Operation and maintenance guiding method and device based on fusion graph neural network

Abstract

The invention discloses an operation and maintenance guiding method and device based on a fusion graph neural network, relates to the technical field of intelligent operation and maintenance, and discloses the operation and maintenance guiding method based on the fusion graph neural network. Comprising the following steps: constructing a target building digital twinborn model according to operation and maintenance work order information, and determining extreme environment scene data according to the operation and maintenance work order information and the target building digital twinborn model; according to the target building digital twin model and the extreme environment scene data, constructing an initial fusion graph neural network; inputting a current maintenance personnel position, environment data and historical operation and maintenance records into the initial fusion graph neural network to obtain a target operation and maintenance path; and generating a navigation instruction according to the target operation and maintenance path to complete operation and maintenance guidance based on the fusion graph neural network. Accurate planning of an operation and maintenance path and targeted navigation instruction generation in an extreme environment are realized, the operation safety of maintenance personnel can be effectively guaranteed, and the operation and maintenance decision-making efficiency and the path planning rationality are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent operation and maintenance, and in particular to an operation and maintenance guiding method and device based on a fusion graph neural network. BACKGROUND

[0002] Traditional intelligent operation and maintenance schemes are usually based on static models, that is, after a work order is triggered, a pre-established three-dimensional model or two-dimensional drawing is called to provide basic equipment positioning and path information for maintenance personnel. Although this method can guide personnel to work in general environments, when sudden high-temperature leaks, dust blockages, or electromagnetic interference occur on site, the model cannot reflect environmental changes in real time, resulting in the need for maintenance personnel to rely on personal experience for judgment, which not only reduces work efficiency but also significantly increases safety risks. The main drawback of the traditional technical solution is the lack of real-time perception and decision support capabilities for dynamic extreme environments, making it difficult to provide safe and efficient navigation and data assistance for maintenance personnel in conditions of severe environmental changes.

[0003] The above content is only used to assist in understanding the technical solutions of the present application and does not represent the acknowledgement of the above content as prior art. SUMMARY

[0004] The main purpose of the present application is to provide an operation and maintenance guiding method and device based on a fusion graph neural network, aiming to solve the technical problem of how to ensure that maintenance personnel can safely and efficiently complete work navigation and decision-making in the case of dynamic path obstruction and incomplete environmental information in extreme environments.

[0005] To achieve the above-mentioned purpose, the present application provides an operation and maintenance guiding method based on a fusion graph neural network, which comprises: constructing a target building digital twin model according to operation and maintenance order information and determining extreme environment scene data according to the operation and maintenance order information and the target building digital twin model; constructing an initial fusion graph neural network according to the target building digital twin model and the extreme environment scene data; inputting the current maintenance personnel position, environmental data, and historical operation and maintenance records into the initial fusion graph neural network to obtain a target operation and maintenance path; generating navigation instructions according to the target operation and maintenance path to complete operation and maintenance guiding based on a fusion graph neural network.

[0006] In addition, to achieve the above-mentioned purpose, the present application also provides an operation and maintenance guiding device based on a fusion graph neural network, which comprises: a data determination module for constructing a target building digital twin model according to operation and maintenance order information and determining extreme environment scene data according to the operation and maintenance order information and the target building digital twin model; a network construction module, configured to construct an initial fusion graph neural network according to the target building digital twin model and the extreme environment scenario data; a path generation module, configured to input the current maintenance personnel position, environment data and historical operation and maintenance records into the initial fusion graph neural network to obtain a target operation and maintenance path; an operation and maintenance guidance module, configured to generate a navigation instruction according to the target operation and maintenance path to complete the operation and maintenance guidance based on the fusion graph neural network.

[0007] In addition, to achieve the above-mentioned purpose, the present application also provides an operation and maintenance guidance device based on a fusion graph neural network, which comprises a memory, a processor and a computer program stored in the memory and executable on the processor, and the computer program is configured to implement the steps of the operation and maintenance guidance method based on the fusion graph neural network as described above.

[0008] In addition, to achieve the above-mentioned purpose, the present application also provides a storage medium, which is a computer readable storage medium, and a computer program is stored on the storage medium, and the computer program is executed by a processor to implement the steps of the operation and maintenance guidance method based on the fusion graph neural network as described above.

[0009] In addition, to achieve the above-mentioned purpose, the present application also provides a computer program product, which comprises a computer program, and the computer program is executed by a processor to implement the steps of the operation and maintenance guidance method based on the fusion graph neural network as described above.

[0010] The one or more technical solutions provided by the present application have at least the following technical effects: Due to the use of the technical means of constructing a target building digital twin model according to operation and maintenance order information and determining extreme environment scenario data, constructing an initial fusion graph neural network combining the model and the scenario data, inputting the current maintenance personnel position, environment data and historical operation and maintenance records into the network to obtain a target operation and maintenance path, and finally generating a navigation instruction based on the path, the problem of lack of real-time safety protection and accurate data support for operation and maintenance personnel in the existing intelligent operation and maintenance system under extreme environment, resulting in difficult route finding, slow decision-making and high safety risk for maintenance personnel is solved. Compared with the existing technology which only relies on static building information model geometric data or two-dimensional drawings and lacks dynamic environment adaptation and intelligent path planning, the present application realizes accurate mapping of extreme environment scenarios through a digital twin model, intelligent planning of paths through a fusion graph neural network integrating multi-dimensional dynamic data, and accurate landing of operation and maintenance guidance through a navigation instruction, finally realizing real-time protection of operation and maintenance personnel safety under extreme environment, dynamic adaptation of operation and maintenance path planning and efficient and accurate operation and maintenance guidance, and reducing the dependence on personal experience of maintenance personnel. BRIEF DESCRIPTION OF DRAWINGS

[0011] The accompanying drawings, which are incorporated herein and constitute part of the specification, illustrate embodiments consistent with the application and, together with the description, function to explain the principles of the application.

[0012] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the accompanying drawings required by the embodiments or prior art description will be briefly introduced. Obviously, those skilled in the art can obtain other drawings according to these drawings without creative effort.

[0013] Figure 1 A flowchart provided by the present application for the operation and maintenance guidance method based on the fusion graph neural network embodiment one; Figure 2 A flowchart provided by the present application for the operation and maintenance guidance method based on the fusion graph neural network embodiment two; Figure 3 A brief flowchart of the operation and maintenance guidance method based on the fusion graph neural network provided by the present application embodiment two; Figure 4 A module structure diagram of the operation and maintenance guidance device based on the fusion graph neural network of the present application embodiment; Figure 5 A device structure diagram of the hardware running environment involved in the operation and maintenance guidance method based on the fusion graph neural network in the present application embodiment.

[0014] The purpose of the present application, functional characteristics and advantages will be further explained with reference to the embodiments and the accompanying drawings. DETAILED DESCRIPTION

[0015] It should be understood that the specific embodiments described herein are only used to explain the technical solutions of the present application and not to limit the present application.

[0016] In order to better understand the technical solutions of the present application, the following will be described in detail in combination with the drawings in the specification and specific embodiments.

[0017] The main solution of the present application embodiment is: constructing a target building digital twin model according to operation and maintenance work order information and determining extreme environment scenario data according to the operation and maintenance work order information and the target building digital twin model; constructing an initial fusion graph neural network according to the target building digital twin model and the extreme environment scenario data; inputting the current maintenance personnel position, environment data and historical operation and maintenance records into the initial fusion graph neural network to obtain a target operation and maintenance path; generating navigation instructions according to the target operation and maintenance path to complete the operation and maintenance guidance based on the fusion graph neural network.

[0018] In this embodiment, for convenience of description, the following describes an identification of the operation and maintenance guidance device based on the fusion graph neural network as an execution subject.

[0019] Since the prior art does not guarantee that the operation and maintenance personnel can safely and efficiently complete the operation navigation and decision-making in the case of dynamic path obstruction and incomplete environmental information in an extreme environment, the present application provides a solution. Since the target building digital twin model is constructed according to the operation order information and the extreme environment scenario data is determined, the initial fusion graph neural network is constructed in combination with the model and the scenario data, the current maintenance personnel position, environmental data and historical operation and maintenance records are input into the network to obtain the target operation and maintenance path, and finally the navigation instruction is generated based on the path, the technical means solves the problem that the existing intelligent operation and maintenance system lacks real-time safety protection and accurate data support for operation and maintenance personnel in an extreme environment, resulting in difficult wayfinding, slow decision-making and high safety risk for maintenance personnel. Compared with the prior art which only relies on static building information model geometric data or two-dimensional drawings and lacks dynamic environment adaptation and intelligent path planning, the present application realizes accurate mapping of the extreme environment scenario through the digital twin model, realizes intelligent planning of the path through the fusion graph neural network, realizes accurate landing of the operation and maintenance guidance through the navigation instruction, and finally realizes real-time protection of operation and maintenance personnel safety, dynamic adaptation of operation and maintenance path planning and efficient and accurate operation and maintenance guidance in an extreme environment, and reduces the dependence on personal experience of maintenance personnel.

[0020] It should be noted that the execution subject of the present embodiment can be a computing service device with data processing, network communication and program running functions, such as a tablet computer, a personal computer, a mobile phone, etc., or an electronic device capable of realizing the above functions, an operation and maintenance guidance device based on a fusion graph neural network, etc. The operation and maintenance guidance device based on a fusion graph neural network is taken as an example to illustrate the present embodiment and the following embodiments.

