A method and system for recovering electric power meter data

Abstract

The present application provides a method and system for recovering data of an electric power meter. The method performs error node perception on the response error data of the target electric power meter, thereby determining the information difference entropy between the characteristics of each error node; performs self-recovery operation on the response error data based on the structural association information of the data between each response module to obtain self-recovery data; determines the recovery granularity of the response error data based on the feature gradient difference and information difference entropy in the self-recovery data; and determines the fault-tolerant recovery cost based on the structural dependency and information difference entropy between the node backup data of each response module; further adjusts the data recovery strategy based on the feedback of the recovery granularity and the fault-tolerant recovery cost, and then performs confidence recovery on the data of the target electric power meter after the response error based on the data recovery strategy adjusted by the feedback. The scheme of the present application can realize the structural association reuse of multimodal data in the electric power meter, thereby improving the confidence of data recovery.

Description

Technical Field

[0001] The present application relates to the field of data recovery technology, and more specifically, to a method and system for recovering data of an electric power meter. Background Art

[0002] As an important tool for monitoring and managing power equipment, electric power meters often need to process multimodal data such as voltage, current, power factor, etc. during operation as the complexity of power systems continues to increase. The accuracy and integrity of this data are crucial to the stable operation of power systems. However, due to the large number of response modules inside the electric power meter and their interdependence, data errors caused by module failures or data transmission problems often occur in actual applications. If these errors are not effectively handled, they may affect the normal operation of the power system and cause safety hazards.

[0003] In the existing technology, traditional data recovery methods often rely on simple error log analysis and preset recovery strategies, but these methods cannot fully consider the complex structure of multimodal data and the dependencies between different response modules, thereby limiting the effectiveness and accuracy of data recovery. As a result, existing data recovery technologies have problems such as poor recovery effect and inflexible recovery strategy when dealing with response errors of electric power meters, and thus cannot meet the high-confidence data recovery requirements of the power system. Therefore, how to achieve structural correlation and reuse of multimodal data in electric power meters to improve the confidence of data recovery has become a difficult problem faced by the industry. Summary of the Invention

[0004] The present application provides a data recovery method and system for an electric power meter, which can realize structural correlation and multiplexing of multimodal data in the electric power meter, thereby improving the confidence of data recovery.

[0005] In a first aspect, the present application provides a method for recovering data from an electric power meter, comprising the following steps:

[0006] Receive a start-up instruction from a target electric power meter, monitor the status identifiers of each response module in the target electric power meter, filter out error logs of each response module according to all status identifiers, and obtain response error data;

[0007] Perform error node perception on the response error data based on a pre-trained error detection model to obtain error node features of each response module, and then determine the information difference entropy between the error node features;

[0008] Determine, based on all error node characteristics, structural association information of data between various response modules when a response error occurs, and then perform a self-recovery operation on the response error data based on the structural association information to obtain self-recovery data;

[0009] Determining the recovery granularity of each error log in the response error data when self-recovering after the response error by using the feature gradient difference and the information difference entropy in the self-recovery data;

[0010] Acquire backup data of the target electric power meter, extract node backup data of each response module from the backup data, and then determine the fault-tolerant recovery cost when using the backup data for multiplexing recovery after a response error based on the structural dependency relationship between the backup data of each node and the information difference entropy;

[0011] When the data fault tolerance rate of the target electric power meter is lower than the standard fault tolerance threshold, the data recovery strategy is adjusted through the recovery granularity and the fault tolerance recovery cost feedback, and then the data of the target electric power meter after the response error is confidently recovered based on the feedback-adjusted data recovery strategy.

[0012] Preferably, the error logs of each response module are filtered out according to all status identifiers, and the response error data obtained specifically includes:

[0013] Based on the pre-trained state model, an abnormal state identification set is obtained from all state identifications;

[0014] Extracting error logs from the response logs of each response module according to the abnormal status identification set;

[0015] A data group consisting of all error logs is used as response error data.

[0016] Preferably, error node perception is performed on the response error data based on a pre-trained error detection model to obtain error node features of each response module, specifically including:

[0017] Using the response error data as initialization parameters of a pre-trained error detection model;

[0018] The error detection model is used to perceive the error nodes in each response module and obtain the error node features of each response module.

[0019] Preferably, determining the information difference entropy between the features of each error node specifically includes:

[0020] Determine the information difference between each error node feature and other error node features;

[0021] All information differences are described by entropy features to obtain the information difference entropy between the features of each error node.

[0022] Preferably, determining the structural association information of data between various response modules when a response error occurs based on all error node characteristics specifically includes:

[0023] Determine the structural association diagram of data interaction between each response module;

[0024] Determine the functional dependency values between error node features through the propagation paths of all error node features in each response module;

[0025] Performing topological adjustment on the edge weights in the structure association graph based on all functional dependency values to obtain a topological structure association graph;

[0026] Structural association information of data between various response modules when a response error occurs is extracted from the topological structure association graph.

[0027] Preferably, the fault-tolerant recovery cost when the backup data is used for multiplexing and recovery after a response error is determined based on the structural dependency relationship between the backup data of each node and the information difference entropy specifically includes:

[0028] Obtain self-recovery data after executing the self-recovery operation;

[0029] Reusing the self-recovery data using the backup data to obtain multiplexed recovery data, and then determining a recovery confidence distance between the multiplexed recovery data and the self-recovery data;

[0030] Determine the structural dependencies between backup data of each node;

[0031] Determining the recovery cost of using the backup data for multiplexing and recovery after a response error by using the structural dependency and the recovery confidence distance;

[0032] The recovery cost is compensated for fault tolerance based on the information difference entropy to obtain a fault tolerance recovery cost when the backup data is used for multiplexing recovery after a response error.

