A transformer monitoring method and system based on cloud-edge fusion

Abstract

The present invention discloses a transformer monitoring method and system based on cloud-edge fusion, and relates to the technical field of transformer monitoring. The method comprises: collecting and acquiring a historical fault database of a target transformer, performing statistical identification, and obtaining fault occurrence parameter information; locating and acquiring multiple transformer monitoring points, and respectively deploying edge sensors and edge computing devices at multiple transformer monitoring points; acquiring multiple point sensor monitoring data streams, performing mapping anomaly analysis, and obtaining multiple point abnormal edge feature information; performing anomaly fusion analysis, and obtaining transformer abnormal fault parameters; performing cluster parameter optimization based on transformer abnormal fault parameters, obtaining operation and maintenance optimization control parameters, and performing fault operation and maintenance management and control on the target transformer based on the operation and maintenance optimization control parameters. The technical problem of insufficient real-time and accuracy of transformer monitoring in the prior art is solved, and the technical effect of improving the real-time and accuracy of transformer monitoring is achieved through cloud-edge fusion technology.

Description

Technical Field

[0001] The present invention relates to the technical field of transformer monitoring, and in particular to a transformer monitoring method and system based on cloud-edge fusion. Background Art

[0002] With the continuous expansion and complexity of the power system, the monitoring and fault diagnosis of the operating status of transformers, as key equipment in power transmission, are particularly important. Traditional transformer monitoring methods usually rely on a single data acquisition method and a centralized data processing mode, which is difficult to meet the requirements of real-time and high efficiency. Especially when facing complex fault modes, traditional methods often find it difficult to accurately locate the fault point or predict potential faults. This limitation easily leads to the accumulation of hidden faults, making the operation and maintenance of transformers more difficult, and when a fault occurs, it may cause a delay in maintenance response, increasing the uncertainty and risk of system operation. Therefore, how to improve the real-time and accuracy of transformer fault monitoring, and how to achieve intelligent management of operation and maintenance through advanced technical means, has become an urgent problem to be solved in the current technical field. Summary of the invention

[0003] The present application provides a transformer monitoring method and system based on cloud-edge fusion, which solves the technical problems of insufficient real-time and accuracy of transformer monitoring in the prior art.

[0004] In view of the above problems, the present application provides a transformer monitoring method and system based on cloud-edge fusion.

[0005] In a first aspect of the present application, a transformer monitoring method based on cloud-edge fusion is provided, the method comprising:

[0006] Acquire a historical fault database of a target transformer, perform statistical identification on the historical fault database, and obtain fault parameter information; locate and acquire multiple transformer monitoring points based on the fault parameter information, and respectively deploy edge sensors and edge computing devices at the multiple transformer monitoring points, wherein the edge sensors include ultra-high frequency sensors, ultrasonic sensors, and high-frequency current transformers; acquire multiple point sensor monitoring data streams through edge sensor monitoring, and perform mapping anomaly analysis on the multiple point sensor monitoring data streams based on the edge computing device to obtain multiple point abnormal edge feature information; integrate and transmit the multiple point abnormal edge feature information to a cloud processor through an optical fiber link for anomaly fusion analysis to obtain transformer abnormal fault parameters; perform cluster parameter optimization based on the transformer abnormal fault parameters to obtain operation and maintenance optimization control parameters, and perform fault operation and maintenance management and control on the target transformer based on the operation and maintenance optimization control parameters.

[0007] A second aspect of the present application provides a transformer monitoring system based on cloud-edge fusion, the system comprising:

[0008] A statistical module, the statistical module is used to collect and obtain a historical fault database of a target transformer, perform statistical identification on the historical fault database, and obtain fault parameter information; a positioning module, the positioning module is used to locate and obtain multiple transformer monitoring points according to the fault parameter information, and respectively deploy edge sensors and edge computing devices at the multiple transformer monitoring points, wherein the edge sensors include ultra-high frequency sensors, ultrasonic sensors, and high-frequency current transformers; a first analysis module, the first analysis module is used to obtain multiple point sensor monitoring data streams through the edge sensor monitoring, and perform mapping abnormality analysis on the multiple point sensor monitoring data streams based on the edge computing device to obtain multiple point abnormal edge feature information; a second analysis module, the second analysis module is used to integrate and transmit the multiple point abnormal edge feature information to the cloud processor through an optical fiber link for abnormal fusion analysis to obtain transformer abnormal fault parameters; an optimization module, the optimization module is used to optimize cluster parameters based on the transformer abnormal fault parameters, obtain operation and maintenance optimization control parameters, and perform fault operation and maintenance management and control on the target transformer based on the operation and maintenance optimization control parameters.

