A Method and System for Detection and Maintenance of High-Voltage GIS Equipment Based on Multi-Sensor Data Fusion

Abstract

This invention discloses a method and system for the detection and maintenance of high-voltage GIS equipment based on multi-sensor data fusion, belonging to the field of equipment detection technology. For high-voltage GIS equipment, a multi-layer operation and maintenance array is determined, and a flexible test circuit is constructed by deploying scenario-based test circuits and developing intelligent detectors. As the equipment operates, the micro-sensor topology triggers self-sensing discrimination based on a first threshold condition, triggers directional reconfiguration and testing of the flexible test circuit based on a second circuit threshold, and directional sensing based on a third sensing threshold. This performs directional risk control probability determination and operation and maintenance plan decision-making based on cross-modal multi-sensor data, determining the target operation and maintenance plan for operation and maintenance management. This addresses the technical problems in existing technologies where equipment detection accuracy, response speed, and decision-making intelligence are limited, making it difficult to support accurate operation and maintenance and risk prediction of high-voltage GIS equipment. It achieves intelligent test configuration for high-voltage GIS equipment, improving the accuracy and response flexibility of equipment operation and maintenance detection.

Description

Technical Field

[0001] This invention relates to the field of equipment testing technology, and specifically to a method and system for testing and maintaining high-voltage GIS equipment based on multi-sensor data fusion. Background Technology

[0002] In the current operation and maintenance of high-voltage gas-insulated switchgear, the equipment status is typically assessed through methods such as regular manual inspections, partial discharge monitoring, gas analysis, and identification of abnormal electrical parameters. However, these methods mostly rely on single sensor signals or post-event data analysis, which cannot perceive the operating status of different components in complex structures in real time. Furthermore, they have low accuracy in identifying sudden faults or latent defects, and suffer from problems such as response lag, incomplete test coverage, and high false positive rates.

[0003] Especially in critical parts of the equipment, the interplay of mechanical, electrical, and gaseous characteristics makes it difficult for existing detection methods to effectively correlate multimodal data and establish a unified decision-making mechanism.

[0004] In summary, the lack of a systematic testing architecture and dynamic response mechanism in existing technologies limits the accuracy of detection, response speed, and decision-making intelligence, making it difficult to support the accurate operation and maintenance and risk prediction of high-voltage GIS equipment. Summary of the Invention

[0005] This application provides a method and system for the detection and maintenance of high-voltage GIS equipment based on multi-sensor data fusion. It is intended to address the technical problem that the lack of a systematic testing architecture and dynamic response mechanism in the existing technology leads to limitations in detection accuracy, response speed and decision intelligence, making it difficult to support the accurate operation and maintenance and risk prediction of high-voltage GIS equipment.

[0006] In view of the above problems, this application provides a method and system for the detection and maintenance of high-voltage GIS equipment based on multi-sensor data fusion.

[0007] In a first aspect, this application provides a method for the detection and maintenance of high-voltage GIS equipment based on multi-sensor data fusion. The method includes: for high-voltage GIS equipment, determining a multi-layer operation and maintenance array by decomposing and integrating the characteristics of equipment components; connecting a dynamic simulation platform, deploying scenario-based test circuits with the multi-layer operation and maintenance array, constructing a flexible test circuit, and developing an intelligent detector, wherein the intelligent detector is connected to the micro-sensor topology and the flexible test circuit; as the high-voltage GIS equipment operates, the micro-sensor topology triggers self-sensing discrimination based on a first threshold condition, and triggers directional reassembly and testing of the flexible test circuit based on a second circuit threshold based on the discrimination result, simultaneously triggering directional sensing based on a third sensing threshold, and determining a second sensor array; for the second sensor array, performing directional risk control probability determination and operation and maintenance plan decision-making based on cross-modal multi-sensor data, determining a target operation and maintenance plan, and performing operation and maintenance management of the high-voltage GIS equipment.

[0008] Secondly, this application provides a high-voltage GIS equipment detection and maintenance system based on multi-sensor data fusion. The system includes: a decomposition and integration unit, used to determine a multi-layer operation and maintenance array by decomposing and integrating the characteristics of the high-voltage GIS equipment components; a construction and development unit, used to connect to a dynamic simulation platform, deploy scenario-based test circuits with the multi-layer operation and maintenance array, construct flexible test circuits, and develop intelligent detectors, wherein the intelligent detectors are connected to the micro-sensor topology and the flexible test circuits; a directional testing unit, used to operate with the high-voltage GIS equipment, wherein the micro-sensor topology triggers self-sensing discrimination based on a first threshold condition, and the discrimination result triggers directional reassembly and testing of the flexible test circuit based on a second circuit threshold, and simultaneously triggers directional sensing based on a third sensor threshold to determine a second sensor array; and an equipment operation and maintenance unit, used to perform directional risk control probability determination and operation and maintenance plan decision-making based on cross-modal multi-sensor data for the second sensor array, determine the target operation and maintenance plan, and perform operation and maintenance management of the high-voltage GIS equipment.

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

[0010] The high-voltage GIS equipment detection and maintenance method based on multi-sensor data fusion provided in this application addresses the technical problem of lacking a systematic test architecture and dynamic response mechanism in existing technologies. This leads to limitations in detection accuracy, response speed, and decision intelligence, hindering the accurate operation and maintenance and risk prediction of high-voltage GIS equipment. The method achieves intelligent test circuit deployment based on scenario-based testing circuits, constructs flexible test circuits, and develops intelligent detectors that operate alongside the high-voltage GIS equipment. The micro-sensor topology triggers self-sensing discrimination based on a first threshold condition. The discrimination result triggers directional reconfiguration and testing of the flexible test circuit based on a second circuit threshold, simultaneously triggering directional sensing based on a third sensing threshold to determine a second sensor array. For the second sensor array, directional risk control probability determination and operation and maintenance plan decision-making based on cross-modal multi-sensor data are performed to determine the target operation and maintenance plan and manage the high-voltage GIS equipment. This method addresses the technical problem in existing technologies where the lack of a systematic test architecture and dynamic response mechanism limits detection accuracy, response speed, and decision intelligence, making it difficult to support accurate operation and maintenance and risk prediction of high-voltage GIS equipment. It achieves intelligent test configuration for high-voltage GIS equipment, improving the accuracy and flexibility of equipment operation and maintenance detection. Attached Figure Description

[0011] Figure 1 This application provides a schematic flowchart of a high-voltage GIS equipment detection and maintenance method based on multi-sensor data fusion;

[0012] Figure 2 This application provides a schematic diagram of the structure of a high-voltage GIS equipment detection and maintenance system based on multi-sensor data fusion.

[0013] Figure labeling: Decomposition and integration unit 11, construction and development unit 12, targeted testing unit 13, equipment operation and maintenance unit 14. Detailed Implementation

[0014] This application provides a method and system for the detection and maintenance of high-voltage GIS equipment based on multi-sensor data fusion. This addresses the technical problem in the prior art where the lack of a systematic testing architecture and dynamic response mechanism limits the detection accuracy, response speed, and decision intelligence, making it difficult to support the accurate operation and maintenance and risk prediction of high-voltage GIS equipment.

