Electric energy meter abnormity diagnosis method, system and equipment based on multi-source consistency verification and confidence fusion, and medium

Abstract

The invention discloses an electric energy meter abnormity diagnosis method, system and device based on multi-source consistency verification and confidence fusion and a medium, and belongs to the field of electric power metering operation and maintaining.The method comprises the steps that multi-source data information of a metering terminal of a low-voltage distribution area is collected, a unified index data set is established, differential data processing is carried out, and a data set is obtained; respectively inputting a preset feature coding network, constructing a standardized feature matrix to measure physical consistency, statistical consistency and communication consistency, and calculating a comprehensive consistency score; whether the metering terminal is abnormal or not is diagnosed according to the multi-source data consistency score through a preset threshold value; and deploying the anomaly diagnosis model in a zone area edge device to realize real-time anomaly diagnosis, periodically updating and issuing parameters by a cloud end, verifying and feeding back an anomaly diagnosis result by an edge end, and forming a cloud-edge collaborative metering anomaly diagnosis closed loop. According to the method, the accuracy and the stability of anomaly recognition are remarkably improved, misjudgment caused by environmental differences is effectively reduced, and the intelligence and the real-time performance of metering operation and maintenance are enhanced.

Description

Technical Field

[0001] This invention relates to the field of power metering and maintenance, specifically to a method, system, device, and medium for diagnosing abnormalities in electricity meters based on multi-source consistency verification and confidence fusion. Background Technology

[0002] In power system operation, electricity meters, as key devices for electricity measurement, are directly related to the economic benefits and reliability of power supply for power companies. Currently, traditional methods for diagnosing electricity meter anomalies often rely on a single data source, such as analyzing only electrical quantities like voltage and current. These methods have limited information dimensions, making it difficult to identify non-electrical issues such as communication anomalies and topology changes. Furthermore, fluctuations in electrical quantities due to noise or environmental interference can easily lead to misjudgments. In addition, some methods use fixed thresholds or static models, lacking self-learning and adaptive capabilities, and are unable to cope with the evolution of anomaly characteristics under new loads and complex power consumption environments.

[0003] Existing technologies include methods for detecting anomalies in electricity meters, but these primarily focus on statistical analysis of electrical quantities and time characteristics, without fully considering communication consistency and physical constraints. Furthermore, while some electricity consumption anomaly detection methods achieve rapid edge diagnosis, their data fusion levels are shallow, and the models lack dynamic confidence correction mechanisms.

[0004] In summary, existing methods for diagnosing anomalies in electricity meters still have shortcomings in terms of multi-source information utilization, adaptive capabilities, and cloud-edge collaborative architecture. They are difficult to balance real-time performance and accuracy. There is an urgent need for an intelligent diagnostic method and system that integrates multi-source consistency verification and confidence fusion to achieve accurate identification and self-evolutionary optimization of the operating status of metering terminals. Summary of the Invention

[0005] In view of the above-mentioned problems, the present invention is proposed.

[0006] Therefore, in order to overcome the problems of single data source, insufficient model adaptability and poor real-time performance of centralized architecture in existing energy meter anomaly diagnosis methods, this invention proposes an energy meter anomaly diagnosis method and system based on multi-source consistency verification and confidence fusion.

[0007] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a method for diagnosing anomalies in electricity meters based on multi-source consistency verification and confidence fusion, comprising, Multi-source data information from metering terminals in low-voltage distribution transformer areas is collected to establish a unified index dataset. Differentiated data processing is performed on the unified index dataset, and the data is input into a preset feature coding network to construct a standardized feature matrix. Multi-source consistency verification is performed on the standardized feature matrix, and a comprehensive consistency score is calculated. Anomaly diagnosis is performed on the comprehensive consistency score, and a preset threshold is used to diagnose whether there are anomalies in the metering terminals. Real-time anomaly diagnosis is performed, parameters are updated and distributed, and anomaly diagnosis results are fed back, forming a cloud-edge collaborative metering anomaly diagnosis closed loop.

