and obtains a snapshot of the interaction data at the time the anomaly occurs; the edge small interaction location model packages the interaction data snapshot and requests the cloud-side large interaction detection model for a global analysis through the network; the cloud-side large interaction detection model receives the data, restores the fault scenario in a digital twin environment, and simulates the propagation and impact of the interaction data to locate the fault source; the cloud-side large interaction detection model comprehensively determines the fault source location, and conducts experiments and simulations on the fault source location to determine the possible cause of the fault; the cloud-side large interaction detection model transmits the inferred fault source location and fault cause feedback to the edge small interaction location model.

[0010] As a preferred solution of the intelligent GIS equipment fault location detection method based on cloud-edge collaboration described in the present invention, the following situations are included in the judgment of interaction anomaly: the judgment criteria include the interaction data deviation exceeding μ+3σ, the interaction response time exceeding 2T e, the interaction mode changes significantly, the interaction link is interrupted or the statistical characteristics of the interaction data change abnormally; the interaction data deviation abnormality is determined by calculating the average value μ and standard deviation σ of the normal data to determine the interaction data deviation threshold as μ+3σ. If the real-time data exceeds the interaction data deviation threshold μ+3σ, it is determined to be an interaction abnormality; the interaction response time abnormality is determined by calculating the average response time T e OK, if the interactive response time exceeds 2T e , it is determined to be an interaction anomaly; if the interaction pattern changes significantly, it is judged by using time series analysis. By analyzing the trend, periodicity and seasonal characteristics of the sequence, a normal interaction pattern is established, and the real-time interaction data is matched with the normal interaction pattern. If there is a mismatch, it is determined to be an interaction anomaly; if the interaction link is interrupted, it is judged by setting a heartbeat detection mechanism for key interaction signals. If the heartbeat times out, it is determined to be an interaction anomaly; if the statistical characteristics of the interaction data change abnormally, it is judged by building a statistical characteristic model to detect whether the real-time data deviates significantly from the established statistical distribution. If there is a significant deviation, it is determined to be an interaction anomaly.

[0011] As a preferred solution of the intelligent GIS equipment fault location detection method based on cloud-edge collaboration described in the present invention, the edge small interaction location model assists in making local and detailed judgments based on feedback to achieve accurate fault location, including the following steps: the edge small interaction location model obtains the fault source location information fed back by the cloud, and focuses on monitoring this fault location, while collecting more real-time interaction data; determines the interaction signal most closely related to the fault through correlation analysis, and constructs a time series model of the most strongly correlated interaction signal to find subtle abnormal change patterns; tracks the signal source to determine the signal source, and filters it in combination with the location information fed back by the cloud, excludes impossible components, and performs accurate fault location; and transmits the positioning results and interaction signal data set feedback to the cloud The large interactive detection model requests the cloud-based large interactive detection model to perform correlation analysis based on global data to determine the fault type and root cause; constructing a time series model of the most correlated interactive signal to find subtle abnormal change patterns includes the following steps: obtaining the interactive signal dataset of the fault source location fed back by the cloud; for each interactive signal, calculating the correlation coefficient between each interactive signal and the fault using the Pearson correlation coefficient method; sorting the correlation coefficients and selecting the top 5 interactive signals from the sorted list as the signals with the strongest correlation with the fault; establishing an ARIMA model based on the most correlated interactive signal, and using the ARIMA model to predict the relevant signals, while adjusting the model according to the prediction effect of the ARIMA model; using the model's prediction error e t Make judgments to detect subtle abnormal change patterns, and analyze the detected abnormal change patterns to obtain more information about the occurrence of the fault.

[0012] As a preferred solution of the intelligent GIS equipment fault location detection method based on cloud-edge collaboration described in the present invention, the correlation coefficient ρ XY The specific formula is as follows:

[0013] Among them, ρ XY represents the correlation coefficient, σ X and σ Y represents the standard deviation of variables X and Y, cov represents the covariance of variables X and Y; ρ XY The value range is between -1 and 1, ρ XY The closer the absolute value of ρ is to 1, the stronger the linear relationship between the two variables is. XY The closer the absolute value is to 0, the weaker the linear relationship between the two variables; when ρ XY When ρ > 0, it indicates that the two variables are positively correlated, that is, when one variable increases, the other variable also increases; when ρ XY When ρ<0, it indicates that the two variables are negatively correlated, that is, the larger the value of one variable is, the smaller the value of the other variable will be. XY =0, indicating that the two variables are nonlinearly correlated, but there are other forms of correlation; when ρ XY =1 and ρ XY = -1, it means that the sample points of the two variables X and Y fall on a straight line.

[0014] As a preferred solution of the intelligent GIS equipment fault location detection method based on cloud-edge collaboration described in the present invention, the ARIMA model includes a lagged difference test model, an intercept test model, and a trend test model. The trend test model eliminates serial correlation by adding a time trend. The specific formula is as follows:

[0015] Among them, t represents the time variable, X i represents any time series, ε t represents white noise, α represents a constant term, and βt represents a trend term.

