A high-voltage switch cabinet visual monitoring and sequence control method and related equipment

Abstract

The application provides a high-voltage switch cabinet visual monitoring and sequential control method and related equipment, and belongs to the technical field of power equipment automation. The method comprises the following steps: establishing a mapping relationship between a sequential control action step and a monitoring object; collecting multi-source state data and packaging the multi-source state data into state tokens containing a confidence value; constructing three types of evidence sets, namely, pre-action admission, in-action verification and post-action confirmation; performing weighted consistency comparison between the admission evidence set and a preset condition matrix, and outputting a hierarchical control signal; autonomously repairing abnormal data through resampling or image focusing instructions; comparing an action trajectory with a standard envelope line by using a time regularization algorithm; realizing proxy state deduction based on a knowledge graph; and generating a sequential control evidence chain record containing full-process decision basis. The scheme solves the problem of the disconnection between monitoring data and dynamic operation processes, realizes the spatio-temporal alignment and confidence fusion of multi-source heterogeneous data, improves the sequential control safety, and reduces the false locking rate.

Description

Technical Field

[0001] This application relates to the field of power equipment automation technology, and in particular to a visual monitoring and sequential control method and related equipment for high-voltage switchgear. Background Technology

[0002] High-voltage switchgear is widely used in power and distribution systems, and its operational safety directly affects the continuity of power supply and distribution. With the popularization of unmanned operation and remote maintenance, high-voltage switchgear has gradually been equipped with various online monitoring units such as temperature and humidity, partial discharge, temperature measurement, mechanical characteristics, and video monitoring, and has introduced one-button sequential control technology to realize the automated operation of components such as circuit breakers, chassis vehicles, and grounding switches, thereby reducing manual intervention.

[0003] In some related technologies, switchgear sequential control operation scenarios typically involve connecting various sensors to a monitoring platform for centralized display and alarm generation. When executing sequential control actions, the system often relies on fixed threshold alarms from a single sensor, or employs a simple dual-confirmation mechanism of location feedback and video feed to determine whether to proceed with the next operation.

[0004] However, the risk sources and safety access conditions faced by switchgear during different sequential control steps (such as chassis vehicle advancement and circuit breaker closing) are dynamically changing. Related technologies typically process various multi-source heterogeneous data (such as electrical parameters, temperature, partial discharge, and images) independently. These data have varying sampling periods and lack deep integration with specific sequential control steps. When a sensor experiences instantaneous data fluctuations due to external interference, or when multiple monitored parameters simultaneously show minor anomalies below the alarm threshold, existing static thresholds and simple confirmation mechanisms cannot comprehensively assess the true risk of the current specific operational step. This disconnect between monitoring data and the dynamic operation process leads to frequent misjudgments during sequential control: either instantaneous fluctuations in a single data point trigger erroneous interlocking, frequently interrupting the normal sequential control process and severely reducing operational efficiency; or failure to promptly intercept multiple hidden defects in the equipment, resulting in operation with defects (missed interlocking), thereby causing equipment damage or even serious safety accidents. Summary of the Invention

[0005] This application provides a method and related equipment for visual monitoring and sequential control of high-voltage switchgear, which is used to realize the spatiotemporal alignment and confidence fusion of multi-source heterogeneous data, thereby improving the safety of sequential control and reducing the false blocking rate.

[0006] Firstly, this application provides a method for visual monitoring and sequential control of high-voltage switchgear based on a multi-source evidence chain. This method includes: establishing a mapping relationship between target sequential control action steps of the high-voltage switchgear and the monitored object; collecting multi-source status data of the monitored object and converting the multi-source status data into status tokens, which contain status category, timestamp, and confidence value; extracting status tokens corresponding to the target sequential control action steps and constructing a pre-action access evidence set, an in-action verification evidence set, and a post-action confirmation evidence set for the target sequential control action steps; comparing the status tokens in the pre-action access evidence set with a preset condition matrix for consistency and outputting a permission execution. The system generates either a pass signal or a delayed verification signal; when an execution permission signal is output, a control command is issued to the actuator corresponding to the target sequential control action step; when a delayed verification signal is output, a resampling command or image focusing command is sent to the monitored object that generated the abnormal status token, the corresponding status token is updated, and a consistency comparison is performed again; during and after the issuance of control commands, the status tokens in the action verification evidence set and the action confirmation evidence set are extracted and compared with the target status model, and an action completion signal or an abnormal interruption signal is output; the target sequential control action steps, status tokens, comparison results, and control commands are associated and stored to generate a sequential control evidence chain record.

[0007] The above embodiments establish a mapping relationship between sequential control actions and monitored objects, and encapsulate multi-source heterogeneous data into status tokens containing timestamps and credibility values, deeply binding discrete monitoring parameters with dynamic operation processes. By utilizing evidence sets from the pre-action, during-action, and post-action stages for access comparison, trajectory verification, and logical confirmation, the system can comprehensively assess the true risks of specific operational steps based on multi-source evidence chains. This reduces the interference of single data anomalies on the overall sequential control process, and improves the safety control capabilities and process traceability of automated equipment operation.

[0008] In conjunction with some embodiments of the first aspect, in some embodiments, the process of collecting multi-source state data of the monitored object and converting the multi-source state data into a state token specifically includes: acquiring the original state data of the monitored object in the current sampling period and extracting the fluctuation feature value of the original state data; retrieving the sensor baseline health and current environmental interference coefficient corresponding to the monitored object; inputting the fluctuation feature value, sensor baseline health and environmental interference coefficient into a preset credibility assessment model to calculate the credibility value of the original state data; mapping the original state data to a state category, and concatenating the timestamp and credibility value to encapsulate and generate a state token.

[0009] The above embodiments extract data fluctuation characteristics and combine sensor baseline health and environmental interference coefficients to calculate a confidence value to encapsulate a status token. This mechanism incorporates hardware status and external environmental influences into the data quality considerations, changing the judgment method that relies solely on raw measurement values. It provides a structured input with confidence attributes for sequential control logic, reducing the risk of false alarms due to sensor aging or sudden environmental changes.

[0010] In conjunction with some embodiments of the first aspect, in some embodiments, the consistency comparison of the state tokens in the pre-action access evidence set with the preset condition matrix, and the output of a permission execution signal or a delayed review signal, specifically includes: parsing the state category of each state token in the pre-action access evidence set, matching the corresponding baseline state requirements and basic weights in the preset condition matrix; adjusting the corresponding basic weights using the credibility values ​​of each state token to obtain effective weights; calculating the comprehensive safety score of the target sequential control action steps based on the matching degree between the state category and the baseline state requirements, combined with the effective weights; outputting a permission execution signal when the comprehensive safety score is greater than or equal to a first safety threshold; and outputting a delayed review signal when the comprehensive safety score is less than the first safety threshold but greater than a second safety threshold.

[0011] The above embodiments construct an access determination mechanism based on comprehensive security scoring by analyzing the state category matching benchmark requirements and basic weights, and dynamically generating effective weights using confidence values. This approach cross-integrates multi-source evidence matching degree and data confidence degree, avoiding the rigid processing of traditional logical thresholds that directly locks out due to slight exceedances of a single non-critical parameter. Combined with multi-level security threshold division, the system can output delayed verification signals for critical states, improving the flexibility and fault tolerance of access assessment.

[0012] In conjunction with some embodiments of the first aspect, in some embodiments, during and after the issuance of control commands, the state tokens in the action verification evidence set and the action confirmation evidence set are extracted and compared with the target state model, and an action completion signal or an abnormal interruption signal is output. Specifically, this includes: during the issuance of control commands, extracting continuous transient state tokens in the action verification evidence set, constructing a time-series action trajectory, comparing the similarity of the time-series action trajectory with the standard envelope in the target state model, and outputting an abnormal interruption signal if the deviation is greater than a preset tolerance; after the issuance of control commands, extracting steady-state state tokens in the action confirmation evidence set, logically verifying them with the topological rules in the target state model, and outputting an action completion signal if the verification passes.

