Intelligent interaction and completion method and device for FMEA knowledge, medium and product

Abstract

The invention discloses an intelligent interaction and completion method and device for FMEA knowledge, a medium and a product, and belongs to the field of data processing. According to the method, the FMEA element information is extracted from the non-structured or semi-structured FMEA document to determine the structured knowledge node, efficient digital conversion of different standard template historical documents is realized, semantic alignment and fusion are performed on the newly extracted knowledge node and the historical structure tree in the FMEA knowledge graph, semantic conflicts are eliminated, and therefore, the reliability of the knowledge graph is improved. And performing integrity verification on the fused knowledge graph by adopting a causal logic chain, automatically generating knowledge node completion suggestion data when identifying that the causal logic chain is missing, presenting a complete knowledge graph and an intelligent completion suggestion to a user through an interactive interface, and receiving a user instruction to update the graph. A closed loop of system intelligent suggestion and manual auditing and decision making is formed, so that FMEA knowledge can be visually viewed and verified.

Description

Technical Field

[0001] This application relates to the field of data processing, and in particular to intelligent interaction and completion methods, devices, media and products for FMEA (Failure Mode and Effects Analysis) knowledge. Background Technology

[0002] Failure Mode and Effects Analysis (FMEA), as a systematic and proactive risk assessment method, plays a crucial role in industrial processes such as product design, manufacturing, and quality control. It significantly reduces product and process risks by systematically identifying potential failure modes, analyzing their effects and causes, and proactively developing improvement measures.

[0003] However, the traditional FMEA implementation process relies heavily on expert experience, and this knowledge is mostly in the form of unstructured or semi-structured documents, leading to a series of challenges:

[0004] Difficulties in digitizing and extracting knowledge: A large amount of historical FMEA knowledge accumulated by enterprises is stored in documents such as Word, Excel or PDF, with different formats, which are difficult for computers to automatically recognize and process, forming "information silos" and hindering the effective digital inheritance and large-scale utilization of knowledge.

[0005] Challenges in Knowledge Integration and Consistency: In cross-departmental and cross-phase FMEA collaboration, the lack of a unified terminology standard can lead to multiple expressions for the same concept. Existing systems lack effective semantic alignment and conflict resolution mechanisms, resulting in ambiguity and inconsistency during knowledge integration, making it difficult to build a high-quality unified knowledge system.

[0006] Lack of logical integrity verification: The core of FMEA lies in the causal logical chain between "structure-function-failure". Current technology fails to fully utilize this unique logic for automated reasoning and verification, making it difficult to detect logical breakpoints or causal gaps in the analysis in a timely manner, thus affecting the depth and quality of FMEA analysis.

[0007] Insufficient intelligent assistance capabilities: When gaps exist in the knowledge graph, the system cannot proactively provide intelligent suggestions for completion. Analysts still need to rely on manual searches and personal experience to fill in the gaps, which is not only inefficient but also increases the workload and subjectivity of the analysis results.

[0008] Poor interaction and application experience: Existing systems generally lack intuitive and user-friendly visual interfaces, making it difficult for users to browse FMEA knowledge globally, explore it in depth, and update it conveniently, thus limiting the effective application of knowledge and continuous iterative optimization.

[0009] In summary, traditional FMEA suffers from difficulties in knowledge digitization, inconsistent terminology, lack of logical verification, insufficient intelligent assistance, and poor interactive experience, which hinders the effective inheritance, integration, and intelligent application of knowledge. Summary of the Invention

[0010] To address the problems of traditional FMEA, such as difficulties in knowledge digitization, inconsistent terminology, lack of logical verification, insufficient intelligent assistance, and poor interactive experience, this paper provides a method, device, medium, and product for intelligent interaction and completion of FMEA knowledge. This method aims to achieve automated extraction and structuring of FMEA knowledge, possess semantic alignment and conflict resolution capabilities, support integrity auditing based on causal logic, and provide intelligent completion suggestions and a user-friendly human-computer interaction interface.

[0011] This application provides an intelligent interaction and completion method for FMEA knowledge, including:

[0012] Obtain unstructured or semi-structured FMEA documents;

[0013] Extract the FMEA element information from the FMEA document to determine the structured knowledge nodes;

[0014] The knowledge nodes are semantically aligned and fused with the historical structure tree in the FMEA knowledge graph, and semantic conflicts are resolved.

[0015] Based on the causal logic chain consisting of structure-function-failure, the integrity of the fused FMEA knowledge graph is verified.

[0016] When a missing link in the causal logic chain is identified, corresponding knowledge node completion suggestion data is generated.

[0017] The FMEA knowledge graph and the completion suggestion data are presented to the user through an interactive interface.

[0018] The FMEA knowledge graph is updated by receiving and updating user instructions.

[0019] Optionally, the step of extracting FMEA element information from the FMEA document and determining structured knowledge nodes includes:

[0020] Identify and parse the text, and / or tables, and / or diagrams, and / or flowcharts in the FMEA document;

[0021] The context of the text, and / or the table, and / or the block diagram, and / or the flowchart are transformed into the association between the knowledge nodes.

[0022] Optionally, the step of semantically aligning and fusing the knowledge nodes with the historical structure tree in the FMEA knowledge graph, and resolving semantic conflicts, includes:

[0023] Using a first language model, the semantic similarity between the knowledge node and the historical structure tree in the FMEA knowledge graph is calculated;

[0024] Based on the semantic similarity, conflicting nodes with the same semantics but inconsistent expressions are identified, and conflict alarm messages are generated.

[0025] In response to the conflict alarm message, a second language model is used to analyze the context of the conflict node, generate standard node terms, and store the standard node terms in the FMEA knowledge graph.

