Fault diagnosis and knowledge elastic index arrangement method and system of scalable graph structure

Abstract

The application discloses a scalable graph structure fault diagnosis and knowledge elastic index arrangement method and system, comprising the following steps: according to a preset rule, a low-code interactive interface is designed based on an open-source graph database to construct a scalable graph structure; according to current fault information, a corresponding fault diagnosis algorithm model is extracted by using the scalable graph structure to perform diagnosis and obtain a current fault type; a diagnosis result of the current fault type is queried in a preset graph database; and according to user feedback on the diagnosis result, the scalable graph structure is updated by using a preset algorithm to complete dynamic index arrangement of a preset recommended diagnosis result. The scalable graph structure is constructed in a low-code semi-automatic mode, the fault diagnosis algorithm knowledge is organized and managed in the form of the graph structure and stored in the graph database, fault query and corresponding diagnosis algorithm recommendation are facilitated, fault diagnosis analysis is assisted, the graph structure is scaled according to user feedback, and the recommended result is dynamically arranged.

Description

Technical Field

[0001] This invention relates to the field of fault diagnosis technology, specifically to a method and system for fault diagnosis and flexible knowledge indexing and arrangement using a scalable graph structure. Background Technology

[0002] Traditional fault diagnosis methods for mechanical equipment primarily rely on the expertise and maintenance experience of specialists. This involves inspecting various aspects of the machine to identify and repair the cause of the fault. However, this expert-driven approach is labor-intensive, time-consuming, and susceptible to subjective biases. With the development of technologies such as the Internet of Things (IoT) and artificial intelligence (AI), data-driven diagnostic methods have become widely used. These methods utilize various state data monitored by sensors to train machine learning algorithms, analyzing the equipment's health status through tasks like classification and regression to identify potential fault types. Examples include using CNN models to analyze visual images to detect surface wear; using VAE models to diagnose fault types in rolling bearings and gears; and employing LSTM models to identify vibration signals from rolling bearing faults. Because data-driven diagnostic methods are not specific to any particular object, offer high diagnostic accuracy, and are convenient and fast to operate, they have achieved a degree of automation and intelligent diagnosis. However, in practice, fault diagnosis algorithms are highly specific; algorithms trained for specific diagnostic problems are only applicable to the analysis of specific types of signal data from specific components.

[0003] Existing technologies utilize knowledge graphs and natural language processing to research and design methods for constructing auxiliary decision-making mechanisms for equipment fault diagnosis in specific domains. These methods achieve fault alarm information parsing and intelligent auxiliary decision-making based on knowledge graphs. Typically, these schemes first construct a knowledge graph of related factors such as equipment fault phenomena, fault types, and underlying causes. Then, they use descriptions of actual equipment fault phenomena to search and match within the fault knowledge graph, thereby assisting in fault type diagnosis. While these schemes can intuitively display the potential relationships between fault phenomena, fault types, and equipment components, knowledge extraction during knowledge graph construction is often difficult. The accuracy of graph diagnosis constructed through automatic entity and relation extraction is not high, and in some cases, the fault log text may lack certain knowledge, preventing the comprehensive extraction of entities and relationships. Furthermore, the implementation of these methods requires training an NLP network model, demanding high coding skills from the designers. Therefore, establishing a scalable graph-structured fault diagnosis knowledge indexing and arrangement method is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0004] Therefore, the present invention provides a scalable graph structure fault diagnosis and knowledge elastic indexing and arrangement method and system, which overcomes the defect of excessive reliance on professional personnel in the fault diagnosis process of equipment in the prior art.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] In a first aspect, embodiments of the present invention provide a method for fault diagnosis and flexible knowledge indexing and arrangement of scalable graph structures, including:

[0007] Based on preset rules, a low-code interactive interface is designed using an open-source graph database to build a scalable graph structure;

[0008] Based on the current fault information, a scalable graph structure is used to extract the corresponding fault diagnosis algorithm model for diagnosis, and the current fault type is obtained.

[0009] Query the diagnostic results for the current fault type in the preset diagram database;

[0010] Based on user feedback on the diagnostic results, the scalable graph structure is updated using a preset algorithm to complete the dynamic indexing and arrangement of preset recommended diagnostic results.

[0011] Optionally, the initial screen of the interactive interface is a node connection diagram of the device and its components. Nodes represent devices or connecting components. Corresponding fault images and text descriptions are uploaded to the nodes to display each component and its connection relationship. Nodes can be added, deleted, and modified through the interactive interface. The content of the operation object involved in the operation is used as the parameter to be filled in through pre-written GQL statements. The interactive interface is used to guide the user to input information. The user needs to focus on the content input of the node and connection relationship without writing GQL code.

