Power distribution equipment fault analysis method and system based on edge computing

Abstract

The application discloses a power distribution equipment fault analysis method and system based on edge calculation, and relates to the field of smart grid. The method comprises the following steps: determining a plurality of power distribution equipment; obtaining historical equipment fault records and determining a plurality of edge feature groups; building a multi-source monitoring matrix for real-time monitoring to obtain a plurality of monitoring data streams; establishing a communication connection between an edge calculation node and a cloud platform and receiving same-frequency analog operation parameters; identifying the monitoring data streams, activating a fault analysis channel if abnormal monitoring features are found, performing fault prediction, and determining predicted fault data; inputting the predicted fault type into a fault operation and maintenance resource library, determining a power distribution control scheme, mapping feedback to the edge calculation node, and executing power distribution control. The technical problems of insufficient real-time and accuracy of data processing and analysis in existing power distribution equipment fault analysis and inability to respond to faults in a timely manner are solved, and the technical effects of improving the real-time and accuracy of fault processing and reducing the impact of faults on the power system are achieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of smart grid, in particular to a power distribution equipment fault analysis method and system based on edge computing. BACKGROUND

[0002] With the rapid development of power systems and the deepening of intelligent transformation, the number of power distribution equipment is increasing, and the operating environment is becoming increasingly complex. This change brings multiple challenges, making fault analysis particularly important. In existing methods, power distribution equipment fault analysis usually relies on centralized data processing and analysis centers. Although this approach can achieve a certain degree of fault detection and diagnosis, the large amount of data and high transmission delay result in insufficient real-time and accuracy of fault analysis and processing, which cannot respond to faults in a timely manner.

[0003] In the related art, the power distribution equipment fault analysis has the technical problem of insufficient real-time and accuracy of data processing and analysis, and cannot respond to faults in a timely manner. SUMMARY

[0004] The present application provides a power distribution equipment fault analysis method and system based on edge computing, which adopts a plurality of edge feature groups, builds a multi-source monitoring matrix, establishes an edge computing node, identifies monitoring data streams, performs fault prediction, determines a power distribution control scheme, and other technical means, thereby achieving the technical effects of improving the real-time and accuracy of fault processing, reducing the impact of faults on power systems, and improving the stability and safety of power distribution.

[0005] The present application provides a power distribution equipment fault analysis method based on edge computing, comprising:

[0006] Identifying a target power distribution system and determining a plurality of power distribution equipment;

[0007] Retrieving and obtaining historical equipment fault records of the plurality of power distribution equipment in a cloud platform, and identifying the historical equipment fault records to determine a plurality of edge feature groups;

[0008] Building a multi-source monitoring matrix based on the plurality of edge feature groups, and performing real-time monitoring on the plurality of power distribution equipment through the multi-source monitoring matrix to obtain a plurality of monitoring data streams, wherein the multi-source monitoring matrix and a plurality of edge computing nodes have a mapping connection relationship;

[0009] Establishing a communication connection between the edge computing node and the cloud platform, and receiving same-frequency analog operating parameters, wherein the cloud platform is embedded with a power distribution twin model of the target power distribution system;

[0010] Within the edge computing node, a monitoring data stream is identified based on the same-frequency analog operation parameters, and if an abnormal monitoring feature is found, a fault analysis channel of the cloud platform is activated, fault prediction is performed based on the monitoring data stream and historical monitoring data stream, and predicted fault data is determined, including a predicted fault type and an abnormal power distribution equipment.

[0011] The predicted fault type is input into a fault operation and maintenance resource library, a power distribution control scheme is determined, and the power distribution control scheme is mapped and fed back to the edge computing node to perform power distribution control of the abnormal power distribution equipment.

[0012] In a possible implementation, at the cloud platform, historical equipment fault records of the plurality of power distribution equipment are retrieved and identified to determine a plurality of edge feature groups, and the following processing is performed:

[0013] A first power distribution equipment is selected from the plurality of power distribution equipment, and at the cloud platform, a first historical equipment fault record is retrieved with the first power distribution equipment as a constraint;

[0014] The first historical equipment fault record is identified to generate a first frequently-occurring fault feature set, wherein each first frequently-occurring fault feature includes a feature proportion;

[0015] A first frequently-occurring fault feature whose feature proportion meets a predetermined proportion threshold is set as a first edge analysis feature, and all first frequently-occurring fault features in the first frequently-occurring fault feature set are set as first edge monitoring features;

[0016] A first edge feature group is formed based on the first edge analysis features and the first edge monitoring features, and is added to the plurality of edge feature groups.

[0017] In a possible implementation, the first historical equipment fault record is identified to generate a first frequently-occurring fault feature set, and the following processing is performed:

[0018] A plurality of first historical equipment fault data in the first historical equipment fault record are sequentially subjected to fault intensity evaluation to determine a plurality of fault intensity coefficients, wherein the fault intensity evaluation indicators include loss degree, correlation range, and recovery time length;

[0019] The plurality of first historical equipment fault data are clustered according to fault types to determine a plurality of fault types and a plurality of frequencies;

[0020] The plurality of frequencies are corrected based on the plurality of fault intensity coefficients to determine a plurality of updated frequency proportions, and a fault type sequence is generated based on the plurality of updated frequency proportions;

[0021] selecting a preset number of fault types in the sequence of fault types as first frequent fault types, and obtaining a first frequent equipment fault data set of the first frequent fault types;

[0022] Based on the first frequent equipment fault data set, a plurality of first frequent fault features are obtained, and a plurality of feature ratios are statistically obtained;

[0023] According to the plurality of first frequent fault features and the plurality of feature ratios, the first frequent fault feature set is constructed.

[0024] In possible implementation manners, the following processing is performed:

[0025] Based on the edge monitoring features of the plurality of edge feature groups, a multi-source monitoring matrix is built;

[0026] Based on the plurality of power distribution equipment, a plurality of edge computing nodes are determined, wherein the power distribution equipment and the edge computing nodes correspond one by one;

[0027] According to the edge analysis features of the plurality of edge feature groups, a feature analysis strategy of the plurality of edge computing nodes is configured.