[0009] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0010] First, the historical fault database of the target transformer is collected and obtained, and the historical fault database is statistically identified to obtain the fault parameter information. Then, according to the fault parameter information, multiple transformer monitoring points are located and obtained, and edge sensors and edge computing devices are deployed at multiple transformer monitoring points respectively, wherein the edge sensors include ultra-high frequency sensors, ultrasonic sensors and high-frequency current transformers. Further, multiple point sensor monitoring data streams are obtained through edge sensor monitoring, and the mapping abnormality analysis of the multiple point sensor monitoring data streams is performed based on the edge computing device to obtain multiple point abnormal edge feature information. Then, the multiple point abnormal edge feature information is integrated and transmitted to the cloud processor through the optical fiber link for abnormal fusion analysis to obtain the transformer abnormal fault parameters. Finally, cluster parameter optimization is performed based on the transformer abnormal fault parameters to obtain the operation and maintenance optimization control parameters, and the target transformer is controlled for fault operation and maintenance based on the operation and maintenance optimization control parameters. The technical problem of insufficient real-time and accuracy of transformer monitoring in the prior art is solved, and the technical effect of improving the real-time and accuracy of transformer monitoring is achieved through cloud-edge fusion technology. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0012] Figure 1 A schematic diagram of a transformer monitoring method based on cloud-edge fusion provided in an embodiment of the present application;

[0013] Figure 2 A schematic diagram of the structure of a transformer monitoring system based on cloud-edge fusion provided in an embodiment of the present application.

[0014] Explanation of the reference numerals: statistical module 11 , positioning module 12 , first analysis module 13 , second analysis module 14 , optimization module 15 . DETAILED DESCRIPTION

[0015] The present application solves the technical problems of insufficient real-time and accuracy of transformer monitoring in the prior art by providing a transformer monitoring method and system based on cloud-edge fusion.

[0016] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0017] It should be noted that the terms "including" and "having" are intended to cover non-exclusive inclusions. For example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules that are not explicitly listed or are inherent to these processes, methods, products or devices.

[0018] Embodiment 1, as Figure 1 As shown, the present application provides a transformer monitoring method based on cloud-edge fusion, wherein the method includes:

[0019] A historical fault database of the target transformer is collected and acquired, and statistical identification is performed on the historical fault database to obtain fault occurrence parameter information.

[0020] The historical fault database of the target transformer is obtained by interacting with the target transformer. The historical fault database contains the fault type, occurrence time, fault location, fault cause, etc. that occurred during the past operation of the transformer. Statistical analysis is performed on the historical fault database to identify parameters such as high-incidence areas of faults and common fault modes. These parameters constitute the fault occurrence parameter information.

[0021] According to the fault parameter information, multiple transformer monitoring points are located and acquired, and edge sensors and edge computing devices are respectively deployed at the multiple transformer monitoring points, wherein the edge sensors include ultra-high frequency sensors, ultrasonic sensors and high-frequency current transformers.

[0022] Based on the fault parameter information, determine the areas or parts of the transformer that need to be monitored. These areas may include the winding part, insulation layer, connection point and other locations prone to faults of the transformer; deploy edge sensors and edge computing devices at these determined monitoring points, where edge sensors include ultra-high frequency sensors, ultrasonic sensors and high-frequency current transformers. Ultra-high frequency sensors are used to detect early signs of discharge phenomena inside electrical equipment, such as partial discharge signals; ultrasonic sensors are used to detect mechanical stress or structural damage, such as loose connection points or mechanical failures inside the transformer; high-frequency current transformers are used to monitor high-frequency current signals, which may be related to electrical faults of the equipment (such as short circuits or arcs).

[0023] Furthermore, the deploying edge sensors at the plurality of transformer monitoring points respectively includes:

[0024] According to the transformer operation characteristic information, monitoring demand analysis is performed on the multiple transformer monitoring points respectively to obtain operation demand information of the multiple monitoring points; according to the operation demand information of the multiple monitoring points, performance parameters of multiple monitoring point sensors are determined, and the multiple monitoring point sensor performance parameters include measurement range, type, accuracy, response time and sensitivity; structural design information of the target transformer is obtained, and three-dimensional modeling is performed on the structural design information to generate a three-dimensional space model of the transformer; according to the multiple monitoring point sensor performance parameters, sensor layout optimization analysis is performed on the transformer three-dimensional space model to obtain multiple monitoring point sensor layout schemes; based on the multiple monitoring point sensor performance parameters and the multiple monitoring point sensor layout schemes, edge sensors are deployed for the multiple transformer monitoring points respectively.

