Transformer operation fault detection method and system

Abstract

The invention is suitable for the technical field of fault detection, and particularly relates to a transformer operation fault detection method and system, and the method comprises the steps: finding out transformers in a preset range, setting a plurality of monitoring indexes, editing an evaluation rule, comparing the monitoring indexes with the evaluation rule, and clustering the transformers into a stable group and a risk group; the method comprises the steps of establishing a global model, collecting operation parameters of transformers, obtaining historical data of all transformer operation parameters in a stable group, training the global model to obtain a parameter group composed of a plurality of sets of model parameters, and enabling each transformer to correspond to one set of model parameters. According to the invention, by constructing a multi-level monitoring architecture, the operation parameters of all transformers can be macroscopically monitored by using the global sensing model, the preliminary screening of anomalies is realized, and when abnormal signals are detected, the local monitoring model is triggered to carry out deep detailed analysis on the related operation parameters, so that the accuracy of the abnormal signals is improved. Therefore, the accuracy and reliability of fault identification are greatly improved.

Description

Technical Field

[0001] The present invention relates to the technical field of fault detection, and in particular to a transformer operation fault detection method and system. Background Art

[0002] Transformer operation failure refers to the phenomenon that the operating status is abnormal or even fails due to the combined effects of electrical, thermal, mechanical and chemical factors during the long-term operation of the transformer. Common operation failures include inter-turn short circuit and inter-phase short circuit caused by insulation aging, moisture or breakdown.

[0003] Under existing data conditions, machine learning models based on historical operating data are usually used to diagnose transformer faults. However, for some transformers with good operating environments and long-term stable equipment status, abnormal data samples are extremely scarce, which makes it difficult for the model to effectively learn unique fault characteristics, affecting the diagnostic accuracy. On the contrary, in some areas with more complex operating environments or severe aging, transformer faults and abnormal events occur frequently, and abnormal data is relatively abundant, providing more effective samples for model training.

[0004] Therefore, “how to use the rich abnormal data from other regions to improve the fault diagnosis capability of this region when abnormal samples are scarce” is the technical problem that the present invention needs to solve. Summary of the Invention

[0005] The purpose of the present invention is to provide a transformer operation fault detection method and system to solve the problem raised in the above background technology: "how to use the rich abnormal data from other regions to improve the fault diagnosis ability of this region when abnormal samples are scarce."

[0006] To achieve the above object, the present invention provides the following technical solutions: A transformer operation fault detection method, the method comprising: Find transformers within a preset range, set several monitoring indicators, edit evaluation rules, compare monitoring indicators and evaluation rules, and cluster transformers into stable and risk groups; Creating a global model, collecting transformer operating parameters, obtaining historical data of the operating parameters of all transformers in the stable group, and training the global model to obtain a parameter group consisting of several sets of model parameters, wherein each transformer corresponds to a set of model parameters, and each model parameter includes at least: a weight coefficient and a threshold; Collect historical data of all transformers in the risk group, select abnormal data, build a training set, and train a pre-built conditional generation model. Extract the operating parameters of each transformer in the stable group under normal conditions to obtain a target parameter set, add random noise, input it into the conditional generation model, output simulated data, train the global model, and calibrate the parameter set. Based on the parameter group, the mean coefficient is calculated and deployed to the global model to obtain a global perception model. The model parameters are deployed to the global model to obtain a local detection model. The global perception model and the local detection model are integrated to generate a multi-level monitoring architecture, and the operating parameters are written in real time.

[0007] Furthermore, the steps of finding transformers within a preset range, setting a number of monitoring indicators, and editing evaluation rules include: Based on the evaluation rules, a fluctuation range is set for each monitoring indicator, and the operating parameters are compared with the fluctuation range to select abnormal equipment; The attribute data of each transformer is collected, wherein the attribute data includes at least: location and model, the attribute data of the abnormal device is written into a preset template, a maintenance task is generated, and sent to a preset terminal.