[0033] Preferably, adjusting the data recovery strategy through feedback of the recovery granularity and the fault-tolerant recovery cost specifically includes:

[0034] Get the policy value of the data recovery policy;

[0035] Determine the state feedback amount of the data recovery strategy by using the recovery granularity and the fault-tolerant recovery cost;

[0036] The data recovery strategy is adjusted based on the strategy value and the state feedback amount.

[0037] In a second aspect, the present application provides an electric power meter data recovery system, comprising:

[0038] an acquisition module configured to, after receiving a start-up instruction from a target electric power meter, monitor the status identifiers of each response module in the target electric power meter, filter out error logs of each response module according to all status identifiers, and obtain response error data;

[0039] A processing module, configured to perform error node perception on the response error data based on a pre-trained error detection model, obtain error node features of each response module, and further determine information difference entropy between the error node features;

[0040] The processing module is further configured to determine, based on all error node characteristics, structural association information of data between various response modules when a response error occurs, and then perform a self-recovery operation on the response error data based on the structural association information to obtain self-recovery data;

[0041] The processing module is further configured to determine, by using the feature gradient difference and the information difference entropy in the self-recovery data, a recovery granularity of each error log in the response error data when self-recovery is performed after the response error;

[0042] The processing module is further configured to obtain backup data of the target electric power meter, extract node backup data of each response module from the backup data, and then determine the fault-tolerant recovery cost when the backup data is used for multiplexing recovery after a response error based on the structural dependency relationship between the backup data of each node and the information difference entropy;

[0043] An execution module is used to adjust the data recovery strategy through the recovery granularity and the fault tolerance recovery cost feedback when the data fault tolerance rate of the target electric power meter is lower than the standard fault tolerance threshold, and then perform confidence recovery on the data of the target electric power meter after the response error based on the feedback-adjusted data recovery strategy.

[0044] In a third aspect, the present application provides a computer device, comprising a memory and a processor, wherein the memory stores a code, and the processor is configured to obtain the code and execute the above-mentioned electric power meter data recovery method.

[0045] In a fourth aspect, the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the above-mentioned electric power meter data recovery method is implemented.

[0046] The technical solutions provided by the embodiments disclosed in this application have the following beneficial effects:

[0047] In the embodiment of the present application, a start-up instruction of a target electric power meter is received, the status identifiers of each response module in the target electric power meter are monitored, and the error logs of each response module are screened out according to all the status identifiers to obtain response error data; error node perception is performed on the response error data based on a pre-trained error detection model to obtain error node features of each response module, and then the information difference entropy between the features of each error node is determined; structural association information of data between each response module when a response error occurs is determined according to all the error node features, and then a self-recovery operation is performed on the response error data according to the structural association information to obtain self-recovery data; and a feature gradient in the self-recovery data is used to obtain the self-recovery data. The degree difference and the information difference entropy determine the recovery granularity of each error log in the response error data when self-recovery is performed after the response error; the backup data of the target electric power meter is obtained, and the node backup data of each response module is extracted from the backup data, and then the structural dependency relationship between the backup data of each node and the information difference entropy are used to determine the fault-tolerant recovery cost when the backup data is used for multiplexing recovery after the response error; when the data fault tolerance rate of the target electric power meter is lower than the standard fault tolerance threshold, the data recovery strategy is adjusted through the recovery granularity and the fault-tolerant recovery cost feedback, and then the data of the target electric power meter after the response error is confidently recovered based on the feedback-adjusted data recovery strategy.

[0048] It can be seen that the present application adjusts the data recovery strategy through feedback of recovery granularity and fault-tolerant recovery cost, and then performs confidence recovery on the data of the target electric power meter after the response error based on the data recovery strategy adjusted by feedback; first, the structural association information of the data between each response module is used to perform a self-recovery operation on the response error data to obtain self-recovery data. By identifying the structural association information of the data between the response modules, the complexity of multimodal data (such as voltage, current, power factor, etc.) and the dependency relationship between each module can be fully considered, so that data recovery is no longer limited to simple log analysis, but uses structured information to perform more effective recovery operations, thereby improving the confidence of data recovery; secondly, the recovery granularity during the self-recovery operation is determined by the feature gradient difference and information difference entropy in the self-recovery data. The recovery granularity reflects the degree of detail of the data state involved in the recovery operation of the response error data during the self-recovery process. By quantifying the recovery granularity, it can be ensured that the recovery operation can accurately repair the target error characteristics, so that the system can accurately identify and understand the structural relationship between the error data, thereby improving data recovery. The accuracy of recovery; then, the fault-tolerant recovery cost during reuse recovery is determined by the structural dependency and information difference entropy between the backup data of each node. The fault-tolerant recovery cost refers to the additional cost added during the system fault-tolerant processing based on the cost of using backup data and self-recovery data for reuse recovery. The quantified fault-tolerant recovery cost can be used to adjust the recovery strategy in time to ensure the stable operation of the electric power meter; finally, the current data recovery strategy is feedback-adjusted by the recovery granularity and the fault-tolerant recovery cost. Through this feedback adjustment mechanism, the data recovery of the electric power meter can be dynamically optimized and adjusted according to various complex situations, thereby overcoming the problem that the traditional data recovery method cannot adaptively adjust the recovery parameters and causes low data recovery confidence, so as to ensure that data can be effectively recovered in various complex situations, and then the data of the target electric power meter after the response error is confidently restored through the feedback-adjusted data recovery strategy; in summary, the present application scheme can realize the structural correlation reuse of multimodal data in the electric power meter, thereby improving the confidence of data recovery. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 is an exemplary flow chart of a method for recovering power meter data according to some embodiments of the present application;

[0050] Figure 2 is a schematic diagram of a process for determining structural association information according to some embodiments of the present application;

[0051] Figure 3 is a schematic diagram of a process for determining restoration granularity according to some embodiments of the present application;

[0052] Figure 4is a schematic diagram of exemplary software modules of an electric power meter data recovery system according to some embodiments of the present application;