[0016] The actual test starts from the trend test model. If the trend test model is a stationary sequence, the test stops, otherwise the intercept test model is continued to be tested; if the intercept test model is a stationary sequence, the test stops, otherwise the lagged difference test model is continued to be tested to see if it is stationary; in the test of the three models, as long as the test result of one model rejects the null hypothesis, the time series is considered to be stationary; if the test results of the three models cannot reject the null hypothesis, the time series is considered to be non-stationary.

[0017] As a preferred solution of the intelligent GIS equipment fault location detection method based on cloud-edge collaboration described in the present invention, the secondary judgment includes the following steps: querying the repair knowledge base according to the fault result output by the model to match the standard repair plan corresponding to the positioning result; simulating according to the real-time status data and the digital twin model to determine the feasibility of the standard plan, and performing personalized optimization and adjustment to form the optimal repair plan; decomposing the optimized repair plan into sequential control instructions, and issuing the control instructions to the edge controller for execution; after receiving the instructions, the edge controller checks the rationality and safety of the instructions to avoid generating new faults; the edge controller executes the maintenance control instructions to complete the automated maintenance process; after the maintenance is completed, a new round of status detection and diagnosis is performed to confirm whether the fault is resolved; if the fault is not resolved, re-positioning analysis is performed until it is confirmed that the problem is resolved; recording the maintenance event and process, and feeding back to the knowledge base for continuous iterative optimization.

[0018] In the second aspect, an embodiment of the present invention provides an intelligent GIS equipment fault location and detection system based on cloud-edge collaboration, which includes a digital twin construction module for establishing a digital twin model on the cloud that includes all modules of the GIS equipment and their interaction relationships, and generating a large amount of fault simulation data by simulating the changes in interaction relationships under different fault conditions; an interaction anomaly detection module for requesting a large interaction detection model on the cloud to perform a global analysis when an interaction anomaly is detected by a small interaction location model on the edge, and the large interaction detection model on the cloud comprehensively determines the location and cause of the fault source, and feeds the results back to the small interaction location model on the edge; an overall fault location module for giving the fault type and root cause based on the location result, and the small interaction location model on the edge verifies the root cause to complete the overall fault location; a secondary judgment module for performing a new round of status detection and diagnosis after the maintenance is completed to confirm whether the fault is resolved, and if the fault is not resolved, re-perform the location analysis until it is confirmed that the problem is resolved.

[0019] In a third aspect, an embodiment of the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program instructions are executed by the processor, the steps of the intelligent GIS equipment fault location and detection method based on cloud-edge collaboration as described in the first aspect of the present invention are implemented.

[0020] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium on which a computer program is stored, wherein: when the computer program instructions are executed by a processor, the steps of the intelligent GIS equipment fault location and detection method based on cloud-edge collaboration as described in the first aspect of the present invention are implemented.

[0021] The beneficial effects of the present invention are as follows: the present invention utilizes the idea of ​​cloud-edge collaboration, and realizes comprehensive equipment monitoring, fault prediction and precise positioning through the combination of digital twin model and abnormal change pattern recognition method; the digital twin model is used to simulate the interaction relationship between the modules of GIS equipment. The application of this model in the field of GIS equipment fault positioning is the first time, which provides new ideas and methods for solving complex equipment failure problems; the cloud-based large interaction detection model constructed by the graph neural network algorithm can monitor and analyze the interaction data of GIS equipment in real time, and improve the accuracy and recall rate of interaction anomaly detection; the use of intelligent algorithms and automatic repair technology can automatically repair equipment faults after fault positioning, so as to improve the reliability and stability of the equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] To more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. Those skilled in the art can also derive other drawings based on these drawings without inventive effort. Among them:

[0023] Figure 1 is a flow chart of the method for fault location and detection of intelligent GIS equipment based on cloud-edge collaboration.

[0024] Figure 2 is a computer equipment diagram of the intelligent GIS equipment fault location and detection method based on cloud-edge collaboration. DETAILED DESCRIPTION

[0025] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0026] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0027] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0028] Example 1

[0029] 1 and 2 , which are the first embodiment of the present invention, provide a method for locating and detecting faults of intelligent GIS equipment based on cloud-edge collaboration, including:

[0030] S1: Establish a digital twin model in the cloud that includes all modules of the GIS equipment and their interaction relationships, and generate a large amount of fault simulation data by simulating the changes in interaction relationships under different fault conditions.

[0031] Specifically, collect design drawings of GIS equipment including mechanical structure diagrams, circuit schematics, control logic diagrams, etc.; use 3D modeling software to build a mechanical model of the equipment based on the structure diagram, including coordinate system, kinematic pairs, structural constraints, etc.; use electrical simulation software to build an electrical model based on the circuit schematic, and set the parameters of electrical components such as power supply, sensors, and actuators; use relevant software to build a control model based on the control logic, write the control program, set the control loop, logical judgment, etc.; combine the above sub-models to build a digital twin model, and debug the normal working status of the system.