[0008] Furthermore, the steps of creating a global model, collecting operating parameters of transformers, obtaining historical data of operating parameters of all transformers in a stable group, and training the global model include: Calculating mathematical characteristics of the operating parameters, wherein the mathematical characteristics include at least: mean, variance, and median; Based on the mathematical characteristics corresponding to each transformer in the stable group, the stable group is refined into several subgroups, and a subgroup-based parameter sharing mechanism is constructed.

[0009] Furthermore, the steps of collecting historical data of all transformers in the risk group, selecting abnormal data, and constructing a training set include: Determine the fault type corresponding to each abnormal data, and divide the training set into several subsets based on the fault type; Insert labels generated by fault types into the subset.

[0010] Furthermore, the steps of calculating the mean coefficient based on the parameter group and deploying it to the global model to obtain the global perception model and deploying the model parameters to the global model to obtain the local detection model include: Collect computing resources within each preset range and select edge nodes; The global perception model is mounted on the edge node, and the local detection model is mounted on the preset cloud node.

[0011] Furthermore, the steps of generating a multi-level monitoring architecture and writing the operating parameters in real time include: Edit federated learning rules and deploy them into the multi-level monitoring architecture; Establish a mapping between preset scopes and multi-level monitoring architecture, and perform federated aggregation.

[0012] Furthermore, the system includes: The clustering module is used to find transformers within a preset range, set several monitoring indicators, edit evaluation rules, compare monitoring indicators and evaluation rules, and cluster transformers into stable groups and risk groups; A module is obtained, which is used to create a global model, collect operating parameters of transformers, obtain historical data of operating parameters of all transformers in the stable group, and train the global model to obtain a parameter group consisting of several sets of model parameters, wherein each transformer corresponds to a set of model parameters, and each model parameter includes at least: a weight coefficient and a threshold; A correction module is used to collect historical data of all transformers in the risk group, select abnormal data, build a training set, and train a pre-built conditional generation model. The operating parameters of each transformer in the stable group under normal conditions are extracted to obtain a target parameter set, and random noise is added. The target parameter set is input into the conditional generation model, and the simulated data is output. The global model is trained to correct the parameter set. A writing module is used to calculate the mean coefficient based on the parameter group and deploy it to the global model to obtain a global perception model, deploy the model parameters to the global model to obtain a local detection model, integrate the global perception model and the local detection model to generate a multi-level monitoring architecture, and write the operating parameters in real time.

[0013] Furthermore, the clustering module includes: A setting unit, configured to set a fluctuation range for each monitoring indicator according to the evaluation rules, compare the operating parameters with the fluctuation range, and select abnormal equipment; The collection unit is used to collect attribute data of each transformer, wherein the attribute data includes at least: location and model, write the attribute data of abnormal equipment into a preset template, generate a maintenance task, and send it to a preset terminal.

[0014] Furthermore, the obtaining module includes: A calculation unit, configured to calculate mathematical characteristics of the operating parameters, wherein the mathematical characteristics include at least a mean, a variance, and a median; The sharing unit is used to refine the stable group into several subgroups according to the mathematical characteristics corresponding to each transformer in the stable group, and to build a parameter sharing mechanism based on the subgroups.

[0015] Furthermore, the correction module includes: A segmentation unit, configured to determine the fault type corresponding to each abnormal data, and based on the fault type, segment the training set into a plurality of subsets; The insertion unit is used to insert the label generated by the fault type into the subset.