[0015] Example 1: As Figure 1 As shown, this application provides a method for the detection and maintenance of high-voltage GIS equipment based on multi-sensor data fusion, the method comprising:

[0016] S1: For high-voltage GIS equipment, a multi-layer operation and maintenance array is determined by decomposing and integrating the characteristics of the equipment components.

[0017] In this embodiment of the invention, a systematic analysis of the high-voltage GIS equipment is first conducted to construct an operation and maintenance data structure that conforms to its structural hierarchy and functional attributes.

[0018] Specifically, in the process of decomposing equipment components, it is necessary to divide them according to the physical structure and functional logic of the equipment, and identify and classify each sub-component as an independent operation and maintenance object to achieve hierarchical modeling of the equipment. Then, based on the above decomposition, the characteristics of each type of equipment component are integrated.

[0019] Specifically, the characteristics of circuit breaker components include mechanical opening and closing operation characteristics and contact arc stability; the characteristics of disconnecting switches include mechanical operation and contact temperature rise; and busbars involve electrical and environmental parameters such as gas tightness, partial discharge characteristics, and SF6 gas concentration variation. To effectively reflect these characteristics, this invention abstracts the typical operating parameters of each type of component into standardized feature units, constructing a feature vector set for operation and maintenance analysis.

[0020] Subsequently, after constructing the feature vector set, a multi-layer operation and maintenance array is further formed through matrix combination. The multi-layer operation and maintenance array proposed in this application refers to a set of feature matrices expanded in two dimensions according to the component structure level (such as circuit breaker layer, isolation layer, busbar layer) and characteristic category (such as mechanical, electrical, and thermal characteristics).

[0021] The array consists of two layers: a first layer, a structure matrix, which identifies each device component and its spatial topological position; and a second layer, a characteristic matrix, which identifies the corresponding operating performance indicators of each device component. This array structure supports horizontal (i.e., intra-layer) expansion to increase feature dimensions, and also supports vertical (i.e., hierarchical) expansion to incorporate other subsystem components, exhibiting good structural scalability and test scenario combination capabilities.

[0022] For example, in a typical 110kV GIS bay unit, circuit breakers and disconnectors can be listed as independent operation and maintenance units in the first layer of the structure array, while their mechanical stroke curves, current fluctuation characteristics and contact temperature rise change data can be extracted as components of the second layer characteristic matrix.

[0023] The multi-layered operation and maintenance array constructed in this way not only has the ability of atomic management, but also provides structured input support for subsequent flexible test circuit deployment, threshold trigger recognition and intelligent decision-making, realizing the basic framework for full-link detection of high-voltage GIS equipment from structural perception to functional response.

[0024] Furthermore, by decomposing and integrating the characteristics of the equipment components, a multi-layer operation and maintenance array is determined. Step S1 of this application includes:

[0025] A first maintenance matrix is ​​determined based on the equipment components of the high-voltage GIS equipment, wherein the first maintenance matrix includes at least circuit breakers, disconnectors, and busbars, and is labeled with component motion characteristics; a second maintenance matrix is ​​determined based on the component characteristics of the equipment components, wherein the second maintenance matrix includes the mechanical and electrical characteristics cascaded to the circuit breakers, the thermal and mechanical characteristics cascaded to the disconnectors, and the partial discharge and gas state characteristics cascaded to the busbars; and a multi-layer maintenance array is determined based on the first and second maintenance matrices, wherein the multi-layer maintenance array is expandable, including intra-layer expansion and hierarchical expansion.

[0026] In this embodiment of the invention, a first operation and maintenance matrix is ​​established, oriented towards the structural dimension of the equipment, using the key components of the high-voltage GIS equipment as basic building blocks. Specifically, the first operation and maintenance matrix refers to a structural matrix with spatial correlation and functional independence, constructed using identifiable functional units in the GIS equipment as array elements. This matrix includes at least three typical sub-components: circuit breakers, disconnect switches, and busbars. These components respectively undertake core functions such as current on / off control, isolation protection, and busbar power transmission support within the GIS system.

[0027] In a preferred embodiment, the above-mentioned components are identified in the first operation and maintenance matrix through a component-location-function triplet structure, and their motion characteristics are highlighted, that is, the dynamic behavior characteristics that the equipment can exhibit in actual operation, such as the closing and opening speed of the circuit breaker, the mechanical opening distance of the disconnecting switch, and the thermal expansion range allowed by the busbar structure.

[0028] Furthermore, based on the specific operational attributes of the equipment components, a second operation and maintenance matrix oriented towards the performance dimension is constructed. This second operation and maintenance matrix is ​​mainly used to describe the changes in physical state and fault characteristic modes during component operation.

[0029] Specifically, for circuit breakers, mechanical characteristics may include the acceleration curve during the release of spring energy and the travel offset of the drive arm; electrical characteristics may include the opening and closing current curves, contact resistance changes, and contact discharge peak values. For disconnectors, thermal characteristics include the temperature rise rate of the joints, steady-state temperature distribution, and heat dissipation rate; mechanical characteristics include changes in the friction of the switching mechanism and the number of opening and closing cycles. For busbars, attention should be paid to their partial discharge characteristics (e.g., PD spectra obtained through ultrasonic or electromagnetic signal detection) and gas state parameters (including SF6 concentration, humidity, and electronegative gas content). These indicators are all added to the second operation and maintenance matrix in vector form, forming a parallel mapping of structural and performance data.

[0030] Subsequently, based on the cascaded combination of the first and second operation and maintenance matrices, a multi-layer operation and maintenance array oriented towards testing and judgment is further formed.

[0031] In this embodiment, the multi-layer operation and maintenance array is the core supporting structure for the subsequent detection and maintenance strategy of this invention, and has two types of expansion capabilities: one is intra-layer expansion, that is, more feature dimensions can be added in the same layer, such as adding pressure release characteristics, insulation degradation coefficient, etc. in the circuit breaker characteristic layer; the other is hierarchical expansion, that is, adding new layers, or adding new component types (such as voltage transformers, grounding switches, surge arresters, etc.) in different layers to form a more complete hierarchical structure model of the entire GIS system.

[0032] In summary, the construction of the multi-layer operation and maintenance array enables this solution to have an expandable, combinable, and identifiable detection logic foundation, laying a structural support for the deployment of flexible test circuits and multi-threshold trigger response mechanisms.

[0033] S2: Connect to the dynamic simulation platform, deploy scenario-based test circuits using the multi-layer operation and maintenance array, construct flexible test circuits, and develop intelligent detectors, wherein the intelligent detectors are connected to the micro-sensing topology and the flexible test circuits.

[0034] In the implementation of this invention, the constructed multi-layer operation and maintenance array is first connected to a dynamic simulation platform. This platform is used to simulate the state response and abnormal characteristic reproduction of high-voltage GIS equipment under different operating conditions. The platform supports multi-component collaborative modeling and dynamic electrical behavior simulation, and can map the structural layer and characteristic layer of the operation and maintenance array to adjustable test objects and test parameters, thereby achieving scenario-based test circuit deployment with structure as the target and characteristics as the indicator.

[0035] Specifically, the scenario-based test circuit refers to the test circuit topology constructed based on the component types and operating characteristics defined in the multi-layer operation and maintenance array, corresponding to typical operating conditions, such as closing impact, electric arcing, gas leakage, etc., as the flexible test circuit.

