Substation three-dimensional visual operation and maintenance management method and system based on digital twinning

Abstract

The invention relates to the technical field of substation operation and maintenance management, in particular to a three-dimensional visual substation operation and maintenance management method and system based on digital twinning. The method comprises the following steps: acquiring real-time working condition data of equipment; acquiring equipment distribution data and space electric field intensity distribution; inputting the equipment distribution data and the historical operation and maintenance data into a physical information neural network model to obtain a health index and a residual service life prediction value; when the real-time working condition data is abnormal or the health index is lower than a preset threshold value, message transmission and node updating are carried out in the substation graph neural network model, a fault source is positioned, and the cascading fault probability is evaluated; and generating an operation and maintenance work order based on the fault source, the health index and the cascading fault probability, and planning a safe operation and maintenance path by using an A-star algorithm. According to the scheme, the limitation that traditional operation and maintenance depend on surface monitoring and experience judgment can be overcome, the depth and the breadth of fault diagnosis are improved, and the safety of field operation is enhanced.

Description

Technical Field

[0001] The present invention relates to the technical field of substation operation and maintenance management, and in particular to a three-dimensional visualized operation and maintenance management method and system for substations based on digital twins. Background Art

[0002] The existing substation operation and maintenance management model primarily relies on regular on-site inspections and preventive maintenance plans. Under this model, O&M personnel must visit the site to obtain equipment status information through manual observation, infrared temperature measurement, and other methods, combining their personal experience to assess equipment health. This approach has numerous drawbacks: First, fixed inspection cycles make it difficult to detect sudden faults and potential defects in real time, resulting in delayed response to faults, which can lead to escalating incidents. Second, the limited data collection method fails to capture multi-dimensional operational status of the equipment, such as electric fields and stress. Consequently, assessments of deeper issues such as equipment aging and damage lack precise physical evidence and rely heavily on qualitative analysis and post-hoc statistics, resulting in limited predictive power. Furthermore, large amounts of monitoring data are dispersed across heterogeneous systems, such as SCADA (Supervisory Control and Data Acquisition) and asset management systems, creating "data silos" that prevent effective integration and analysis, making it difficult to gain a holistic understanding of the station's operational status and accurately predict the risk of cascading failures.

[0003] With the development of digital technology, 3D visualization and some intelligent technologies have begun to be applied to substation operation and maintenance. For example, some systems construct static 3D geometric models of substations to display equipment layout and basic information, making space management more intuitive. At the same time, some fault diagnosis methods based on expert systems or simple data thresholds have also been introduced for preliminary alarm analysis. However, these technologies still have significant limitations. Existing 3D models are mostly "digital prototypes" that lack deep coupling with real-time equipment operating data and inherent physical properties. They are unable to dynamically and accurately simulate the multi-field distribution of electrical, thermal, and mechanical fields under real loads, and therefore cannot serve as accurate input for equipment health assessment and life prediction. Furthermore, fault diagnosis models are often based on fixed rule bases, making them difficult to handle complex and unforeseen failure modes. In particular, they lack effective traceability and deduction capabilities for cascading failures caused by inter-equipment interactions. Regarding operation and maintenance path planning, traditional planning algorithms primarily focus on the shortest distance, ignoring safety risk factors such as the dynamic changes in spatial electric field strength in high-voltage environments, posing a potential threat to the personal safety of on-site operation and maintenance personnel. Therefore, how to build a closed-loop operation and maintenance management system that integrates real-time perception, precise simulation, intelligent prediction, fault tracing and safety planning is a technical problem that needs to be solved urgently in the current substation field. Summary of the Invention

[0004] The purpose of the present invention is to propose a three-dimensional visual operation and maintenance management method and system for substations based on digital twins to solve the problems of inaccurate prediction results, poor fault tracing, and poor on-site operation safety in the prior art. To this end, the present invention provides solutions in the following two aspects.