[0008] As a preferred embodiment of the energy meter anomaly diagnosis method based on multi-source consistency verification and confidence fusion described in this invention, the multi-source data information includes electrical quantity data, communication metadata, and transformer area metering terminal topology information. A unified index dataset is created based on data timestamps from multiple sources.

[0009] As a preferred embodiment of the energy meter anomaly diagnosis method based on multi-source consistency verification and confidence fusion described in this invention, the differential data processing includes: adopting differential data processing for the type differences of multi-source data; Standardize and normalize electrical quantity data; The topology information of the metering terminal in the distribution area is converted into a binary vector.

[0010] As a preferred embodiment of the energy meter anomaly diagnosis method based on multi-source consistency verification and confidence fusion described in this invention, the construction of the standardized feature matrix includes combining the electrical quantity data, communication metadata, and topology information of each metering terminal in the low-voltage distribution substation after data processing within a time window to form a feature vector; All metering terminal features are spliced ​​together to form a standardized feature matrix.

[0011] As a preferred embodiment of the energy meter anomaly diagnosis method based on multi-source consistency verification and confidence fusion described in this invention, wherein: the calculation of the comprehensive consistency score includes, and the multi-source consistency verification includes physical consistency, statistical consistency, and communication element consistency; The physical consistency score A physical constraint model for the measurement results of the metering terminal is established, and the key parameters of voltage, current, and power are verified and calculated. in, Let t be the t-th physical quantity actually measured. This is a reference value calculated based on rated parameters or historical steady-state data, where n is the total number of parameters measured by the metering terminal. The statistical consistency score Based on time series stability and population distribution similarity, the mean Pearson correlation coefficient among the feature vectors of each terminal is calculated: The communication element consistency score Communication reliability is calculated based on communication metadata metrics, including latency and packet loss rate. in, For data time intervals, This represents the average drift within a time window under normal historical communication conditions. , These are the weighting coefficients for time drift and packet loss rate. , This is the feature vector after data processing, where N is the total number of metering terminals. For packet loss rate, For data time, This refers to the data collection time.

[0012] The preferred technical solution in the embodiments of the present invention has the following beneficial effects: by using confidence fusion of multi-source consistency scores and threshold discrimination, the accuracy and reliability of anomaly diagnosis are improved, and subjective misjudgment is reduced.

[0013] As a preferred embodiment of the energy meter anomaly diagnosis method based on multi-source consistency verification and confidence fusion described in this invention, the step of diagnosing whether the metering terminal has an anomaly through a preset threshold includes: the metering terminal anomaly diagnosis standard is a threshold discrimination method; and the consistency scores are fused with confidence to obtain a comprehensive consistency score. ,in, For Sigmoid mapping function, , , The multi-source consistency score weighting coefficients are assigned to scores with different consistency levels; when the overall consistency score is... Greater than the set threshold If the reading is normal, the metering terminal is diagnosed as normal; otherwise, it is diagnosed as abnormal.

[0014] The preferred technical solution in the embodiments of the present invention has the following beneficial effects: by using edge feedback and cloud threshold correction, the diagnostic model is dynamically optimized, the false positive and false negative rates are reduced, and the real-time diagnostic accuracy is improved.

[0015] As a preferred embodiment of the energy meter anomaly diagnosis method based on multi-source consistency verification and confidence fusion described in this invention, the real-time anomaly diagnosis includes: verifying and feeding back the anomaly diagnosis result; the edge verification and feedback results will affect the anomaly diagnosis threshold; after receiving the edge result feedback, the cloud calculates the model diagnosis accuracy. Where TP represents the number of correctly identified abnormal samples, TN represents the number of correctly identified normal samples, FP represents normal samples that were misclassified as abnormal, and FN represents abnormal samples that were misclassified as normal. Implement anomaly diagnosis threshold correction strategy: in, This is a correction factor; it is negative when a case is misjudged and positive when a case is missed. To achieve the normalization rate of the target abnormal diagnosis, The current rate of normal diagnosis of abnormalities, The corrected threshold. The threshold before correction is used to construct a cloud-edge collaborative metering anomaly diagnosis closed loop. The consistency score weight coefficient and diagnosis threshold are periodically issued by the cloud. Metering personnel perform real-time anomaly diagnosis through the edge devices of the substation where the anomaly diagnosis model is deployed, and verify and provide feedback on the anomaly diagnosis results in real time. The cloud corrects the diagnosis threshold and reissues it, forming a cloud-edge collaborative metering anomaly diagnosis closed loop.