[0013] The above embodiments introduce a similarity comparison between the time-series motion trajectory and the standard envelope during the control command issuance process, and perform topological rule logic verification after the action. This closed-loop verification mechanism, which combines transient continuous features with steady-state result features, enables the system to output abnormal interruption signals in a timely manner based on the deviation when the mechanism jams or the action is not in place. This mechanism expands the monitoring dimension from the end position feedback to the entire action cycle, improving the ability to intercept hidden mechanical faults.

[0014] In conjunction with some embodiments of the first aspect, in some embodiments, the similarity comparison between the time-series action trajectory and the standard envelope in the target state model specifically includes: using a time warping algorithm to calculate the shortest warped path distance between the time-series action trajectory and the standard envelope; using the shortest warped path distance as the deviation, determining whether the deviation is greater than a preset tolerance, wherein the continuous transient state token includes at least one of a motor drive current token, a mechanical vibration frequency domain token, and an operating mechanism travel token.

[0015] The above embodiments use a time warping algorithm to calculate the shortest warped path distance between the motion trajectory and the standard envelope as the deviation. This algorithm eliminates trajectory misalignment interference caused by fluctuations in the mechanism's motion speed through non-linear alignment of the time axis. When applied to token comparisons such as motor current or vibration, it improves the objectivity of transient feature similarity assessment and avoids false interruptions caused by differences in normal operation time.

[0016] In conjunction with some embodiments of the first aspect, in some embodiments, when the delayed verification signal is output, a resampling instruction or image focusing instruction is sent to the monitoring object that generated the abnormal status token, the corresponding status token is updated, and a consistency comparison is re-triggered. Specifically, this includes: parsing the type of the abnormal status token that caused the delayed verification signal; if the abnormal status token is a visual image type, then sending an image focusing instruction containing target preset position adjustment and autofocus parameters to the corresponding pan-tilt camera that is the monitoring object, obtaining a new image after sharpening processing to update the corresponding status token; if the abnormal status token is a sensor value type, then sending a resampling instruction to increase the sampling frequency to the corresponding sensor that is the monitoring object, obtaining multiple sample values ​​within the sliding time window and performing filtering and smoothing processing, updating the corresponding status token with the smoothed value; replacing the original abnormal status token in the admission evidence set before the action with the updated status token, and re-triggering the consistency comparison.

[0017] The above embodiments issue image focusing or resampling commands specifically based on the type of abnormal token. This mechanism transforms one-way data reception into closed-loop feedback control with active detection capabilities, acquiring clear images through gimbal adjustments or eliminating numerical glitches through high-frequency sampling filtering. By replacing the original abnormal data with updated, high-quality tokens and re-comparing them, the system gains self-healing capabilities, reducing the frequency of manual intervention.

[0018] In conjunction with some embodiments of the first aspect, in some embodiments, after updating the corresponding state token and re-performing the consistency comparison, the method further includes: if a delayed verification signal is still output after re-performing the consistency comparison, then in a preset association knowledge graph, searching for a proxy monitoring object that has a physical coupling relationship with the monitoring object that generated the abnormal state token; extracting the state token of the proxy monitoring object, calculating the proxy deduction state and proxy credibility; if the proxy credibility meets the preset requirements and the proxy deduction state meets the requirements of the preset condition matrix, then outputting a permission execution signal; otherwise, outputting a manual intervention lockout signal.

[0019] The above embodiments introduce a proxy inference mechanism based on an associated knowledge graph. By mining the physical coupling relationships between components, the inference state and credibility are calculated using the proxy object's state token. This approach constructs a bypass verification channel when a directly monitored node fails, relying on the health data of associated nodes for logical verification. While ensuring the bottom line of security access, it improves the sequential control execution continuity rate of the system under local sensor failure conditions.

[0020] In a second aspect, embodiments of this application provide a high-voltage switchgear visualization monitoring and sequential control device, which includes: one or more processors and a memory; the memory is coupled to the one or more processors, and the memory is used to store computer program code, which includes computer instructions, and the one or more processors call the computer instructions to cause the high-voltage switchgear visualization monitoring and sequential control device to perform the method described in the first aspect and any possible implementation thereof.

[0021] Thirdly, embodiments of this application provide a computer program product containing instructions that, when the computer program product is run on a high-voltage switchgear visual monitoring and sequential control device, cause the high-voltage switchgear visual monitoring and sequential control device to execute the method described in the first aspect and any possible implementation thereof.

[0022] Fourthly, embodiments of this application provide a computer-readable storage medium including instructions that, when executed on a high-voltage switchgear visualization monitoring and sequential control device, cause the high-voltage switchgear visualization monitoring and sequential control device to perform the method described in the first aspect and any possible implementation thereof.

[0023] Understandably, the high-voltage switchgear visualization monitoring and sequential control equipment provided in the second aspect, the computer program product provided in the third aspect, and the computer storage medium provided in the fourth aspect are all used to execute the methods provided in the embodiments of this application. Therefore, the beneficial effects they can achieve can be referred to the beneficial effects in the corresponding methods, and will not be repeated here.

[0024] One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages:

[0025] 1. This application establishes a mapping relationship between sequential control actions and monitored objects, encapsulates multi-source data into trusted status tokens, and deeply binds discrete monitoring parameters with dynamic operations. By combining evidence sets from the pre-action, during-action, and post-action stages for comprehensive analysis, it reduces the interference of single data anomalies on the sequential control process and improves safety control capabilities.

[0026] 2. This application constructs an access control mechanism based on comprehensive security scoring by parsing the state category matching benchmark requirements and dynamically generating effective weights using credibility values. This avoids the rigid handling of traditional logic thresholds that directly locks out systems due to slight exceedances of a single non-critical parameter, thus improving the system's fault tolerance.

[0027] 3. This application constructs a closed-loop verification mechanism by introducing similarity comparison between the motion trajectory and the standard envelope, as well as post-motion logic verification. This expands the monitoring dimension from the endpoint location feedback to the entire motion cycle, improving the ability to intercept hidden mechanical faults and multiple defects. Attached Figure Description

[0028] Figure 1 This is a flowchart illustrating a high-voltage switchgear visualization monitoring and sequential control method in an embodiment of this application.

[0029] Figure 2 This is a schematic diagram of the physical device structure of a high-voltage switchgear visualization monitoring and sequential control equipment in the embodiments of this application. Detailed Implementation

[0030] The terminology used in the following embodiments of this application is for the purpose of describing particular embodiments only and is not intended to be limiting of this application. As used in the specification of this application, the singular expressions “a,” “an,” “the,” “the,” and “this” are intended to include the plural expressions as well, unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used in this application refers to any or all possible combinations including one or more of the listed items.

[0031] Hereinafter, the terms "first" and "second" are used for descriptive purposes only and should not be construed as implying or suggesting relative importance or implicitly indicating the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature, and in the description of the embodiments of this application, unless otherwise stated, "multiple" means two or more.

[0032] In the specific application scenario of this application, the KYN28A intelligent online monitoring system is used as a typical implementation carrier. This system is an integrated intelligent monitoring solution designed for KYN28A high-voltage switchgear. Its product composition includes ten functional units: a visual centralized control screen, a one-button sequential control controller, a microcomputer integrated protection device, a three-phase multi-function instrument, a temperature and humidity sensor, a wireless temperature measurement system, a wireless partial discharge sensor, a camera, a wireless infrared temperature measurement sensor, and a mechanical characteristic sensor. These units work together to form a closed-loop system of "sensing-computing-controlling-interaction".

