Visual disassembly platform for boiler fan of thermal power plant based on digital twinning technology

Abstract

The application discloses a power plant boiler fan visual disassembly platform based on digital twin technology, and relates to the technical field of power equipment maintenance.The platform comprises a data fusion module, a model construction module, a visual disassembly module and an intelligent management module.The data fusion module is used to construct a maintenance knowledge graph.The model construction module performs enhanced semantic labeling on a constructed three-dimensional fan model, associates the three-dimensional fan model with the maintenance knowledge graph, and obtains an identifier mapping table.The visual disassembly module performs rendering and interaction of the three-dimensional fan model, and updates the three-dimensional fan model in real time based on the identifier mapping table, and simultaneously decomposes and visually presents the disassembly process based on the maintenance knowledge graph.The intelligent management module performs intelligent management based on the maintenance knowledge graph and the identifier mapping table.The application realizes visual disassembly of the fan, improves the efficiency and accuracy of maintenance work, and reduces the maintenance cost and resource consumption.

Description

Technical Field

[0001] This application relates to the field of power equipment maintenance technology, specifically to a visual disassembly and assembly platform for boiler fans in thermal power plants based on digital twin technology. Background Technology

[0002] Currently, boiler fans, as key production equipment in thermal power plants, directly determine the reliability and continuity of the entire power generation system through their safe and stable operation, serving as a crucial foundation for ensuring power supply. Power generation costs have become a core determinant of power plants' competitiveness in grid bidding. Furthermore, within the total life-cycle cost structure of thermal power equipment, the maintenance and repair costs of easily worn-out rotating equipment such as fans, coal mills, circulating pumps, and motors continue to rise. The current traditional maintenance methods, primarily based on preventative maintenance, lack specificity and flexibility, making them ill-suited to the demands of high-quality development in the power market and improved business efficiency for enterprises.

[0003] The current maintenance system for wind turbines in thermal power plants generally adopts a model centered on planned maintenance, supplemented by routine troubleshooting and post-incident maintenance. However, this model presents several problems in practice: First, maintenance records largely rely on paper ledgers, and equipment technical data is stored in a scattered manner without effective correlation, resulting in insufficient focus in maintenance work and cumbersome and time-consuming fault location processes. Second, the maintenance of wind turbines and other rotating equipment relies heavily on external collaboration, leading to high maintenance costs and delays in emergency service response due to the dominance of the original manufacturer or specialized maintenance companies in terms of technology and pricing. This seriously affects the efficiency of equipment failure handling; thirdly, there are no unified industry standards and specifications for the disassembly, assembly and replacement of wind turbine components, making it difficult to ensure the consistency of maintenance processes, directly leading to inconsistent maintenance quality and difficulty in effective control, which in turn poses a potential risk to the long-term stable and reliable operation of the equipment; fourthly, maintenance personnel mainly rely on equipment alarm information and traditional maintenance manuals for their work, and the maintenance process lacks intuitive and visual guidance, which not only prolongs the skill training cycle of maintenance personnel, but also restricts the overall improvement of maintenance efficiency and quality.

[0004] The information disclosed in this background section is intended only to enhance the understanding of the overall background of this application and should not be construed as an admission or in any way implying that the information constitutes prior art known to those skilled in the art. Summary of the Invention

[0005] The purpose of this application is to provide a visual disassembly and assembly platform for boiler fans in thermal power plants based on digital twin technology, which can solve the problems raised in the background technology to a certain extent, realize the visualization of the fan disassembly and assembly process, thereby improving the efficiency of maintenance operations and reducing maintenance costs.

[0006] To achieve the above objectives, this application provides the following technical solution:

[0007] This application provides a visual disassembly and assembly platform for boiler fans in thermal power plants based on digital twin technology, including a data fusion module, a model building module, a visual disassembly and assembly module, and an intelligent management module;

[0008] The data fusion module is used to acquire multi-source heterogeneous data and construct a maintenance knowledge graph based on the multi-source heterogeneous data;

[0009] The model building module is used to construct a three-dimensional wind turbine model and perform enhanced semantic annotation on the three-dimensional wind turbine model based on multi-source heterogeneous data; the model building module is also used to associate the three-dimensional wind turbine model with the maintenance knowledge graph to obtain an identifier mapping table;

[0010] The visualization disassembly and assembly module is used to render and interact with the 3D wind turbine model, and to update and synchronize the 3D wind turbine model in real time based on the identifier mapping table and multi-source heterogeneous data; the visualization disassembly and assembly module also decomposes and visualizes the disassembly and assembly process based on the maintenance knowledge graph;

[0011] The intelligent management module is based on a maintenance knowledge graph and an identifier mapping table to manage users and permissions, work orders and processes, and faults and decisions.