[0028] In possible implementation manners, in the edge computing node, based on the same frequency simulation operation parameters, a monitoring data stream is identified, and the following processing is performed:

[0029] Based on the target power distribution system, the power distribution twin model is synchronously operated;

[0030] In a preset time window, same frequency simulation operation data of the power distribution twin model is obtained;

[0031] Based on the edge analysis features of the plurality of edge feature groups, the same frequency simulation operation data is extracted to obtain same frequency simulation operation parameters, wherein the same frequency simulation operation parameters and the edge computing nodes correspond one by one;

[0032] In the edge computing node, based on the feature analysis strategy, the monitoring data stream is extracted to determine monitoring feature data;

[0033] The same frequency simulation operation parameters and the monitoring feature data are compared and analyzed to determine an abnormal feature deviation set, and a feature type with an abnormal feature deviation greater than a deviation tolerance interval is set as the abnormal monitoring feature.

[0034] In possible implementation manners, the following processing is performed:

[0035] If no abnormal monitoring feature is found, in the edge computing node, the monitoring data stream is preprocessed based on a predetermined data processing strategy to obtain standard monitoring data stream, wherein the predetermined data processing strategy includes data cleaning, format conversion, and dimensionless processing;

[0036] The standard monitoring data stream and the data monitoring time node are input into the cloud platform for mapping and association storage.

[0037] In a possible implementation, if an abnormal monitoring feature is found, a fault analysis channel of the cloud platform is activated, fault prediction is performed based on the monitoring data stream and historical monitoring data stream, predicted fault data is determined, and the following processing is performed:

[0038] The power distribution equipment corresponding to the abnormal monitoring feature is determined, and matching in the fault analysis channel is performed based on the power distribution equipment to determine a matching analysis subchannel, wherein the fault analysis channel is constructed based on a feedforward neural network and includes multiple analysis subchannels, and each analysis subchannel corresponds to one power distribution equipment;

[0039] In the cloud platform, abnormal monitoring data stream of the abnormal monitoring feature and historical monitoring data stream are mapped and obtained;

[0040] The abnormal monitoring data stream is compared with the same-frequency analog operation data for deviation, and an abnormal deviation feature set is determined;

[0041] The abnormal deviation feature set is analyzed for a change trend based on the historical monitoring data stream, and an abnormal deviation trend set is determined;

[0042] The abnormal deviation feature set and the abnormal deviation trend set are input into the matching analysis subchannel for prediction, and the predicted fault data is output.

[0043] The application also provides a power distribution equipment fault analysis system based on edge computing, including:

[0044] A power distribution equipment determination module is configured to identify a target power distribution system and determine multiple power distribution equipments;

[0045] An edge feature group determination module is configured to retrieve historical equipment fault records of the multiple power distribution equipments in a cloud platform, identify the historical equipment fault records, and determine multiple edge feature groups;

[0046] A monitoring data stream acquisition module is configured to build a multi-source monitoring matrix based on the multiple edge feature groups, perform real-time monitoring on the multiple power distribution equipments through the multi-source monitoring matrix, and obtain multiple monitoring data streams, wherein the multi-source monitoring matrix and multiple edge computing nodes have a mapping connection relationship;

[0047] A same-frequency analog operation parameter receiving module is configured to establish a communication connection between the edge computing node and the cloud platform and receive same-frequency analog operation parameters, wherein the cloud platform is embedded with a power distribution twin model of the target power distribution system.

[0048] A fault prediction module is configured to identify monitoring data streams based on the same-frequency analog operation parameters in the edge computing node, activate a fault analysis channel of the cloud platform if abnormal monitoring features are found, perform fault prediction based on the monitoring data streams and historical monitoring data streams, and determine predicted fault data including a predicted fault type and an abnormal power distribution device.

[0049] A power distribution control module is configured to input the predicted fault type into a fault operation and maintenance resource library, determine a power distribution control scheme, map the power distribution control scheme back to the edge computing node, and perform power distribution control of the abnormal power distribution device.

[0050] The power distribution device fault analysis method and system based on edge computing provided in the present application can identify a target power distribution system, determine a plurality of power distribution devices, retrieve historical device fault records of the plurality of power distribution devices in a cloud platform, identify the historical device fault records, determine a plurality of edge feature groups, build a multi-source monitoring matrix based on the plurality of edge feature groups, perform real-time monitoring of the plurality of power distribution devices through the multi-source monitoring matrix to obtain a plurality of monitoring data streams, wherein the multi-source monitoring matrix and a plurality of edge computing nodes have a mapping connection relationship, then establish a communication connection between the edge computing node and the cloud platform and receive same-frequency analog operation parameters, wherein the cloud platform is embedded with a power distribution twin model of the target power distribution system, then identify monitoring data streams based on the same-frequency analog operation parameters in the edge computing node, activate a fault analysis channel of the cloud platform if abnormal monitoring features are found, perform fault prediction based on the monitoring data streams and historical monitoring data streams, determine predicted fault data including a predicted fault type and an abnormal power distribution device, finally input the predicted fault type into a fault operation and maintenance resource library, determine a power distribution control scheme, map the power distribution control scheme back to the edge computing node, and perform power distribution control of the abnormal power distribution device, thereby improving the real-time performance and accuracy of fault handling, reducing the impact of faults on the power system, and improving the stability and safety of power distribution. BRIEF DESCRIPTION OF DRAWINGS

[0051] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings of the embodiments of the present application will be briefly introduced below, and the flowcharts are used to illustrate the operations performed by the system according to the embodiments of the present application in the present application. It should be understood that the foregoing or the following operations are not necessarily performed in sequence. On the contrary, various steps can be processed in reverse order or simultaneously according to needs. Meanwhile, other operations can be added to these processes, or one or more steps of operations can be removed from these processes.

[0052] Figure 1 The flowchart of the power distribution equipment fault analysis method based on edge computing provided for the embodiments of the present application is shown in the figure.

[0053] Figure 2 The structural schematic diagram of the power distribution equipment fault analysis system based on edge computing provided for the embodiments of the present application is shown in the figure.

[0054] Legend: power distribution equipment determination module 10, edge feature group determination module 20, monitoring data flow acquisition module 30, same frequency analog operation parameter receiving module 40, fault prediction module 50, power distribution control module 60. DETAILED DESCRIPTION

[0055] The above description is only a summary of the technical solutions of the present application, in order to more clearly understand the technical means of the present application, the content of the specification can be implemented, and in order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the following specific embodiments of the present application are described.

[0056] In order to make the purposes, technical solutions and advantages of the present application more clear, the present application will be further described in detail below in combination with the drawings, and the described embodiments should not be regarded as limiting the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0057] In the following description, "some embodiments" are referred to, which describe a subset of all possible embodiments, but it can be understood that "some embodiments" can be the same subset or different subsets of all possible embodiments, and can be combined with each other without conflict, and the term "first\second" referred to is only to distinguish similar objects, and does not represent a specific order for the objects. The terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or server including a series of steps or units does not have to be limited to those steps or units clearly listed, but can include other steps or modules not clearly listed or inherent to these processes, methods, products or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as understood by those skilled in the art to which the present application belongs. The terms used herein are only for the purpose of describing the embodiments of the present application.

