Wind turbine generator fault diagnosis and repair method based on knowledge graph and related device

By building a knowledge graph in the field of wind turbines, extracting fault characteristics and automatically generating repair solutions, the problems of insufficient generalization ability and interpretability of fault diagnosis in the existing technology are solved, and efficient diagnosis and repair of wind turbine failures are achieved, and complex adaptation is achieved. Working conditions and improve the accuracy of fault detection.

CN118940002APending Publication Date: 2024-11-12HUANENG HEZHANG WIND POWER CO LTD +1
View PDF 0 Cites 4 Cited by

Patent Information

Application Number
CN202411034003.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-30
Publication Date
2024-11-12

AI Technical Summary

Technical Problem

The existing wind turbine fault diagnosis methods are insufficient in generalization capabilities under complex operating conditions, difficult to adapt to changes, and difficult to obtain interpretability and repair measures of diagnostic results, and lack knowledge fusion and synergistic effects.

Method used

Using a knowledge graph-based method, a knowledge graph in the field of wind turbines is constructed. By collecting operating data of key equipment, fault characteristics are extracted, and repair solutions are automatically generated using graph pattern matching and rule reasoning, and the knowledge graph is optimized through active learning and incremental learning mechanisms. .

Benefits of technology

It realizes interpretability diagnosis and operational repair solutions for wind turbine faults, adapts to complex working conditions, improves the accuracy and robustness of fault detection, combines massive data and expert experience, and forms knowledge, data, algorithms and Collaborative closed loop of personnel.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118940002A_ABST
    Figure CN118940002A_ABST
Patent Text Reader

Abstract

The invention discloses a wind turbine generator fault diagnosis and repair method based on a knowledge graph and a related device, and the method comprises the steps: constructing a wind turbine generator domain knowledge graph, and collecting the operation data of key equipment of a wind turbine generator; fault features in the operation data of the key equipment in the wind turbine generator are extracted, fault classification and intelligent diagnosis are carried out on the fault features, and the diagnosed fault types are mapped to the knowledge graph; in the mapped knowledge graph, automatically generating a repair scheme by using graph pattern matching and rule reasoning technologies; and based on the repair scheme, optimizing the knowledge graph by using active learning and incremental learning mechanisms. According to the method, the knowledge graph is constructed, and the knowledge graph can be continuously updated and optimized by utilizing active learning and incremental learning mechanisms so as to adapt to continuously changing working conditions.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the technical field of wind turbine equipment operation and maintenance, and specifically relates to a wind turbine fault diagnosis and repair method based on a knowledge graph and related devices. Background Art

[0002] Wind power generation is a method of generating electricity by converting mechanical energy into electrical energy through a generator. As a clean and renewable form of energy utilization, wind power generation plays an increasingly important role in the adjustment of the global energy structure. With the continuous advancement of wind power technology and the continuous growth of wind power installed capacity, the operation and maintenance of wind farms have become increasingly prominent.

[0003] Wind turbines are composed of key components such as blades, gearboxes, generators, and converters. Wind turbines work in harsh outdoor environments for a long time, and are subject to the influence of external factors such as wind, temperature, and humidity. They are prone to various mechanical and electrical failures, which not only cause the wind turbine to shut down and reduce power generation efficiency, but may also cause safety accidents, resulting in equipment damage and economic losses. Therefore, it is of great significance to monitor the status of wind turbines and realize early diagnosis and rapid repair of faults.