[0033] At the functional division level, the aforementioned components can be categorized into three types according to their responsibilities within the system. The first type is the monitoring object, responsible for sensing the operating status of the switchgear and generating raw status data: the microcomputer-based integrated protection device collects the three-phase current of the circuit breaker (A, B, C) in real time, possesses three-stage overcurrent protection functions, and continuously reports current status data to the central control panel; the three-phase multi-function instrument collects complete electrical parameters such as three-phase voltage, current, active / reactive power, power factor, and energy metering in real time, providing a rich source of numerical data; temperature and humidity sensors collect temperature and humidity data from the circuit breaker compartment and cable compartment respectively, and link with the heating device in the control cabinet to actively suppress the risk of condensation; the wireless temperature measurement system consists of one temperature receiver and nine wireless temperature sensors, providing full coverage monitoring of the upper and lower contacts of each phase of the circuit breaker (A, B, C) and the contact temperature of the ABC busbars, achieving monitoring of key current-carrying nodes. The system features comprehensive temperature measurement without blind spots. A wireless partial discharge sensor, consisting of a receiver and a sensor, monitors partial discharge signals within high-voltage electrical equipment in real time to assess insulation status and provide early warnings of potential insulation faults. Cameras are deployed in zones according to the monitoring area, covering the circuit breaker operation screen, the electric chassis vehicle's entry and exit position screen, and the electric grounding switch opening screen, providing image-based status tokens for each sequential control step. A wireless infrared temperature sensor specifically monitors the infrared thermal image temperature distribution in cable compartments, capturing abnormal thermal fields caused by poor contact or overload. A mechanical characteristic sensor monitors the mechanical performance of the high-voltage circuit breaker operating mechanism online, collecting sequential data such as the operating mechanism's stroke timing, motor drive current, and vibration spectrum, providing continuous transient characteristic evidence for constructing a verification evidence set during operation.

[0034] The second category consists of actuators, responsible for receiving control commands and driving the switchgear to complete physical operations: the circuit breaker operating mechanism is responsible for closing and opening the circuit breaker; the electric chassis trolley drive mechanism is responsible for driving the chassis trolley between the working position and the test position; and the electric grounding switch actuator is responsible for closing and opening the grounding switch. The sequential actions of these three types of actuators are uniformly managed and driven by a one-button sequential control controller. As the execution control hub, the one-button sequential control controller issues action commands to the aforementioned actuators sequentially according to preset operating logic and provides real-time feedback on the response status of each mechanism to the central control panel. It is the key intermediate layer that translates sequential control decisions into physical operations.

[0035] The third category is the display / computing / remote communication unit, centrally managed by a visual control panel. This panel aggregates status data reported by all sensors via the RS485 communication bus, performs data parsing, status token encapsulation, evidence set construction, consistency comparison calculations, and sequential control logic decisions. It provides human-machine interaction, real-time status display, hierarchical alarm display, and local operation authorization functions through a touchscreen interface. Simultaneously, it supports communication connections with PC-based monitoring platforms and mobile apps via wired or wireless networks, synchronously pushing real-time status data, alarm events, and evidence chain records to remote locations. This enables remote monitoring and operation authorization by maintenance personnel, meeting the engineering requirements for unattended and remote maintenance. The core customer requirement is to integrate all monitoring and operation functions into a unified screen interface, supporting direct local touchscreen operation, centralized remote backend management, and real-time mobile access. This allows maintenance personnel to control all operating parameters of the switchgear and execute sequential control operations from anywhere, achieving centralized, visual, and integrated management of high-voltage switchgear.

[0036] At the hardware connectivity level, all ten functional units are connected to the visual control panel via RS485 communication lines, forming a converged communication topology centered on the control panel. Specifically, the wireless temperature sensor and its receiver, and the wireless partial discharge sensor and its receiver, first complete the local wireless acquisition of terminal sensor data via a wireless link (such as ZigBee or a self-developed near-field wireless protocol), and then the receiver forwards the data to the control panel via the RS485 interface, achieving flexible sensing coverage in areas with limited installation and wiring. The RS485 bus carries the uplink traffic of all sensor data within the cabinet and is the main communication channel for the convergence of multi-source status data. Externally, the visual control panel establishes a bidirectional data channel with the PC monitoring platform and mobile APP via wired Ethernet or Wi-Fi wireless links, transmitting real-time status, alarm events, and evidence chain records uplink, and receiving remote operation authorization and configuration commands downlink, supporting maintenance personnel to perform status monitoring and sequential control operations anytime, anywhere.

[0037] At the system deployment architecture level, the intelligent online monitoring KYN28A adopts a two-level collaborative architecture of "local cabinet-remote cloud". The cabinet-side visual control screen has the ability to independently complete full data collection, status token encapsulation, evidence set construction, access comparison and judgment, control command issuance, local alarm display, and local cache writing of sequential control evidence chain. Even when network communication is interrupted, it can still independently run the entire sequential control process, ensuring the local autonomy of core functions. The remote PC monitoring platform and mobile APP undertake auxiliary management functions such as remote status overview, historical curve query, alarm information push and reception, sequential control task authorization and approval, and evidence chain record retrieval. After the network is restored, it automatically performs bidirectional data synchronization with the cabinet-side control screen. The local cache data retention period is no less than 180 days, taking into account compliance audit requirements and the continuous accumulation of historical normal action samples.

[0038] It is worth noting that the visual control panel maintains an independent set of data quality labels for each type of access sensor, including clock offset compensation parameters for that sensor channel, sampling baseline health score, and modal confidence coefficient. Based on this channel-level quality labeling mechanism, raw data from different sensors using different physical units and sampling periods are not simply spliced ​​together after being imported into the control panel. Instead, they undergo channel-by-channel quality assessment and clock alignment processing, and are then encapsulated into structured status tokens carrying confidence attributes. This ensures the distinguishability and comparability of multi-source heterogeneous data fusion input from the data access layer, laying a data quality foundation for the subsequent dynamic admission scoring mechanism based on effective weights.

[0039] To facilitate understanding, the method provided in this implementation will be described in detail below, using the above scenario as an example. Please refer to [link / reference]. Figure 1 This is a flowchart illustrating a high-voltage switchgear visualization monitoring and sequential control method in an embodiment of this application.

[0040] S101. Establish the mapping relationship between the target sequential control action steps of the high-voltage switchgear and the monitoring objects.

[0041] Among them, the target sequential control action step refers to a specific operation unit in the automated operation sequence of high-voltage switchgear, such as circuit breaker closing / opening, chassis vehicle advancing to the working position / exiting to the test position, grounding switch closing / opening, etc.; the monitoring object refers to the data acquisition device that senses the operating status of the switchgear, including temperature and humidity sensors, wireless temperature measurement units, wireless partial discharge sensors, mechanical characteristic sensors, three-phase multi-function instruments, wireless infrared temperature measurement devices, and cameras, which can generate numerical, image, or sequential status data respectively; the mapping relationship represents the set of monitoring objects on which a specific sequential control step depends, which is a one-to-many relationship, and the set of monitoring objects called by different steps is different.

