CN-122026602-A - Multi-source running state diagnosis and visual decision-making method and system based on station control layer

CN122026602ACN 122026602 ACN122026602 ACN 122026602ACN-122026602-A

Abstract

The application provides a station control layer-based multisource running state diagnosis and visual decision-making method and system, and relates to the technical field of smart grids; the method comprises the steps of identifying an abnormal state of a comprehensive feature set at the network edge side of a power supply place, generating a preliminary diagnosis event set, compressing physical field features of a non-abnormal state into a low-dimensional feature vector, uploading the low-dimensional feature vector and the preliminary diagnosis event set to a station control layer, correlating the preliminary diagnosis event set with the low-dimensional feature vector, generating a multi-dimensional state portrait, combining a space coordinate set generated by utilizing a digital twin model to form target state data, finally generating a dynamic three-dimensional view, and carrying out multi-source running state diagnosis of the station control layer, thereby realizing the fine quantitative evaluation of the health state of power supply equipment and the visual diagnosis of a fault evolution process.

Inventors

MENG YUEFENG
ZHOU ZHENG

Assignees

斯普屹科技(北京)有限公司

Dates

Publication Date: 20260512
Application Date: 20260413

Claims (10)

1. The multi-source running state diagnosis and visual decision method based on the station control layer is characterized by comprising the following steps of: Acquiring electric quantity, mechanical vibration signals, sound wave signals and gas concentration of key electric equipment in a power supply place to form a comprehensive feature set; Identifying the abnormal state of the comprehensive feature set at the network edge side of a power supply place to generate a preliminary diagnosis event set carrying a space region identifier, simultaneously compressing physical field features of a non-abnormal state into a low-dimensional feature vector carrying a time interval identifier, and uploading the preliminary diagnosis event set and the low-dimensional feature vector to a station control layer; Performing space-time correlation processing on the preliminary diagnosis event set and the low-dimensional feature vector to generate a multi-dimensional state portrait; Generating a space coordinate set by utilizing a pre-constructed digital twin model, and correlating a health quantized value and a fault evolution path of the multi-dimensional state portrait with the space coordinate set to form target state data; And generating a dynamic three-dimensional view by utilizing the digital twin model according to the target state data so as to diagnose the multi-source running state of the station control layer.
2. The method of claim 1, wherein identifying the abnormal state of the comprehensive feature set at the network edge side of the power supply site to generate a preliminary diagnostic event set carrying a spatial region identification comprises: Dividing each key power equipment in the power supply place into different adjacent equipment groups according to the distance comparison result of the preset distance threshold value and each key power equipment in the power supply place, and distributing space area identifiers for each adjacent equipment group; Receiving comprehensive feature sets of each key power device at a network edge side node of a power supply place, determining adjacent device groups to which each comprehensive feature set belongs, and arranging the comprehensive feature sets of the same adjacent device groups to form intra-group feature sequences of band group identifiers corresponding to each adjacent device group; calculating the deviation between the comprehensive feature set of each key power equipment in each group and the corresponding historical feature fluctuation range by using the network edge side node, and marking the comprehensive feature set with the deviation exceeding a set threshold value as single feature abnormal data; Calculating a historical characteristic association mode of each adjacent equipment group by using the network edge side node, and identifying association characteristic abnormal data by combining a preset intra-group association rule; And integrating the single characteristic abnormal data and the associated characteristic abnormal data into comprehensive abnormal characteristic data, and associating the installation coordinates of the key power equipment corresponding to the comprehensive abnormal characteristic data, the space region identifiers of the adjacent equipment groups and the original time stamps to form a preliminary diagnosis event set with preliminary diagnosis information carrying the space region identifiers.
3. The method of claim 1, wherein performing a spatiotemporal correlation process on the set of preliminary diagnostic events and the low-dimensional feature vector to generate a multi-dimensional state representation comprises: classifying the preliminary diagnosis event set and the low-dimensional feature vector according to the time interval identifier and the space region identifier to form an integrated data set; Extracting an association mode of an abnormal event and a normal characteristic from prestored historical operation patterns and historical state data matched with the integrated data set so as to establish an association benchmark; Identifying the relevance of the abnormal event and the normal feature in the integrated data set through a preset feature relevance rule so as to generate an event relevance feature; Analyzing the deviation degree of the low-dimensional feature vector and the historical feature change trend in the integrated data set through a preset feature evolution rule to generate feature trend features; And fusing the event correlation features, the feature trend features, the abnormal parameters of the preliminary diagnosis event set and the core parameters of the low-dimensional feature vector to generate a multi-dimensional state portrait.