[0005] In a first aspect, the present invention provides a three-dimensional visual operation and maintenance management method for a substation based on digital twins, comprising: Real-time operating condition data of equipment at monitoring points within the substation is obtained; in a pre-constructed electric-thermal-mechanical multi-physics field finite element model coupled with the equipment's three-dimensional geometric model, the real-time operating condition data is solved as boundary conditions to obtain the equipment's current electric field, temperature, thermal stress distribution data, and the spatial electric field intensity distribution in the operation and maintenance activity area; the equipment distribution data and historical operation and maintenance data are input into a pre-trained physical information neural network model to calculate a health index characterizing the degree of equipment aging and a predicted remaining service life; when the real-time operating condition data is abnormal or the health index is lower than a preset threshold, message transmission and node updates are performed in the pre-constructed substation graph neural network model to locate the root cause of the fault and assess the probability of cascading failures; an operation and maintenance work order is generated based on the root cause of the fault, the health index, and the probability of cascading failures, and a safe operation and maintenance path is planned using the A-star algorithm, where the path cost function of the A-star algorithm is calculated based on the movement distance of the operation and maintenance personnel and the spatial electric field intensity distribution at the path location.

[0006] Preferably, the real-time operating condition data includes: real-time current, voltage, winding temperature and partial discharge amount.

[0007] Preferably, in the pre-constructed electric-thermal-force multi-physics field finite element model coupled with the three-dimensional geometric model of the equipment, the real-time operating condition data is used as the boundary condition to be solved, including: using the real-time voltage of the equipment terminal obtained by the monitoring system as the potential boundary condition; converting the real-time current of the monitored equipment into an internal heat source, and using the ambient temperature as the convection heat transfer boundary condition of the outer surface of the model; using the equipment temperature distribution obtained by the thermal field solution as the heat load, and applying displacement constraints to the fixed parts of the equipment as the force field boundary condition.

[0008] Preferably, the inputting of equipment distribution data and historical operation and maintenance data into a pre-trained physical information neural network model includes: taking the current electric field, temperature, thermal stress distribution data and historical operation and maintenance data of the equipment as model input; outputting a health index representing the degree of equipment aging through the physical information neural network model, and predicting the remaining service life of the equipment based on the time series change trend of the health index.

[0009] Preferably, the message transmission and node update in the pre-built substation graph neural network model include: defining the equipment in the substation as nodes of the graph, and defining the electrical or physical associations between the equipment as edges; when the equipment health index corresponding to the node is lower than a preset first threshold or the real-time operating condition data is abnormal, the node sends a fault message containing its own status information to its adjacent nodes; the adjacent nodes update their own chain failure probability according to the preset association rules and edge weights, and after a preset number of message transmissions, output a set of equipment whose chain failure probability exceeds a preset second threshold.

[0010] Preferably, the feature vector of the node includes equipment type, health index and real-time operating condition data.

[0011] Preferably, the use of the A-star algorithm to plan a safe operation and maintenance path includes: discretizing the operation and maintenance area into grid nodes for path planning; the path cost function of the A-star algorithm comprehensively considers the moving distance and the spatial electric field strength of the path location, wherein the path cost of moving from one node to an adjacent node is positively correlated with the distance between the nodes and the average spatial electric field strength, and the influence of the spatial electric field strength is adjusted by a preset weight coefficient.

[0012] Preferably, it also includes: in the three-dimensional visualization model of the substation, the electric field, temperature, thermal stress distribution data, health index and remaining service life prediction value of the equipment are superimposed, rendered and dynamically displayed.

[0013] Preferably, the operation and maintenance work order includes: information of the equipment to be repaired, a fault description based on fault root location, a recommended operation and maintenance level based on health index assessment, and three-dimensional visual navigation information of the planned safe operation and maintenance path.