[0025] Specifically, according to the operating characteristic information of the transformer, a detailed monitoring demand analysis is performed on each monitoring point. The operating characteristic information includes key parameters such as voltage, current, temperature, vibration, etc. By analyzing the operating characteristic information, the operating demand information of each monitoring point can be determined, such as which parameters need to be monitored, the frequency and range of monitoring, etc.; according to the operating demand information of multiple monitoring points, appropriate sensor performance parameters are determined for each monitoring point, including measurement range (the range of parameters that the sensor can monitor), type (type of sensor), accuracy (accuracy of sensor measurement), response time (the speed at which the sensor responds to external changes) and sensitivity (the sensor's ability to sense small changes); by obtaining the structural design information of the target transformer (such as the transformer's appearance, internal structure, component distribution, etc.), and performing three-dimensional modeling, the three-dimensional transformer Spatial model: The three-dimensional spatial model of the transformer intuitively displays the spatial distribution inside and outside the transformer; using the determined sensor performance parameters, the three-dimensional spatial model is optimized for sensor layout analysis, that is, combining the sensor performance parameters with the structural characteristics of the transformer to ensure that the sensor can be deployed at the most suitable monitoring point location, thereby maximizing monitoring efficiency and effectiveness; based on the layout optimization analysis, a sensor layout plan for multiple monitoring points is generated. The sensor layout plan includes the sensor installation location, quantity, direction, etc., to ensure that the monitoring area can be fully covered; according to the generated sensor layout plan and performance parameter requirements, the corresponding edge sensor is deployed at each transformer monitoring point. Through this precise deployment, it is ensured that each monitoring point can obtain high-quality monitoring data in real time, and provide a reliable data basis for subsequent edge computing and fault analysis.

[0026] Furthermore, the obtaining of a plurality of sensor layout schemes for monitoring points includes:

[0027] A preliminary analysis of the sensor layout of the transformer three-dimensional space model is performed according to the sensor performance parameters of the multiple monitoring points to obtain the initial layout parameters of the multiple monitoring points; a set of sensor layout effect indicators is obtained, and a simulation evaluation is performed on the initial layout parameters of the multiple monitoring points based on the sensor layout effect indicator set to obtain transformer monitoring effect information; layout parameter optimization analysis is performed based on the transformer monitoring effect information to determine the optimization direction of the sensor layout parameters; the initial layout parameters of the multiple monitoring points are optimized, adjusted and compared according to the sensor layout parameter optimization direction to obtain the sensor layout scheme of the multiple monitoring points.

[0028] Specifically, based on the performance parameters of sensors at multiple monitoring points, a preliminary analysis of the sensor layout of the transformer three-dimensional spatial model is performed, and multiple initial layout parameters of the monitoring points are generated, including the preliminary position, direction and installation angle of the sensor at each monitoring point; a set of indicators for evaluating the effect of sensor layout is predefined, and the sensor layout effect indicator set includes monitoring coverage, data accuracy, fault detection sensitivity, etc. These indicators will be used to evaluate the effectiveness of the initial layout parameters; the initial layout parameters obtained from the preliminary analysis are simulated and evaluated using the sensor layout effect indicator set, and the performance of the sensor layout of each monitoring point in actual monitoring is evaluated through tests in a virtual environment. The simulation evaluation results will generate transformer monitoring effect information. , reflecting the advantages and disadvantages of the layout, including whether the sensors can effectively cover the key areas and whether they can meet the real-time monitoring needs; based on the monitoring effect information obtained by simulation, the layout parameters are optimized and analyzed. The goal of the optimization analysis is to determine how to adjust the sensor layout (change the position of the sensor, adjust the installation angle, increase or decrease certain sensors, etc.) to improve the monitoring effect; according to the determined optimization direction, the initial layout parameters are optimized and adjusted. Through repeated comparison and optimization, the optimal layout solution can be gradually approached. Each time the optimized layout parameters are re-simulated and evaluated to verify the optimization effect and ensure that the layout adjustment can indeed improve the monitoring effect; after multiple optimizations and evaluations, the sensor layout solution for multiple monitoring points is finally obtained.

[0029] Multiple point sensor monitoring data streams are acquired through the edge sensor monitoring, and mapping anomaly analysis is performed on the multiple point sensor monitoring data streams based on the edge computing device to obtain multiple point abnormal edge feature information.

[0030] By deploying edge sensors at multiple monitoring points of the transformer, the monitoring data stream of each point can be obtained in real time. The monitoring data streams of multiple point sensors are transmitted to the edge computing device for mapping anomaly analysis, identifying abnormal signals that may exist in the data stream, and obtaining abnormal edge feature information of multiple points. The abnormal edge feature information includes abnormal current peak, abnormal vibration frequency, temperature changes beyond the normal range, etc.

