A data bias feedback-based electric energy meter data intelligent analysis method and system

Abstract

The application discloses a kind of based on data bias feedback electric energy meter data intelligent analysis method and system, it is related to electric power fault detection technical field, including obtaining electric power system information, according to electric power system information, with electric power system electrical circuit as foundation, electric energy meter is regarded as topological node, obtains electric energy meter topological relationship graph, based on electric power system information, obtains electric energy meter historical data.The application is through data difference coefficient, on the basis of considering the influence of electric energy meter association, accurately assesses the data variation range of electric energy meter, provides data basis for subsequent abnormal data identification, establishes electric energy meter association feature chain based on electric energy meter topological relationship graph to classify electric power system area, improves electric power fault detection efficiency, detects electric power fault through electric energy meter real-time data and data difference coefficient, improves the accuracy and timeliness of electric power fault detection, ensures the stability of electric power system.

Description

Technical Field

[0001] This invention relates to the field of power fault detection technology, specifically to an intelligent analysis method and system for electricity meter data based on data deviation feedback. Background Technology

[0002] Energy consumption is closely related to our production and daily life, especially the use of electrical energy. my country's power distribution network is characterized by complex line structures, diverse and changeable environments, frequent and complex faults, and a large workload for maintenance. When a fault occurs in a section of a power distribution line, it is necessary to check the location of the fault section by section, which is not only labor-intensive but also delays repair time and affects the reliability of power supply. Therefore, in actual use, electricity meters are used to monitor the status of the power system. The stability of the electricity meter's operation and the accuracy of its measurement directly affect our safe use of electricity and the power grid company's distribution management.

[0003] Currently, power fault detection suffers from several limitations, including the inability to accurately analyze electricity meter data, accurately analyze anomalies in different meters, promptly identify the location of power faults based on anomalies, and detect faults in a timely manner. Existing technologies often set fixed alarm thresholds for electricity meters to detect power faults. However, anomalies in electricity meter data are propagating; a power fault in one location can often cause anomalies in multiple meters. If fixed standards are used for power fault detection, it is impossible to locate the fault in a timely manner, affecting the stability and reliability of the power system. Analyzing each electricity meter individually not only involves a large amount of data processing but also results in low detection efficiency, impacting the timeliness of power fault detection. Summary of the Invention

[0004] To address the aforementioned technical problems, this paper provides an intelligent analysis method and system for electricity meter data based on data deviation feedback. This technical solution solves the problems mentioned in the background technology, such as the inability to accurately analyze electricity meter data, the inability to accurately analyze data anomalies in different electricity meters, the inability to promptly identify the location of power faults based on electricity meter data anomalies, and the inability to detect power faults in a timely manner. In existing technologies, fixed alarm thresholds are often set for electricity meters to detect power faults. However, electricity meter data anomalies are propagating; a power fault in one location can often cause data anomalies in multiple electricity meters. If power fault detection is performed directly using fixed standards, it is impossible to locate power faults in a timely manner, affecting the stability and reliability of the power system. If each electricity meter is analyzed individually, not only is the data processing volume large, but the detection efficiency is also low, affecting the timeliness of power fault detection.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows: A smart analysis method for electricity meter data based on data deviation feedback includes: Obtain power system information, which includes power system area information and electricity meter location information; Based on power system information, and taking the power system electrical lines as the foundation, the electricity meter is used as the topology node to obtain the electricity meter topology diagram; Based on power system information, historical data of electricity meters are obtained, including historical electrical data and historical time information for each electricity meter. Based on historical data of electricity meters and the topology diagram of electricity meters, obtain the data difference coefficient corresponding to each electricity meter; Based on the data difference coefficient, an energy meter association feature chain is established on the basis of the energy meter topology diagram to obtain the energy meter association feature information. Based on power system information and electricity meter association feature information, the regions corresponding to the electricity meters in each electricity meter association feature chain are divided into the same region to obtain power system region classification information; Based on the power system regional classification information, obtain the real-time data of the electricity meters corresponding to each region; Based on the real-time data from the electricity meter, determine whether a power failure has occurred. If so, obtain the location information of the power failure based on the real-time data from the electricity meter and the data difference coefficient. Power faults are detected based on their location information and power system regional classification information.

[0006] Preferably, the step of obtaining the data difference coefficient for each electricity meter based on historical electricity meter data and the electricity meter topology diagram specifically includes: Based on the historical data of the electricity meter, the historical data of the electricity meter corresponding to the same historical time is divided into the same dataset to obtain the historical dataset of the electricity meter. Based on power system information, electrical data thresholds for electricity meters are obtained. These electrical data thresholds represent the maximum and minimum electrical data values ​​of each electricity meter during normal operation in the power system. Based on the electrical data thresholds of the electricity meter and the historical data set of the electricity meter, the historical electrical data of each electricity meter that exceeds the electrical data threshold is regarded as the historical abnormal data of that electricity meter. Based on historical abnormal data, the historical dataset of electricity meters corresponding to the historical abnormal data is used as the abnormal dataset of electricity meters. Based on the topology diagram of the electricity meter, the topology node corresponding to the electricity meter is taken as the feature node, and the topology node associated with the feature node is taken as the associated topology node corresponding to the feature node. Based on the associated topology node corresponding to each feature node, the associated topology node pointed to by the feature node is taken as the influence topology node of the feature node, and the associated topology node pointing to the feature node is taken as the starting topology node of the feature node. Based on the feature nodes and the corresponding influencing topology nodes, and using the electricity meter anomaly dataset as a basis, obtain the synchronous anomaly dataset corresponding to each influencing topology node. The synchronous anomaly dataset indicates that the influencing topology node and the feature node both have historical anomaly data at the same historical time. The number of synchronization anomaly datasets is taken as the number of anomaly synchronizations that affect each topology node and feature node; The number of abnormal data sets of electricity meters corresponding to each feature node is used as the target number of abnormalities. The ratio of the number of abnormal synchronizations affecting topology nodes and feature nodes to the target number of abnormal synchronizations is used as the abnormal synchronization coefficient of the affected topology node. Based on the abnormal data set of electricity meters, the threshold of the abnormal synchronization coefficient is obtained; The data difference coefficient is obtained based on the abnormal synchronization coefficient threshold.

[0007] Preferably, obtaining the abnormal synchronization coefficient threshold based on the abnormal data set of electricity meters specifically includes: Based on the abnormal electricity meter dataset, obtain the electricity meter information corresponding to the historical abnormal data; Use the electricity meters corresponding to historical abnormal data as target electricity meters and obtain the topology node information corresponding to the target electricity meters; Based on the topology node information corresponding to the target energy meter, obtain the associated topology node information of the topology node corresponding to the target energy meter; Target energy meters whose associated topology nodes do not contain the starting topology node are designated as feature energy meters. If all target energy meters have an initial topology node, then the target energy meter with the smallest number of initial topology nodes is selected as the feature energy meter. Based on the characteristic energy meter, obtain the abnormal synchronization coefficient of the topology node corresponding to the characteristic energy meter; The abnormal data set of the electricity meter corresponding to the historical abnormal data of the characteristic electricity meter is used as the characteristic abnormal data set. Based on the feature anomaly dataset, the topology node of the electricity meter corresponding to the historical anomaly data in each feature anomaly dataset is taken as the anomaly topology node of that feature anomaly dataset. Based on the influencing topology nodes corresponding to the characteristic energy meters, the abnormal topology node with the largest number of abnormal topology nodes between it and the influencing topology node is taken as the influencing calibration node corresponding to the influencing topology node. The number of abnormal topology nodes between each affected topology node and its corresponding affected calibration node is taken as the maximum depth value of that affected topology node. Based on the abnormal synchronization coefficient and maximum depth value of the affected topology nodes corresponding to the characteristic energy meters, obtain the abnormal synchronization characteristic index corresponding to each characteristic energy meter; The average of the abnormal synchronization characteristic indexes of all characteristic energy meters is used as the threshold for the abnormal synchronization coefficient. Specifically, the abnormal synchronization characteristic index is: In the formula, This is an abnormal synchronization characteristic index. Let be the abnormal synchronization coefficient of the k-th influencing topology node corresponding to the characteristic energy meter. The influence coefficient of the k-th influencing topology node corresponding to the characteristic energy meter. This represents the maximum depth value of the k-th influencing topology node corresponding to the characteristic energy meter. , representing the attenuation factor.