[0016] The beneficial effects of the preferred technical solution in the embodiments of the present invention are as follows: by circulating parameters from the cloud and feedback from the edge, the diagnostic parameters can be continuously self-learned and adaptively adjusted, thereby enhancing the long-term stability and adaptability of the system.

[0017] Another objective of this invention is to provide an energy meter anomaly diagnosis system based on multi-source consistency verification and confidence fusion.

[0018] To solve the above-mentioned technical problems, the present invention provides the following technical solution: an energy meter anomaly diagnosis system based on multi-source consistency verification and confidence fusion, comprising: a multi-source data acquisition module, a feature encoding module, a consistency verification module, an anomaly diagnosis module, a confidence update module, an edge deployment module, and a cloud management module; The multi-source data acquisition module collects multi-source data information from metering terminals in low-voltage distribution transformer areas and establishes a unified index dataset. The feature encoding module performs differentiated data processing on the unified index dataset and inputs it into a preset feature encoding network to construct a standardized feature matrix. The consistency verification module performs multi-source consistency verification on the standardized feature matrix and calculates the comprehensive consistency score. The anomaly diagnosis module performs anomaly diagnosis based on the overall consistency score and diagnoses whether the metering terminal is abnormal by using a preset threshold. The edge deployment module and the cloud management module perform real-time anomaly diagnosis, update and distribute parameters, and provide feedback on anomaly diagnosis results, forming a closed loop for metering anomaly diagnosis through cloud-edge collaboration.

[0019] The present invention provides a computer device, including a memory and a processor, wherein the memory stores a computer program, characterized in that the processor executes the computer program to implement the steps of the above-described method for diagnosing anomalies in energy meters based on multi-source consistency verification and confidence fusion.

[0020] The present invention provides a computer-readable storage medium having a computer program stored thereon, characterized in that, when the computer program is executed by a processor, it implements the steps of the above-described method for diagnosing anomalies in electricity meters based on multi-source consistency verification and confidence fusion.

[0021] The beneficial effects of this invention are as follows: This invention establishes three types of measurement models: physical consistency, statistical consistency and communication consistency. These models can simultaneously verify data from the perspectives of physical constraints, temporal stability and communication reliability. This overcomes the shortcomings of traditional methods, such as strong dependence on local features and susceptibility to noise interference, thereby improving the robustness of the model in complex operating scenarios.

[0022] The confidence fusion and cloud-edge collaboration mechanism proposed in this invention enables dynamic adaptive correction of diagnostic model parameters. By combining real-time edge diagnosis with periodic cloud updates, the model can automatically adjust the consistency score weights and thresholds based on actual feedback, effectively mitigating misjudgments and missed judgments caused by differences in the field environment.

[0023] This invention enables real-time anomaly diagnosis in low-voltage distribution area metering systems, significantly improving the intelligence, real-time capabilities, and self-learning abilities of metering devices, and providing efficient and reliable technical support for power metering operation and maintenance, distribution automation, and power data quality control. Attached Figure Description

[0024] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0025] Figure 1 The above is a flowchart of an overall method for diagnosing energy meter anomalies based on multi-source consistency verification and confidence fusion, which is provided as an embodiment of the present invention. Detailed Implementation