[0012] As a preferred embodiment of the digital twin-based visualization disassembly and assembly platform for boiler fans in thermal power plants described in this application, the model building module is configured with a 3D fan model building strategy for building a 3D fan model; the 3D fan model building strategy specifically includes:

[0013] Identify the key areas for maintenance and, based on these key areas and the working principle of the fan, identify the core functional units of the fan; according to the design drawings and bill of materials information of each core functional unit, decompose the corresponding core functional units step by step to the basic component level to obtain different parts;

[0014] Set the main driving parameters, derived parameters and corresponding calculation rules for each component, and set the baseline value of the main driving parameters to obtain the parameterized component template;

[0015] The constraint relationships between the parameterized component templates are determined, and a geometric constraint network is constructed based on the constraint relationships. Based on the geometric constraint network, all component templates are organized according to the hierarchical structure of the bill of materials information to obtain a three-dimensional wind turbine model.

[0016] As a preferred embodiment of the digital twin-based visualization disassembly and assembly platform for boiler fans in thermal power plants described in this application, the geometric constraint network is constructed as follows:

[0017] Based on the design drawings, obtain the geometric shape and theoretical assembly position of each component template, and analyze the assembly relationship between each component template;

[0018] Analyze the physical fit type between any two component templates with assembly relationship, and map the physical fit type to the corresponding constraint condition;

[0019] The constraints include positioning constraints and connection constraints; the positioning constraints include concentric or coaxial constraints, coplanar or fitting constraints, and distance constraints; the connection constraints include bolt connection constraints, pin connection constraints, and key connection constraints.

[0020] A geometric constraint network is constructed using the component templates as nodes and the constraints between the component templates as edges.

[0021] As a preferred embodiment of the digital twin-based visualization disassembly and assembly platform for boiler fans in thermal power plants described in this application, the model building module is configured with an enhanced semantic annotation strategy for performing enhanced semantic annotation on the 3D fan model; the enhanced semantic annotation strategy specifically includes:

[0022] Obtain the assembly process document and identify the key entities and corresponding process parameters in the assembly process document;

[0023] By using a similarity matching strategy, a synonym mapping table is established between the key entity and the names and feature terms of each component in the 3D wind turbine model. Based on the synonym mapping table, the process parameters are added to the corresponding component templates.

[0024] Based on the core functional units and assembly process documents, assembly process nodes are divided, and based on the assembly process nodes, information is labeled on the component templates of the three-dimensional wind turbine model.

[0025] The information annotation includes geometric dimension annotation, process dimension annotation, fault association annotation, and standard reference annotation.

[0026] As a preferred embodiment of the digital twin-based visualization disassembly and assembly platform for boiler fans in thermal power plants described in this application, the visualization disassembly and assembly module is configured with a rendering interaction strategy for rendering and interacting with the 3D fan model; the rendering interaction strategy specifically includes:

[0027] Based on the geometric constraint network and the geometric dimension annotation of each component template, computer-aided design software is used to perform feature modeling on the three-dimensional wind turbine model to obtain the basic three-dimensional model.

[0028] The basic 3D model is imported into a digital content creation tool, and based on the digital content creation tool, the complex surfaces and appearance details in the basic 3D model are refined and optimized to obtain an optimized 3D model.

[0029] To optimize the 3D model, a high dynamic range texture map is created, and the optimized 3D model is associated with the corresponding texture map to generate a GLTF file;

[0030] Based on the GLTF file, a 3D visualization scene is constructed using a 3D engine that integrates WebGL.

[0031] As a preferred embodiment of the digital twin-based visualization disassembly and assembly platform for boiler fans in thermal power plants described in this application, the visualization disassembly and assembly module is configured with a 3D fan model update strategy for real-time updating and synchronizing the 3D fan model; the 3D fan model update strategy specifically includes:

[0032] Receive real-time operating data and, based on the identifier mapping table, map the real-time operating data to the corresponding components and their main drive parameters in the three-dimensional wind turbine model;

[0033] Extract the actual values ​​of the main driving parameters and calculate the actual deviation between the actual values ​​of the main driving parameters and the reference values; use the actual deviation as a known input quantity and input it into the main driving parameters corresponding to the geometric constraint network;

[0034] Based on the constraint relationships defined in the geometric constraint network, starting from the component corresponding to the main driving parameter, and using the actual value of the main driving parameter as the driving source, constraint propagation calculation is performed to obtain the updated main driving parameter.

[0035] The updated main drive parameters are passed back to the corresponding component template, and the 3D wind turbine model is updated in the 3D visualization scene based on the component template.

[0036] As a preferred embodiment of the digital twin-based visualization disassembly and assembly platform for boiler fans in thermal power plants described in this application, the visualization disassembly and assembly module is further configured with a disassembly and assembly process decomposition strategy for decomposing and visualizing the disassembly and assembly process; the disassembly and assembly process decomposition strategy specifically includes:

[0037] All process entities associated with the target wind turbine equipment entity are obtained from the maintenance knowledge graph, and a directed graph of disassembly and assembly processes is constructed based on the sequential relationship between the process entities.

[0038] Receive the user's disassembly and assembly instructions, and locate the starting process node in the directed graph of the disassembly and assembly process based on the disassembly and assembly instructions; extract the process sequence based on the starting process node and according to the sequential relationship of the directed graph of the disassembly and assembly process.

[0039] In the three-dimensional visualization scene, based on the identifier mapping table, each process entity in the process sequence is mapped to the corresponding component or assembly process node in the three-dimensional wind turbine model.