[0036] In one feasible embodiment proposed in this application, programmable flexible connection units, such as controllable switch modules, current source / voltage source modules, etc., are combined and reconfigured to support dynamic loading of different test scripts, thereby realizing the simulation of operating conditions and fault reproduction for a specific array element or multiple combined elements.

[0037] For example, in a test scenario targeting the deterioration of circuit breaker contacts, circuit elements that simulate arc current fluctuations and mechanical hysteresis delays can be invoked to form a test path for electromechanical coupling failure.

[0038] Subsequently, after completing the deployment of the test circuit, an intelligent detector was further developed. This intelligent detector is used to discriminate, make decisions, and generate response commands from the sensor data during the testing process, forming a logical closed loop with the micro-sensor topology and flexible test circuit.

[0039] In this application, the micro-sensor topology is a network of sensor units deployed on the surface or embedded in the structure of high-voltage GIS equipment, capable of acquiring multi-source sensor information such as temperature, vibration, partial discharge, current, voltage, and gas concentration in real time. It can be implemented by connecting to a flexible test circuit via a bus structure to achieve synchronized test initiation, status feedback, and anomaly marking.

[0040] For example, when the flexible test circuit activates a certain operation and maintenance scenario simulation, the intelligent detector listens to the feedback data of the micro-sensor topology in real time, and judges whether the test has triggered an effective response based on the preset first-order threshold rules (such as temperature rise threshold > 15℃ / min, partial discharge amplitude > 5mV), so that the test plan can be dynamically adjusted or abnormal behavior can be recorded.

[0041] In summary, the deployment of intelligent detectors enables a closed-loop detection mechanism that integrates test-driven, perception-response, and judgment processes, significantly improving the accuracy of GIS equipment fault identification and the relevance of operation and maintenance strategies.

[0042] Furthermore, by deploying scenario-based test circuits using the multi-layer operation and maintenance array to construct flexible test circuits, step S2 of this application includes:

[0043] Based on the multi-layer operation and maintenance array, an operation and maintenance scenario set is defined, wherein the operation and maintenance scenario is defined with a focus on testing based on a single array element or a group of array elements; based on the operation and maintenance scenario set, a flexible test circuit is determined through simulation and optimization, wherein the flexible test circuit is an external microcircuit of the high-voltage GIS equipment, and the flexible test circuit is based on the scenario-based reorganization of circuit loops based on line switches.

[0044] In the implementation of this invention, based on the already constructed multi-layer operation and maintenance array, a combination of scenarios oriented towards specific operation and maintenance objectives is further defined, thus forming an operation and maintenance scenario set.

[0045] The aforementioned operation and maintenance scenario set refers to a collection of scenario units for scenario-based testing, which are selected from a multi-layer operation and maintenance array based on different testing and maintenance needs, including a single array element (such as a single circuit breaker) or a group of multiple array elements (such as a combination of circuit breaker, disconnector, and busbar).

[0046] In the preferred implementation process, each scenario unit has a clear fault concern direction or risk category, that is, the focus of the operation and maintenance scenario.

[0047] For example, when the focus is on partial discharge, a test scenario is constructed centered on the busbar and guided by the characteristics of partial discharge; when the focus is on slow contact action, a test scenario is constructed centered on the circuit breaker and focused on its mechanical behavior characteristics.

[0048] Furthermore, after establishing the set of operation and maintenance scenarios, a dynamic simulation platform is used to perform simulation analysis and circuit behavior modeling for each scenario. Preferably, historical fault cases and boundary parameter optimization can be combined to determine the test circuit structure suitable for this type of scenario.

[0049] In this process, the generated test circuit is not a fixed circuit topology, but a flexible test circuit that can be dynamically reconfigured according to the needs of the scenario.

[0050] In one specific embodiment, the flexible test circuit is an auxiliary detection microcircuit system installed outside the high-voltage GIS equipment. It can form a logical coupling relationship with the main equipment system through structural connectors (such as detachable cable terminals) and functional units (such as high-impedance sensing loads and switch control chips).

[0051] In this application, the core feature of the flexible test circuit is the scenario-based reconfiguration of circuit loops based on line switches. This feature refers to the automatic triggering of the on / off operation of circuit connection switches during the test execution phase, based on the array elements involved in the current operation and maintenance scenario, to form a current or voltage loop structure that matches the scenario.

[0052] For example, in a test scenario simulating circuit breaker tripping and jamming, the control logic closes the high-frequency interference source port in the closed loop and disconnects the mechanical drive load port in order to inject abnormal waveforms and monitor feedback; in a test scenario simulating busbar partial discharge characteristics, the high-impedance acquisition branch is opened and a high-voltage coupling unit is connected in parallel to achieve accurate capture of partial discharge signals.

[0053] In summary, by defining the operation and maintenance scenario set and reorganizing the flexible test circuit in a scenario-based manner, not only was a targeted and precisely coupled test path constructed, but also dynamic circuit-level support was provided for the subsequent threshold triggering mechanism and intelligent detection logic, forming a closed-loop test system that links structure, scenario, and circuit.

[0054] Furthermore, prior to developing the intelligent detector, step S2 of this application includes:

[0055] Based on the multi-layer operation and maintenance array, a micro-sensor topology based on directional testing is deployed; the operation and maintenance records of the high-voltage GIS equipment are retrieved, and critical value mining is performed under abnormal operating conditions, which is converted into risk control thresholds based on the micro-sensor topology, wherein the micro-sensor topology and the risk control thresholds correspond one-to-one; the micro-sensor topology and the risk control thresholds are cascaded to deploy the first threshold condition.

[0056] In this embodiment of the invention, based on the structural and characteristic information of the multi-layer operation and maintenance array, a sensor deployment work is carried out for a specific test target to form a micro-sensor topology oriented towards directional testing.

[0057] Specifically, the micro-sensor topology refers to deploying micro-sensors in an array on the surface or adjacent area of ​​key components of high-voltage GIS equipment. Optional sensor types include, but are not limited to, temperature sensors, vibration acceleration sensors, ultrasonic partial discharge sensors, current and voltage probes, and gas detection probes. The sensor deployment method depends on the structural arrangement of each component in the array, fault-prone points, and key maintenance scenarios to ensure that the deployed sensors can achieve high sensitivity, low latency, and well-defined spatial positioning for feature acquisition.

[0058] For example, high-response Hall displacement sensors and thermistor sensors are deployed in the area of ​​the disconnector switch moving mechanism; and electronegative gas concentration detection probes and partial discharge electromagnetic acquisition arrays are deployed in the SF6 gas chamber area of ​​the busbar cylinder to achieve precise coverage of key points.

[0059] Furthermore, historical operation and maintenance data records of high-voltage GIS equipment are retrieved, especially datasets related to abnormal operating conditions, to conduct abnormal operating condition analysis and critical value mining.

[0060] In this application, the abnormal operating conditions refer to non-standard operating states of equipment during operation, such as closing / opening delays, sudden temperature rises, abnormal current fluctuations, and sudden changes in partial discharge intensity. This invention identifies the critical change intervals of characteristics before and after an anomaly by combining multi-dimensional feature correlation analysis and typical fault sample extraction with indicators such as time-series features, frequency domain distribution, and waveform distortion. Based on this, a risk control threshold set for single-point sensors is formed.