[0012] Optionally, the low-code interface for open-source graph databases includes:

[0013] Add nodes for devices and their components to the initial interface. Enter the name and description of the device in the device node and the name and description of the component in the component node to display the connection relationship between the devices and components.

[0014] Select the node of the preset component, and follow the prompts to fill in the preset fault type, the corresponding phenomenon description, the preset cause of the fault, the solution, the data type of the sensor signal, and the fault diagnosis algorithm model in the pop-up sub-node option input field.

[0015] Optionally, the preset rules include: the connection relationship between faulty equipment and components, the relationship between faulty components and sensor signal data types, the relationship between algorithm models and fault phenomena, the relationship between algorithm models and data types, the relationship between fault types and fault phenomena, the relationship between fault types and preset causes, and the relationship between fault types and solutions.

[0016] Optionally, the connection relationship between the faulty device and the component, and the relationship between the faulty component and the data type of the sensor signal are both matching correspondences.

[0017] Optionally, based on the current fault information, a corresponding fault diagnosis algorithm model is extracted using a scalable graph structure for diagnosis, including:

[0018] Users can use a scalable graph structure to query and match the corresponding fault diagnosis algorithm model based on the current fault description information. The description information includes: fault location, fault phenomenon, and sensor data type.

[0019] Calculate the semantic similarity between the fault phenomena in the graph structure and the description of the current input fault phenomenon.

[0020] Determine whether the fault phenomenon and the corresponding algorithm model match;

[0021] Based on whether there is a match, the corresponding fault diagnosis algorithm model is extracted by multiplying the semantic similarity and the matching confidence.

[0022] Optionally, based on user feedback on the diagnostic results, a preset algorithm is used to update the scalable graph structure to complete the dynamic indexing and arrangement of preset recommended diagnostic results, including:

[0023] Based on the diagnostic results of the current fault type retrieved from the graph database, the diagnostic results include: the cause of the fault and the solution;

[0024] The diagnostic results are scored based on their accuracy in relation to the algorithm diagnostic model's recommendations, the degree of consistency with the causes, and the accuracy of the solutions.

[0025] Based on the preset scoring rules, the confidence level is updated, and the dynamic indexing and arrangement of the preset recommended diagnostic results are completed.

[0026] Secondly, embodiments of the present invention provide a fault diagnosis and knowledge elastic indexing and orchestration system with a scalable graph structure, comprising:

[0027] The scalable graph structure building module is used to design low-code interactive interfaces and build scalable graph structures based on open-source graph databases according to preset rules.

[0028] The fault type diagnosis module is used to extract the corresponding fault diagnosis algorithm model based on the current fault information using a scalable graph structure to diagnose the current fault type.

[0029] The diagnostic result query module is used to query the diagnostic results of the current fault type in the preset diagram database.

[0030] The indexing module is used to update the scalable graph structure based on user feedback on diagnostic results using a preset algorithm, thereby completing the dynamic indexing and arrangement of preset recommended diagnostic results.

[0031] Thirdly, embodiments of the present invention provide a terminal, including: at least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to cause the at least one processor to execute the fault diagnosis and knowledge elastic indexing and arrangement method of the scalable graph structure described in the first aspect of the present invention.

[0032] Fourthly, embodiments of the present invention provide a computer-readable storage medium storing computer instructions for causing the computer to execute the fault diagnosis and knowledge elastic indexing and arrangement method for scalable graph structures described in the first aspect of the present invention.

[0033] The technical solution of this invention has the following advantages:

[0034] 1. The fault diagnosis and knowledge elastic indexing and arrangement method and system for scalable graph structures provided by the present invention guides professionals to input fault diagnosis algorithm knowledge through a low-code interactive interface and stores it in a graph database, shielding users from writing GQL code. The guided interactive interface can improve the convenience and efficiency of inputting information when building scalable graph structures.

[0035] 2. The interactive interface provided by this invention guides professionals to semi-automatically input entities and relationships to construct a scalable graph structure. The expression of knowledge can be more targeted and complete, which helps to solve the problem of low accuracy in graph diagnosis when knowledge entities are missing or incompletely extracted.