[0040] As a preferred embodiment of the digital twin-based visualization disassembly and assembly platform for boiler fans in thermal power plants described in this application, the disassembly and assembly process decomposition strategy further includes:

[0041] In the preset interface area of ​​the 3D visualization scene, the process entities and required resource entities of the current disassembly / assembly step are displayed;

[0042] The driver engineering view controller visually highlights the parts operated in the current process step; if the current process step is disassembly, the driver explodes the view to display the parts locally or globally.

[0043] If the current process step is assembly, then demonstrate the assembly process in reverse.

[0044] For any step in the process sequence, based on the identifier mapping table, a deep information query is performed on the components involved in the process step, and the information annotation associated with the maintenance knowledge graph is returned.

[0045] As a preferred embodiment of the digital twin-based visualization disassembly and assembly platform for boiler fans in thermal power plants described in this application, the 3D engine is configured with different view controllers, including a view navigation controller, a model manipulation controller, and an engineering view controller.

[0046] The view navigation controller can be used to translate, rotate, and zoom the 3D visualization scene.

[0047] When a user selects a specific component by clicking or selecting a box, the user can translate and rotate the component using the model manipulation controller. The transformation matrix of translation and rotation is restricted by the constraint relationship in the geometric constraint network.

[0048] The engineering view controller provides sectional and exploded views of the 3D wind turbine model;

[0049] The sectional view dynamically trims the 3D wind turbine model using a custom sectional plane; the exploded view decomposes the components at each level according to the assembly logic of the 3D wind turbine model.

[0050] As a preferred embodiment of the digital twin-based visualized disassembly and assembly platform for boiler fans in thermal power plants described in this application, the data fusion module is configured with a maintenance knowledge graph construction strategy; the maintenance knowledge graph construction specifically includes:

[0051] Named entity recognition is performed on the multi-source heterogeneous data, and core entities are extracted, including equipment entities, fault entities, process entities, resource entities, and standard entities.

[0052] A deep learning-based relation extraction model is used to extract semantic relationships between core entities from the multi-source heterogeneous data, and the relation types are identified based on the semantic relationships; the relation types include compositional relations, causal relations, sequential relations, dependent relations, and mutually exclusive relations.

[0053] Using the core entities as nodes and the relationship types between the core entities as edges, and adding data sources to each core entity and edge, a maintenance knowledge graph is obtained.

[0054] Compared with the prior art, the beneficial effects achieved by the present invention are as follows:

[0055] By constructing a 3D wind turbine model and a maintenance knowledge graph, maintenance personnel can quickly locate faults and obtain the disassembly and assembly sequence without repeatedly consulting paper documents, effectively shortening maintenance downtime and improving maintenance efficiency. The constructed 3D wind turbine model undergoes enhanced semantic annotation, improving operational accuracy. Linking the 3D wind turbine model with the maintenance knowledge graph yields an identifier mapping table. The 3D wind turbine model is rendered and interacted with, and updated in real-time based on the identifier mapping table. Simultaneously, the disassembly and assembly process is decomposed and visualized based on the maintenance knowledge graph, improving the accuracy and efficiency of complex disassembly and assembly operations while reducing maintenance costs and resource consumption. Attached Figure Description

[0056] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein:

[0057] Figure 1 A structural diagram of the visualization disassembly and assembly platform for boiler fans in thermal power plants based on digital twin technology provided in this application;

[0058] Figure 2 The flowchart for constructing the three-dimensional wind turbine model provided in this application;

[0059] Figure 3 A schematic diagram of the three-dimensional wind turbine model provided in this application. Detailed Implementation

[0060] The technical solution of this application will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments and specific features in the embodiments are detailed descriptions of the technical solution of this application, rather than limitations thereof. In the absence of conflict, the embodiments and technical features in the embodiments can be combined with each other.

[0061] like Figure 1 As shown in the figure, this embodiment introduces a visualization disassembly and assembly platform for boiler fans in thermal power plants based on digital twin technology, including a data fusion module, a model building module, a visualization disassembly and assembly module, and an intelligent management module;

[0062] The data fusion module is used to acquire multi-source heterogeneous data and preprocess the multi-source heterogeneous data. The multi-source heterogeneous data includes design and manufacturing data, historical operation and maintenance data, real-time operation data, and spatial scanning data. The design and manufacturing data includes wind turbine design drawings and bill of materials information. The historical operation and maintenance data includes work order records, fault reports, spare parts replacement records, etc. The real-time operation data is acquired by collecting sensor time-series data through an industrial IoT gateway, including vibration data, temperature data, and pressure data. The spatial scanning data is acquired by using a laser scanner to acquire high-density point cloud data of the wind turbine.

[0063] The specific method for preprocessing the multi-source heterogeneous data is as follows: the running data is cleaned and normalized to remove outliers and fill in missing values; the spatial scanning data is denoised and registered as point clouds, and a three-dimensional mesh model with the same scale as the wind turbine is generated.