[0021] Based on this, the present application provides an operation and maintenance guidance method based on a fusion graph neural network, which is described below with reference to Figure 1 , Figure 1 The present application provides a flowchart of the first embodiment of the operation and maintenance guidance method based on a fusion graph neural network.

[0022] In the present embodiment, the operation and maintenance guidance method based on a fusion graph neural network includes steps S10-S40: Step S10, constructing a target building digital twin model according to operation order information and determining extreme environment scenario data according to the operation order information and the target building digital twin model; It should be noted that the operation order information is a data set recording the key content of the operation and maintenance task, usually containing information such as fault device identification, fault type, operation and maintenance area range, etc. These information is the basis for starting the operation and maintenance process and determining the operation and maintenance target.

[0023] In addition, the target building digital twin model is a digital model built based on the actual structure and equipment information of the building. It can map the physical spatial structure of the building, the three-dimensional coordinates of the equipment, the equipment relationship, etc., providing visualization and data support for subsequent operation and maintenance analysis.

[0024] Furthermore, extreme environment scenario data describes the characteristics of extreme environments within the operation and maintenance area, including the deployment locations of environmental sensors, the distribution of risk areas, and environmental parameter thresholds. This data is used to reflect the specific conditions of extreme environments such as high temperature and high dust.

[0025] Understandably, the faulty equipment identifier, fault type, and maintenance area range are extracted from the maintenance work order information. Based on this extracted information, the sub-model corresponding to the maintenance area range in the building's overall digital twin model is retrieved to supplement the detailed 3D model data of the faulty equipment and the associated equipment model data, thereby constructing the target building's digital twin model. Based on the target building's digital twin model, the specific location area of ​​the faulty equipment in the model is located, i.e., the model location area.

[0026] Furthermore, by querying historical and real-time environmental data of the model's location area, the extreme environmental types of the area are determined, such as high temperature, high dust, strong electromagnetic fields, and low temperature. Combining the spatial structure data of the target building's digital twin model with the determined extreme environmental types, the distribution patterns of environmental parameters and the locations of risk points are analyzed. Historical and real-time data of environmental sensors in the area are retrieved to determine the safety and risk thresholds of environmental parameters. Finally, extreme environmental scenario data containing sensor locations, risk area coordinates, and environmental parameter thresholds are obtained.

[0027] In one feasible implementation, step S10 may include steps S11 to S15: Step S11: Extract the faulty device identifier, fault type, and maintenance area from the maintenance work order information; It should be noted that the faulty equipment identifier is a code or identification information used to uniquely identify the faulty equipment. This identifier can accurately locate the specific equipment that needs to be repaired, avoiding confusion between maintenance tasks for different equipment.

[0028] Furthermore, the fault type is a classification description of the problem that occurs in the faulty equipment, such as electrical faults, mechanical faults, etc. Different fault types correspond to different maintenance strategies and data requirements, providing direction for subsequent operation and maintenance preparation.

[0029] It can be understood that the operation and maintenance area range is a space range that needs to carry out operation and maintenance work around the fault equipment, which is usually determined as a certain radius of area with the fault equipment as the center to clearly define the spatial boundary of operation and maintenance work; the three types of key information of fault equipment identification, fault type and operation and maintenance area range are extracted one by one from the operation and maintenance order information, which provides core input data for subsequent construction of target building digital twin model.

[0030] Step S12, constructing a target building digital twin model according to the fault equipment identification, the fault type and the operation and maintenance area range; It should be noted that the target building digital twin model contains three-dimensional coordinates, spatial position relationship and building structure and other digital information of the fault equipment, which are the basis for positioning the model position area of the fault equipment.

[0031] It can be understood that the operation and maintenance area range is used to filter out the sub-model related to the operation and maintenance task from the building global digital twin model, avoiding loading global redundant data and improving the model construction efficiency; according to the extracted fault equipment identification, fault type and operation and maintenance area range, the sub-model corresponding to the operation and maintenance area range in the building global digital twin model is first called.

[0032] Further, based on the fault equipment identification, the basic model of the fault equipment is located in the sub-model, and then according to the fault type, the detailed three-dimensional model data of the fault equipment such as internal component structure, key parameter annotation and the like are supplemented, and the associated equipment model data related to the fault equipment is also supplemented, so as to finally construct the target building digital twin model.

[0033] Combining the operation and maintenance core information to construct the target building digital twin model can focus on the specific area and equipment required for operation and maintenance, ensure that the model contains complete spatial and equipment association information and no redundant data interference, and provide accurate digital carrier for subsequent positioning of fault equipment position and analysis of extreme environment.

[0034] Step S13, locating the model position area of the fault equipment of the operation and maintenance order information according to the target building digital twin model; It can be understood that in the target building digital twin model, the corresponding fault equipment three-dimensional model is found according to the fault equipment identification; taking the three-dimensional model as the core, combining the spatial range required for maintenance operation in the operation and maintenance scene, the area containing the fault equipment and the surrounding associated space is determined, which is the model position area of the fault equipment.

[0035] By accurately positioning the model position area of the fault equipment through the target building digital twin model, the operation and maintenance analysis range can be focused on the space directly related to the fault equipment, avoiding invalid analysis of irrelevant areas, and providing clear spatial range for subsequent accurate determination of extreme environment type and analysis of environment parameters.

[0036] Step S14, determine the extreme environment type of the model location area; It should be noted that the extreme environment type represents the type of environmental parameter that needs to be analyzed, such as high temperature environment for temperature parameter, high dust environment for dust concentration parameter, to ensure that the analysis focuses on key risk factors.

[0037] It can be understood that the extreme environment scenario data is a data set describing the extreme environment characteristics of the model location area, including sensor location, risk area coordinates, environment parameter threshold, etc., which is used to support the subsequent construction of the initial fusion graph neural network and the planning of the operation and maintenance path; combined with the spatial structure data of the target building digital twin model, the distribution law of the environment parameter corresponding to the extreme environment type in the model location area is analyzed, such as the trend of temperature spreading outward along the equipment heat dissipation port in high temperature environment, and the corner position where dust is easy to accumulate in high dust environment.

[0038] Step S15, according to the target building digital twin model and the extreme environment type, analyzing the environment parameter to obtain the extreme environment scenario data.

[0039] It can be understood that the extreme environment scenario data is a data set describing the extreme environment characteristics of the model location area, including sensor location, risk area coordinates, environment parameter threshold, etc., which is used to support the subsequent construction of the initial fusion graph neural network and the planning of the operation and maintenance path; combined with the spatial structure data of the target building digital twin model, the distribution law of the environment parameter corresponding to the extreme environment type in the model location area is analyzed, such as the trend of temperature spreading outward along the equipment heat dissipation port in high temperature environment, and the corner position where dust is easy to accumulate in high dust environment.

[0040] Further, according to the distribution law, the risk point position in the region is determined, such as high temperature risk point, dust accumulation risk point; at the same time, the historical data and real-time data of the environmental sensor in the model location area are called, combined with the operation and maintenance safety standard, to determine the safety threshold and risk threshold of each environment parameter; integrate sensor location, risk area coordinates, environment parameter threshold, etc. Information, finally get the extreme environment scenario data.

[0041] Based on the target building digital twin model and the extreme environment type, the environment parameter is analyzed, which can accurately obtain the extreme environment characteristic data of the model location area, and provide comprehensive and practical environment data support for defining nodes and edges when constructing the initial fusion graph neural network, and evaluating risk cost when planning the target operation and maintenance path.

[0042] Step S20, according to the target building digital twin model and the extreme environment scenario data, constructing an initial fusion graph neural network; It should be noted that the initial fusion graph neural network is the initial model of fusion graph neural network technology. It integrates multi-dimensional data such as buildings, environment, and equipment in the form of a graph structure, and has the ability to extract features and make decisions from multi-source data. It is the core model for generating target operation and maintenance paths.

[0043] Additionally, the geometric data of the building information model is the data describing the spatial form of the building and equipment in the digital twin model of the target building. It includes the three-dimensional coordinates of the equipment, the dimensions of the building passages, the location of the walls, etc., and is used to build the spatial foundation for the initial fusion graph neural network.

[0044] Furthermore, non-geometric data in Building Information Modeling (BIM) refers to the data describing the non-spatial attributes of buildings and equipment in the target building's digital twin model. This includes information such as equipment materials, service life, and maintenance cycles, and is used to supplement the attribute information of the initial fusion graph neural network.

[0045] Understandably, device association data describes the connection relationships between devices, including power supply relationships, data flow transmission relationships, etc., and is used to reflect the logical relationships between device nodes in the initial fusion graph neural network; environmental sensor deployment location is information recorded in the extreme environment scenario data about the installation location of environmental sensors, usually presented in three-dimensional coordinates, and is used to determine the location of sensor nodes in the initial fusion graph neural network.

[0046] Furthermore, the risk area distribution data is information describing the range of high-risk areas in extreme environment scenario data, including the boundary coordinates and risk levels of areas such as high temperature areas and high dust areas, which is used to construct virtual risk nodes in the initial fusion graph neural network. The initial fusion graph node is the basic building block of the initial fusion graph neural network, including maintenance personnel starting candidate nodes, target equipment nodes, sensor nodes, and virtual risk nodes. Each node carries the feature information of the corresponding entity or region.

[0047] It is understandable that the initial fusion graph edges are the elements connecting the nodes of the initial fusion graph, including physical access edges and device logical association edges. Physical access edges are used to represent the passable physical paths, and device logical association edges are used to mark the direction of energy flow or data flow between devices. Geometric data of building information model, non-geometric data of building information model and device association data are extracted from the target building digital twin model. Environmental sensor deployment location and risk area distribution data are extracted from extreme environment scenario data.