[0058] The embodiments of the present application provide a power distribution equipment fault analysis method based on edge computing, as shown in the method comprises: Figure 1 The method comprises the following steps:

[0059] Step S100, identifying a target power distribution system and determining a plurality of power distribution equipment. Specifically, based on user request, system configuration or specific demand, a target power distribution system, i.e. a specific power distribution system that needs to be analyzed, is determined, wherein the power distribution system refers to the part of the power system for distributing electric energy, including high-voltage, medium-voltage and low-voltage power distribution equipment and its connecting lines. In the target power distribution system, all related power distribution equipment is listed and identified, wherein the power distribution equipment is an electrical device for realizing the functions of electric energy distribution, control, protection, etc., such as transformer, switch cabinet, capacitor, etc.

[0060] Step S200, retrieving and acquiring historical equipment fault records of the plurality of power distribution equipment on a cloud platform, and identifying the historical equipment fault records to determine a plurality of edge feature groups. Wherein the cloud platform is a remote server cluster providing data storage, computing services and application management. Specifically, in the database of the cloud platform, according to the identification information of the power distribution equipment, the historical equipment fault records of these power distribution equipment are retrieved, data processing, analysis and mining methods are used to analyze the historical equipment fault records, identify the feature information related to equipment failure, such as failure time, failure type, failure location, etc., and based on the identified feature information, a plurality of edge feature groups are formed.

[0061] In a possible implementation, step S200 further includes step S201, selecting a first power distribution device from the plurality of power distribution devices, and retrieving a first historical device failure record in the cloud platform with the first power distribution device as a constraint. Specifically, one power distribution device is randomly selected from the plurality of power distribution devices determined in step S100 as the first power distribution device, which is any specific device selected from the plurality of power distribution devices. In the database of the cloud platform, the device identifier, time range, and other conditions of the first power distribution device are used to retrieve the first historical device failure record related to the first power distribution device. Step S202, identifying the first historical device failure record to generate a first frequently occurring failure feature set, wherein each first frequently occurring failure feature includes a feature proportion. Specifically, the first historical device failure record is counted and analyzed to identify different types of failure features, and the proportion of each failure feature in the total failure record, i.e., the feature proportion, is calculated. The identified failure features and their feature proportions are combined into the first frequently occurring failure feature set, i.e., the first frequently occurring failure feature set is a set containing the first power distribution device frequently occurring failure features and their occurrence proportions. Step S203, selecting the first frequently occurring failure feature whose feature proportion meets a predetermined proportion threshold as a first edge analysis feature, and selecting all first frequently occurring failure features in the first frequently occurring failure feature set as first edge monitoring features. Specifically, a predetermined proportion threshold is set to filter important analysis features. From the first frequently occurring failure feature set, features with a feature proportion exceeding the predetermined proportion threshold are selected as first edge analysis features, which are key features for in-depth analysis. All features in the first frequently occurring failure feature set are selected as first edge monitoring features for comprehensive monitoring of all frequently occurring failure features of the first power distribution device. Step S204, assembling a first edge feature group based on the first edge analysis features and the first edge monitoring features, and adding it to the plurality of edge feature groups. Specifically, the first edge analysis features and the first edge monitoring features are combined into a feature group, i.e., the first edge feature group (a feature group containing analysis features and monitoring features of the first power distribution device), and the first edge feature group is added to the plurality of edge feature groups. The edge feature groups of the remaining power distribution devices are obtained in a similar manner to obtaining the first edge feature group, and are added to the plurality of edge feature groups. This implementation reduces unnecessary calculation and analysis burden by filtering key analysis features while maintaining comprehensive monitoring, achieving the technical effects of improving data processing efficiency, enhancing the accuracy and reliability of failure analysis.

[0062] In a possible implementation, step S202 further includes step S2021 of sequentially performing failure intensity evaluation on the plurality of first historical equipment failure data in the first historical equipment failure record to determine a plurality of failure intensity coefficients, wherein the failure intensity evaluation indexes include loss degree, correlation range, and recovery time length. Specifically, each of the first historical equipment failure data in the first historical equipment failure record is traversed, and for each of the first historical equipment failure data, failure intensity evaluation is performed using the three evaluation indexes of loss degree, correlation range, and recovery time length, wherein the loss degree refers to direct or indirect economic loss, performance decline, or other forms of negative impact suffered by the first power distribution equipment when the failure occurs, including cost increase or benefit reduction caused by production interruption, data loss, equipment damage, and the like; the correlation range refers to the range of failure impact, including affected components, modules, systems, or business processes, and the like; and the recovery time length refers to the time required for the first power distribution equipment to recover to normal operation from the occurrence of the failure, including the time required for each link of failure detection, positioning, repair, and verification. According to the evaluation result, a coefficient representing the failure intensity, i.e., a failure intensity coefficient, is calculated, and the failure intensity coefficient is a numerical value quantitatively representing the severity of the failure. Step S2022 includes clustering the plurality of first historical equipment failure data according to the failure types to determine a plurality of failure types and a plurality of frequencies. Specifically, the failure type information in each of the first historical equipment failure data is extracted, and a clustering algorithm (such as K-means, hierarchical clustering, or the like) is used to cluster the plurality of first historical equipment failure data according to the nature, cause, or performance of the failure, to form a plurality of failure types, and the occurrence frequency of each failure type in the first historical equipment failure record is counted. Step S2023 includes correcting the plurality of frequencies based on the plurality of failure intensity coefficients to determine a plurality of updated frequency proportions, and generating a failure type sequence based on the plurality of updated frequency proportions. Specifically, the frequency of each failure type is combined with the failure intensity coefficient corresponding thereto to perform weighted processing to obtain an updated frequency, the proportion of each updated frequency is calculated, the failure type sequence is sorted in descending order of the updated frequency proportion, and the failure type sequence is generated. Step S2024 includes selecting a preset number of failure types in the failure type sequence as first frequently-occurring failure types, and obtaining a first frequently-occurring equipment failure data set of the first frequently-occurring failure types. Specifically, a number threshold is determined according to actual requirements or a preset rule, a preset number of failure types in the failure type sequence are selected as the first frequently-occurring failure types, i.e., the first frequently-occurring failure types are failure types ranking in the front in the failure type sequence and having a higher frequency, and the failure data related to the first frequently-occurring failure types is extracted from the first historical equipment failure record to form the first frequently-occurring equipment failure data set. Step S2025 includes obtaining a plurality of first frequently-occurring failure features based on the first frequently-occurring equipment failure data set, and counting a plurality of feature proportions.Specifically, feature extraction is performed on the first frequently-occurring equipment failure data set to obtain a plurality of failure-related features, and the frequency or proportion of each feature appearing in the first frequently-occurring equipment failure data set is counted. In step S2026, the first frequently-occurring failure feature set is constructed according to the plurality of first frequently-occurring failure features and the plurality of feature proportions. Specifically, the extracted first frequently-occurring failure features and the corresponding feature proportions are sorted to construct a first frequently-occurring failure feature set containing the frequently-occurring failure features and the feature proportions. This implementation focuses on the failure features of the first frequently-occurring failure type, which has a greater impact on the stability, safety and operating efficiency of the system, avoids processing and analyzing a large amount of non-critical failure data, and achieves the technical effect of improving the efficiency of data processing and analysis.