[0031] Furthermore, the step of obtaining abnormal edge feature information of multiple points includes:

[0032] Through the integrated built-in chip of the edge computing device, the sensor data processing multi-channel is called, and the sensor data processing multi-channel includes a data preprocessing node and a data anomaly analysis node; the multiple point sensor monitoring data streams are mapped and allocated with the sensor data processing multi-channel to obtain calibration data processing channel information; based on the calibration data processing channel information, the multiple point sensor monitoring data streams are analyzed for data anomalies in parallel to obtain the multiple point abnormal edge feature information.

[0033] Specifically, the edge computing device integrates an efficient built-in chip, which can realize parallel processing of sensor data by calling the sensor data processing multi-channel of the chip; the sensor data processing multi-channel includes a data preprocessing node and a data anomaly analysis node, wherein the data preprocessing node is used to perform preliminary processing on the original sensor data, such as data cleaning, denoising, normalization, etc., to ensure data quality; the data anomaly analysis node is used to perform anomaly analysis on the preprocessed data. The sensor monitoring data stream obtained from multiple monitoring points is allocated to each node of the sensor data processing multi-channel, and the allocated data is recorded to obtain the calibration data processing channel information; based on the calibration data processing channel information, the data stream of multiple monitoring points is analyzed in parallel for data anomalies, that is, the data of different monitoring points are analyzed simultaneously, including identifying abnormal fluctuations in parameters such as current, voltage, temperature, or detecting abnormal changes in vibration frequency, and then obtaining abnormal edge feature information of multiple points, and the abnormal edge feature information of the point includes the identified abnormal peaks, frequency fluctuations, temperature exceeding the standard, etc.

[0034] Furthermore, the calling of sensor data processing multiple channels includes:

[0035] Determine a set of monitoring data types according to the edge sensor, perform noise characteristic analysis on the monitoring data type set, and obtain a monitoring data noise characteristic set; perform filtering algorithm configuration and standardization program association based on the monitoring data noise characteristic set to obtain a data preprocessing node; perform anomaly detection demand analysis on the monitoring data type set, construct an abnormal feature task list, and obtain a data anomaly analysis node based on the abnormal feature task list training; construct the sensor data processing multi-channel based on the data preprocessing node and the data anomaly analysis node, and store the sensor data processing multi-channel configuration in the integrated built-in chip.

[0036] Specifically, according to the data collected by the edge sensor, the set of monitoring data types that need to be processed is determined. The monitoring data type set includes various sensor data, such as current, voltage, temperature, vibration signals, etc. Each type of data has different characteristics and processing requirements; for the determined monitoring data type set, noise characteristic analysis is performed to identify the noise mode and characteristics in the data, such as environmental noise, electromagnetic interference, etc., and then the monitoring data noise characteristic set is obtained. The monitoring data noise characteristic set contains the noise characteristic information of each data type; based on the monitoring data noise characteristic set, appropriate filtering algorithms are configured to remove noise in the data and improve the clarity and accuracy of the data. At the same time, standardized procedures are configured for different types of data, such as spatial alignment, normalization, dimensionality reduction, etc., to ensure the uniformity of data format and facilitate subsequent analysis. These configurations ultimately construct a data preprocessing node to ensure that the original data has been cleaned and standardized before entering deep analysis. Anomaly detection requirements analysis is performed on the monitoring data type set to identify and determine the anomaly types and characteristics that need to be detected, such as anomaly location, intensity, duration, etc., so as to construct an anomaly feature task list; according to the anomaly feature task list, the anomaly feature analysis model set is obtained by training historical data separately and the analysis node configuration is performed to obtain a data anomaly analysis node. By combining the preprocessing node and the anomaly analysis node, a sensor data processing multi-channel is formed. The sensor data processing multi-channel integrates the data preprocessing and anomaly analysis functions to ensure that data can be processed efficiently and accurately throughout the entire process from collection to analysis. After completing the construction of the sensor data processing multi-channel, the sensor data processing multi-channel is stored in the integrated built-in chip of the edge computing device.

[0037] The abnormal edge feature information of the multiple points is integrated and transmitted to the cloud processor through the optical fiber link for abnormal fusion analysis to obtain the abnormal fault parameters of the transformer.

[0038] The abnormal feature information of multiple monitoring points is integrated, including the unification of data formats, synchronization of timestamps, and deduplication and integration of information, to ensure that all feature information can form a complete data set; the integrated abnormal feature information is transmitted to the cloud processor via a high-speed optical fiber link. The high bandwidth and low latency of the optical fiber link ensure that large amounts of abnormal information can be quickly and reliably transmitted to the cloud; on the cloud processor, the received multi-point abnormal edge feature information is fused and analyzed to obtain the abnormal fault parameters of the transformer.