[0004] At present, wind turbine fault diagnosis mainly adopts methods based on signal processing and physical modeling. For example, by analyzing the monitoring signals of wind turbine vibration, temperature and oil quality, a fault feature library is constructed, and fault classification and location are realized by using algorithms such as pattern recognition and machine learning. These methods have achieved certain diagnostic effects under specific working conditions, but there are still some problems and challenges: first, the fault modes of wind turbines are complex and changeable, and the existing diagnostic models have insufficient generalization ability and are difficult to adapt to the changing working conditions; second, the interpretability of the diagnostic results is not strong, and it is difficult to obtain the cause analysis and repair measures of the fault; third, the massive operating data of wind turbines and the expert experience knowledge are scattered and independent, lacking the means of integration, and it is difficult to exert synergistic effects. Therefore, the wind power industry urgently needs a new type of fault diagnosis and repair technology that can integrate the domain knowledge of wind turbines, cope with complex working conditions, and provide interpretable diagnostic results and operable repair solutions. Summary of the invention

[0005] In order to solve one of the above technical defects, the present invention provides a wind turbine fault diagnosis and repair method based on knowledge graph and related devices. In order to achieve the above purpose, the present invention adopts the following technical solutions to achieve it:

[0006] In a first aspect, a wind turbine fault diagnosis and repair method based on a knowledge graph is provided, comprising:

[0007] Build a knowledge graph in the field of wind turbines and collect operating data of key equipment of wind turbines;

[0008] Extracting fault features from the operation data of the key equipment in the wind turbine generator set, classifying the fault features and performing intelligent diagnosis, and mapping the diagnosed fault types to the knowledge graph;

[0009] In the mapped knowledge graph, a repair solution is automatically generated using graph pattern matching and rule reasoning technology;

[0010] Based on the repair solution, the knowledge graph is optimized using active learning and incremental learning mechanisms.

[0011] Furthermore, the construction of the wind turbine domain knowledge graph and the collection of operating data of key equipment of the wind turbine specifically include:

[0012] Define the concepts, attributes and relationships of wind turbine ontology, use PDF triples to represent the multi-dimensional knowledge of wind turbine component composition, operating conditions, failure modes, diagnostic methods and repair measures, and store them in the knowledge graph database. Then use the TransE knowledge representation learning model to embed entities and relationships into a low-dimensional vector space by minimizing the energy function, and mine the implicit semantic associations between knowledge.

[0013] The key equipment includes blades, a gearbox and a generator, and sensors are used to collect the operating data of the blades, the gearbox and the generator in real time.

[0014] Furthermore, the extracting of fault features from the operation data of the key equipment in the wind turbine generator set, classifying the fault features and performing intelligent diagnosis, and mapping the diagnosed fault types to the knowledge graph specifically includes:

[0015] In the collected operation data of the blades, gearboxes and generators, a one-dimensional convolutional neural network is used to extract multi-scale fault features in the operation data;

[0016] The multi-scale fault features are classified and intelligently diagnosed using a support vector machine classification model, the diagnosed fault types are mapped to a knowledge graph, matched and associated with the fault patterns in the knowledge graph, and the semantic ontology descriptions of the diagnosed faults are matched in the knowledge graph.

[0017] Furthermore, the fault types include blade cracking, gear wear and bearing overheating.

[0018] Furthermore, in the mapped knowledge graph, a repair solution is automatically generated using graph pattern matching and rule reasoning technology, specifically including:

[0019] Apply graph pattern matching and rule reasoning techniques on the knowledge graph to explore the possible causes of the diagnosed faults, use the AHP multi-criteria decision-making method to evaluate and optimize the alternative repair measures, and obtain the priority ranking of the repair measures;

[0020] Based on the comprehensive reasoning and optimization results, a repair plan is automatically generated that includes specific implementation steps and resource requirements.

[0021] Furthermore, based on the repair solution, the knowledge graph is optimized using active learning and incremental learning mechanisms, specifically:

[0022] Through active learning strategies, the quality of the knowledge graph is evaluated, and experts are actively asked about the knowledge with low confidence in the graph to expand and update the knowledge graph;

[0023] Develop a knowledge graph visualization interface to present knowledge in the form of an interactive network diagram for users to explore and browse.

[0024] Furthermore, in the process of developing the knowledge graph visualization interface, a mobile APP is added for interaction.