[0016] Compared with the prior art, the present invention has the following beneficial effects: By setting monitoring indicators, key parameters can be determined, the operating status of each transformer can be accurately identified, and detection efficiency can be improved. By building a global model and setting model parameters, each transformer can be equipped with personalized parameters to reflect its unique operating characteristics and environmental differences, achieve personalized diagnosis, and improve the accuracy of fault detection. By building a conditional generation model and using the target parameter set for training, the data from the abnormal data-rich area can be used to train the conditional generation model, so that the conditional generation model has the ability to simulate abnormal data and improve the accuracy of the simulated abnormal data. By correcting the parameter group, the simulated abnormal data can be used to train the global model, thereby improving the global model's ability to recognize abnormal data and alleviating the problem of abnormal data scarcity. By building a multi-level monitoring architecture, the global perception model can be used to perform macro-monitoring of the operating parameters of all transformers, preliminarily screen for abnormalities, and when an abnormal signal is detected, the local detection model is triggered to perform in-depth and detailed analysis of the relevant operating parameters, thereby greatly improving the accuracy and reliability of fault identification. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 A flowchart of a transformer operation fault detection method provided by an embodiment of the present invention; Figure 2 A first sub-flow chart of the transformer operation fault detection method provided by an embodiment of the present invention; Figure 3 A second sub-flow chart of the transformer operation fault detection method provided by an embodiment of the present invention; Figure 4 A third sub-flow chart of the transformer operation fault detection method provided by an embodiment of the present invention; Figure 5 A fourth sub-flow chart of the transformer operation fault detection method provided by an embodiment of the present invention; Figure 6 A block diagram of a transformer operation fault detection system provided by an embodiment of the present invention; Figure 7 A block diagram of a clustering module in a transformer operation fault detection system provided by an embodiment of the present invention; Figure 8A block diagram of the components of a module in a transformer operation fault detection system provided by an embodiment of the present invention; Figure 9 A block diagram of the composition of a correction module in a transformer operation fault detection system provided in an embodiment of the present invention; Figure 10 This is a block diagram of the composition of the writing module in the transformer operation fault detection system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0018] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0019] In Example 1, Figure 1 The following is a flowchart of the transformer operation fault detection method provided by an embodiment of the present invention, which is described in detail as follows: S100: Find transformers within a preset range, set several monitoring indicators, edit evaluation rules, compare the monitoring indicators and evaluation rules, and cluster the transformers into stable groups and risk groups.

[0020] Find transformers within a preset range. The preset range here refers to the range where fault detection is required, for example, a substation or administrative area. Set several key monitoring indicators for the transformer, such as temperature, load, current, voltage fluctuation, and partial discharge, and formulate evaluation rules. The evaluation rules are usually based on industry standards, historical equipment performance, and safety thresholds. The evaluation rules can also be jointly formulated by the equipment manufacturer and the operation and maintenance team. The evaluation rules are mainly used to determine abnormalities in the monitoring indicators. By comparing the real-time values ​​of the monitoring indicators with the evaluation rules, the health status of each transformer is identified and the transformers are divided into stable and risk groups.

[0021] For example, suppose an electric power operation and maintenance company has designated an area as a preset range in a city's power grid, found 100 transformers in the area, and set monitoring indicators, specifically: transformer oil temperature, load current, and partial discharge value. The real-time value of each monitoring indicator is collected, and the evaluation rule corresponding to the transformer oil temperature is: if the temperature exceeds 85°C and lasts for more than 15 minutes, there is a safety hazard risk; if the load current exceeds 90% of the rated capacity, it is considered high-load operation; and if the partial discharge value exceeds a certain threshold, it indicates a potential fault. Among them, 10 transformers have experienced safety hazard risks or potential faults as specified in the evaluation rules in the past three days. These 10 transformers are divided into the risk group, and the remaining 90 transformers are divided into the stable group.

[0022] S200: Create a global model, collect operating parameters of transformers, obtain historical data of operating parameters of all transformers in a stable group, and train the global model to obtain a parameter group consisting of several sets of model parameters, wherein each transformer corresponds to a set of model parameters, and each model parameter includes at least: a weight coefficient and a threshold.

[0023] Multiple convolution kernels are used to extract local features of time series data or image data, and these features are fused through multi-layer convolution and pooling operations to capture information of different scales and levels. A global model is constructed through feature integration and abstract representation. The global model is a fault analysis model, which is mainly used to identify potential anomalies in operating parameters. The operating parameters of all transformers are collected. It should be noted that the operating parameters here are the real-time values ​​of the monitoring indicators mentioned above. The historical operating data of all transformers in the stable group under different operating states and environmental conditions are continuously collected, and the global model is trained. During the training process, the weight of each convolution kernel in the model is repeatedly adjusted, and the threshold of each monitoring indicator is fine-tuned. The set composed of weights and thresholds is defined as model parameters, where each transformer corresponds to a set of model parameters. The model parameters corresponding to all transformers are integrated to obtain a parameter group.