[0036] 3. Based on user feedback, knowledge is supplemented and reduced, and the confidence level of the correlation can be adaptively updated and optimized according to user ratings, realizing dynamic arrangement of recommendation results and improving the personalization of recommendations. Attached Figure Description

[0037] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0038] Figure 1 A flowchart illustrating a specific example of a fault diagnosis and knowledge elastic indexing and arrangement method for a scalable graph structure provided in an embodiment of the present invention;

[0039] Figure 2This is a schematic diagram illustrating a specific example of building a scalable graph structure through a low-code interactive interface, as provided in an embodiment of the present invention.

[0040] Figure 3 This is a flowchart of the interactive interface operation provided in an embodiment of the present invention;

[0041] Figure 4 A schematic diagram of a knowledge entry interface provided in an embodiment of the present invention;

[0042] Figure 5 A flowchart for querying the current fault status and recommending a suitable algorithm model, provided in an embodiment of the present invention;

[0043] Figure 6 This is a schematic diagram illustrating an implementation example of the structural scaling and dynamic orchestration recommendation results provided in this invention.

[0044] Figure 7 This is a module composition diagram of a scalable graph structure fault diagnosis and knowledge elastic indexing and orchestration system provided in an embodiment of the present invention;

[0045] Figure 8 This is a composition diagram of a specific example of a terminal provided in an embodiment of the present invention. Detailed Implementation

[0046] The technical solution of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0047] In the description of this invention, it should be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing the invention and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0048] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can also refer to the internal connection of two components; and they can refer to a wireless connection or a wired connection. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0049] Furthermore, the technical features involved in the different embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0050] Example 1

[0051] This invention provides a scalable graph structure-based fault diagnosis and knowledge elastic indexing and arrangement method. It organizes and manages various knowledge such as fault phenomena, faulty components, fault types, and diagnostic algorithms in equipment fault diagnosis in a graph structure and stores them in a graph database in a low-code semi-automatic manner. This facilitates fault querying and recommendation of corresponding diagnostic algorithms to assist in fault diagnosis and analysis. Furthermore, it dynamically arranges the graph structure and recommendation results based on user feedback.

[0052] like Figure 1 As shown, it includes the following steps:

[0053] Step S1: Based on preset rules, design a low-code interactive interface and build a scalable graph structure using an open-source graph database.

[0054] In this embodiment of the invention, the initial screen of the interactive interface is a node connection diagram of the device and its components. Nodes represent devices or connecting components. Corresponding fault images and text descriptions are uploaded to the nodes to display each component and its connection relationship. Nodes can be added, deleted, and modified through the interactive interface.

[0055] In this embodiment of the invention, the low-code interactive interface of the open-source graph database includes: adding nodes of devices and their components on the initial interface; entering the name and description of the device in the device node; entering the name and description of the component in the component node; and displaying the connection relationship between the devices and components. Selecting a node of a preset component allows users to follow the guidance prompts and fill in the preset fault type, the corresponding phenomenon description, the preset cause of the fault, the solution, the data type of the sensor signal, and the fault diagnosis algorithm model in the pop-up child node option input field.

[0056] In this embodiment of the invention, the preset rules include: the connection relationship between faulty devices and components, the relationship between faulty components and sensor signal data types, the relationship between algorithm models and fault phenomena, the relationship between algorithm models and data types, the relationship between fault types and fault phenomena, the relationship between fault types and preset causes, and the relationship between fault types and solutions. Among these, the connection relationship between faulty devices and components, and the relationship between faulty components and sensor signal data types, are all matching and corresponding relationships.

[0057] In this embodiment of the invention, addressing the issues of insufficient knowledge management and difficulty in constructing knowledge graphs in current intelligent fault diagnosis of equipment, an interactive interface for operating graph databases in a low-code manner is designed. The interface shields users from writing GQL code and provides visual aids and prompts to guide users in filling in relevant content. Commonly used GQL statements for graph construction, querying, and modification are pre-written. The content of the objects involved in the operation is used as parameters to be filled in. A guided UI interface based on visual components receives the input parameters, assembles them into complete statements, and then executes them.

[0058] Users only need to focus on the content and operational logic of entities and relationships, without having to write GQL code. They can construct graph structures based on information such as nodes, relationships, attributes, and labels in the guided input graph structure, improving the convenience and efficiency of information entry when constructing the graph.

[0059] After the graph is created, users can input the objects to be queried through the interactive interface. These objects are then passed as parameters to the graph database for searching, resulting in visual results returned in tabular or graphical form. Furthermore, because pre-written GQL statements for modification operations are provided, entities and relationships can be modified and expanded during subsequent use to support the scaling of the graph structure.