[0008] Preferably, obtaining the data difference coefficient based on the abnormal synchronization coefficient threshold specifically includes: Based on the starting topology node corresponding to each feature node, the starting topology node with an abnormal synchronization coefficient higher than the abnormal synchronization coefficient threshold is taken as the strongly correlated node corresponding to that feature node. Based on the feature nodes and the strongly correlated nodes corresponding to each feature node, obtain the abnormal deviation degree corresponding to the feature nodes and the strongly correlated nodes; Based on the abnormal deviation degree corresponding to the feature node and the strongly correlated node, and taking the abnormal synchronization coefficient corresponding to the strongly correlated node as the basis, the data difference coefficient corresponding to each feature node is obtained, that is, the data difference coefficient of the electricity meter corresponding to the feature node. Specifically, the abnormal deviation is: In the formula, F represents the degree of abnormal deviation. For the i-th historical outlier, For reference electrical data, This represents the maximum value of the electrical data. Let n be the minimum value of electrical data, and n be the total number of historical anomalies. The specific data difference coefficient is: In the formula, This represents the data difference coefficient of the x-th electricity meter. This indicates the degree of abnormal deviation of the x-th electricity meter. This represents the abnormal deviation of the j-th strongly correlated node of the x-th electricity meter. This represents the abnormal synchronization coefficient of the j-th strongly correlated node of the x-th energy meter. The total number of strongly correlated nodes for the x-th electricity meter Let represent the transmission coefficient of the j-th strongly correlated node of the x-th energy meter.

[0009] Preferably, the step of dividing the regions corresponding to the electricity meters in each electricity meter association feature chain into the same region based on power system information and electricity meter association feature information to obtain power system region classification information specifically includes: S501: Based on the historical data of the electricity meter and the electrical data threshold of the electricity meter, the maximum value of the difference between the historical electrical data and the benchmark electrical data corresponding to each electricity meter is taken as the electrical identification amplitude corresponding to the electricity meter. S502: Based on the electricity meter topology graph, take any electricity meter as the starting node of the path and obtain the influencing topology node corresponding to the starting node in the electricity meter topology graph. S503: The ratio of the data difference coefficient corresponding to the topology node to the data difference coefficient corresponding to the path starting node is used as the similarity difference coefficient; S504: Based on the electrical identification amplitude and the similarity difference coefficient corresponding to the affected topology node, determine whether the affected topology node meets the association characteristics; If the ratio of the electrical identification amplitude of the influencing topology node to the electrical identification amplitude of the path starting node exceeds the similarity difference coefficient, then the influencing topology node does not meet the association characteristics; if the ratio of the electrical identification amplitude of the influencing topology node to the electrical identification amplitude of the path starting node does not exceed the similarity difference coefficient, then the influencing topology node is regarded as the first associated node. S505: Based on the first-level associated nodes, obtain the influence topology nodes corresponding to each first-level associated node. Repeat steps S503-S504 and take the influence topology nodes that meet the association characteristics as the second-level associated nodes. S506: Repeat step S505 until no new associated nodes can be added, forming an energy meter associated feature chain with the path start node as the root, then proceed to S507. S507: Based on the electricity meter topology diagram, remove the topology nodes in the electricity meter association feature chain, and repeat steps S502-506 for the remaining topology nodes until all electricity meters are traversed to obtain the electricity meter association feature information. S508: Based on the electricity meter association feature information, divide the regions corresponding to all electricity meters in each electricity meter association feature chain into the same region to obtain power system region classification information.

[0010] Preferably, the step of determining whether a power failure has occurred based on real-time data from the electricity meter specifically includes: The difference between the real-time data of the electricity meter and the reference electrical data is used as the real-time deviation value for each electricity meter. Based on the electrical data threshold of the electricity meter, the difference between the maximum and minimum electrical data values ​​of each electricity meter during normal operation in the power system is used as the benchmark deviation value. The ratio of the real-time deviation value to the reference deviation value is used as the real-time deviation coefficient of the energy meter. Based on the real-time deviation coefficient and data difference coefficient of each electricity meter, it is determined whether the electricity meter is abnormal. If the real-time deviation coefficient of the electricity meter exceeds the data difference coefficient, the electricity meter is regarded as a faulty electricity meter and the faulty electricity meter information is obtained. Based on the associated feature information of electricity meters and the information of faulty and abnormal electricity meters, the associated feature chains of all suspected faulty associated meters are extracted as abnormal associated feature chains. The ratio of the number of faulty energy meters in the anomaly association feature chain to the total number of energy meters in the anomaly association feature chain is used as the regional anomaly coefficient. Based on the power system regional classification information, the power coefficient region to which the anomaly association feature chain corresponding to the maximum value of the regional anomaly coefficient belongs is taken as the fault region. Based on the data difference coefficient, the product of the real-time deviation coefficient of the faulty energy meter and the data difference coefficient in the abnormal association feature chain is used as the fault identification coefficient. The faulty energy meter corresponding to the maximum value of the fault identification coefficient is used as the location of the power fault, and the power fault location information is obtained.

[0011] Furthermore, a smart data analysis system for electricity meters based on data deviation feedback is proposed to implement the detection method described above, including: The main control module is used to establish an energy meter association feature chain based on the energy meter topology diagram according to the data difference coefficient, obtain energy meter association feature information, divide the area corresponding to each energy meter in the energy meter association feature chain into the same area according to the power system information and the energy meter association feature information, obtain power system area classification information, determine whether the energy meter is abnormal according to the real-time deviation coefficient and data difference coefficient of each energy meter, extract the association feature chain of all suspected faulty associated tables as the abnormal association feature chain according to the energy meter association feature information and the faulty energy meter information, take the product of the energy meter real-time deviation coefficient and the data difference coefficient of the faulty energy meter in the abnormal association feature chain as the fault identification coefficient, take the faulty energy meter corresponding to the maximum value of the fault identification coefficient as the power fault location, obtain power fault location information, and detect power faults according to the power fault location information and the power system area classification information. The information acquisition module is used to acquire power system information, including power system regional information and electricity meter location information. Based on the power system information, the module uses electricity meters as topology nodes based on the power system electrical lines to acquire an electricity meter topology diagram. Based on the power system information, the module acquires historical data of the electricity meters, including historical electrical data and historical time information corresponding to each electricity meter. Based on the power system regional classification information, the module acquires real-time data of the electricity meters corresponding to each region. The evaluation module is used to divide the historical data of electricity meters corresponding to the same historical time into the same dataset based on the historical time, obtain the historical dataset of electricity meters, obtain the abnormal synchronization coefficient based on the historical dataset of electricity meters, obtain the abnormal deviation degree corresponding to the feature node and the strongly correlated node corresponding to each feature node based on the feature node and the strongly correlated node corresponding to each feature node, and obtain the data difference coefficient corresponding to each feature node based on the abnormal synchronization coefficient corresponding to the strongly correlated node, that is, the data difference coefficient of the electricity meter corresponding to the feature node. The display module interacts with the main control module and is used to output and display the topology diagram of the electricity meter, the data difference coefficient, the electricity meter association characteristic information, the power system area classification information, the real-time data of the electricity meter, and the power fault location information.