[0014] In the second aspect, a three-dimensional visualization operation and maintenance management system for a substation based on digital twins includes: a processor and a memory, the memory stores a computer program, and the processor executes the computer program to implement the above-mentioned three-dimensional visualization operation and maintenance management method for a substation based on digital twins.

[0015] The beneficial effects of the present invention are as follows: by combining real-time operating data with the electric-thermal-mechanical multi-physics field finite element model, the present invention realizes the accurate quantitative analysis of the internal operating status of the equipment, such as the electric field, temperature, thermal stress, etc., and overcomes the limitations of traditional operation and maintenance that rely on surface monitoring and empirical judgment. Based on the precise physical state and the integration of historical data, the physical information neural network can reliably predict the health status and remaining life of the equipment, providing a scientific basis for implementing forward-looking maintenance strategies and optimizing asset management, and effectively reducing the risk of unplanned downtime. In addition, the use of graph neural networks to model the relationship between equipment can deeply locate the root cause of the fault and evaluate the risk of chain failures, thereby improving the depth and breadth of fault diagnosis. By planning the operation and maintenance path based on the spatial electric field strength as the key constraint, not only the work efficiency is improved, but also the personal safety of the operation and maintenance personnel is placed first, significantly enhancing the safety of on-site operations. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 The following is a schematic diagram showing the steps of the three-dimensional visual operation and maintenance management method of the substation based on digital twin in this embodiment; Figure 2 The structural block diagram of the three-dimensional visual operation and maintenance management system of the substation based on digital twin in this embodiment is schematically shown. DETAILED DESCRIPTION

[0017] The technical solutions in the embodiments of the present invention will be described clearly and completely below with reference to the accompanying drawings in the embodiments of the present invention.

[0018] like Figure 1 As shown, the three-dimensional visual operation and maintenance management method of the substation based on digital twin in this embodiment includes the following steps: Step S1: Acquire real-time operating data of equipment at monitoring points within the substation.

[0019] Specifically, by deploying current transformers, voltage transformers, fiber optic temperature sensors, and partial discharge sensors on key equipment such as transformers and circuit breakers, real-time current, voltage, winding temperature, and partial discharge data of the equipment are collected. Using the IEC61850 communication protocol, the data is transmitted in real time via industrial Ethernet to the InfluxDB time series database on the central server for storage and management.

[0020] Step S2: In a pre-built electric-thermal-mechanical multi-physics field finite element model coupled with the three-dimensional geometric model of the equipment, the real-time operating condition data is solved as a boundary condition to obtain the current electric field, temperature, thermal stress distribution data of the equipment and the spatial electric field intensity distribution of the operation and maintenance activity area.

[0021] Specifically, a precise 3D geometric model of equipment, such as transformers, was constructed using SolidWorks software and imported into COMSOL Multiphysics. The real-time load current and voltage were used as excitation sources in the electric field module to calculate the Joule heat loss and dielectric loss within the equipment. This heat loss was then input into the heat transfer module as a heat source, with the ambient temperature set as the convective heat transfer boundary to determine the temperature distribution of each component.

[0022] The real-time voltage at the equipment terminals obtained by the monitoring system is used as the potential boundary condition; the real-time current of the monitored equipment is converted into an internal heat source, and the ambient temperature is used as the convection heat transfer boundary condition on the outer surface of the model; the equipment temperature distribution obtained by the thermal field solution is used as the heat load, and displacement constraints are imposed on the fixed components of the equipment as force field boundary conditions.