[0048] Further, according to the extracted building information model geometric data, building information model non-geometric data, device association data, environmental sensor deployment location and risk area distribution data, define initial fusion graph nodes and initial fusion graph edges; set static attributes and dynamic attributes for physical passing edges, the static attributes include path length and basic passing cost, the dynamic attributes are used to associate environmental data calculation results, and device logical association edges are marked with energy flow or data flow direction between devices; based on the defined initial fusion graph nodes and initial fusion graph edges, the initial fusion graph neural network is constructed.

[0049] In a feasible implementation, step S20 can include steps S21-S24: Step S21, according to the target building digital twin model, determine building information model geometric data, building information model non-geometric data and device association data; It should be noted that the building information model geometric data is data describing the spatial form of the building and the device, including device three-dimensional coordinates, building passage size, wall position and the like, which are used to define the spatial position of the initial fusion graph node and the physical length of the initial fusion graph edge in the subsequent.

[0050] Further, the building information model non-geometric data is data describing the non-spatial attributes of the building and the device, including device material, service life, maintenance cycle and the like, which are used to supplement the attribute information of the initial fusion graph node and improve the completeness of the node features.

[0051] It can be understood that the device association data is data describing the connection relationship between devices, including power supply relationship between devices, data flow transmission relationship and the like, which are used to define the device logical association edge in the initial fusion graph in the subsequent; the building information model geometric data, the building information model non-geometric data and the device association data are filtered and extracted from the target building digital twin model, providing data support for subsequent definition of initial fusion graph nodes and edges.

[0052] Step S22, according to the extreme environment scenario data, obtain environmental sensor deployment location and risk area distribution data; It should be noted that the environmental sensor deployment location is information recording the installation location of the environmental sensor in the physical space, usually in the form of three-dimensional coordinates, which are used to define the position of the sensor node in the initial fusion graph in the subsequent, ensuring that the sensor node is consistent with the actual sensor spatial position.

[0053] Further, the risk area distribution data is information describing the range of high-risk areas, including boundary coordinates and risk levels of high-temperature areas, high-dust areas and the like, which are used to define the position and risk attributes of the virtual risk node in the initial fusion graph in the subsequent, reflecting the risk distribution of the extreme environment.

[0054] It can be understood that the environmental sensor deployment position and risk area distribution data are filtered and extracted from the extreme environment scene data, and redundant data irrelevant to the initial fusion graph node and edge definition are excluded, so as to ensure that the extracted data can directly support the construction of the subsequent initial fusion graph structure.

[0055] Focusing on extracting key environmental data from extreme environment scene data can ensure that the initial fusion graph neural network constructed subsequently contains risk and perception information of extreme environment, so that the model has adaptation ability to extreme environment, and provides environmental data support for subsequent generation of safe target operation path.

[0056] In step S23, the initial fusion graph node and the initial fusion graph edge are defined according to the building information model geometry data, the building information model non-geometry data, the equipment correlation data, the environmental sensor deployment position and the risk area distribution data, wherein the initial fusion graph node includes a maintenance personnel starting candidate node, a target equipment node, a sensor node and a virtual risk node, and the initial fusion graph edge includes a physical passing edge and an equipment logical correlation edge. It should be noted that the initial fusion graph node is a basic constituting unit of the initial fusion graph neural network, the maintenance personnel starting candidate node is a node corresponding to a possible initial position of a maintenance personnel, the target equipment node is a node corresponding to a fault equipment, the sensor node is a node corresponding to an environmental sensor, and the virtual risk node is a node corresponding to a high-risk area.

[0057] Further, the initial fusion graph edge is an element connecting the initial fusion graph node, the physical passing edge is used to represent a passable physical path, and is connected to nodes related to passing, such as the maintenance personnel starting candidate node and the target equipment node, and the equipment logical correlation edge is used to mark the direction of energy flow or data flow between devices, and is connected to device nodes having logical correlation.

[0058] It can be understood that the spatial coordinates of each node and the path length of the physical passing edge are determined according to the building information model geometry data, and the nodes are added with attributes such as material and service life in combination with the building information model non-geometry data; the equipment logical correlation edge and the logical direction of the edge are defined according to the equipment correlation data; the coordinates of the sensor node are defined according to the environmental sensor deployment position, and the coordinates and risk level of the virtual risk node are defined according to the risk area distribution data, so as to finally complete the definition of the initial fusion graph node and the initial fusion graph edge.

[0059] In step S24, the initial fusion graph neural network is constructed according to the initial fusion graph node and the initial fusion graph edge.

[0060] It can be understood that the attribute information of the initial fusion graph node such as the space coordinates, the material, the risk level is converted into a node feature vector to form a node feature matrix; the attribute information of the initial fusion graph edge such as the path length, the logical direction is converted into an edge feature vector to form an edge feature matrix; the node feature matrix and the edge feature matrix are taken as initial inputs, a network architecture including a graph attention module and a decision module is built, parameters of each module are configured, and the initial fusion graph neural network is constructed.

[0061] In step S30, the current maintenance personnel position, the environment data and the historical operation and maintenance record are input into the initial fusion graph neural network to obtain a target operation and maintenance path. It should be noted that the current maintenance personnel position is real-time spatial position information of the maintenance personnel in the operation and maintenance site, which is usually obtained through ultra wide band positioning technology (UWB), Bluetooth low energy technology (BLE) or 5th generation mobile communication technology (5G), and is used to determine the current node in the initial fusion graph neural network.

[0062] In addition, the environment data is data reflecting the real-time environment condition of the operation and maintenance site, including temperature, dust concentration, harmful gas concentration and other parameters, which are collected by environment sensors and used to evaluate the environmental risk of the path.

[0063] Further, the historical operation and maintenance record is data recording information related to past operation and maintenance tasks, including path congestion frequency, fault handling time and other contents, which are used to provide historical experience reference for current path planning.

[0064] It can be understood that the target operation and maintenance path is the optimal path from the current position of the maintenance personnel to the target device node output by the initial fusion graph neural network, which comprehensively considers the time cost and the environmental risk cost and can ensure the safe and efficient arrival of the maintenance personnel at the target device; the current maintenance personnel position is obtained through positioning technology and is mapped to the current node in the initial fusion graph neural network; the environment data collected by the environment sensors are collected, and the temperature, dust concentration, harmful gas concentration and other data are respectively associated with the corresponding sensor nodes in the initial fusion graph neural network.

[0065] Further, the historical operation and maintenance record is data recording information related to past operation and maintenance tasks, including path congestion frequency, fault handling time and other contents, which are used to provide historical experience reference for current path planning.

[0066] It can be understood that the comprehensive cost of each candidate path is calculated, the comprehensive cost includes the time cost based on the path length and the risk cost based on the environmental data, and the candidate path with the lowest comprehensive cost is selected as the target operation and maintenance path.

[0067] In step S40, navigation instructions are generated according to the target operation and maintenance path to complete the operation and maintenance guidance based on the fusion graph neural network.

[0068] It should be noted that the navigation instructions are instruction information for guiding the maintenance personnel to carry out operation and maintenance work according to the target operation and maintenance path, including first perspective path guidance instructions and data pop-up window display instructions. These instructions are presented to the maintenance personnel through a digital twin mobile terminal.

[0069] In addition, the digital twin mobile terminal is a terminal device used by the maintenance personnel in the operation and maintenance site, such as an industrial tablet or an augmented reality (AR) glasses, which is used to receive the target operation and maintenance path, display the navigation instructions, and feed back the operation and maintenance site information.

[0070] Further, the operation and maintenance guidance based on the fusion graph neural network is to guide the maintenance personnel throughout the operation and maintenance process relying on the target operation and maintenance path and the navigation instructions generated by the fusion graph neural network, aiming to ensure the safety and efficiency of the operation and maintenance process.

[0071] It can be understood that the target operation and maintenance path output by the initial fusion graph neural network is obtained; the navigation instructions adapted to the digital twin mobile terminal are generated according to the target operation and maintenance path, the first perspective path guidance instructions in the navigation instructions are used to prompt the maintenance personnel to turn, go straight or avoid actions, and the data pop-up window display instructions are used to control the digital twin mobile terminal to display the operation and maintenance required data in layers.

[0072] Further, the generated navigation instructions are sent to the digital twin mobile terminal, the maintenance personnel receive and view the navigation instructions through the digital twin mobile terminal, and go to the target device to carry out operation and maintenance work according to the instructions, thereby completing the operation and maintenance guidance based on the fusion graph neural network.

[0073] The navigation instructions generated based on the target operation and maintenance path can intuitively provide path guidance and data support for the maintenance personnel, avoid the maintenance personnel relying on personal experience to explore in extreme environments, reduce safety risks, and at the same time improve the efficiency and accuracy of operation and maintenance work, and ensure the practicality and reliability of the operation and maintenance guidance.

[0074] In a feasible implementation, step S40 can include steps S41-S43: In step S41, an operation and maintenance perspective is generated according to the target operation and maintenance path, and the operation and maintenance perspective is sent to the digital twin mobile terminal, so that the digital twin mobile terminal feeds back the operation and maintenance site information according to the target operation and maintenance path. It should be noted that the operation and maintenance perspective is a visual display angle generated according to the target operation and maintenance path, which is usually the first perspective, can simulate the visual effect when the maintenance personnel walk along the path, and let the maintenance personnel directly see the path direction and the surrounding environment.