[0061] Each sensing node is set with an over-limit threshold range corresponding to its physical measurement results. For example, the abnormal identification threshold of a certain temperature sensor is: a heating rate greater than 8℃ / min or a peak value higher than 105℃ is judged as a suspicious state; the abnormal threshold of a certain partial discharge sensor is that the discharge pulse intensity continuously exceeds 2mV and the interval is less than 5ms.

[0062] Subsequently, the aforementioned risk control thresholds are mapped point by point to the micro-sensor topology, thus forming a one-to-one mapping relationship to ensure that each sensing node is bound to at least one set of risk control judgment parameters.

[0063] Finally, based on the binding relationship between microsensor topology and risk control threshold, a first threshold condition is constructed, which is the initial sensing limit judgment rule used to trigger subsequent test logic and discrimination mechanism during device operation.

[0064] Optionally, the first threshold condition consists of threshold judgment logic with multiple independent branches, forming a low-level but high-frequency state detection mechanism suitable for online real-time monitoring.

[0065] In this application, based on the first threshold condition, potential abnormal signals can be captured and test circuits or alarm modules can be activated immediately without human intervention, thereby constructing an active detection chain and providing a reliable starting point for subsequent flexible test circuit reconstruction and intelligent detector driving.

[0066] Furthermore, in developing an intelligent detector, step S2 of this application includes:

[0067] Based on the flexible test circuit, a second circuit threshold is defined, wherein the second circuit threshold performs circuit loop reassembly based on the test scenario; based on the micro-sensor topology, a third sensing threshold is defined, wherein the third sensing threshold performs sensing orientation triggering based on the test scenario; using the first threshold condition, the second circuit threshold, and the third sensing threshold, drive training based on the underlying operation and maintenance test logic is performed to determine the intelligent detector.

[0068] In this embodiment of the invention, based on the aforementioned constructed flexible test circuit, a corresponding circuit test loop is further defined for each type of test scenario, that is, a second circuit threshold is constructed.

[0069] In this application, the second circuit threshold refers to the loop structure and electrical parameter configuration required to activate the flexible test circuit under different test scenarios. The threshold includes not only the triggering conditions and circuit structure of the logic switch, but also the configuration instructions and tolerance ranges of circuit levels such as voltage, current, frequency, and phase.

[0070] For example, in a test scenario simulating abnormal contact resistance of a circuit breaker contact, the circuit loops to be activated include low-frequency current injection and voltage drop monitoring. In this scenario, the second circuit threshold is defined as follows: when the scenario is activated, the low-frequency injection branch is automatically closed, the sampling loop is opened, and the excitation signal amplitude is set to 10A, the frequency to 100Hz, and the duration is not less than 2 seconds. In the technical solution of this invention, the second circuit threshold logically corresponds to a specific circuit test path, ensuring that the test circuit can be reconstructed according to a preset scheme.

[0071] Furthermore, a third sensing threshold is defined for the microsensor response behavior corresponding to each type of test scenario. This third sensing threshold controls the directional triggering mechanism of the microsensor topology during the test process, ensuring that specific sensor nodes are preferentially activated in the target test scenario, thereby forming a structured data acquisition path.

[0072] In the preferred example, the third sensing threshold includes not only parameters such as the time sequence of sensor activation, sampling frequency, and channel bandwidth, but also the instantaneous discrimination criteria for abnormal responses.

[0073] For example, in the scenario of overheating of the disconnector switch, the third sensor threshold will activate the high-frequency sampling mode of the thermal resistance group in that area, increasing the sampling frequency from the usual 0.5Hz to 2Hz, and setting the alarm threshold as a temperature rise rate greater than 5℃ / min or a steady-state temperature higher than 95℃; while the sensors in other areas remain in normal state to achieve resource focus and data redundancy control.

[0074] Subsequently, after establishing three threshold mechanisms—the first threshold condition (for real-time operational sensing), the second circuit threshold (for testing circuit control), and the third sensing threshold (for sensing activation regulation)—further training based on underlying operation and maintenance testing logic is conducted. This training process employs a combined approach of multi-round simulation testing and real-world data-driven methods: first, a typical test sample set is constructed, including normal state data, typical fault state data, and boundary condition data; then, in a dynamic simulation platform, the aforementioned threshold logic is combined to drive the flexible circuit and sensing topology to operate collaboratively; within each test cycle, circuit behavior, sensing response, and threshold exceedance results are collected to construct a triplet training sample of test behavior—data response—judgment result.

[0075] In summary, through multiple rounds of training and threshold-linked control, an intelligent detector is ultimately constructed. During testing, it can determine the scene matching degree, fault level, and test validity in real time based on threshold conditions, and output circuit control commands or anomaly reporting signals. It possesses high adaptability and scene recognition capabilities, and is the key execution module in this invention for achieving on-demand detection, accurate testing, and autonomous response.

[0076] S3: As the high-voltage GIS equipment operates, the micro-sensor topology triggers self-sensing discrimination based on the first threshold condition, and triggers directional reassembly and testing of the flexible test circuit based on the second circuit threshold based on the discrimination result, and simultaneously triggers directional sensing based on the third sensing threshold to determine the second sensor array.

[0077] During the implementation of this invention, when the high-voltage GIS equipment is in normal operation, the micro-sensor topology deployed at key locations in the equipment structure continuously performs data acquisition tasks. The micro-sensor topology encompasses multiple different types of miniature sensor units, including temperature sensors, vibration sensors, partial discharge sensors, current / voltage probes, and gas concentration sensing modules, all of which are connected to the detection network via a bus protocol and upload data at preset intervals or in an event-driven manner.

[0078] During equipment operation, the sensor data stream is evaluated in real time based on a first threshold condition, i.e., the self-sensing discrimination logic is executed. The first threshold condition includes the judgment of exceeding the limit of key parameters in the sensor sampling results.

[0079] For example, if the temperature rise rate of a certain channel exceeds 8℃ / min, or a partial discharge sensor detects a continuous high-intensity pulse signal. If the data of any node in the micro-sensor topology meets the risk control threshold condition it is bound to, it is identified as a potential risk trigger, automatically generating a self-sensing discrimination result, and driving the downstream module based on this result.

[0080] Subsequently, based on the above judgment results, the second circuit threshold control mechanism is immediately triggered to perform directional reconfiguration and scenario testing on the flexible test circuit. The directional reconfiguration refers to dynamically selecting and closing a specific loop in the flexible test circuit according to the identified target test scenario.

[0081] For example, after determining that the circuit breaker may have abnormal contact resistance, based on the circuit thresholds associated with this scenario, low-frequency current injection, voltage drop acquisition loops, and auxiliary switching links are activated to form a closed-loop circuit structure capable of performing targeted testing on the circuit breaker. Preferably, during test execution, the switching state, power parameters, and load state of all channels are dynamically configured to perfectly match the test scenario.

[0082] At the same time, the third sensing threshold mechanism is activated to execute the directional sensing trigger logic, which is used to control the sensing nodes in the sensing topology that are related to the current test scenario to collect data at a higher sampling frequency, finer resolution or special working mode.

[0083] For example, in a partial discharge detection scenario for a busbar, the high-bandwidth sampling mode of the partial discharge electromagnetic probe is activated, and the transient sampling window is set to 20μs with a total sampling period of 5 seconds. Other non-critical nodes enter a low-power monitoring state to improve resource utilization efficiency.