[0063] In step S300, a multi-source monitoring matrix is built based on the plurality of edge feature groups, and the plurality of power distribution equipment are monitored in real time through the multi-source monitoring matrix to obtain a plurality of monitoring data streams, wherein the multi-source monitoring matrix and the plurality of edge computing nodes have a mapping connection relationship. Specifically, an edge feature group is used to design and build a multi-source monitoring matrix for real-time monitoring of power distribution equipment. Each row of the multi-source monitoring matrix can represent a monitoring point or a data source, and each column can represent a specific feature or index. The mapping connection relationship between the multi-source monitoring matrix and the plurality of edge computing nodes is established, that is, each edge computing node is responsible for collecting and processing data from a specific monitoring point and transmitting the data to the multi-source monitoring matrix in real time.

[0064] In a possible implementation, step S300 further includes step S301 of constructing a multi-source monitoring matrix based on the edge monitoring features of the plurality of edge feature groups. Specifically, edge monitoring features are selected from the plurality of edge feature groups, and the structure of the multi-source monitoring matrix is designed according to the extracted edge monitoring features. Step S302 includes determining a plurality of edge computing nodes based on the plurality of power distribution devices, wherein the power distribution devices and the edge computing nodes are in one-to-one correspondence. Specifically, all power distribution devices that need to be monitored are listed to form a device list, and the number and position of edge computing nodes that need to be deployed are determined according to the device list. The edge computing nodes are computing nodes deployed near the power distribution devices, used for real-time collection and processing of device data, and can reduce the burden of the central server. Each power distribution device corresponds to one edge computing node, so each power distribution device has a corresponding edge computing node for data processing and analysis. The hardware and software environment of the edge computing node is configured, and the configured edge computing node is deployed to the corresponding position and connected with the power distribution device. Step S303 includes configuring a feature analysis strategy of the plurality of edge computing nodes according to the edge analysis features of the plurality of edge feature groups. Specifically, edge analysis features are extracted from the edge feature groups, and a corresponding feature analysis strategy is formulated for each edge computing node according to the extracted edge analysis features. The feature analysis strategy is a data processing and analysis rule or algorithm formulated for a specific feature, used to extract valuable information from raw data, including data preprocessing rules, fault identification algorithms, pattern recognition models, etc. The formulated feature analysis strategy is configured into the corresponding edge computing node, so that the edge computing node can perform data analysis and processing according to the feature analysis strategy. This implementation can achieve comprehensive coverage and real-time monitoring of the power distribution devices by constructing a multi-source monitoring matrix based on edge monitoring features, reduce the burden of the central server by configuring corresponding edge computing nodes for each power distribution device, improve data processing efficiency, ensure the pertinence and effectiveness of data analysis by configuring feature analysis strategies according to edge analysis features, and achieve the technical effect of more accurately identifying potential faults and abnormal states.

[0065] At step S400, the edge computing node establishes a communication connection with the cloud platform and receives the same-frequency analog operation parameters, wherein the cloud platform is embedded with a power distribution twin model of the target power distribution system. The power distribution twin model is a digital model established based on the target power distribution system and can simulate the operation state and behavior of the target power distribution system. The same-frequency analog operation parameters are parameters simulated by the cloud platform based on the power distribution twin model and synchronized with the operation state of the target power distribution system. Specifically, a communication protocol used between the edge computing node and the cloud platform is determined, network parameters, including an IP address, a port number, network authentication information, etc., are configured on the edge computing node, and a communication module on the edge computing node is started to establish a connection with the cloud platform. A data monitoring service is started on the edge computing node to receive data from the cloud platform. The cloud platform simulates same-frequency operation parameters, including electrical quantities such as voltage, current, and power factor, and non-electrical quantities such as temperature and humidity, based on the power distribution twin model, and sends the operation parameters to the edge computing node.

[0066] Step S500, in the edge computing node, based on the same frequency analog operation parameter identification monitoring data flow, if the abnormal monitoring feature is found, activate the fault analysis channel of the cloud platform, based on the monitoring data flow and historical monitoring data flow for fault prediction, determine the predicted fault data, the predicted fault data includes the predicted fault type and the abnormal power distribution equipment. Wherein, the abnormal monitoring feature refers to the feature detected in the monitoring data flow which is inconsistent with the normal state; the fault analysis channel is a functional module or channel in the cloud platform for receiving the monitoring data flow sent by the edge computing node, and performing fault analysis and prediction. Specifically, the edge computing node continuously receives the monitoring data flow from the multi-source monitoring matrix, the monitoring data flow contains the real-time operation state information of multiple power distribution equipment, the edge computing node uses the built-in algorithm or model to analyze and identify the received same frequency analog operation parameter, and extracts the key operation index and state information in the form of monitoring data flow. The edge computing node uses the same frequency analog operation parameter as a reference to analyze the monitoring data flow in real time, compares the current data with the historical data, the preset threshold or the pattern recognition algorithm, detects whether there is an abnormal monitoring feature, the abnormal monitoring feature includes abnormal fluctuation or out-of-range of current, voltage, temperature and other parameters, or abnormal change of power distribution equipment operation state, etc. Once the abnormal monitoring feature is detected, the edge computing node immediately sends a request to activate the fault analysis channel to the cloud platform through the established communication connection, and the cloud platform activates the corresponding fault analysis channel after receiving the request. The edge computing node sends the detected abnormal monitoring data flow and related historical monitoring data flow to the cloud platform, and the cloud platform uses the embedded power distribution twin model, machine learning algorithm or other analysis technology to comprehensively analyze the received data flow, including pattern recognition, trend analysis, probability prediction and other data analysis, to predict the possible fault type and the affected abnormal power distribution equipment. The cloud platform determines the predicted fault data according to the result of fault prediction, including the type of predicted fault (such as short circuit, overload, equipment aging, etc.) and the specific information of the affected abnormal power distribution equipment (such as equipment location, model, number, etc.).