[0075] Further, the digital twin mobile terminal is a terminal device used by maintenance personnel in the operation and maintenance site, such as an industrial tablet or an augmented reality glasses, which has the functions of receiving perspective data, collecting site information, and feeding back.

[0076] It can be understood that the operation and maintenance site information is the data collected and fed back by the digital twin mobile terminal according to the target operation and maintenance path, including the position deviation data of the maintenance personnel and the target operation and maintenance path, real-time collection data of the site environment parameters, etc.; the perspective data generated based on the target operation and maintenance path is sent to the digital twin mobile terminal, after receiving, through the positioning module and the environment collection module of the digital twin mobile terminal, the deviation of the current position of the maintenance personnel and the path, the real-time temperature of the site, etc. Operation and maintenance site information, and feed these information back to the background.

[0077] Step S42, determining the target operation and maintenance data package according to the operation and maintenance site information; It should be noted that the target operation and maintenance data package is a targeted data set arranged according to the operation and maintenance site information and the operation and maintenance demand, which contains device basic information, fault related data, risk response prompt, etc., and can provide accurate on-site decision support for maintenance personnel.

[0078] Further, the determination of the target operation and maintenance data package needs to be combined with the specific situation of the operation and maintenance site information. If the position deviation exceeds the preset value, the path correction data is supplemented, and if the environmental parameter exceeds the threshold value, the risk warning data is supplemented, while the device ID, core parameter and other basic data are retained.

[0079] It can be understood that the operation and maintenance site information fed back by the digital twin mobile terminal is analyzed first to see if the deviation of the position of the maintenance personnel and the target operation and maintenance path is within a reasonable range. If the deviation exceeds 1 meter, path correction prompt data is added. Then analyze whether the site environmental parameters exceed the safety threshold. If the temperature is too high, add high-temperature protection and operation precautions. Finally, integrate the device ID, current core parameter, and historical maintenance records related to the current fault to form the target operation and maintenance data package.

[0080] According to the real-time operation and maintenance site information, the target operation and maintenance data package is determined, which can ensure that the data package not only meets the actual needs of the site, but also has no redundant data interference, so that the maintenance personnel can quickly obtain key support information, avoid decision delay caused by complex data, and improve the operation and maintenance efficiency.

[0081] Step S43, generating navigation instructions according to the target operation and maintenance data package to complete the operation and maintenance guidance based on the fusion graph neural network.

[0082] It can be understood that, according to the path correction data in the target operation data packet and the target operation path, the first perspective path guiding instruction such as "turn left 5 meters ahead of the path" and "quickly pass through the high-temperature area" is generated; according to the device ID, core parameter and historical maintenance record in the data packet, a hierarchical pop-up window display instruction is generated to control the digital twin mobile terminal to display the device ID and core parameter on the first layer, the historical maintenance record on the second layer, and the risk warning data on the third layer; the two types of instructions are integrated into complete navigation instructions and sent to the digital twin mobile terminal, and the maintenance personnel proceed and operate according to the instructions, and finally complete the operation guidance based on the fusion graph neural network.

[0083] Further, the operation guidance based on the fusion graph neural network is to generate the target operation path and the navigation instruction relying on the initial fusion graph neural network, to guide the maintenance personnel from the start to the completion of the operation, and aims to ensure safety and efficiency.

[0084] The targeted navigation instruction generated based on the target operation data packet can provide integrated guidance of path guidance and data support for the maintenance personnel, avoid the maintenance personnel from relying on personal experience to explore in extreme environments, reduce safety risks, and at the same time ensure that the operation operation accurately interfaces with the fault demand and improves the operation task completion quality.

[0085] Based on the first embodiment of the present application, in the second embodiment of the present application, the same or similar contents as the above embodiment one can refer to the above introduction, and the following will not be repeated. On this basis, please refer to Figure 2 , the step S30 of the operation guidance method based on the fusion graph neural network comprises steps S31-S34: Step S31, inputting the current maintenance personnel position, environment data and historical operation record into the initial fusion graph neural network to determine the current node and the physical passing edge of the current node. It should be noted that the current maintenance personnel position is the real-time spatial position information of the maintenance personnel in the operation site, which is usually obtained through ultra wide band positioning technology (Ultra Wide Band, UWB), Bluetooth low power technology (Bluetooth Low Energy, BLE) or 5th generation mobile communication technology (5th Generation Mobile Communication Technology, 5G), and is used to locate the corresponding node in the initial fusion graph neural network.

[0086] In addition, the environment data is data reflecting the real-time environment of the operation site, including temperature, dust concentration, harmful gas concentration and other parameters, which are collected by environmental sensors and used to associate sensor nodes and physical passing edges in the initial fusion graph neural network.

[0087] Further, the historical operation record is data recording past operation task related information, containing path congestion frequency, fault processing time consumption and the like, and is used for supplementing historical attribute information for the physical passing edge in the initial fusion graph neural network.

[0088] Further, the physical passing edge is a passable path edge connecting the current node and the surrounding node in the initial fusion graph neural network, and the average length and the distribution density thereof determine the rationality of the preset hop range, ensuring that the effective path within the hop range can cover the safe field of view of the maintenance personnel.

[0089] It can be understood that the current node is a node in the initial fusion graph neural network matched with the current maintenance personnel position, and the physical passing edge is an edge representing a passable physical path connecting the current node and the surrounding node in the initial fusion graph neural network; the current maintenance personnel position, the environmental data and the historical operation record are input into the initial fusion graph neural network, the network first maps the current maintenance personnel position to the corresponding current node, then associates the environmental data to the sensor nodes around the current node and the connected physical passing edges, and associates the historical operation record to the corresponding physical passing edges, and finally determines the current node and the physical passing edges of the current node.

[0090] In a feasible implementation manner, the step S31 can include steps S311-S314: In step S311, the current temperature, dust concentration and harmful gas concentration are determined according to the environmental data. It should be noted that the current temperature is the real-time air or device surrounding temperature value of the operation site, which is used to judge whether the environment is in an extreme state of high temperature or low temperature, and directly affects the operation safety of the maintenance personnel and the running state of the equipment.

[0091] Further, the dust concentration is the content value of the suspended dust in the air of the operation site, and high dust concentration will affect the visibility and the health of the personnel, which is an important index for evaluating the path passing risk; the harmful gas concentration is the content value of the gas harmful to the human body or the equipment in the air of the operation site, and exceeding the safety threshold will directly threaten the safety of the personnel and needs to be monitored.

[0092] It can be understood that the current temperature, dust concentration and harmful gas concentration are screened and extracted from the collected environmental data, and the redundant information irrelevant to the risk assessment in the environmental data is removed, so as to provide accurate environmental risk data for subsequent determination of the current node and the physical passing edge.

[0093] In step S312, the current maintenance personnel position, the current temperature, the dust concentration and the harmful gas concentration are input into the initial fusion graph neural network to determine the current node. It should be noted that the current maintenance personnel position is the real-time spatial coordinates of the maintenance personnel in the operation and maintenance site, which is used to find the corresponding spatial position anchor point in the initial fusion graph neural network.

[0094] In addition, the current temperature, dust concentration and harmful gas concentration are parameters reflecting the real-time risk of the operation and maintenance site. These parameters can assist the initial fusion graph neural network in judging the node risk attribute of the environment around the maintenance personnel, and ensure that the current node matches the actual environment state.

[0095] Further, the current node is the node in the initial fusion graph neural network that matches the real-time position of the maintenance personnel and the risk of the surrounding environment. This node not only contains spatial coordinate information, but also associates the risk characteristics of the surrounding environment, and is the starting point of subsequent path planning.

[0096] It can be understood that the spatial coordinates of the current maintenance personnel position are input into the initial fusion graph neural network together with the risk parameters such as current temperature, dust concentration and harmful gas concentration; the network matches the possible node range through the spatial coordinates, and then compares the risk parameters with the preset environmental risk attributes of each node to finally determine the current node that completely matches the actual position of the maintenance personnel and the surrounding environment.

[0097] Step S313, determining path congestion frequency and fault handling time data according to the historical operation and maintenance records; It should be noted that the path congestion frequency is the number of times that a certain physical path is blocked by personnel or equipment, which reflects the historical traffic efficiency of the path. The higher the frequency, the more likely the path is to delay traffic.

[0098] Further, the fault handling time data is a record of the time spent in handling faults in the surrounding or associated equipment of a certain physical path during the historical operation and maintenance process, which can indirectly reflect the fault occurrence rate and maintenance complexity of the equipment around the path, providing an empirical reference for current path selection.

[0099] It can be understood that the data related to the physical path is selected from the historical operation and maintenance records, classified and counted according to the path identifier, and the path congestion frequency corresponding to each path is extracted, and the fault handling time data of the equipment around each path is sorted to form historical attribute data for subsequent determination of physical traffic edges.

[0100] Step S314, inputting the path congestion frequency and fault handling time data into the initial fusion graph neural network to determine the physical traffic edge of the current node.

[0101] It should be noted that the physical passing edge of the current node is the edge connecting the current node and other nodes in the initial fusion graph neural network, which represents the passable physical path. The attributes of these edges not only include real-time environmental risks, but also need to be combined with historical operation data to improve efficiency and fault-related attributes.

[0102] Further, the initial fusion graph neural network has data correlation capability, which can match the input historical data with the existing physical passing edges in the network, supplement the corresponding historical attributes for each edge, and determine which edges around the current node are the preferred path edges with high historical passing efficiency and low fault correlation.