[0025] In the second aspect, a wind turbine fault diagnosis and repair system based on knowledge graph is provided, including:

[0026] The knowledge graph construction module is used to integrate the design parameters, operation data and historical fault multi-source heterogeneous data of wind turbines to build a knowledge graph in the field of wind turbines;

[0027] Data acquisition module, used to collect operating data of key equipment of wind turbines;

[0028] A fault feature extraction module is used to extract fault features from the operating data of key equipment of the wind turbine;

[0029] A fault diagnosis module is used to classify the extracted fault features and perform intelligent diagnosis on the fault type;

[0030] The fault repair decision module is used to infer the cause of the fault and optimize the repair strategy on the knowledge graph, and automatically generate a repair decision plan;

[0031] Knowledge self-learning module, used to continuously accumulate fault knowledge characteristics and perform self-optimization;

[0032] The human-computer interaction module is used to provide knowledge graph visualization, natural language question and answer, diagnostic report generation, and mobile App interaction functions.

[0033] In a third aspect, an electronic device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable in the processor, wherein when the processor executes the computer program, the steps of the wind turbine fault diagnosis and repair method based on the knowledge graph as described in any one of the above items are implemented.

[0034] In a fourth aspect, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the wind turbine fault diagnosis and repair method based on the knowledge graph as described in any one of the above items is implemented.

[0035] Compared with the prior art, the present invention has the following beneficial effects:

[0036] 1. This method constructs a knowledge graph and uses active learning and incremental learning mechanisms to enable the knowledge graph to be continuously updated and optimized to adapt to changing working conditions; secondly, the fault characteristics of wind turbines are classified and intelligently diagnosed, and mapped to the knowledge graph to automatically generate repair plans, thereby enhancing the interpretability of fault diagnosis.

[0037] 2. By defining the concepts, attributes and relationships of wind turbine ontology and using PDF triples to represent them, the multi-dimensional knowledge of wind turbines can be fully integrated. The construction of this knowledge graph makes information query on wind turbines more efficient, and can quickly retrieve various information related to wind turbines, thereby improving decision-making efficiency and accuracy. Secondly, by using the TransE knowledge representation learning model, entities and relationships are embedded into a low-dimensional vector space by minimizing the energy function, and the implicit semantic associations between knowledge can be mined. This association mining helps to discover the potential connections between different components, fault modes, diagnostic methods and repair measures, providing strong support for the intelligent maintenance and management of wind turbines.

[0038] 3. Through the convolution operation, local features in the data can be captured, and through the pooling operation, the dimension of the data can be reduced while retaining the key features. Multi-scale feature extraction means that the network can capture fault modes at different time scales and improve the accuracy and robustness of fault detection. Secondly, the support vector machine classification model is used to classify the extracted multi-scale fault features to achieve intelligent diagnosis. The diagnosed fault type is mapped to the knowledge graph. By mapping the fault type to the knowledge graph, the structured storage and query of the fault information can be achieved, which is convenient for subsequent fault analysis and processing. At the same time, the fault mode in the knowledge graph can provide rich background knowledge and experience support for intelligent diagnosis, improving the accuracy and reliability of diagnosis.

[0039] 4. Through intelligent diagnosis of specific fault types, such as blade cracking, gear wear and bearing overheating, these faults can be identified more accurately, the false alarm rate and missed alarm rate can be reduced, and the targetedness and accuracy of fault detection can be improved.

[0040] 5. With the knowledge graph as the hub, the massive heterogeneous data of wind turbines and the expert experience knowledge are semantically integrated, and machine learning, reasoning decision-making and human-computer interaction technology are connected to form a collaborative closed loop of knowledge, data, algorithms and personnel; in addition, by innovatively introducing active learning and incremental learning mechanisms, the ability of continuous learning and self-optimization is given, and the quality of the knowledge graph can be actively evaluated, uncertain knowledge can be asked from experts and the graph can be dynamically expanded; the fault diagnosis model can also be regularly adjusted and optimized according to new fault cases, so that it can keep pace with the times and adapt to the dynamic evolution of wind turbine fault characteristics.