[0042] This step is executed during system initialization or sequential control task configuration, deeply binding multi-source data acquisition with the safety access requirements of specific operation steps. This avoids the inefficient processing of indiscriminately applying all monitoring quantities to the judgment logic of all operation steps. Specifically, the system can construct a "sequential control step - monitoring object" mapping table based on the factory cabinet template, primary wiring topology diagram, or sensor installation list. Taking the "circuit breaker closing" step as an example, the corresponding monitoring objects include a three-phase multi-function instrument (providing current and voltage status tokens), a camera (providing circuit breaker open / close position image tokens), a mechanical characteristic sensor (providing operating mechanism travel and vibration tokens), a wireless partial discharge sensor, and a temperature and humidity sensor. The monitoring objects corresponding to the "chassis vehicle advancing to the working position" step include a camera (chassis vehicle entry and exit images), a mechanical characteristic sensor (drive current and vibration), and a position sensor. This mapping relationship clarifies the safety concern domain of each operation step and serves as a structured index for subsequent dynamic extraction of evidence sets step by step. It can be flexibly adjusted according to cabinet type, operating status, or maintenance conditions.

[0043] S102. Collect multi-source status data of the monitored object and convert the multi-source status data into status tokens.

[0044] Multi-source status data refers to heterogeneous data sets from multiple monitoring objects, using different physical dimensions and sampling mechanisms. These are categorized into three types based on their format: numerical (e.g., scalar data with periodic sampling such as RMS current, temperature and humidity measurements, and partial discharge amplitude), image-based (e.g., images of circuit breaker opening / closing positions captured by cameras, and images of chassis entering / exiting positions), and sequence-based (e.g., time-series curves of operating mechanism travel and vibration spectrum sequences collected by mechanical characteristic sensors). Status tokens are data objects that unify and structurally encapsulate the aforementioned raw heterogeneous data. Their fields must include at least: object ID (identifying the monitoring object from which the data originates), data type (numerical / image / sequence), status category (a semantic label obtained through mapping), timestamp (collection time), confidence value (a confidence attribute reflecting data quality), and original value summary (for audit traceability).

[0045] This step is executed periodically after each sequential control task is triggered. It converts the raw physical quantities generated by each monitored object into semantic data units with unified interfaces and quality attributes, providing standardized input for subsequent evidence set construction and consistency comparison. Specifically, the system collects raw data of each monitored object in parallel for the current period. After preprocessing such as clock synchronization and time alignment (establishing a time window centered on the trigger time of the sequential control step and performing nearest neighbor matching or interpolation aggregation on data from different sampling periods), missing value imputation, and outlier removal, tokens are generated through state category mapping. For example, image recognition results are mapped to "circuit breaker open / closed / intermediate state", temperature values ​​are mapped to "normal / warning / abnormal" according to preset intervals, and partial discharge amplitude is mapped to "low risk / medium risk / high risk".

[0046] In some preferred embodiments, the above steps specifically include: acquiring the original state data of the monitored object in the current sampling period, and extracting the fluctuation characteristic value of the original state data; retrieving the sensor baseline health and the current environmental interference coefficient corresponding to the monitored object;

[0047] The fluctuation characteristic value, sensor baseline health and environmental interference coefficient are input into the preset credibility assessment model to calculate the credibility value of the original state data; the original state data is mapped to the state category, and the timestamp and credibility value are concatenated to encapsulate and generate a state token.

[0048] Among them, the fluctuation characteristic value refers to the index that quantifies the stability of the original state data within the current sampling window. For numerical data, it can be the sliding window variance, the absolute value of the slope, or the peak-to-peak value. For image data, it can be the fluctuation amplitude of the image sharpness score (such as the gradient variance calculated based on the Laplacian operator) or the target recognition confidence. The sensor baseline health refers to the comprehensive score of the current operational reliability of a specific sensor, with a value range of [0, 1]. It can be calculated based on historical calibration deviation statistics, recent self-test results, cumulative offline time, and recent fault records. The environmental interference coefficient refers to the potential impact of the current acquisition environment on data quality. Its sources include sudden changes in light intensity, electromagnetic interference intensity, condensation inside the cabinet, cabinet door opening status, and operational vibration, etc. It is converted into an interference factor in the range of [0, 1] through a level mapping (low / medium / high).

[0049] This step assigns objective data quality attributes to each token, enabling the sequential access control judgment to distinguish between two scenarios: "data truly reflects the device status" and "data is distorted due to interference," thus avoiding false alarms caused by sensor aging or sudden environmental changes. Specifically, the confidence value is calculated using a weighted fusion method, for example: Confidence = α × (1 − fluctuation normalized value) + β × sensor baseline health + γ × (1 − environmental interference coefficient), where α, β, and γ are configurable weight coefficients and their sum is 1, with an output value range of [0, 1]. Subsequently, the raw data is mapped to status categories through threshold grading or a classification model, and encapsulated together with the timestamp and confidence value to form the final token. This mechanism allows the token to carry both semantic state and confidence attributes, providing a quantifiable data quality foundation for subsequent dynamic scoring based on effective weights.

[0050] S103. Extract the status tokens corresponding to the target sequential control action steps, and construct the pre-action access evidence set, in-action verification evidence set, and post-action confirmation evidence set for the target sequential control action steps respectively.

[0051] Among them, the pre-action access evidence set refers to the set of state tokens that are in a stable sampling state within a time window before the triggering of the target sequential control action step (e.g., 5 to 30 seconds before triggering), reflecting the initial safety conditions before the equipment performs the operation; the in-action verification evidence set refers to the set of transient state tokens that are continuously collected at a high frequency from the issuance of the control command to the completion of the mechanism action, used to characterize the dynamic process characteristics of the mechanism's movement; and the post-action confirmation evidence set refers to the set of state tokens that remain stable for a period of time after the mechanism action is completed (e.g., continuously stable for 3 to 10 seconds), used to confirm whether the final state of the operation meets the expected topology requirements.

[0052] This step is executed after token collection but before access determination is triggered. Its core design concept is to divide the continuous token flow into three dedicated sets based on time stages and action semantics, serving the three control nodes of "pre-access verification, in-process trajectory monitoring, and post-result confirmation," forming a closed-loop process. Specifically, taking "circuit breaker closing" as an example: the pre-action access evidence set includes circuit breaker position image tokens, grounding switch status tokens, partial discharge risk level tokens, temperature and humidity tokens, and chassis position tokens; the in-action verification evidence set includes operating mechanism travel timing tokens, motor drive current tokens, and mechanical vibration frequency domain tokens; the post-action confirmation evidence set includes circuit breaker closed position image tokens and three-phase current tokens. Each evidence set has minimum component requirements. If any token is missing, the system will determine that the evidence is incomplete and trigger corresponding supplementary collection or blocking logic, ensuring comprehensive coverage of the three-stage control without any omissions.

[0053] S104. Compare the status tokens in the pre-action access evidence set with the preset condition matrix to ensure consistency, and output an execution permission signal or a delayed review signal.

[0054] Among them, the preset condition matrix refers to a multi-dimensional safety access rule table predefined for various sequential control action steps. Its structure is based on sequential control steps as rows (such as circuit breaker closing, chassis vehicle advancement, etc.) and the baseline state requirements, basic weights, and judgment thresholds of each monitored object as columns. Each matrix element records the qualified state standard and its importance weight of a specific monitored object under the corresponding sequential control step. Consistency comparison refers to the process of matching the state category of each token in the pre-action access evidence set with the corresponding baseline state requirements in the matrix, and outputting a comprehensive score by combining the matching results and weights.

