An intelligent operation and maintenance management system for substation secondary equipment loop

Abstract

The application relates to an intelligent operation and maintenance management system of a substation secondary equipment loop. The technical scheme is as follows: an internet of things sensing layer, an edge computing layer, a mobile operation layer and a platform application layer are included, and a complete closed loop is formed through a standard communication protocol. The intelligent sensing terminal of the internet of things sensing layer realizes real-time monitoring of terminal-level electrical parameters and physical states; the edge computing layer is provided with a digital twin engine integrating multiple models, a three-dimensional virtual model is constructed and state updating is driven; the mobile operation layer provides visual operation guidance in combination with ultra-wideband positioning and augmented reality technology; the platform application layer integrates intelligent diagnosis, operation management and other modules, fault identification and prediction are realized based on a multi-source data fusion algorithm, and the system is continuously optimized through a knowledge depositing mechanism. The application solves the pain points of the traditional operation and maintenance, such as monitoring blind spots, extensive operation and information islands, realizes digital and intelligent management of the whole life cycle of the secondary loop, and improves operation and maintenance efficiency and safety level.

Description

Technical Field

[0001] This invention belongs to the field of power system automation and digitalization technology, and relates to an intelligent operation and maintenance management system for secondary equipment circuits in substations. Background Technology

[0002] The secondary equipment circuits in substations are the "nerve center" for the safe and stable operation of a power system, and the reliability of their terminal connections directly affects the correct operation of protection, measurement and control, and communication functions. Traditional operation and maintenance management of substation secondary circuits suffers from the following prominent pain points:

[0003] 1. Blind spots in condition monitoring: There is a lack of effective online monitoring methods for the condition of the terminals themselves (such as contact resistance, temperature, looseness), relying on regular manual inspections, which cannot provide real-time early warning.

[0004] 2. The work process is rough: maintenance, calibration and other operations rely on paper drawings and manual memory, which poses risks such as difficulty in finding points, low efficiency and easy error. In addition, the work process lacks digital records and traceability.

[0005] 3. Severe information silos: Design institute drawings, construction unit installation records, and operation and maintenance unit maintenance files are scattered and isolated, resulting in a broken data chain throughout the equipment's entire lifecycle.

[0006] 4. Weak safety management: The "five preventions" for secondary circuit operations mainly target primary equipment. The prevention of misoperation in the disconnection and connection of secondary circuits relies on personnel experience and safety measures tickets, and lacks technical mandatory constraints.

[0007] 5. Difficulty in knowledge transfer: Expert experience is difficult to digitize and preserve, and the training cycle for new employees is long and costly.

[0008] While there are scattered solutions for wireless temperature measurement or online monitoring in existing technologies, none of them have deeply integrated the intelligent sensing at the physical layer, the data fusion at the network layer, the digital twin at the platform layer, and the workflow reengineering at the application layer from an ecosystem perspective to form a closed-loop intelligent operation and maintenance management model.

[0009] Therefore, the present invention provides an intelligent operation and maintenance management system for the secondary equipment circuits of substations to solve the above problems. Summary of the Invention

[0010] In view of the problems existing in the prior art, the present invention discloses an intelligent operation and maintenance management system for secondary equipment circuits in substations. The technical solution adopted includes an Internet of Things sensing layer, an edge computing layer, a mobile operation layer, and a platform application layer.

[0011] The IoT sensing layer is deployed inside the secondary equipment cabinet of the substation and includes multiple intelligent sensing terminals. The intelligent sensing terminals are integrated inside the traditional wiring terminals. Their housings are physically isolated into a power cavity and an intelligent cavity. The power cavity is equipped with a conductive connection mechanism and a contact point temperature sensor. The intelligent cavity is equipped with a main control unit, a non-invasive current sensor, a vibration sensor, and a communication unit, which are used to monitor the electrical parameters flowing through the terminals and the physical state of the connection points in real time and to realize digital communication.

[0012] The edge computing layer is deployed locally at the substation and includes at least one edge server and network equipment. The edge server is used to aggregate and process data from the IoT sensing layer and deploys a digital twin engine. This digital twin engine is a software middleware that integrates a real-time data-driven model, a physical rule model, and a geometric model. Its core modules include:

[0013] 1. Real-time synchronization and mapping engine: Apache Kafka / Apache Pulsar is used as the data pipeline, and Redis is used as the real-time state database. Based on the publish-subscribe pattern and in-memory computing, it realizes zero-latency mapping between physical terminals and virtual models;

[0014] 2. 3D Geometry Rendering and Lightweight Engine: 3D visualization is achieved through Three.js / Unity / Unreal Engine, and massive numbers of terminals are drawn efficiently using glTF format models and instantiation rendering technology;

[0015] 3. Physics and Rule Model Engine: Based on graph theory algorithms, a topology connection model is established; combined with Kirchhoff's laws, an electrical quantity propagation model is constructed; and a terminal state logic model is defined through a finite state machine.