[0024] In an actual production environment, due to the different geographical locations, system architectures, and load conditions of transformers, there are certain slight differences in the thresholds of various monitoring indicators. Therefore, after the global model is trained based on their respective historical operating data, the parameters obtained will also be different. In this application, the global model is similar to the initial model. The historical data of each transformer is used to train the initial model to obtain a set of parameters. At the same time, the threshold of each monitoring indicator is adjusted according to the geographical location, system architecture, and load conditions of the transformer. The set of parameters and thresholds obtained after training is defined as model parameters. Each transformer corresponds to a set of model parameters. When in use, the set of model parameters is written into the initialized global model to obtain a unique and personalized fault analysis model.

[0025] S300: Collect historical data of all transformers in the risk group, select abnormal data, build a training set, and train the pre-built conditional generation model, extract the operating parameters of each transformer in the stable group under normal conditions, obtain the target parameter set, add random noise, input it into the conditional generation model, output simulated data, train the global model, and correct the parameter set.

[0026] Collect historical data of all transformers in the risk group, and define historical data with safety hazards or potential fault conditions as abnormal data. Integrate all abnormal data to generate a training set; create a conditional generation model. The conditional generation model can simulate the corresponding abnormal data under specific conditions (geographical location, system architecture, and load conditions, etc.) according to the operating parameters of the transformer in the normal state. The conditional generation model simulates the operating parameters by learning the distribution of normal data. The conditional generation model can be a conditional generation adversarial network in the existing technology, or it can be obtained by conditionally constraining or embedding the ordinary generation model and then adjusting it; for example, in transformer monitoring, the condition is the operating parameters of the transformer, and simulated data that meets the current state characteristics is generated. Through training with a large amount of real data, the conditional generation model can more accurately simulate the operating parameters under abnormal conditions.

[0027] The operating parameters of each transformer in the stable group under normal conditions are collected and integrated to generate a target parameter set. Random noise is added to the target parameter set to simulate the fluctuations of the transformers in the stable group in the actual environment. The target parameter set with noise is input into the conditional generation model. The distribution characteristics of the operating parameters under the learned stable state are used to output simulation data, where the simulation data refers to the simulated operating parameters of the transformers in the stable group. The simulation data is used to train the global model and calibrate the corresponding parameter set.

[0028] In the present application, after determining the model parameters corresponding to each transformer, the operating parameters of each transformer in the stable group under normal conditions are used, combined with the conditions to generate a model, and the abnormal data of the transformers in the stable group are simulated. Each transformer in the stable group corresponds to a set of abnormal data. The corresponding global model is trained using this abnormal data, and the model parameters are adjusted. In other words, using real stable data and simulated abnormal data to train the global model can effectively make up for the problem of insufficient number of abnormal samples of transformers in the stable group, and improve the generalization ability of the global model in identifying abnormal states. Finally, through continuous iterative training, the global model can accurately capture the characteristic changes of the transformer under different operating conditions, and improve the accuracy of fault identification.

[0029] S400: Based on the parameter group, the mean coefficient is calculated and deployed to the global model to obtain a global perception model, the model parameters are deployed to the global model to obtain a local detection model, the global perception model and the local detection model are integrated to generate a multi-level monitoring architecture, and the operating parameters are written in real time.

[0030] The average value of the weight coefficients in all parameter groups is calculated and defined as the mean coefficient. The mean coefficient can highlight the commonality of the overall operating status. The calculated mean coefficient is used as the new weight coefficient and deployed to the initialized global model to obtain the global perception model. The global perception model can roughly monitor and preliminarily warn the operating status of the transformer within the entire preset range, and has strong generalization ability; the global model is initialized again, and the model parameters corresponding to each transformer in the stable group are deployed to the global model to obtain the local detection model. The local detection model has strong pertinence and flexibility, and can be combined with the geographical location, system architecture, and load conditions of the specific transformer to deeply analyze the subtle abnormal characteristics of the transformer, effectively identify early fault signals, and integrate the global perception model and the local detection model to generate a multi-level monitoring architecture.