[0026] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0027] Example 1, referring to Figure 1 This is one embodiment of the present invention, which provides a method for diagnosing anomalies in electricity meters based on multi-source consistency verification and confidence fusion, including: S100: Collect multi-source data information from metering terminals in low-voltage distribution transformer areas and establish a unified index dataset; S200. Perform differentiated data processing on the unified index dataset and input them into the preset feature encoding network to construct a standardized feature matrix; S300. Perform multi-source consistency verification on the standardized feature matrix and calculate the overall consistency score. S400: Perform anomaly diagnosis on the overall consistency score, and diagnose whether there is anomaly in the metering terminal by using a preset threshold. S500 performs real-time anomaly diagnosis, updates and distributes parameters, and provides feedback on anomaly diagnosis results, forming a cloud-edge collaborative metering anomaly diagnosis closed loop. It should be noted that existing methods for diagnosing abnormalities in electricity meters often suffer from a variety of defects, including the use of single data sources leading to frequent misjudgments and missed judgments, rigid diagnostic models lacking adaptability, and insufficient cloud-edge collaboration resulting in a lack of closed-loop management. These defects limit their accuracy and long-term effectiveness under complex actual working conditions.

[0028] Therefore, in response to the aforementioned problems, through steps S100-S500, this invention achieves a multi-dimensional consistency assessment of the operating status of metering terminals by integrating electrical quantities, communication metadata, and topology information, effectively improving the accuracy and robustness of diagnosis and providing technical support for proactive operation and maintenance of intelligent metering systems. This invention can be widely applied in fields such as intelligent metering in distribution substations and the operation and maintenance of metering equipment.

[0029] Example 2, refer to Figure 1 This is one embodiment of the present invention, which provides a method for diagnosing anomalies in electricity meters based on multi-source consistency verification and confidence fusion, including: In this embodiment of the invention, step S100 involves collecting multi-source data information from low-voltage distribution transformer area metering terminals and establishing a unified index dataset, including the following steps: S101: S101, Multi-source data information includes: Electrical quantity data includes three-phase voltage , , Three-phase current value , , Phase angle , , Active power P, power factor CT ratio ; Communication metadata includes data time. Collection time Packet loss rate ; The topology information of the metering terminal in the distribution area includes the metering terminal ID, the name of the upstream line, and the association relationship.

[0030] In this embodiment of the invention, step S200 involves differential data processing of the unified index dataset and inputting it into a preset feature encoding network to construct a standardized feature matrix, including the following steps S201-S203: S201. Differentiated data processing refers to the use of differentiated data processing methods to address the differences in the types of multi-source data. In embodiments of the present invention, the standardization and normalization of electrical quantity data includes the following steps A1-A2: A1. For continuous numerical electrical quantity data, Z-score standardization is used to eliminate the influence of dimensions on subsequent feature extraction. The calculation formula is as follows: Where x represents the original electrical quantity data. This represents the average of this type of data over historical periods during normal operation. This represents the standard deviation of the historical period during normal operation.

[0031] A2. Continuous numerical electrical quantity data include three-phase voltage, current, active power, etc.

[0032] In an optional implementation, the standardization process in S201 can be a decimal scaling standardization process. For electrical quantity data, the number of decimal places to be moved is determined based on the maximum absolute value of each electrical quantity data. Then, each data point is divided by a power of 10 so that the data falls within the interval [-1, 1].

[0033] In embodiments of the present invention, the standardization and normalization of electrical quantity data includes the following steps B1-B2: B1. Proportional values ​​are normalized using the min-max method, and the calculation formula is as follows: in, , These are the minimum and maximum values ​​of this type of data within a preset historical period, respectively.

[0034] B2. Proportional values ​​include power factor, packet loss rate, etc.

[0035] In an optional implementation, the normalization process in S201 can be mean normalization, whereby for all proportional values, the mean, minimum, and maximum values ​​of each feature data are calculated, and each data value is scaled to adjust the data range to the [-1, 1] interval.

[0036] S202. The topology information of the metering terminal in the distribution area is converted into a binary vector using one-hot encoding to eliminate the deviation and interference introduced by the identification difference.

[0037] S203. Construct the standardized feature matrix, obtained by the following method: Each metering terminal in the low-voltage distribution transformer area will be within the time window. The electrical quantity data, communication metadata, and topology information after internal data processing are combined to form a feature vector: in, For electrical quantity characteristics, For communication characteristics, It is a topological feature.