[0084] In summary, the output data set of the directional sensing process constitutes the second sensing array, which is composed of real-time sensing data collected by the target sensing node based on the flexible circuit reconstruction and directional testing scenario execution. It has the characteristics of strong scenario correspondence, high state resolution, and clear anomaly correlation.

[0085] Subsequently, the second sensor array will serve as an important input basis for subsequent fault probability analysis, intelligent contingency plan generation, and operation and maintenance decisions, realizing the evolution of the detection closed loop from perception to discrimination to response to deep detection.

[0086] Furthermore, to trigger the self-sensing discrimination based on the first threshold condition, step S3 of this application includes:

[0087] As the high-voltage GIS equipment operates, the micro-sensor topology synchronously performs sensing detection to determine the first sensor array; based on the first threshold condition, the first sensor array is subjected to limit-crossing judgment based on the risk control threshold to determine the risk control sensing part; if the risk control sensing part is not empty, a target test scenario is defined, wherein the target test scenario belongs to the operation and maintenance scenario set.

[0088] In this embodiment of the invention, when the high-voltage GIS equipment is in online operation, the micro-sensor topology will perform real-time sensing detection in a continuous and synchronous manner.

[0089] In this application, the micro-sensor topology is a sensor network deployed at key parts of the equipment, covering multiple functional units such as circuit breakers, disconnect switches, and busbars, and corresponding to multiple dimensions of operating status parameters such as temperature, current, voltage, partial discharge, and gas state. Optionally, each sensor node is uniformly sampled and controlled by a centralized control unit, and the data is synchronized using a unified timestamp, thereby obtaining a set of sensor data covering multiple parts and multiple status parameters of the GIS equipment, thus forming the first sensor array.

[0090] That is, the first sensor array refers to the set of raw operating status data collected by all deployed microsensor topology nodes within a certain synchronous detection cycle. Preferably, the array structure is grouped according to sensor type, installation location, and physical quantity category. Each group of data includes real-time measurement values ​​at multiple time points, such as 10 sets of temperature values ​​collected by a temperature sensor on a circuit breaker within 10 seconds, or a discharge peak sequence collected by a partial discharge sensor on a busbar within 50ms.

[0091] Subsequently, an out-of-limit determination based on a first threshold condition is performed on the first sensor array. The first threshold condition covers the boundary values ​​of various sensors during normal operation and the initial stage of a fault.

[0092] For example, if the threshold of the circuit breaker temperature sensing channel is 105℃ or the heating rate is greater than 8℃ / min, or the threshold of the partial discharge channel is an instantaneous pulse amplitude greater than 2mV and lasting for more than 10ms, then when the output data of a certain sensor in the first sensing array exceeds the corresponding threshold, it will be identified as exceeding the limit.

[0093] Furthermore, all sensor channels that exceed limits constitute the risk control sensing section. This section is structurally a subset of the first sensor array, but it has a high risk indication significance. If the risk control sensing section in the judgment result is not empty, that is, if one or more sensor data meet the limit judgment condition, then it is judged that there is a potential equipment risk.

[0094] Furthermore, based on the sensor location, data type, and corresponding device unit of the risk control sensing section, a target test scenario is defined. In the implementation scheme of this application, the target test scenario is the scenario that best matches the current sensing anomaly, selected from a predefined set of operation and maintenance scenarios, and has a clear test path, triggering mechanism, and judgment logic.

[0095] For example, when the risk control sensing part involves the over-limit of the heating channel and contact vibration channel of the disconnector switch, the target test scenario is defined as the test scenario of the degradation of the thermo-mechanical coupling characteristics of the disconnector switch by performing approximate matching in the operation and maintenance scenario set, and the corresponding test circuit and sensor high-frequency response path are prepared to be started.

[0096] In summary, a continuous logical chain was realized, starting from raw multi-channel sensor data, proceeding through risk assessment and scene recognition, to test strategy invocation, thus constructing a basic input path for subsequent targeted flexible testing and intelligent discrimination.

[0097] Furthermore, triggering the directional reassembly and testing of the flexible test circuit based on the second circuit threshold, step S3 of this application includes:

[0098] The target test scenario is identified, and a circuit loop based on the target test scenario is matched and determined according to the second circuit threshold. Based on the circuit loop, state adjustment based on the line switch is performed to determine the switch instruction set. Based on the switch instruction set, the loop of the flexible test circuit is adjusted to determine the target test circuit. The target test circuit is driven to perform directional scenario testing on the high-voltage GIS equipment.

[0099] In this embodiment of the invention, after a specific target test scenario is determined based on the risk control sensing part, a scenario circuit mapping operation is first performed, that is, the target test scenario is identified, and the circuit loop corresponding to the scenario is matched and determined according to the second circuit threshold.

[0100] In this application, the second circuit threshold is pre-bound to each operation and maintenance scenario, defining the electrical paths, load units and control nodes to be activated for different test purposes.

[0101] For example, in the circuit breaker contact resistance abnormality test scenario, the second circuit threshold specifies that a low-frequency excitation current source, a parallel voltage sampling circuit, and a temperature rise detection module must be activated. By comparing the target test scenario with the threshold rules, the complete test circuit structure required for this scenario can be identified.

[0102] Subsequently, based on the current and expected connection states of each node in the circuit structure, state adjustment based on line switches is performed, and a set of switching instructions is generated.

[0103] The circuit switches mentioned refer to the logic control switches between the various circuit branches in the flexible test circuit, including solid-state relays, electromagnetic contactors, or programmable electronic switches. By comparing the existing circuit state with the requirements of the target test circuit, a set of execution operation instructions is formed, including the closing / opening command of each switch, the execution sequence, and the delay time, for example: closing K3, opening K1 and K2, and delaying the start of the P1 power module by 2 seconds.

[0104] Next, the flexible test circuit is adjusted according to the aforementioned switch instruction set. During instruction execution, the flexible circuit will automatically switch the connection relationships of circuit branches and adjust the internal load state, filtering module, signal acquisition branch, etc., so that the current circuit structure fully matches the requirements of the target test scenario.

[0105] In summary, to ensure the integrity and safety of electrical behavior paths, for example, in a high-voltage simulation environment, signal acquisition channel isolation is performed first, and then the main circuit is switched on and off to prevent false triggering and arc damage.

[0106] Furthermore, after completing the circuit reconstruction, the target test circuit is formed, which is a circuit topology that has been configured and whose structure and logic correspond to the target test scenario. Subsequently, the test-driven mechanism is activated to drive the target test circuit to perform targeted scenario testing.

[0107] In one specific implementation, the driving process includes steps such as excitation signal emission, response data acquisition, and scene feature recognition.

[0108] For example, in the busbar partial discharge degradation test scenario, the drive module will emit a high-frequency pulse test signal, which will be injected into the test point through capacitive coupling, and at the same time activate the high-bandwidth partial discharge sampling probe to collect abnormal discharge signals; in the disconnector heating test scenario, a low-frequency steady-state load current will be applied, and the temperature rise rate of its thermocouple circuit will be monitored.

[0109] In summary, a closed-loop logic of threshold identification, circuit reconfiguration, path control, and test execution, driven by the target test scenario, has been realized. This provides high-voltage GIS equipment with customized and highly accurate fault verification capabilities and structural condition detection methods, effectively supporting the implementation of intelligent operation and maintenance strategies.