[0032] S2: Extract the interaction data between components, build a cloud-based large-scale interaction detection model based on graph neural networks, optimize the model hyperparameters, and deploy the optimized large-scale interaction detection model to the cloud platform.

[0033] Specifically, from the normal and faulty samples collected by the digital twins, the interaction features between components are extracted, including time series data such as voltage, current, temperature, and vibration; a connection topology diagram between components is constructed, and the direction of information flow and energy flow is marked; a graph neural network algorithm is used to build a model with components as nodes and interactions as edges; hyperparameters such as the number of network layers and node types are adjusted, and a network structure that can handle time series relationships is designed; a graph form of the sample set is constructed, and the interaction features are configured as node vector inputs; model training is performed to minimize the classification loss function to obtain a detection model for interaction anomalies; the model's classification accuracy, recall rate and other indicators are evaluated on the test set; the optimal hyperparameter configuration is obtained by repeatedly adjusting parameters and comparing different structures; the optimized large interaction detection model is deployed on the cloud server and encapsulated into a prediction API interface; the cloud-based large interaction detection model receives the interaction data uploaded in real time by the edge small interaction positioning model to predict interaction anomalies.

[0034] S3: Deploy sensors locally on GIS devices to collect component interaction data and build an edge small interaction positioning model to analyze the interaction data.

[0035] Preferably, determine the locations of key components based on the GIS equipment structure diagram and select appropriate sensors, such as current, voltage, and vibration sensors; formulate a sensor distribution plan, deploy various sensors to corresponding locations, and set sensor range and accuracy parameters; write a driver to collect real-time monitoring data from each sensor.

[0036] Furthermore, an edge small interaction positioning model is built at the edge of the device to realize the aggregation and analysis of collected data; a topological connection structure between components is constructed, and a connection model of information flow and control flow is set; an interaction monitoring algorithm is designed to analyze the real-time interaction data between components; when the interaction data exceeds the judgment standard, it is judged as abnormal and the positioning analysis program is started; the topological structure and real-time data are integrated to perform fault isolation and positioning reasoning; the faulty component is accurately judged and the fault positioning result is output.

[0037] S4: When the edge small interaction positioning model detects an interaction anomaly, it requests the cloud-based large interaction detection model to perform a global analysis. The cloud-based large interaction detection model comprehensively determines the fault source location and cause, and feeds the results back to the edge small interaction positioning model.

[0038] Specifically, the following steps are included:

[0039] S4.1: The edge small interaction positioning model continuously detects the interaction data between components. When the interaction data meets the judgment criteria, the edge small interaction positioning model determines that an interaction anomaly has occurred and obtains a snapshot of the interaction data at the time the anomaly occurs.

[0040] Specifically, the judgment criteria include interaction data deviation exceeding μ+3σ and interaction response time exceeding 2T. e , the interaction mode changes significantly, the interaction link is interrupted, or the statistical characteristics of the interaction data change abnormally.

[0041] Specifically, the interaction data deviation anomaly is determined by calculating the average value μ and standard deviation σ of normal data to determine the interaction data deviation threshold as μ+3σ. If the real-time data exceeds the interaction data deviation threshold μ+3σ, it is determined to be interaction anomaly. The interaction response time anomaly is determined by calculating the average response time T e OK, if the interactive response time exceeds 2T e , it is determined to be an interaction anomaly, where the average response time T e is the average value of historical response time; significant changes in the interaction pattern are judged by time series analysis. By analyzing the trend, periodicity, and seasonality of the sequence, a normal interaction pattern is established, and the real-time interaction data is matched with the normal interaction pattern. If there is a mismatch, it is judged as an interaction anomaly; interruptions in the interaction link are judged by setting a heartbeat detection mechanism for key interaction signals. If the heartbeat times out, it is judged as an interaction anomaly; abnormal changes in the statistical characteristics of the interaction data are judged by building a statistical characteristic model to detect whether the real-time data deviates significantly from the established statistical distribution. If there is a significant deviation, it is judged as an interaction anomaly.

[0042] S4.2: Package the interaction data snapshots and request the cloud-based large interaction detection model through the network for global analysis.

[0043] S4.3: The cloud-based large-scale interactive detection model receives the data, restores the fault scenario in the digital twin environment, and simulates the propagation and impact of the interactive data to locate the source of the fault.

[0044] S4.4: The cloud-based large-scale interactive detection model comprehensively determines the fault source location and conducts experiments and simulations on the fault source location to determine the possible cause of the fault.

[0045] Preferably, after the fault source is located, experiments and simulations are conducted on the fault source; the problems and goals to be solved are determined, and the fault locations and interaction processes that need to be experimented and simulated are clearly defined; a suitable experimental plan is designed based on the characteristics of the fault location and interaction process; specific working environments, input conditions or other parameter settings are simulated according to the requirements of the fault location and interaction process, and experimental data, phenomena and changes are recorded and observed; sensors, recorders and other equipment are used to collect data generated during the experiment, and the collected experimental data are analyzed and processed to discover patterns, trends and correlations; possible causes of the fault or abnormal behavior are found by observing and analyzing the data; and possible causes of the fault are verified and confirmed based on the experimental results.