[0064] The data fusion module is also used to construct a maintenance knowledge graph based on the multi-source heterogeneous data; the construction method of the maintenance knowledge graph is as follows:

[0065] A natural language processing model is used to perform named entity recognition on the multi-source heterogeneous data and extract core entities; the core entities include equipment entities, fault entities, process entities, resource entities, and standard entities;

[0066] Specifically, the equipment entity refers to the fan system and all its physical components, including the entire unit, functional components, and basic parts; such as the blower and main shaft; the fault entity refers to abnormal states, performance degradation, or functional failure modes that occur during the operation or maintenance of the fan; such as excessive vibration or oil leakage; the process entity refers to independent operating steps or activity units in the maintenance work that can be decomposed and have clear start and end points, such as removing the coupling cover or measuring the radial clearance of the bearing; the resource entity refers to the various tools, materials, spare parts, and personnel roles required to support the smooth progress of the maintenance work; and the standard entity refers to the technical specifications, industry standards, enterprise procedures, or specific requirements of the manufacturer on which the maintenance work is based.

[0067] A deep learning-based relation extraction model is used to extract semantic relationships between core entities from the multi-source heterogeneous data, and the relation types are identified based on the semantic relationships. The relation types include: compositional relationships describing the overall and partial structures of equipment entities; causal relationships describing one entity causing another entity; sequential relationships describing the logical order that must be strictly followed between two process entities; resource entity requirements describing the necessary or recommended use of any process entity; and mutual exclusion relationships describing the prohibition of parallel or adjacent execution of two process entities from safety or technical logic.

[0068] Using the core entities as nodes and the relationship types between the core entities as edges, and adding data sources to each core entity and edge, a maintenance knowledge graph is obtained.

[0069] The model building module uses a parametric design method to construct a 3D wind turbine model and performs enhanced semantic annotation on the 3D wind turbine model based on multi-source heterogeneous data; such as Figure 2 and Figure 3 As shown, the specific method for constructing the three-dimensional wind turbine model is as follows:

[0070] The key areas of maintenance are identified, and based on the working principle and key areas of maintenance of the fan, the core functional units of the fan are identified, including the impeller system, bearing housing assembly, hydraulic adjustment mechanism, coupling assembly, and casing. The key areas of maintenance refer to the critical parts or links in the fan equipment that have a high failure rate, high maintenance complexity, and significant impact on system operation.

[0071] Based on the design drawings and bill of materials for each core functional unit, the corresponding core functional unit is decomposed down to the basic component level to obtain different parts;

[0072] Each component is configured with its main drive parameters, derived parameters, and corresponding calculation rules. A baseline value for the main drive parameters is also set to obtain a parameterized component template. The main drive parameters are independent variables that determine the core geometry and performance of the component. The derived parameters and their corresponding calculation rules are variables and calculation rules automatically derived from the main drive parameters through defined mathematical relationships or geometric constraints. For example, for a centrifugal impeller, the impeller diameter, number of blades, and inlet installation angle are determined as the main drive parameters, and the impeller outlet width calculated from the impeller diameter is used as a derived parameter. The calculation relationship between the impeller diameter and the impeller outlet width is used as the corresponding calculation rule.

[0073] The constraint relationships between the parameterized component templates are determined, and a geometric constraint network is constructed based on these constraints. The geometric constraint mesh is constructed as follows:

[0074] Based on the design drawings, the geometric shape and theoretical assembly position of each component template are obtained, and the assembly relationship between each component template is analyzed. The assembly relationship refers to a physically definite relative position, movement or connection relationship with engineering significance, which has direct constraints and guiding significance for the assembly, disassembly, operation or maintenance of the equipment.

[0075] Analyze the physical fit type between any two component templates with assembly relationship, and map the physical fit type to the corresponding constraint condition;

[0076] Specifically, the constraints include positioning constraints and connection constraints; the positioning constraints are used to determine the relative position and relative direction between components; the positioning constraints include: concentric or coaxial constraints acting between cylindrical surfaces of shafts and holes, bearings and shaft shoulders to ensure alignment of rotation centers; coplanar or fitting constraints acting between mating planes such as flanges and end caps to ensure sealing or force transmission; distance constraints defining the distance between two planes or axes, such as the clearance between an impeller and a housing; and angular constraints defining a fixed angle between feature surfaces, such as the blade mounting angle.

[0077] Optionally, the connection constraints are used to define detachable or non-detachable connection methods, including: bolt connection constraints defining bolt distribution circle diameter, number of bolts, and thread fit; pin connection constraints defining pin hole position and fit tolerance; and key connection constraints defining keyway size and fit.

[0078] A geometric constraint network is constructed using the component templates as nodes and the constraints between the component templates as edges.

[0079] Based on the geometric constraint network, all component templates are organized according to the hierarchical structure of the bill of materials information to obtain a three-dimensional wind turbine model.

[0080] The specific method for enhancing semantic annotation of the three-dimensional wind turbine model is as follows:

[0081] The assembly process document is obtained, and natural language processing technology is used to identify the key entities and corresponding process parameters in the assembly process document. Through a similarity matching strategy, a synonym mapping table is established between the key entities and the names and feature terms of each component in the 3D wind turbine model. Based on the synonym mapping table, the process parameters are added to the corresponding component templates.