[0012] Optionally, the main control module specifically includes: The control unit is used to determine whether an energy meter is abnormal based on the real-time deviation coefficient and data difference coefficient of each energy meter; extract the association feature chain of all suspected fault association tables as an abnormal association feature chain based on the energy meter association feature information and faulty energy meter information; take the product of the real-time deviation coefficient and the data difference coefficient of the faulty energy meter in the abnormal association feature chain as the fault identification coefficient based on the data difference coefficient; take the faulty energy meter corresponding to the maximum value of the fault identification coefficient as the power fault location; obtain the power fault location information; and detect the power fault based on the power fault location information and the power system area classification information. An information receiving unit, which interacts with the information acquisition module and the evaluation module, is used to receive data and transmit it to the region division unit. The region division unit is used to establish an energy meter association feature chain based on the energy meter topology diagram according to the data difference coefficient, obtain energy meter association feature information, and divide the region corresponding to each energy meter in the energy meter association feature chain into the same region according to the power system information and the energy meter association feature information, thereby obtaining power system region classification information.

[0013] Optionally, the information acquisition module specifically includes: The first acquisition unit is used to acquire power system information, which includes power system area information and electricity meter location information. Based on the power system information, the electricity meter is used as a topology node based on the power system electrical lines to acquire an electricity meter topology diagram. The second acquisition unit is used to acquire historical data of electricity meters based on power system information. The historical data of electricity meters includes historical electrical data and historical time information corresponding to each electricity meter. According to the power system regional classification information, the real-time data of electricity meters corresponding to each region is acquired.

[0014] Optionally, the evaluation module specifically includes: The first evaluation unit is used to divide the historical data of the electricity meter corresponding to the same historical time into the same dataset based on the historical time, obtain the historical dataset of the electricity meter, and obtain the abnormal synchronization coefficient based on the historical dataset of the electricity meter. The second evaluation unit is used to obtain the abnormal deviation degree corresponding to the feature node and the strongly correlated node corresponding to each feature node based on the feature node and the strongly correlated node corresponding to each feature node, and to obtain the data difference coefficient corresponding to each feature node based on the abnormal synchronization coefficient corresponding to the strongly correlated node, that is, the data difference coefficient of the electricity meter corresponding to the feature node.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention proposes an intelligent analysis method and system for electricity meter data based on data deviation feedback. By using a data difference coefficient and considering the influence of electricity meter correlation, the method accurately assesses the magnitude of data fluctuations in electricity meters, providing a data foundation for subsequent abnormal data identification. By establishing an electricity meter correlation feature chain based on the electricity meter topology diagram, the method classifies power system regions, improving the efficiency of power fault detection. By using real-time electricity meter data and the data difference coefficient, the method detects power faults, improving the accuracy and timeliness of power fault detection and ensuring the stability of the power system. Attached Figure Description

[0016] Figure 1 This is a flowchart of an intelligent analysis method for electricity meter data based on data deviation feedback proposed in this invention. Figure 2 This is a flowchart of the process for obtaining the abnormal synchronization coefficient in this invention; Figure 3 This is a flowchart of the process for obtaining the abnormal synchronization coefficient threshold in this invention; Figure 4 This is a flowchart of the process for obtaining regional classification information of the power system in this invention; Figure 5 This is a block diagram of an intelligent analysis system for electricity meter data based on data deviation feedback proposed in this invention. Detailed Implementation

[0017] The following description is intended to disclose the invention and enable those skilled in the art to implement it. The preferred embodiments described below are merely examples, and other obvious variations will occur to those skilled in the art.

[0018] Reference Figure 1 - Figure 4 As shown in the figure, an intelligent analysis method for electricity meter data based on data deviation feedback in an embodiment of the present invention includes: Obtain power system information, which includes power system area information and electricity meter location information; Based on power system information, and taking the power system electrical lines as the foundation, the electricity meter is used as the topology node to obtain the electricity meter topology diagram; Based on power system information, historical data of electricity meters are obtained, including historical electrical data and historical time information for each electricity meter. Based on historical data of electricity meters and the topology diagram of electricity meters, obtain the data difference coefficient corresponding to each electricity meter; Specifically, based on historical data from electricity meters and the electricity meter topology diagram, the data difference coefficient for each electricity meter is obtained, including: Based on the historical data of the electricity meter, the historical data of the electricity meter corresponding to the same historical time is divided into the same dataset to obtain the historical dataset of the electricity meter. Based on power system information, electrical data thresholds for electricity meters are obtained. These electrical data thresholds represent the maximum and minimum electrical data values ​​of each electricity meter during normal operation in the power system. Based on the electrical data thresholds of the electricity meter and the historical data set of the electricity meter, the historical electrical data of each electricity meter that exceeds the electrical data threshold is regarded as the historical abnormal data of that electricity meter. Based on historical abnormal data, the historical dataset of electricity meters corresponding to the historical abnormal data is used as the abnormal dataset of electricity meters. Based on the topology diagram of the electricity meter, the topology node corresponding to the electricity meter is taken as the feature node, and the topology node associated with the feature node is taken as the associated topology node corresponding to the feature node. Based on the associated topology node corresponding to each feature node, the associated topology node pointed to by the feature node is taken as the influence topology node of the feature node, and the associated topology node pointing to the feature node is taken as the starting topology node of the feature node. Based on the feature nodes and the corresponding influencing topology nodes, and using the electricity meter anomaly dataset as a basis, obtain the synchronous anomaly dataset corresponding to each influencing topology node. The synchronous anomaly dataset indicates that the influencing topology node and the feature node both have historical anomaly data at the same historical time. The number of synchronization anomaly datasets is taken as the number of anomaly synchronizations that affect each topology node and feature node; The number of abnormal data sets of electricity meters corresponding to each feature node is used as the target number of abnormalities. The ratio of the number of abnormal synchronizations affecting topology nodes and feature nodes to the target number of abnormal synchronizations is used as the abnormal synchronization coefficient of the affected topology node. Based on the abnormal data set of electricity meters, the threshold of the abnormal synchronization coefficient is obtained; The data difference coefficient is obtained based on the abnormal synchronization coefficient threshold.