[0046] S4.5: Transmit the inferred fault source location and fault cause feedback to the edge small interaction positioning model.

[0047] S5: The edge small-interaction positioning model assists in making detailed local judgments based on the feedback to achieve accurate fault location, and sends the positioning results to the cloud-based large-interaction detection model for correlation analysis.

[0048] Specifically, the following steps are included:

[0049] S5.1: The edge small interaction positioning model obtains the fault source location information fed back by the cloud, and focuses on monitoring this fault location while collecting more real-time interaction data.

[0050] S5.2: Determine the interaction signal that is most strongly correlated with the fault through correlation analysis, and construct a time series model of the most strongly correlated interaction signal to find subtle abnormal change patterns.

[0051] Specifically, the following steps are included:

[0052] S5.2.1: Obtain the interaction signal dataset of the fault source location fed back by the cloud.

[0053] S5.2.2: For each interaction signal, calculate the correlation coefficient between each interaction signal and the fault using the Pearson correlation coefficient method.

[0054] Specifically, the correlation coefficient ρ XY The specific formula is as follows:

[0055] Among them, ρ XY represents the correlation coefficient, σ X and σ Y represents the standard deviation of variables X and Y, and cov represents the covariance of variables X and Y.

[0056] It should be noted that ρ XY The value range is between -1 and 1, ρ XY The closer the absolute value of ρ is to 1, the stronger the linear relationship between the two variables is; XY The closer the absolute value is to 0, the weaker the linear relationship between the two variables; when ρ XY When ρ > 0, it indicates that there is a positive correlation between the two variables, that is, when one variable increases, the other variable also increases; when ρ XY When ρ<0, it indicates that there is a negative correlation between the two variables, that is, the larger the value of one variable is, the smaller the value of the other variable will be. XY =0, indicating that the two variables are nonlinearly correlated, but there may be other forms of correlation; when ρ XY =1 and ρ XY = -1, which means that the two X and Y can be well described by the straight line equation, and all sample points fall well on a straight line.

[0057] S5.2.3: Sort the correlation coefficients and select the top five interaction signals from the sorted list as the signals most strongly correlated with the fault.

[0058] Specifically, obtain all the correlation coefficients calculated in S5.2.2, and record each interaction signal and its corresponding correlation coefficient in a list; sort the interaction signals in the list from high to low according to the size of the correlation coefficient, and the larger the correlation coefficient, the stronger the correlation between the signal and the fault; select the first 5 interaction signals from the sorted list as the signals with the strongest correlation with the fault; for the selected interaction signals, conduct further analysis and research to determine their relationship with the fault.

[0059] S5.2.4: Build an ARIMA model based on the most strongly correlated interaction signals, use the ARIMA model to predict the correlated signals, and adjust the model based on the ARIMA model's prediction performance.

[0060] Specifically, collect the time series data with the strongest correlation signal, use ADF to test the stationarity of the time series; if the time series is not stationary, perform difference processing to obtain a stationary series, and perform a white noise test on the stationary series; if the stationary series is white noise, determine the p and q orders of the ARIMA model; use ACK (autocorrelation diagram) and PAC (partial correlation diagram) to preliminarily determine the values ​​of p and q; establish an ARIMA (p, d, q) model, and estimate the model parameters using the maximum likelihood method; use the AIC and BIC criteria to determine the optimal model, adjust the parameters of p and q; generate the model prediction value Y~ t , and calculate the prediction error e t ; If the prediction error e t If it does not conform to white noise, adjust the modeling parameters and rebuild the ARIMA model.

[0061] It should be noted that the ADF test of the stationarity of the time series is achieved through three models, among which the lagged difference test model plus the lag period eliminates serial correlation. The specific formula is as follows:

[0062] Among them, t represents the time variable, X i represents any time series, ε t Represents white noise.

[0063] The intercept test model adds an intercept term to eliminate serial correlation. The specific formula is as follows:

[0064] Among them, t represents the time variable, X i represents any time series, ε t represents white noise, and α represents a constant term.

[0065] The trend test model plus time trend eliminates serial correlation. The specific formula is as follows:

[0066] Among them, t represents the time variable, X i represents any time series, ε t represents white noise, α represents a constant term, and βt represents a trend term.

[0067] Furthermore, the actual test starts from the trend test model. If the trend test model is a stationary sequence, the test stops, otherwise, the intercept test model is continued to be tested; if the intercept test model is a stationary sequence, the test stops, otherwise, the lagged difference test model is continued to be tested to see if it is stationary; various model tests also have corresponding critical value tables; in the test of the three models, as long as the test result of one model rejects the null hypothesis, the time series is considered to be stationary; when the test results of the three models cannot reject the null hypothesis, the time series is considered to be non-stationary.