[0041] 6. Through the knowledge graph visualization interface, you can intuitively explore wind power field knowledge, use natural language to conduct fault diagnosis questions and answers with the system, automatically generate standardized diagnostic reports, and provide mobile APP to support remote diagnosis and real-time monitoring.

[0042] 7. This system can cope with complex working conditions, quickly and accurately discover and locate wind turbine faults, explore the inherent causes of the faults, and provide an operational optimal repair plan. This system will play an important role in improving the availability of wind turbines and ensuring the safe and economical operation of wind farms. It also establishes a new technical paradigm for intelligent operation and maintenance in the field of new energy. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0044] Figure 1 A flow chart of the wind turbine fault diagnosis and repair method based on knowledge graph provided by the present invention;

[0045] Figure 2 Schematic diagram of the wind turbine fault diagnosis and repair system based on knowledge graph provided by the present invention.

[0046] Figure 3 A schematic diagram of the category hierarchy of the wind turbine generator set body in the wind turbine generator set fault diagnosis and repair system based on the knowledge graph provided by the present invention;

[0047] Figure 4A schematic diagram of a network structure for extracting fault features in a wind turbine fault diagnosis and repair system based on a knowledge graph provided by the present invention;

[0048] Figure 5 A schematic diagram of blade cracking fault cause matching in the wind turbine fault diagnosis and repair system based on the knowledge graph provided by the present invention. DETAILED DESCRIPTION

[0049] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0050] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0051] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, further definition and explanation thereof is not required in subsequent drawings.

[0052] In the description of the embodiments of the present invention, it should be noted that if the terms "upper", "lower", "horizontal", "inner", etc. indicate an orientation or positional relationship based on the orientation or positional relationship shown in the drawings, or the orientation or positional relationship in which the product of the invention is usually placed when in use, it is only for the convenience of describing the present invention and simplifying the description, and does not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention. In addition, the terms "first", "second", etc. are only used to distinguish the description, and cannot be understood as indicating or implying relative importance.

[0053] In addition, if the term "horizontal" appears, it does not mean that the component must be absolutely horizontal, but can be slightly tilted. For example, "horizontal" only means that its direction is more horizontal than "vertical", which does not mean that the structure must be completely horizontal, but can be slightly tilted.

[0054] In the description of the embodiments of the present invention, it is also necessary to explain that, unless otherwise clearly specified and limited, the terms "set", "install", "connect", and "connect" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal connection of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0055] The present invention is further described in detail below in conjunction with the accompanying drawings:

[0056] The embodiment of the present invention provides a wind turbine fault diagnosis and repair method based on knowledge graph and related devices.

[0057] like Figure 1 As shown, in the first aspect, an embodiment of the present invention provides a wind turbine fault diagnosis and repair method based on a knowledge graph, comprising:

[0058] S101. Construct a knowledge graph in the field of wind turbines and collect operating data of key equipment of wind turbines. For example, before diagnosing and repairing wind turbine faults, first construct a knowledge graph in the field of wind turbines, define the ontological concepts and attributes of wind turbines and the relationships between the various components of wind turbines, use PDF triples to represent the multi-dimensional knowledge of wind turbine component composition, operating conditions, fault modes, diagnostic methods and repair measures, and store them in a knowledge graph database. Then, use the TransE knowledge representation learning model to embed entities and relationships into a low-dimensional vector space by minimizing the energy function, and mine implicit semantic associations between knowledge. Among them, the key equipment of wind turbines includes blades, gearboxes and generators, and sensors are used to collect the operating data of blades, gearboxes and generators in real time.