[0103] It can be understood that the path congestion frequency and fault handling time data are input into the initial fusion graph neural network, and the network matches the physical passing edges around the current node one by one according to the path identification in the data; add the path congestion frequency and fault handling time attribute labels to each matched physical passing edge, combine these historical attributes with real-time environmental attributes, and finally determine the physical passing edge of the current node.

[0104] Step S32, determining a preset hop range according to the physical passing edge, and performing feature extraction on the subgraph within the preset hop range of the current node to obtain target subgraph features; It should be noted that the preset hop range is a node hop count range around the current node for filtering subgraphs, which is usually set to 5 hops in this embodiment. This range can cover the area within 50 meters around the current node, which can include enough path selection and avoid reducing the operation efficiency due to too large subgraphs.

[0105] Further, the subgraph is a local graph structure within the preset hop range in the initial fusion graph neural network, which includes nodes and edges within the range. The target subgraph feature is the extraction and integration result of the node attributes, edge attributes and associated relationships in the subgraph, which is used for subsequent candidate path generation.

[0106] Further, the target subgraph feature includes the attributes and associated information of the nodes and edges in the subgraph. These information are the core basis for the initial fusion graph neural network to generate candidate paths, which can support the network to analyze the feasibility and rationality of the path.

[0107] It can be understood that the preset hop range is determined according to the average length of the physical passing edge of the current node and the safety requirement in the extreme environment; all nodes and edges within the preset hop range are selected to form a subgraph with the current node as the center; the feature extraction module of the initial fusion graph neural network is called to extract the spatial coordinates, risk levels and other attributes of the nodes in the subgraph, and the path length, dynamic risk weight and other attributes of the edges. These attributes are aggregated to finally obtain the target subgraph feature.

[0108] In an implementable embodiment, step S32 can include steps S321-S325: Step S321, a preset perception range is acquired, and a preset hop range is determined according to the physical passing edge and the preset perception range; It should be noted that the preset perception range is a physical space range set in combination with the safety field of view and operation demand of the operation and maintenance personnel in an extreme environment, and is usually a 50-meter area around the current node. This range can cover the effective environment that can be perceived by the personnel, and can also avoid the operation redundancy caused by a too large subsequent subgraph.

[0109] Further, the preset hop range is the number of node jumps drawn around the current node, which is calculated by dividing the preset perception range by the average length of the physical passing edge, and is used to clearly define the node screening boundary of the subsequent subgraph.

[0110] It can be understood that the preset perception range is first acquired, and then the average length of the physical passing edge around the current node is counted. The number of node jumps is obtained by dividing the preset perception range by the average length, which is the preset hop range. It is ensured that the subgraph within the preset hop range can cover the space area corresponding to the preset perception range.

[0111] Step S322, the subgraph node set is obtained by expanding from the current node according to the preset hop range, and the edge set of the subgraph node set is determined; It should be noted that the subgraph node set is a set of all initial fusion graph nodes within the preset hop range, including the current node, the peripheral path node, the sensor node and the virtual risk node, etc. These nodes jointly constitute the node basis of the subgraph.

[0112] Further, the edge set is a set of initial fusion graph edges connecting nodes in the subgraph node set, including the physical passing edge and the device logical association edge. The edge set and the subgraph node set jointly constitute the complete structure of the subgraph.

[0113] It can be understood that the current node is taken as the starting point, and the preset hop range is sequentially jumped outwards. Each jump includes the corresponding node until all nodes within the preset hop range are covered, forming the subgraph node set. Then, all edges connecting any two nodes in the subgraph node set are screened out to form the edge set, so as to obtain the subgraph basic structure including nodes and edges.

[0114] Step S323, the static attributes of the subgraph node set are extracted, and the initial features are determined according to the static attributes, wherein the static attributes include spatial coordinates and component materials; It should be noted that the subgraph node set includes all nodes within the preset hop range, and each node carries preset static attributes. These attributes are inherent characteristics of the node and do not change with the environment, and are the basis for extracting the initial features.

[0115] In addition, the spatial coordinates in the static attribute are three-dimensional position data of the node in the initial fusion graph neural network, which is used to reflect the spatial distribution relationship of the node and is a spatial basis for path planning; the component material is a material attribute of a node corresponding entity (such as equipment, channel wall), and different materials have different tolerances to extreme environments, which affects the risk association attribute of the node.

[0116] Further, the initial feature is a digital representation of the static attribute of the subgraph node set, which converts the spatial coordinates, component materials and other attributes into a vector form recognizable by the neural network, providing a basic feature carrier for subsequent combination with dynamic risk factors.

[0117] It can be understood that the spatial coordinates and component material information of each node in the subgraph node set are extracted one by one, and redundant attributes unrelated to the initial feature in the node are removed; the spatial coordinates are converted into a numerical vector, and the component material is converted into a numerical label according to a preset rule (such as assigning a value of 1 to high-temperature-resistant material and a value of 0 to ordinary material); the spatial coordinate vector and the material label of each node are integrated to form the initial feature of each node, and finally the initial feature set of the subgraph node set is obtained.

[0118] In step S324, the dynamic risk factor of the initial feature is calculated, and the dynamic risk factor is distributed to the edge set of the subgraph node set according to a preset attenuation rule to obtain a time sequence feature. It should be noted that the dynamic risk factor is a dynamic numerical value calculated based on real-time environmental data, which reflects the risk of the node's surrounding environment, and the value is updated with the change of the environment, which is used to correct the risk attribute of the initial feature.

[0119] In addition, the preset attenuation rule is a rule for describing the change of risk with distance, and in this embodiment, a Gaussian attenuation rule is adopted, that is, the risk value decreases with the increase of the distance between the node and the risk source, which conforms to the actual situation of risk diffusion in the physical world.

[0120] Further, the time sequence feature is the edge feature containing the dynamic risk attribute obtained after the dynamic risk factor is distributed to the edge set according to the preset attenuation rule, which can reflect the real-time risk state of the edge and provide dynamic data support for subsequent subgraph update.

[0121] It can be understood that according to the current temperature, dust concentration and other real-time environmental data, the dynamic risk factor of the surrounding of each node is calculated combined with the spatial coordinates of each node; according to the preset attenuation rule, the contribution value of the dynamic risk factor of each node at both ends of the edge to the edge is calculated, and the contribution value is added as the dynamic risk attribute of the edge; the dynamic risk attribute of the edge and the original static attribute of the edge are integrated to form the time sequence feature of each edge, and finally the time sequence feature set of the edge set is obtained.

[0122] Step S325, updating the subgraph node set according to the timing characteristics, to obtain target subgraph characteristics.

[0123] It should be noted that the target subgraph characteristics are the updated subgraph overall characteristics, which include the characteristics of the subgraph node set and the timing characteristics of the edge set. The two are associated through the connection relationship between nodes and edges to form structured subgraph characteristic data for subsequent generation of candidate paths.

[0124] Further, updating the subgraph node set is to reversely associate the timing characteristics of the edge set to the corresponding nodes, so that the node characteristics include the dynamic risk information of the surrounding edges, ensuring that the node characteristics can reflect the risk state of the associated path.

[0125] It can be understood that according to the timing characteristics of each edge in the edge set, the two nodes connected by the edge are found; the dynamic risk attribute of the edge is distributed to the corresponding two nodes according to a preset weight, and the initial characteristics of the nodes are updated; the characteristics of the updated nodes and the timing characteristics of the edges are integrated to form a complete feature set including the node-edge association relationship, static attributes and dynamic risk attributes. The set is the target subgraph characteristics.

[0126] Step S33, generating a candidate path set from the current node to the target device node according to the target subgraph characteristics; It should be noted that the candidate path set includes multiple potential paths from the current node to the target device node, and the comprehensive cost of each candidate path is the core indicator for measuring the pros and cons of the path, which comprehensively considers the time consumption and risk level of the path to ensure that the selected path has both efficiency and safety.

[0127] Further, the target device node is a device to be maintained or a destination node.

[0128] It can be understood that the target subgraph characteristics are input into the path search module of the initial fusion graph neural network, and the module takes the current node as the starting point and the target device node as the end point, combines the passing attributes of the edges in the subgraph and the risk attributes of the nodes, and analyzes the connected paths between nodes. Set path screening conditions, such as excluding edges with dynamic risk weight exceeding a threshold and limiting the total length of the path to no more than 1.5 times the straight-line distance from the current node to the target device node; According to the screening conditions, 3-5 differentiated paths are selected from all connected paths to form a candidate path set.

[0129] In one possible implementation, step S33 can include steps S331-S333: Step S331, calculating a moving expected return according to the target subgraph characteristics; It should be noted that the target subgraph feature is structured data containing the updated features of the subgraph node set and the edge set time sequence feature, wherein the node feature contains the spatial coordinates, the component material and the dynamic risk information of the associated edge, and the edge feature contains the path length, the dynamic risk weight and the like, and these features are the core basis for calculating the moving expected return.

[0130] In addition, the moving expected return is an estimated value of the comprehensive income that can be obtained in the future after moving from the current node to the surrounding neighbor nodes, which comprehensively considers the path efficiency (such as time cost) and safety risk (such as dynamic risk weight) after moving, and the higher the value, the more optimal the comprehensive income of the moving direction.

[0131] Further, the calculation of the moving expected return needs to rely on the decision module of the initial fusion graph neural network, which will operate on the node and edge features in the target subgraph feature, and convert the feature data into a quantitative return value to provide a decision basis for subsequent path search.