[0082] In this embodiment, the similarity matching strategy is as follows: calculate the semantic similarity between each key entity in the assembly process file and the name or feature noun of each component in the 3D wind turbine model, and extract the maximum semantic similarity value; if the maximum semantic similarity value is greater than a preset similarity threshold, then the key entity and the component name or feature noun corresponding to the maximum semantic similarity value are determined to be synonyms.

[0083] Based on the core functional units and assembly process documents, assembly process nodes are divided, and based on the assembly process nodes, model definition technology is used to annotate the component templates of the three-dimensional wind turbine model with information, including geometric dimension annotation, process dimension annotation, fault association annotation and standard reference annotation.

[0084] In this embodiment, the assembly process nodes are divided as follows: based on the principle of not breaking down key cooperation relationships and not crossing functional boundaries, the overall disassembly and assembly process of the wind turbine is divided into a stepped process node, including a first-level process node, a second-level process node, and a third-level process node; the first-level process node corresponds to the overall disassembly and assembly of the core functional unit, the second-level process node corresponds to the disassembly and assembly of the sub-assemblies under the core functional unit, and the third-level process node corresponds to the specific operation of basic components.

[0085] Specifically, the geometric dimension annotation involves marking geometric dimensions, form and position tolerances, surface roughness, welding symbols, etc., on the three-dimensional geometric features of the component template; the geometric dimensions include shaft diameter, hole diameter, flange thickness, etc., and the form and position tolerances include roundness, coaxiality, parallelism, etc.; the process dimension annotation involves binding the process parameters extracted from the assembly process document to the corresponding component or assembly process node using the MBD annotation tool; the fault association annotation involves marking historical high-frequency fault information on the three-dimensional model template through causal relationship mapping in the maintenance knowledge graph; and the standard reference annotation involves marking the standard entity number and key clause content on the corresponding component or process node in the form of hyperlinks.

[0086] The model building module is also used to associate the 3D wind turbine model with the maintenance knowledge graph to obtain an identifier mapping table; the specific method of associating the 3D wind turbine model with the maintenance knowledge graph is as follows:

[0087] In the three-dimensional wind turbine model, a unique digital identifier is assigned to each component template, assembly process node, and key geometric feature with enhanced semantic annotation, and a unique entity identifier is established for the core entity in the maintenance knowledge graph.

[0088] The numerical identifiers of the three-dimensional wind turbine model are associated with the entity identifiers of the maintenance knowledge graph to obtain a bidirectional queryable identifier mapping table.

[0089] The main driving parameters, derived parameters, and geometric dimension annotation information of each component template are extracted from the three-dimensional wind turbine model to obtain static attributes. Based on the identifier mapping table, the static attributes are synchronized to the corresponding equipment entities in the maintenance knowledge graph to enrich the attribute description of the equipment entities.

[0090] Extract the fault phenomenon description and occurrence frequency of fault entities, the tool specifications and spare part models of resource entities, and the standard clause content of standard entities from the maintenance knowledge graph to obtain dynamic attributes; synchronize the dynamic attributes to the fault association annotation and standard reference annotation of the corresponding component template to achieve bidirectional flow and integration of the geometric attributes of the three-dimensional model and maintenance domain knowledge;

[0091] Extract all process parameter annotations under the same assembly process node, and associate the process parameter annotations with the corresponding process entities in the maintenance knowledge graph of the assembly process node.

[0092] The visualization assembly / disassembly module is used for rendering and interacting with the 3D wind turbine model, and updates and synchronizes the 3D wind turbine model in real time based on the identifier mapping table and real-time operating data; the specific methods for rendering and interacting with the 3D wind turbine model are as follows:

[0093] Based on the geometric constraint network and the geometric dimension annotation of each component template, computer-aided design software is used to perform feature modeling on the three-dimensional wind turbine model to ensure that the dimensions of each component are accurate and the assembly relationship meets the constraint conditions, thus obtaining the basic three-dimensional model.

[0094] The basic 3D model is imported into a digital content creation tool; based on the digital content creation tool, the complex curved surfaces and appearance details in the basic 3D model are refined and optimized to repair the flaws of the 3D wind turbine model and enhance the visual realism, resulting in an optimized 3D model.

[0095] In the digital content creation tool, a high dynamic range texture map is created to optimize the 3D model; the optimized 3D model is associated with the corresponding texture map, and the associated optimized 3D model and texture map are exported as a GLTF file; the texture map is used to achieve physically based rendering effects, including albedo channel, normal channel, metallicity channel and roughness channel;

[0096] The 3D engine, which integrates WebGL (Web Graphics Library), loads and parses the GLTF file and constructs a 3D visualization scene. Based on the node hierarchy, geometric data, material and texture information in the GLTF file, the 3D engine uses a physically based rendering shader to perform realistic lighting calculations to achieve high-fidelity 3D wind turbine model rendering.

[0097] Specifically, the 3D engine provides multiple view controllers, including a view navigation controller, a model manipulation controller, and an engineering view controller. The view navigation controller is used to realize translation, rotation, and scaling of the 3D visualization scene. The model manipulation controller supports independent operation on selected parts. When a user selects a specific part by clicking or selecting a box, the model manipulation controller allows free transformation of the part by translation and rotation, and the transformation matrix is ​​restricted by the constraint relationship in the geometric constraint network.