[0039] Furthermore, the obtaining of abnormal fault parameters of the transformer includes:

[0040] Obtain transformer fault factor information, wherein the transformer fault factor information includes fault type, fault location, fault degree and impact range; extract fault factors from the abnormal edge feature information of the multiple points according to the transformer fault factor information, and obtain multiple fault-related factor feature sets; perform node fault classification and associated topology analysis on the historical fault database based on the transformer fault factor information, and construct a transformer cascade cause fault tree; investigate the fault generation causes of the multiple fault-related factor feature sets based on the transformer cascade cause fault tree, and obtain the transformer abnormal fault parameters.

[0041] Specifically, comprehensive transformer fault factor information is obtained, including fault type (such as overload, short circuit, insulation damage, etc.), fault location (the specific location or component where the fault occurs, such as winding, insulation layer, connection point, etc.), fault severity (such as minor fault, critical fault or serious fault) and impact range (the range that the fault may affect, such as a single component, the entire transformer); based on the transformer fault factor information, the abnormal edge feature information of multiple points is analyzed, the features related to the fault are extracted, and these features are classified and sorted according to factors such as fault type, location, severity, impact range, etc., to form multiple fault-related factor feature sets.

[0042] The historical fault database is deeply analyzed using the transformer fault factor information. Specifically, the faults in the historical data are classified, and through the associated topological analysis, a topological structure representing the fault relationship between the nodes of the transformer is constructed, that is, the transformer cascade cause fault tree. The transformer cascade cause fault tree can reveal the causal relationship between different faults and the potential fault propagation path. Based on the constructed transformer cascade cause fault tree, the multiple fault association factor feature sets currently extracted are checked and analyzed; by checking the nodes and paths of the fault tree layer by layer, the root causes that may cause the current abnormality are checked; the results of fault factor extraction and cause troubleshooting are combined to generate transformer abnormal fault parameters, including specific fault categories (such as winding overheating), the precise location of the fault, the severity level of the fault, and the possible scope of impact.

[0043] Cluster parameter optimization is performed based on the abnormal fault parameters of the transformer to obtain operation and maintenance optimization control parameters, and fault operation and maintenance management and control are performed on the target transformer based on the operation and maintenance optimization control parameters.

[0044] Based on the obtained transformer abnormal fault parameters, cluster parameter optimization is performed. The goal of cluster parameter optimization is to find the optimal operation and maintenance strategy by analyzing the data of multiple transformers or multiple fault instances; based on the cluster parameter optimization, operation and maintenance optimization control parameters are generated; based on the generated operation and maintenance optimization control parameters, fault operation and maintenance control of the target transformer is performed.

[0045] Furthermore, obtaining the operation and maintenance optimization control parameters includes:

[0046] The transformer expert system is used to perform operation and maintenance strategy analysis on the abnormal fault parameters of the transformer to obtain multiple initial transformer operation and maintenance strategy parameters; the multiple initial transformer operation and maintenance strategy parameters are evaluated and selected to determine a parent strategy parameter set; based on the parent strategy parameter set, cross-combination and random perturbation are performed to perform child expansion and update on the multiple initial transformer operation and maintenance strategy parameters to obtain a transformer operation and maintenance strategy parameter space; a global optimization comparison is performed in the transformer operation and maintenance strategy parameter space to obtain the operation and maintenance optimization control parameters.

[0047] Specifically, the transformer expert system is used to analyze the extracted transformer abnormal fault parameters. The transformer expert system covers multiple experts. Based on the rich transformer operation and maintenance knowledge and experience of these experts, the operation and maintenance strategy analysis of the transformer abnormal fault parameters is carried out, and multiple initial transformer operation and maintenance strategy parameters are formulated; based on multiple indicators, such as the effectiveness of the strategy, the implementation cost, the impact on the life of the transformer, etc., multiple initial transformer operation and maintenance strategy parameters are evaluated, the strategy parameters with the best performance are selected, and these parameters are combined into a parent strategy parameter set; based on the parent strategy parameter set, the child strategy is generated by cross-combination and random perturbation methods. Cross-combination refers to the re-combination of the parent strategy parameters to form a new strategy parameter set; random perturbation is a small random adjustment based on the existing parameters to explore new optimization space. Through these operations, the transformer operation and maintenance strategy parameters are continuously expanded and updated to form a larger and more diverse transformer operation and maintenance strategy parameter space; in the generated transformer operation and maintenance strategy parameter space, global optimization is performed to find the best combination in terms of various indicators by traversing and comparing all possible operation and maintenance strategy parameters; after global optimization comparison, a set of operation and maintenance optimization control parameters is finally determined.