[0019] In this scheme, the influence topology nodes (downstream related nodes) and the starting topology nodes (upstream related nodes) of a feature node (target energy meter) are distinguished by the energy meter topology diagram. By combining the synchronization anomaly dataset to analyze the number of abnormal synchronizations, the propagation path of the fault in the electrical topology can be clearly identified. For example, when an energy meter (feature node) malfunctions, the abnormal synchronization coefficient of the downstream influence topology nodes is calculated, providing an analytical basis for the correlation impact on the energy meter data and providing data support for defining the fault range. Historical abnormal data is filtered based on the energy meter electrical data threshold, and the abnormal correlation of related nodes is quantified by the abnormal synchronization coefficient. By combining the historical data of the energy meter with the topology relationship, the scattered energy meter data is transformed into a collaborative analysis object with spatial correlation, avoiding the limitations of single energy meter data. Key related nodes can be quickly identified in the early stage of a fault, shortening the response time of fault detection and providing direction for subsequent accurate troubleshooting. The calculation of the abnormal synchronization coefficient combines the topology correlation with the abnormal synchronization of the time dimension, forming a quantifiable "difference index". For example, a high synchronization coefficient between the starting topology node (upstream node) and the characteristic node may indicate that the fault originates from the upstream line; while a low synchronization coefficient affecting the topology node (downstream node) may point to a fault in the characteristic node itself or a local branch, providing clear quantitative clues for fault tracing.

[0020] Understandably, existing technologies often analyze anomalies in electricity meter data by examining the fluctuation range of the meter readings. However, in reality, anomalies in electricity meter data often interact with each other, causing even meters that were not initially abnormal to show anomalies. Alternatively, two anomalies can have their effects on other meters cumulatively, affecting the accuracy of electricity meter data analysis and making it impossible to accurately assess the condition of the electricity meter data.

[0021] Specifically, based on the abnormal data set of electricity meters, the threshold of the abnormal synchronization coefficient is obtained, including: Based on the abnormal electricity meter dataset, obtain the electricity meter information corresponding to the historical abnormal data; Use the electricity meters corresponding to historical abnormal data as target electricity meters and obtain the topology node information corresponding to the target electricity meters; Based on the topology node information corresponding to the target energy meter, obtain the associated topology node information of the topology node corresponding to the target energy meter; Target energy meters whose associated topology nodes do not contain the starting topology node are designated as feature energy meters. If all target energy meters have an initial topology node, then the target energy meter with the smallest number of initial topology nodes is selected as the feature energy meter. Based on the characteristic energy meter, obtain the abnormal synchronization coefficient of the topology node corresponding to the characteristic energy meter; The abnormal data set of the electricity meter corresponding to the historical abnormal data of the characteristic electricity meter is used as the characteristic abnormal data set. Based on the feature anomaly dataset, the topology node of the electricity meter corresponding to the historical anomaly data in each feature anomaly dataset is taken as the anomaly topology node of that feature anomaly dataset. Based on the influencing topology nodes corresponding to the characteristic energy meters, the abnormal topology node with the largest number of abnormal topology nodes between it and the influencing topology node is taken as the influencing calibration node corresponding to the influencing topology node. The number of abnormal topology nodes between each affected topology node and its corresponding affected calibration node is taken as the maximum depth value of that affected topology node. Based on the abnormal synchronization coefficient and maximum depth value of the affected topology nodes corresponding to the characteristic energy meters, obtain the abnormal synchronization characteristic index corresponding to each characteristic energy meter; The average of the abnormal synchronization characteristic indexes of all characteristic energy meters is used as the threshold for the abnormal synchronization coefficient. Specifically, the abnormal synchronization characteristic index is: In the formula, This is an abnormal synchronization characteristic index. Let be the abnormal synchronization coefficient of the k-th influencing topology node corresponding to the characteristic energy meter. The influence coefficient of the k-th influencing topology node corresponding to the characteristic energy meter. This represents the maximum depth value of the k-th influencing topology node corresponding to the characteristic energy meter. , representing the attenuation factor.

[0022] In this scheme, the topological relationships of electricity meters in the power system are complex, with multiple starting topological nodes intersecting and influencing each other. By screening characteristic electricity meters with "no starting topological node (or the fewest starting nodes)," the scheme focuses on relatively independent nodes in the fault propagation chain that are less affected by upstream interference. This avoids the confusion in anomaly coefficient calculation caused by multi-source topological interference, laying the foundation for subsequent accurate analysis. For example, in complex distribution area topologies, characteristic electricity meters located at the end of branches are selected, and their anomalies are more easily traced back to local faults. The anomaly data of these characteristic electricity meters are closer to "primary anomalies" (rather than secondary anomalies affected by upstream fault propagation). Based on the analysis of anomaly synchronization coefficients, the initial trigger point of the fault can be located more accurately, reducing misjudgments of "propagation interference" during fault tracing. When power faults (such as short circuits and overloads) propagate in the topology, the range and intensity of their impact decrease with node distance. By introducing depth values ​​and influence coefficients, the anomaly synchronization coefficient threshold can reflect this attenuation law, making fault detection more consistent with the actual physical process and improving the adaptability of the threshold to fault characteristics. The anomaly synchronization coefficient threshold serves as the core judgment standard for fault detection, providing a quantitative basis for subsequent fault identification, location, and early warning.

[0023] It is important to note the attenuation factor ( The core function of this algorithm is to adjust the weight of the "maximum depth value affecting topology nodes" in the calculation of the abnormal synchronization characteristic index. In power systems, the impact of faults (such as short circuits and overloads) propagates along electrical lines, but its intensity naturally weakens as the number of node intervals (maximum depth value) increases. In similar algorithms for power system fault location and topology correlation analysis, the value of the "distance attenuation factor" generally follows the middle range of 0.3 to 0.7. In this scheme, we take... As a decay factor, it ensures that the abnormal synchronization characteristic index can truly reflect the node association strength.

[0024] Specifically, based on the abnormal synchronization coefficient threshold, the data difference coefficient is obtained, including: Based on the starting topology node corresponding to each feature node, the starting topology node with an abnormal synchronization coefficient higher than the abnormal synchronization coefficient threshold is taken as the strongly correlated node corresponding to that feature node. Based on the feature nodes and the strongly correlated nodes corresponding to each feature node, obtain the abnormal deviation degree corresponding to the feature nodes and the strongly correlated nodes; Based on the abnormal deviation degree corresponding to the feature node and the strongly correlated node, and taking the abnormal synchronization coefficient corresponding to the strongly correlated node as the basis, the data difference coefficient corresponding to each feature node is obtained, that is, the data difference coefficient of the electricity meter corresponding to the feature node. Specifically, the abnormal deviation is: In the formula, F represents the degree of abnormal deviation. For the i-th historical outlier, For reference electrical data, This represents the maximum value of the electrical data. Let n be the minimum value of electrical data, and n be the total number of historical anomalies. The specific collaborative deviation coefficient is: In the formula, This represents the coordination deviation coefficient of the x-th electricity meter. This indicates the degree of abnormal deviation of the x-th electricity meter. This represents the abnormal deviation of the j-th strongly correlated node of the x-th electricity meter. This represents the abnormal synchronization coefficient of the j-th strongly correlated node of the x-th energy meter. The total number of strongly correlated nodes for the x-th electricity meter Let represent the transmission coefficient of the j-th strongly correlated node of the x-th energy meter.

[0025] This scheme filters strongly correlated nodes (those affecting the topology nodes above the threshold) by using an anomaly synchronization coefficient threshold, focusing on the nodes with the "closest correlation" in fault propagation. For example, when a certain electricity meter (the starting topology node) malfunctions, downstream nodes may also malfunction synchronously with the starting topology node due to a line short circuit (synchronization coefficient exceeding the threshold). The starting topology node is then identified as a strongly correlated node of the downstream node (the feature node). This filtering allows fault analysis to focus from "broad topology correlation" to "core fault chains," reducing interference from irrelevant nodes and improving fault location accuracy. Power faults (such as short circuits and ground faults) have "topology conduction" (faults propagate along the line, and associated nodes become synchronously abnormal) and "historical tendency" (older equipment has a higher frequency of malfunctions). This scheme captures "conductivity" through the collaborative deviation coefficient and "tendency" through the data anomaly frequency, allowing the data difference coefficient to adapt to the characteristics of power faults. This upgrades electricity meter data from "isolated outliers" to "topology collaborative fault characteristics," providing physically interpretable and multi-dimensional fusion-based discrimination criteria for intelligent power fault detection, significantly improving the accuracy and intelligence of fault detection.