[0023] The multiphysics coupling solution simulates the device's real-world operating environment by setting precise boundary conditions. For example, for a 110 kV circuit breaker, the online monitoring system measures a real-time phase voltage of 110.2 kV. This value is directly applied to the corresponding surface of the conductive rod in the model, serving as the potential excitation for the electric field analysis. Simultaneously, the current flowing through the conductor is monitored by a current transformer, measuring 1500 amperes. Based on Joule's law and the conductor's resistivity, the heat generated per unit volume is calculated. This power is then applied to the conductor domain as a heat source. The model's outer shell is set to an ambient temperature of 30 degrees Celsius and a convective heat transfer coefficient of 8 watts per square meter Kelvin based on daily meteorological data to simulate natural convection heat dissipation. After completing the electric and thermal field analyses, the results serve as input for the subsequent force field analysis. The 3D temperature distribution obtained in the previous step might show that the internal contact temperature of the device is as high as 90 degrees Celsius, while the external flange temperature is only 40 degrees Celsius. This uneven temperature field causes thermal expansion and contraction of the material. Temperature distribution data is applied as a thermal load to the entire equipment model. To simulate the actual installation of the equipment within the substation, the base portion of the model connected to the ground foundation is fully constrained, meaning that its displacement in all directions is zero. Stress-strain analysis based on this constraint accurately calculates the thermal stresses caused by temperature differences. For example, stress concentrations as high as 20 MPa can occur at the connection between insulators and metal components.

[0024] In step S3, the equipment distribution data and historical operation and maintenance data are input into a pre-trained physical information neural network model to calculate the health index and remaining service life prediction value that characterize the degree of equipment aging.

[0025] Specifically, the equipment's current electric field, temperature, thermal stress distribution data, and historical operation and maintenance data are used as model inputs. The physical information neural network model outputs a health index that represents the degree of equipment aging, and based on the time series change trend of the health index, the remaining service life of the equipment is predicted.

[0026] A physical-information neural network is used to fuse multi-source data to assess equipment status. The input data is multidimensional. For example, current simulation analysis data for a main transformer bushing might include a maximum electric field strength of 4 kilovolts per millimeter, a maximum operating temperature of 85 degrees Celsius, and a maximum thermal stress of 18 MPa. Furthermore, historical O&M data, such as the equipment's 12-year operating age, the most recent dielectric loss test result of 0.5%, and three fault tripping records within the past five years, are quantified and used as input feature vectors.

[0027] After processing these complex inputs, the pre-trained physical information neural network model outputs a standardized health index ranging from 0 to 1. For example, the current value is 0.72, where 1 represents a pristine condition and 0 represents complete failure. The health index is continuously recorded, forming a time series curve. Assuming that the casing's health index has steadily decreased from 0.85 to 0.72 over the past two years, by applying a long short-term memory network or similar time series prediction algorithm to extrapolate this downward trend, the model can predict that its health index will drop to the preset failure threshold of 0.5 in approximately 4.5 years, concluding that its remaining service life is 4.5 years.

[0028] Step S4: When the real-time operating condition data is abnormal or the health index is lower than a preset threshold, message transmission and node update are performed in the pre-built substation graph neural network model to locate the root cause of the fault and evaluate the probability of cascading failures.

[0029] Specifically, all primary equipment in the substation, such as transformers, circuit breakers, and disconnectors, are treated as nodes in the graph, and the electrical connections between them are used as edges to construct a topological graph of the entire station. The feature vector of each node contains its device type, health index, and real-time operating condition data. When the health index of a transformer node falls below 0.6, the graph convolutional network (GCN) model is triggered. The GCN model aggregates information about neighboring nodes, such as connected circuit breakers and busbars, through multi-layer information transmission, updates node representations, and finally calculates the probability of each device in the graph being the initial fault source of this event through the output layer. It further infers the probability value of the fault being transmitted to upstream or downstream devices, causing a cascading trip.

[0030] Real-time operating data anomalies refer to collected data exceeding the preset normal range, such as voltage fluctuation exceeding ±5%, temperature exceeding 85 degrees Celsius, etc.

[0031] In an optional embodiment, message transmission and node updates are performed in a pre-built substation graph neural network model, including: defining the equipment in the substation as nodes of the graph, and defining the electrical or physical associations between the equipment as edges; when the equipment health index corresponding to the node is lower than a preset first threshold or the real-time operating condition data is abnormal, the node sends a fault message containing its own status information to its adjacent nodes; the adjacent nodes update their own chain failure probabilities according to preset association rules and edge weights, and after a preset number of message transmissions, output a set of equipment whose chain failure probability exceeds a preset second threshold.