[0068] It should be noted that the null hypothesis is H0: δ = 0, that is, there is a unit root.

[0069] S5.2.5: Use the model's prediction error e t Make judgments to detect subtle abnormal change patterns, and analyze the detected abnormal change patterns to obtain more information about the occurrence of the fault.

[0070] Specifically, the prediction error e between the model prediction value and the actual observation value is calculated t ; For the prediction error e t Perform white noise test; if the prediction error e t If it is consistent with white noise, it means that the model fitting effect is good, otherwise it means that there are abnormal changes; when the prediction error e t When it does not conform to white noise, abnormal change pattern recognition is performed, and the residual time series diagram is observed to determine the location, time and duration of the abnormal change. At the same time, the deviation degree of the residual is calculated to determine the degree of abnormality; the identified abnormal change pattern is analyzed to determine whether it represents a fault signal of the system.

[0071] S5.3: Trace the signal to determine its source, and filter it based on the location information fed back by the cloud to eliminate impossible components and accurately locate the fault.

[0072] Specifically, based on the detected fault signal and sensor layout, a preliminary judgment is made on the possible source direction of the abnormal signal; based on the amplitude and frequency characteristics of the abnormal signal fluctuation, and with reference to the working characteristics of different components, the possible sources of the abnormal signal are further narrowed down; temporary sensors are installed at specific locations to collect more relevant signal data, and the performance of the abnormal signal on these sensors is observed to determine the precise source of the abnormal signal; the determined fault signal source is matched with the historical fault location records in the cloud platform database to screen out the matching fault location; the system feedback results of the location fault possibility analysis based on expert experience are combined, and the impossible fault locations judged by the system are excluded; the consistency of the results of multiple models is compared and integrated to determine the final fault location; the final fault location is inspected on-site, and the actual fault situation is checked to verify the correctness of the fault location; the relationship between the fault location and the fault signal source is marked and fed back to the cloud platform, and the database is continuously improved to improve the efficiency of subsequent positioning.

[0073] S5.4: Feedback the positioning results and the interactive signal data set to the cloud-based large interactive detection model, and request the cloud-based large interactive detection model to perform correlation analysis based on the global data to determine the fault type and root cause.

[0074] S6: The cloud-based large-interaction detection model gives the fault type and root cause based on the positioning results, and the edge-based small-interaction positioning model verifies the root cause to complete the overall fault positioning.

[0075] Specifically, the following steps are included:

[0076] S6.1: The cloud-based large-interaction detection model receives the precise fault location data fed back by the edge-based small-interaction positioning model, and constructs the associated topological subgraph of this location on the digital twin platform.

[0077] S6.2: By simulating the propagation effects under different fault conditions, comprehensively analyze and determine the most likely fault mode, and determine the matching fault type based on the fault mode.

[0078] Specifically, different types of faults are injected into the subgraph to observe the fault propagation effect; the propagation effect caused by different fault modes under given fault location conditions is analyzed; the fault mode that is most likely to match the actual situation is determined, and the matching fault type is determined based on the fault mode database.

[0079] S6.3: Use complex system analysis techniques to determine the most likely fault chain, analyze the fault chain, derive the root cause of the fault, and feed the inferred fault type and root cause back to the edge small interaction location model.

[0080] Preferably, the key position of the fault site in the topology map is analyzed to construct a possible fault chain; the most likely fault chain is determined through fault propagation simulation, and degradation analysis is performed step by step along the fault chain to infer the most likely root cause of the fault.

[0081] S6.4: After receiving the reason, the edge small interaction positioning model checks whether the real-time monitoring data supports the reason.

[0082] Specifically, if it is found that the real-time data does not match the cause, the problem that needs to be analyzed will be re-proposed and the cloud-edge collaborative analysis will be re-performed; if the data verification passes, the given fault location result will be confirmed to complete the overall fault accurate location.

[0083] S7: Based on the fault results output by multiple models, the repair knowledge base is queried to determine the optimization plan, and control instructions are issued to automate the repair process. The system performs a secondary judgment to ensure that the fault is resolved.

[0084] Furthermore, the repair knowledge base is queried based on the fault results output by the model to match the standard repair plan corresponding to the positioning result. The knowledge base is a summary of repair strategies for different types of faults based on historical cases; simulation is performed based on real-time status data and digital twin models to determine the feasibility of the standard plan, and personalized optimization and adjustment are performed to form the optimal repair plan; the optimized repair plan is decomposed into sequential control instructions and sent to the edge controller for execution. The control instructions include switching, restarting, and reconfiguring device modules; after receiving the instructions, the edge controller checks the rationality and safety of the instructions to avoid new faults; the edge controller executes the maintenance control instructions to complete the automated maintenance process.

[0085] Furthermore, after the maintenance is completed, a new round of status detection and diagnosis is carried out to confirm whether the fault has been resolved; if the fault is not resolved, the positioning analysis is re-performed until it is confirmed that the problem has been resolved; the maintenance event and process are recorded and fed back to the knowledge base for continuous iterative optimization.