[0059] S102. Extract fault features from the operating data of the key equipment in the wind turbine generator set, classify the fault features and perform intelligent diagnosis, and map the diagnosed fault types to the knowledge graph; illustratively, in the collected operating data of the blades, gearboxes and generators, a one-dimensional convolutional neural network is used to extract multi-scale fault features in the operating data to capture local patterns and high-order semantics in the operating data; then, a support vector machine classification model is used to classify the multi-scale fault features and perform intelligent diagnosis, and the fault types include blade cracking, gear wear and bearing overheating; the diagnosed fault type is mapped to the knowledge graph, matched and associated with the fault pattern in the knowledge graph, and the semantic ontology description of the diagnosed fault is matched in the knowledge graph.

[0060] S103. In the mapped knowledge graph, graph pattern matching and rule reasoning techniques are used to automatically generate a repair plan. Exemplarily, graph pattern matching and rule reasoning techniques are applied to the knowledge graph to explore the possible causes of the diagnosed faults, and the AHP multi-criteria decision-making method is used to evaluate and optimize the alternative repair measures to obtain a priority ranking of the repair measures. Finally, the reasoning and optimization results are integrated to automatically generate a repair plan including specific implementation steps and resource requirements.

[0061] S104. Based on the repair scheme, the knowledge graph is optimized using active learning and incremental learning mechanisms. Exemplarily, the knowledge graph quality is evaluated through active learning strategies, and experts are actively asked about knowledge with low confidence in the graph to expand and update the knowledge graph; in addition, the fault diagnosis model is incrementally trained regularly based on newly collected fault case data to adapt it to the dynamic changes in the distribution of fault data. In the process of developing a knowledge graph visualization interface, a knowledge graph visualization interface is developed to present knowledge in the form of an interactive network diagram for users to explore and browse. Finally, natural language processing technology is introduced to allow users to use everyday language for fault diagnosis questions and answers, automatically generate standardized reports based on the diagnosis results, and add a mobile APP for interaction to achieve remote diagnosis and real-time monitoring.

[0062] This method constructs a knowledge graph and uses active learning and incremental learning mechanisms to enable the knowledge graph to be continuously updated and optimized to adapt to changing working conditions. Secondly, the fault characteristics of wind turbines are classified and intelligently diagnosed, and mapped to the knowledge graph to automatically generate repair plans, thereby enhancing the interpretability of fault diagnosis.

[0063] The following is an example of wind turbine blade cracking.

[0064] First, a knowledge graph in the field of wind turbine blades was constructed, and core concepts such as blade components, materials, working conditions, and faults were defined, as well as key relationships such as fault occurrence, signs, causes, diagnosis, and repair. Then, heterogeneous data such as blade fault cases, sensor data, and design parameters were semantically mapped and fused to form a high-quality knowledge base for blade cracking faults. During fault diagnosis, 1D-CNN was used to extract the multi-scale time-frequency features of blade vibration signals to characterize the abnormal vibration mode caused by blade cracking. SVM was then used to classify the features, determine the severity of blade cracking faults, and match and associate the diagnosed fault types with the fault concepts in the knowledge graph. When making fault repair decisions, graph pattern matching was applied to the knowledge graph to find possible causes of blade cracking, such as material defects, fatigue fracture, overload damage, etc., and then AHP was used to perform multi-criteria evaluation on common repair measures such as crack repair and blade replacement to obtain the priority ranking of repair measures, and finally the optimal repair solution for blade cracking faults was given.

[0065] like Figure 2 As shown, in the second aspect, an embodiment of the present invention provides a wind turbine fault diagnosis and repair system based on a knowledge graph, comprising:

[0066] The knowledge graph construction module is used to integrate the design parameters, operation data and historical fault multi-source heterogeneous data of wind turbines to construct a knowledge graph in the field of wind turbines; the knowledge graph construction module includes functional units such as ontology definition, knowledge fusion, knowledge storage and knowledge representation learning.

[0067] like Figure 3 As shown, the ontology definition unit is used to define the wind turbine domain ontology, standardize core concepts, attributes and relationships, and provide a partial category hierarchy of the wind turbine ontology.