[0026] Optionally, the step of using a first language model to calculate the semantic similarity between the knowledge node and the historical structure tree in the FMEA knowledge graph includes:

[0027] The attribute set of the knowledge node is vectorized using the first language model to generate a first feature vector;

[0028] Obtain the attribute set of the historical structure tree in the FMEA knowledge graph and generate a second feature vector;

[0029] Calculate the semantic similarity score between the first feature vector and the second feature vector;

[0030] If the semantic similarity score is greater than a preset threshold, then the semantic association between the knowledge node and the historical structure tree is determined.

[0031] The semantic similarity scores are stored in a similarity matrix;

[0032] Based on the similarity matrix, potential conflict relationships between the knowledge nodes and the historical structure tree are determined.

[0033] Optionally, the step of identifying conflicting nodes with the same semantics but inconsistent expressions based on the semantic similarity and generating conflict alarm messages includes:

[0034] Extract the node pairs in the similarity matrix whose semantic similarity scores are higher than the preset threshold;

[0035] For each node pair, the attribute set of the knowledge node is compared with that of the historical structure tree to determine whether there is a semantic conflict due to inconsistent expression; if a semantic conflict exists, the knowledge node is marked as a conflict node.

[0036] Generate a conflict alarm message, which includes the identifier of the conflicting node, the conflict type, and the conflict description;

[0037] The conflict alarm message is stored in the log database.

[0038] Optionally, when a missing causal logic chain is identified, generating corresponding knowledge node completion suggestion data includes:

[0039] From the FMEA knowledge graph, extract the upstream and downstream node information that is related to the missing part of the causal logic chain to form the reasoning context;

[0040] The completion suggestion data is generated by using a third language model based on the reasoning context.

[0041] Optionally, when a missing causal logic chain is identified, corresponding knowledge node completion suggestion data is generated, and the process further includes:

[0042] The fourth language model is used to predict the score value after the completion suggestion data is adopted based on historical data in the FMEA knowledge graph;

[0043] The step of presenting the FMEA knowledge graph and the completion suggestion data to the user through an interactive interface also includes:

[0044] The rating values ​​are presented.

[0045] This application also provides an electronic device comprising: one or more processors; and a memory storing computer program instructions, which, when executed, cause the processor to perform the steps of the method described above.

[0046] This application also provides a computer-readable medium having computer program instructions stored thereon, which can be executed by a processor to implement the method described above.

[0047] This application also provides a computer program product, including a computer program / instructions that, when executed by a processor, implement the steps of the method described above.

[0048] The beneficial effects of the above technical solution are as follows:

[0049] In this technical solution, the intelligent interaction and completion method for FMEA knowledge first acquires unstructured or semi-structured FMEA documents and extracts FMEA element information to determine structured knowledge nodes. This achieves efficient digital conversion of historical documents from different standard templates, solving the fundamental problem that a large amount of FMEA knowledge cannot be transferred into the intelligent system. Subsequently, by semantically aligning and fusing the newly extracted knowledge nodes with the historical structure tree in the FMEA knowledge graph and resolving semantic conflicts, it effectively ensures that knowledge from different departments or periods with differing expressions can be integrated into a unified and consistent knowledge system. Furthermore, based on a causal logic chain composed of "structure-function-failure," the completeness of the fused knowledge graph is verified. When a missing causal logic chain is identified, knowledge node completion suggestions are automatically generated, significantly improving the quality and logical rigor of the FMEA analysis itself. Finally, the complete knowledge graph and intelligent completion suggestions are presented to the user through an interactive interface, and user commands are received to update the graph, forming a closed loop of "system intelligent suggestions, manual review and decision-making." This not only allows FMEA knowledge to be viewed and verified intuitively but also supports its continuous iterative optimization. In summary, this application significantly improves knowledge management efficiency and analysis quality. Attached Figure Description

[0050] One or more embodiments are illustrated by way of example with reference numerals in the accompanying drawings. These illustrations do not constitute a limitation on the embodiments. Elements with the same reference numerals in the drawings are denoted as similar elements. Unless otherwise stated, the figures in the drawings are not to be limited by scale.

[0051] Figure 1 A flowchart illustrating one embodiment of the intelligent interaction and completion method for FMEA knowledge described in this application;

[0052] Figure 2 A flowchart illustrating an embodiment for determining structured knowledge nodes in FMEA element information extracted from FMEA documents for this application;

[0053] Figure 3 This is a flowchart of one embodiment of the semantic conflict resolution method in this application;

[0054] Figure 4 This is an exemplary structural diagram of the electronic device of this application. Detailed Implementation

[0055] The advantages of this application are further illustrated below with reference to the accompanying drawings and specific embodiments.

[0056] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numerals in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this disclosure as detailed in the appended claims.

[0057] The terminology used in this disclosure is for the purpose of describing particular embodiments only and is not intended to be limiting of the disclosure. The singular forms “a,” “the,” and “the” as used in this disclosure and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any and all possible combinations of one or more of the associated listed items.

[0058] It should be understood that although the terms first, second, third, etc., may be used in this disclosure to describe various information, such information should not be limited to these terms. These terms are used only to distinguish information of the same type from one another. For example, without departing from the scope of this disclosure, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the word "if" as used herein may be interpreted as "when," "when," or "in response to determination."

[0059] In the description of this application, it should be understood that the numerical labels before the steps do not indicate the order of the steps, but are only used to facilitate the description of this application and to distinguish each step, and therefore should not be construed as a limitation of this application.