[0055] This step is triggered immediately after the pre-action access evidence set is constructed. It is a core safety gating step before sequential control execution, using multi-source evidence for comprehensive evaluation and alarm, thereby simultaneously reducing the false locking rate caused by instantaneous data fluctuations and the risk of missed locking caused by multiple minor anomalies going undetected. Specifically, taking the "circuit breaker closing" step as an example, the baseline state requirements in the preset condition matrix may include: the circuit breaker is in the open position (higher basic weight), the grounding switch is in the open state (higher basic weight), the chassis is in the working position (medium basic weight), the partial discharge risk level is low (medium basic weight), and the video recognition image is unobstructed and the target position is correct (lower basic weight). After the comprehensive score is output, it is handled in three levels: when the score reaches or exceeds the first safety threshold, a permission execution signal is output; when the score is between the first and second safety thresholds, a delayed review signal is output; when the score is below the second safety threshold, a manual intervention locking signal is directly output, realizing hierarchical closed-loop control. Each threshold and weight can be configured independently for different cabinet types and sequential control steps.

[0056] In some preferred embodiments, the above steps specifically include: parsing the state category of each state token in the pre-action admission evidence set, matching the corresponding baseline state requirements and basic weights in a preset condition matrix; and adjusting the corresponding basic weights using the credibility values ​​of each state token to obtain effective weights.

[0057] Based on the matching degree between the state category and the baseline state requirements, and combined with the effective weights, a comprehensive safety score for the target sequential control action steps is calculated. When the comprehensive safety score is greater than or equal to the first safety threshold, a permission execution signal is output. When the comprehensive safety score is less than the first safety threshold but greater than the second safety threshold, a delayed review signal is output.

[0058] Among them, the basic weight refers to the fixed importance coefficient pre-assigned to each monitoring object in the preset condition matrix, reflecting the contribution of the correctness of the monitoring object's status to the overall safety of the target sequential control steps; the effective weight refers to the weight of the actual participation in the scoring after the basic weight is dynamically adjusted by the token credibility value, and the calculation method is: effective weight = basic weight × credibility value. When the credibility of a token is low, the contribution of the monitoring object to the comprehensive score is reduced accordingly to prevent low-quality data from dominating the admission conclusion; the comprehensive safety score is the score obtained by multiplying the matching degree of each token status category with the baseline status requirement (1 for a perfect match, 0 for a perfect mismatch, or [0,1] continuous values ​​using fuzzy membership) by the corresponding effective weight and then summing the scores, with the value range being [0,1].

[0059] The core of this step lies in embedding data quality factors into the admission score, constructing a two-dimensional cross-fusion of "state matching degree × data confidence degree," thus solving the rigidity problem in traditional logic gating where the entire operation is forcibly locked due to insufficient confidence or slight violation of a single non-critical parameter. Specifically, the comprehensive security score = ∑ i (match degree) i ×Effective Weight i The score dynamically changes based on the matching degree of each token's status and the data quality. When the score is greater than or equal to the first security threshold (e.g., 0.85), the admission conditions are deemed sufficient, and a permission execution signal is output. When the score is between the second security threshold (e.g., 0.60) and the first security threshold, it is determined that there is suspicious data but the direct locking standard has not yet been met, and a delayed review signal is output to trigger the active data update process. When the score is lower than the second security threshold, it is determined that the overall risk is unacceptable, and a manual intervention locking signal is directly output. Each threshold can be configured independently for different cabinet types and sequential control steps to ensure the differentiated adaptability of the evaluation standards.

[0060] S105. When the output permission execution signal is given, a control command is issued to the actuator corresponding to the target sequential control action step.

[0061] Among them, the actuator refers to the electromechanical device that receives control commands and drives the switchgear to complete physical operation actions, including at least the circuit breaker operating mechanism (driving the circuit breaker to complete the closing or opening action), the electric chassis vehicle driving mechanism (driving the chassis vehicle to move between the working position and the test position), and the electric grounding switch actuator (driving the grounding switch to complete the closing or opening action); the control command is the data packet of operation command issued to the actuator, and its key fields include at least: target action type (closing / opening / advancing / exiting, etc.), target position parameters, timeout protection time limit (to prevent the equipment from being damaged by continuous power supply after the mechanism is stuck), and interlocking mark (a status flag indicating that the command has passed the anti-misoperation interlocking verification).

[0062] This step, triggered immediately after the comprehensive safety score reaches the first safety threshold and a permission execution signal is output, is a crucial step in the system's transformation of multi-source evidence analysis conclusions into actual physical operations. Specifically, before formally issuing control commands, the system first sends a handshake confirmation command to the corresponding actuator via a one-click sequential control controller to verify the communication link and the actuator's readiness status. Simultaneously, it supports a dual authorization confirmation mechanism via local visual control screen operation and a remote backend platform, ensuring the legitimacy of the command source and the clarity of operational permissions. After the control command is forwarded to the corresponding actuator by the sequential control controller, the system simultaneously starts a timeout timer. If no action feedback is received from the actuator within the preset time limit, the timeout protection is automatically triggered, and the current step is set to an abnormal state to prevent equipment damage caused by continuous energization of the actuator. Furthermore, simultaneously with the issuance of control commands, the system immediately switches to the in-action verification phase, beginning to extract continuous transient tokens from the in-action verification evidence set, achieving seamless integration from access determination to action process monitoring, ensuring the continuity of the entire process control.

[0063] S106. When outputting a delayed verification signal, send a resampling command or image focusing command to the monitoring object that generated the abnormal status token, update the corresponding status token, and re-perform a consistency comparison.

[0064] Among them, the delayed review signal refers to the intermediate judgment result output by the system when the comprehensive security score is between the first security threshold and the second security threshold, indicating that the current access evidence has local data quality deficiencies but the overall risk has not yet reached the critical state of direct closure; the resampling command is a control command issued to sensor numerical monitoring objects, requiring the sampling frequency to be increased within a specified time window to obtain multiple statistical samples, and the key parameters include the target sampling frequency multiplier and the window duration; the image focusing command is an image quality optimization command issued to image monitoring objects, containing the target preset position number, autofocus trigger flag and exposure compensation parameters.

[0065] This step is triggered immediately after the S104 outputs the delayed verification signal. Its core function is to transform the system's one-way passive reception of monitored object data into a closed-loop feedback control with active detection capabilities. This eliminates the problem of unnecessary manual locking caused by low reliability due to sensor momentary jitter or image blur. Specifically, the system first identifies the type of abnormal token causing insufficient scoring: for image-based monitored objects, it issues an image focusing command containing the target's preset position number and autofocus trigger flag, driving the gimbal to switch to the optimal shooting angle and activating exposure optimization and supplemental lighting enhancement; for sensor-based monitored objects, it increases the sampling frequency to 3 to 5 times the normal sampling rate, acquires multiple high-frequency samples within a sliding time window, performs median filtering and smoothing, repackages the token with the smoothed result, replaces the original abnormal token, and automatically triggers a re-comparison.

[0066] In some preferred embodiments, the above steps specifically include: parsing the type of the status token that caused the delay review signal to be abnormal; if the abnormal status token is a visual image type, then sending an image focusing instruction containing target preset position adjustment and autofocus parameters to the corresponding PTZ camera that is the monitoring object, and obtaining a new image after sharpening to update the corresponding status token.

[0067] If the abnormal status token is a sensor value type, a resampling command to increase the sampling frequency is sent to the corresponding sensor that is being monitored. Multiple sampled values ​​within the sliding time window are obtained and filtered and smoothed. The smoothed value is used to update the corresponding status token. The updated status token replaces the original abnormal status token in the admission evidence set before the action, and the consistency comparison is retried.