[0048] In summary, the embodiments of the present application have at least the following technical effects:

[0049] First, the historical fault database of the target transformer is collected and obtained, and the historical fault database is statistically identified to obtain the fault parameter information. Then, according to the fault parameter information, multiple transformer monitoring points are located and obtained, and edge sensors and edge computing devices are deployed at multiple transformer monitoring points respectively, wherein the edge sensors include ultra-high frequency sensors, ultrasonic sensors and high-frequency current transformers. Further, multiple point sensor monitoring data streams are obtained through edge sensor monitoring, and the mapping abnormality analysis of the multiple point sensor monitoring data streams is performed based on the edge computing device to obtain multiple point abnormal edge feature information. Then, the multiple point abnormal edge feature information is integrated and transmitted to the cloud processor through the optical fiber link for abnormal fusion analysis to obtain the transformer abnormal fault parameters. Finally, cluster parameter optimization is performed based on the transformer abnormal fault parameters to obtain the operation and maintenance optimization control parameters, and the target transformer is controlled for fault operation and maintenance based on the operation and maintenance optimization control parameters. The technical problem of insufficient real-time and accuracy of transformer monitoring in the prior art is solved, and the technical effect of improving the real-time and accuracy of transformer monitoring is achieved through cloud-edge fusion technology.

[0050] Embodiment 2 is based on the same inventive concept as a transformer monitoring method based on cloud-edge fusion in the above embodiment. Figure 2 As shown, the present application provides a transformer monitoring system based on cloud-edge fusion, wherein the system includes:

[0051] A statistical module 11, the statistical module 11 is used to collect and obtain a historical fault database of a target transformer, perform statistical identification on the historical fault database, and obtain fault parameter information; a positioning module 12, the positioning module 12 is used to locate and obtain multiple transformer monitoring points according to the fault parameter information, and respectively deploy edge sensors and edge computing devices at the multiple transformer monitoring points, wherein the edge sensors include ultra-high frequency sensors, ultrasonic sensors, and high-frequency current transformers; a first analysis module 13, the first analysis module 13 is used to obtain multiple point sensor monitoring data streams through the edge sensor monitoring, and perform mapping abnormality analysis on the multiple point sensor monitoring data streams based on the edge computing device to obtain multiple point abnormal edge feature information; a second analysis module 14, the second analysis module 14 is used to integrate and transmit the multiple point abnormal edge feature information to the cloud processor through an optical fiber link for abnormal fusion analysis to obtain transformer abnormal fault parameters; an optimization module 15, the optimization module 15 is used to optimize cluster parameters based on the transformer abnormal fault parameters, obtain operation and maintenance optimization control parameters, and perform fault operation and maintenance management and control on the target transformer based on the operation and maintenance optimization control parameters.

[0052] Furthermore, the positioning module 12 is used to perform the following method:

[0053] According to the transformer operation characteristic information, monitoring demand analysis is performed on the multiple transformer monitoring points respectively to obtain operation demand information of the multiple monitoring points; according to the operation demand information of the multiple monitoring points, performance parameters of multiple monitoring point sensors are determined, and the multiple monitoring point sensor performance parameters include measurement range, type, accuracy, response time and sensitivity; structural design information of the target transformer is obtained, and three-dimensional modeling is performed on the structural design information to generate a three-dimensional space model of the transformer; according to the multiple monitoring point sensor performance parameters, sensor layout optimization analysis is performed on the transformer three-dimensional space model to obtain multiple monitoring point sensor layout schemes; based on the multiple monitoring point sensor performance parameters and the multiple monitoring point sensor layout schemes, edge sensors are deployed for the multiple transformer monitoring points respectively.

[0054] Furthermore, the positioning module 12 is used to perform the following method:

[0055] A preliminary analysis of the sensor layout of the transformer three-dimensional space model is performed according to the sensor performance parameters of the multiple monitoring points to obtain the initial layout parameters of the multiple monitoring points; a set of sensor layout effect indicators is obtained, and a simulation evaluation is performed on the initial layout parameters of the multiple monitoring points based on the sensor layout effect indicator set to obtain transformer monitoring effect information; layout parameter optimization analysis is performed based on the transformer monitoring effect information to determine the optimization direction of the sensor layout parameters; the initial layout parameters of the multiple monitoring points are optimized, adjusted and compared according to the sensor layout parameter optimization direction to obtain the sensor layout scheme of the multiple monitoring points.

[0056] Furthermore, the first analysis module 13 is used to perform the following method:

[0057] Through the integrated built-in chip of the edge computing device, the sensor data processing multi-channel is called, and the sensor data processing multi-channel includes a data preprocessing node and a data anomaly analysis node; the multiple point sensor monitoring data streams are mapped and allocated with the sensor data processing multi-channel to obtain calibration data processing channel information; based on the calibration data processing channel information, the multiple point sensor monitoring data streams are analyzed for data anomalies in parallel to obtain the multiple point abnormal edge feature information.