[0086] Furthermore, this embodiment also provides an intelligent GIS equipment fault location and detection system based on cloud-edge collaboration, including a digital twin construction module, which is used to establish a digital twin model on the cloud that includes all modules of the GIS equipment and their interaction relationships, and generate a large amount of fault simulation data by simulating the changes in interaction relationships under different fault conditions; an interaction anomaly detection module, which is used to request the cloud-based large interaction detection model to perform a global analysis when the edge small interaction location model detects an interaction anomaly, and the cloud-based large interaction detection model comprehensively determines the fault source location and fault cause, and feeds the results back to the edge small interaction location model; an overall fault location module, which is used to give the fault type and root cause based on the location result, and the edge small interaction location model verifies the root cause to complete the overall fault location; a secondary judgment module, which is used to perform a new round of status detection and diagnosis after the maintenance is completed to confirm whether the fault has been resolved, and if the fault has not been resolved, re-perform the location analysis until it is confirmed that the problem has been resolved.

[0087] This embodiment also provides a computer device suitable for the intelligent GIS device fault location detection method, including a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute computer-executable instructions to implement the intelligent GIS device fault location detection method proposed in the above embodiment.

[0088] The computer device may be a terminal, comprising a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device comprises a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner may be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display or an electronic ink display screen, and the input device of the computer device may be a touch layer covering the display screen, or a button, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse.

[0089] This embodiment further provides a storage medium storing a computer program, which, when executed by a processor, implements the method for locating and detecting faults of intelligent GIS equipment as proposed in the above embodiment.

[0090] In summary, the present invention utilizes the idea of ​​cloud-edge collaboration, and realizes comprehensive equipment monitoring, fault prediction and precise positioning by combining the digital twin model and the abnormal change pattern recognition method; the digital twin model is used to simulate the interaction relationship between the various modules of the GIS equipment. The application of this model in the field of GIS equipment fault positioning is the first time, which provides new ideas and methods for solving complex equipment failure problems; the cloud-based large-scale interaction detection model constructed using the graph neural network algorithm can monitor and analyze the interaction data of GIS equipment in real time, and improve the accuracy and recall rate of interaction anomaly detection; using intelligent algorithms and automated repair technology, it can automatically repair equipment faults after fault positioning to improve the reliability and stability of the equipment.

[0091] Example 2

[0092] Referring to Figures 1 and 2, the second embodiment of the present invention provides a method for fault location and detection of intelligent GIS equipment based on cloud-edge collaboration. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and simulation experiments.

[0093] Specifically, GIS design drawings were collected and their simulation models were built in the digital twin. A large amount of fault simulation data was generated by simulating the changes in the interaction relationships under different fault conditions. 100 sets of fault data and normal data were collected, and some of the data are shown in Tables 1 and 2.

[0094] Table 1 Normal data of some GIS equipment

[0095] Table 2 Part of GIS equipment failure data

[0096] Furthermore, using the data in Tables 1 and 2, we extracted sensor outputs and actuator inputs as features and constructed a cloud-based large-scale interaction detection model based on a graph neural network. Sensors were placed at the device edge to collect normal operating data in real time, and a small-scale interaction localization model was constructed. When sensor output exceeded a threshold, an abnormality was identified.

[0097] Furthermore, the edge small-interaction positioning model detected an abnormal output of 11.8V from sensor A and sent the data to the cloud platform. The cloud-based large-interaction detection model analyzed the data within the digital twin and determined that the fault source was sensor A. The edge small-interaction positioning model collected 100 sets of sensor A data and found that the output fluctuated between 11 and 13V. The ADF distribution was used to test the stationarity of the time series. The ADF distribution critical value table was used to determine that the time series was stable and that the stationary series was white noise. Therefore, an ARIMA model was constructed, detecting that the error did not conform to white noise. The abnormal change pattern was then identified, confirming that sensor A was faulty. The ADF distribution critical value table is shown in Table 3.

[0098] Table 3 ADF distribution critical value table

[0099] Furthermore, the cloud-based large-scale interactive detection model receives the positioning results, simulates fault propagation in the digital twin, and determines that the fault type is sensor A deviation drift; based on the deviation drift, it queries the repair knowledge base and matches the standard repair plan corresponding to the deviation drift. At the same time, it simulates based on the real-time status data and the digital twin model to verify the feasibility of the standard plan, and performs personalized optimization and adjustment to form the optimal repair plan; the optimal repair plan is decomposed into sequential control instructions, which are issued to adjust the calibration parameters of sensor A. After the maintenance is completed, a second confirmation is performed to confirm that the fault has been resolved.

[0100] Furthermore, a comparison table of the detection and positioning of the present invention and traditional GIS equipment is shown in Table 4.