[0016] The mobile operation layer includes a mobile operation terminal and personnel positioning tags. The mobile operation terminal is equipped with an operation guidance application, which can receive operation instructions from the edge computing layer and provide cabinet positioning, terminal highlighting recognition, virtual wiring diagram overlay, and operation step visualization guidance in combination with augmented reality technology.

[0017] The platform application layer, deployed on the main site or in the cloud, includes a job management module, an intelligent diagnosis module, an asset management module, a knowledge base module, and a job safety control module. The intelligent diagnosis module integrates three core algorithms: an overload and abnormal load identification algorithm based on real-time current values ​​and historical trends; a temperature rise model deviation analysis algorithm based on the square of the current and measured contact point temperature; and a mechanical loosening identification algorithm based on vibration sensor spectral characteristics and a machine learning classifier. The job safety control module is used for dynamically authorizing operation permissions, real-time monitoring of misoperation risks, and early warning.

[0018] Each layer is connected through a standard communication protocol, forming a complete closed loop from status awareness, data processing, job guidance to intelligent decision-making.

[0019] As a preferred embodiment of the present invention, the configuration of the operation safety management module includes:

[0020] Before starting work: Verify the operator's identity, the validity of the electronic work ticket, and the work time window; only unlock the software "logic lock" for the smart sensing terminals involved in the work order.

[0021] During operation: Authorization of the work area is confirmed through personnel positioning; interaction authentication between intelligent tools and terminals prevents accidental contact with non-target terminals; and non-standard operations are identified through video behavior analysis.

[0022] After the operation: The system automatically verifies the integrity of the circuit, closes the "logic lock" after confirming that there are no errors, and generates a digital operation report containing safety verification records.

[0023] As a preferred embodiment of the present invention, the following steps are included:

[0024] S1: Status monitoring and data aggregation: Each intelligent sensing terminal continuously monitors status parameters such as current, contact point temperature, and vibration, and uploads them to the edge server through the industrial network;

[0025] S2: Digital Twin Construction and Update: The digital twin engine of the edge server dynamically binds real-time monitoring data, historical operation and maintenance data and corresponding entities in the 3D virtual model based on the received data, drives the virtual model status update and calculates the terminal health score.

[0026] S3: Intelligent Diagnosis and Work Order Triggering: The intelligent diagnosis module at the platform application layer, based on multi-source data fusion analysis, identifies potential faults or predicts maintenance needs through the following algorithms, and automatically generates preventative maintenance work orders:

[0027] 1. Overload and abnormal load identification algorithm:

[0028] The following algorithm is used to integrate dynamic threshold early warning:

[0029] ;

[0030] in, and It is the rolling average and standard deviation of historical current in the same period (such as the same time period of each day), and k is the sensitivity coefficient (taken as 2-3). This threshold can adapt to the periodic changes of the load.

[0031] The following algorithm is used for cumulative and trend anomaly detection:

[0032] ;

[0033] in, This represents the normalized deviation between the current value and the expected value. To allow for small deviations, when If the preset threshold H is exceeded, an abnormal trend alarm will be triggered;

[0034] Waveform anomaly identification and feature extraction analysis: Calculate the harmonic distortion rate (THD), waveform factor and other features for each power frequency cycle at the edge side; establish a feature baseline under normal operating conditions; when real-time features exceed the baseline range, combine with threshold judgment to identify waveform distortion anomalies.

[0035] 2. Temperature rise model deviation analysis algorithm:

[0036] A baseline temperature rise model is established, and based on Joule's law and heat conduction theory, the temperature rise at the contact point is determined under ideal clean and tight conditions. With the square of the current It exhibits a linear relationship and is affected by ambient temperature. Influence:

[0037] ;

[0038] in The theoretical thermal resistance of the contact point (determined by the terminal material and design). : Ambient temperature influence coefficient;

[0039] Deviation analysis between measured values ​​and model values: Real-time data acquisition I, , ; Calculate the deviation index D:

[0040] ;

[0041] Analysis logic:

[0042] D ≈ 0: The state is healthy, consistent with the theoretical model;

[0043] D > 0 and continues to increase slowly: This indicates an increase in actual thermal resistance, which is an early sign of oxidation, contamination or minor loosening of the contact surface;

[0044] The growth slope of D is used to predict remaining lifetime (RUL).