[0026] In this embodiment, a synchronization anomaly dataset of feature nodes and their corresponding strongly correlated nodes is extracted. This dataset represents historical anomalies that occurred at the same historical time for both the feature node and its corresponding strongly correlated node. Each sub-item in the synchronization anomaly dataset represents a synchronous anomaly that occurred between the corresponding feature node and its corresponding strongly correlated node at a certain moment. This synchronization anomaly dataset is then passed into the transfer function, i.e.: In the formula, Indicates the first in the synchronization anomaly dataset The abnormal deviation of the feature nodes corresponding to each sub-item Indicates the first in the synchronization anomaly dataset Historical anomaly data for the feature nodes corresponding to each sub-item Represents a constant term. Indicates the first in the synchronization anomaly dataset Among the sub-items, the one with the abnormal synchronization with the feature node is the Abnormal deviation of strongly correlated nodes For the first The transmission coefficients of strongly correlated nodes to feature nodes.

[0027] It is important to note that the first The calculation method for the abnormal deviation of strongly correlated nodes is the same as that for feature nodes. Substitute each item in the synchronous abnormal dataset into the transfer function, construct a simultaneous equation to fit the data in the synchronous abnormal dataset, and obtain the transfer coefficient between the feature node and each strongly correlated node. Based on the data difference coefficient, an energy meter association feature chain is established on the basis of the energy meter topology diagram to obtain the energy meter association feature information. Based on power system information and electricity meter association feature information, the regions corresponding to the electricity meters in each electricity meter association feature chain are divided into the same region to obtain power system region classification information; Specifically, based on power system information and electricity meter association feature information, the regions corresponding to each electricity meter in each electricity meter association feature chain are divided into the same region to obtain power system region classification information, including: S501: Based on the historical data of the electricity meter and the electrical data threshold of the electricity meter, the maximum value of the difference between the historical electrical data and the benchmark electrical data corresponding to each electricity meter is taken as the electrical identification amplitude corresponding to the electricity meter. S502: Based on the electricity meter topology graph, take any electricity meter as the starting node of the path and obtain the influencing topology node corresponding to the starting node in the electricity meter topology graph. S503: The ratio of the data difference coefficient corresponding to the topology node to the data difference coefficient corresponding to the path starting node is used as the similarity difference coefficient; S504: Based on the electrical identification amplitude and the similarity difference coefficient corresponding to the affected topology node, determine whether the affected topology node meets the association characteristics; If the ratio of the electrical identification amplitude of the influencing topology node to the electrical identification amplitude of the path starting node exceeds the similarity difference coefficient, then the influencing topology node does not meet the association characteristics; if the ratio of the electrical identification amplitude of the influencing topology node to the electrical identification amplitude of the path starting node does not exceed the similarity difference coefficient, then the influencing topology node is regarded as the first associated node. S505: Based on the first-level associated nodes, obtain the influence topology nodes corresponding to each first-level associated node. Repeat steps S503-S504 and take the influence topology nodes that meet the association characteristics as the second-level associated nodes. S506: Repeat step S505 until no new associated nodes can be added, forming an energy meter associated feature chain with the path start node as the root, then proceed to S507. S507: Based on the electricity meter topology diagram, remove the topology nodes in the electricity meter association feature chain, and repeat steps S502-506 for the remaining topology nodes until all electricity meters are traversed to obtain the electricity meter association feature information. S508: Based on the electricity meter association feature information, divide the regions corresponding to all electricity meters in each electricity meter association feature chain into the same region to obtain power system region classification information.

[0028] In this scheme, associated nodes are screened layer by layer based on electrical identification amplitude (the difference between historical data of the electricity meter and the benchmark) and similarity difference coefficient (the ratio of data difference coefficients). For example, if an electricity meter at the head end of a line is abnormal, the nodes associated with the fault propagation (first and second-level nodes that meet the association characteristics) are accurately identified by calculating the ratio of electrical identification amplitude and data difference coefficient between the downstream node and the head end. An association feature chain from the fault source to the affected node is constructed, clearly presenting the fault propagation path and solving the problem of "fuzzy associated nodes" in traditional topology analysis. Power faults (such as short circuits and overloads) propagate along the line topology. This method uses "electrical identification amplitude + similarity difference coefficient" dual-dimensional verification to ensure that the association feature chain is consistent with the actual fault propagation law. For example, in a short circuit fault, the abnormal current will propagate from the fault point to the upstream and downstream of the line. The association feature chain can capture this propagation, making the fault analysis fit the physical characteristics of the power system. The corresponding areas of the electricity meters in the association feature chain are classified into the same area, upgrading the power system fault detection from "scattered electricity meter monitoring" to "regional collaborative management". When an energy meter in a certain area malfunctions, the source of the fault (the starting node of the chain) can be quickly located based on the regional correlation feature chain, and the affected range (subsequent nodes of the chain) can be predicted, thereby improving the targeting and efficiency of fault handling and avoiding the waste of resources caused by indiscriminate investigation.

[0029] It should be noted that in step S505, the selection of the second-level associated node is achieved by using the electrical identification amplitude of the topology node corresponding to the first-level associated node and the electrical identification amplitude of the first-level associated node. At this time, the first-level associated node has been incorporated into the electricity meter associated feature chain. The electrical identification amplitude of the first-level associated node is the product of the path starting node and the similarity difference coefficient. That is, the electrical identification amplitude of each level of associated node is the product of the previous level of associated node and the similarity difference coefficient. In step S507, the topology nodes in the electricity meter associated feature chain are removed according to the electricity meter topology relationship diagram. Specifically, the topology nodes that have formed the electricity meter associated feature chain are removed.

[0030] Based on the power system regional classification information, obtain the real-time data of the electricity meters corresponding to each region; Based on the real-time data from the electricity meter, determine whether a power failure has occurred. If so, obtain the location information of the power failure based on the real-time data from the electricity meter and the data difference coefficient. Specifically, based on real-time data from the electricity meter, it is determined whether a power outage has occurred, including: The difference between the real-time data of the electricity meter and the reference electrical data is used as the real-time deviation value for each electricity meter. Based on the electrical data threshold of the electricity meter, the difference between the maximum and minimum electrical data values ​​of each electricity meter during normal operation in the power system is used as the benchmark deviation value. The ratio of the real-time deviation value to the reference deviation value is used as the real-time deviation coefficient of the energy meter. Based on the real-time deviation coefficient and data difference coefficient of each electricity meter, it is determined whether the electricity meter is abnormal. If the real-time deviation coefficient of the electricity meter exceeds the data difference coefficient, the electricity meter is regarded as a faulty electricity meter and the faulty electricity meter information is obtained. Based on the associated feature information of electricity meters and the information of faulty and abnormal electricity meters, the associated feature chains of all suspected faulty associated meters are extracted as abnormal associated feature chains. The ratio of the number of faulty energy meters in the anomaly association feature chain to the total number of energy meters in the anomaly association feature chain is used as the regional anomaly coefficient. Based on the power system regional classification information, the power coefficient region to which the anomaly association feature chain corresponding to the maximum value of the regional anomaly coefficient belongs is taken as the fault region. Based on the data difference coefficient, the product of the real-time deviation coefficient of the faulty energy meter and the data difference coefficient in the abnormal association feature chain is used as the fault identification coefficient. The faulty energy meter corresponding to the maximum value of the fault identification coefficient is used as the location of the power fault, and the power fault location information is obtained.