[0060] Example 1

[0061] This application addresses the challenges of traditional FMEA (Fast-To-Earn) knowledge acquisition, integration, and intelligent application, which are hampered by difficulties in knowledge digitization, inconsistent terminology, lack of logical verification, insufficient intelligent assistance, and poor interactive experience. It proposes an intelligent interaction and completion method for FMEA knowledge. (See reference...) Figure 1 The intelligent interaction and completion method for FMEA knowledge provided in this embodiment mainly includes the following steps:

[0062] S1. Obtain unstructured or semi-structured FMEA documents;

[0063] In this embodiment, the FMEA document can be text data containing FMEA-related information obtained from an external data source. The text data may include unstructured documents and / or semi-structured tables. The text data is preprocessed to extract FMEA element information, which includes components, functions, failure modes, effects, and measures. The FMEA element information is parsed using natural language processing technology to generate a structured data format. The structured data format is converted into knowledge nodes, which include node attributes and node relationships. The knowledge nodes are validated to determine whether they meet the structural requirements of an FMEA knowledge graph. If the knowledge nodes meet the requirements, they are stored in a temporary database.

[0064] In practical applications, FMEA documents can be internally accumulated failure mode and impact analysis reports, such as Word documents, Excel spreadsheets, or PDF files, etc.

[0065] S2. Extract the FMEA element information from the FMEA document and determine the structured knowledge nodes;

[0066] The purpose of step S2 is to process information in different formats within the FMEA document and transform it into structured knowledge nodes and relationships. For details, please refer to... Figure 2 Step S2 shown may include the following steps:

[0067] S21. Identify and parse the text, and / or tables, and / or diagrams, and / or flowcharts in the FMEA document;

[0068] Step S21 involves recognizing and parsing the input FMEA document. Specifically: for text content, text parsing technology is used to identify and segment continuous semantic paragraphs and key fields. For table content, a table structure recognition algorithm is used to parse the row and column structure of the table and locate the table header and data cells. For block diagrams and flowcharts, image recognition technology is used to extract graphic elements (such as boxes and arrows), as well as the connections between elements and text annotations.

[0069] Step S21 enables unified parsing and information extraction of heterogeneous information in unstructured or semi-structured FMEA documents.

[0070] S22. Transform the context of the text, and / or the table, and / or the block diagram, and / or the flowchart into the association relationship between the knowledge nodes.

[0071] As an example rather than a limitation, information such as product structure name, function description, failure mode, cause, and consequence is extracted from the FMEA table as knowledge nodes, and logical connections between these nodes are established based on the row and column relationships of the table.

[0072] Step S22, based on the parsing results of step S21, transforms the original information into structured knowledge nodes and establishes the relationships between them:

[0073] Textual context association: For entities extracted from text paragraphs (such as "a certain valve"), the association is determined by analyzing the grammatical structure of the sentence in which it is located (such as subject-verb-object relationship). For example, from the sentence "The main function of a certain valve is to regulate flow", "a certain valve" (structural node) and "regulate flow" (functional node) can be extracted, and the association relationship of "having function" can be established.

[0074] Table row and column association: Utilizing the inherent structure of tables, data in the same row or column is directly mapped to related nodes. As an example, not a limitation, in a standard FMEA table, the system extracts multiple elements from a row of data, such as "pump body" (product structure), "providing stable pressure" (function), "pressure fluctuation" (failure mode), "seal wear" (cause), and "system performance degradation" (consequence), as independent knowledge nodes. Based on their coexistence in the same row, a complete causal logical chain of "structure-function-failure mode-cause-consequence" is established.

[0075] Topological relationship transformation: For block diagrams and flowcharts, the connection relationships between their graphical elements (such as arrow directions) are directly transformed into semantic associations between knowledge nodes. For example, if the block diagram "Component A" is identified from the flowchart and points to the block diagram "Component B" via an arrow, then the association relationship "Component A - connects to - Component B" can be established.

[0076] The method described in this embodiment can be used to process FMEA documents in different formats and templates within an enterprise, efficiently and accurately extract key risk knowledge elements from them, and transform them into structured nodes and relationships that can be used to build a knowledge graph, thus laying a reliable data foundation for subsequent semantic alignment and intelligent analysis.

[0077] In this embodiment, step S2 constructs an attribute set for knowledge nodes based on the extracted FMEA element information. The attribute set includes component identifier, functional description, and failure mode. For each knowledge node, its hierarchical relationship in the FMEA knowledge graph is determined. The FMEA element information is matched with the node template of the FMEA knowledge graph using a preset mapping rule. If the match is successful, the corresponding knowledge node is generated. If the match fails, an anomaly report is generated and the inconsistent FMEA element information is recorded. The generated knowledge node is stored in the candidate node set of the FMEA knowledge graph.

[0078] S3. Semantically align and fuse the knowledge nodes with the historical structure tree in the FMEA knowledge graph, and resolve semantic conflicts;

[0079] Further, see Figure 3 Step S3 shown may include the following steps:

[0080] S31. Using a first language model, calculate the semantic similarity between the knowledge node and the historical structure tree in the FMEA knowledge graph;

[0081] In this embodiment, the first language model is a semantic similarity model. The first language model adopts a large language model.

[0082] The purpose of step S31 is to leverage the deep semantic understanding capabilities of large language models to achieve accurate semantic similarity calculation and conflict relationship identification. Specifically, step S31 may include the following steps:

[0083] S311. Vectorize the attribute set of the knowledge node using the first language model to generate a first feature vector;

[0084] In step S311, the attribute set of the knowledge nodes to be fused (including textual information such as node name, function description, and system to which they belong) is taken as input and encoded using a first language model. In this embodiment, the first language model adopts a large language model based on the Transformer architecture, which performs deep semantic parsing of the input text through a self-attention mechanism and outputs a high-dimensional, dense numerical vector, namely the first feature vector, which can effectively represent the deep semantic features of the knowledge nodes.

[0085] S312. Obtain the attribute set of the historical structure tree in the FMEA knowledge graph and generate a second feature vector;

[0086] In step S312, the attribute set corresponding to the historical structure tree is obtained from the FMEA knowledge graph. The same first language model and encoding method as in S311 are used to generate the corresponding second feature vector to ensure that new and old knowledge are measured in a unified vector space.