[0058] Furthermore, the first analysis module 13 is used to perform the following method:

[0059] Determine a set of monitoring data types according to the edge sensor, perform noise characteristic analysis on the monitoring data type set, and obtain a monitoring data noise characteristic set; perform filtering algorithm configuration and standardization program association based on the monitoring data noise characteristic set to obtain a data preprocessing node; perform anomaly detection demand analysis on the monitoring data type set, construct an abnormal feature task list, and obtain a data anomaly analysis node based on the abnormal feature task list training; construct the sensor data processing multi-channel based on the data preprocessing node and the data anomaly analysis node, and store the sensor data processing multi-channel configuration in the integrated built-in chip.

[0060] Furthermore, the second analysis module 14 is used to perform the following method:

[0061] Obtain transformer fault factor information, wherein the transformer fault factor information includes fault type, fault location, fault degree and impact range; extract fault factors from the abnormal edge feature information of the multiple points according to the transformer fault factor information, and obtain multiple fault-related factor feature sets; perform node fault classification and associated topology analysis on the historical fault database based on the transformer fault factor information, and construct a transformer cascade cause fault tree; investigate the fault generation causes of the multiple fault-related factor feature sets based on the transformer cascade cause fault tree, and obtain the transformer abnormal fault parameters.

[0062] Furthermore, the optimization module 15 is used to perform the following method:

[0063] The transformer expert system is used to perform operation and maintenance strategy analysis on the abnormal fault parameters of the transformer to obtain multiple initial transformer operation and maintenance strategy parameters; the multiple initial transformer operation and maintenance strategy parameters are evaluated and selected to determine a parent strategy parameter set; based on the parent strategy parameter set, cross-combination and random perturbation are performed to perform child expansion and update on the multiple initial transformer operation and maintenance strategy parameters to obtain a transformer operation and maintenance strategy parameter space; a global optimization comparison is performed in the transformer operation and maintenance strategy parameter space to obtain the operation and maintenance optimization control parameters.

[0064] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description and does not represent the advantages and disadvantages of the embodiments. And the above-mentioned specific embodiments of this specification are described. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0065] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

[0066] This specification and drawings are merely exemplary illustrations of the present application and are deemed to cover any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, a person skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application intends to include these modifications and variations.

Claims

1. A transformer monitoring method based on cloud-edge fusion, characterized in that: The method comprises: Collect and obtain a historical fault database of the target transformer, perform statistical identification on the historical fault database, and obtain fault occurrence parameter information; According to the fault parameter information, multiple transformer monitoring points are located and acquired, and edge sensors and edge computing devices are respectively deployed at the multiple transformer monitoring points, wherein the edge sensors include ultra-high frequency sensors, ultrasonic sensors, and high-frequency current transformers; Acquire multiple point sensor monitoring data streams through the edge sensor monitoring, and perform mapping anomaly analysis on the multiple point sensor monitoring data streams based on the edge computing device to obtain multiple point abnormal edge feature information; The abnormal edge feature information of the multiple points is integrated and transmitted to the cloud processor through the optical fiber link for abnormal fusion analysis to obtain the abnormal fault parameters of the transformer; Perform cluster parameter optimization based on the abnormal fault parameters of the transformer to obtain operation and maintenance optimization control parameters, and perform fault operation and maintenance management and control on the target transformer based on the operation and maintenance optimization control parameters; The step of obtaining the abnormal fault parameters of the transformer includes: Obtaining transformer fault factor information, wherein the transformer fault factor information includes fault type, fault location, fault degree and impact range; Extracting fault factors from the abnormal edge feature information of the plurality of points according to the transformer fault factor information to obtain a plurality of fault-related factor feature sets; Based on the transformer fault factor information, node fault classification and associated topology analysis are performed on the historical fault database to construct a transformer cascade cause fault tree; Based on the transformer cascade cause fault tree, the fault generation causes of the multiple fault-related factor feature sets are investigated to obtain the transformer abnormal fault parameters; The obtaining of the operation and maintenance optimization control parameters includes: Using a transformer expert system to perform operation and maintenance strategy analysis on the abnormal fault parameters of the transformer to obtain a plurality of initial transformer operation and maintenance strategy parameters; Evaluate and select the multiple initial transformer operation and maintenance strategy parameters to determine a parent strategy parameter set; Based on the parent generation strategy parameter set, cross-combination and random disturbance are performed, and the multiple initial transformer operation and maintenance strategy parameters are expanded and updated in the child generation to obtain the transformer operation and maintenance strategy parameter space; A global optimization comparison is performed in the transformer operation and maintenance strategy parameter space to obtain the operation and maintenance optimization control parameters.