[0053] Figure 5 It is a structural diagram of a computer device using a method for recovering electric power meter data according to some embodiments of the present application. DETAILED DESCRIPTION

[0054] The core of this application is to perform error node perception on the response error data of the target electric power meter, and then determine the information difference entropy between the characteristics of each error node; perform self-recovery operations on the response error data based on the structural association information of the data between each response module to obtain self-recovery data; determine the recovery granularity of the response error data through the feature gradient difference and information difference entropy in the self-recovery data; then determine the fault-tolerant recovery cost based on the structural dependency and information difference entropy between the node backup data of each response module; further adjust the data recovery strategy through feedback of the recovery granularity and the fault-tolerant recovery cost, and then perform confidence recovery on the data of the target electric power meter after the response error based on the data recovery strategy adjusted by feedback; the scheme of this application can realize the structural association reuse of multimodal data in the electric power meter, thereby improving the confidence of data recovery.

[0055] In order to better understand the above technical solution, the following will be combined with the accompanying drawings and specific implementation methods to describe the above technical solution in detail. Figure 1 , which is an exemplary flow chart of a method for recovering power meter data according to some embodiments of the present application. The power meter data recovery method 100 mainly includes the following steps:

[0056] In step 101, a start instruction of a target electric power meter is received, the status identifiers of each response module in the target electric power meter are monitored, and the error logs of each response module are screened out according to all the status identifiers to obtain response error data.

[0057] In specific implementation, receiving the start-up instruction of the target electric power meter and monitoring the status identification of each response module in the target electric power meter can be achieved in the following manner, namely: starting the target electric power meter immediately after receiving the start-up instruction of the target electric power meter, monitoring the status of each response module in the target electric power meter in real time through the abnormality monitoring unit in the target electric power meter, and marking the monitored status quantity to obtain the status identification of each response module. In this application, the status identification can be used to quickly read whether the response status of the corresponding response module is abnormal, which can improve the monitoring efficiency of the electric power meter in generating erroneous data when detecting the operating status of the power equipment; it should be noted that the response module includes: a current measurement module, a voltage measurement module, a power factor measurement module, a reactive power measurement module and a data storage module. In other embodiments, the response module may also include other power parameter measurement modules, which are not limited here.

[0058] In some embodiments, the error logs of each response module are filtered out according to all status identifiers, and the response error data is obtained by the following steps:

[0059] Based on the pre-trained state model, an abnormal state identification set is obtained from all state identifications;

[0060] Extracting error logs from the response logs of each response module according to the abnormal status identification set;

[0061] A data group consisting of all error logs is used as response error data.

[0062] It should be noted that the state model in the present application can be pre-trained using a large amount of historical response data, and the network architecture of the state model can adopt a convolutional neural network. In other embodiments, other network architectures can also be adopted, which are not limited here; in specific implementation, the abnormal state identification set obtained by dividing all state identifications based on the pre-trained state model can be implemented in the following manner, namely: all state identifications can be input into the pre-trained state model, and then the state model is used to divide each state identification into normal state identification and abnormal state identification, and the set consisting of all normal state identifications is used as the normal state identification set, and the set consisting of all abnormal state identifications is used as the abnormal state identification set; according to the abnormal state identification set, the error log is extracted from the response log of each response module, which can be implemented in the following manner, namely: for each response module, all response logs that identify the state of the response module as abnormal are used as the error log of the response module, and then the error logs of all response modules are obtained; the data group consisting of all error logs is used as the response error data, which can be implemented in the following manner, namely: all error logs can be arranged into data groups in chronological order according to the occurrence time in the timestamp, and the arranged data group is used as the response error data.

[0063] In step 102, error node perception is performed on the response error data based on a pre-trained error detection model to obtain error node features of each response module, and then determine the information difference entropy between the error node features.

[0064] In some embodiments, error node perception is performed on the response error data based on a pre-trained error detection model to obtain error node features of each response module. The following steps can be used to achieve this:

[0065] Using the response error data as initialization parameters of a pre-trained error detection model;

[0066] The error detection model is used to perceive the error nodes in each response module and obtain the error node features of each response module.

[0067] It should be noted that the error detection model in this application is a model for automatically identifying and extracting error features. The error detection model is usually pre-trained based on a large amount of historical response data based on machine learning or deep learning technology. It can analyze the input data and detect potential error features therein; it should also be noted that the error node feature in this application refers to the feature of the specific error point attribute that can be identified in the response module. These features usually include: the type of error, the frequency of occurrence and the specific location. The error node feature can help the system quickly locate the location attribute information of the error; in specific implementation, for each response module, first, all error logs corresponding to the response module in the response error data can be input into the pre-trained error detection model for error node perception, and then the perception result of the error detection model is used as the error node feature of the response module, and then the error node features of all response modules are obtained. It should also be noted that the error node feature of each response module can be one or more.

[0068] In some embodiments, determining the information difference entropy between each error node feature may be achieved by using the following steps:

[0069] Determine the information difference between each error node feature and other error node features;

[0070] All information differences are described by entropy features to obtain the information difference entropy between the features of each error node.

[0071] It should be noted that the information difference degree in this application is an indicator for measuring the degree of information difference between the features of two error nodes; the information difference entropy in this application is an indicator for measuring the degree of confusion of the information difference between the features of the error nodes, which reflects the degree of diversity of the features of the error nodes in the detection system. The larger the information difference entropy, the greater the information difference between the features of the error nodes, and the higher the diversity of the features of the error nodes in the detection system.