[0132] It can be understood that the features of the current node, the features of the surrounding neighbor nodes and the time sequence features of the physical passing edges between the current node and the neighbor nodes are extracted from the target subgraph feature; these features are input into the decision module, and the module evaluates the time cost saving amount and risk reduction amount of moving to each neighbor node through a preset algorithm, and outputs the moving expected return corresponding to each neighbor node.

[0133] Step S332, setting path search constraints according to the dynamic risk weight threshold of the physical passing edge and the path total length threshold; It should be noted that the dynamic risk weight threshold of the physical passing edge is a critical value preset for judging whether the edge has passing safety, which is determined based on the extreme environment safety standard. If the dynamic risk weight of the edge is greater than the threshold, it means that the path risk corresponding to the edge is too high and is not suitable for passing.

[0134] In addition, the path total length threshold is a critical value preset for limiting the maximum length of the candidate path, which is usually set to 1.5 times the straight-line distance from the current node to the target device node. This threshold can avoid the situation that the path is too long to cause low operation and maintenance efficiency, and ensure that the candidate path has safety and efficiency.

[0135] Further, the path search constraint condition is a screening rule composed of the dynamic risk weight threshold and the path total length threshold. The path generated in the subsequent path search process needs to meet both conditions at the same time to be included in the candidate path set, so as to ensure the basic safety and efficiency of the candidate path.

[0136] It can be understood that the dynamic risk weight threshold of the physical passing edge is determined according to the extreme environment type, such as setting the threshold to 0.8 in a high-temperature environment and setting the threshold to 0.7 in a high-dust environment; the straight-line distance from the current node to the target device node is calculated, and the path total length threshold is determined by a ratio of 1.5 times; the physical passing edge dynamic risk weight ≤ the dynamic risk weight threshold and the path total length ≤ the path total length threshold are taken as two core rules to form the path search constraint condition.

[0137] In step S333, path search is performed according to the greedy strategy, the moving expected return and the constraint condition, taking the current node as the starting point and the target device node as the end point, to generate a candidate path set.

[0138] It can be understood that, taking the current node as the starting point, neighbor nodes around the current node that satisfy the constraint condition are first screened out; from these neighbor nodes, the node with the highest moving expected return is selected as the next moving node according to the greedy strategy; taking the node as a new starting point, the above screening and selection process is repeated until the target device node is reached, forming a complete path; the above search process is repeated multiple times to generate 3-5 different complete paths to form the candidate path set.

[0139] For the tthiteration, when the current path is extended to the neighbor node to generate a new path, the attributes are calculated as follows:

[0140] In the formula, is the cumulative risk of the new path; is the cumulative risk of the original path; is the path dynamic risk function, that is, the real-time risk weight of the edge connecting and (from the risk quantification in step one); is the end node of the original path; is the neighbor node of .

[0141]

[0142] In the formula, is the cumulative time of the new path; is the cumulative time of the original path; is the path passing time function, that is, the static passing time of the edge connecting and ; is the end node of the original path; is the neighbor node of .

[0143]

[0144] In the formula, The cumulative expected return for the new path; This represents the cumulative expected return along the original path. The expected return function for the mobile activity (from the IRL / Q network); The cumulative risk of the new path; This represents the maximum acceptable cumulative risk constraint for the path. The dynamic attenuation coefficient ( (This is used to control the penalty effect of accumulated risk on returns). It is a natural exponential function.

[0145] Furthermore, it is necessary to filter the set of valid paths. The formula for calculating the set of valid paths is as follows:

[0146] In the formula, For the first The set of all legal and feasible paths expanded by round-by-round iteration; This represents the maximum acceptable cumulative risk constraint for the path. This is the maximum acceptable cumulative time constraint for the path.

[0147]

[0148] In the formula, For the first The current bundle (best path set) in each round of iteration; To select a function; The cumulative expected return of the path; Beam Width indicates the number of data points used to determine the beam width. The optimal number of paths to retain in descending order. Finally, by integrating, we obtain the final set of paths that contain all paths that reach the goal and satisfy the constraints.

[0149] Step S34: Calculate the comprehensive cost of the candidate paths in the candidate path set, and determine the target operation and maintenance path based on the comprehensive cost.

[0150] It should be noted that the overall cost consists of time cost and risk cost. Time cost is calculated based on path length and travel speed, reflecting the time taken for the path. Risk cost is calculated based on the dynamic risk weights of the edges, reflecting the degree of environmental risk of the path.

[0151] Furthermore, the target maintenance path is the path with the lowest overall cost selected from the candidate path set. This path can guide maintenance personnel to the target equipment node with the least time and risk cost in extreme environments.

[0152] It can be understood that the time cost of each candidate path is calculated, the total length of all physical passing edges included in the path is summed up, combined with the average passing speed in the extreme environment, and the time cost is obtained by dividing the total length by the average passing speed; the risk cost of each candidate path is calculated, the dynamic risk weight of all physical passing edges included in the path is summed up to obtain the risk cost; the weight coefficient of the time cost and the risk cost is set according to the extreme environment level, the risk cost weight is higher in the high-risk environment, and the comprehensive cost of each candidate path is calculated by weighted summation; the comprehensive costs of all candidate paths are compared, and the path with the lowest comprehensive cost is selected as the target operation and maintenance path.

[0153] In a possible implementation, step S34 can include steps S341-S345: Step S341, defining a comprehensive cost index according to the path level characteristics of the candidate path, the comprehensive cost index including time cost, safety risk and historical reliability; It should be noted that the time cost is an index for evaluating the efficiency of path passing, which is related to the total length of the path and the passing speed; the safety risk is an index for evaluating the environmental risk of the path, which is related to the dynamic risk weight of the physical passing edge; the historical reliability is an index for evaluating the past performance of the path, which is related to the path congestion frequency and fault handling time in the historical operation record.

[0154] It can be understood that according to the path level characteristics of each candidate path, the core evaluation dimensions of the path in efficiency, safety and historical performance are determined; the time cost, safety risk and historical reliability are defined as three components of the comprehensive cost index, each component corresponding to a type of key information in the path level characteristics, forming a complete multi-dimensional evaluation index system.

[0155] Defining a multi-dimensional comprehensive cost index based on path level characteristics can avoid one-sidedness of single index evaluation of the path, ensure that subsequent scoring calculation can comprehensively consider the efficiency, safety and historical reliability of the path, and provide a comprehensive evaluation basis for selecting the optimal path.

[0156] Step S342, scoring calculation of the candidate paths in the candidate path set according to the comprehensive cost index, to obtain the comprehensive cost; It should be noted that the time cost score is calculated by the path total length and the average passing speed in the extreme environment, the shorter the path total length and the more adaptive the passing speed, the lower the time cost score, and the lower the score represents the better efficiency; the safety risk score is calculated by summing up the dynamic risk weight of all physical passing edges in the path, the lower the weight sum, the lower the safety risk score, and the lower the score represents the better safety.

[0157] Further, the historical reliability score is calculated by the path congestion frequency and the fault handling time data. The less the congestion frequency and the shorter the fault handling time, the lower the historical reliability score. The lower the score, the better the historical performance. The comprehensive cost is the result of the weighted sum of the scores of the three indicators according to the preset weight. The lower the comprehensive cost, the better the overall path.

[0158] It can be understood that the time cost score of each candidate path is calculated by dividing the total length of the path by the average travel speed to get a time value, and then converting it into a score according to the preset grading rule. The safety risk score is calculated by summing the dynamic risk weights of all physical travel edges in the path, which is directly used as the safety risk score. The historical reliability score is calculated by summing the congestion frequency and the fault handling time according to the preset proportion, which is used as the historical reliability score. The weights of the three scores are set according to the extreme environment level, and the weighted sum is used to get the comprehensive cost of each candidate path.

[0159] By quantifying the scores, the comprehensive cost indicator is converted into a calculable comprehensive cost, which can objectively and accurately measure the overall advantages and disadvantages of each candidate path, avoiding the deviation caused by subjective judgment and providing a quantitative decision basis for subsequent path selection.

[0160] Step S343, selecting an initial operation and maintenance path according to the comprehensive cost under the preset safety constraint; It should be noted that the preset safety constraint is a rule that is set in advance and is used to ensure the basic safety of the initial operation and maintenance path. This rule is based on the extreme environment safety standard.

[0161] In addition, the initial operation and maintenance path is a path that is selected from the candidate path set and meets the preset safety constraint and has a relatively low comprehensive cost. This path is the basis for subsequent path value evaluation and needs to meet the basic safety requirements before further value evaluation.

[0162] Further, selecting the initial operation and maintenance path requires checking each path in the candidate path set for safety constraints, excluding paths that do not meet the safety requirements, and then selecting the path with the lowest comprehensive cost from the remaining paths as the initial operation and maintenance path, ensuring that the path has both safety and preliminary optimality.

[0163] It can be understood that each path in the candidate path set is checked one by one to see if it meets the preset safety constraint. If the safety risk score of a path exceeds the safety risk threshold, or there is a physical travel edge with a dynamic risk weight greater than the threshold, it is excluded. From the remaining paths that meet the safety constraint, the path with the lowest comprehensive cost is selected as the initial operation and maintenance path.

[0164] Step S344, performing path value evaluation according to the initial operation and maintenance path to obtain a path value; It should be noted that the path value is a quantitative value for measuring the added value of the initial operation and maintenance path. The higher the value, the more additional support the path can provide for operation and maintenance, such as the presence of emergency equipment storage points around the path, the ability of the path to quickly reach the location of associated equipment, and the like.