[0038] All metering terminal features are concatenated to form a standardized feature matrix: Where N is the number of metering terminals in the distribution area, and d is the feature dimension.

[0039] In an embodiment of the present invention, step S300 involves performing a multi-source consistency check on the standardized feature matrix and calculating a comprehensive consistency score, including the following steps S301-S302: S301. Multi-source consistency verification includes physical consistency, statistical consistency, and communication element consistency. The physical consistency score A physical constraint model for the measurement results of the metering terminal is established, and the key parameters of voltage, current, and power are verified and calculated. in, Let t be the t-th physical quantity actually measured. This is a reference value calculated based on rated parameters or historical steady-state data, where n is the total number of parameters measured by the metering terminal. S302, the statistical consistency score Based on time series stability and population distribution similarity, the mean Pearson correlation coefficient among the feature vectors of each terminal is calculated: S303, the consistency score of the communication element Communication reliability is calculated based on communication metadata metrics, including latency and packet loss rate. in, For data time intervals, This represents the average drift within a time window under normal historical communication conditions. , These are the weighting coefficients for time drift and packet loss rate. , This is the feature vector after data processing.

[0040] In this embodiment of the invention, step S400 involves performing anomaly diagnosis on the overall consistency score and diagnosing whether the metering terminal is abnormal by using a preset threshold. This includes the following steps S401-S402: S401. The diagnostic standard for metering terminals is a threshold discrimination method, which integrates the consistency scores with confidence to obtain a comprehensive consistency score. ,in, For Sigmoid mapping function, , , Multi-source consistency score weighting coefficients are assigned to scores with different levels of consistency. S402, When the overall consistency score Greater than the set threshold If the reading is normal, the metering terminal is diagnosed as normal; otherwise, it is diagnosed as abnormal.

[0041] In an optional implementation, the diagnosis of whether the metering terminal is abnormal in S400 can be based on majority voting. An independent threshold is set for each consistency score, and each consistency score is compared to see if it exceeds its corresponding threshold. If at least two consistency scores exceed the threshold, the metering terminal is diagnosed as normal; otherwise, it is diagnosed as abnormal. However, this implementation may misjudge due to the imbalance of consistency scores.

[0042] In another optional implementation, the S400 method for diagnosing whether a metering terminal is abnormal can also be based on anomaly diagnosis through historical trend comparison. This involves calculating the overall consistency score for the current time window, obtaining the overall consistency score sequence of the metering terminal within a historical normal period, and calculating the average or median of the historical scores as a reference benchmark. The current overall consistency score is then compared with the historical reference benchmark. If the current score is not lower than 90% of the historical benchmark, the metering terminal is diagnosed as having no abnormality; otherwise, it is diagnosed as abnormal. However, because the historical benchmark is updated later, this can lead to a detection delay.

[0043] In an embodiment of the present invention, real-time anomaly diagnosis is performed in S500, the parameters are updated and the anomaly diagnosis results are fed back, forming a cloud-edge collaborative metering anomaly diagnosis closed loop, including the following steps S501-S502: In an embodiment of the present invention, S501, feeding back the abnormal diagnosis result, includes the following steps C1-C2: C1. Verify and feedback the abnormal diagnosis results. The edge verification and feedback results will affect the abnormal diagnosis threshold. C2. After receiving the results from the edge, the cloud calculates the model's diagnostic accuracy: Where TP represents the number of correctly identified abnormal samples, TN represents the number of correctly identified normal samples, FP represents normal samples that were misclassified as abnormal, and FN represents abnormal samples that were misclassified as normal. In an optional implementation, the feedback mechanism for anomaly diagnosis results in S501 can be based on a dynamic threshold adjustment mechanism. The edge sends the anomaly diagnosis results to the cloud, along with a timestamp and terminal ID. The cloud calculates the false positive rate and false negative rate of anomaly diagnosis based on recent feedback results. If the false positive rate exceeds a preset upper limit, the anomaly diagnosis threshold is automatically increased; if the false negative rate exceeds a preset upper limit, the anomaly diagnosis threshold is automatically decreased. The cloud sends the updated threshold to the edge for subsequent real-time diagnosis, forming a closed-loop adjustment. However, this is not suitable for situations with large data fluctuations, which may cause threshold oscillations.