[0098] The engineering view controller provides sectional and exploded views of the 3D wind turbine model. The sectional view dynamically trims the 3D wind turbine model through a custom cutting plane to display its internal structure. The exploded view decomposes the components at each level according to the assembly logic of the 3D wind turbine model. The assembly logic is extracted from the assembly process document.

[0099] The specific method for real-time updating and synchronizing the 3D wind turbine model based on the identifier mapping table and real-time operating data is as follows:

[0100] Receive real-time operating data and, based on the identifier mapping table, map the real-time operating data to the corresponding components and their main drive parameters in the three-dimensional wind turbine model;

[0101] Extract the actual values ​​of the main drive parameters and calculate the actual deviation between the actual values ​​of the main drive parameters and the reference values, i.e., the actual value minus the reference value;

[0102] The actual deviation is used as a known input quantity and input into the main driving parameters corresponding to the geometric constraint network, thereby changing the numerical state of the corresponding main driving parameter nodes.

[0103] Based on the constraint relationships defined in the geometric constraint network, starting from the component corresponding to the main driving parameter, and using the actual value of the main driving parameter as the driving source, constraint propagation calculation is performed; the constraint propagation calculation solves and updates the main driving parameters of all component templates associated with the component according to the positioning constraints in the geometric constraint network, and obtains a set of updated main driving parameters;

[0104] The updated main drive parameters are passed back to the corresponding component template, and the 3D wind turbine model is updated in the 3D visualization scene based on the component template.

[0105] The visualization disassembly and assembly module also decomposes and visualizes the disassembly and assembly process based on the maintenance knowledge graph; the specific methods for decomposing and visualizing the disassembly and assembly process based on the maintenance knowledge graph are as follows:

[0106] The maintenance knowledge graph queries all process entities associated with the target wind turbine equipment entity, and constructs a directed graph of disassembly and assembly processes with clear sequential dependencies based on the sequential relationships between process entities; the nodes in the directed graph of disassembly and assembly processes represent process entities, and the edges represent sequential relationships.

[0107] Receive the user's disassembly and assembly instructions; traverse the directed graph of the disassembly and assembly process based on the disassembly and assembly instructions, and locate the starting process node; based on the starting process node and according to the sequential relationship of the directed graph of the disassembly and assembly process, extract a linear and executable process sequence.

[0108] In the three-dimensional visualization scene, according to the identifier mapping table, each process entity in the process sequence is mapped to the corresponding component or assembly process node in the three-dimensional wind turbine model, and the process entity and required resource entity of the current disassembly and assembly step are dynamically displayed in the preset interface area of ​​the three-dimensional visualization scene, such as the side information panel.

[0109] The engineering view controller is driven to visually highlight the set of parts operated in the current process step; if the current process step is disassembly, the exploded view is driven to display the partial or global decomposition; if the current process step is assembly, the assembly process is demonstrated in reverse.

[0110] For any step in the process sequence, the user can pause the guidance interactively and, based on the identifier mapping table, initiate a deep information query for the components involved in the process step, and return the information annotations associated with the maintenance knowledge graph.

[0111] The intelligent management module, based on the maintenance knowledge graph and identifier mapping table, realizes user and permission management, work order and process management, and fault and decision management. The intelligent management module assigns a unique identity to each user and associates it with the user's department, position and skill level information to form a basic user profile. Based on the resource entities and standard entities in the maintenance knowledge graph, permission levels are divided, including basic permissions, intermediate permissions, advanced permissions and training permissions.

[0112] Specifically, the basic permissions are for maintenance personnel, allowing them to access the regular visualization and interactive functions of the 3D wind turbine model, the process sequence corresponding to the current work order, the associated resource entities and standard entities, and to execute disassembly and assembly operation records and upload data; the regular visualization and interactive functions include viewing the translation, rotation, scaling, section view and exploded view of the 3D wind turbine model, as well as viewing the geometric dimension annotations and process dimension annotations corresponding to the selected parts;

[0113] The intermediate-level permissions are for technical experts. In addition to the basic permissions, they include access to the fault diagnosis function, the ability to edit fault entities in the maintenance knowledge graph, and the ability to make suggestions on adjusting process parameters on work orders. They can also view the detailed causes and historical data of fault association annotations in the 3D wind turbine model, as well as the complete content of standard reference annotations.

[0114] The advanced permissions are for administrators, who have full access to functions, including user profile management, permission allocation, work order approval, standard entity updates, data statistical analysis, and other operation permissions. They can also modify the enhanced semantic annotation information of the 3D wind turbine model and the association relationships between core entities in the maintenance knowledge graph.

[0115] The training permissions are for trainees and only allow access to the standard visualization and interactive functions of the 3D visualization scene and 3D wind turbine model, as well as process dimension annotations. They do not have permission to operate actual work orders or modify data, and cannot view unauthorized fault cases or standard entity details.

[0116] The implementation method of the work order and process management is as follows:

[0117] The system receives basic work order information input by the user, including wind turbine model, maintenance type, arrival time, planned completion time, and maintenance scope. Based on the identifier mapping table, it retrieves the corresponding process entities and standard entities from the maintenance knowledge graph and automatically generates an initial work order. The initial work order includes a recommended process sequence, standard working hours for each process, and a list of required resource entities. The user can adjust the content of the initial work order based on the actual site conditions to obtain the final work order.