[0165] Further, the path value evaluation needs to be combined with the operation and maintenance scene requirements, and scored from the distance between the path and the emergency resource, the correlation between the path and the associated equipment, and the sufficiency of the operation space along the path, and then the weighted sum of the scores of each dimension is obtained to obtain the path value.

[0166] It can be understood that the location of the emergency resource around the initial operation and maintenance path is queried, the shortest distance between the path and the emergency resource is calculated, the closer the distance, the higher the score; whether the initial operation and maintenance path can quickly reach the associated equipment of the fault equipment is analyzed, the stronger the correlation, the higher the score; whether the operation space along the path meets the maintenance equipment requirements is evaluated, the more sufficient the space, the higher the score; and the scores of each dimension are weighted and summed according to the preset weight to obtain the path value of the initial operation and maintenance path.

[0167] Step S345, determining the target operation and maintenance path according to the path value.

[0168] Further, the determination of the target operation and maintenance path needs to preset a path value threshold first. If the path value of the initial operation and maintenance path is greater than or equal to the threshold, it means that the added value meets the standard, and the initial operation and maintenance path can be directly used as the target operation and maintenance path. If it is lower than the threshold, the path with the second lowest comprehensive cost from the candidate paths that meet the safety constraints needs to be selected for value evaluation until a path with a value that meets the standard is found.

[0169] It can be understood that the path value of the initial operation and maintenance path is compared with the preset path value threshold. If the path value is greater than or equal to the threshold, the initial operation and maintenance path is determined as the target operation and maintenance path. If the path value is less than the threshold, the path with the second lowest comprehensive cost from the candidate paths that meet the safety constraints is selected as a new initial operation and maintenance path, and the value evaluation step is repeated until a path with a value that meets the standard is found, which is determined as the target operation and maintenance path.

[0170] The determination of the target operation and maintenance path based on the path value can ensure that the final target path is not only safe and efficient, but also has a high added value, which can provide more support for operation and maintenance, avoid problems such as insufficient emergency resources and insufficient operation space in the operation and maintenance process due to neglect of added value, and improve the overall success rate and safety of operation and maintenance tasks.

[0171] By way of example, in order to facilitate understanding of the implementation process of the operation and maintenance guidance method based on the fusion graph neural network obtained after the above-mentioned embodiment one, please refer to Figure 3 , Figure 3A brief flowchart of a maintenance guide method based on a fusion graph neural network is provided, specifically: Starting from dynamic perception, sensor readings need to be mutated to capture real-time changes in the environment, such as temperature, dust concentration, etc. Next, the risk diffusion function is calculated through sensor readings and maintenance worker positions, and a path subgraph is generated, which uses sensor data to assess risks in the environment and builds a dynamic subgraph reflecting the current environmental conditions. Then, through the path subgraph, the meta-path attention and node fusion representation are generated, and the graph neural network technology is used to analyze the paths and nodes in the subgraph, generating a representation that integrates environmental risks and path information to facilitate the next step of decision-making. Perform comprehensive cost calculation to evaluate the cost of all possible paths, including time cost and risk cost, to determine the optimal path. By comparing the costs of different paths, the optimal next action is selected to ensure that maintenance personnel can safely and efficiently reach the target location. The calculated optimal path and related data are pushed to the digital twin terminal of the maintenance personnel, such as AR glasses or industrial tablets, to provide real-time navigation and data support.

[0172] It should be noted that the above examples are only for understanding the present application and do not constitute a limitation on the maintenance guide method based on the fusion graph neural network of the present application. More forms of simple transformation based on this technical concept are within the scope of protection of the present application.

[0173] The present application also provides a maintenance guide device based on a fusion graph neural network, please refer to Figure 4 , the maintenance guide device based on the fusion graph neural network comprises: A data determination module 10 is configured to construct a target building digital twin model according to maintenance order information and determine extreme environment scenario data according to the maintenance order information and the target building digital twin model; A network construction module 20 is configured to construct an initial fusion graph neural network according to the target building digital twin model and the extreme environment scenario data; A path generation module 30 is configured to input the current maintenance personnel position, environmental data and historical maintenance records into the initial fusion graph neural network to obtain a target maintenance path; A maintenance guide module 40 is configured to generate navigation instructions according to the target maintenance path to complete the maintenance guide based on the fusion graph neural network.

[0174] The operation and maintenance guiding device based on the fusion graph neural network provided in the application adopts the operation and maintenance guiding method based on the fusion graph neural network in the above embodiment, and can solve the technical problem of how to ensure that the operation and maintenance personnel can safely and efficiently complete job navigation and decision-making in the case of dynamic path obstruction and incomplete environmental information in an extreme environment. Compared with the prior art, the operation and maintenance guiding device based on the fusion graph neural network provided in the application has the same beneficial effects as the operation and maintenance guiding method based on the fusion graph neural network provided in the above embodiment, and other technical features in the operation and maintenance guiding device based on the fusion graph neural network are the same as the features disclosed in the above embodiment method, which will not be repeated here.

[0175] The application provides an operation and maintenance guiding device based on a fusion graph neural network. The operation and maintenance guiding device based on the fusion graph neural network comprises at least one processor and a memory in communication connection with the at least one processor. The memory stores instructions executable by the at least one processor. The instructions are executed by the at least one processor to enable the at least one processor to perform the operation and maintenance guiding method based on the fusion graph neural network in the above embodiment one.

[0176] Reference will be made to the following description Figure 5 which shows a structural schematic diagram of the operation and maintenance guiding device based on the fusion graph neural network suitable for being used to implement the embodiments of the application. The operation and maintenance guiding device based on the fusion graph neural network in the embodiments of the application can include but is not limited to mobile terminals such as mobile phones, notebook computers, digital broadcast receivers, PDAs (Personal Digital Assistant), PADs (Portable Application Description), PMPs (Portable Media Player), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), and the like, and fixed terminals such as digital TVs, desktop computers, and the like. Figure 5 The operation and maintenance guiding device based on the fusion graph neural network shown is only an example, and should not bring any limitation to the functions and use range of the embodiments of the application.

[0177] As Figure 5As shown, the operation and maintenance guidance device based on the fusion graph neural network can include a processing apparatus 1001 (for example, a central processing unit, a graphics processing unit, etc.) that can perform various appropriate actions and processes according to a program stored in a ROM (Read Only Memory) 1002 or a program loaded from a storage apparatus 1003 into a RAM (Random Access Memory) 1004. In the RAM 1004, various programs and data required for operation and maintenance guidance device based on the fusion graph neural network are also stored. The processing apparatus 1001, the ROM 1002, and the RAM 1004 are connected to each other through a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Generally, the following systems can be connected to the I / O interface 1006: an input apparatus 1007 including, for example, a touch panel, a touch pad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; an output apparatus 1008 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; the storage apparatus 1003 including, for example, a magnetic tape, a hard disk, etc.; and a communication apparatus 1009. The communication apparatus 1009 can allow the operation and maintenance guidance device based on the fusion graph neural network to communicate with other devices wirelessly or by wire to exchange data. Although the operation and maintenance guidance device based on the fusion graph neural network with various systems is shown in the figure, it should be understood that all the shown systems are not required to be implemented or possessed. More or less systems can be alternatively implemented or possessed.

[0178] In particular, according to embodiments of the present disclosure, the processes described above with reference to the flowcharts can be implemented as a computer software program. For example, embodiments of the present disclosure include a computer program product comprising a computer program carrying computer program code for executing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network through the communication apparatus, or installed from the storage apparatus 1003, or installed from the ROM 1002. When the computer program is executed by the processing apparatus 1001, the above-mentioned functions defined in the methods of the embodiments of the present disclosure are performed.

[0179] The operation and maintenance guiding device based on the fusion graph neural network provided in the application adopts the operation and maintenance guiding method based on the fusion graph neural network in the above embodiment, and can solve the technical problem of how to ensure that the operation and maintenance personnel can safely and efficiently complete job navigation and decision-making in the case of dynamic path obstruction and incomplete environmental information in an extreme environment. Compared with the prior art, the operation and maintenance guiding device based on the fusion graph neural network provided in the application has the same beneficial effects as the operation and maintenance guiding method based on the fusion graph neural network provided in the above embodiment, and other technical features in the operation and maintenance guiding device based on the fusion graph neural network are the same as the features disclosed in the above embodiment method, and will not be repeated here.

[0180] It should be understood that parts of the present application can be realized by hardware, software, firmware or a combination thereof. In the description of the above embodiments, specific features, structures, materials or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0181] The above is merely specific implementation of the present application, but the protection scope of the present application is not limited thereto, any person skilled in the art can easily think of changes or replacements within the technical scope disclosed in the present application, which should be covered in the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

[0182] The present application provides a computer readable storage medium having stored thereon computer readable program instructions (i.e. computer programs) for performing the operation and maintenance guiding method based on the fusion graph neural network in the above embodiment.

[0183] The computer readable storage medium provided in the present application may, for example, be a U disk, but is not limited to an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, system, or device, or any combination of the above. More specific examples of computer readable storage media can include, but are not limited to, an electrical connection having one or more wires, a portable computer diskette, a hard disk, a RAM (Random Access Memory), a ROM (Read Only Memory), an erasable programmable ROM (Erasable Programmable Read Only Memory or flash memory, EPROM), an optical fiber, a CD-ROM (CD-Read Only Memory), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present embodiment, the 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, system, or device. The program code contained on the computer readable storage medium can be transmitted in any suitable medium, including but not limited to electrical wires, optical cables, RF (Radio Frequency), etc., or any suitable combination of the above.