[0032] Specifically, the entire substation is abstracted into a graph structure, in which each circuit breaker, transformer, or disconnector is a separate node. For example, if a circuit breaker CB01 is directly connected to the main transformer T01 on the main wiring diagram, an edge is established between the CB01 and T01 nodes. The weight of this edge can be set based on the strength of the connection—for example, the weight of the main circuit connection is set to 0.9, while the weight of the connection to the control circuit is set to 0.4—to reflect the impact of different connections on fault propagation. When the health index of the main transformer T01 drops to 0.5 due to internal overheating, falling below the first threshold of 0.6, the T01 node is activated and sends a fault message to all its neighboring nodes, including CB01, stating that T01 is overheating and has a health index of 0.5. Upon receiving this message, CB01, based on its built-in expert rules, determines that severe transformer overheating increases the risk of circuit breaker misoperation or failure, and increases its own cascading failure probability from the initial 0.05 to 0.3. After several rounds of message transmission and probability updates across the entire site, all devices with a probability of cascading failure exceeding the second threshold of 0.5, such as the set {CB01, DS02}, are listed as a potential high-risk device group for cascading failures.

[0033] In step S5, an operation and maintenance work order is generated based on the root cause of the fault, the health index, and the probability of cascading failures, and a safe operation and maintenance path is planned using the A-star algorithm. The path cost function of the A-star algorithm is calculated based on the distance the operation and maintenance personnel move and the spatial electric field intensity distribution at the path location.

[0034] Specifically, based on the device with the highest probability of being the root cause of the fault, such as the No. 3 main transformer, an electronic operation and maintenance work order containing the device number, fault description, recommended maintenance measures, and cascading failure risk warning is automatically generated. At the same time, on the three-dimensional grid map of the substation, the A-star pathfinding algorithm is started from the station entrance to the location of the No. 3 main transformer. When calculating the cost of each step of movement, the weighted sum of the physical length of the path segment and the average electric field strength value of the spatial position of the path segment is used as the total cost, where the electric field strength data comes from the multi-physics field simulation results. Avoid areas with high electric field strength and plan an operation and maintenance route that takes into account both the shortest distance and the highest safety.

[0035] In an optional embodiment, the use of the A-star algorithm to plan a safe operation and maintenance path includes: discretizing the operation and maintenance area into grid nodes for path planning; the path cost function of the A-star algorithm comprehensively considers the moving distance and the spatial electric field strength at the path location, wherein the path cost of moving from one node to an adjacent node is positively correlated with the distance between the nodes and the average spatial electric field strength, and the influence of the spatial electric field strength is adjusted by a preset weight coefficient.

[0036] Specifically, to achieve safe path planning, the substation's operation and maintenance personnel's activity area is divided into one-meter square grids. The A-star algorithm evaluates the cost of each step when searching for a path from a starting point to a destination. This cost function incorporates safety factors in addition to distance. For example, the cost formula for moving from grid point A to adjacent grid point B is: ; Among them, the weight coefficient is an adjustable parameter, such as setting it to 0.2, which is used to balance the distance and electric field risks. For example, the operation and maintenance personnel need to walk from point A to point B, and there are two channels of similar length to choose from. Channel one is away from high-voltage equipment, and the average spatial electric field strength on its path is 1 kilovolt per meter. Channel two needs to pass under a running knife switch, and the average electric field strength is as high as 10 kilovolts per meter. For one-step movement on channel one, the cost is 0.2×1=0.2. For one-step movement on channel two, the cost is 0.2×10=2. The A-star algorithm will significantly tend to choose the path through channel one, even if the total distance is slightly longer, it can ensure that personnel are always in a safe area with low electric field strength.