[0067] In a possible implementation, in the edge computing node, the monitoring data stream is identified based on the same-frequency simulation operation parameters, and step S500 further includes the following steps. Step S501: based on the target power distribution system, the power distribution twin model is synchronously operated. Specifically, the power distribution twin model is started on the cloud platform, so that the power distribution twin model is kept synchronized with the actual operation state of the target power distribution system. The power distribution twin model simulates the operation state synchronized with the target power distribution system according to the target power distribution system data received in real time. Step S502: within a preset time window, same-frequency simulation operation data of the power distribution twin model is obtained. The same-frequency simulation operation data is the operation data simulated by the power distribution twin model within the preset time window and synchronized with the target power distribution system. Specifically, a preset time window is set, and simulation operation data is collected from the power distribution twin model within the preset time window. Step S503: the same-frequency simulation operation data is extracted based on the edge analysis features of the plurality of edge feature groups, to obtain same-frequency simulation operation parameters, wherein the same-frequency simulation operation parameters and the edge computing node are in one-to-one correspondence. Specifically, the same-frequency simulation operation data is extracted and processed by using the edge analysis features of the edge feature groups, and the extracted data is the same-frequency simulation operation parameters. Step S504: in the edge computing node, the monitoring data stream is extracted based on the feature analysis strategy, to determine the monitoring feature data. Specifically, the edge computing node receives real-time monitoring data stream from the multi-source monitoring matrix, and applies the configured feature analysis strategy (such as threshold analysis, pattern recognition, etc.) to extract and process the monitoring data stream, to determine the monitoring feature data related to the power distribution equipment state. Step S505: the same-frequency simulation operation parameters and the monitoring feature data are compared and analyzed, to determine an abnormal feature deviation set, and a feature type with an abnormal feature deviation greater than a deviation tolerance interval is extracted and set as the abnormal monitoring feature. Specifically, the same-frequency simulation operation parameters and the monitoring feature data are compared and analyzed, the deviation between the two is calculated, and the abnormal feature deviation set is determined. A deviation tolerance interval is set, which is used to judge whether the deviation exceeds the normal range. The feature type with an abnormal feature deviation exceeding the deviation tolerance interval is extracted, and the feature type is set as the abnormal monitoring feature. The abnormal monitoring feature indicates the fault or abnormal state of the power distribution equipment. In actual application, due to the change of equipment operation state and environmental conditions, some deviations are normal, and some deviations indicate potential faults. This implementation filters out the normal deviations by setting the deviation tolerance interval, and only focuses on the deviations that may indicate faults, thereby achieving the technical effects of improving the efficiency and accuracy of fault detection.

[0068] In a possible implementation, the step S500 further includes a step S506: if no abnormal monitoring feature is found, performing data preprocessing on the monitoring data stream based on a predetermined data processing strategy to obtain a standard monitoring data stream at the edge computing node, wherein the predetermined data processing strategy includes data cleaning, format conversion, and dimensionless processing. Specifically, after performing the step of identifying the abnormal monitoring feature, the edge computing node first checks whether the abnormal monitoring feature is found. If no abnormal monitoring feature is found, the edge computing node performs preprocessing on the monitoring data stream according to the predetermined data processing strategy. The predetermined data processing strategy is a set of preconfigured data processing methods and steps, which are used for cleaning, converting, and standardizing the monitoring data stream, including data cleaning (removing noise, repeated, or invalid data in the monitoring data stream), format conversion (converting the format of the monitoring data stream into a standard format that can be efficiently processed by the cloud platform), and dimensionless processing (eliminating the influence of dimension and magnitude). After the above data preprocessing step, the edge computing node obtains a standard monitoring data stream that is cleaned, formatted, and dimensionless. The step S507: inputting the standard monitoring data stream and the data monitoring time node to the cloud platform for mapping and associated storage. Specifically, the edge computing node sends the preprocessed standard monitoring data stream and the corresponding data monitoring time node to the cloud platform. After receiving the data, the cloud platform performs mapping and associated storage according to the correspondence between the data monitoring time node and the standard monitoring data stream, that is, the cloud platform stores the monitoring data at each moment in chronological order and establishes the association between the data. This implementation performs preprocessing and mapping and associated storage on the monitoring data stream in the case where no abnormal monitoring feature is found, which is used to establish a long-term data archive and achieves the technical effect of providing strong data support for the operation management and fault prevention of the target power distribution system.

[0069] In a possible implementation, if an abnormal monitoring feature is found, a fault analysis channel of the cloud platform is activated, a fault prediction is made based on the monitoring data stream and historical monitoring data stream, prediction fault data is determined, and step S500 further includes step S508 of determining the power distribution equipment corresponding to the abnormal monitoring feature, and matching in the fault analysis channel based on the power distribution equipment to determine a matching analysis subchannel. Specifically, by analyzing the abnormal monitoring feature, it is determined which power distribution equipment is associated with the abnormal monitoring feature. The cloud platform is built-in with a fault analysis channel based on a feedforward neural network (a neural network structure in which information is unidirectionally transmitted from an input layer to an output layer without feedback loop). The fault analysis channel includes a plurality of analysis subchannels, each of which corresponds to a power distribution equipment. According to the determined power distribution equipment, a corresponding matching analysis subchannel is found in the fault analysis channel. Step S509: In the cloud platform, the abnormal monitoring data stream in which the abnormal monitoring feature is obtained, and the historical monitoring data stream are mapped. Specifically, the cloud platform maps and extracts the corresponding abnormal monitoring data stream from the stored data according to the identification of the abnormal monitoring feature. At the same time, the cloud platform also obtains the historical monitoring data stream related to the corresponding power distribution equipment. Step S510: Deviation comparison is made between the abnormal monitoring data stream and the same frequency simulation running data to determine an abnormal deviation feature set. Specifically, the extracted abnormal monitoring data stream is compared with the same frequency simulation running data of the power distribution twin model. Through the comparison, the deviation between the two is identified, and the deviation features are combined to form the abnormal deviation feature set. Step S511: Change trend analysis is made on the abnormal deviation feature set based on the historical monitoring data stream to determine an abnormal deviation trend set. Specifically, the historical monitoring data stream is used to perform time series analysis on the abnormal deviation feature set to identify the change trend, and the abnormal deviation change trend obtained by the analysis is combined to form the abnormal deviation trend set. Step S512: The abnormal deviation feature set and the abnormal deviation trend set are input into the matching analysis subchannel for prediction, and the prediction fault data is output. Specifically, the abnormal deviation feature set and the abnormal deviation trend set determined in steps S510 and S511 are input into the matching analysis subchannel determined in step S508. The matching analysis subchannel processes and analyzes the input data based on the learning ability of the feedforward neural network, and finally outputs the prediction fault data. This implementation comprehensively analyzes the nature and evolution of the abnormal feature by using the comparison of the same frequency simulation running data and the change trend analysis of the historical monitoring data, thereby achieving the technical effects of improving the accuracy and reliability of fault prediction.