[0031] This solution achieves accurate identification and location of power faults through collaborative analysis of real-time data and historical features. It replaces traditional fixed threshold judgments with a comparison of "real-time deviation coefficient (the degree of deviation between real-time data and the benchmark)" and "data difference coefficient (historical collaborative anomaly characteristics)." For example, if a power meter's historical data shows a large normal fluctuation range (high data difference coefficient), even if the real-time deviation value is slightly higher, as long as it does not exceed the data difference coefficient, it will not be misjudged as an anomaly. This solves the problem that fixed thresholds are difficult to adapt to the characteristics of different power meters. By incorporating historical collaborative features (such as topological correlation anomaly patterns) into the data difference coefficient, if a power meter's real-time deviation coefficient exceeds the data difference coefficient, it indicates that its anomaly not only exceeds the standard but also conforms to the collaborative characteristics of historical faults (such as consistency with the anomaly patterns of associated nodes), making it more likely a genuine fault. This reduces false alarms caused by instantaneous interference (such as short-term fluctuations caused by lightning strikes).

[0032] Power faults are detected based on their location information and power system regional classification information.

[0033] Reference Figure 5 As shown, further, combining the above-mentioned intelligent analysis method for electricity meter data based on data deviation feedback, an intelligent analysis system for electricity meter data based on data deviation feedback is proposed, including: The main control module is used to establish an energy meter association feature chain based on the energy meter topology diagram according to the data difference coefficient, obtain energy meter association feature information, divide the area corresponding to each energy meter in the energy meter association feature chain into the same area according to the power system information and the energy meter association feature information, obtain power system area classification information, determine whether the energy meter is abnormal according to the real-time deviation coefficient and data difference coefficient of each energy meter, extract the association feature chain of all suspected faulty associated tables as the abnormal association feature chain according to the energy meter association feature information and the faulty energy meter information, take the product of the energy meter real-time deviation coefficient and the data difference coefficient of the faulty energy meter in the abnormal association feature chain as the fault identification coefficient, take the faulty energy meter corresponding to the maximum value of the fault identification coefficient as the power fault location, obtain power fault location information, and detect power faults according to the power fault location information and the power system area classification information. The information acquisition module is used to acquire power system information, including power system regional information and electricity meter location information. Based on the power system information, the module uses electricity meters as topology nodes based on the power system electrical lines to acquire an electricity meter topology diagram. Based on the power system information, the module acquires historical data of the electricity meters, including historical electrical data and historical time information corresponding to each electricity meter. Based on the power system regional classification information, the module acquires real-time data of the electricity meters corresponding to each region. The evaluation module is used to divide the historical data of electricity meters corresponding to the same historical time into the same dataset based on the historical time, obtain the historical dataset of electricity meters, obtain the abnormal synchronization coefficient based on the historical dataset of electricity meters, obtain the abnormal deviation degree corresponding to the feature node and the strongly correlated node corresponding to each feature node based on the feature node and the strongly correlated node corresponding to each feature node, and obtain the data difference coefficient corresponding to each feature node, i.e., the data difference coefficient of the electricity meter corresponding to the feature node, based on the abnormal synchronization coefficient corresponding to the strongly correlated node and the abnormal deviation degree corresponding to the feature node and the strongly correlated node. The display module interacts with the main control module and is used to output and display the topology diagram of the electricity meter, the data difference coefficient, the electricity meter association characteristic information, the power system area classification information, the real-time data of the electricity meter, and the power fault location information.

[0034] The main control module specifically includes: The control unit is used to determine whether an energy meter is abnormal based on the real-time deviation coefficient and data difference coefficient of each energy meter; extract the association feature chain of all suspected fault association tables as an abnormal association feature chain based on the energy meter association feature information and faulty energy meter information; take the product of the real-time deviation coefficient and the data difference coefficient of the faulty energy meter in the abnormal association feature chain as the fault identification coefficient based on the data difference coefficient; take the faulty energy meter corresponding to the maximum value of the fault identification coefficient as the power fault location; obtain the power fault location information; and detect the power fault based on the power fault location information and the power system area classification information. An information receiving unit, which interacts with the information acquisition module and the evaluation module, is used to receive data and transmit it to the region division unit. The region division unit is used to establish an energy meter association feature chain based on the energy meter topology diagram according to the data difference coefficient, obtain energy meter association feature information, and divide the region corresponding to each energy meter in the energy meter association feature chain into the same region according to the power system information and the energy meter association feature information, thereby obtaining power system region classification information.

[0035] The information acquisition module specifically includes: The first acquisition unit is used to acquire power system information, which includes power system area information and electricity meter location information. Based on the power system information, the electricity meter is used as a topology node based on the power system electrical lines to acquire an electricity meter topology diagram. The second acquisition unit is used to acquire historical data of electricity meters based on power system information. The historical data of electricity meters includes historical electrical data and historical time information corresponding to each electricity meter. According to the power system regional classification information, the real-time data of electricity meters corresponding to each region is acquired.

[0036] The evaluation module specifically includes: The first evaluation unit is used to divide the historical data of the electricity meter corresponding to the same historical time into the same dataset based on the historical time, obtain the historical dataset of the electricity meter, and obtain the abnormal synchronization coefficient based on the historical dataset of the electricity meter. The second evaluation unit is used to obtain the abnormal deviation degree corresponding to the feature node and the strongly correlated node corresponding to each feature node based on the feature node and the strongly correlated node corresponding to each feature node, and to obtain the data difference coefficient corresponding to each feature node based on the abnormal synchronization coefficient corresponding to the strongly correlated node, that is, the data difference coefficient of the electricity meter corresponding to the feature node.

[0037] In summary, the advantages of this invention are as follows: By using historical data from electricity meters and a topological diagram of the electricity meters, a data difference coefficient is obtained for each electricity meter. Based on this coefficient, and considering the influence of meter correlations, the magnitude of data fluctuations is accurately assessed, providing a data foundation for subsequent anomaly identification. Furthermore, by using the data difference coefficient and the topological diagram of the electricity meters, a chain of electricity meter correlation features is established to obtain related feature information. This related feature information is then used to classify power system regions, improving the efficiency of power fault detection. Finally, by using real-time electricity meter data and the data difference coefficient, power faults are detected, improving the accuracy and timeliness of power fault detection and ensuring the stability of the power system.

[0038] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claimed invention. The scope of protection claimed by the appended claims and their equivalents is defined.

Claims

1. A method for intelligent analysis of electricity meter data based on data deviation feedback, characterized in that, include: Obtain power system information, which includes power system area information and electricity meter location information; Based on power system information, and taking the power system electrical lines as the foundation, the electricity meter is used as the topology node to obtain the electricity meter topology diagram; Based on power system information, historical data of electricity meters are obtained, including historical electrical data and historical time information for each electricity meter. Based on historical data of electricity meters and the topology diagram of electricity meters, obtain the data difference coefficient corresponding to each electricity meter; Based on the data difference coefficient, an energy meter association feature chain is established on the basis of the energy meter topology diagram to obtain the energy meter association feature information. Based on power system information and electricity meter association feature information, the regions corresponding to the electricity meters in each electricity meter association feature chain are divided into the same region to obtain power system region classification information; Based on the power system regional classification information, obtain the real-time data of the electricity meters corresponding to each region; Based on the real-time data from the electricity meter, determine whether a power failure has occurred. If so, obtain the location information of the power failure based on the real-time data from the electricity meter and the data difference coefficient. Power faults are detected based on their location information and power system regional classification information.