[0037] In an optional embodiment, the method further includes: in a three-dimensional visualization model of the substation, superimposing, rendering and dynamically displaying the electric field, temperature, thermal stress distribution data, health index and remaining service life prediction value of the equipment.

[0038] Specifically, users see a 3D digital twin model of the substation that is identical to the actual substation. When users click on a GIS device in the model, the device's internal temperature distribution is immediately rendered as a cloud map, smoothly transitioning from blue, representing low temperatures, to red, representing hotspots. Users can also switch the view from the temperature cloud map to an electric field intensity distribution cloud map, clearly showing the concentration of electric fields at specific locations on high-voltage conductors and insulation. In addition to visualizing physical field data, key evaluation indicators are also dynamically displayed. For example, an information panel pops up next to the 3D model, indicating the selected device's current health index of 0.81 and a predicted remaining service life of 6.2 years. This data is automatically refreshed every few minutes as real-time data and model calculations are continuously processed in the background. Through the integrated visualization platform, operations and maintenance managers can clearly understand the health status and future trends of each device, from a macro to a micro level.

[0039] In an optional embodiment, the operation and maintenance work order includes: equipment information to be repaired, fault description based on fault root cause location, recommended operation and maintenance level based on health index assessment, and three-dimensional visual navigation information of the planned safe operation and maintenance path.

[0040] Specifically, when a device is determined to require maintenance, a detailed electronic work order is automatically generated. The work order begins with the device to be repaired, such as circuit breaker No. 2 on phase B. This is followed by a description of the fault, which provides the root cause based on physical simulation analysis. For example, an abnormal stress of 15 MPa in the operating mechanism linkage caused the closing time to be extended by 20 milliseconds, posing a risk of refusal to operate. The work order also provides clear handling recommendations and safety instructions. Because the circuit breaker's health index has dropped to 0.55, the recommended maintenance level is classified as urgent, requiring maintenance to be completed within 24 hours. A three-dimensional navigation module is embedded in the work order. When maintenance personnel open the work order on-site on a tablet, they can see the optimal safe path from their current location to circuit breaker No. 2 on phase B highlighted in the three-dimensional model. This guides them to avoid high electric field areas near other equipment, enabling safe and accurate on-site operations.

[0041] The present invention also provides a three-dimensional visual operation and maintenance management system for substations based on digital twins, such as Figure 2 As shown, it includes a processor and a memory, the memory stores a computer program, the processor can interact with the memory, can call the computer program (for example, through a bus), and then the processor executes the computer program. When the computer program is executed by the processor, the three-dimensional visualization operation and maintenance management method of the substation based on digital twin of the above embodiment is implemented.

[0042] In the present invention, the aforementioned memory may 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. For example, a computer-readable storage medium may be any suitable magnetic storage medium or magneto-optical storage medium.

[0043] In the description of this specification, “a plurality of” means at least two, for example, two, three or more, etc., unless otherwise clearly defined.

[0044] Although this specification has shown and described several embodiments of the present invention, it is obvious to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will conceive of many modifications, changes and substitutions without departing from the idea and spirit of the present invention.

Claims

1. A three-dimensional visual operation and maintenance management method for substations based on digital twins, characterized by: The following steps are involved: Obtain real-time operating data of equipment at monitoring points within the substation; In a pre-built electric-thermal-mechanical multi-physics finite element model coupled with the equipment's three-dimensional geometric model, the real-time operating condition data is used as boundary conditions to solve the problem, obtaining the equipment's current electric field, temperature, and thermal stress distribution data, as well as the spatial electric field intensity distribution in the operation and maintenance activity area. Input equipment distribution data and historical operation and maintenance data into a pre-trained physical information neural network model to calculate the health index representing the degree of equipment aging and the remaining service life prediction value; When the real-time operating condition data is abnormal or the health index is lower than a preset threshold, message transmission and node updates are performed in the pre-built substation graph neural network model to locate the root cause of the fault and assess the probability of cascading failures; An operation and maintenance work order is generated based on the root cause of the fault, the health index, and the probability of cascading failures. A safe operation and maintenance path is planned using the A-star algorithm. The path cost function of the A-star algorithm is calculated based on the distance the operator moves and the spatial electric field intensity distribution at the path location.