2. A transformer monitoring method based on cloud-edge fusion as claimed in claim 1, characterized in that: The edge sensors are respectively deployed at the plurality of transformer monitoring points, including: Performing monitoring demand analysis on the plurality of transformer monitoring points respectively according to the transformer operation characteristic information to obtain operation demand information of the plurality of monitoring points; Determining, according to the operation requirement information of the plurality of monitoring points, performance parameters of sensors at the plurality of monitoring points, wherein the performance parameters of sensors at the plurality of monitoring points include measurement range, type, accuracy, response time and sensitivity; Acquire structural design information of the target transformer, perform three-dimensional modeling on the structural design information, and generate a three-dimensional spatial model of the transformer; Performing sensor layout optimization analysis on the three-dimensional model of the transformer according to the performance parameters of the sensors at the multiple monitoring points to obtain a layout plan for the sensors at the multiple monitoring points; Based on the performance parameters of the sensors at the multiple monitoring points and the layout schemes of the sensors at the multiple monitoring points, edge sensors are deployed at the multiple transformer monitoring points respectively.

3. A transformer monitoring method based on cloud-edge fusion as claimed in claim 2, characterized in that: The step of obtaining a layout plan of sensors at multiple monitoring points includes: Performing a preliminary analysis of the sensor layout of the three-dimensional model of the transformer according to the sensor performance parameters of the multiple monitoring points to obtain initial layout parameters of the multiple monitoring points; Acquire a sensor layout effect indicator set, and perform simulation evaluation on the initial layout parameters of the plurality of monitoring points based on the sensor layout effect indicator set to obtain transformer monitoring effect information; Performing layout parameter optimization analysis based on the transformer monitoring effect information to determine the sensor layout parameter optimization direction; The initial layout parameters of the plurality of monitoring points are optimized, adjusted and compared according to the sensor layout parameter optimization direction to obtain the sensor layout scheme of the plurality of monitoring points.

4. The transformer monitoring method based on cloud-edge fusion according to claim 1, characterized in that: The step of obtaining abnormal edge feature information of multiple points includes: Invoke a sensor data processing multi-channel through the integrated built-in chip of the edge computing device, wherein the sensor data processing multi-channel includes a data preprocessing node and a data anomaly analysis node; Mapping and allocating the monitoring data streams of the multiple point sensors to the sensor data processing multi-channels to obtain calibration data processing channel information; Based on the calibration data processing channel information, data anomaly parallel analysis is performed on the monitoring data streams of the multiple point sensors to obtain abnormal edge feature information of the multiple points.

5. A transformer monitoring method based on cloud-edge fusion as claimed in claim 4, characterized in that: The calling of sensor data processing multiple channels includes: Determine a set of monitoring data types according to the edge sensor, perform noise characteristic analysis on the set of monitoring data types, and obtain a set of monitoring data noise characteristics; Performing filtering algorithm configuration and standardization program association based on the monitoring data noise characteristic set to obtain a data preprocessing node; Performing anomaly detection demand analysis on the monitoring data type set, constructing an anomaly feature task list, and obtaining a data anomaly analysis node based on training of the anomaly feature task list; Based on the data preprocessing node and the data anomaly analysis node, the sensor data processing multi-channel is constructed, and the sensor data processing multi-channel configuration is stored in the integrated built-in chip.

6. A transformer monitoring system based on cloud-edge fusion, characterized in that: A transformer monitoring method based on cloud-edge fusion for implementing any one of claims 1 to 5, the system comprising: A statistical module, the statistical module is used to collect and obtain a historical fault database of a target transformer, perform statistical identification on the historical fault database, and obtain fault occurrence parameter information; A positioning module, the positioning module is used to locate and obtain multiple transformer monitoring points according to the fault occurrence parameter information, and to deploy edge sensors and edge computing devices at the multiple transformer monitoring points respectively, wherein the edge sensors include ultra-high frequency sensors, ultrasonic sensors, and high-frequency current transformers; A first analysis module, wherein the first analysis module is used to obtain a plurality of point sensor monitoring data streams through the edge sensor monitoring, and perform mapping anomaly analysis on the plurality of point sensor monitoring data streams based on the edge computing device to obtain a plurality of point abnormal edge feature information; A second analysis module, the second analysis module is used to integrate and transmit the abnormal edge feature information of the multiple points to the cloud processor through an optical fiber link for abnormal fusion analysis to obtain abnormal fault parameters of the transformer; An optimization module, the optimization module is used to optimize cluster parameters based on the abnormal fault parameters of the transformer, obtain operation and maintenance optimization control parameters, and perform fault operation and maintenance management and control on the target transformer based on the operation and maintenance optimization control parameters.

Images