[0118] Based on the work order, the on-site operation data is synchronized to the corresponding process entity through an identifier mapping table, and the execution status of each process is dynamically displayed in a preset interface area of ​​the 3D visualization scene; the execution status includes not started, in progress, and completed; the on-site operation data includes the completion time of the corresponding process, the actual resources used, etc.

[0119] After all procedures in the work order are completed, the entire process data of the work order is extracted, including the execution records of each procedure, resource consumption list, fault handling records, 3D wind turbine model update log, etc., and a standardized maintenance report is generated. Through an identifier mapping table, the standardized maintenance report is associated and stored with the equipment entity corresponding to the 3D wind turbine model, the relevant fault entity and procedure entity in the maintenance knowledge graph.

[0120] The fault and decision management is implemented as follows:

[0121] Collect abnormal data that exceeds the standard threshold in real-time operation data, such as excessive vibration value or abnormal temperature rise, and map the abnormal data to the corresponding components and fault entities in the maintenance knowledge graph of the three-dimensional wind turbine model through an identifier mapping table;

[0122] Using the faulty entity as the search criteria, equipment entities, process entities, resource entities, and standard entities that have a causal relationship with the faulty entity are extracted from the maintenance knowledge graph. A fault diagnosis report is generated based on the retrieved equipment entities, process entities, resource entities, and standard entities. The fault diagnosis report includes possible causes of the fault, related components, recommended maintenance process sequence, a list of required tools and spare parts, and reference standard clauses.

[0123] In the 3D visualization scene, the components corresponding to abnormal data are highlighted through the engineering view controller, and the implementation steps of the recommended maintenance sequence are dynamically demonstrated. At the same time, the standard reference annotations are associated through the identifier mapping table, allowing users to quickly view the details of relevant standard clauses.

[0124] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0125] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of protection of this application, and these forms are all within the protection scope of this application.

Claims

1. A visual disassembly and assembly platform for boiler fans in thermal power plants based on digital twin technology, characterized in that, It includes a data fusion module, a model building module, a visualization assembly / disassembly module, and an intelligent management module; The data fusion module is used to acquire multi-source heterogeneous data and construct a maintenance knowledge graph based on the multi-source heterogeneous data; The model building module is used to build a 3D wind turbine model and perform enhanced semantic annotation on the 3D wind turbine model based on multi-source heterogeneous data; The model building module is also used to associate the three-dimensional wind turbine model with the maintenance knowledge graph to obtain an identifier mapping table; The visualization disassembly and assembly module is used to render and interact with the 3D wind turbine model, and to update and synchronize the 3D wind turbine model in real time based on the identifier mapping table and multi-source heterogeneous data; the visualization disassembly and assembly module also decomposes and visualizes the disassembly and assembly process based on the maintenance knowledge graph; The intelligent management module is based on a maintenance knowledge graph and an identifier mapping table to manage users and permissions, work orders and processes, and faults and decisions. The model building module is configured with a 3D wind turbine model building strategy for building 3D wind turbine models; The specific strategies for constructing the three-dimensional wind turbine model include: Identify the key areas for maintenance and, based on these key areas and the working principle of the fan, identify the core functional units of the fan; according to the design drawings and bill of materials information of each core functional unit, decompose the corresponding core functional units step by step to the basic component level to obtain different parts; Set the main driving parameters, derived parameters and corresponding calculation rules for each component, and set the baseline value of the main driving parameters to obtain the parameterized component template; The constraint relationships between the parameterized component templates are determined, and a geometric constraint network is constructed based on the constraint relationships. Based on the geometric constraint network, all component templates are organized according to the hierarchical structure of the bill of materials information to obtain a three-dimensional wind turbine model. The geometric constraint network is constructed as follows: Based on the design drawings, obtain the geometric shape and theoretical assembly position of each component template, and analyze the assembly relationship between each component template; Analyze the physical fit type between any two component templates with assembly relationship, and map the physical fit type to the corresponding constraint condition; The constraints include positioning constraints and connection constraints; the positioning constraints include concentric or coaxial constraints, coplanar or fitting constraints, and distance constraints; the connection constraints include bolt connection constraints, pin connection constraints, and key connection constraints. A geometric constraint network is constructed using the component templates as nodes and the constraints between the component templates as edges.

2. The visualized disassembly and assembly platform for boiler fans in thermal power plants based on digital twin technology as described in claim 1, characterized in that, The model building module is configured with an enhanced semantic annotation strategy for performing enhanced semantic annotation on the 3D wind turbine model; The enhanced semantic annotation strategy specifically includes: Obtain the assembly process document and identify the key entities and corresponding process parameters in the assembly process document; By using a similarity matching strategy, a synonym mapping table is established between the key entity and the names and feature terms of each component in the 3D wind turbine model. Based on the synonym mapping table, the process parameters are added to the corresponding component templates. Based on the core functional units and assembly process documents, assembly process nodes are divided, and based on the assembly process nodes, information is labeled on the component templates of the three-dimensional wind turbine model. The information annotation includes geometric dimension annotation, process dimension annotation, fault association annotation, and standard reference annotation.