4. The method of claim 3, wherein fusing the event-related features, the feature trend features, the anomaly parameters of the preliminary diagnostic event set, and the core parameters of the low-dimensional feature vector to generate a multi-dimensional state representation comprises: extracting the association strength value, the association feature type and the association occurrence frequency of the abnormal event and the normal feature from the event association feature to form an event association parameter set; extracting the deviation amplitude, the deviation duration and the deviation change rate of the low-dimensional feature vector and the historical feature change trend from the feature trend features to form a trend parameter set; Extracting an abnormal characteristic type, an abnormal value and an abnormal occurrence moment from the abnormal parameters of the preliminary diagnosis event set to form an abnormal state parameter set; Extracting a reference value of a normal feature, a feature fluctuation amplitude and an inter-feature correlation coefficient from core parameters of the low-dimensional feature vector to form a normal state parameter set; selecting target parameters related to the running state of the equipment from the event-related parameter set, the trend parameter set, the abnormal state parameter set and the normal state parameter set according to a preset parameter extraction rule; and integrating all target parameters according to a preset weight distribution table to generate comprehensive state data, and adding the space region identifiers and the time stamps of the key power equipment to the comprehensive state data to form a multi-dimensional state portrait.
5. The method of claim 1, wherein generating a set of spatial coordinates using a pre-constructed digital twinning model, correlating health quantification values and fault evolution paths of the multi-dimensional state representation with the set of spatial coordinates to form target state data, comprising: generating a space coordinate set taking a reference origin of a power supply place as a coordinate starting point according to the space structure parameters in the pre-constructed digital twin model; Extracting a device unique identifier, a health degree quantized value and a fault evolution path of each key power device from the multi-dimensional state portrait; associating the unique device identifier with the virtual device coordinates in the space coordinate set to establish an association mapping table so as to determine virtual target device coordinates corresponding to each key power device; And carrying out association storage on the health degree quantized value, the fault evolution path and the virtual coordinates of the target equipment to form target state data covering a power supply place.
6. The method of claim 1, wherein generating a dynamic three-dimensional view from the target state data using the digital twin model for multisource operational state diagnostics of a plant floor comprises: Converting the health degree quantized value in the target state data into a corresponding visual identifier according to a preset health degree visualization rule, and adding the visual identifier to a corresponding device virtual component in a three-dimensional infrastructure in the digital twin model based on target device virtual coordinates of each key power device to obtain a device virtual component with the visual identifier; Generating a dynamic track line of a fault evolution path in the three-dimensional basic structure according to a preset path rule and an association mapping table, and generating an initial three-dimensional view by combining the virtual component of the equipment with the visual identifier; And loading a preset interaction control function to update the visual identification and the fault track display form in the initial three-dimensional view based on the interaction control function to form an adjustable three-dimensional view so as to perform collaborative diagnosis of a station control layer on a multi-source running state.
7. The method of claim 6, wherein generating a dynamic trajectory of a fault evolution path in the three-dimensional infrastructure in combination with the device virtual part with visual identification according to a preset path rule and an associated mapping table, comprises: analyzing the target state data to obtain fault path analysis information, wherein the fault path analysis information comprises the occurrence time of equipment history abnormal events, associated equipment identifiers and characteristic change nodes; Matching the associated equipment identifier with corresponding virtual coordinates of target equipment based on an associated mapping table so as to determine key node coordinates of a fault evolution path; According to the occurrence time of the historical abnormal event of the equipment, the key node coordinates are arranged to form a node coordinate list; Generating a dynamic track line according to attribute parameters in a preset path rule and the node coordinate list; generating a node identification set based on the characteristic change nodes, and combining the dynamic track line, the three-dimensional basic structure and the equipment virtual part with the visual identification to form a target superposition body; and calling a preset synthesis rule, and carrying out layer priority sequencing on the elements in the target stack to generate an initial three-dimensional view.
8. The system for diagnosing and visually deciding the multisource running state based on the station control layer is characterized by comprising the following components: The acquisition module is used for acquiring the electric quantity, the mechanical vibration signal, the sound wave signal and the gas concentration of key electric equipment in the power supply place so as to form a comprehensive feature set; The generation module is used for identifying the abnormal state of the comprehensive feature set at the network edge side of the power supply place based on a physical proximity principle so as to generate a preliminary diagnosis event set carrying a space region identifier, compressing physical field features of a non-abnormal state into a low-dimensional feature vector carrying a time interval identifier, and uploading the preliminary diagnosis event set and the low-dimensional feature vector to a station control layer; The association module is used for carrying out space-time association processing on the preliminary diagnosis event set and the low-dimensional feature vector to generate a multi-dimensional state portrait; the binding module is used for generating a space coordinate set by utilizing the pre-constructed digital twin model, and correlating the health quantized value and the fault evolution path of the multi-dimensional state portrait with the space coordinate set to form target state data; And the diagnosis module is used for generating a dynamic three-dimensional view by utilizing the digital twin model according to the target state data so as to diagnose the multi-source running state of the station control layer.
9. An electronic device, comprising: A memory for storing a computer program; a processor for implementing the steps of the station control layer based multisource operation state diagnosis and visualization decision method according to any one of claims 1 to 7 when executing said computer program.
10. A computer readable storage medium, wherein a computer program is stored in the computer readable storage medium, and when the computer program is executed by a processor, the method for diagnosing and visually deciding a multi-source running state based on a station control layer according to any one of claims 1 to 7 can be implemented.