[0184] The above computer readable storage medium can be included in the operation and maintenance guidance device based on the fusion graph neural network; or can exist separately and not be assembled into the operation and maintenance guidance device based on the fusion graph neural network.

[0185] The above computer readable storage medium carries one or more programs, which, when executed by the operation and maintenance guidance device based on the fusion graph neural network, cause the operation and maintenance guidance device based on the fusion graph neural network to: ddd Computer program code for carrying out operations of the present application can be written in one or more programming languages or combinations of the above, including an object-oriented programming language such as Java, Smalltalk, C++, or a conventional procedural programming language such as "C" language or a similar programming language. The program code can be executed entirely on a user computer, partially on a user computer, as a separate software package, partially on a user computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer can be connected to the user computer through any kind of network, including a LAN (Local Area Network) or a WAN (Wide Area Network), or can be connected to an external computer (for example, through the Internet using an Internet service provider).

[0186] The flow diagrams and the block diagrams in the drawings are illustrations of possible architectures, functions, and operations for systems, methods, and computer program products according to various embodiments of the present application. In this regard, each block in the flow diagrams or block diagrams can represent a module, a segment, or a portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that in some alternative implementations, the functions noted in the blocks can occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and / or flow diagrams, and combinations thereof, can be implemented by special purpose hardware-based systems that perform the specified functions or operations, or combinations of special purpose hardware and computer instructions.

[0187] The modules involved in the embodiments of the present application can be implemented in the form of software or in the form of hardware. In some cases, the name of the module does not constitute a limitation on the unit itself.

[0188] The readable storage medium provided by the present application is a computer readable storage medium, which stores computer readable program instructions (i.e., a computer program) for executing the above-mentioned operation and maintenance guidance method based on the fusion graph neural network, and can solve the technical problem of how to ensure that the operation and maintenance personnel can safely and efficiently complete the job navigation and decision-making in the case of dynamic path obstruction and incomplete environmental information in an extreme environment. Compared with the prior art, the computer readable storage medium provided by the present application has the same beneficial effects as the operation and maintenance guidance method based on the fusion graph neural network provided by the above-mentioned embodiments, and will not be described here.

[0189] The present application also provides a computer program product comprising a computer program, which, when executed by a processor, implements the steps of the operation and maintenance guidance method based on the fusion graph neural network as described above.

[0190] The computer program product provided by the present application can solve the technical problem of how to ensure that the operation and maintenance personnel can safely and efficiently complete the job navigation and decision-making in the case of dynamic path obstruction and incomplete environmental information in an extreme environment. Compared with the prior art, the computer program product provided by the present application has the same beneficial effects as the operation and maintenance guidance method based on the fusion graph neural network provided by the above-mentioned embodiments, and will not be described here.

[0191] The above merely describes some embodiments of the present application, and does not limit the patent scope of the present application. Any equivalent structural transformation, direct / indirect application in other related technical fields, or the like, which is made based on the technical concept of the present application and the content of the specification and drawings, is included in the patent protection scope of the present application.

Claims

1. An operation and maintenance guidance method based on a fusion graph neural network, characterized in that, The method comprises: constructing a target building digital twin model according to operation and maintenance work order information and determining extreme environment scenario data according to the operation and maintenance work order information and the target building digital twin model; constructing an initial fusion graph neural network according to the target building digital twin model and the extreme environment scenario data; inputting current maintenance personnel position, environment data and historical operation and maintenance records into the initial fusion graph neural network to obtain a target operation and maintenance path; generating navigation instructions according to the target operation and maintenance path to complete operation and maintenance guidance based on the fusion graph neural network.

2. The method of claim 1, wherein, The step of inputting the current maintenance personnel position, environment data and historical operation and maintenance records into the initial fusion graph neural network to obtain the target operation and maintenance path comprises: determining a current node and physical passing edges of the current node by inputting the current maintenance personnel position, environment data and historical operation and maintenance records into the initial fusion graph neural network; determining a preset hop range according to the physical passing edges, and performing feature extraction on subgraphs within the preset hop range of the current node to obtain target subgraph features; generating a candidate path set from the current node to a target device node according to the target subgraph features; calculating the comprehensive cost of the candidate paths in the candidate path set, and determining a target operation and maintenance path according to the comprehensive cost.

3. The method of claim 2, wherein, The step of determining the current node and the physical passing edges of the current node by inputting the current maintenance personnel position, environment data and historical operation and maintenance records into the initial fusion graph neural network comprises: determining current temperature, dust concentration and harmful gas concentration according to environment data; determining a current node by inputting the current maintenance personnel position, the current temperature, the dust concentration and the harmful gas concentration into the initial fusion graph neural network; determining path congestion frequency and fault handling time data according to historical operation and maintenance records; inputting the path congestion frequency and fault handling time data into the initial fusion graph neural network to determine the physical passing edges of the current node.

4. The method of claim 2, wherein, The step of determining a preset hop range according to the physical passing edges, and performing feature extraction on subgraphs within the preset hop range of the current node to obtain target subgraph features comprises: acquiring a preset perception range, determining a preset hop range according to the physical passing edges and the preset perception range; expanding from the current node according to the preset hop range to obtain a subgraph node set and determine an edge set of the subgraph node set; extracting static attributes of the subgraph node set, and determining initial features according to the static attributes, wherein the static attributes include spatial coordinates and component materials; calculating dynamic risk factors of the initial features, and distributing the dynamic risk factors to the edge set of the subgraph node set according to a preset decay rule to obtain time sequence features; updating the subgraph node set according to the time sequence features to obtain target subgraph features.

5. The method of claim 2, wherein, The step of generating a candidate path set from the current node to a target device node according to the target subgraph features comprises: calculating a moving expected return according to the target subgraph features; setting a path search constraint condition according to a dynamic risk weight threshold of the physical passing edge and a path total length threshold. According to the current node as the starting point, the target device node as the end point, the path search is performed according to the greedy strategy, the mobile expected return and the constraint condition, and a candidate path set is generated.

6. The method of claim 2, wherein, The step of calculating the comprehensive cost of the candidate paths in the candidate path set and determining the target operation and maintenance path according to the comprehensive cost comprises: According to the path level characteristics of the candidate paths, a comprehensive cost index is defined, and the comprehensive cost index comprises time cost, safety risk and historical reliability; According to the comprehensive cost index, the candidate paths in the candidate path set are scored and calculated to obtain a comprehensive cost; Under a preset safety constraint, an initial operation and maintenance path is selected according to the comprehensive cost; According to the initial operation and maintenance path, path value evaluation is performed to obtain a path value; According to the path value, a target operation and maintenance path is determined.

7. The method of claim 1, wherein, The step of generating a navigation instruction according to the target operation and maintenance path to complete the operation and maintenance guidance based on the fusion graph neural network comprises: According to the target operation and maintenance path, an operation and maintenance perspective is generated and sent to a digital twin mobile terminal, so that the digital twin mobile terminal feeds back operation and maintenance site information according to the target operation and maintenance path; According to the operation and maintenance site information, a target operation and maintenance data packet is determined; According to the target operation and maintenance data packet, a navigation instruction is generated to complete the operation and maintenance guidance based on the fusion graph neural network.

8. The method of claim 1, wherein, The step of constructing an initial fusion graph neural network according to the target building digital twin model and the extreme environment scenario data comprises: According to the target building digital twin model, building information model geometric data, building information model non-geometric data and equipment association data are determined; According to the extreme environment scenario data, environment sensor deployment position and risk area distribution data are obtained; According to the building information model geometric data, the building information model non-geometric data, the equipment association data, the environment sensor deployment position and the risk area distribution data, an initial fusion graph node and an initial fusion graph edge are defined, wherein the initial fusion graph node comprises a maintenance personnel starting candidate node, a target device node, a sensor node and a virtual risk node, and the initial fusion graph edge comprises a physical passing edge and a device logical association edge; According to the initial fusion graph node and the initial fusion graph edge, an initial fusion graph neural network is constructed.

9. The method of claim 1, wherein, The step of constructing a target building digital twin model according to operation and maintenance work order information and determining extreme environment scenario data according to the operation and maintenance work order information and the target building digital twin model comprises: From the operation and maintenance work order information, a fault device identifier, a fault type and an operation and maintenance area range are extracted; According to the fault device identifier, the fault type and the operation and maintenance area range, a target building digital twin model is constructed; According to the target building digital twin model, a model position area of a fault device of the operation and maintenance work order information is located; An extreme environment type of the model position area is determined; According to the target building digital twin model and the extreme environment type, environmental parameters are analyzed to obtain extreme environment scenario data.

10. An operation and maintenance guiding device based on a fusion graph neural network, characterized in that, The device comprises: The data determination module is configured to construct a target building digital twin model according to operation and maintenance work order information and determine extreme environment scenario data according to the operation and maintenance work order information and the target building digital twin model; The network construction module is configured to construct an initial fusion graph neural network according to the target building digital twin model and the extreme environment scenario data; The path generation module is configured to input a current maintenance personnel position, environment data and historical operation and maintenance records into the initial fusion graph neural network to obtain a target operation and maintenance path; The operation and maintenance guidance module is configured to generate a navigation instruction according to the target operation and maintenance path to complete operation and maintenance guidance based on the fusion graph neural network.

Images