[0060] In this embodiment, Neo4j is chosen as the graph database. It uses a graph structure of nodes and edges to represent data. Each node and edge has a unique identifier and a set of attributes, which can be basic types, arrays, or custom types. It also uses memory-based indexes and CQL, a query language specifically designed for graph databases, improving query efficiency and flexibility. CQL uses a graphical representation to query data, providing a simple way to query complex relationships. Furthermore, Neo4j provides JavaScript support for UI MVC frameworks and supports both the Cypher API and the Native Java API for developing Java applications.

[0061] like Figure 2 As shown, this is a schematic diagram illustrating how a professional can construct a scalable graph structure using the low-code interactive interface proposed in this embodiment of the invention. The professional is someone familiar with open-source graph databases.

[0062] The design concept of the low-code interactive interface is as follows: First, write CQL code to perform common CRUD functions or call its API to execute CQL. Use the information such as entities, relationships, attributes, tags and query conditions involved as parameters to be filled in. Receive the corresponding content entered by the user in the interactive interface and pass it to the fields to be filled in to form a complete CQL statement. Then run the statement to realize the functions of graph construction and query.

[0063] The UI can be developed based on Neo4j visualization components such as Neo4j Browser, a web-based Neo4j visualization tool that supports interactive queries and visualizations. The front-end can be developed using the JavaScript API and Cypher language provided by Neo4j Browser and embedded into the UI. The UI accesses Neo4j Browser's functionality through the JavaScript API, such as querying data, adding nodes, and setting relationships, and uses Neo4j Browser's visualization features to obtain visualization results in graph and table formats. Furthermore, other Neo4j visualization components that support UI development can also be considered, such as Neo4j Bloom, APOC, and the vis.js library.

[0064] In addition to basic components such as parameter input boxes and a run button, the UI also incorporates guided prompts to assist users in inputting relevant information. For example, the initial screen displays a node connection diagram of the device and common faulty components. Nodes represent the device and its connected components, allowing users to upload corresponding images and text descriptions. The interface directly displays each component and its connections, and users can add, delete, and modify nodes. Then, when entering a fault description for a component, selecting a node will bring up several sub-node options. Each sub-node option has a label and description indicating the required information (such as fault phenomenon, sensor signal data type, diagnostic algorithm model, fault type, and recommended solution). The UI collects the entered information and passes it as parameters to the corresponding Cypher statements to construct the graph structure of algorithmic knowledge.

[0065] In one specific embodiment, such as Figure 3 The example interactive interface operation flow is shown in this embodiment.

[0066] Guided by the interactive interface, professionals input relevant entities and relationships to construct a graph structure of fault diagnosis algorithm knowledge. Entity descriptions are input as entity nodes, and the connections between entities are input as relationships (edges of the graph structure). These are then input into Neo4j, and the corresponding CQL statements are executed to construct the graph structure of fault diagnosis algorithm knowledge and store it in a graph database. Through the interactive interface, professionals do not need to consider the CQL code for manipulating the graph database; they only need to focus on the entity descriptions and the logical relationships between entities. Guided by the interface, they can semi-automatically input relevant entities and relationships according to the input rules to construct a scalable graph structure.

[0067] Example of a specific data entry process: First, add device nodes and their component nodes on the initial interface. Enter the device name and description information for the device node, and the component name and description information for the component node, thus displaying the connection relationship. Then, select a specific component node and follow the guided prompts to fill in common fault types and their descriptions of fault phenomena, possible causes and recommended solutions for specific fault types, sensor signal data types, and commonly used algorithm models in the pop-up sub-node option input fields. The above process completes the input of entity content. Next, it is necessary to set the relationships between entities according to preset input rules, mainly including: the connection relationship between faulty devices and components, the relationship between faulty components and sensor signal data types, the relationship between algorithm models and fault phenomena and data types, the relationship between fault types and fault phenomena, the relationship between fault types and possible causes, and the relationship between fault types and recommended solutions. Among them, the connection relationship between faulty equipment and components, and the relationship between faulty components and sensor signal data types are matching correspondences, which have been established during the above entity entry process, specifically represented by the characters "the equipment includes...components, and the sensor data types commonly collected by the components are..."; the relationship between the algorithm model and the fault phenomenon refers to the degree to which a certain algorithm model is recommended for diagnosis of a specific fault phenomenon, measured by the matching confidence score, the initial value of which is preset by professionals based on actual conditions and experience; the relationship between the algorithm model and the data type is a matching correspondence, specifically represented by the characters "the input data types of the model are..."; the relationship between the fault type and the fault phenomenon refers to the degree of conformity between the specific fault phenomenon and the characteristic manifestation of a certain fault type, measured by the confidence score of the relationship, the magnitude of which is preset by professionals based on actual conditions and experience; similarly, the relationship between the fault type and the possible causes of the fault and the recommended solutions is also like this: for a specific fault type, professionals fill in several possible causes and recommended solutions based on historical experience, and represent the relationship between them and the fault type with a confidence score, the magnitude of which is preset and can be updated later based on the effect and user rating feedback.