2. The intelligent analysis method for electricity meter data based on data deviation feedback according to claim 1, characterized in that, The step of obtaining the data difference coefficient for each electricity meter based on historical data and the electricity meter topology diagram specifically includes: Based on the historical data of the electricity meter, the historical data of the electricity meter corresponding to the same historical time is divided into the same dataset to obtain the historical dataset of the electricity meter. Based on power system information, electrical data thresholds for electricity meters are obtained. These electrical data thresholds represent the maximum and minimum electrical data values ​​of each electricity meter during normal operation in the power system. Based on the electrical data thresholds of the electricity meter and the historical data set of the electricity meter, the historical electrical data of each electricity meter that exceeds the electrical data threshold is regarded as the historical abnormal data of that electricity meter. Based on historical abnormal data, the historical dataset of electricity meters corresponding to the historical abnormal data is used as the abnormal dataset of electricity meters. Based on the topology diagram of the electricity meter, the topology node corresponding to the electricity meter is taken as the feature node, and the topology node associated with the feature node is taken as the associated topology node corresponding to the feature node. Based on the associated topology node corresponding to each feature node, the associated topology node pointed to by the feature node is taken as the influence topology node of the feature node, and the associated topology node pointing to the feature node is taken as the starting topology node of the feature node. Based on the feature nodes and the corresponding influencing topology nodes, and using the electricity meter anomaly dataset as a basis, obtain the synchronous anomaly dataset corresponding to each influencing topology node. The synchronous anomaly dataset indicates that the influencing topology node and the feature node both have historical anomaly data at the same historical time. The number of synchronization anomaly datasets is taken as the number of anomaly synchronizations that affect each topology node and feature node; The number of abnormal data sets of electricity meters corresponding to each feature node is used as the target number of abnormalities. The ratio of the number of abnormal synchronizations affecting topology nodes and feature nodes to the target number of abnormal synchronizations is used as the abnormal synchronization coefficient of the affected topology node. Based on the abnormal data set of electricity meters, the threshold of the abnormal synchronization coefficient is obtained; The data difference coefficient is obtained based on the abnormal synchronization coefficient threshold.

3. The intelligent analysis method for electricity meter data based on data deviation feedback according to claim 2, characterized in that, The process of obtaining the abnormal synchronization coefficient threshold based on the abnormal data set of electricity meters specifically includes: Based on the abnormal electricity meter dataset, obtain the electricity meter information corresponding to the historical abnormal data; Use the electricity meters corresponding to historical abnormal data as target electricity meters and obtain the topology node information corresponding to the target electricity meters; Based on the topology node information corresponding to the target energy meter, obtain the associated topology node information of the topology node corresponding to the target energy meter; Target energy meters whose associated topology nodes do not contain the starting topology node are designated as feature energy meters. If all target energy meters have an initial topology node, then the target energy meter with the smallest number of initial topology nodes is selected as the feature energy meter. Based on the characteristic energy meter, obtain the abnormal synchronization coefficient of the topology node corresponding to the characteristic energy meter; The abnormal data set of the electricity meter corresponding to the historical abnormal data of the characteristic electricity meter is used as the characteristic abnormal data set. Based on the feature anomaly dataset, the topology node of the electricity meter corresponding to the historical anomaly data in each feature anomaly dataset is taken as the anomaly topology node of that feature anomaly dataset. Based on the influencing topology nodes corresponding to the characteristic energy meters, the abnormal topology node with the largest number of abnormal topology nodes between it and the influencing topology node is taken as the influencing calibration node corresponding to the influencing topology node. The number of abnormal topology nodes between each affected topology node and its corresponding affected calibration node is taken as the maximum depth value of that affected topology node. Based on the abnormal synchronization coefficient and maximum depth value of the affected topology nodes corresponding to the characteristic energy meters, obtain the abnormal synchronization characteristic index corresponding to each characteristic energy meter; The average of the abnormal synchronization characteristic indexes of all characteristic energy meters is used as the threshold for the abnormal synchronization coefficient. Specifically, the abnormal synchronization characteristic index is: In the formula, This is an abnormal synchronization characteristic index. Let be the abnormal synchronization coefficient of the k-th influencing topology node corresponding to the characteristic energy meter. The influence coefficient of the k-th influencing topology node corresponding to the characteristic energy meter. This represents the maximum depth value of the k-th influencing topology node corresponding to the characteristic energy meter. , representing the attenuation factor.

4. The intelligent analysis method for electricity meter data based on data deviation feedback according to claim 3, characterized in that, The step of obtaining the data difference coefficient based on the abnormal synchronization coefficient threshold specifically includes: Based on the starting topology node corresponding to each feature node, the starting topology node with an abnormal synchronization coefficient higher than the abnormal synchronization coefficient threshold is taken as the strongly correlated node corresponding to that feature node. Based on the feature nodes and the strongly correlated nodes corresponding to each feature node, obtain the abnormal deviation degree corresponding to the feature nodes and the strongly correlated nodes; Based on the abnormal deviation degree corresponding to the feature node and the strongly correlated node, and taking the abnormal synchronization coefficient corresponding to the strongly correlated node as the basis, the data difference coefficient corresponding to each feature node is obtained, that is, the data difference coefficient of the electricity meter corresponding to the feature node. Specifically, the abnormal deviation is: In the formula, F represents the degree of abnormal deviation. For the i-th historical outlier, For reference electrical data, This represents the maximum value of the electrical data. Let n be the minimum value of electrical data, and n be the total number of historical anomalies. The specific data difference coefficient is: In the formula, This represents the data difference coefficient of the x-th electricity meter. This indicates the degree of abnormal deviation of the x-th electricity meter. This represents the abnormal deviation of the j-th strongly correlated node of the x-th electricity meter. This represents the abnormal synchronization coefficient of the j-th strongly correlated node of the x-th energy meter. The total number of strongly correlated nodes for the x-th electricity meter Let represent the transmission coefficient of the j-th strongly correlated node of the x-th energy meter.