[0068] Among them, target preset position adjustment refers to switching the pan-tilt camera lens to the optimal shooting angle and focal length parameter combination pre-calibrated for a specific monitoring target (such as the status sign of the circuit breaker operating handle, the chassis vehicle position indicator area), ensuring that the target area is completely and clearly presented in the effective area of ​​the image; sliding time window refers to a continuous data acquisition interval with the current sampling time as the right boundary and extending to the historical direction for a fixed duration, and eliminating the influence of instantaneous noise by aggregating multiple sample values ​​within the window in a statistical manner; filtering and smoothing processing refers to the mathematical processing operation of denoising and smoothing multiple sample values ​​within the window in sequence, and typical implementations include median filtering (effectively suppressing spike pulse interference), moving average (suitable for smoothing high-frequency random noise), or Kalman filtering (suitable for dynamic quantity estimation with motion trends).

[0069] This step addresses the distortion mechanisms of both image-based and sensor-based numerical abnormal tokens, and designs differentiated active repair strategies accordingly. The following section provides a detailed explanation using typical monitoring scenarios of the KYN28A. Specifically, for visual image anomalies, taking chassis vehicle entry and exit position monitoring as an example: when the camera image becomes blurry due to condensation inside the cabinet or operational vibration, and the confidence level of chassis vehicle position recognition falls below the lower confidence limit, the system sends an image focusing command to the pan-tilt unit, including the corresponding preset position angle, autofocus trigger, and exposure compensation coefficient; after the camera completes the adjustment, the system acquires a clear new image and re-executes the recognition to update the recognition result and encapsulate a new token; for sensor value anomalies, taking wireless partial discharge amplitude measurement as an example: if data spikes occur due to electromagnetic interference caused by operations inside the cabinet, the system increases the sampling frequency to 3 to 5 times the original, acquires multiple measurement values ​​within a sliding time window of 500ms to 2000ms, first eliminates spikes through median filtering, and then performs sliding average smoothing, updating the token with the processed result; the original abnormal token is stored in the evidence chain with a "replaced" mark to ensure the complete traceability of the operation process, and then automatically re-triggers the consistency comparison.

[0070] In other embodiments of this application, after updating the corresponding state token and re-performing the consistency comparison, the method further includes:

[0071] If a delayed verification signal is still output after a new consistency comparison, then in the preset associated knowledge graph, a proxy monitoring object that has a physical coupling relationship with the monitoring object that generated the abnormal status token is searched.

[0072] Extract the status token of the agent monitoring object, and calculate the agent's inferred status and agent credibility;

[0073] If the agent's credibility meets the preset requirements and the agent's deduction state meets the requirements of the preset condition matrix, then an execution permission signal is output; otherwise, a manual intervention lockout signal is output.

[0074] Among them, the association knowledge graph refers to a pre-constructed structured knowledge network that reflects the multi-dimensional associations between various components of the switchgear and sensors. Its node types include equipment component nodes (such as circuit breaker main contacts, operating mechanisms, busbars, and chassis), sensor nodes (each monitored object), and status attribute nodes. The relationship types cover electrical coupling relationships (such as the thermal effect relationship between main contact temperature and load current), mechanical transmission relationships (such as the causal relationship between the operating mechanism stroke and the circuit breaker open / close position), and spatial proximity relationships. Its construction is based on the design BOM, primary wiring diagram, sensor installation list, and cabinet layout. The proxy monitoring object refers to an alternative monitoring node in the knowledge graph that has a physical coupling relationship with the target monitoring object that generates the abnormal token and whose own token credibility meets the requirements. The proxy inference state refers to the calculation result of indirectly inferring the state that the target abnormal object should be in through the token data of the proxy object. The proxy credibility refers to the quantitative score of the reliability of the proxy inference conclusion.

[0075] This step is triggered after the active repair process (resampling or image focusing) still fails to generate a qualified token. It serves as a fallback verification channel when direct monitoring fails. The core of this step is to rely on knowledge graphs to mine the physical coupling relationship between components and use relevant health node data for logical verification when there is a local sensor failure. This ensures that the safety baseline is not lowered and maintains the continuity of sequential control execution. Specifically, taking the example of a camera being unable to update the open / close position image token during the circuit breaker closing process due to continuous obstruction: The system starts from the camera node in the knowledge graph, searches for proxy nodes with mechanical transmission and electrical coupling relationships with it, and determines that the operating mechanism travel sensor token and mechanical vibration sensor token can be used as proxy basis; the proxy deduction status is calculated by weighted voting (e.g., "travel is in place and vibration sequence meets the opening characteristics, the circuit breaker is deduced to be in the open position"), and the weighted average of the credibility of the two proxy tokens is used as the proxy credibility; if the proxy credibility exceeds the preset threshold (e.g., 0.75) and the proxy deduction status is consistent with the preset condition matrix requirements, a permission execution signal is output; otherwise, a manual intervention interlock signal is output, a pop-up alarm is pushed to the visual control screen simultaneously, and the sequential control task is locked, requiring maintenance personnel to confirm on-site before continuing the operation.

[0076] S107. During and after the issuance of control commands, extract the status tokens from the verification evidence set during the action and the confirmation evidence set after the action, compare them with the target status model, and output the action completion signal or the abnormal interruption signal.

[0077] Among them, the target state model refers to a reference model that is pre-established for a specific sequential control action step and describes the evolution law of the equipment state during normal execution. It is divided into two categories: transient model and steady-state model. The transient model describes the normal trajectory range of each physical quantity changing with time during the execution of the mechanism action in the form of a standard envelope, and is used to capture the continuous dynamic characteristics during the action process. The steady-state model describes the logical constraint relationship that each monitored object token should satisfy after the action is completed in the form of a set of topological rules, and is used to verify the consistency and correctness of the final state of the operation. The action completion signal indicates that the target sequential control action step has been executed safely and completely and the final state has passed the logical verification. The abnormal interruption signal indicates that an abnormal deviation is detected during or after the action is executed, and the operation must be stopped immediately.

[0078] This step operates continuously throughout the entire process of control command issuance and in the two time phases following its completion. Its core lies in expanding the traditional single-point confirmation method, which relies solely on endpoint position feedback, into a dual closed loop covering the entire action cycle: time-series trajectory monitoring and final-state logic verification. This allows for timely interception of hidden mechanical faults such as mechanism jamming, refusal to move, and incomplete travel. Specifically, during control command issuance, the system continuously extracts continuous transient tokens such as motor drive current tokens, operating mechanism travel tokens, and mechanical vibration frequency domain tokens at a high sampling frequency. These tokens are then spliced ​​into a time-series action trajectory and compared in real-time with a standard envelope. If the trajectory deviation exceeds the preset tolerance (e.g., during chassis vehicle propulsion, the drive current continuously exceeds the normal peak range during the expected descent phase due to mechanism jamming), an abnormal interruption signal is immediately output, and the control command is cut off, triggering a protection action. After the action is completed, the system extracts the steady-state token set and performs logic verification against each topology rule. For example, after the circuit breaker is closed, the closed position image token status category should be "closed," and the open position token status category should be "open." Only when all rules pass verification is an action completion signal output, triggering the next sequential control step.

[0079] In some preferred embodiments, the above steps specifically include:

[0080] During the issuance of control commands, the continuous transient state tokens in the verification evidence set of the action are extracted, a time series action trajectory is constructed, and the similarity of the time series action trajectory with the standard envelope in the target state model is compared. If the deviation is greater than the preset tolerance, an abnormal interruption signal is output.

[0081] After issuing the control command, the steady-state state token in the action confirmation evidence set is extracted and logically verified against the topological rules in the target state model. If the verification passes, the action completion signal is output.

[0082] Among them, the standard envelope refers to the time series reference boundary presented in the form of a tolerance band above and below the mean, based on the time series statistical analysis results of historical normal operation samples of the same type of switchgear. It reflects the reasonable fluctuation range of each physical quantity over time under normal execution conditions for a specific sequential control action. Its source can be one or more of the following: factory calibration data, historical normal operation sample statistics, or template migration of the same type of switchgear. The topology rule refers to the formal constraint description of the logical relationship between the tokens of each monitoring object after the sequential control action is completed, based on the primary wiring safety constraints and electrical operation procedures. Each rule corresponds to an electrical or positional state logic that should be satisfied.