[0068] like Figure 4 The example algorithm knowledge input interface shown is based on this embodiment.

[0069] Step S2: Based on the current fault information, use a scalable graph structure to extract the corresponding fault diagnosis algorithm model for diagnosis, and obtain the current fault type.

[0070] In this embodiment of the invention, based on the current fault information, a scalable graph structure is used to extract the corresponding fault diagnosis algorithm model for diagnosis. This includes: the user using the scalable graph structure to query and match the corresponding fault diagnosis algorithm model based on the current fault description information. The description information includes: fault location, fault phenomenon, and sensor data type. This is merely an example and not a limitation; in actual applications, the corresponding description information should be obtained according to the actual situation. The semantic similarity between the fault phenomenon in the graph structure and the description of the currently input fault phenomenon is calculated. The matching of the fault phenomenon and the corresponding algorithm model is determined. Based on whether they match, the corresponding fault diagnosis algorithm model is extracted according to the product of semantic similarity and matching confidence.

[0071] like Figure 5 The diagram shows a flowchart of querying the current fault condition and recommending suitable algorithm models. The graph structure recommends (potentially multiple) diagnostic algorithm models suitable for the current fault condition for fault diagnosis.

[0072] The query recommendation rules are as follows: Matching is performed on the fault location and its corresponding signal data type. Semantic similarity is calculated for the fault phenomena, and the matching confidence of the algorithmic model relationship between the fault phenomena and their corresponding phenomena is considered. Algorithm models are recommended based on whether there is a match and the product of similarity and matching confidence. The matching degree of each algorithm model with the current fault situation is sorted, and a predetermined number of algorithm models are selected from high to low as results and recommended to the user. To facilitate the calculation of semantic similarity, records about fault phenomena can be pre-subdivided into multiple structured fields, and the similarity of each field is calculated separately and then averaged as the numerical value.

[0073] Step S3: Query the diagnostic results of the current fault type in the preset diagram database.

[0074] In this embodiment of the invention, the diagnostic results include: the cause of the fault and the solution. This is only an example and not a limitation. In actual applications, the corresponding diagnostic results should be obtained according to the actual situation. After the specific fault type is diagnosed using the algorithm model, the possible causes of the fault and recommended solutions can be queried in the graph database. Users can score the algorithm model's recommendations based on their accuracy, the degree of conformity of the cause, and the effectiveness of the recommended solutions. The data feedback is used as historical scores to update relevant confidence levels (such as the confidence level of the relationship between the fault type and the inferred cause and recommended solution, and the matching confidence level of the relationship between the algorithm model and the fault phenomenon).

[0075] Step S4: Based on user feedback on the diagnostic results, update the scalable graph structure using a preset algorithm to complete the dynamic indexing and arrangement of the preset recommended diagnostic results.

[0076] In this embodiment of the invention, based on user feedback on the diagnostic results, a scalable graph structure is updated using a preset algorithm to complete the dynamic indexing and arrangement of preset recommended diagnostic results. This includes: querying the graph database for diagnostic results of the current fault type, where the diagnostic results include the cause of the fault and the solution; scoring the diagnostic results based on their accuracy against the algorithm's diagnostic model recommendations, the degree of consistency with the cause, and the accuracy of the solution; and updating the confidence level according to preset scoring rules to complete the dynamic indexing and arrangement of the preset recommended diagnostic results.

[0077] After diagnosing the specific fault type using the algorithm model, users can query the graph database for possible causes and recommended solutions. Users can rate the algorithm model's recommendations based on their accuracy, the degree of agreement with the causes, and the effectiveness of the recommended solutions. This feedback data serves as historical scores, used to update relevant confidence levels (such as the confidence level of the relationship between the fault type and the inferred causes and recommended solutions, and the confidence level of the match between the algorithm model and the fault phenomenon).