[0044] In another optional implementation, the feedback of abnormal diagnosis results in S501 can also be based on a feedback mechanism that updates the weight coefficients. The edge feeds back the abnormal diagnosis results and the actual operating status to the cloud. The cloud calculates the matching degree between each consistency score and the diagnosis results over a period of time, dynamically adjusts the weight coefficients of the multi-source consistency scores based on the matching degree, recalculates the weights in the comprehensive consistency score formula, and sends the updated weight coefficients to the edge to optimize the calculation of the comprehensive consistency score. However, this method cannot respond to sudden anomalies in a timely manner, resulting in a diagnosis delay.

[0045] S502, Implement anomaly diagnosis threshold correction strategy: in, This is a correction factor; it is negative when a case is misjudged and positive when a case is missed. To achieve the normalization rate of the target abnormal diagnosis, The current rate of normal diagnosis of abnormalities, The corrected threshold. The threshold before correction; The cloud-edge collaborative metering anomaly diagnosis closed loop is constructed by periodically issuing consistency score weighting coefficients and diagnostic thresholds from the cloud. Metering personnel perform real-time anomaly diagnosis through the edge devices of the transformer area where the anomaly diagnosis model is deployed, and verify and provide feedback on the anomaly diagnosis results in real time. The cloud corrects the diagnostic thresholds and reissues them, thus forming a cloud-edge collaborative metering anomaly diagnosis closed loop.

[0046] Example 3 is an embodiment of the present invention, which provides a virtual power plant profit-driven optimization scheduling method. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through experiments.

[0047] In this embodiment, the present invention is applied to a low-voltage distribution transformer area with a total of 48 metering terminals. The system consists of a multi-source data acquisition module, a feature encoding module, a consistency verification module, an anomaly diagnosis module, a confidence update module, an edge deployment module, and a cloud management module.

[0048] First, the multi-source data acquisition module collects multi-source data from each metering terminal every 15 minutes, as shown in Table 1. Table 1. Multi-source data collected from various metering terminals

[0049] Z-score standardization is applied to continuous features such as voltage, current, and active power, with μ and σ representing the mean and standard deviation of monthly data during normal operation, respectively. Min-max normalization is performed on proportional features such as power factor and packet loss rate. One-hot encoding is used for topology identifiers. The final result is a dimensionless array. The standardized feature matrix.

[0050] The consistency verification module calculates the consistency scores for the three categories respectively, and then uses the confidence fusion function to obtain the overall consistency score. ,when The metering terminal was immediately identified as malfunctioning.

[0051] For example, terminal T002 experienced power fluctuations and increased communication latency during the period from 10:00:00 to 11:00:00 on October 1, 2025. Calculations show that: When an anomaly is diagnosed, the cloud-edge collaboration mechanism enables adaptive correction: edge devices periodically report abnormal samples, the cloud adjusts the threshold based on the actual accuracy, and the parameters are reissued in the next cycle, thus forming a dynamic closed-loop update.

[0052] Example 4 is an embodiment of the present invention. The above is an illustrative scheme of a method for diagnosing energy meter anomalies based on multi-source consistency verification and confidence fusion. It should be noted that the technical solution of an energy meter anomaly diagnosis system based on multi-source consistency verification and confidence fusion and the technical solution of the above-described method for diagnosing energy meter anomalies based on multi-source consistency verification and confidence fusion belong to the same concept. Details not described in detail in the technical solution of the energy meter anomaly diagnosis system based on multi-source consistency verification and confidence fusion in this embodiment can be found in the description of the above-described method for diagnosing energy meter anomalies based on multi-source consistency verification and confidence fusion.