[0031] In the multi-level monitoring architecture, the global perception model is responsible for comprehensive and rough monitoring and trend analysis of the transformer operating status within the entire preset range, quickly identifying potential risk areas, while the local detection model conducts in-depth anomaly detection and fault diagnosis for a single transformer, combining its specific operating environment and historical behavior data. The multi-level monitoring architecture refers to a progressive monitoring system formed by a macro-monitoring layer composed of the global perception model and a micro-detection layer composed of local detection models. This multi-level monitoring architecture is used to analyze and process the real-time operating parameters generated by the transformer.

[0032] In Example 2, Figure 2 The following is a detailed description of the steps of finding transformers within a preset range, setting several monitoring indicators, and editing evaluation rules. S101: Based on the evaluation rules, a fluctuation range of each monitoring indicator is set, and the operating parameters are compared with the fluctuation range to select abnormal equipment.

[0033] According to the evaluation rules, a corresponding fluctuation range is set for each monitoring indicator, and transformers whose operating parameters exceed the fluctuation range are defined as abnormal equipment.

[0034] S102: Collect attribute data of each transformer, wherein the attribute data includes at least location and model, write the attribute data of abnormal equipment into a preset template, generate a maintenance task, and send it to a preset terminal.

[0035] The attribute data of each transformer is collected. When an abnormal device is identified, the attribute data of the abnormal device is written into the preset task template. The template pre-defines the key fields required for the maintenance task, such as equipment location, model, fault type and recommended treatment plan. By writing the content in the attribute data into the corresponding position in the preset template, a maintenance task is generated and sent to the preset terminal, where the preset terminal can be the mobile workstation, tablet computer or mobile phone terminal of the operation and maintenance personnel.

[0036] In Example 3, Figure 3 The implementation process of the transformer operation fault detection method provided by an embodiment of the present invention is shown. The steps of creating a global model, collecting transformer operating parameters, obtaining historical data of operating parameters of all transformers in a stable group, and training the global model are described in detail below: S201: Calculate mathematical characteristics of operating parameters, wherein the mathematical characteristics at least include: mean, variance, and median.

[0037] The mean, variance and median value of each monitoring indicator in the operating parameters are calculated. In the specific calculation process, the mathematical characteristics of the operating parameters in each time window can be calculated by constructing a time window and multi-point sampling.

[0038] S202: Based on the mathematical characteristics corresponding to each transformer in the stable group, the stable group is refined into several subgroups, and a parameter sharing mechanism based on the subgroups is constructed.

[0039] Based on mathematical characteristics, the transformers in the stable group are further divided into multiple subgroups. For example, transformers whose temperature, load, and RMS current are within preset ranges are merged into one subgroup. At the same time, a parameter sharing mechanism is established. The parameter sharing mechanism refers to the unified management and application of model parameters within the subgroup, so that transformers in the same group can share calculation results and model parameters in model training, operating status prediction, and anomaly detection. This reduces repeated calculations, improves the overall operation analysis efficiency and maintenance decision-making level of the stable group, and improves the generalization ability of the global model.

[0040] In Example 4, Figure 4 The following is a detailed description of the steps for collecting historical data of all transformers in the risk group, selecting abnormal data, and constructing a training set. S301: Determine the fault type corresponding to each abnormal data, and divide the training set into several subsets based on the fault type.

[0041] The fault type corresponding to each abnormal data is determined, such as winding short circuit, insulation aging, high oil temperature or partial discharge abnormality, and the data containing different fault types are divided into categories from the original training set to form several relatively independent subsets, each of which contains only abnormal data of one fault type.

[0042] S302: Insert a label generated by the fault type into the subset.

[0043] Inserting labels generated by fault types into subsets can intuitively display the fault type corresponding to each subset; by splitting the training set, the convergence speed and training accuracy of the global model can be greatly improved, and the fault identification ability of the global model can be enhanced.