[0083] This step breaks down the closed-loop verification of the action into two independent verification stages: process monitoring and final state confirmation. This avoids the risk of missed detection caused by "abnormalities in the action process but the endpoint position just happens to be in place", and achieves dual independent protection for process safety and correct results.

[0084] Specifically, during the process monitoring phase, a multi-dimensional time-series action trajectory is constructed using the motor drive current token as the core, supplemented by the operating mechanism travel token and the mechanical vibration frequency domain token. This trajectory is then compared with a standard envelope for similarity analysis. If the trajectory remains within the envelope, the action continues. If the deviation exceeds the preset tolerance (e.g., mechanism jamming causing the current to remain high even when it should be decreasing), an abnormal interruption signal is immediately output, and protection actions are triggered simultaneously. The jamming time and deviation are recorded for subsequent analysis. During the final state confirmation phase, the topology rules must include at least the following typical entries: After the circuit breaker is closed, the closing position image token's status category is "closed" and the opening position token is "open," with the two token states mutually confirming each other; after the chassis reaches the working position, both the position sensor token and the camera image recognition token indicate "working position" and are consistent with the control command target position; under the grounding switch closing state, the circuit breaker token must not be in the "closed" state. An action completion signal is output only if all the above rules are met; any non-compliance results in an abnormal alarm and blocks subsequent steps.

[0085] In other embodiments, a similarity comparison is performed between the time-series action trajectory and the standard envelope in the target state model, specifically including:

[0086] The time warping algorithm is used to calculate the shortest warped path distance between the time series action trajectory and the standard envelope.

[0087] The shortest regular path distance is used as the deviation degree. It is determined whether the deviation degree is greater than the preset tolerance. The continuous transient state token includes at least one of the motor drive current token, mechanical vibration frequency domain token and operating mechanism travel token.

[0088] Among them, the time warping algorithm refers to the algorithm that calculates the morphological similarity between two time series by non-linearly bending and aligning the time axis. The typical implementation is Dynamic Time Warping (DTW), which constructs the cumulative cost matrix between the two series through dynamic programming and backtracks to solve the optimal alignment path. The cumulative cost of the obtained path is the "shortest warping path distance", which is used to quantify the degree of morphological difference between the two series after time elastic alignment. The preset tolerance refers to the acceptable upper limit of deviation pre-set based on the distribution of the shortest warping path distance of historical normal action samples (such as taking the historical sample mean plus 2 standard deviations). It can be updated regularly according to cabinet age or seasonal temperature and humidity to maintain adaptability.

[0089] This step introduces DTW (Time-Divergence Wave) instead of point-by-point Euclidean distance comparison because, in actual operation, the execution speed of the same type of action naturally fluctuates due to differences in ambient temperature, lubrication level, or load conditions. If point-by-point comparison is performed at fixed time points, normal speed differences will cause time axis misalignment, misjudging healthy actions as abnormal. DTW eliminates the misalignment interference caused by speed fluctuations through non-linear alignment of the time axis, ensuring that the deviation only responds to actual abnormalities (jamming, refusal to move, oscillation, etc.) rather than normal speed changes, thereby effectively reducing the false interruption rate.

[0090] Specifically, before comparison, the system performs normalization and resampling preprocessing on the motion trajectory sequence and the standard envelope to unify the time axis resolution and amplitude range. Taking the motor drive current token during the chassis vehicle's propulsion process as an example, the current sequence of normal action presents a typical three-segment form of "starting peak - constant speed platform - deceleration upon arrival". DTW can flexibly align the current curves under different action durations to the standard envelope and calculate the shortest regular path distance. If the mechanism jams, the current will show a continuous abnormally high value in the constant speed segment, the shape will be severely deviated, and the path distance will exceed the preset tolerance. The system will immediately output an abnormal interruption signal. If the time axis is shifted as a whole due to only slightly slower operation speed, and the path distance is still within the tolerance range after DTW alignment, the system will determine that the action is normal and avoid false interruptions.

[0091] S108. Link and store the target sequential control action steps, status tokens, comparison results, and control commands to generate a sequential control evidence chain record.

[0092] Among them, associated storage refers to binding and storing multiple interrelated data objects with explicit relation keys to ensure that the causal relationship and logical connection between each data unit are completely solidified at the storage level and can be restored and replayed; the sequential control evidence chain record refers to a structured traceability file formed by linking the status tokens, comparison results at each stage, control instructions and disposal conclusions throughout the entire life cycle of the operation step as the core unit. Its role is to provide complete and inseparable original data evidence for post-event analysis, responsibility tracing, model sample accumulation and compliance audit.

[0093] This step is executed immediately after each sequential control action step is completed or abnormally interrupted, solidifying all-element data generated during the system's dynamic decision-making process into a permanent evidence archive with high integrity and traceability. Specifically, the associated storage uses a composite master index of step number and execution timestamp, and the bound fields include at least: sequential control step identifier (step number, associated task ID, and operator authorization identifier), details of each token in the pre-action access evidence set (object ID, data type, status category, timestamp, credibility value, original value summary, and replacement marker), consistency comparison results (comprehensive security score, details of each token's matching, effective weight distribution, and output signal type), control instruction records (instruction content, issuance time, timeout limit, and local / remote authorization identifier), and action process comparison results (deviation time series, topology verification results, and final output signal type). The storage structure can adopt a combination of a relational master table (storing step-level associated metadata) and a time-series database (storing token time-series data), supporting rapid retrieval and complete playback of the entire decision-making basis for a sequential control operation by step number or time range.

[0094] Records can be appended with digest hash values ​​to prevent tampering and enhance audit non-repudiation; local caching prioritizes real-time writing, automatically synchronizing with the backend platform after network recovery, and the log retention period is no less than 180 days, balancing compliance audit requirements with the need for continuous accumulation of historical normal action samples.

[0095] The following describes the high-voltage switchgear visualization monitoring and sequential control equipment in the embodiments of this invention from the perspective of hardware processing. Please refer to [link / reference needed]. Figure 2 This is a schematic diagram of the physical device structure of a high-voltage switchgear visualization monitoring and sequential control equipment in the embodiments of this application.

[0096] It should be noted that, Figure 2 The structure of the high-voltage switchgear visualization monitoring and sequential control equipment shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of the present invention.

[0097] like Figure 2As shown, the high-voltage switchgear visualization monitoring and sequential control device includes a central processing unit (CPU) 201, which can perform various appropriate actions and processes based on programs stored in read-only memory (ROM) 202 or programs loaded from storage section 208 into random access memory (RAM) 203, such as performing the methods described in the above embodiments. The RAM 203 also stores various programs and data required for system operation. The CPU 201, ROM 202, and RAM 203 are interconnected via bus 204. An input / output (I / O) interface 205 is also connected to bus 204.

[0098] The following components are connected to I / O interface 205: input section 206 including audio input devices, push-button switches, etc.; output section 207 including a liquid crystal display (LCD) and audio output devices, indicator lights, etc.; storage section 208 including a hard disk, etc.; and communication section 209 including a network interface card such as a LAN (Local Area Network) card, modem, etc. Communication section 209 performs communication processing via a network such as the Internet. Drive 210 is also connected to I / O interface 205 as needed. Removable media 211, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., are installed on drive 210 as needed so that computer programs read from them can be installed into storage section 208 as needed.