[0087] S313. Calculate the semantic similarity score between the first feature vector and the second feature vector;

[0088] In step S313, the cosine similarity between the first feature vector and the second feature vector is calculated to obtain a semantic similarity score. The score ranges from [-1, 1]. The closer the score is to 1, the more consistent the directions of the two nodes are in the semantic space, and the higher the semantic similarity.

[0089] S314. If the semantic similarity score is greater than a preset threshold, then the semantic association between the knowledge node and the historical structure tree is determined;

[0090] In step S314, the calculated semantic similarity score is compared with a preset threshold (set to 0.75 in this embodiment). If the score is greater than the threshold, it is determined that the current knowledge node has a semantic relationship with the historical structure tree, that is, the two are highly related semantically and may point to the same entity or concept.

[0091] S315. Store the semantic similarity score into a similarity matrix;

[0092] In step S315, the semantic similarity score between each new knowledge node and each historical structure tree node is systematically stored in a two-dimensional similarity matrix. The rows of this matrix are the identifiers of new knowledge nodes, the columns are the identifiers of historical structure tree nodes, and the cell values ​​are the corresponding similarity scores, providing a complete data foundation for subsequent conflict analysis.

[0093] S316. Based on the similarity matrix, determine the potential conflict relationships between the knowledge nodes and the historical structure tree.

[0094] In step S316, based on the constructed similarity matrix, all node pairs with scores higher than a preset threshold are automatically scanned. The attribute sets of these node pairs are compared in a refined manner. If significant differences are found in the node names or key attribute descriptions but the semantics are highly similar (e.g., "sealing ring" and "O-ring"), they are marked as having a potential conflict relationship, i.e., "synonymous" conflict, and a corresponding conflict warning is generated.

[0095] This embodiment transforms text attributes into vectors using a large language model for similarity calculation, which goes beyond surface string matching to achieve deep semantic association recognition. By constructing a similarity matrix to systematically manage the association strength between all nodes, it provides a scientific and quantitative basis for accurately discovering potential terminology conflicts, significantly improving the automation and accuracy of knowledge fusion.

[0096] In this embodiment, by employing a first language model to perform high-dimensional vectorization representation of knowledge nodes, the system can deeply understand the semantic connotations of FMEA professional terms, thereby achieving accurate semantic association recognition that surpasses keyword matching. Furthermore, by constructing a similarity matrix to systematically locate potential conflicts, it can proactively and accurately discover inconsistencies in terms such as "homonyms" or "synonyms," making implicit contradictions that originally relied on manual intervention explicit, and issuing warnings before fusion, fundamentally avoiding the damage to the quality of the knowledge system caused by terminology confusion. These steps together automate key links in knowledge fusion, which not only greatly improves integration efficiency, but also ensures the consistency, accuracy, and authority of the fused knowledge graph through machine-assisted decision-making, laying a high-quality data foundation for subsequent intelligent applications.

[0097] S32. Based on the semantic similarity, identify conflicting nodes that have the same semantics but inconsistent expressions, and generate conflict alarm messages;

[0098] Step S32 ensures terminology consistency during the knowledge fusion process through a systematic conflict detection mechanism. Specifically, step S32 may include the following steps:

[0099] S321. Extract the node pairs in the similarity matrix whose semantic similarity scores are higher than the preset threshold;

[0100] In step S321, all node pairs with semantic similarity scores higher than a preset threshold (0.75 in this embodiment) are automatically selected from the semantic similarity matrix. These node pairs are highly correlated in the semantic space and are potential conflict detection targets.

[0101] S322. For each node pair, compare the attribute set of the knowledge node with that of the historical structure tree to determine whether there is a semantic conflict due to inconsistent expression; if there is a semantic conflict, mark the knowledge node as a conflict node.

[0102] In step S322, a refined comparison is performed on each pair of highly similar nodes:

[0103] The core fields in the node attribute set, including node name, functional description, and technical parameters, are compared, and a text difference analysis algorithm is used to identify inconsistencies in the descriptions. Conflict types are determined according to preset rules, including: synonyms (same meaning but different names, such as "sealing ring" and "O-ring"); synonyms with different meanings (same name but different meanings, such as "controller" referring to different model parts); and attribute contradictions (conflicting key parameters, such as inconsistent rated voltage labeling). When any of the above situations is detected, the new knowledge node is marked as a conflicting node.

[0104] S323. Generate a conflict alarm message, the conflict alarm message including the identifier of the conflicting node, the conflict type and the conflict description;

[0105] In step S323, for each conflict node, the system automatically generates a structured conflict alarm message, which includes: Conflict node identifier: a unique ID of the new knowledge node; Conflict type: a specific conflict category (synonyms / nominal opposites / attribute contradictions); Conflict description: a detailed description of the problem and a comparison of differences; Associated historical nodes: information on the historical tree structure nodes that conflict with it; Confidence level: the degree of conflict confirmation based on similarity scores.

[0106] S324. Store the conflict alarm message in the log database.

[0107] In step S324, the generated conflict alarm message is stored in the log database in a standardized format, including: timestamp: time of conflict occurrence; processing status: pending / processed / ignored; operation record: subsequent processing trajectory; a complete conflict management ledger is established, supporting historical query and statistical analysis.

[0108] This embodiment achieves automatic identification and standardized alarms for terminology inconsistencies through a conflict detection process. It transforms quality checks, which rely on manual experience, into a traceable and quantifiable automated process, significantly improving the quality control level of knowledge fusion and providing a reliable guarantee for building a high-quality unified knowledge system. In step S32, by extracting nodes with high semantic similarity from the similarity matrix and performing refined attribute comparisons, conflicting nodes that are "semantically identical but inconsistent in expression" can be accurately identified. This achieves automated discovery and marking of implicit knowledge contradictions such as "synonyms." Furthermore, by generating and storing alarm messages containing conflict details, potential terminology confusion problems are systematized and made explicit, providing analysts with clear handling guidelines. This effectively intercepts data contradictions caused by inconsistent expression before knowledge fusion, significantly improving the internal consistency and reliability of the FMEA knowledge system.