[0044] In Example 5, Figure 5 The implementation process of the transformer operation fault detection method provided by an embodiment of the present invention is shown. The following details the steps of calculating the mean coefficient based on the parameter group, deploying it to the global model, obtaining the global perception model, and deploying the model parameters to the global model to obtain the local detection model. S401: Collect computing resources within each preset range and select edge nodes.

[0045] Computing resources within a preset range are collected and counted. Computing resources include but are not limited to servers, edge computing devices, industrial gateways, and smart terminals with computing capabilities. Nodes with low latency, high reliability, and high availability are selected as edge nodes to process transformer operating parameters close to the data source.

[0046] S402: Mount the global perception model to the edge node, and mount the local detection model to the preset cloud node.

[0047] The global perception model is mounted on the edge node to meet the requirements of low latency and bandwidth saving, ensure the timeliness of fault monitoring, and deploy a larger number of local detection models with more intensive computation to the preset cloud nodes (in this application, the number of local detection models is the same as the number of transformers or subgroups): on this basis, the global perception model can be lightweight processed and run on the edge device as a lightweight service; the cloud carries the computing tasks of the personalized local detection model, and uses stronger computing power to perform detailed analysis and processing of the initially detected fault data.

[0048] In Example 6, Figure 5 The implementation process of the transformer operation fault detection method provided by the embodiment of the present invention is shown. The steps of generating a multi-level monitoring architecture and writing the operating parameters in real time are described in detail below: S403: Edit the federated learning rules and deploy them into the multi-level monitoring architecture.

[0049] When using operating parameters to update the model parameters of the global perception model, the operating parameters are not uploaded. Instead, only the model parameters corresponding to each transformer are uploaded. The federated learning rule means that only model parameters are uploaded, not operating parameters.

[0050] S404: Establish a mapping between the preset scope and the multi-level monitoring architecture, and perform federation aggregation.

[0051] Within each preset range, a corresponding multi-level monitoring architecture is constructed, and the model parameters of the corresponding global perception model are updated; after the update, the model parameters of the global perception model within all preset ranges are calculated again to obtain new model parameters. Federal aggregation refers to: merging the model parameters distributed in different preset ranges and obtained through local training according to certain rules to obtain a new global perception model; in other embodiments of the present application, the global perception model corresponds to the preset range (assuming the preset range is a city), and the new global perception model corresponds to all preset ranges (a city or a larger area).

[0052] Figure 6 The following is a structural block diagram of a transformer operation fault detection system according to an embodiment of the present invention. The transformer operation fault detection system 1 includes: Clustering module 11 is used to find transformers within a preset range, set several monitoring indicators, edit evaluation rules, compare monitoring indicators and evaluation rules, and cluster transformers into stable groups and risk groups; Obtaining module 12, for creating a global model, collecting operating parameters of the transformers, obtaining historical data of the operating parameters of all transformers in the stable group, and training the global model to obtain a parameter group consisting of several sets of model parameters, wherein each transformer corresponds to a set of model parameters, and each model parameter includes at least: a weight coefficient and a threshold; The correction module 13 is used to collect historical data of all transformers in the risk group, select abnormal data, build a training set, and train the pre-built conditional generation model. The operating parameters of each transformer in the stable group under normal conditions are extracted to obtain a target parameter set, and random noise is added. The target parameter set is input into the conditional generation model, and the simulated data is output. The global model is trained to correct the parameter set. The writing module 14 is used to calculate the mean coefficient based on the parameter group and deploy it to the global model to obtain a global perception model, deploy the model parameters to the global model to obtain a local detection model, integrate the global perception model and the local detection model to generate a multi-level monitoring architecture, and write the operating parameters in real time.

[0053] Figure 7 The following is a structural block diagram of a transformer operation fault detection system according to an embodiment of the present invention. The clustering module 11 includes: The setting unit 111 is used to set the fluctuation range of each monitoring indicator according to the evaluation rule, compare the operating parameters with the fluctuation range, and select abnormal equipment; The collection unit 112 is configured to collect attribute data of each transformer, wherein the attribute data includes at least location and model, write the attribute data of the abnormal device into a preset template, generate a maintenance task, and send it to a preset terminal.