[0053] This embodiment provides an energy meter anomaly diagnosis system based on multi-source consistency verification and confidence fusion, including: a multi-source data acquisition module, a feature encoding module, a consistency verification module, an anomaly diagnosis module, a confidence update module, an edge deployment module, and a cloud management module; The multi-source data acquisition module collects multi-source data information from metering terminals in low-voltage distribution transformer areas and establishes a unified index dataset. The feature encoding module performs differentiated data processing on the unified index dataset and inputs it into a preset feature encoding network to construct a standardized feature matrix. The consistency verification module performs multi-source consistency verification on the standardized feature matrix and calculates the comprehensive consistency score. The anomaly diagnosis module performs anomaly diagnosis based on the overall consistency score and diagnoses whether the metering terminal is abnormal by using a preset threshold. The edge deployment module and the cloud management module perform real-time anomaly diagnosis, update and distribute parameters, and provide feedback on anomaly diagnosis results, forming a closed loop for metering anomaly diagnosis through cloud-edge collaboration.

[0054] This embodiment also provides an electronic device applicable to a method for diagnosing energy meter anomalies based on multi-source consistency verification and confidence fusion, comprising: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the method for diagnosing energy meter anomalies based on multi-source consistency verification and confidence fusion as proposed in the above embodiment.

[0055] This embodiment also provides a storage medium storing a computer program that, when executed by a processor, implements a method for diagnosing energy meter anomalies based on multi-source consistency verification and confidence fusion as proposed in the above embodiment.

[0056] The storage medium proposed in this embodiment belongs to the same inventive concept as the method for diagnosing anomalies in electricity meters based on multi-source consistency verification and confidence fusion proposed in the above embodiments. Technical details not described in detail in this embodiment can be found in the above embodiments, and this embodiment has the same beneficial effects as the above embodiments.

[0057] Based on the above description of the implementation methods, those skilled in the art can clearly understand that the present invention can be implemented using software and necessary general-purpose hardware, and of course, it can also be implemented using hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk, or optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods of the various embodiments of the present invention.

[0058] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for diagnosing anomalies in electricity meters based on multi-source consistency verification and confidence fusion, characterized in that: include, Collect multi-source data information from metering terminals in low-voltage distribution transformer areas and establish a unified index dataset; Differentiated data processing is performed on the unified index dataset, and the data is then input into a preset feature encoding network to construct a standardized feature matrix. Perform multi-source consistency verification on the standardized feature matrix and calculate the overall consistency score; The overall consistency score is used for anomaly diagnosis, and a preset threshold is used to diagnose whether there are any anomalies in the metering terminal. Real-time anomaly diagnosis is performed, parameters are updated and distributed, and anomaly diagnosis results are fed back, forming a closed loop for metering anomaly diagnosis through cloud-edge collaboration.

2. The method for diagnosing energy meter anomalies based on multi-source consistency verification and confidence fusion as described in claim 1, characterized in that: The multi-source data information includes electrical quantity data, communication metadata, and transformer area metering terminal topology information; A unified index dataset is created based on data timestamps from multiple sources.

3. The method for diagnosing energy meter anomalies based on multi-source consistency verification and confidence fusion as described in claim 2, characterized in that: The differentiated data processing includes applying differentiated data processing to address the differences in the types of multi-source data; Standardize and normalize electrical quantity data; The topology information of the metering terminal in the distribution area is converted into a binary vector.

4. The method for diagnosing energy meter anomalies based on multi-source consistency verification and confidence fusion as described in claim 3, characterized in that: The construction of the standardized feature matrix includes combining the electrical quantity data, communication metadata, and topology information of each metering terminal in the low-voltage distribution substation after data processing within a time window to form a feature vector. All metering terminal features are spliced ​​together to form a standardized feature matrix.