[0109] S33. In response to the conflict alarm message, a second language model is used to analyze the context of the conflict node, generate standard node terms, and store the standard node terms in the FMEA knowledge graph.

[0110] In this embodiment, the second language model is a terminology standardization model, and the second language model adopts a large language model.

[0111] Step S33 resolves semantic conflicts in knowledge fusion through intelligent terminology standardization. Further, step S33 may include the following steps:

[0112] S331. Obtain the context information of the conflicting node, the context information including the attribute set of the conflicting node and the attributes of the associated node;

[0113] Step S331 obtains the complete context information of the conflicting node, including: the set of attributes of the conflicting node itself (node ​​name, function description, technical parameters, etc.); the attribute information of directly related nodes (parent and child nodes, sibling nodes and related functions / failed nodes); the upstream and downstream node relationship network of the causal chain; forming a complete semantic analysis context environment.

[0114] S332. Perform semantic analysis on the context information using a second language model to generate a set of candidate standard terms;

[0115] Step S332 inputs the collected context information into the second language model (in this embodiment, a terminology standardization model based on a large language model is used). The model generates a set of candidate standard terms through the following steps: performing deep semantic understanding of the context to identify the technical field and conceptual category; performing pattern matching based on the domain terminology library and existing standard terms; generating 3-5 candidate standard terms that best fit the context semantics and sorting them by confidence level.

[0116] S333. Select standard node terms from the candidate standard terminology set according to the preset terminology standard rules;

[0117] Step S333 filters the candidate set according to preset terminology standardization rules: prioritizes terms that are completely matched with the enterprise standard terminology library; if no complete match is found, selects terms that are closest to industry standards (such as ISO and GB series standards); and ensures consistency with relevant terms in the existing knowledge graph through terminology consistency checks.

[0118] S334. Map the standard node terms to the original terms of the conflicting nodes;

[0119] Step S334 establishes a bidirectional mapping relationship between the selected standard node terminology and the original terminology of the conflicting node, including: synonym mapping: establishing a synonym relationship between "standard terminology and original terminology"; version record: retaining the original terminology as a historical version; usage scenario annotation: recording the specific context in which the terminology applies.

[0120] S335. Update the mapped standard node terms to the FMEA knowledge graph;

[0121] Step S335 updates the mapped standard node terms to the FMEA knowledge graph: replaces non-standard terms in conflicting nodes; updates term references in related relationships; and maintains the topology of the knowledge graph unchanged.

[0122] S336. Generate term update records and store them in the metadata table of the FMEA knowledge graph.

[0123] Step S336 generates a complete terminology update record and stores it in the metadata table of the FMEA knowledge graph. The record includes: pre-standardization terms and post-standardization terms; standardization timestamp and executing user; the version of the terminology specification rules adopted; the candidate terminology set and the confidence score of each term; and establishes a complete terminology evolution tracing system.

[0124] This embodiment achieves intelligent resolution of semantic conflicts through a context-aware terminology standardization process. It maintains the terminology standardization of the knowledge graph while preserving the continuity and traceability of knowledge through complete mapping records, providing technical support for building a high-quality enterprise standard knowledge system. In step S33, by introducing a terminology standardization model to perform deep semantic analysis of the context of conflict nodes, candidate standard terms can be intelligently generated based on domain knowledge. Combined with preset specifications, terminology selection and mapping are automatically completed, achieving an automated closed loop from conflict identification to terminology unification. This process not only significantly improves the efficiency of terminology standardization and avoids the subjectivity and lag of manual unification, but also ensures the consistency, traceability, and authority of the FMEA knowledge graph during its continuous evolution by maintaining complete terminology update records, laying a solid foundation for the long-term maintenance and reliable application of knowledge.

[0125] S4. Based on the causal logic chain consisting of structure-function-failure, perform integrity verification on the fused FMEA knowledge graph;

[0126] In this embodiment, step S4 checks whether there are logical missing situations in the knowledge graph, such as structural nodes not having corresponding functional descriptions, functional nodes not having corresponding failure modes, or failure modes not having causes or consequences.

[0127] Step S4 ensures the integrity and logical consistency of the knowledge graph through systematic logical chain traversal and verification. Specifically, step S4 may include the following steps:

[0128] S41. Traverse the FMEA knowledge graph and extract the causal logic chain containing structure, function, and failure;

[0129] Step S41 performs a full graph traversal of the fused FMEA knowledge graph to identify and extract all causal logical chains containing nodes of three types: structure, function, and failure. Specifically, a graph traversal algorithm is used, starting from the root-level structural node and performing a depth-first search along the standard path of "structure-function-failure mode-failure cause-failure impact," recording all complete and partially complete logical paths.

[0130] S42. For each causal logic chain, determine whether there are any missing knowledge nodes;

[0131] Step S42 verifies each extracted causal logic chain according to the preset integrity rules: check whether each structural node is associated with at least one functional node; check whether each functional node is associated with at least one failure mode node; check whether each failure mode node is associated with both the failure cause and failure impact nodes; when any logic chain is found to have missing nodes or interrupted associations, record the type and location information of the missing nodes.

[0132] S43. If a missing node exists, infer the set of attributes of the missing node based on the context of the causal logical chain;

[0133] Step S43: For the detected logical missing, the system performs intelligent inference based on the context of the causal logical chain: analyzes the position and role of the missing node in the logical chain; extracts the attribute features and relationship patterns of upstream and downstream related nodes; refers to existing logical chain patterns with similar structures or functions; and generates a set of candidate attributes for the missing node, including core attributes such as node name, type, and description.