[0072] In specific implementation, determining the information difference between each error node feature and other error node features can be achieved in the following manner, namely: selecting an error node feature as the selected error node feature, selecting one from the other error node features as the target error node feature, using the Euclidean distance between the selected error node feature and the target error node feature as the information difference distance between the selected error node feature and the target error node feature, continuing to determine the information difference distance between the selected error node feature and the remaining target error node features, and then using the average of all information difference distances as the information difference between the selected error node feature and other error node features, repeating the above steps, continuing to determine the information difference between the remaining selected error node features and other error node features, thereby obtaining the information difference between each error node feature and other error node features; performing entropy feature description on all information differences, and obtaining the information difference entropy between each error node feature can be achieved in the following manner, namely: bringing all information differences into the information entropy calculation formula, and using the calculated information entropy as the information difference entropy between each error node feature.

[0073] In step 103, the structural association information of the data between the various response modules when the response error occurs is determined according to all the error node characteristics, and then a self-recovery operation is performed on the response error data based on the structural association information to obtain self-recovery data.

[0074] In some embodiments, reference Figure 2 As shown in the figure, this figure is a schematic diagram of the process of determining structural association information in some embodiments of the present application. In this embodiment, the structural association information of data between various response modules when a response error occurs is determined based on all error node characteristics, which can be implemented by the following steps:

[0075] In step 1031, a structural association diagram of data interaction between various response modules is determined;

[0076] In step 1032, the functional dependency values between the error node features are determined through the propagation paths of all error node features in each response module;

[0077] In step 1033, topological adjustment is performed on the edge weights in the structure association graph based on all functional dependency values to obtain a topological structure association graph;

[0078] In step 1034, structural association information of data between various response modules when a response error occurs is extracted from the topological structure association graph.

[0079] It should be noted that the structural association diagram in this application is a graphical model for representing the relationship between various response modules. The main function of the structural association diagram is to visualize and analyze the mutual relationship between various response modules, wherein the structural association diagram is usually composed of nodes and edges. In the structural association diagram, each node represents an independent response module, and the edges between nodes represent the dependency relationship between response modules. The edges have directionality (indicating the direction of data flow) and weight (indicating the strength of the dependency relationship); the functional dependency value in this application is an indicator used to quantify the data interaction intensity between the functions of each response module; the structural association information in this application refers to the comprehensive information formed between different response modules based on their mutual relationship, dependency and interaction mode. It describes the association structure between various response modules, including the dependency, interaction mode and overall topology between response modules. By understanding and utilizing structural association information, data recovery efficiency can be improved.

[0080] In a specific implementation, determining the structural association diagram of data interaction between each response module can be achieved in the following manner, namely: first, analyzing the data flow between each response module, for example, the voltage measurement module transmits data to the power factor measurement module, and the power factor measurement module then transmits the result to the data storage module, and the data flow path is used as the edge between the nodes. Then, each response module is regarded as a node, and the structural diagram composed of all nodes and the edges between the nodes is used as the structural association diagram of data interaction between the response modules; determining the functional dependency value between the error node features through the propagation path of all error node features in each response module can be achieved in the following manner, namely: first, selecting an error node feature as the selected error node feature, obtaining the propagation path of the selected error node feature in each response module, selecting an error node feature from other error node features as the target error node feature, obtaining the propagation path of the target error node feature in each response module, calculating the similarity between the propagation path of the selected error node feature and the propagation path of the target error node feature, and using the similarity as the functional dependency value between the selected error node feature and the target error node feature. Repeating the above steps to obtain the functional dependency values between all selected error node features and the target error node feature.

[0081] In addition, the edge weights in the structural association graph are topologically adjusted based on all functional dependency values, and the topological structure association graph can be obtained in the following manner, namely: two nodes are obtained from the structural association graph, and all error node features of the response modules corresponding to the two nodes are obtained, and then the functional dependency values between all error node features are obtained, and the weights of all edges between the two nodes in the structural association graph are adjusted by the functional dependency values to obtain the topological structure between the two nodes, and the above steps are repeated to obtain the topological structure between all nodes of the structural association graph, and then the image composed of all topological structures is used as the topological structure association graph; the structural association information of the data between each response module when a response error occurs is extracted from the topological structure association graph, which can be implemented in the following manner, namely: the graph algorithm in the prior art (such as PageRank) can be used to identify the association information of the data between each node (response module) in the topological structure association graph, and then the extracted association information is used as the structural association information of the data between each response module when a response error occurs.

[0082] It should be noted that in the analysis of error data characteristics, the structural association diagram can be used to determine the propagation path of error characteristics between various modules. By adjusting the weights and topological structure in the diagram, the error data recovery process can be optimized to ensure that errors on the critical path are recovered first. By analyzing the topological structure association diagram, it is possible to intuitively identify the close dependencies between response modules and which response modules may affect each other in the event of an error. This provides a clear basis for subsequent self-recovery operations and fault-tolerant recovery.

[0083] In some embodiments, performing a self-recovery operation on the response error data based on the structure association information to obtain the self-recovery data may be achieved by using the following steps:

[0084] Initialize the data recovery model;

[0085] Using the response error data as initialization parameters of the data recovery model;

[0086] Using the structural association information as a constraint parameter of the data recovery model;

[0087] The error features in the response error data are predicted and recovered using the data recovery model to obtain self-recovery data.

[0088] It should be noted that the data recovery model in this application is a machine learning model for repairing erroneous data. Its data recovery model can use machine learning algorithms (such as regression models, neural networks) to learn how to detect and repair erroneous data from historical response data. Through the training data set, the model can automatically identify error features and recover them. It can handle complex and changeable error patterns, has strong adaptability, and can also continuously optimize data recovery performance through training data.

[0089] In specific implementation, first, a data recovery model can be pre-trained using historical response data, and then the response error data can be used as the initialization parameter of the data recovery model and an initialization operation can be performed. Then, the structural association information can be set as the constraint parameter of the data recovery model, and the error features in the response error data can be further regressed and predicted using the data recovery model. The original error features can then be replaced with the self-recovery features obtained by the regression prediction, thereby using the replaced response error data as the self-recovery data.