[0078] like Figure 6 The diagram illustrates an implementation example of the graph structure scaling and dynamic arrangement of recommendation results according to an embodiment of the present invention. Based on user feedback, the content and structure of the graph are expanded and reduced, and the confidence level of the relationships is bidirectionally bound by the designer and the user, enabling adaptive updates and optimizations, thereby achieving dynamic arrangement of recommendation results.

[0079] Users can record any errors or new fault information they discover in the graph database during actual use and provide feedback to professionals. Professionals can then add, delete, and modify entities and relationships through a low-code interface, ensuring the scalability of the graph structure. Furthermore, the priority of the graph database's recommendation results is related to the confidence level of relationships within the graph structure. This value is bidirectionally linked between the designers and users and can be updated and optimized based on user ratings. For example, the relationships between fault type and suspected causes, type and recommended solutions, type and fault symptoms, and algorithm models and fault symptoms are weighted as confidence levels, with initial values ​​set by professionals based on their expertise. After users utilize the graph structure to assist in fault diagnosis, they can rate the accuracy of the suspected causes, the effectiveness of the recommended solutions, the similarity of the fault symptoms, and the accuracy of the algorithm model's recommendations based on the assistance provided.

[0080] The rating feedback is used to update the confidence level. The update principle is that if a user's rating is high, it means the result is useful, and the weight can be increased for continued recommendations; if the rating is low, the weight is decreased, reducing future recommendations. For example, this example introduces the sigmoid activation function, which is commonly used in machine learning. Its function curve is smooth and monotonically increasing, and it is often used as the activation function of neural networks to map the independent variable to the range of 0 to 1. The mathematical expression is as follows, where the independent variable x represents the real value of the input:

[0081]

[0082] Assume there are three possible user ratings (good, average, poor), and let n represent the number of updates (counting from 0). W i n This represents the value of the i-th confidence level after the nth update. The confidence level can be updated according to the following formula:

[0083]

[0084] P i n The meaning of this is to make the confidence level of the i-th element W. i n The corresponding sigmoid function independent variable, K i P represents the step size during the i-th confidence update. i 0 (Initial confidence level) and K i The value of is pre-configured. 'm' represents the user's rating; 'm' is 1 for a good rating, 0 for a medium rating, and -1 for a poor rating. The update formula utilizes the monotonicity and compression of the sigmoid function to achieve adaptive weight updates: when the user rating is good, the weight increases, and the recommendation level of the corresponding result is enhanced; when the rating is medium, it remains unchanged; and when the rating is poor, the confidence level decreases, and the recommendation level of the corresponding result weakens. Furthermore, because the sigmoid function compresses the function value range to [0,1], the confidence level always remains non-negative.

[0085] On the one hand, professionals can modify the entities and relationships in the graph structure based on the knowledge and feedback suggestions provided by users; on the other hand, the relationship confidence can be adaptively updated and optimized based on user ratings, thereby affecting the priority of recommendation results and adjusting the recommendation order according to the actual usage of users, thus improving the personalization of recommendations. Through the above two aspects, the scaling of the graph structure and the dynamic arrangement of recommendation results can be achieved.

[0086] The scalable graph structure-based fault diagnosis and knowledge elastic indexing method provided in this invention organizes and manages various knowledge in equipment fault diagnosis, such as fault phenomena, faulty components, fault types, and diagnostic algorithms, in the form of a graph structure and stores them in a graph database in a low-code, semi-automatic manner. This facilitates fault querying and the recommendation of corresponding diagnostic algorithms to assist in fault diagnosis and analysis. The graph structure is scaled up and down and the recommendation results are dynamically arranged based on user feedback. This helps solve problems in current intelligent equipment fault diagnosis, such as insufficient knowledge management, difficulties in automatically extracting entities and relationships from knowledge graphs leading to incomplete extraction and low diagnostic accuracy, and poor scalability and real-time update of knowledge graphs. It improves the accuracy and comprehensiveness of knowledge recommendation and enhances the usability and personalization of reasoning and prediction results.

[0087] Example 2

[0088] This invention provides a fault diagnosis and flexible knowledge indexing and orchestration system with a scalable graph structure, such as... Figure 7 As shown, it includes:

[0089] The scalable graph structure building module 1 is used to design a low-code interactive interface based on an open-source graph database according to preset rules and to build a scalable graph structure. This module executes the method described in step S1 of embodiment 1, which will not be repeated here.

[0090] The fault type diagnosis module 2 is used to extract the corresponding fault diagnosis algorithm model based on the current fault information using a scalable graph structure to diagnose the current fault type. This module executes the method described in step S2 of embodiment 1, which will not be repeated here.