3. The visualized disassembly and assembly platform for boiler fans in thermal power plants based on digital twin technology as described in claim 2, characterized in that, The visualization assembly / disassembly module is configured with a rendering and interaction strategy for rendering and interacting with the 3D wind turbine model. The rendering interaction strategy specifically includes: Based on the geometric constraint network and the geometric dimension annotation of each component template, computer-aided design software is used to perform feature modeling on the three-dimensional wind turbine model to obtain the basic three-dimensional model. The basic 3D model is imported into a digital content creation tool, and based on the digital content creation tool, the complex surfaces and appearance details in the basic 3D model are refined and optimized to obtain an optimized 3D model. To optimize the 3D model, a high dynamic range texture map is created, and the optimized 3D model is associated with the corresponding texture map to generate a GLTF file; Based on the GLTF file, a 3D visualization scene is constructed using a 3D engine that integrates WebGL.

4. The visualized disassembly and assembly platform for boiler fans in thermal power plants based on digital twin technology as described in claim 3, characterized in that, The visualization disassembly and assembly module is equipped with a 3D wind turbine model update strategy for real-time updating and synchronizing the 3D wind turbine model. The specific update strategy for the three-dimensional wind turbine model includes: Receive real-time operating data and, based on the identifier mapping table, map the real-time operating data to the corresponding components and their main drive parameters in the three-dimensional wind turbine model; Extract the actual values ​​of the main driving parameters and calculate the actual deviation between the actual values ​​of the main driving parameters and the reference values; use the actual deviation as a known input quantity and input it into the main driving parameters corresponding to the geometric constraint network; Based on the constraint relationships defined in the geometric constraint network, starting from the component corresponding to the main driving parameter, and using the actual value of the main driving parameter as the driving source, constraint propagation calculation is performed to obtain the updated main driving parameter. The updated main drive parameters are passed back to the corresponding component template, and the 3D wind turbine model is updated in the 3D visualization scene based on the component template.

5. The visualized disassembly and assembly platform for boiler fans in thermal power plants based on digital twin technology according to claim 4, characterized in that, The visualization disassembly / assembly module is also configured with a disassembly / assembly process decomposition strategy, used to decompose and visualize the disassembly / assembly process; the disassembly / assembly process decomposition strategy specifically includes: All process entities associated with the target wind turbine equipment entity are obtained from the maintenance knowledge graph, and a directed graph of disassembly and assembly processes is constructed based on the sequential relationship between the process entities. Receive the user's disassembly and assembly instructions, and locate the starting process node in the directed graph of the disassembly and assembly process based on the disassembly and assembly instructions; extract the process sequence based on the starting process node and according to the sequential relationship of the directed graph of the disassembly and assembly process. In the three-dimensional visualization scene, based on the identifier mapping table, each process entity in the process sequence is mapped to the corresponding component or assembly process node in the three-dimensional wind turbine model.

6. The visualized disassembly and assembly platform for boiler fans in thermal power plants based on digital twin technology as described in claim 5, characterized in that, The disassembly and assembly process decomposition strategy also includes: In the preset interface area of ​​the 3D visualization scene, the process entities and required resource entities of the current disassembly / assembly step are displayed; The driver engineering view controller visually highlights the parts operated in the current process step; if the current process step is disassembly, the driver explodes the view to display the parts locally or globally. If the current process step is assembly, then demonstrate the assembly process in reverse. For any step in the process sequence, based on the identifier mapping table, a deep information query is performed on the components involved in the process step, and the information annotation associated with the maintenance knowledge graph is returned.

7. The visualized disassembly and assembly platform for boiler fans in thermal power plants based on digital twin technology according to claim 6, characterized in that, The 3D engine is equipped with different view controllers, including a view navigation controller, a model manipulation controller, and an engineering view controller; The view navigation controller can be used to translate, rotate, and zoom the 3D visualization scene. When a user selects a specific component by clicking or selecting a box, the user can translate and rotate the component using the model manipulation controller. The transformation matrix of translation and rotation is restricted by the constraint relationship in the geometric constraint network. The engineering view controller provides sectional and exploded views of the 3D wind turbine model; The sectional view dynamically trims the 3D wind turbine model using a custom sectional plane; the exploded view decomposes the components at each level according to the assembly logic of the 3D wind turbine model.

8. The visualized disassembly and assembly platform for boiler fans in thermal power plants based on digital twin technology according to claim 7, characterized in that, The data fusion module is configured with a maintenance knowledge graph construction strategy; the maintenance knowledge graph construction specifically includes: Named entity recognition is performed on the multi-source heterogeneous data, and core entities are extracted, including equipment entities, fault entities, process entities, resource entities, and standard entities. A deep learning-based relation extraction model is used to extract semantic relationships between core entities from the multi-source heterogeneous data, and the relation types are identified based on the semantic relationships; the relation types include compositional relations, causal relations, sequential relations, dependent relations, and mutually exclusive relations. Using the core entities as nodes and the relationship types between the core entities as edges, and adding data sources to each core entity and edge, a maintenance knowledge graph is obtained.

Images