[0054] Figure 8 The structure block diagram of the transformer operation fault detection system provided by an embodiment of the present invention is shown. The obtaining module 12 includes: A calculation unit 121 is configured to calculate mathematical characteristics of the operating parameters, wherein the mathematical characteristics include at least a mean, a variance, and a median; The sharing unit 122 is used to refine the stable group into several subgroups according to the mathematical characteristics corresponding to each transformer in the stable group, and to establish a parameter sharing mechanism based on the subgroups.

[0055] Figure 9 The following is a structural block diagram of a transformer operation fault detection system according to an embodiment of the present invention. The correction module 13 includes: A segmentation unit 131 is configured to determine the fault type corresponding to each abnormal data, and based on the fault type, segment the training set into a plurality of subsets; The inserting unit 132 is configured to insert a label generated by the fault type into the subset.

[0056] Figure 10 The following is a structural block diagram of a transformer operation fault detection system according to an embodiment of the present invention. The writing module 14 includes: A selection unit 141 is configured to collect computing resources within each preset range and select edge nodes; The mounting unit 142 is used to mount the global perception model to the edge node and the local detection model to the preset cloud node; An editing unit 143, configured to edit federated learning rules and deploy them into the multi-level monitoring architecture; The aggregation unit 144 is used to establish a mapping between a preset range and a multi-level monitoring architecture, and perform federation aggregation.

[0057] The clustering module 11 is mainly used to complete step S100, the obtaining module 12 is mainly used to complete step S200, the correction module 13 is mainly used to complete step S300, and the writing module 14 is mainly used to complete step S400; The setting unit 111 is mainly used to complete step S101, and the collection unit 112 is mainly used to complete step S102; The calculation unit 121 is mainly used to complete step S201, and the sharing unit 122 is mainly used to complete step S202; The segmentation unit 131 is mainly used to complete step S301, and the insertion unit 132 is mainly used to complete step S302; The selection unit 141 is mainly used to complete step S401, the mounting unit 142 is mainly used to complete step S402, the editing unit 143 is mainly used to complete step S403, and the aggregation unit 144 is mainly used to complete step S404.

[0058] The technical features of the above-mentioned embodiments can be combined arbitrarily. In order to make the description concise, not all possible combinations of the technical features in the above-mentioned embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0059] The above-described embodiments merely illustrate several implementations of the present invention, and while their descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that a person skilled in the art would be able to make numerous variations and improvements without departing from the spirit of the present invention, all of which fall within the scope of protection of the present invention. Therefore, the scope of protection of the present invention shall be determined by the appended claims.

[0060] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A transformer operation fault detection method, characterized in that: The method comprises: Find transformers within a preset range, set several monitoring indicators, edit evaluation rules, compare monitoring indicators and evaluation rules, and cluster transformers into stable and risk groups; Creating a global model, collecting transformer operating parameters, obtaining historical data of the operating parameters of all transformers in the stable group, and training the global model to obtain a parameter group consisting of several sets of model parameters, wherein each transformer corresponds to a set of model parameters, and each model parameter includes at least: a weight coefficient and a threshold; Collect historical data of all transformers in the risk group, select abnormal data, build a training set, and train a pre-built conditional generation model. Extract the operating parameters of each transformer in the stable group under normal conditions to obtain a target parameter set, add random noise, input it into the conditional generation model, output simulated data, train the global model, and calibrate the parameter set. Based on the parameter group, the mean coefficient is calculated and deployed to the global model to obtain a global perception model. The model parameters are deployed to the global model to obtain a local detection model. The global perception model and the local detection model are integrated to generate a multi-level monitoring architecture, and the operating parameters are written in real time.