[0070] Step S600, input the predicted fault type into the fault operation and maintenance resource library, determine the power distribution control scheme, and map the power distribution control scheme to the edge computing node for feedback, and perform power distribution control of the abnormal power distribution equipment. Wherein, the fault operation and maintenance resource library is a database for storing and managing fault handling strategies and power distribution control schemes, which provides specific countermeasures for fault prediction results; the power distribution control scheme is a specific control strategy and operation step formulated for a specific fault type, which is used to restore the normal operation of the power distribution equipment or reduce the impact of the fault on the system. Specifically, the predicted fault type in the predicted fault data is input into the fault operation and maintenance resource library, and the corresponding power distribution control scheme is retrieved and matched in the fault operation and maintenance resource library according to the input predicted fault type, including adjusting device parameters, switching standby power supply, limiting load and other specific measures. The cloud platform maps and feeds back the power distribution control scheme to the corresponding edge computing node to perform the corresponding control operation, adjusts or intervenes the abnormal power distribution equipment, and eliminates the hidden danger or reduces the impact of the fault. The embodiments of the application adopt technical means such as determining multiple edge feature groups, building a multi-source monitoring matrix, establishing an edge computing node, identifying monitoring data flow, fault prediction, and determining a power distribution control scheme, which achieves the technical effects of improving the real-time and accuracy of fault handling, reducing the impact of faults on the power system, and improving the stability and safety of power distribution.

[0071] In the foregoing, with reference to Figure 1 The power distribution equipment fault analysis method based on edge computing according to the embodiments of the application is described in detail. Next, the power distribution equipment fault analysis system based on edge computing according to the embodiments of the application will be described with reference to Figure 2 The power distribution equipment fault analysis system based on edge computing according to the embodiments of the application is described in detail. Next, the power distribution equipment fault analysis system based on edge computing according to the embodiments of the application will be described with reference to

[0072] The power distribution equipment fault analysis system based on edge computing according to the embodiments of the application is used to solve the technical problem of insufficient real-time and accuracy of data processing and analysis in existing power distribution equipment fault analysis, which cannot respond to faults in time, and achieves the technical effects of improving the real-time and accuracy of fault handling, reducing the impact of faults on the power system, and improving the stability and safety of power distribution. The power distribution equipment fault analysis system based on edge computing includes: a power distribution equipment determination module 10, an edge feature group determination module 20, a monitoring data flow acquisition module 30, a same frequency analog operation parameter receiving module 40, a fault prediction module 50, and a power distribution control module 60.

[0073] The power distribution equipment determination module 10 is used to identify a target power distribution system and determine a plurality of power distribution equipment;

[0074] The edge feature group determination module 20 is used to retrieve and acquire historical equipment fault records of the plurality of power distribution equipment in the cloud platform, identify the historical equipment fault records, and determine a plurality of edge feature groups;

[0075] The monitoring data stream acquisition module 30 is configured to build a multi-source monitoring matrix based on the plurality of edge feature groups, and to monitor the plurality of power distribution devices in real time through the multi-source monitoring matrix to obtain a plurality of monitoring data streams, wherein the multi-source monitoring matrix is in a mapping connection relationship with a plurality of edge computing nodes.

[0076] The same-frequency analog operation parameter receiving module 40 is configured to establish a communication connection between the edge computing node and the cloud platform, and to receive same-frequency analog operation parameters, wherein the cloud platform is embedded with a power distribution twin model of the target power distribution system.

[0077] The fault prediction module 50 is configured to identify monitoring data streams based on the same-frequency analog operation parameters in the edge computing node, to activate a fault analysis channel of the cloud platform if abnormal monitoring features are found, to perform fault prediction based on the monitoring data streams and historical monitoring data streams, and to determine predicted fault data, wherein the predicted fault data includes a predicted fault type and an abnormal power distribution device.

[0078] The power distribution control module 60 is configured to input the predicted fault type into a fault operation and maintenance resource library, to determine a power distribution control scheme, and to map and feed back the power distribution control scheme to the edge computing node to perform power distribution control of the abnormal power distribution device.

[0079] Next, the specific configuration of the edge feature group determination module 20 will be described in detail. As described above, in the cloud platform, historical device fault records of the plurality of power distribution devices are retrieved and obtained, and the historical device fault records are identified to determine a plurality of edge feature groups. The edge feature group determination module 20 can further include: a first historical device fault record acquisition unit configured to select a first power distribution device from the plurality of power distribution devices, and to retrieve and obtain first historical device fault records in the cloud platform with the first power distribution device as a constraint; a first frequently-occurring fault feature set generation unit configured to identify the first historical device fault records and to generate a first frequently-occurring fault feature set, wherein each first frequently-occurring fault feature contains a feature proportion; a first edge monitoring feature setting unit configured to select a first frequently-occurring fault feature with a feature proportion satisfying a predetermined proportion threshold as a first edge analysis feature, and to select all first frequently-occurring fault features in the first frequently-occurring fault feature set as first edge monitoring features; and a first edge feature group assembly unit configured to assemble a first edge feature group based on the first edge analysis features and the first edge monitoring features, and to add the first edge feature group to the plurality of edge feature groups.