[0090] It should be noted that the extracted structural association information in this application provides effective guidance for the precise positioning of errors and the rapid recovery of data, thereby reducing the time and complexity of error recovery, and through structural association information.

[0091] In step 104, the recovery granularity of each error log in the response error data when self-recovering after the response error is performed is determined by using the feature gradient difference and the information difference entropy in the self-recovery data.

[0092] In some embodiments, reference Figure 3 As shown in FIG, this figure is a schematic diagram of a process for determining the recovery granularity in some embodiments of the present application. In this embodiment, the recovery granularity of each error log in the response error data when self-recovering after the response error is performed is determined by using the feature gradient difference and the information difference entropy in the self-recovery data. The following steps can be used to implement it:

[0093] In step 1041, for each error log in the response error data, a plurality of self-recovery features of the error log are extracted from the self-recovery data;

[0094] In step 1042, the gradient difference values between the self-recovery features are determined, and the set of all gradient difference values is used as the feature gradient difference in the self-recovery data;

[0095] In step 1043, the granularity difference between each self-recovery feature is determined by using each gradient difference value in the feature gradient difference and the information difference entropy;

[0096] In step 1044, the recovery granularity of the error log when self-recovering after responding to the error is determined based on all granularity differences, thereby obtaining the recovery granularity of each error log in the response error data when self-recovering after responding to the error.

[0097] It should be noted that the gradient difference value in this application is an indicator for measuring the relative rate of change between self-recovery features. The gradient difference value helps to identify which self-recovery features change most significantly during the recovery process, thereby improving the accuracy of quantifying the recovery granularity; the feature gradient difference is an indicator for measuring the degree of difference between the change rates of different self-recovery features in the data space. The feature gradient difference can reflect the degree of change of each self-recovery feature over time or when an event occurs, as well as the relative relationship of these changes in the feature space; the recovery granularity refers to the degree of detail of the data state involved in the recovery operation during the data self-recovery process. By quantifying the recovery granularity, it can be ensured that the recovery operation can accurately repair the target error feature while reducing the impact on other parts of the system.

[0098] In specific implementation, the gradient difference value between each self-recovery feature can be determined in the following manner, namely: the numerical differentiation method in the prior art can be used to differentiate the numerical value in each self-recovery feature to obtain the gradient value of each self-recovery feature, and then the absolute difference between the gradient values of each self-recovery feature is used as the gradient difference value between the corresponding self-recovery features; the granularity difference between each self-recovery feature is determined by each gradient difference value in the feature gradient difference and the information difference entropy, which can be implemented in the following manner, namely: a self-recovery feature is selected as the selected self-recovery feature, the gradient difference value between the selected self-recovery feature and other self-recovery features is obtained, and the standard deviation of all gradient difference values is used as the gradient difference standard deviation, and the ratio of the information difference entropy to the gradient difference standard deviation is further used as the ratio of the selected self-recovery feature to the other self-recovery features. granularity differences between them, repeat the above steps to continue to determine the granularity differences between the remaining selected self-recovery features and other self-recovery features; determining the recovery granularity of the error log when self-recovering after responding to an error based on all granularity differences can be achieved in the following way, namely: performing information entropy feature description on all granularity differences, and using the obtained information entropy as the recovery granularity of the error log when self-recovering after responding to an error. It should be noted that information entropy has good adaptability and can be dynamically adjusted according to real-time changes in system status. The error characteristics and granularity differences in different error logs may change over time. These changes can be evaluated in real time through information entropy, thereby dynamically adjusting the recovery granularity. This dynamic adjustment capability makes the system self-recovery process more flexible and adaptable, and can make timely adjustments according to actual conditions, thereby improving the efficiency and effectiveness of system self-recovery.

[0099] In step 105, the backup data of the target electric power meter is obtained, and the node backup data of each response module is extracted from the backup data. Then, the structural dependency relationship between the backup data of each node and the information difference entropy are used to determine the fault-tolerant recovery cost when the backup data is used for multiplexing recovery after a response error.

[0100] It should be noted that obtaining the backup data of the target power meter in this application refers to obtaining the backup data of the target power meter from a cloud database. In other embodiments, the backup data of the target power meter can also be obtained from other distributed backup databases. This is not limited here. The backup data here is the data collected by the power meter directly transmitted to the cloud database through Internet technology, and does not pass through the local storage unit of the power meter.

[0101] It should also be noted that extracting the node backup data of each response module from the backup data refers to intercepting the backup data at the error node in each response module from the backup data. In specific implementation, the timestamp information of the error node can be matched with the timestamp information of the backup data, the backup data at the matching node can be intercepted, and the intercepted node data can be used as the node backup data of the corresponding response module.

[0102] In some embodiments, the fault-tolerant recovery cost of using the backup data for multiplexing recovery after a response error is determined based on the structural dependency relationship between the backup data of each node and the information difference entropy can be implemented by the following steps:

[0103] Obtain self-recovery data after executing the self-recovery operation;

[0104] Reusing the self-recovery data using the backup data to obtain multiplexed recovery data, and then determining a recovery confidence distance between the multiplexed recovery data and the self-recovery data;

[0105] Determine the structural dependencies between backup data of each node;

[0106] Determining the recovery cost of using the backup data for multiplexing and recovery after a response error by using the structural dependency and the recovery confidence distance;

[0107] The recovery cost is compensated for fault tolerance based on the information difference entropy to obtain a fault tolerance recovery cost when the backup data is used for multiplexing recovery after a response error.