[0101] Table 4 Comparison of detection and positioning between the present invention and traditional GIS equipment

[0102] Preferably, the results in Table 4 show that the present invention has significant advantages in multiple aspects such as fault detection accuracy, positioning accuracy, response time, repair time and maintenance cost. The present invention adopts a detection and positioning concept based on cloud-edge collaboration, which improves the analysis and judgment capabilities through digital twins, multi-source data fusion and other methods. Therefore, the fault detection accuracy and positioning accuracy are improved by 11% and 10% respectively compared with traditional methods; the response time and repair time of the present invention are more advantageous than traditional methods. These advantages greatly shorten the downtime of faulty equipment and significantly improve system availability; from the perspective of maintenance costs, the present invention can reduce 23%, while traditional methods have no significant effect.

[0103] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for locating and detecting faults of intelligent GIS equipment based on cloud-edge collaboration, characterized in that: include, Establish a digital twin model in the cloud that includes all modules of GIS equipment and their interactions, and generate a large amount of fault simulation data by simulating the changes in the interactions under different fault conditions; Extract the interaction data between components, build a cloud-based large-scale interaction detection model based on graph neural networks, optimize the model hyperparameters, and deploy the optimized large-scale interaction detection model to the cloud platform; Deploy sensors on local GIS devices to collect component interaction data, and build edge small interaction positioning models to analyze interaction data; When the edge small interaction positioning model detects an interaction anomaly, it requests the cloud-based large interaction detection model to perform a global analysis. The cloud-based large interaction detection model comprehensively determines the fault source location and fault cause, and feeds the results back to the edge small interaction positioning model. The edge small interaction positioning model assists in making detailed local judgments based on the feedback to achieve accurate fault positioning, and sends the positioning results to the cloud large interaction detection model to request correlation analysis; The cloud-based large-interaction detection model gives the fault type and root cause based on the location results, and the edge-based small-interaction location model verifies the root cause to complete the overall fault location. Based on the fault results output by multiple models, the repair knowledge base is queried to determine the optimization plan, and control instructions are issued to automate the repair process. The system performs a secondary judgment to ensure that the fault is resolved.

2. The intelligent GIS equipment fault location detection method based on cloud-edge collaboration as claimed in claim 1, characterized in that: The requesting the cloud-based large interaction detection model to perform global analysis includes the following steps: The edge small interaction positioning model continuously detects the interaction data between components. When the interaction data meets the judgment criteria, the edge small interaction positioning model determines that an interaction anomaly occurs and obtains a snapshot of the interaction data at the time the anomaly occurs. The edge small interaction location model packages the interaction data snapshots and requests the cloud-based large interaction detection model through the network for global analysis; The cloud-based large interactive detection model receives the data, restores the fault scenario in the digital twin environment, and simulates the propagation and impact of the interactive data to locate the source of the fault; The cloud-based large-scale interactive detection model comprehensively determines the fault source location, and conducts experiments and simulations on the fault source location to determine the possible cause of the fault; The cloud-based large-interaction detection model transmits the inferred fault source location and fault cause feedback to the edge small-interaction positioning model.

3. The intelligent GIS equipment fault location detection method based on cloud-edge collaboration as claimed in claim 2 is characterized by: The judgment that an interaction abnormality occurs includes the following situations: The judgment criteria include interaction data deviation exceeding μ+3σ and interaction response time exceeding 2T. e , the interaction mode changes significantly, the interaction link is interrupted, or the statistical characteristics of the interaction data change abnormally; The interaction data deviation anomaly is determined by calculating the average value μ and standard deviation σ of normal data to determine the interaction data deviation threshold as μ+3σ. If the real-time data exceeds the interaction data deviation threshold μ+3σ, it is determined to be interaction anomaly. The interaction response time anomaly is calculated by the average response time T e OK, if the interactive response time exceeds 2T e , it is judged as interaction abnormality; The significant change of interaction mode is judged by time series analysis. By analyzing the trend, periodicity and seasonality characteristics of the sequence, a normal interaction mode is established, and the real-time interaction data is matched with the normal interaction mode. If there is a mismatch, it is judged as an abnormal interaction. Interruptions in the interaction process are judged by setting up a heartbeat detection mechanism for key interaction signals. If the heartbeat times out, it is considered an interaction anomaly. Abnormal changes in the statistical characteristics of interactive data are judged by building a statistical characteristic model to detect whether the real-time data deviates significantly from the established statistical distribution. If there is a significant deviation, it is judged as an interactive anomaly.