5. The method for diagnosing energy meter anomalies based on multi-source consistency verification and confidence fusion as described in claim 4, characterized in that: The calculation of the overall consistency score includes multi-source consistency verification, which includes physical consistency, statistical consistency, and communication element consistency. The physical consistency score A physical constraint model for the measurement results of the metering terminal is established, and the key parameters of voltage, current, and power are verified and calculated. in, Let t be the t-th physical quantity actually measured. This is a reference value calculated based on rated parameters or historical steady-state data, where n is the total number of parameters measured by the metering terminal. The statistical consistency score Based on time series stability and population distribution similarity, the mean Pearson correlation coefficient among the feature vectors of each terminal is calculated: The communication element consistency score Communication reliability is calculated based on communication metadata metrics, including latency and packet loss rate. in, For data time intervals, This represents the average drift within a time window under normal historical communication conditions. , These are the weighting coefficients for time drift and packet loss rate. , This is the feature vector after data processing, where N is the total number of metering terminals. For packet loss rate, For data time, This refers to the data collection time.

6. The method for diagnosing energy meter anomalies based on multi-source consistency verification and confidence fusion as described in claim 5, characterized in that: The step of diagnosing whether a metering terminal is abnormal by using a preset threshold includes using a threshold discrimination method as the diagnostic criterion for metering terminal abnormalities, and performing confidence fusion on the consistency scores to obtain a comprehensive consistency score. ,in, For Sigmoid mapping function, , , The multi-source consistency score weighting coefficients are assigned to scores with different consistency levels; when the overall consistency score is... Greater than the set threshold If the reading is normal, the metering terminal is diagnosed as normal; otherwise, it is diagnosed as abnormal.

7. The method for diagnosing energy meter anomalies based on multi-source consistency verification and confidence fusion as described in claim 6, characterized in that: The real-time anomaly diagnosis includes verifying and feeding back the anomaly diagnosis results. The edge verification and feedback results will affect the anomaly diagnosis threshold. After receiving the edge result feedback, the cloud calculates the model's diagnostic accuracy. Where TP represents the number of correctly identified abnormal samples, TN represents the number of correctly identified normal samples, FP represents normal samples that were misclassified as abnormal, and FN represents abnormal samples that were misclassified as normal. Implement anomaly diagnosis threshold correction strategy: in, This is a correction factor; it is negative when a case is misjudged and positive when a case is missed. To achieve the normalization rate of the target abnormal diagnosis, The current rate of normal diagnosis of abnormalities, The corrected threshold. The threshold before correction is used to construct a cloud-edge collaborative metering anomaly diagnosis closed loop. The consistency score weight coefficient and diagnosis threshold are periodically issued by the cloud. Metering personnel perform real-time anomaly diagnosis through the edge devices of the substation where the anomaly diagnosis model is deployed, and verify and provide feedback on the anomaly diagnosis results in real time. The cloud corrects the diagnosis threshold and reissues it, forming a cloud-edge collaborative metering anomaly diagnosis closed loop.

8. A power meter anomaly diagnosis system based on multi-source consistency verification and confidence fusion, employing the power meter anomaly diagnosis method based on multi-source consistency verification and confidence fusion as described in any one of claims 1 to 7, characterized in that, include: Multi-source data acquisition module, feature encoding module, consistency verification module, anomaly diagnosis module, confidence update module, edge deployment module, and cloud management module; The multi-source data acquisition module collects multi-source data information from metering terminals in low-voltage distribution transformer areas and establishes a unified index dataset. The feature encoding module performs differentiated data processing on the unified index dataset and inputs it into a preset feature encoding network to construct a standardized feature matrix. The consistency verification module performs multi-source consistency verification on the standardized feature matrix and calculates the comprehensive consistency score. The anomaly diagnosis module performs anomaly diagnosis based on the overall consistency score and diagnoses whether the metering terminal is abnormal by using a preset threshold. The edge deployment module and the cloud management module perform real-time anomaly diagnosis, update and distribute parameters, and provide feedback on anomaly diagnosis results, forming a closed loop for metering anomaly diagnosis through cloud-edge collaboration.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method for diagnosing abnormalities in an energy meter based on multi-source consistency verification and confidence fusion 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 the processor, it implements the steps of the method for diagnosing abnormalities in an energy meter based on multi-source consistency verification and confidence fusion as described in any one of claims 1 to 7.

Images