[0108] In a specific implementation, the backup data is used to reuse and restore the self-recovery data to obtain the reused recovery data, and then the recovery confidence distance between the reused recovery data and the self-recovery data is determined. This can be achieved by the following method: first, all node backup data are obtained from the backup data, and all self-recovery features are obtained from the self-recovery data, wherein there is a one-to-one correspondence between the node backup data and the self-recovery features, and then the corresponding node backup data and self-recovery features are marked as feature correspondence groups. Then, for each feature correspondence group, the Euclidean distance between the node backup data and the self-recovery features in the feature correspondence group is used as the feature distance. When the feature distance is greater than a preset feature distance threshold, the node backup data is replaced with the self-recovery feature. When the feature distance is less than or equal to the preset feature distance threshold, the self-recovery feature is not replaced, and then the self-recovery data after the replacement is used as the reused recovery data. Finally, the Euclidean distance between the reused recovery data and the self-recovery data can be used as the recovery confidence distance between the reused recovery data and the self-recovery data.

[0109] It should be noted that the recovery confidence distance in the present application is an indicator for measuring the degree of difference between the reused recovery data and the self-recovery data. The recovery confidence distance can be used to quantify the effect of the reused recovery operation, and then quantify the degree of change of the reused recovery data relative to the self-recovery data, which helps to judge the effectiveness of the reused recovery and the degree of improvement it brings to the system status. It should also be noted that the feature distance threshold preset in the present application can be preset by the Euclidean distance between the node backup data and the self-recovery feature in a large number of feature corresponding groups. The feature distance threshold preset in the present application is the average value of all Euclidean distances, which will not be elaborated here.

[0110] In specific implementation, determining the structural dependency between the backup data of each node can be achieved in the following manner, namely: the backup data of each node can be input into the existing Graphviz graph visualization model, and then a dependency graph can be generated through Graphviz, and the connecting edges between each node can be used as the degree of dependency between the node backup data, so that the connection relationship of all nodes in the dependency graph can be used as the structural dependency between the backup data of each node; determining the recovery cost of using the backup data for reuse and recovery after a response error through the structural dependency and the recovery confidence distance can be achieved in the following manner: first, the connecting edges between each node in the structural dependency can be obtained, and the length of the connecting edge can be used as an indicator of the dependency between the node backup data. The longer the connecting edge, the weaker the dependency between the node backup data, and the shorter the connecting edge, the stronger the dependency between the node backup data. Then, the average length of all connecting edges can be used as the average dependency, and further, the product of the average dependency and the recovery confidence distance can be used as the recovery cost of using the backup data for reuse and recovery after a response error.

[0111] It should be noted that the recovery cost in this application is an indicator of the cost or resources required to use the backup data for reuse and recovery after a response error. The larger the recovery cost, the more cost or resources are required to use the backup data for reuse and recovery after a response error. The smaller the recovery cost, the less cost or resources are required to use the backup data for reuse and recovery after a response error.

[0112] In specific implementation, the recovery cost is fault-tolerantly compensated based on the information difference entropy, and the fault-tolerant recovery cost when the backup data is used for reuse recovery after the response error is obtained can be implemented in the following way: the opposite of the information difference entropy can be used as the opposite value of the difference entropy, and the exponential function value of the opposite value of the difference entropy with the natural logarithm as the base is used as the fault-tolerant compensation coefficient, and then the product of the fault-tolerant compensation coefficient and the recovery cost is used as the fault-tolerant recovery cost. It should be noted that the solution of the present application improves the flexibility and adaptability of the system in the face of different fault situations. For situations with large differences, the fault-tolerant compensation coefficient will automatically increase, and the calculation of the recovery cost will be improved, thereby ensuring that the system can adapt to more complex recovery needs. For situations with small differences, the compensation coefficient is low, reducing unnecessary recovery overhead.

[0113] It should be noted that the fault-tolerant recovery cost in this application refers to the additional cost added during the system fault-tolerant processing process based on the cost required for reuse recovery using backup data and self-recovery data. The recovery cost is compensated for fault tolerance through information differential entropy, so that the final cost not only reflects the direct cost of the recovery operation, but also includes additional costs such as adaptive adjustment and redundant processing performed by the system to adapt to complex environments.

[0114] In step 106, when the data fault tolerance rate of the target electric power meter is lower than the standard fault tolerance threshold, the data recovery strategy is adjusted through the recovery granularity and the fault tolerance recovery cost feedback, and then the data of the target electric power meter after the response error is confidently recovered based on the feedback-adjusted data recovery strategy.

[0115] It should be noted that the present application can monitor the data fault tolerance rate of the target electric power meter in real time, and then obtain the current data fault tolerance rate. When the data fault tolerance rate is lower than the set standard fault tolerance threshold, the data recovery strategy adjustment process is triggered. The standard fault tolerance threshold can be set according to the historically monitored data fault tolerance rate, that is, the average value of all historical data fault tolerance rates can be used as the standard fault tolerance threshold, which is not limited here.

[0116] In some embodiments, adjusting the data recovery strategy based on the recovery granularity and the fault-tolerant recovery cost feedback may be achieved by the following steps:

[0117] Get the policy value of the data recovery policy;

[0118] Determine the state feedback amount of the data recovery strategy by using the recovery granularity and the fault-tolerant recovery cost;

[0119] The data recovery strategy is adjusted based on the strategy value and the state feedback amount.

[0120] It should be noted that the policy value in this application is a set of preset parameters used by the current system for data recovery operations, including the execution method, recovery level and resource allocation priority of the recovery operation; the state feedback quantity in this application is a measurement indicator used to evaluate and adjust the current operating state of the system, and the policy parameters of the current data recovery strategy can be adjusted according to the size of the state feedback quantity.