[0091] The diagnostic result query module 3 is used to query the diagnostic results of the current fault type in the preset diagram database; this module executes the method described in step S3 of embodiment 1, which will not be repeated here.

[0092] The index arrangement module 4 is used to update the scalable graph structure based on the user's feedback on the diagnostic results using a preset algorithm, and to complete the dynamic index arrangement of the preset recommended diagnostic results. This module executes the method described in step S4 of embodiment 1, which will not be repeated here.

[0093] This invention provides a scalable graph-structured fault diagnosis and knowledge elastic indexing and orchestration system. By establishing a scalable fault diagnosis algorithm knowledge indexing and orchestration system, it can quickly recommend diagnostic algorithms suitable for the current fault phenomenon, and accurately and comprehensively analyze the location of the fault and the related causes of the fault, thereby reducing the reliance on professional personnel during equipment operation and maintenance.

[0094] Example 3

[0095] This invention provides a terminal, such as... Figure 8 As shown, the system includes: at least one processor 401, such as a CPU (Central Processing Unit), at least one communication interface 403, a memory 404, and at least one communication bus 402. The communication bus 402 is used to enable communication between these components. The communication interface 403 may include a display screen or a keyboard; optionally, the communication interface 403 may also include a standard wired interface or a wireless interface. The memory 404 may be high-speed RAM (Random Access Memory) or non-volatile memory, such as at least one disk storage device. Optionally, the memory 404 may also be at least one storage device located remotely from the processor 401. The processor 401 can execute the fault diagnosis and flexible knowledge indexing method for the scalable graph structure in Embodiment 1. The memory 404 stores a set of program code, and the processor 401 calls the program code stored in the memory 404 to execute the fault diagnosis and flexible knowledge indexing method for the scalable graph structure in Embodiment 1. The communication bus 402 can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The communication bus 402 can be divided into an address bus, a data bus, and a control bus, etc. For ease of representation, Figure 8 The bus is represented by a single line, but this does not mean that there is only one bus or one type of bus. Memory 404 may include volatile memory, such as random-access memory (RAM); memory may also include non-volatile memory, such as flash memory, hard disk drive (HDD), or solid-state drive (SSD); memory 404 may also include combinations of the above types of memory. Processor 401 may be a central processing unit (CPU), a network processor (NP), or a combination of a CPU and an NP.

[0096] The memory 404 may include volatile memory, such as random-access memory (RAM); the memory may also include non-volatile memory, such as flash memory, hard disk drive (HDD) or solid-state drive (SSD); the memory 404 may also include a combination of the above types of memory.

[0097] The processor 401 can be a central processing unit (CPU), a network processor (NP), or a combination of CPU and NP.

[0098] The processor 401 may further include a hardware chip. This hardware chip may be an application-specific integrated circuit (ASIC), a programmable logic device (PLD), or a combination thereof. The PLD may be a complex programmable logic device (CPLD), a field-programmable gate array (FPGA), a generic array logic (GAL), or any combination thereof.

[0099] Optionally, the memory 404 is also used to store program instructions. The processor 401 can call the program instructions to implement the fault diagnosis and knowledge elastic indexing arrangement method of the scalable graph structure as described in Embodiment 1 of this application.

[0100] This invention also provides a computer-readable storage medium storing computer-executable instructions that can execute the fault diagnosis and knowledge elastic indexing arrangement method for the scalable graph structure in Embodiment 1. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk drive (HDD), or solid-state drive (SSD), etc.; the storage medium may also include combinations of the above types of memory.