2. The three-dimensional visual operation and maintenance management method of substation based on digital twin according to claim 1 is characterized in that: The real-time operating condition data includes: real-time current, voltage, winding temperature and partial discharge amount.

3. The three-dimensional visual operation and maintenance management method of substation based on digital twin according to claim 1 is characterized in that: Solving the real-time operating condition data as boundary conditions in a pre-built electric-thermal-mechanical multi-physics field finite element model coupled with a three-dimensional geometric model of the equipment includes: The real-time voltage of the equipment terminals obtained by the monitoring system is used as the potential boundary condition; The real-time current of the monitored equipment is converted into an internal heat source, and the ambient temperature is used as the convection heat transfer boundary condition on the outer surface of the model; The equipment temperature distribution obtained by thermal field solution is used as thermal load, and displacement constraints are imposed on the fixed parts of the equipment as force field boundary conditions.

4. The three-dimensional visual operation and maintenance management method of substation based on digital twin according to claim 1 is characterized in that: Inputting equipment distribution data and historical operation and maintenance data into a pre-trained physical information neural network model includes: Using the current electric field, temperature, thermal stress distribution data of the equipment and historical operation and maintenance data as model inputs; The physical information neural network model outputs a health index representing the degree of aging of the equipment, and based on the time series change trend of the health index, the remaining service life of the equipment is predicted.

5. The three-dimensional visual operation and maintenance management method of substation based on digital twin according to claim 1 is characterized in that: The message transmission and node update in the pre-built substation graph neural network model include: The equipment in the substation is defined as the nodes of the graph, and the electrical or physical connections between the equipment are defined as edges; When the device health index corresponding to the node is lower than a preset first threshold or the real-time operating condition data is abnormal, the node sends a fault message containing its own status information to its adjacent nodes; The adjacent nodes update their own cascading failure probabilities according to the preset association rules and edge weights, and after a preset number of message transmissions, output a set of devices whose cascading failure probabilities exceed a preset second threshold.

6. The three-dimensional visual operation and maintenance management method of substation based on digital twin according to claim 5 is characterized in that: The feature vector of the node includes equipment type, health index and real-time working condition data.

7. The three-dimensional visual operation and maintenance management method of substation based on digital twin according to claim 1 is characterized in that: The use of the A-star algorithm to plan a safe operation and maintenance path includes: Discretize the operation and maintenance area into grid nodes for path planning; The path cost function of the A-star algorithm comprehensively considers the movement distance and the spatial electric field strength of the path location. The path cost of moving from one node to an adjacent node is positively correlated with the distance between the nodes and the average spatial electric field strength, and the influence of the spatial electric field strength is adjusted by a preset weight coefficient.

8. The three-dimensional visual operation and maintenance management method of substation based on digital twin according to claim 1 is characterized in that: Also includes: In the three-dimensional visualization model of the substation, the electric field, temperature, thermal stress distribution data, health index and remaining service life prediction value of the equipment are superimposed, rendered and dynamically displayed.

9. The three-dimensional visual operation and maintenance management method of substation based on digital twin according to claim 1 is characterized in that: The operation and maintenance work order includes: equipment information to be repaired, fault description based on fault root location, recommended operation and maintenance level based on health index assessment, and three-dimensional visual navigation information of the planned safe operation and maintenance path.

10. The three-dimensional visual operation and maintenance management system of substation based on digital twin is characterized by: include: A processor and a memory, the memory storing a computer program, and the processor executing the computer program to implement the three-dimensional visualization operation and maintenance management method of a substation based on digital twins as described in any one of claims 1 to 9.

Images