[0099] In particular, according to embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present invention include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing computer programs for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 209, and / or installed from removable medium 211. When the computer program is executed by central processing unit (CPU) 201, it performs the various functions defined in the present invention.

[0100] It should be noted that specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), flash memory, optical fiber, portable compact disc read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this invention, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0101] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present invention. Each block in a flowchart or block diagram may represent a module, program segment, or portion of code, which contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those shown in the drawings.

[0102] Specifically, the high-voltage switchgear visualization monitoring and sequential control device of this embodiment includes a processor and a memory. The memory stores a computer program. When the computer program is executed by the processor, it implements the high-voltage switchgear visualization monitoring and sequential control method provided in the above embodiment.

[0103] In another aspect, the present invention also provides a computer-readable storage medium, which may be included in the high-voltage switchgear visualization monitoring and sequential control device described in the above embodiments; or it may exist independently and not assembled into the high-voltage switchgear visualization monitoring and sequential control device. The storage medium carries one or more computer programs, which, when executed by a processor of the high-voltage switchgear visualization monitoring and sequential control device, cause the high-voltage switchgear visualization monitoring and sequential control device to implement the high-voltage switchgear visualization monitoring and sequential control method provided in the above embodiments.

[0104] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit it. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

[0105] As used in the above embodiments, depending on the context, the term "when..." can be interpreted as meaning "if...", "after...", "in response to determining...", or "in response to detecting...". Similarly, depending on the context, the phrase "when determining..." or "if (the stated condition or event) is interpreted as meaning "if determining...", "in response to determining...", "when (the stated condition or event) is detected", or "in response to detecting (the stated condition or event)".

[0106] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. This program can be stored in a computer-readable storage medium, and when executed, it can include the processes described in the above method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as ROM or random access memory (RAM), magnetic disks, or optical disks.

Claims

1. A method for visual monitoring and sequential control of high-voltage switchgear based on multi-source evidence chains, characterized in that, The method includes: Establish a mapping relationship between the target sequential control action steps of high-voltage switchgear and the monitored objects; Collect multi-source status data of the monitored object, and convert the multi-source status data into a status token, the status token containing a status category, a timestamp, and a confidence value; Extract the status tokens corresponding to the target sequential control action steps, and construct the pre-action access evidence set, the in-action verification evidence set, and the post-action confirmation evidence set for the target sequential control action steps respectively; The status token in the pre-action access evidence set is compared with the preset condition matrix for consistency, and an execution permission signal or a delayed review signal is output. When the permission execution signal is output, a control command is issued to the actuator corresponding to the target sequential control action step; When the delayed verification signal is output, a resampling command or image focusing command is sent to the monitoring object that generated the abnormal status token, the corresponding status token is updated and the consistency comparison is performed again. During and after the issuance of the control command, the status tokens in the verification evidence set and the confirmation evidence set after the action are extracted and compared with the target status model, and the action completion signal or abnormal interruption signal is output. The target sequential control action steps, the status token, the comparison result, and the control command are associated and stored to generate a sequential control evidence chain record.

2. The method according to claim 1, characterized in that, The process of collecting multi-source status data of the monitored object and converting the multi-source status data into status tokens specifically includes: Obtain the raw state data of the monitored object in the current sampling period, and extract the fluctuation characteristic values ​​of the raw state data; Retrieve the baseline health status of the sensor corresponding to the monitored object and the current environmental interference coefficient; The fluctuation characteristic value, the sensor baseline health value, and the environmental interference coefficient are input into a preset credibility assessment model to calculate the credibility value of the original state data. The original state data is mapped to the state category, and the timestamp and the credibility value are concatenated to encapsulate and generate the state token.

3. The method according to claim 1, characterized in that, The step of comparing the status token in the pre-action access evidence set with the preset condition matrix for consistency and outputting a permission execution signal or a delayed review signal specifically includes: The state category of each state token in the pre-action admission evidence set is analyzed, and the corresponding baseline state requirements and basic weights are matched in the preset condition matrix. The corresponding base weights are adjusted using the credibility values ​​of each state token to obtain effective weights; Based on the matching degree between the state category and the baseline state requirement, and combined with the effective weight, the comprehensive safety score of the target sequential control action steps is calculated; When the comprehensive security score is greater than or equal to the first security threshold, the permission execution signal is output; When the overall security score is less than the first security threshold but greater than the second security threshold, the delayed verification signal is output.

4. The method according to claim 1, characterized in that, During and after the issuance of the control command, the state tokens in the verification evidence set during the action and the confirmation evidence set after the action are extracted and compared with the target state model, and an action completion signal or an abnormal interruption signal is output. Specifically, this includes: During the issuance of the control command, the state tokens of continuous transient states in the verification evidence set of the action are extracted, a time series action trajectory is constructed, and the similarity of the time series action trajectory with the standard envelope in the target state model is compared. If the deviation is greater than the preset tolerance, the abnormal interruption signal is output. After the control command is issued, the steady-state state token in the action confirmation evidence set is extracted and logically verified against the topological rules in the target state model. If the verification passes, the action completion signal is output.

5. The method according to claim 4, characterized in that, The step of comparing the similarity between the time-series action trajectory and the standard envelope in the target state model specifically includes: The time warping algorithm is used to calculate the shortest warped path distance between the time series action trajectory and the standard envelope. The shortest regularized path distance is used as the deviation. It is determined whether the deviation is greater than the preset tolerance. The continuous transient state token includes at least one of motor drive current token, mechanical vibration frequency domain token, and operating mechanism travel token.

6. The method according to claim 1, characterized in that, When the delayed verification signal is output, a resampling command or image focusing command is sent to the monitoring object that generated the abnormal status token, the corresponding status token is updated, and the consistency comparison is performed again. Specifically, this includes: Analyze the type of state token that caused the abnormality in the delayed verification signal; If the abnormal status token is a visual image type, then send the image focusing instruction containing target preset position adjustment and autofocus parameters to the corresponding PTZ camera that is the monitoring object, and obtain a new image after sharpening to update the corresponding status token. If the abnormal status token is a sensor value type, then a resampling instruction to increase the sampling frequency is sent to the corresponding sensor that is the monitoring object, to obtain multiple sampled values ​​within the sliding time window and to perform filtering and smoothing processing, and to update the corresponding status token with the smoothed value. Replace the original abnormal state token in the pre-action admission evidence set with the updated state token, and re-trigger the consistency comparison.

7. The method according to claim 6, characterized in that, After the steps of updating the corresponding state token and re-performing the consistency comparison, the method further includes: If the delayed verification signal is still output after the consistency comparison is performed again, then in the preset associated knowledge graph, a proxy monitoring object that has a physical coupling relationship with the monitoring object that generated the abnormal state token is searched. Extract the status token of the agent monitoring object, and calculate the agent's inferred status and agent credibility; If the agent's credibility meets the preset requirements and the agent's deduction state meets the requirements of the preset condition matrix, then the permission execution signal is output; otherwise, a manual intervention lockout signal is output.

8. A high-voltage switchgear visualization monitoring and sequential control device, characterized in that, The high-voltage switchgear visual monitoring and sequential control device includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and the one or more processors call the computer instructions to cause the high-voltage switchgear visual monitoring and sequential control device to perform the method as described in any one of claims 1-7.

9. A computer-readable storage medium comprising instructions, characterized in that, When the instruction is executed on the high-voltage switchgear visual monitoring and sequential control device, the high-voltage switchgear visual monitoring and sequential control device performs the method as described in any one of claims 1-7.

10. A computer program product, characterized in that, When the computer program product is run on the high-voltage switchgear visual monitoring and sequential control device, the high-voltage switchgear visual monitoring and sequential control device performs the method as described in any one of claims 1-7.

Images