[0068] The knowledge fusion unit is used to map multi-source heterogeneous data into the ontology to realize the semantic integration of knowledge. The fused knowledge is represented in the form of RDF triples, such as (blade_1, material, fiberglass), (blade_1, cutting_length, 5m), etc.

[0069] The knowledge storage unit is used to import knowledge triples into the graph database Neo4j to build a wind turbine knowledge graph. Neo4j supports graphical storage, query and reasoning of knowledge.

[0070] The knowledge representation learning unit uses knowledge representation learning models such as TransE to learn low-dimensional vector representations of entities and relationships in the knowledge graph by minimizing the energy function:

[0071]

[0072] Among them, S is the set of all correct triplets (h, r, t) in the knowledge graph, γ is the regularization hyperparameter, and d is the L1 norm or L2 norm. Through embedding representation, the implicit semantic associations between entities can be mined.

[0073] Data acquisition module, used to collect operating data of key equipment of wind turbines;

[0074] The fault feature extraction module is used to extract fault features from the operating data of key equipment of wind turbines. The fault feature extraction module includes a feature extraction unit, which uses a one-dimensional convolutional neural network (1D-CNN) to extract multi-scale fault features from the time series data (such as vibration, temperature, etc.) of wind turbines. 1D-CNN automatically learns local patterns and high-order semantic features through convolution and pooling operations. Figure 4 The network structure of 1D-CNN is shown in Figure 2.

[0075] The fault diagnosis module is used to classify the extracted fault features and perform intelligent diagnosis of the fault type; the fault diagnosis module includes a fault classification unit and an ontology matching unit. The fault classification unit uses a classification model such as a support vector machine (SVM) to diagnose the fault type according to the extracted fault features. SVM separates different types of fault samples in a high-dimensional feature space by finding the optimal classification hyperplane. Given an input feature vector x, the classification function of SVM is:

[0076]

[0077] Among them, x i is the support vector, y i Its corresponding class label, α i is the Lagrange multiplier, κ is the kernel function, and b is the bias term.

[0078] The ontology matching unit matches the diagnosed fault type to the corresponding ontology concept in the knowledge graph, and associates and queries the related knowledge such as the attributes, causes, symptoms, etc. of the fault type.

[0079] The fault repair decision module is used to perform fault cause reasoning and repair strategy optimization on the knowledge graph, and automatically generate repair decision plans; the fault repair decision module includes a cause reasoning unit, a strategy optimization unit and a plan generation unit.

[0080] The cause reasoning unit uses graph pattern matching, rule reasoning and other technologies on the knowledge graph to find possible causes of the diagnosed faults. For example, for a blade cracking fault, graph pattern matching can be used to find that its cause may be "material defect", "fatigue accumulation", etc. Figure 5 shown.

[0081] The strategy optimization unit uses multi-criteria decision-making methods such as the analytic hierarchy process (AHP) to comprehensively evaluate and optimize the alternative repair measures. AHP first constructs a hierarchical model of the decision problem, and then calculates the criterion weights and the scheme ranking by pairwise comparison. Suppose there are m evaluation criteria and n alternative schemes, then the ranking value of the jth scheme under the i-th criterion is:

[0082]

[0083] Among them, a ij is the importance score of the jth solution under the i-th criterion, a kj is the score of the jth solution under the kth criterion. Finally, the priority vector of the repair measures is obtained by combining the ranking values ​​of each criterion:

[0084] W j =A·w j

[0085] Where A is the criterion weight vector, w j =(w 1j ,…,w mj ) T is the ranking value vector of the jth solution. j The larger the value, the higher the comprehensive evaluation of plan j and the higher the priority.

[0086] The solution generation unit automatically generates a fault repair plan based on the reasoning and optimization results, including specific repair measures, implementation steps, required resources, etc.

[0087] The knowledge self-learning module is used to continuously accumulate fault knowledge features and perform self-optimization; the knowledge self-learning module includes an active learning unit and an incremental learning unit.