[0080] The first historical equipment failure record is identified, and a first frequently-occurring failure feature set is generated. The first frequently-occurring failure feature set generation unit can further include: a failure intensity evaluation subunit configured to sequentially perform failure intensity evaluation on a plurality of first historical equipment failure data in the first historical equipment failure record, and determine a plurality of failure intensity coefficients, wherein the failure intensity evaluation indicators include loss degree, correlation range, and recovery time length; a clustering subunit configured to cluster the plurality of first historical equipment failure data according to failure types, and determine a plurality of failure types and a plurality of frequencies; a failure type sequence generation subunit configured to correct the plurality of frequencies based on the plurality of failure intensity coefficients, determine a plurality of updated frequency proportions, and generate a failure type sequence based on the plurality of updated frequency proportions; a first frequently-occurring failure type setting subunit configured to select a pre-set number of failure types in the failure type sequence as first frequently-occurring failure types, and obtain a first frequently-occurring equipment failure data set of the first frequently-occurring failure types; a feature proportion acquisition subunit configured to acquire a plurality of first frequently-occurring failure features based on the first frequently-occurring equipment failure data set, and statistically obtain a plurality of feature proportions; and a first frequently-occurring failure feature set construction subunit configured to construct the first frequently-occurring failure feature set according to the plurality of first frequently-occurring failure features and the plurality of feature proportions.

[0081] In the following, the specific configuration of the monitoring data stream acquisition module 30 will be described in detail. As described above, the monitoring data stream acquisition module 30 can further include: a multi-source monitoring matrix building unit configured to build a multi-source monitoring matrix based on the edge monitoring features of the plurality of edge feature groups; an edge computing node determination unit configured to determine a plurality of edge computing nodes based on the power distribution equipment, wherein the power distribution equipment and the edge computing nodes correspond to each other in a one-to-one manner; and a feature analysis strategy configuration unit configured to configure the feature analysis strategies of the plurality of edge computing nodes according to the edge analysis features of the plurality of edge feature groups.

[0082] Below, the specific configuration of the fault prediction module 50 will be described in detail. As described above, based on the identification of the monitoring data stream in the edge computing node based on the same frequency simulation operation parameters, the fault prediction module 50 can further include: a power distribution twin model operation unit for synchronously operating the power distribution twin model based on the target power distribution system; a same frequency simulation operation data acquisition unit for acquiring the same frequency simulation operation data of the power distribution twin model within a preset time window; a same frequency simulation operation parameter acquisition unit for extracting the same frequency simulation operation data based on the edge analysis features of the plurality of edge feature groups, to obtain same frequency simulation operation parameters, wherein the same frequency simulation operation parameters and the edge computing node correspond one-to-one; a monitoring feature data determination unit for extracting monitoring data stream based on feature analysis strategy in the edge computing node, to determine monitoring feature data; an abnormal monitoring feature setting unit for comparing and analyzing the same frequency simulation operation parameters and the monitoring feature data, to determine an abnormal feature deviation set, and extracting the feature type with an abnormal feature deviation greater than the deviation tolerance interval as the abnormal monitoring feature.

[0083] Wherein, the fault prediction module 50 can further include: a standard monitoring data stream acquisition unit for, if no abnormal monitoring feature is found, performing data preprocessing on the monitoring data stream based on a predetermined data processing strategy in the edge computing node, to obtain a standard monitoring data stream, wherein the predetermined data processing strategy includes data cleaning, format conversion, and dimensionless processing; a data input unit for inputting the standard monitoring data stream and the data monitoring time node into the cloud platform for mapping and associated storage.

[0084] Wherein, if an abnormal monitoring feature is found, the fault analysis channel of the cloud platform is activated, fault prediction is performed based on the monitoring data stream and the historical monitoring data stream, to determine predicted fault data, and the fault prediction module 50 can further include: a fault analysis channel matching unit for determining the power distribution equipment corresponding to the abnormal monitoring feature, and matching in the fault analysis channel based on the power distribution equipment, to determine a matching analysis sub-channel, wherein the fault analysis channel is constructed based on a feedforward neural network, includes a plurality of analysis sub-channels, and each analysis sub-channel corresponds to a power distribution equipment; a data stream acquisition unit for mapping and acquiring the abnormal monitoring data stream of the abnormal monitoring feature and the historical monitoring data stream in the cloud platform; a deviation comparison unit for comparing and analyzing the abnormal monitoring data stream and the same frequency simulation operation data, to determine an abnormal deviation feature set; a change trend analysis unit for performing change trend analysis on the abnormal deviation feature set based on the historical monitoring data stream, to determine an abnormal deviation trend set; a prediction unit for inputting the abnormal deviation feature set and the abnormal deviation trend set into the matching analysis sub-channel for prediction, to output the predicted fault data.

[0085] The power distribution equipment fault analysis system based on edge computing provided by the embodiment of the application can execute the power distribution equipment fault analysis method based on edge computing provided by any embodiment of the application, and has the corresponding function modules and beneficial effects of the execution method.

[0086] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, however, any number of different modules can be used and run on the user terminal and / or server, and the various units and modules are only divided according to the functional logic, but are not limited to the above division, as long as the corresponding functions can be realized; in addition, the specific names of the functional units are only for the convenience of mutual differentiation, and do not limit the protection scope of the present application.

[0087] The above specific embodiments do not constitute a limitation on the protection scope of the present application. Those skilled in the art should understand that various modifications, combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present application shall be included in the protection scope of the present application. In some cases, the actions or steps described in the present application can be executed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are possible or can be advantageous.