Description

Multi-source running state diagnosis and visual decision-making method and system based on station control layer Technical Field The application relates to the technical field of smart grids, in particular to a station control layer-based multisource running state diagnosis and visual decision method and system. Background In modern power supply places, with the continuous expansion of the power grid scale and the continuous improvement of the complexity of power equipment, efficient and accurate state monitoring and fault diagnosis on key power equipment become important technical requirements for ensuring safe and stable operation of the power grid. In particular, in the site-controlled layer, data from different sources, such as electrical quantities, mechanical vibration signals, etc., need to be processed to achieve a comprehensive assessment of the health status of the electrical equipment. The application scene requires that the running state of the equipment can be monitored in real time, potential faults can be found in time, measures can be rapidly taken for maintenance, and therefore large-area power failure accidents caused by the equipment faults are avoided. Currently, a mainstream scheme aiming at the technical requirements is to adopt a distributed monitoring system based on the internet of things technology, wherein the system collects various operation parameters of power equipment by deploying a large number of sensors, and performs preliminary data analysis and anomaly detection nearby a site by utilizing an edge computing technology. However, the existing schemes still have some remarkable limitations, for example, when facing multi-source heterogeneous data, the digital model of the system is often too simplified to provide enough fine quantitative indicators of equipment health and fault evolution path prediction, which affect the accuracy and reliability of fault diagnosis, and the visualization tools are mostly limited to two-dimensional display, which limit the understanding and judgment of operators on complex spatial information, and are not beneficial to quick decision-making and the like. Disclosure of Invention The application aims to provide a multisource running state diagnosis and visual decision-making method and system based on a station control layer, which are used for solving the problems that a digital model is too simplified, the accuracy and reliability of fault diagnosis are insufficient, and a visual tool limits the understanding and judgment of operators on complex space information, is not beneficial to quick decision-making and the like in the prior art. In order to solve the technical problems, in a first aspect, the present application provides a station control layer-based multisource operation state diagnosis and visualization decision method, which includes: Acquiring electric quantity, mechanical vibration signals, sound wave signals and gas concentration of key electric equipment in a power supply place to form a comprehensive feature set; Identifying the abnormal state of the comprehensive feature set at the network edge side of a power supply place to generate a preliminary diagnosis event set carrying a space region identifier, simultaneously compressing physical field features of a non-abnormal state into a low-dimensional feature vector carrying a time interval identifier, and uploading the preliminary diagnosis event set and the low-dimensional feature vector to a station control layer; Performing space-time correlation processing on the preliminary diagnosis event set and the low-dimensional feature vector to generate a multi-dimensional state portrait; Generating a space coordinate set by utilizing a pre-constructed digital twin model, and correlating a health quantized value and a fault evolution path of the multi-dimensional state portrait with the space coordinate set to form target state data; And generating a dynamic three-dimensional view by utilizing the digital twin model according to the target state data so as to diagnose the multi-source running state of the station control layer. Optionally, identifying an abnormal state of the comprehensive feature set at a network edge side of the power supply site to generate a preliminary diagnosis event set carrying a spatial region identifier, including: Dividing each key power equipment in the power supply place into different adjacent equipment groups according to the distance comparison result of the preset distance threshold value and each key power equipment in the power supply place, and distributing space area identifiers for each adjacent equipment group; Receiving comprehensive feature sets of each key power device at a network edge side node of a power supply place, determining adjacent device groups to which each comprehensive feature set belongs, and arranging the comprehensive feature sets of the same adjacent device groups to form intra-group feature sequences of band group identif