[0121] In the specific implementation, first, a policy adjustment model can be initialized (for example: a policy adjustment model based on the support vector machine algorithm). The core function of the policy adjustment model is to automatically adjust the recovery strategy according to the feedback information of the actual data recovery situation to improve the effect and efficiency of data recovery, and then read the policy value of the current data recovery strategy through the policy adjustment model; secondly, the recovery granularity and the fault-tolerant recovery cost are used as state parameters of the policy adjustment model, and the state feedback amount of the data recovery strategy is obtained by analyzing and calculating the state function built into the policy adjustment model. As a preferred embodiment, the state feedback amount = F(A*recovery granularity, B*fault-tolerant recovery cost, C), where F(A*recovery granularity, B*fault-tolerant recovery cost, C) represents the state function, A*recovery granularity, B*fault-tolerant recovery cost, C The cost and C are independent variables, A and B are weight coefficients, and C is a state feedback constant. The state feedback constant refers to the minimum adjustment step to ensure feedback adjustment. In this application, the value of A can be set to 0.5, the value of B can be set to 0.5, and the value of C can be set to 0.01. In other embodiments, it can be adjusted according to specific needs and is not limited here. Finally, the parameters in the policy value are adjusted by the state feedback amount to make it more in line with actual recovery needs. For example, if the state feedback amount shows that the recovery granularity is too fine but the cost is too high, the policy value can be adjusted to reduce the recovery accuracy and reduce resource consumption. Conversely, if the fault-tolerant recovery cost is low but the recovery effect is insufficient, the recovery intensity can be increased, and then through multiple iterations, the policy value can be continuously fine-tuned to gradually reach the optimal configuration, thereby achieving the best data recovery effect.

[0122] It should be noted that the data recovery strategy based on feedback adjustment in this application performs confidence recovery of the data of the target power meter after a response error, which means that the data of the target power meter after a response error is recovered through an optimized data recovery strategy to improve the confidence of data recovery. Confidence recovery achieves high-confidence data recovery of the target power meter after a response error through the data recovery strategy adjusted by feedback. Its core lies in ensuring the accuracy of the recovered data through dynamic adjustment, and ensuring the reliability of the system in the face of data errors. In specific implementation, the data recovery strategy after feedback adjustment can be collected to recover the response error data, thereby improving the confidence of data recovery.

[0123] It should also be noted that this application can continuously optimize the data recovery strategy during the data recovery process through a feedback mechanism, ensuring that the data recovery operation achieves the best balance between resource consumption and recovery effect.

[0124] On the other hand, in some embodiments, the present application provides an electric power meter data recovery system, referring to Figure 4, which is a schematic diagram of exemplary software modules of an electric power meter data recovery system according to some embodiments of the present application. The electric power meter data recovery system 400 includes: an acquisition module 401, a processing module 402, and an execution module 403, which are described as follows:

[0125] Acquisition module 401, in this application, acquisition module 401 is mainly used to monitor the status identifiers of each response module in the target power meter after receiving the start instruction of the target power meter, filter out the error logs of each response module according to all the status identifiers, and obtain response error data;

[0126] Processing module 402, in this application, is used to perform error node perception on the response error data based on a pre-trained error detection model, obtain error node features of each response module, and then determine the information difference entropy between each error node feature;

[0127] The processing module 402 in the present application is further configured to determine, based on all error node characteristics, structural association information of data between various response modules when a response error occurs, and then perform a self-recovery operation on the response error data based on the structural association information to obtain self-recovery data;

[0128] The processing module 402 in the present application is further configured to determine the recovery granularity of each error log in the response error data when self-recovering after the response error by using the feature gradient difference and the information difference entropy in the self-recovery data;

[0129] The processing module 402 in the present application is further configured to obtain backup data of the target electric power meter, extract node backup data of each response module from the backup data, and then determine the fault-tolerant recovery cost when the backup data is reused and recovered after a response error based on the structural dependency relationship between the backup data of each node and the information difference entropy;

[0130] Execution module 403. In this application, execution module 403 is mainly used to adjust the data recovery strategy through the recovery granularity and the fault tolerance recovery cost feedback when the data fault tolerance rate of the target power meter is lower than the standard fault tolerance threshold, and then perform confidence recovery on the data of the target power meter after the response error based on the feedback-adjusted data recovery strategy.

[0131] In addition, the present application also provides a computer device, which includes a memory and a processor, wherein the memory stores a code, and the processor is configured to obtain the code and execute the above-mentioned electric power meter data recovery method.

[0132] In some embodiments, reference Figure 5, which is a schematic diagram of the structure of a computer device using a method for recovering power meter data according to some embodiments of the present application. The method for recovering power meter data in the above embodiment can be Figure 5 The computer device 500 shown in FIG. 5 is implemented as shown in FIG. 5 . The computer device 500 includes at least one processor 501 , a communication bus 502 , a memory 503 , and at least one communication interface 504 .

[0133] The processor 501 may be a general-purpose central processing unit (CPU) or an application-specific integrated circuit (ASIC).

[0134] The communication bus 502 may be used to transmit information between the aforementioned components.

[0135] The memory 503 may be a read-only memory (ROM) or other static storage device that can store static information and instructions, a random access memory (RAM) or other dynamic storage device that can store information and instructions, or an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, an optical disc storage (including a compact disc, laser disc, optical disc, digital versatile disc, Blu-ray disc, etc.), a magnetic disk or other magnetic storage device, or any other medium that can be used to carry or store desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited thereto. The memory 503 may be independent and connected to the processor 501 via the communication bus 502. The memory 503 may also be integrated with the processor 501.

[0136] Memory 503 is used to store program code for executing the solution of the present application, and is controlled by processor 501 for execution. Processor 501 is used to execute the program code stored in memory 503. The program code may include one or more software modules. The power meter data recovery method in the above embodiment can be implemented by processor 501 and one or more software modules in the program code in memory 503.

[0137] The communication interface 504 uses any transceiver or other device for communicating with other devices or communication networks, such as Ethernet, radio access network (RAN), wireless local area network (WLAN), etc.

[0138] In a specific implementation, as an embodiment, a computer device may include multiple processors, each of which may be a single-core (single-CPU) processor or a multi-core (multi-CPU) processor. The processor herein may refer to one or more devices, circuits, and / or processing cores for processing data (e.g., computer program instructions).