2. The transformer operation fault detection method according to claim 1, characterized in that: The steps of finding transformers within a preset range, setting a number of monitoring indicators, and editing evaluation rules include: Based on the evaluation rules, a fluctuation range is set for each monitoring indicator, and the operating parameters are compared with the fluctuation range to select abnormal equipment; The attribute data of each transformer is collected, wherein the attribute data includes at least: location and model, the attribute data of the abnormal device is written into a preset template, a maintenance task is generated, and sent to a preset terminal.

3. The transformer operation fault detection method according to claim 1, characterized in that: The steps of creating a global model, collecting operating parameters of transformers, obtaining historical data of operating parameters of all transformers in a stable group, and training the global model include: Calculating mathematical characteristics of the operating parameters, wherein the mathematical characteristics include at least: mean, variance, and median; Based on the mathematical characteristics corresponding to each transformer in the stable group, the stable group is refined into several subgroups, and a subgroup-based parameter sharing mechanism is constructed.

4. The transformer operation fault detection method according to claim 1, characterized in that: The steps of collecting historical data of all transformers in the risk group, selecting abnormal data, and constructing a training set include: Determine the fault type corresponding to each abnormal data, and divide the training set into several subsets based on the fault type; Insert labels generated by fault types into the subset.

5. The transformer operation fault detection method according to claim 1, characterized in that: The steps of calculating the mean coefficient based on the parameter group and deploying it to the global model to obtain the global perception model, and deploying the model parameters to the global model to obtain the local detection model include: Collect computing resources within each preset range and select edge nodes; The global perception model is mounted on the edge node, and the local detection model is mounted on the preset cloud node.

6. The transformer operation fault detection method according to claim 5, characterized in that: The steps of generating a multi-level monitoring architecture and writing the operating parameters in real time include: Edit federated learning rules and deploy them into the multi-level monitoring architecture; Establish a mapping between preset scopes and multi-level monitoring architecture, and perform federated aggregation.

7. A transformer operation fault detection system, characterized in that: The system comprises: The clustering module is used to find transformers within a preset range, set several monitoring indicators, edit evaluation rules, compare monitoring indicators and evaluation rules, and cluster transformers into stable groups and risk groups; A module is obtained, which is used to create a global model, collect operating parameters of transformers, obtain historical data of operating parameters of all transformers in the stable group, and train the global model to obtain a parameter group consisting of several sets of model parameters, wherein each transformer corresponds to a set of model parameters, and each model parameter includes at least: a weight coefficient and a threshold; A correction module is used to collect historical data of all transformers in the risk group, select abnormal data, build a training set, and train a pre-built conditional generation model. The operating parameters of each transformer in the stable group under normal conditions are extracted to obtain a target parameter set, and random noise is added. The target parameter set is input into the conditional generation model, and the simulated data is output. The global model is trained to correct the parameter set. A writing module is used to calculate the mean coefficient based on the parameter group and deploy it to the global model to obtain a global perception model, deploy the model parameters to the global model to obtain a local detection model, integrate the global perception model and the local detection model to generate a multi-level monitoring architecture, and write the operating parameters in real time.

8. The transformer operation fault detection system according to claim 7, characterized in that: The clustering module includes: A setting unit, configured to set a fluctuation range for each monitoring indicator according to the evaluation rules, compare the operating parameters with the fluctuation range, and select abnormal equipment; The collection unit is used to collect attribute data of each transformer, wherein the attribute data includes at least: location and model, write the attribute data of abnormal equipment into a preset template, generate a maintenance task, and send it to a preset terminal.

9. The transformer operation fault detection system according to claim 7, characterized in that: The obtaining module includes: A calculation unit, configured to calculate mathematical characteristics of the operating parameters, wherein the mathematical characteristics include at least a mean, a variance, and a median; The sharing unit is used to refine the stable group into several subgroups according to the mathematical characteristics corresponding to each transformer in the stable group, and to build a parameter sharing mechanism based on the subgroups.

10. The transformer operation fault detection system according to claim 7, characterized in that: The correction module includes: A segmentation unit, configured to determine the fault type corresponding to each abnormal data, and based on the fault type, segment the training set into a plurality of subsets; The insertion unit is used to insert the label generated by the fault type into the subset.

Images