[0110] Furthermore, triggering directional sensing based on a third sensing threshold, step S3 of this application includes:

[0111] Identify the target test scenario, and determine the scene sensing topology based on the target test scenario according to the third sensing threshold; as the target test circuit is triggered, the scene sensing topology synchronously performs sensing triggering to determine the second sensing array.

[0112] In the implementation of this invention, after identifying a specific target test scenario and completing the structural reorganization of the flexible test circuit, a matching process for the third sensing threshold is further executed to determine a sensor activation strategy that is compatible with the test scenario.

[0113] Specifically, based on the third sensing threshold, i.e. the set of sensing activation conditions set for various test scenarios, scenario-based matching is performed to determine the scenario sensing topology based on the target test scenario. That is, the sub-topology structure selected on demand from the complete micro-sensing topology has a specific node composition and sensing strategy to serve the anomaly monitoring needs of the current scenario.

[0114] For example, in a circuit breaker mechanical characteristic anomaly test scenario, the matched scenario sensing topology will include a high-precision accelerometer, a displacement encoder, and a contact temperature rise sensor deployed in the mechanism housing; while in a busbar drum partial discharge anomaly test scenario, an ultrasonic sensor, an electromagnetic radiation detector, and a gas composition analysis probe distributed in the busbar cavity will be selected. In a preferred embodiment, the third sensing threshold also specifies the specific sampling frequency, activation conditions, and feature extraction strategy for the scenario. For example, in this scenario, the sampling frequency of a certain temperature sensor is increased from 1Hz to 5Hz, and a fast threshold judgment mechanism is activated.

[0115] With the formal activation of the aforementioned target test circuit, i.e., the execution of test actions such as current excitation, voltage loading, simulated waveform injection, or load switching, all sensing nodes in the scene sensing topology are simultaneously activated, i.e., the sensing triggering process is executed. This triggering mechanism is achievable by relying on event synchronization mechanisms and sampling timing control. For example, timestamp constraints are used to ensure strict alignment between sensing data and test actions, thereby forming highly consistent test response data.

[0116] For example, while the flexible test circuit applies a gradually increasing current to the disconnecting switch, the temperature sensor collects the temperature rise changes in the contact area at a high frequency, and the vibration sensor captures the weak vibration changes of the actuating mechanism, ensuring that the multidimensional state response after the test stimulus can be accurately reproduced.

[0117] The data generated by the above tests constitutes the second sensor array. This array contains only the data collected by the target sensor nodes activated in the current test scenario, and is scenario-specific, response-specific, and of high analytical value. The data in the second sensor array is organized using a timestamp synchronization method for easy subsequent analysis.

[0118] In summary, a closed-loop linkage between testing behavior and data perception has been achieved, which not only improves the accuracy and sensitivity of fault detection, but also provides a high-quality decision data source for the subsequent generation of operation and maintenance strategies based on intelligent algorithms.

[0119] S4: For the second sensor array, perform directional risk control probability determination and operation and maintenance plan decision-making based on cross-modal multi-sensor data, determine the target operation and maintenance plan, and perform operation and maintenance management on the high-voltage GIS equipment.

[0120] In the implementation of this invention, after completing the flexible circuit excitation and scene sensing topology triggering of the target test scenario, the corresponding second sensor array has been obtained. To achieve accurate risk assessment and intelligent operation and maintenance suggestion output, the array is further subjected to directional risk control probability determination and operation and maintenance plan decision-making based on cross-modal multi-sensor data.

[0121] First, preprocessing and feature extraction are performed on the various types of data in the second sensor array to construct a multimodal feature vector. Optional cross-modal data features include, but are not limited to: the slope of the temperature rise rate curve, the amplitude of the dominant frequency of the vibration spectrum, the temporal density of the partial discharge pulse, the abnormal change rate of gas components, and the morphological factor of the arc characteristic waveform. Preferably, a unified embedding structure of an adversarial neural network under fusion constraints is adopted to encode and map the data of each modality, construct a unified expression space, and retain the correlation information between each sensor channel.

[0122] After completing modality fusion, a targeted risk control probability determination is further introduced, which involves inputting the fused feature vectors into the trained adversarial generative network and outputting the probability of occurrence of risk events in each scenario.

[0123] For example, in the test scenario of abnormal heating and hysteresis of disconnecting switch, it can be determined that there is a 70% probability of loose thermal connection, a 20% probability of mechanical aging, and a 10% probability of normal fluctuation in the current state. Based on the historical evolution trend, the fault level and priority handling suggestions are output.

[0124] Based on the above judgment results, the system further invokes the preset maintenance plan library and operating procedures to execute the matching decision process of the maintenance plan and automatically generate a structured target maintenance plan. For example, when the partial discharge anomaly is judged to be of moderate intensity, the target maintenance plan may include measures such as advancing the next maintenance to 15 days, increasing the frequency of partial discharge monitoring to once every 12 hours, and deploying portable partial discharge locators to perform on-site verification.

[0125] Finally, the generated target operation and maintenance plan is sent to the GIS equipment management system or connected with the operation and maintenance scheduling system of the back-end main station to form a closed-loop operation and maintenance management logic for the current abnormal behavior.

[0126] In summary, this invention achieves closed-loop control throughout the entire process, from testing and incentives to real-time perception, risk identification, and solution development. This ensures that high-voltage GIS equipment can achieve intelligent, autonomous, and scenario-customized operation and maintenance management capabilities with the support of multi-source data, significantly improving system security and reliability.

[0127] Furthermore, to determine the target operation and maintenance solution, step S4 of this application includes:

[0128] A cross-modal adversarial generative network is introduced, wherein the cross-modal adversarial generative network takes sensor data as input, directional risk control probability as the first output, and operation and maintenance plan as the second output; based on the cross-modal adversarial generative network, the second sensor array is analyzed to determine the target operation and maintenance plan.

[0129] In a further implementation of the present invention, in order to improve the risk identification accuracy and the intelligence level of operation and maintenance suggestion generation of the second sensor array in complex test scenarios, a cross-modal adversarial generative network is introduced, namely the deep learning structure constructed in this technical solution for multi-source heterogeneous sensor data processing and decision reasoning, which has the dual objective collaborative optimization capability of multi-modal data modeling, adversarial training and operation and maintenance output.

[0130] Specifically, the network takes a second sensor array obtained during scenario testing as input. To achieve unified data modeling and effective inference, it first performs feature encoding on various types of sensor data by executing a multi-channel feature extraction encoder to extract high-order semantic association features between modalities. For example, a modality-specific convolutional encoder.

[0131] Subsequently, based on the architecture of fault detection and contingency reasoning, sample integration is performed according to the principle of adversarial training. Training samples are determined and training is supervised until convergence. The constructed encoder, fault detection node and contingency reasoning node are cascaded to form the cross-modal adversarial generative network.

[0132] The system generates two outputs: targeted risk control probability and operational contingency plan content. Specifically, the first output is the targeted risk control probability, which is the probability distribution of different risk events in the current test scenario, such as contact ablation, gas leakage, and mechanical lag, and is represented by multi-class soft labels. The second output is the operational contingency plan, which includes the corresponding handling strategies, suggested measures, and expected response levels, expressed in the form of structured text or multi-dimensional labels.