Claims

1. A power distribution equipment fault analysis method based on edge computing, characterized by, The method comprises: identifying a target power distribution system and determining a plurality of power distribution devices; retrieving historical device failure records of the plurality of power distribution devices on a cloud platform, identifying the historical device failure records, and determining a plurality of edge feature groups; building a multi-source monitoring matrix based on the plurality of edge feature groups, performing real-time monitoring on the plurality of power distribution devices through the multi-source monitoring matrix, and obtaining a plurality of monitoring data streams, wherein the multi-source monitoring matrix is in a mapping connection relationship with a plurality of edge computing nodes; establishing a communication connection between the edge computing node and the cloud platform, and receiving a same-frequency analog operation parameter, wherein the cloud platform is embedded with a power distribution twin model of the target power distribution system; in the edge computing node, identifying monitoring data streams based on the same-frequency analog operation parameter, activating a fault analysis channel of the cloud platform if an abnormal monitoring feature is found, performing fault prediction based on the monitoring data streams and historical monitoring data streams, determining predicted fault data, and the predicted fault data including a predicted fault type and an abnormal power distribution device; inputting the predicted fault type into a fault operation and maintenance resource library, determining a power distribution control scheme, and mapping and feeding back the power distribution control scheme to the edge computing node to perform power distribution control on the abnormal power distribution device; retrieving historical device failure records of the plurality of power distribution devices on the cloud platform, identifying the historical device failure records, and determining a plurality of edge feature groups, comprising: selecting a first power distribution device from the plurality of power distribution devices, and retrieving a first historical device failure record in the cloud platform with the first power distribution device as a constraint; identifying the first historical device failure record to generate a first frequently-occurring fault feature set, wherein each first frequently-occurring fault feature contains a feature proportion; selecting a first frequently-occurring fault feature with a feature proportion satisfying a predetermined proportion threshold as a first edge analysis feature, and selecting all first frequently-occurring fault features in the first frequently-occurring fault feature set as first edge monitoring features; building a first edge feature group based on the first edge analysis features and the first edge monitoring features, and adding the first edge feature group to the plurality of edge feature groups; The method further comprises: building a multi-source monitoring matrix based on edge monitoring features of the plurality of edge feature groups; determining a plurality of edge computing nodes based on the plurality of power distribution devices, wherein the power distribution devices and the edge computing nodes correspond one-to-one; configuring a feature analysis strategy of the plurality of edge computing nodes according to edge analysis features of the plurality of edge feature groups.

2. The edge-computing-based power distribution equipment fault analysis method according to claim 1, characterized in that, identifying the first historical device failure record to generate a first frequently-occurring fault feature set, comprising: sequentially performing fault intensity evaluation on a plurality of first historical device failure data in the first historical device failure record to determine a plurality of fault intensity coefficients, wherein the fault intensity evaluation indicators include loss degree, correlation range, and recovery time; clustering the plurality of first historical device failure data according to fault types to determine a plurality of fault types and a plurality of frequencies; correcting the plurality of frequencies based on the plurality of fault intensity coefficients to determine a plurality of updated frequency proportions, and generating a fault type sequence based on the plurality of updated frequency proportions; selecting a preset number of fault types in the sequence of fault types as first frequently-occurring fault types, and obtaining a first frequently-occurring equipment fault data set of the first frequently-occurring fault types; based on the first frequently-occurring equipment fault data set, obtaining a plurality of first frequently-occurring fault features, and statistically obtaining a plurality of feature proportions; constructing the first frequently-occurring fault feature set according to the plurality of first frequently-occurring fault features and the plurality of feature proportions. 3.The edge-computing-based power distribution equipment fault analysis method according to claim 1, characterized in that, In the edge computing node, based on the same frequency simulation operation parameter, a monitoring data stream is identified, including: based on the target power distribution system, synchronously running the power distribution twin model; in a preset time window, obtaining same frequency simulation operation data of the power distribution twin model; based on the edge analysis features of the plurality of edge feature groups, extracting the same frequency simulation operation data to obtain same frequency simulation operation parameters, wherein the same frequency simulation operation parameters and the edge computing nodes are one-to-one corresponding; in the edge computing node, based on the feature analysis strategy, extracting the monitoring data stream to determine the monitoring feature data; comparing and analyzing the same frequency simulation operation parameters and the monitoring feature data to determine an abnormal feature deviation set, and extracting a feature type with an abnormal feature deviation greater than a deviation tolerance interval as the abnormal monitoring feature.

4. The edge-computing-based power distribution equipment fault analysis method according to claim 3, characterized in that, The method further comprises: if no abnormal monitoring feature is found, in the edge computing node, based on a predetermined data processing strategy, data preprocessing is performed on the monitoring data stream to obtain standard monitoring data stream, wherein the predetermined data processing strategy includes data cleaning, format conversion and dimensionless processing; inputting the standard monitoring data stream and the data monitoring time node into the cloud platform for mapping and correlation storage.

5. The edge-computing-based power distribution equipment fault analysis method according to claim 3, characterized in that, if an abnormal monitoring feature is found, activating the fault analysis channel of the cloud platform, based on the monitoring data stream and the historical monitoring data stream, performing fault prediction to determine the predicted fault data, including: determining the power distribution equipment corresponding to the abnormal monitoring feature, and based on the power distribution equipment, matching in the fault analysis channel to determine a matching analysis subchannel, wherein the fault analysis channel is constructed based on a feedforward neural network, including a plurality of analysis subchannels, and each analysis subchannel corresponds to a power distribution equipment; in the cloud platform, mapping and obtaining the abnormal monitoring data stream of the abnormal monitoring feature and the historical monitoring data stream; comparing and analyzing the abnormal monitoring data stream and the same frequency simulation operation data to determine an abnormal deviation feature set; based on the historical monitoring data stream, performing trend analysis on the abnormal deviation feature set to determine an abnormal deviation trend set; inputting the abnormal deviation feature set and the abnormal deviation trend set into the matching analysis subchannel for prediction, and outputting the predicted fault data.

6. A power distribution equipment fault analysis system based on edge computing, characterized by, The system is used to implement the edge computing-based power distribution equipment fault analysis method of any one of claims 1-5, and the system comprises: a power distribution equipment determination module, which is used to identify a target power distribution system and determine a plurality of power distribution equipments; an edge feature group determination module, which is used to retrieve and obtain historical equipment fault records of the plurality of power distribution equipments in the cloud platform, and identify the historical equipment fault records to determine a plurality of edge feature groups; A monitoring data stream acquisition module is configured to build a multi-source monitoring matrix based on the plurality of edge feature groups, to perform real-time monitoring on the plurality of power distribution devices through the multi-source monitoring matrix, and to obtain a plurality of monitoring data streams, wherein the multi-source monitoring matrix is in a mapping connection relationship with a plurality of edge computing nodes; A same-frequency analog operation parameter receiving module is configured to establish a communication connection between the edge computing node and the cloud platform, and to receive same-frequency analog operation parameters, wherein the cloud platform is embedded with a power distribution twin model of the target power distribution system; A fault prediction module is configured to identify monitoring data streams based on the same-frequency analog operation parameters in the edge computing node, to activate a fault analysis channel of the cloud platform if abnormal monitoring features are found, to perform fault prediction based on the monitoring data streams and historical monitoring data streams, to determine predicted fault data, and to determine the predicted fault data including a predicted fault type and an abnormal power distribution device; A power distribution control module is configured to input the predicted fault type into a fault operation and maintenance resource library, to determine a power distribution control scheme, to map and feed back the power distribution control scheme to the edge computing node, and to perform power distribution control on the abnormal power distribution device.

Images