[0101] Obviously, the above embodiments are merely illustrative examples for clear explanation and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations here. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A method for fault diagnosis and flexible knowledge indexing and arrangement using a scalable graph structure, characterized in that, include: Based on preset rules, a low-code interactive interface is designed using an open-source graph database to build a scalable graph structure; Based on the current fault information, a scalable graph structure is used to extract the corresponding fault diagnosis algorithm model for diagnosis, and the current fault type is obtained. Query the diagnostic results for the current fault type in the preset diagram database; Based on user feedback on the diagnostic results, the scalable graph structure is updated using a preset algorithm to complete the dynamic indexing and arrangement of the preset recommended diagnostic results. Based on the current fault information, a scalable graph structure is used to extract the corresponding fault diagnosis algorithm model for diagnosis, including: Users can use a scalable graph structure to query and match the corresponding fault diagnosis algorithm model based on the current fault description information. The description information includes: fault location, fault phenomenon, and sensor data type. Calculate the semantic similarity between the fault phenomena in the graph structure and the description of the current input fault phenomenon. Determine whether the fault phenomenon and the corresponding algorithm model match; Based on whether a match is found, the corresponding fault diagnosis algorithm model is extracted by the product of semantic similarity and matching confidence. Based on user feedback on the diagnostic results, a scalable graph structure is updated using a preset algorithm to complete the dynamic indexing and arrangement of preset recommended diagnostic results, including: Based on the diagnostic results of the current fault type retrieved from the graph database, the diagnostic results include: the cause of the fault and the solution; The diagnostic results are scored based on their accuracy in relation to the algorithm diagnostic model's recommendations, the degree of consistency with the causes, and the accuracy of the solutions. Based on the preset scoring rules, the confidence level is updated, and the dynamic indexing and arrangement of the preset recommended diagnostic results are completed.

2. The fault diagnosis and flexible knowledge indexing method for scalable graph structures according to claim 1, characterized in that, The initial screen of the interactive interface is a node connection diagram of the device and its components. Nodes represent devices or connecting components. Users can upload corresponding fault images and text descriptions to the nodes, which display the various components and their connection relationships. Users can add, delete, and modify nodes through the interactive interface. Using pre-written GQL statements, the content of the operation objects is used as parameters to be filled in. The interactive interface is used to guide users to input information. Users need to focus on the content input of nodes and connection relationships without writing GQL code.

3. The fault diagnosis and flexible knowledge indexing method for scalable graph structures according to claim 2, characterized in that, The low-code interface for open-source graph databases includes: Add nodes for devices and their components to the initial interface. Enter the name and description of the device in the device node and the name and description of the component in the component node to display the connection relationship between the devices and components. Select the node of the preset component, and follow the prompts to fill in the preset fault type, the corresponding phenomenon description, the preset cause of the fault, the solution, the data type of the sensor signal, and the fault diagnosis algorithm model in the pop-up sub-node option input field.

4. The fault diagnosis and flexible knowledge indexing method for scalable graph structures according to claim 3, characterized in that, The preset rules include: the connection relationship between faulty equipment and components, the relationship between faulty components and sensor signal data types, the relationship between algorithm models and fault phenomena, the relationship between algorithm models and data types, the relationship between fault types and fault phenomena, the relationship between fault types and preset causes, and the relationship between fault types and solutions.

5. In the fault diagnosis and knowledge elastic indexing method of the scalable graph structure according to claim 4, the connection relationship between faulty equipment and components, and the relationship between faulty components and sensor signal data types are all matching correspondence relationships.

6. A fault diagnosis and knowledge elastic indexing and arrangement system with a scalable graph structure, characterized in that, include: The scalable graph structure building module is used to design low-code interactive interfaces and build scalable graph structures based on open-source graph databases according to preset rules. The fault type diagnosis module is used to extract the corresponding fault diagnosis algorithm model based on the current fault information using a scalable graph structure for diagnosis. The process includes: users querying and matching corresponding fault diagnosis algorithm models using a scalable graph structure based on current fault description information, where the description information includes: fault location, fault phenomenon, and sensor data type; calculating the semantic similarity between the fault phenomenon in the graph structure and the description of the currently input fault phenomenon; determining whether the fault phenomenon and the corresponding algorithm model match; and extracting the corresponding fault diagnosis algorithm model based on the product of semantic similarity and matching confidence, depending on whether they match. The diagnostic result query module is used to query the diagnostic results of the current fault type in the preset diagram database. The indexing and arrangement module is used to update the scalable graph structure based on user feedback on diagnostic results using a preset algorithm, and to complete the dynamic indexing and arrangement of preset recommended diagnostic results. This includes: diagnostic results for the current fault type retrieved from the graph database, wherein the diagnostic results include: the cause of the fault and the solution; scoring the diagnostic results based on the accuracy of the algorithm diagnostic model recommendations, the degree of consistency of the cause, and the accuracy of the solution; and updating the confidence level according to preset scoring rules to complete the dynamic indexing and arrangement of preset recommended diagnostic results.

7. A terminal, characterized in that, include: At least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to cause the at least one processor to perform the fault diagnosis and knowledge elastic indexing and orchestration method for a scalable graph structure according to any one of claims 1-5.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing the computer to execute the fault diagnosis and knowledge elastic indexing and arrangement method for the scalable graph structure as described in any one of claims 1-5.

Images