[0133] During the optimal training process, the discriminative architecture performs adversarial training by comparing the generated results with historical labeled data, thereby enhancing the generation accuracy and generalization ability and ensuring that it can maintain reliable inference performance in multiple scenarios, multiple devices, and dynamic states.

[0134] Subsequently, the second sensor array obtained from the test is input, a fast parallel inference process is executed, and the risk control probability judgment result and operation and maintenance plan suggestions under the current device status are output.

[0135] For example, given a set of data from a partial discharge test of a busbar, the network can output that the probability of partial discharge type A is 0.76, indicating that the equipment status is within the tolerable range, but recommends a second verification within two weeks, along with specific operational suggestions such as using a portable ultrasonic partial discharge detector for targeted investigation and increasing the SF6 gas sampling frequency.

[0136] Ultimately, based on the output results and combined with the geographic location information, maintenance cycle, and historical work orders of the GIS equipment, actions such as remote control operations, work dispatching, or early warning pushes are triggered. This significantly enhances the decision-making accuracy and adaptive processing capabilities in complex data environments, propelling high-voltage GIS equipment from responsive operation and maintenance to predictive and proactive intelligent operation and maintenance.

[0137] The high-voltage GIS equipment detection and maintenance method based on multi-sensor data fusion provided in this application has the following technical effects:

[0138] 1. Construction of a Multi-Layer Operation and Maintenance Array: Decompose high-voltage GIS equipment components, integrate component characteristics to form a scalable multi-layer operation and maintenance array, and clarify the operation and maintenance priorities of each component. Systematically analyze the equipment structure and characteristics to provide precise targets and directions for subsequent testing and maintenance, thereby improving the targeting of operation and maintenance.

[0139] 2. Flexible Testing and Intelligent Detection System: A flexible testing circuit is constructed by combining a multi-layer operation and maintenance array. An intelligent detector integrating multiple thresholds is developed, and a micro-sensor topology is linked to achieve scenario-based testing and directional sensing. The testing circuit and sensing method can be dynamically adjusted according to the equipment's operating status, improving the flexibility and accuracy of testing.

[0140] 3. Cross-modal data fusion decision-making: Multi-sensor data is analyzed using cross-modal adversarial generative networks to determine risk control probabilities and make operational and maintenance plans. This effectively integrates different types of sensor data, improving the accuracy of risk identification and operational and maintenance plan development, and ensuring the safe and stable operation of high-voltage GIS equipment.

[0141] Example 2: Based on the same inventive concept as the high-voltage GIS equipment detection and maintenance method based on multi-sensor data fusion in the foregoing examples, such as... Figure 2 As shown, this application provides a high-voltage GIS equipment inspection and maintenance system based on multi-sensor data fusion, the system comprising:

[0142] The decomposition and integration unit 11 is used to determine a multi-layer operation and maintenance array by decomposing and integrating the characteristics of equipment components for high-voltage GIS equipment.

[0143] A development unit 12 is constructed to connect to the dynamic simulation platform, deploy scenario-based test circuits using the multi-layer operation and maintenance array, construct flexible test circuits, and develop intelligent detectors, wherein the intelligent detectors are connected to the micro-sensing topology and the flexible test circuits.

[0144] The directional testing unit 13 is used to operate with the high-voltage GIS equipment. The micro-sensor topology triggers self-sensing discrimination based on the first threshold condition, and triggers the directional reassembly and testing of the flexible test circuit based on the second circuit threshold based on the discrimination result. Simultaneously, it triggers directional sensing based on the third sensing threshold to determine the second sensor array.

[0145] The equipment operation and maintenance unit 14 is used to perform directional risk control probability determination and operation and maintenance plan decision-making for the second sensor array, determine the target operation and maintenance plan, and perform operation and maintenance management of the high-voltage GIS equipment.

[0146] Furthermore, the decomposition and integration unit 11 performs the following steps: Based on the equipment components of the high-voltage GIS equipment, a first maintenance matrix is ​​determined, wherein the first maintenance matrix includes at least circuit breakers, disconnectors, and busbars, and is labeled with component motion characteristics; based on the component characteristics of the equipment components, a second maintenance matrix is ​​determined, wherein the second maintenance matrix includes mechanical and electrical characteristics cascaded to circuit breakers, thermal and mechanical characteristics cascaded to disconnectors, and partial discharge and gas state characteristics cascaded to busbars; based on the first and second maintenance matrices, the multi-layer maintenance array is determined, wherein the multi-layer maintenance array is expandable, including intra-layer expansion and hierarchical expansion.

[0147] Furthermore, the construction and development unit 12 performs the following steps: defining an operation and maintenance scenario set based on the multi-layer operation and maintenance array, wherein the operation and maintenance scenario is defined with an emphasis on testing based on a single array element or a group of array elements; determining a flexible test circuit based on the operation and maintenance scenario set through simulation and optimization, wherein the flexible test circuit is an external microcircuit of the high-voltage GIS equipment, and the flexible test circuit is based on the scenario-based reorganization of circuit loops using line switches.

[0148] Furthermore, the construction and development unit 12 performs the following steps: based on the multi-layer operation and maintenance array, deploy a micro-sensor topology based on directional testing guidance; retrieve the operation and maintenance records of the high-voltage GIS equipment, perform critical value mining under abnormal operating conditions, and convert them into risk control thresholds based on the micro-sensor topology, wherein the micro-sensor topology and the risk control thresholds correspond one-to-one; cascade the micro-sensor topology and the risk control thresholds to deploy the first threshold condition.

[0149] Furthermore, the construction and development unit 12 performs the following steps: defining a second circuit threshold based on the flexible test circuit, wherein the second circuit threshold performs circuit loop reassembly based on the test scenario; defining a third sensing threshold based on the micro-sensor topology, wherein the third sensing threshold performs sensing orientation triggering based on the test scenario; and performing driven training based on the underlying operation and maintenance test logic using the first threshold condition, the second circuit threshold, and the third sensing threshold to determine the intelligent detector.

[0150] Furthermore, the directional testing unit 13 performs the following steps: as the high-voltage GIS equipment operates, the micro-sensor topology synchronously performs sensing detection to determine the first sensor array; based on the first threshold condition, the first sensor array is subjected to over-limit judgment based on the risk control threshold to determine the risk control sensing part; if the risk control sensing part is not empty, a target test scenario is defined, wherein the target test scenario belongs to the operation and maintenance scenario set.

[0151] Furthermore, the directional testing unit 13 performs the following steps: identifying the target test scenario; matching and determining a circuit loop based on the target test scenario according to the second circuit threshold; performing state adjustment based on the line switch according to the circuit loop to determine a switch instruction set; adjusting the loop of the flexible test circuit according to the switch instruction set to determine the target test circuit; and driving the target test circuit to perform directional scenario testing on the high-voltage GIS equipment.

[0152] Furthermore, the orientation test unit 13 performs the following steps: identifying the target test scene, matching and determining the scene sensing topology based on the target test scene according to the third sensing threshold; and synchronously triggering the scene sensing topology to determine the second sensing array as the target test circuit is triggered.

[0153] Furthermore, the equipment operation and maintenance unit 14 performs the following steps: introducing a cross-modal adversarial generation network, wherein the cross-modal adversarial generation network takes sensor data as input, directional risk control probability as the first output, and operation and maintenance plan as the second output; and analyzes the second sensor array according to the cross-modal adversarial generation network to determine the target operation and maintenance plan.