[0088] The active learning unit evaluates the quality of the knowledge graph, proactively discovers areas with low confidence in the graph, and asks experts for relevant knowledge to expand and update the knowledge graph. The quality of the knowledge graph can be measured by the confidence score of the knowledge:

[0089] confidence(h,r,t)=σ(γ-d(h+r,t))

[0090] Among them, σ is the Sigmoid function, (h, r, t) is the knowledge triplet, and for the triplet whose confidence is lower than the threshold, the system will generate corresponding questions and ask experts to confirm or correct it.

[0091] The incremental learning unit regularly conducts incremental training on the existing machine learning model based on the newly collected fault case data of the wind farm to adapt it to changes in data distribution. Incremental learning fine-tunes the original model parameters by optimizing the loss function of the model on new data:

[0092]

[0093] Among them, θ is the model parameter, D new is a new fault case dataset, L is the loss function, f θ is the machine learning model, and λ is the regularization coefficient. Through incremental learning, the diagnosis model can keep pace with the times and adapt to the dynamic changes of wind turbine fault characteristics.

[0094] The human-computer interaction module is used to provide knowledge graph visualization, natural language question and answer, diagnosis report generation, and mobile App interaction functions. The human-computer interaction module includes a knowledge graph visualization unit, a natural language question and answer unit, a diagnosis report generation unit, and a mobile App unit.

[0095] The knowledge graph visualization unit uses visualization libraries such as D3.js to present the wind turbine knowledge graph in the form of an interactive network diagram. Users can explore knowledge by clicking, dragging, zooming, and other operations.

[0096] The natural language question answering unit introduces natural language processing and semantic analysis technology, allowing users to ask questions related to fault diagnosis using everyday language. The system generates answers through knowledge graph query and reasoning.

[0097] The diagnostic report generation unit automatically generates a standardized fault diagnosis report based on the fault diagnosis results, which includes fault overview, fault cause analysis, and repair measures suggestions.

[0098] The mobile App unit develops mobile applications for wind turbine fault diagnosis, realizing functions such as remote diagnosis, real-time monitoring, and knowledge query on mobile devices.

[0099] The wind turbine fault diagnosis and repair system driven by the above knowledge graph realizes the semantic representation and fusion of fault knowledge, intelligent diagnosis and repair decision-making, autonomous learning and updating, human-computer collaborative interaction and other functions. With the help of knowledge graphs to connect heterogeneous data and model algorithms, this system provides a new paradigm and idea for the intelligent operation and maintenance of wind turbines.

[0100] In a third aspect, an embodiment of the present invention provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable in the processor, wherein the processor implements the steps of the wind turbine fault diagnosis and repair method based on the knowledge graph as described above when executing the computer program.

[0101] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the wind turbine fault diagnosis and repair method based on the knowledge graph as described above is implemented.

[0102] The above embodiments are preferred implementation modes of the present invention, but the implementation modes of the present invention are not limited to the above embodiments. Any other changes, modifications, substitutions, combinations, and simplifications that do not deviate from the spirit and principles of the present invention should be equivalent replacement methods and are included in the protection scope of the present invention.

[0103] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented 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.

[0104] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0105] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0106] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0107] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A wind turbine fault diagnosis and repair method based on knowledge graph, characterized in that: include: Build a knowledge graph in the field of wind turbines and collect operating data of key equipment of wind turbines; Extracting fault features from the operation data of the key equipment in the wind turbine generator set, classifying the fault features and performing intelligent diagnosis, and mapping the diagnosed fault types to the knowledge graph; In the mapped knowledge graph, a repair solution is automatically generated using graph pattern matching and rule reasoning technology; Based on the repair solution, the knowledge graph is optimized using active learning and incremental learning mechanisms.