[0134] S44. Generate knowledge node completion suggestions, wherein the completion suggestions include the attributes of the missing node and the insertion position;

[0135] In step S44, based on the inference results, the system automatically generates structured knowledge node completion suggestions. Each suggestion includes: the complete set of attributes of the missing node; the exact insertion position in the knowledge graph; the definition of the relationship with upstream and downstream nodes; the completion confidence score; and the explanation of the reasoning basis.

[0136] S45. Verify the logical consistency of the completion suggestion using preset verification rules;

[0137] Step S45 rigorously verifies the completion suggestions using preset verification rules: Semantic consistency check: ensures that the new node is semantically consistent with the context; Logical conflict detection: avoids circular dependencies or contradictory relationships; Integrity verification: ensures that the completed logical chain conforms to the FMEA standard specification; Risk assessment: predicts the impact of the completion node on the existing risk priority.

[0138] S46. Store the completed suggestions that pass the verification to the completion suggestion database.

[0139] Step S46 stores the verified completion suggestions in a standardized format to the completion suggestion database, including: the unique identifier and generation timestamp of the suggestion; the corresponding missing logical chain information; the complete node attributes and relationship definitions; the verification result and confidence score; and the processing status (pending review / adopted / rejected).

[0140] This embodiment, through automated causal logic chain integrity verification, can systematically discover logical gaps in the knowledge graph and provide verified intelligent completion suggestions, significantly improving the completeness and reliability of FMEA analysis and providing a more comprehensive knowledge foundation for quality risk management. Simultaneously, rigorous logical consistency verification ensures that all completion suggestions remain consistent with the existing knowledge system. In step S4, by systematically traversing and verifying the "structure-function-failure" causal logic chain in the knowledge graph, logical breakpoints such as missing functions or failure modes without corresponding causes can be automatically and accurately identified. Based on contextual intelligence, complete and accurately positioned completion suggestions are generated. This mechanism transforms the previously reliant and easily overlooked logical integrity review into an automated and fully-covered intelligent process. This not only greatly improves the rigor and review efficiency of FMEA analysis but also fundamentally guarantees the quality and reliability of the completed knowledge by verifying the logical consistency of suggestions through preset rules, significantly enhancing the logical completeness and decision support value of the FMEA knowledge graph.

[0141] S5. When a missing link in the causal logic chain is identified, corresponding knowledge node completion suggestion data is generated;

[0142] Step S5 leverages the powerful reasoning capabilities of a large language model to generate high-quality knowledge completion suggestions. Further, step S5 may include the following steps:

[0143] S51. Extract upstream and downstream node information that are related to the missing part of the causal logic chain from the FMEA knowledge graph to form a reasoning context;

[0144] Step S51: When the system identifies a missing causal logic chain, it first extracts upstream and downstream node information closely related to the missing location from the FMEA knowledge graph to construct a rich reasoning context. Specifically, this includes: extracting the direct upstream node of the missing location (e.g., for a missing "failure mode," extracting its corresponding "function" node); extracting the direct downstream node of the missing location (e.g., for a missing "failure cause," extracting its corresponding "failure impact" node); extracting analogical references for similar nodes (e.g., typical failure modes corresponding to similar functions); and extracting relevant technical specifications and requirements document fragments.

[0145] S52. The completion suggestion data is generated by using a third language model based on the reasoning context.

[0146] The third language model in this embodiment adopts a large language model.

[0147] Furthermore, S52 inputs the constructed reasoning context as a prompt to the third language model (a large language model is used in this embodiment). The model generates completion suggestion data based on the following process: Semantic understanding and reasoning: The model deeply understands the technical semantics and logical relationships in the context and identifies the necessary features of the missing node; Multiple candidate generation: Based on domain knowledge and logical reasoning, multiple possible completion suggestion candidates are generated; Quality assessment: Each candidate suggestion is internally evaluated, and the generation confidence score is calculated. Structured output: Standardized completion suggestion data is output, including: a complete attribute description of the suggestion node; the association method with upstream and downstream nodes; a brief description of the generation basis and reasoning process; confidence score and quality assessment index.

[0148] This embodiment fully leverages the deep reasoning capabilities of a large language model to generate high-quality completion suggestions that conform to technical logic based on the contextual information of the knowledge graph. This significantly improves the accuracy and efficiency of knowledge completion, effectively solving the problems of low efficiency and poor consistency caused by reliance on human experience in traditional methods. In a preferred embodiment, the third language model is specifically trained on FMEA domain texts, possessing deep domain knowledge, ensuring that the generated completion suggestions not only conform to technical specifications but also remain consistent with the enterprise's specific terminology system.

[0149] In this embodiment, when a functional node is found to lack a corresponding failure mode, the system will generate possible failure mode suggestions based on the description of the function and the historical failure modes of similar functions.

[0150] In step S5, by employing a large language model to perform deep reasoning on the upstream and downstream contexts of the missing causal logic chain, accurate and reasonable knowledge node completion suggestions can be intelligently generated, achieving a leap from "passive discovery of missing information" to "proactive intelligent completion." This method effectively utilizes the historical patterns and semantic associations contained in the knowledge graph, significantly reducing the over-reliance on personal experience in the completion work. It not only greatly improves the efficiency of FMEA knowledge construction and improvement, but also ensures the quality and logical consistency of the completed content through data-driven intelligent suggestions, enhancing the completeness and reliability of the knowledge system.

[0151] S6. Present the FMEA knowledge graph and the completion suggestion data to the user through an interactive interface;

[0152] In this embodiment, users can view the structure of the knowledge graph, the relationships between nodes, and the completion suggestions generated by the system through the interface.