[0154] Through the foregoing detailed description of the high-voltage GIS equipment detection and maintenance method based on multi-sensor data fusion, those skilled in the art can clearly understand the high-voltage GIS equipment detection and maintenance method and system based on multi-sensor data fusion in this embodiment. As for the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and relevant parts can be referred to the method section description.

[0155] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for the detection and maintenance of high-voltage GIS equipment based on multi-sensor data fusion, characterized in that, The method includes: For high-voltage GIS equipment, a multi-layer operation and maintenance array is determined by decomposing and integrating the characteristics of the equipment components. Connect to the dynamic simulation platform, deploy scenario-based test circuits using the multi-layer operation and maintenance array, construct flexible test circuits, and develop intelligent detectors, wherein the intelligent detectors are connected to the micro-sensing topology and the flexible test circuits; As the high-voltage GIS equipment operates, the micro-sensor topology triggers self-sensing discrimination based on a first threshold condition, and triggers directional reassembly and testing of the flexible test circuit based on a second circuit threshold based on the discrimination result, and simultaneously triggers directional sensing based on a third sensing threshold to determine the second sensor array. For the second sensor array, perform directional risk control probability determination and operation and maintenance plan decision-making based on cross-modal multi-sensor data, determine the target operation and maintenance plan, and perform operation and maintenance management on the high-voltage GIS equipment; The deployment of scenario-based test circuits using the multi-layer operation and maintenance array to construct flexible test circuits includes: Based on the multi-layer operation and maintenance array, an operation and maintenance scenario set is defined, wherein the operation and maintenance scenario is defined with an emphasis on testing based on a single array element or a group of array elements. Based on the aforementioned operation and maintenance scenario set, a flexible test circuit is determined through simulation and optimization. The flexible test circuit is an external microcircuit of the high-voltage GIS equipment, and the flexible test circuit is based on the scenario-based reconfiguration of circuit loops using line switches.

2. The high-voltage GIS equipment inspection and maintenance method based on multi-sensor data fusion as described in claim 1, characterized in that, By decomposing and integrating the characteristics of equipment components, a multi-layered operation and maintenance array is determined, including: A first operation and maintenance matrix is ​​determined based on the equipment components of the high-voltage GIS equipment. The first operation and maintenance matrix includes at least circuit breakers, disconnect switches and busbars, and is marked with the movement characteristics of the components. A second operation and maintenance matrix is ​​determined based on the component characteristics of the equipment components. The second operation and maintenance matrix includes the mechanical and electrical characteristics cascaded to the circuit breaker, the thermal and mechanical characteristics cascaded to the disconnecting switch, and the partial discharge and gas state characteristics cascaded to the busbar. Based on the first and second operation and maintenance matrices, the multi-layer operation and maintenance array is determined, wherein the multi-layer operation and maintenance array is scalable, including intra-layer expansion and hierarchical expansion.

3. The high-voltage GIS equipment inspection and maintenance method based on multi-sensor data fusion as described in claim 1, characterized in that, Before developing a smart detector, the following should be included: Based on the aforementioned multi-layer operation and maintenance array, a micro-sensor topology based on directional testing is deployed; The operation and maintenance records of the high-voltage GIS equipment are retrieved, and critical value mining is performed under abnormal operating conditions. The results are then converted into risk control thresholds based on the micro-sensor topology, wherein the micro-sensor topology and the risk control thresholds correspond one-to-one. The microsensor topology is cascaded with the risk control threshold to deploy the first threshold condition.

4. The high-voltage GIS equipment inspection and maintenance method based on multi-sensor data fusion as described in claim 3, characterized in that, Develop intelligent detectors, including: Based on the flexible test circuit, a second circuit threshold is defined, wherein the second circuit threshold performs circuit loop reassembly based on the test scenario; Based on the microsensor topology, a third sensing threshold is defined, wherein the third sensing threshold performs sensing orientation triggering based on the test scenario; Using the first threshold condition, the second circuit threshold, and the third sensing threshold, drive training based on the underlying operation and maintenance test logic is carried out to determine the intelligent detector.

5. The high-voltage GIS equipment inspection and maintenance method based on multi-sensor data fusion as described in claim 1, characterized in that, Triggering self-sensing discrimination based on the first threshold condition includes: As the high-voltage GIS equipment operates, the micro-sensor topology synchronously performs sensing and detection to determine the first sensor array; Based on the first threshold condition, the first sensor array is subjected to over-limit determination based on the risk control threshold to determine the risk control sensing part; If the risk control sensing part is not empty, define a target test scenario, wherein the target test scenario belongs to the operation and maintenance scenario set.

6. The high-voltage GIS equipment inspection and maintenance method based on multi-sensor data fusion as described in claim 5, characterized in that, Triggering directional reconfiguration and testing of flexible test circuits based on a second circuit threshold includes: Identify the target test scenario, and determine the circuit loop based on the target test scenario according to the second circuit threshold; Based on the circuit loop, perform state adjustment based on the line switch to determine the switch instruction set; According to the switching instruction set, the flexible test circuit is loop adjusted to determine the target test circuit; The target test circuit is driven to perform directional scene testing on the high-voltage GIS equipment.

7. The high-voltage GIS equipment inspection and maintenance method based on multi-sensor data fusion as described in claim 6, characterized in that, Triggering directional sensing based on a third sensing threshold includes: Identify the target test scenario, and determine the scene sensing topology based on the target test scenario according to the third sensing threshold; Upon triggering of the target test circuit, the scene sensing topology synchronously triggers perception to determine the second sensing array.

8. The high-voltage GIS equipment inspection and maintenance method based on multi-sensor data fusion as described in claim 1, characterized in that, Determine the target operation and maintenance plan, including: A cross-modal adversarial generative network is introduced, wherein the cross-modal adversarial generative network takes sensor data as input, directional risk control probability as the first output, and operation and maintenance plan as the second output; Based on the cross-modal adversarial generative network, the second sensor array is analyzed to determine the target operation and maintenance plan.

9. A high-voltage GIS equipment inspection and maintenance system based on multi-sensor data fusion, characterized in that, The system is used to perform the high-voltage GIS equipment inspection and maintenance method based on multi-sensor data fusion as described in any one of claims 1-8, the system comprising: The decomposition and integration unit is used to determine a multi-layer operation and maintenance array by decomposing and integrating the characteristics of equipment components for high-voltage GIS equipment. A development unit is constructed to connect to the dynamic simulation platform, deploy scenario-based test circuits using the multi-layer operation and maintenance array, construct flexible test circuits, and develop intelligent detectors, wherein the intelligent detectors are connected to the micro-sensing topology and the flexible test circuits. The directional testing unit is used to operate with the high-voltage GIS equipment. The micro-sensor topology triggers self-sensing discrimination based on a first threshold condition. The discrimination result triggers directional reassembly and testing of the flexible test circuit based on a second circuit threshold. Simultaneously, it triggers directional sensing based on a third sensing threshold to determine the second sensor array. The equipment operation and maintenance unit is used to perform directional risk control probability determination and operation and maintenance plan decision-making for the second sensor array, determine the target operation and maintenance plan, and perform operation and maintenance management of the high-voltage GIS equipment.

Images