[0045] D > 0 and a sudden step: indicates that the fastening screw has suddenly loosened or that an external heat source has affected it;

[0046] D < 0: This is theoretically rare and indicates a malfunction in the temperature sensor or an abnormal improvement in cooling conditions.

[0047] 3. Mechanical loosening identification algorithm: Extract time-domain features, frequency-domain features, and bearing fault frequency band energy from vibration data. Use a gradient boosting decision tree or random forest classification model to output the loosening state and severity level. The algorithm flow is as follows:

[0048] 3.1 Data Preprocessing and Feature Extraction (Completed at the Edge):

[0049] Sampling: Triaxial vibration sensor, sampling frequency ≥ 2kHz;

[0050] Feature extraction (data window per 1 second): Temporal features: Root mean square (RMS), peak value, peak-to-peak value, kurtosis;

[0051] Frequency domain characteristics (after FFT transformation): Within the range of 100Hz-1kHz (avoiding power frequency and its harmonics), calculate the frequency and amplitude of the first N significant spectral peaks, the spectral centroid, and the spectral entropy.

[0052] Bearing fault characteristic frequency band energy: Calculate the energy ratio of several narrow bands related to the screw loosening frequency;

[0053] 3.2 Classification Model (Trained on the edge server, registered for inference at the gateway):

[0054] Algorithm selection: Gradient boosting decision trees (such as LightGBM) or random forests, which perform well on structured features, are lightweight, and are suitable for edge deployment;

[0055] Label preparation: Label the feature data using historical data or simulation experiments (setting "tight", "slightly loose", "significantly loose" states);

[0056] Model output: It not only outputs a binary classification of "normal / loose", but also outputs the probability of looseness or the severity level;

[0057] S4: Work Navigation and Execution: Operators receive work orders via mobile work terminals. The system guides them to the target cabinet using ultra-wideband positioning, highlights the target terminals using augmented reality technology, and overlays virtual wiring diagrams and operation steps. It also verifies the identity of intelligent tools and terminals to prevent misoperation and automatically records key operation data.

[0058] S5: Knowledge Accumulation and Model Optimization: Structure and store operation process data, result data, and handling experience into a knowledge base. Extract fault keywords and associate them with standard fault modes through natural language processing. When the system prediction is inconsistent with the expert judgment, update and optimize the diagnostic model through active learning or ensemble learning voting. At the same time, explore the potential relationship between new feature combinations and faults, and generate suspected new rules for expert review.

[0059] As a preferred embodiment of the present invention, in step S4, the anti-misoperation verification specifically involves: when the intelligent tool approaches an unauthorized intelligent sensing terminal, the mobile operation terminal issues an audible and visual warning; the disconnection and reconnection operation is only permitted after the tool correctly identifies the target terminal identification code.

[0060] As a preferred embodiment of the present invention, it further includes a computer-readable storage medium having a computer program stored thereon, wherein the program, when executed by a processor, implements steps S1-S5.

[0061] The beneficial effects of this invention are:

[0062] Terminal-level status can be detected: By using intelligent sensing terminals, the monitoring reach can be extended to the last connection point of the secondary circuit. Combined with multi-dimensional sensors and precise algorithms, early and accurate detection of hidden faults such as increased contact resistance and mechanical loosening can be achieved, filling monitoring blind spots.

[0063] Digitalization of work processes: By integrating digital twin and augmented reality technologies, abstract drawings and procedures are transformed into intuitive visual guidance. Combined with intelligent tools to prevent errors and verify them, the threshold for operation is greatly reduced, the risk of errors is reduced, and the efficiency and security of operation and maintenance are improved.