4. The intelligent GIS equipment fault location detection method based on cloud-edge collaboration as claimed in claim 1, characterized in that: The edge small interaction positioning model assists in making local detailed judgments based on feedback to achieve accurate fault positioning, including the following steps: The edge small interaction positioning model obtains the fault source location information fed back by the cloud, and focuses on monitoring the fault location while collecting more real-time interaction data. Determine the interaction signal most strongly associated with the fault through correlation analysis, and build a time series model of the most strongly associated interaction signal to find subtle abnormal change patterns; Trace the signal source to determine the signal source, and filter it based on the location information fed back by the cloud, exclude impossible components, and accurately locate the fault; Transmit the positioning results and interaction signal data set feedback to the cloud-based large interaction detection model, and request the cloud-based large interaction detection model to perform correlation analysis based on global data to determine the fault type and root cause; The method of constructing a time series model of the strongest correlated interaction signals to find subtle abnormal change patterns includes the following steps: Obtain the interactive signal data set of the fault source location fed back by the cloud; For each interaction signal, the correlation coefficient between each interaction signal and the fault is calculated by Pearson correlation coefficient method; Sort the correlation coefficients and select the top 5 interaction signals from the sorted list as the signals most strongly correlated with the fault; Establish an ARIMA model based on the most strongly correlated interactive signal, use the ARIMA model to predict the relevant signals, and adjust the model according to the prediction effect of the ARIMA model; Using the model's prediction error e t Judgments are made to detect subtle abnormal change patterns, and the detected abnormal change patterns are analyzed to obtain more information about the occurrence of the fault.

5. The intelligent GIS equipment fault location detection method based on cloud-edge collaboration as claimed in claim 4 is characterized by: The correlation coefficient ρ XY The specific formula is as follows: Among them, ρ XY represents the correlation coefficient, σ X and σ Y represents the standard deviation of variables X and Y, and cov represents the covariance of variables X and Y; ρ XY The value range is between -1 and 1, ρ XY The closer the absolute value of is to 1, the stronger the linear relationship between the two variables is. XY The closer the absolute value is to 0, the weaker the linear relationship between the two variables; When XY When > 0, it indicates that the two variables are positively correlated, that is, when one variable increases, the other variable also increases; When XY When <0, it indicates that the two variables are negatively correlated, that is, the larger the value of one variable, the smaller the value of the other variable. When XY =0, indicating that the two variables are nonlinearly related, but there are other ways of correlation; When XY =1 and ρ XY = -1, it means that the sample points of the two variables X and Y fall on a straight line.

6. The intelligent GIS equipment fault location detection method based on cloud-edge collaboration as claimed in claim 4 is characterized by: The ARIMA model includes a lagged difference test model, an intercept test model and a trend test model. The trend test model eliminates serial correlation by adding a time trend. The specific formula is as follows: Among them, t represents the time variable, X i represents any time series, ε t represents white noise, α represents a constant term, and βt represents a trend term; The actual test starts from the trend test model. If the trend test model is a stable sequence, the test stops, otherwise the intercept test model continues to be tested; If the intercept test model is a stable series, the test stops, otherwise continue to test whether the lagged difference test model is stable; In the tests of the three models, as long as the test result of one model rejects the null hypothesis, the time series is considered to be stationary; If the test results of the three models cannot reject the null hypothesis, the time series is considered to be non-stationary.

7. The intelligent GIS equipment fault location detection method based on cloud-edge collaboration as claimed in claim 1, characterized in that: The secondary judgment comprises the following steps: Query the repair knowledge base according to the fault results output by the model to match the standard repair solution corresponding to the positioning result; Simulate based on real-time status data and digital twin models to determine the feasibility of standard solutions, and make personalized optimization and adjustments to form the best repair solution; Decompose the optimized repair plan into sequential control instructions and send the control instructions to the edge controller for execution; After receiving the instruction, the edge controller checks the rationality and safety of the instruction to avoid new faults; The edge controller executes maintenance control instructions to complete the automated maintenance process; After the repair is completed, a new round of status detection and diagnosis is carried out to confirm whether the fault has been resolved; If the fault is not resolved, re-perform location analysis until it is confirmed that the problem is resolved; Record the maintenance event and process, and provide feedback to the knowledge base for continuous iteration and optimization.

8. A smart GIS device fault location detection system based on cloud-edge collaboration, based on the smart GIS device fault location detection method based on cloud-edge collaboration according to any one of claims 1 to 7, characterized in that: Also includes, The digital twin construction module is used to build a digital twin model on the cloud that includes all modules of the GIS equipment and their interactive relationships, and to generate a large amount of fault simulation data by simulating the changes in interactive relationships under different fault conditions; The interaction anomaly detection module is used to request the cloud-based large interaction detection model to perform a global analysis when the edge small interaction positioning model detects an interaction anomaly. The cloud-based large interaction detection model comprehensively determines the fault source location and fault cause, and feeds the results back to the edge small interaction positioning model. The overall fault location module is used to give the fault type and root cause based on the location results. The edge small interaction location model verifies the root cause to complete the overall fault location; The secondary judgment module is used to perform a new round of status detection and diagnosis after the maintenance is completed to confirm whether the fault has been resolved. If the fault has not been resolved, re-positioning and analysis will be performed until it is confirmed that the problem has been resolved.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, it implements the steps of the intelligent GIS equipment fault location detection method based on cloud-edge collaboration as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the intelligent GIS equipment fault location detection method based on cloud-edge collaboration described in any one of claims 1 to 7 are implemented.

Images