[0139] The aforementioned computer device may be a general-purpose computer device or a dedicated computer device. In a specific implementation, the computer device may be a desktop computer, a portable computer, a network server, a personal digital assistant (PDA), a mobile phone, a tablet computer, a wireless terminal device, a communication device, or an embedded device. The embodiments of the present application do not limit the type of computer device.

[0140] In addition, the present application also provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the above-mentioned electric power meter data recovery method is implemented.

[0141] Although the preferred embodiments of the present application have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present application.

[0142] Obviously, those skilled in the art may make various changes and modifications to this application without departing from the spirit and scope of this application. Thus, if these modifications and variations of this application fall within the scope of the claims of this application and their equivalents, this application is intended to include these modifications and variations.

Claims

1. A method for recovering data from an electric power meter, characterized in that: The steps include: Receive a start-up instruction from a target electric power meter, monitor the status identifiers of each response module in the target electric power meter, filter out error logs of each response module according to all status identifiers, and obtain response error data; Perform error node perception on the response error data based on a pre-trained error detection model to obtain error node features of each response module, and then determine the information difference entropy between the error node features; Determine, based on all error node characteristics, structural association information of data between various response modules when a response error occurs, and then perform a self-recovery operation on the response error data based on the structural association information to obtain self-recovery data; Determining the recovery granularity of each error log in the response error data when self-recovering after the response error by using the feature gradient difference and the information difference entropy in the self-recovery data; Acquire backup data of the target electric power meter, extract node backup data of each response module from the backup data, and then determine the fault-tolerant recovery cost when using the backup data for multiplexing recovery after a response error based on the structural dependency relationship between the backup data of each node and the information difference entropy; When the data fault tolerance rate of the target electric power meter is lower than the standard fault tolerance threshold, the data recovery strategy is adjusted through the recovery granularity and the fault tolerance recovery cost feedback, and then the data of the target electric power meter after the response error is confidently recovered based on the feedback-adjusted data recovery strategy.

2. The method according to claim 1, wherein Filter the error logs of each response module based on all status identifiers, and obtain the response error data including: Based on the pre-trained state model, an abnormal state identification set is obtained from all state identifications; Extracting error logs from the response logs of each response module according to the abnormal status identification set; A data group consisting of all error logs is used as response error data.

3. The method according to claim 1, wherein The error node perception of the response error data is performed based on the pre-trained error detection model, and the error node features of each response module are obtained, which specifically include: Using the response error data as initialization parameters of a pre-trained error detection model; The error detection model is used to perceive the error nodes in each response module and obtain the error node features of each response module.

4. The method according to claim 1, wherein Determining the information difference entropy between the features of each error node specifically includes: Determine the information difference between each error node feature and other error node features; All information differences are described by entropy features to obtain the information difference entropy between the features of each error node.

5. The method according to claim 1, wherein The structural association information of the data between the various response modules when a response error occurs is determined based on all error node characteristics, specifically including: Determine the structural association diagram of data interaction between each response module; Determine the functional dependency values between error node features through the propagation paths of all error node features in each response module; Performing topological adjustment on the edge weights in the structure association graph based on all functional dependency values to obtain a topological structure association graph; Structural association information of data between various response modules when a response error occurs is extracted from the topological structure association graph.

6. The method according to claim 1, wherein The fault-tolerant recovery cost when the backup data is used for multiplexing and recovery after a response error is determined based on the structural dependency relationship between the backup data of each node and the information difference entropy specifically includes: Obtain self-recovery data after executing the self-recovery operation; Reusing the self-recovery data using the backup data to obtain multiplexed recovery data, and then determining a recovery confidence distance between the multiplexed recovery data and the self-recovery data; Determine the structural dependencies between backup data of each node; Determining the recovery cost of using the backup data for multiplexing and recovery after a response error by using the structural dependency and the recovery confidence distance; The recovery cost is compensated for fault tolerance based on the information difference entropy to obtain a fault tolerance recovery cost when the backup data is used for multiplexing recovery after a response error.

7. The method according to claim 1, wherein Adjusting the data recovery strategy through feedback of the recovery granularity and the fault-tolerant recovery cost specifically includes: Get the policy value of the data recovery policy; Determine the state feedback amount of the data recovery strategy by using the recovery granularity and the fault-tolerant recovery cost; The data recovery strategy is adjusted based on the strategy value and the state feedback amount.

8. An electric power meter data recovery system, characterized in that: include: an acquisition module configured to, after receiving a start-up instruction from a target electric power meter, monitor the status identifiers of each response module in the target electric power meter, filter out error logs of each response module according to all status identifiers, and obtain response error data; A processing module, configured to perform error node perception on the response error data based on a pre-trained error detection model, obtain error node features of each response module, and further determine information difference entropy between the error node features; The processing module is further configured to determine, based on all error node characteristics, structural association information of data between various response modules when a response error occurs, and then perform a self-recovery operation on the response error data based on the structural association information to obtain self-recovery data; The processing module is further configured to determine, by using the feature gradient difference and the information difference entropy in the self-recovery data, a recovery granularity of each error log in the response error data when self-recovery is performed after the response error; The processing module is further configured to obtain backup data of the target electric power meter, extract node backup data of each response module from the backup data, and then determine the fault-tolerant recovery cost when the backup data is used for multiplexing recovery after a response error based on the structural dependency relationship between the backup data of each node and the information difference entropy; An execution module is used to adjust the data recovery strategy through the recovery granularity and the fault tolerance recovery cost feedback when the data fault tolerance rate of the target electric power meter is lower than the standard fault tolerance threshold, and then perform confidence recovery on the data of the target electric power meter after the response error based on the feedback-adjusted data recovery strategy.

9. A computer device comprising a memory and a processor, wherein the memory stores a code, wherein: The processor is configured to obtain the code and execute the electric power meter data recovery method according to any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the electric power meter data recovery method according to any one of claims 1 to 7 is implemented.

Images