[0064] Closed-loop management throughout the entire lifecycle: Connecting data flows across the design, installation, and operation and maintenance stages, establishing a unique "digital ID card" and full lifecycle file for each terminal, and achieving traceability, assessment, and predictability of assets;

[0065] Self-learning intelligent ecosystem: Construct an automated evolution loop of "data-information-knowledge", continuously optimize diagnostic models and operation standards through case accumulation, incremental learning and rule mining, so that the system can continuously improve its intelligence level over time;

[0066] Safety management and control have been comprehensively upgraded: the concept of "five protections for primary equipment" has been extended to secondary circuits. Through dynamic access control, real-time process monitoring, and mandatory operational constraints, a three-dimensional safety protection network has been built to effectively avoid accidents caused by misoperation. Attached Figure Description

[0067] Figure 1 This is a schematic diagram of the overall system architecture of the present invention;

[0068] Figure 2 This is a schematic cross-sectional view of the intelligent sensing terminal in this invention;

[0069] Figure 3 This is a flowchart of the operation navigation process based on digital twin and AR in this invention;

[0070] Figure 4 This is a schematic diagram of the intelligent diagnosis and maintenance decision-making closed loop in this invention.

[0071] Figure 5 This is a logical diagram illustrating the operation safety management in an embodiment of the present invention;

[0072] Figure 6 This is a logic diagram for data interaction and business closed loop.

[0073] Figure descriptions: 1-Conductive connection mechanism; 2-Contact point temperature sensor; 3-Main control unit; 4-Non-invasive current sensor; 5-Vibration sensor; 6-Communication unit. Detailed Implementation

[0074] Example 1: System Deployment and Data Flow

[0075] Reference Figure 1 The system was deployed in the protection room of a 220kV substation. First, the critical circuit terminals in the protection and control cabinets were replaced with intelligent sensing terminals (such as the integrated terminals described in the background section). Each cabinet was equipped with a regional aggregation gateway, which collected data from all intelligent terminals within its cabinet via an RS-485 bus. Each aggregation gateway was connected to an edge server deployed in the secondary equipment room via the substation's industrial fiber optic ring network. The edge server ran a digital twin engine, constructing a detailed 3D model containing all intelligent terminals based on the substation's SCD file and BIM model. Maintenance personnel were equipped with explosion-proof tablets as mobile work terminals and wore UWB positioning tags. The main station platform communicated with the substation's edge server via the power dispatch data network.

[0076] Example 2: Preventive Maintenance Workflow

[0077] Reference Figure 3 and Figure 4 After the system had been running for a period of time, the intelligent diagnostic module detected that the contact point temperature of a certain intelligent terminal in the "Differential Protection A-phase Current Circuit of Main Transformer No. 1" was showing a slow upward trend, and the temperature rise model deviation exceeded the threshold, causing the health score to drop to 75%. The system automatically generated a "Preventive Maintenance - Check and Tighten Wiring Terminals" work order and pushed it to the relevant work team.

[0078] The foreman dispatched a work order to worker Zhang San via a mobile work terminal. Upon arriving at the substation, Zhang San opened the work app on the terminal and scanned the work order's QR code. The system first guided Zhang San to the correct protection cabinet using UWB positioning. Next, Zhang San used the terminal's camera to scan the cabinet; the AR function automatically identified the nameplate on the cabinet and highlighted the area where the target terminal was located. Further alignment with the terminal block revealed the target terminal (and its adjacent terminals for misalignment) framed in a special color on the screen, while a virtual wiring diagram, historical temperature curve, and the SOP animation for this operation were overlaid on the screen.

[0079] Zhang San uses a smart torque wrench with Bluetooth communication. When he attempts to tighten a non-target terminal, the terminal immediately vibrates and issues a voice warning: "Non-operational object, please confirm!" Only after he places the wrench on the target terminal and the system verifies the terminal's identification code via the wrench's integrated RFID reader does the terminal display an "Operational" prompt. After Zhang San completes the tightening according to the SOP, the terminal prompts "Operation completed, awaiting verification." The system then automatically monitors the terminal's temperature change trend over the following period. Once the abnormal trend disappears, the system automatically updates the work order status to "Completed" and records the entire operation process (including location path, pre- and post-operation data comparison, executor, and timestamp) in the terminal's digital asset file.

[0080] Example 3: Work Safety Management

[0081] Reference Figure 5 The work safety management module intervenes throughout the entire process before, during, and after work. Before work, the system verifies the operator's identity, the validity of the electronic work order, and the work time window, and only unlocks the software "logic lock" on specific smart terminals involved in the work order. During work, personnel positioning ensures they are in authorized areas; intelligent tools and terminal interaction authentication prevent accidental contact with other terminals; and video behavior analysis (optional) identifies non-standard behaviors such as not wearing insulated gloves. After work, the system automatically verifies the integrity of the circuit (e.g., by measuring circuit resistance or verifying signals), and only closes the "logic lock" on the circuit after confirming that everything is correct, generating a digital work report containing all safety verification records.