[0153] S7. Receive and update the FMEA knowledge graph according to user instructions.

[0154] In this embodiment, users can accept or reject system-generated completion suggestions, modify node attributes, or manually add new knowledge nodes. These operations will update the FMEA knowledge graph in real time.

[0155] In step S7, the user can receive confirmation instructions for the completion suggestions or custom editing instructions issued by the user through drag / drop / link operations via the visual interface; in response to the confirmation instructions or custom editing instructions, the FMEA database is updated.

[0156] Step S7 establishes a human-computer collaborative closed loop of "system intelligent recommendation and human review and decision-making" by receiving and responding to user confirmation, modification, or custom editing instructions for completion suggestions. This fully utilizes the system's computational efficiency and reasoning capabilities while ensuring the dominant role of human expert experience in key decisions. This interactive mechanism not only makes the FMEA knowledge graph update process intuitive and efficient but also guarantees the accuracy and professionalism of all knowledge entered into the database, ultimately achieving dynamic optimization and reliable evolution of the knowledge graph through continuous human-computer interaction.

[0157] In this embodiment, the intelligent interaction and completion method for FMEA knowledge first acquires unstructured or semi-structured FMEA documents and extracts FMEA element information to determine structured knowledge nodes. This achieves efficient digital conversion of historical documents from different standard templates, solving the fundamental problem that a large amount of FMEA knowledge cannot be transferred into the intelligent system. Subsequently, by semantically aligning and fusing the newly extracted knowledge nodes with the historical structure tree in the FMEA knowledge graph and resolving semantic conflicts, it effectively ensures that knowledge from different departments or periods with different expressions can be integrated into a unified and consistent knowledge system. Furthermore, based on the causal logic chain composed of "structure-function-failure", the integrity of the fused knowledge graph is verified. When a missing causal logic chain is identified, knowledge node completion suggestions are automatically generated, significantly improving the quality and logical rigor of the FMEA analysis itself. Finally, the complete knowledge graph and intelligent completion suggestions are presented to the user through an interactive interface, and user instructions are received to update the graph, forming a closed loop of "system intelligent suggestions and manual review and decision-making". This not only allows FMEA knowledge to be viewed and verified intuitively but also supports its continuous iterative optimization. In summary, this application significantly improves knowledge management efficiency and analysis quality.

[0158] Example 2

[0159] In this embodiment, after performing step S5, the following may also be included:

[0160] A. Using a fourth language model based on historical data in the FMEA knowledge graph, predict the score value after the completion suggestion data is adopted; proceed to step S6.

[0161] Furthermore, in step S6, while presenting the FMEA knowledge graph and the completion suggestion data to the user through an interactive interface, the scoring value can also be presented.

[0162] In this embodiment, the score may include predictions of the severity, frequency, and detectability of the newly added node, and provide corresponding confidence intervals.

[0163] The fourth language model in this embodiment is a risk prediction model, which uses a large language model.

[0164] In this embodiment, by introducing a risk prediction model based on historical data, the risk level of the supplementary suggestions can be assessed in a forward-looking manner, providing key quantitative basis for core decisions. By automatically predicting and displaying the scores and confidence intervals of the new nodes in terms of severity, frequency, and detectability, this function effectively reduces the subjective dependence of risk assessment on personal experience, enabling users to identify high-priority risks and assess the reliability of suggestions. This significantly improves the scientific nature and decision-making efficiency of the FMEA knowledge supplementation process, ensuring that the knowledge graph maintains high-quality risk insight throughout its evolution.

[0165] Furthermore, some embodiments of this application also provide an electronic device. The electronic device can be various forms of digital computer, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, etc. The electronic device can also be various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices.

[0166] The electronic device includes: one or more processors; and a memory storing computer program instructions that, when executed, cause the processor to perform the steps of the methods provided in any one or more of the above embodiments. Figure 4 An exemplary structural diagram of the electronic device is disclosed. For example... Figure 4 As shown, the electronic device includes one or more processors 1101, a memory 1102, and interfaces for connecting the components, including high-speed interfaces and low-speed interfaces. The components are interconnected via different buses and can be mounted on a common motherboard or otherwise as required. The processors can process instructions executed within the electronic device, including instructions stored in or on memory to display graphical information of a GUI on an external input / output device (such as a display device coupled to the interface). In some other embodiments, multiple processors and / or multiple buses can be used with multiple memories and multiple memory modules, if desired. Similarly, multiple electronic devices can be connected, each providing some of the necessary operations (e.g., as a server array, a group of blade servers, or a multiprocessor system). The components, their connections and relationships, and their functions shown herein are merely examples and are not intended to limit the implementation of the present application described and / or claimed herein.

[0167] The electronic device may further include an input device 1103 and an output device 1104. The processor 1101, memory 1102, input device 1103, and output device 1104 may be connected via a bus or other means. Figure 4 Taking the example of a connection between China and Israel via a bus.

[0168] Input device 1103 can receive input numerical or character information, and generate key signal inputs related to user settings and function control of the electronic device, such as a touch screen, keypad, mouse, trackpad, touchpad, joystick, one or more mouse buttons, trackball, joystick, etc. Output device 1104 may include a display device, auxiliary lighting device (e.g., LED), and haptic feedback device (e.g., vibration motor). The display device may include, but is not limited to, a liquid crystal display (LCD), a light-emitting diode (LED) display, and a plasma display. In some embodiments, the display device may be a touch screen.

[0169] To provide interaction with the user, the electronic device can be a computer. The computer has: a display device for displaying information to the user (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0170] In this embodiment, a computer-readable medium stores a computer program / instructions that, when executed by a processor, implement the steps of the methods provided in any one or more of the above embodiments. This computer-readable medium may be included in the electronic device described in the above embodiments; or it may exist independently and not assembled into that device. The aforementioned computer-readable medium carries one or more computer-readable instructions.