5. The intelligent analysis method for electricity meter data based on data deviation feedback according to claim 4, characterized in that, The step of dividing the region corresponding to each energy meter in each energy meter association feature chain into the same region based on power system information and energy meter association feature information to obtain power system region classification information specifically includes: S501: Based on the historical data of the electricity meter and the electrical data threshold of the electricity meter, the maximum value of the difference between the historical electrical data and the benchmark electrical data corresponding to each electricity meter is taken as the electrical identification amplitude corresponding to the electricity meter. S502: Based on the electricity meter topology graph, take any electricity meter as the starting node of the path and obtain the influencing topology node corresponding to the starting node in the electricity meter topology graph. S503: The ratio of the data difference coefficient corresponding to the topology node to the data difference coefficient corresponding to the path starting node is used as the similarity difference coefficient; S504: Based on the electrical identification amplitude and the similarity difference coefficient corresponding to the affected topology node, determine whether the affected topology node meets the association characteristics; If the ratio of the electrical identification amplitude of the influencing topology node to the electrical identification amplitude of the path starting node exceeds the similarity difference coefficient, then the influencing topology node does not meet the association characteristics; if the ratio of the electrical identification amplitude of the influencing topology node to the electrical identification amplitude of the path starting node does not exceed the similarity difference coefficient, then the influencing topology node is regarded as the first associated node. S505: Based on the first-level associated nodes, obtain the influence topology nodes corresponding to each first-level associated node. Repeat steps S503-S504 and take the influence topology nodes that meet the association characteristics as the second-level associated nodes. S506: Repeat step S505 until no new associated nodes can be added, forming an energy meter associated feature chain with the path start node as the root, then proceed to S507. S507: Based on the electricity meter topology diagram, remove the topology nodes in the electricity meter association feature chain, and repeat steps S502-506 for the remaining topology nodes until all electricity meters are traversed to obtain the electricity meter association feature information. S508: Based on the electricity meter association feature information, divide the regions corresponding to all electricity meters in each electricity meter association feature chain into the same region to obtain power system region classification information.

6. The intelligent analysis method for electricity meter data based on data deviation feedback according to claim 5, characterized in that, The step of determining whether a power failure has occurred based on real-time data from the electricity meter specifically includes: The difference between the real-time data of the electricity meter and the reference electrical data is used as the real-time deviation value for each electricity meter. Based on the electrical data threshold of the electricity meter, the difference between the maximum and minimum electrical data values ​​of each electricity meter during normal operation in the power system is used as the benchmark deviation value. The ratio of the real-time deviation value to the reference deviation value is used as the real-time deviation coefficient of the energy meter. Based on the real-time deviation coefficient and data difference coefficient of each electricity meter, it is determined whether the electricity meter is abnormal. If the real-time deviation coefficient of the electricity meter exceeds the data difference coefficient, the electricity meter is regarded as a faulty electricity meter and the faulty electricity meter information is obtained. Based on the associated feature information of electricity meters and the information of faulty and abnormal electricity meters, the associated feature chains of all suspected faulty associated meters are extracted as abnormal associated feature chains. The ratio of the number of faulty energy meters in the anomaly association feature chain to the total number of energy meters in the anomaly association feature chain is used as the regional anomaly coefficient. Based on the power system regional classification information, the power coefficient region to which the anomaly association feature chain corresponding to the maximum value of the regional anomaly coefficient belongs is taken as the fault region. Based on the data difference coefficient, the product of the real-time deviation coefficient of the faulty energy meter and the data difference coefficient in the abnormal association feature chain is used as the fault identification coefficient. The faulty energy meter corresponding to the maximum value of the fault identification coefficient is used as the location of the power fault, and the power fault location information is obtained.

7. A smart analysis system for electricity meter data based on data deviation feedback, used to implement the analysis method as described in any one of claims 1-6, characterized in that, include: The main control module is used to establish an energy meter association feature chain based on the energy meter topology diagram according to the data difference coefficient, obtain energy meter association feature information, divide the area corresponding to each energy meter in the energy meter association feature chain into the same area according to the power system information and the energy meter association feature information, obtain power system area classification information, determine whether the energy meter is abnormal according to the real-time deviation coefficient and data difference coefficient of each energy meter, extract the association feature chain of all suspected faulty associated tables as the abnormal association feature chain according to the energy meter association feature information and the faulty energy meter information, take the product of the energy meter real-time deviation coefficient and the data difference coefficient of the faulty energy meter in the abnormal association feature chain as the fault identification coefficient, take the faulty energy meter corresponding to the maximum value of the fault identification coefficient as the power fault location, obtain power fault location information, and detect power faults according to the power fault location information and the power system area classification information. The information acquisition module is used to acquire power system information, including power system regional information and electricity meter location information. Based on the power system information, the module uses electricity meters as topology nodes based on the power system electrical lines to acquire an electricity meter topology diagram. Based on the power system information, the module acquires historical data of the electricity meters, including historical electrical data and historical time information corresponding to each electricity meter. Based on the power system regional classification information, the module acquires real-time data of the electricity meters corresponding to each region. The evaluation module is used to divide the historical data of electricity meters corresponding to the same historical time into the same dataset based on the historical time, obtain the historical dataset of electricity meters, obtain the abnormal synchronization coefficient based on the historical dataset of electricity meters, obtain the abnormal deviation degree corresponding to the feature node and the strongly correlated node corresponding to each feature node based on the feature node and the strongly correlated node corresponding to each feature node, and obtain the data difference coefficient corresponding to each feature node, i.e., the data difference coefficient of the electricity meter corresponding to the feature node, based on the abnormal synchronization coefficient corresponding to the strongly correlated node and the abnormal deviation degree corresponding to the feature node and the strongly correlated node. The display module interacts with the main control module and is used to output and display the topology diagram of the electricity meter, the data difference coefficient, the electricity meter association characteristic information, the power system area classification information, the real-time data of the electricity meter, and the power fault location information.

8. The intelligent analysis system for electricity meter data based on data deviation feedback according to claim 7, characterized in that, The main control module specifically includes: The control unit is used to determine whether an energy meter is abnormal based on the real-time deviation coefficient and data difference coefficient of each energy meter; extract the association feature chain of all suspected fault association tables as an abnormal association feature chain based on the energy meter association feature information and faulty energy meter information; take the product of the real-time deviation coefficient and the data difference coefficient of the faulty energy meter in the abnormal association feature chain as the fault identification coefficient based on the data difference coefficient; take the faulty energy meter corresponding to the maximum value of the fault identification coefficient as the power fault location; obtain the power fault location information; and detect the power fault based on the power fault location information and the power system area classification information. An information receiving unit, which interacts with the information acquisition module and the evaluation module, is used to receive data and transmit it to the region division unit. The region division unit is used to establish an energy meter association feature chain based on the energy meter topology diagram according to the data difference coefficient, obtain energy meter association feature information, and divide the region corresponding to each energy meter in the energy meter association feature chain into the same region according to the power system information and the energy meter association feature information, thereby obtaining power system region classification information.

9. The intelligent analysis system for electricity meter data based on data deviation feedback according to claim 7, characterized in that, The information acquisition module specifically includes: The first acquisition unit is used to acquire power system information, which includes power system area information and electricity meter location information. Based on the power system information, the electricity meter is used as a topology node based on the power system electrical lines to acquire an electricity meter topology diagram. The second acquisition unit is used to acquire historical data of electricity meters based on power system information. The historical data of electricity meters includes historical electrical data and historical time information corresponding to each electricity meter. According to the power system regional classification information, the real-time data of electricity meters corresponding to each region is acquired.

10. The intelligent analysis system for electricity meter data based on data deviation feedback according to claim 7, characterized in that, The evaluation module specifically includes: The first evaluation unit is used to divide the historical data of the electricity meter corresponding to the same historical time into the same dataset based on the historical time, obtain the historical dataset of the electricity meter, and obtain the abnormal synchronization coefficient based on the historical dataset of the electricity meter. The second evaluation unit is used to obtain the abnormal deviation degree corresponding to the feature node and the strongly correlated node corresponding to each feature node based on the feature node and the strongly correlated node corresponding to each feature node, and to obtain the data difference coefficient corresponding to each feature node, i.e. the data difference coefficient of the electricity meter corresponding to the feature node, based on the abnormal deviation degree corresponding to the feature node and the strongly correlated node, and using the abnormal synchronization coefficient corresponding to the strongly correlated node as a basis.

Images