2. The diagnosis and repair method according to claim 1, characterized in that: The construction of the wind turbine domain knowledge graph and the collection of operating data of key equipment of the wind turbine specifically include: Define the concepts, attributes and relationships of wind turbine ontology, use PDF triples to represent the multi-dimensional knowledge of wind turbine component composition, operating conditions, failure modes, diagnostic methods and repair measures, and store them in the knowledge graph database. Then use the TransE knowledge representation learning model to embed entities and relationships into a low-dimensional vector space by minimizing the energy function, and mine the implicit semantic associations between knowledge. The key equipment includes blades, a gearbox and a generator, and sensors are used to collect the operating data of the blades, the gearbox and the generator in real time.

3. The diagnosis and repair method according to claim 1, characterized in that: The extracting of fault features from the operation data of the key equipment in the wind turbine generator set, classifying the fault features and performing intelligent diagnosis, and mapping the diagnosed fault types to the knowledge graph specifically includes: In the collected operation data of the blades, gearboxes and generators, a one-dimensional convolutional neural network is used to extract multi-scale fault features in the operation data; The multi-scale fault features are classified and intelligently diagnosed using a support vector machine classification model, the diagnosed fault types are mapped to a knowledge graph, matched and associated with the fault patterns in the knowledge graph, and the semantic ontology descriptions of the diagnosed faults are matched in the knowledge graph.

4. The diagnosis and repair method according to claim 3, characterized in that: The types of failures described include blade cracking, gear wear and bearing overheating.

5. The diagnosis and repair method according to claim 1, characterized in that: In the mapped knowledge graph, graph pattern matching and rule reasoning techniques are used to automatically generate repair solutions, including: Apply graph pattern matching and rule reasoning techniques on the knowledge graph to explore the possible causes of the diagnosed faults, use the AHP multi-criteria decision-making method to evaluate and optimize the alternative repair measures, and obtain the priority ranking of the repair measures; Based on the comprehensive reasoning and optimization results, a repair plan is automatically generated that includes specific implementation steps and resource requirements.

6. The diagnosis and repair method according to claim 1, characterized in that: Based on the repair solution, the knowledge graph is optimized using active learning and incremental learning mechanisms, specifically: Through active learning strategies, the quality of the knowledge graph is evaluated, and experts are actively asked about the knowledge with low confidence in the graph to expand and update the knowledge graph; Develop a knowledge graph visualization interface to present knowledge in the form of an interactive network diagram for users to explore and browse.

7. The diagnosis and repair method according to claim 6, characterized in that: In the process of developing the knowledge graph visualization interface, a mobile APP is added for interaction.

8. A wind turbine fault diagnosis and repair system based on knowledge graph, characterized in that: include: The knowledge graph construction module is used to integrate the design parameters, operation data and historical fault multi-source heterogeneous data of wind turbines to build a knowledge graph in the field of wind turbines; Data acquisition module, used to collect operating data of key equipment of wind turbines; A fault feature extraction module is used to extract fault features from the operating data of key equipment of the wind turbine; A fault diagnosis module is used to classify the extracted fault features and perform intelligent diagnosis on the fault type; The fault repair decision module is used to infer the cause of the fault and optimize the repair strategy on the knowledge graph, and automatically generate a repair decision plan; Knowledge self-learning module, used to continuously accumulate fault knowledge characteristics and perform self-optimization; The human-computer interaction module is used to provide knowledge graph visualization, natural language question and answer, diagnostic report generation, and mobile App interaction functions.

9. An electronic device, characterized in that: The method comprises a memory, a processor and a computer program stored in the memory and executable in the processor, wherein the processor implements the steps of the wind turbine fault diagnosis and repair method based on knowledge graph as described in any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the wind turbine fault diagnosis and repair method based on knowledge graph according to any one of claims 1 to 7 is implemented.

Citation Information

Cited By

  • Fault diagnosis method for drive motor

    CN119199527A

  • Forest fire prevention method based on smoke recognition

    CN119811055A

  • Power equipment health state monitoring method based on multiple modes

    CN120277594A

  • Knowledge graph construction method based on active learning and incremental learning

    CN121351972A