[0082] Electrical connection methods or structures not described in detail in this article are existing technologies.

[0083] While the specific embodiments of the present invention have been described in detail above, the present invention is not limited to the above embodiments. Within the scope of knowledge possessed by those skilled in the art, various changes can be made without departing from the spirit of the present invention, and modifications or variations without creative effort are still within the protection scope of the present invention.

Claims

1. An intelligent operation and maintenance management system for secondary equipment circuits in a substation, characterized in that: It includes the IoT sensing layer, edge computing layer, mobile operation layer, and platform application layer; The IoT sensing layer is deployed in the secondary equipment cabinet of the substation and includes multiple intelligent sensing terminals. The intelligent sensing terminals are integrated inside the traditional wiring terminals. The housing is physically isolated into a power cavity and an intelligent cavity. The power cavity is equipped with a conductive connection mechanism (1) and a contact point temperature sensor (2). The intelligent cavity is equipped with a main control unit (3), a non-invasive current sensor (4), a vibration sensor (5), and a communication unit (6) for real-time monitoring of electrical parameters flowing through the terminals, physical status of connection points, and digital communication. The edge computing layer is deployed locally at the substation and includes at least one edge server and network equipment. The edge server is used to aggregate and process data from the IoT sensing layer and deploys a digital twin engine. This digital twin engine is a software middleware that integrates a real-time data-driven model, a physical rule model, and a geometric model. Its core modules include:

1. Real-time synchronization and mapping engine: Apache Kafka / Apache Pulsar is used as the data pipeline, and Redis is used as the real-time state database. Based on the publish-subscribe pattern and in-memory computing, it realizes zero-latency mapping between physical terminals and virtual models; 2. 3D Geometry Rendering and Lightweight Engine: 3D visualization is achieved through Three.js / Unity / Unreal Engine, and massive numbers of terminals are drawn efficiently using glTF format models and instantiation rendering technology; 3. Physics and Rule Model Engine: Based on graph theory algorithms, a topology connection model is established; combined with Kirchhoff's laws, an electrical quantity propagation model is constructed; and a terminal state logic model is defined through a finite state machine. The mobile operation layer includes a mobile operation terminal and personnel positioning tags. The mobile operation terminal is equipped with an operation guidance application, which can receive operation instructions from the edge computing layer and provide cabinet positioning, terminal highlighting recognition, virtual wiring diagram overlay, and operation step visualization guidance in combination with augmented reality technology. The platform application layer, deployed on the main site or in the cloud, includes a job management module, an intelligent diagnosis module, an asset management module, a knowledge base module, and a job safety control module. The intelligent diagnosis module integrates three core algorithms: an overload and abnormal load identification algorithm based on real-time current values ​​and historical trends; a temperature rise model deviation analysis algorithm based on the square of the current and measured contact point temperature; and a mechanical loosening identification algorithm based on vibration sensor spectral characteristics and a machine learning classifier. The job safety control module is used for dynamically authorizing operation permissions, real-time monitoring of misoperation risks, and early warning. Each layer is connected through a standard communication protocol, forming a complete closed loop from status awareness, data processing, job guidance to intelligent decision-making.

2. The intelligent operation and maintenance management system for secondary equipment circuits in a substation according to claim 1, characterized in that: The configuration of the operation safety management module includes: Before starting work: Verify the operator's identity, the validity of the electronic work ticket, and the work time window; only unlock the software "logic lock" on the smart sensing terminals involved in the work order. During operation: Authorization of the work area is confirmed through personnel positioning; interaction authentication between intelligent tools and terminals prevents accidental contact with non-target terminals; and non-standard operations are identified through video behavior analysis. After the operation: The system automatically verifies the integrity of the circuit, closes the "logic lock" after confirming that there are no errors, and generates a digital operation report containing safety verification records.

3. The intelligent operation and maintenance management system for substation secondary equipment circuits according to claim 1, characterized in that, It also includes the following steps: S1: Status monitoring and data aggregation: Each intelligent sensing terminal continuously monitors status parameters such as current, contact point temperature, and vibration, and uploads them to the edge server through the industrial network; S2: Digital Twin Construction and Update: The digital twin engine of the edge server dynamically binds real-time monitoring data, historical operation and maintenance data and corresponding entities in the 3D virtual model based on the received data, drives the virtual model status update and calculates the terminal health score. S3: Intelligent Diagnosis and Work Order Triggering: The intelligent diagnosis module at the platform application layer, based on multi-source data fusion analysis, identifies potential faults or predicts maintenance needs through the following algorithms, and automatically generates preventative maintenance work orders:

1. Overload and abnormal load identification algorithm: The following algorithm is used to integrate dynamic threshold early warning: ； in, and It is the rolling average and standard deviation of the historical current in the same period, and k is the sensitivity coefficient (taken as 2-3). This threshold can adapt to the periodic changes of the load. The following algorithm is used for cumulative and trend anomaly detection: ； in, This represents the normalized deviation between the current value and the expected value. To allow for small deviations, when If the preset threshold H is exceeded, an abnormal trend alarm will be triggered; Waveform anomaly identification and feature extraction analysis: Calculate the harmonic distortion rate (THD), waveform factor and other features for each power frequency cycle at the edge side; establish a feature baseline under normal operating conditions; when real-time features exceed the baseline range, combine with threshold judgment to identify waveform distortion anomalies.

2. Temperature rise model deviation analysis algorithm: A baseline temperature rise model is established, and based on Joule's law and heat conduction theory, the temperature rise at the contact point is determined under ideal clean and tight conditions. With the square of the current It exhibits a linear relationship and is affected by ambient temperature. Influence: ； in The theoretical thermal resistance of the contact point (determined by the terminal material and design). : Ambient temperature influence coefficient; Deviation analysis between measured values ​​and model values: Real-time data acquisition I, , ; Calculate the deviation index D: ； Analysis logic: D ≈ 0: The state is healthy, consistent with the theoretical model; D > 0 and continues to increase slowly: This indicates an increase in actual thermal resistance, which is an early sign of oxidation, contamination, or minor loosening of the contact surface; the growth slope of D is used to predict the remaining life (RUL). D > 0 and a sudden step: indicates that the fastening screw has suddenly loosened or that an external heat source has affected it; D < 0: This is theoretically rare and indicates a malfunction in the temperature sensor or an abnormal improvement in cooling conditions.

3. Mechanical loosening identification algorithm: Extract time-domain features, frequency-domain features, and bearing fault frequency band energy from vibration data. Use a gradient boosting decision tree or random forest classification model to output the loosening state and severity level. The algorithm flow is as follows: 3.1 Data Preprocessing and Feature Extraction (Completed at the Edge): Sampling: Triaxial vibration sensor, sampling frequency ≥ 2kHz; Feature extraction (data window per 1 second): Temporal features: Root mean square (RMS), peak value, peak-to-peak value, kurtosis; Frequency domain characteristics (after FFT transformation): Within the range of 100Hz-1kHz (avoiding power frequency and its harmonics), calculate the frequency and amplitude of the first N significant spectral peaks, the spectral centroid, and the spectral entropy. Bearing fault characteristic frequency band energy: Calculate the energy ratio of several narrow bands related to the screw loosening frequency; 3.2 Classification Model (Trained on the edge server, registered for inference at the gateway): Algorithm selection: Gradient boosting decision tree or random forest, which are effective for structured features, lightweight, and suitable for edge deployment; Label preparation: Label the feature data using historical data or simulated experiments (setting "tight", "slightly loose", "significantly loose" states); Model output: It not only outputs a binary classification of "normal / loose", but also outputs the probability of looseness or the severity level; S4: Work Navigation and Execution: Operators receive work orders via mobile work terminals. The system guides them to the target cabinet using ultra-wideband positioning, highlights the target terminals using augmented reality technology, and overlays virtual wiring diagrams and operation steps. It also verifies the identity of intelligent tools and terminals to prevent misoperation and automatically records key operation data. S5: Knowledge Accumulation and Model Optimization: Structure and store operation process data, result data, and handling experience into a knowledge base. Extract fault keywords and associate them with standard fault modes through natural language processing. When the system prediction is inconsistent with the expert judgment, update and optimize the diagnostic model through active learning or ensemble learning voting. At the same time, explore the potential relationship between new feature combinations and faults, and generate suspected new rules for expert review.

4. The intelligent operation and maintenance management system for substation secondary equipment circuits according to claim 3, characterized in that: In step S4, the anti-misoperation verification specifically involves: when the smart tool approaches a smart sensing terminal that is not authorized by the work order, the mobile operation terminal issues an audible and visual warning; the disconnection and reconnection operation is only allowed after the tool correctly identifies the target terminal identification code.

5. The intelligent operation and maintenance management system for substation secondary equipment circuits according to claim 3, characterized in that: It also includes a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements steps S1-S5.

Images