[0171] The memory 1102 can serve as a non-transitory computer-readable storage medium, used to store non-transitory software programs, non-transitory computer-executable programs, and modules. The processor 1101 executes various functional applications and data processing of the server by running the non-transitory software programs, instructions, and modules stored in the memory 1102, thereby implementing the program instructions / modules corresponding to the methods provided in any one or more of the embodiments described above in this application.

[0172] The memory 1102 may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function; the data storage area may store data created based on the use of the electronic device. Furthermore, the memory 1102 may include high-speed random access memory and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, the memory 1102 may optionally include memory remotely located relative to the processor 1101, and these remote memories can be connected to the electronic device via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0173] It should be noted that the computer-readable medium described in this application can be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this application, a computer-readable medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0174] Computer-readable media include permanent and non-permanent, removable and non-removable media, which can store information by any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, read-only optical disc (CD-ROM), digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transfer medium that can be used to store information accessible by a computing device.

[0175] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0176] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. For example, it can be implemented using an application-specific integrated circuit (ASIC), a general-purpose computer, or any other similar hardware device. In some embodiments, the software program of this application can be executed by a processor to implement the steps or functions described above. Similarly, the software program of this application (including related data structures) can be stored in a computer-readable recording medium, such as RAM memory, magnetic or optical drives, floppy disks, or similar devices. Furthermore, some steps or functions of this application can be implemented in hardware, for example, as circuitry that works with a processor to perform the various steps or functions.

[0177] The computer program product provided in this application includes one or more computer programs / instructions. When executed by a processor, these computer programs / instructions generate, in whole or in part, the processes or functions described in this application. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions may be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium may be any available medium accessible to a computer or a data storage device such as a server or data center that integrates one or more available media. The available medium may be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid-state disk (SSD)).

[0178] The flowcharts or block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of devices, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented using a dedicated hardware-specific system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0179] The scope of this application is defined by the appended claims rather than the foregoing description, and is therefore intended to encompass all variations falling within the meaning and scope of equivalents of the claims. No reference numerals in the claims should be construed as limiting the scope of the claims. Furthermore, it is clear that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units or devices recited in a device claim may also be implemented by a single unit or device in software or hardware. Terms such as "first," "second," etc., are used only for distinguishing descriptions and do not indicate any particular order, nor should they be construed as indicating or implying relative importance.

[0180] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily made by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims, and the above embodiments should be regarded as exemplary and non-limiting.

Claims

1. A method for intelligent interaction and completion of FMEA knowledge, characterized in that, include: Obtain unstructured or semi-structured FMEA documents; Extract the FMEA element information from the FMEA document to determine the structured knowledge nodes; The knowledge nodes are semantically aligned and fused with the historical structure tree in the FMEA knowledge graph, and semantic conflicts are resolved. Based on the causal logic chain consisting of structure-function-failure, the integrity of the fused FMEA knowledge graph is verified. When a missing link in the causal logic chain is identified, corresponding knowledge node completion suggestion data is generated. The FMEA knowledge graph and the completion suggestion data are presented to the user through an interactive interface. The FMEA knowledge graph is updated by receiving and updating user instructions.

2. The intelligent interaction and completion method for FMEA knowledge according to claim 1, characterized in that, The step of extracting FMEA element information from the FMEA document and determining structured knowledge nodes includes: Identify and parse the text, and / or tables, and / or diagrams, and / or flowcharts in the FMEA document; The context of the text, and / or the table, and / or the block diagram, and / or the flowchart are transformed into the association between the knowledge nodes.

3. The intelligent interaction and completion method for FMEA knowledge according to claim 1, characterized in that, The step of semantically aligning and fusing the knowledge nodes with the historical structure tree in the FMEA knowledge graph, and resolving semantic conflicts, includes: Using a first language model, the semantic similarity between the knowledge node and the historical structure tree in the FMEA knowledge graph is calculated; Based on the semantic similarity, conflicting nodes with the same semantics but inconsistent expressions are identified, and conflict alarm messages are generated. In response to the conflict alarm message, a second language model is used to analyze the context of the conflict node, generate standard node terms, and store the standard node terms in the FMEA knowledge graph.

4. The intelligent interaction and completion method for FMEA knowledge according to claim 3, characterized in that, The step of employing a first language model to calculate the semantic similarity between the knowledge node and the historical structure tree in the FMEA knowledge graph includes: The attribute set of the knowledge node is vectorized using the first language model to generate a first feature vector; Obtain the attribute set of the historical structure tree in the FMEA knowledge graph and generate a second feature vector; Calculate the semantic similarity score between the first feature vector and the second feature vector; If the semantic similarity score is greater than a preset threshold, then the semantic association between the knowledge node and the historical structure tree is determined. The semantic similarity scores are stored in a similarity matrix; Based on the similarity matrix, potential conflict relationships between the knowledge nodes and the historical structure tree are determined.

5. The intelligent interaction and completion method for FMEA knowledge according to claim 4, characterized in that, The process of identifying conflicting nodes with the same semantics but inconsistent expressions based on the semantic similarity and generating conflict alarm messages includes: Extract the node pairs in the similarity matrix whose semantic similarity scores are higher than the preset threshold; For each node pair, the attribute set of the knowledge node is compared with that of the historical structure tree to determine whether there is a semantic conflict due to inconsistent expression; if a semantic conflict exists, the knowledge node is marked as a conflict node. Generate a conflict alarm message, which includes the identifier of the conflicting node, the conflict type, and the conflict description; The conflict alarm message is stored in the log database.

6. The intelligent interaction and completion method for FMEA knowledge according to claim 1, characterized in that, When a missing link in the causal logic chain is identified, corresponding knowledge node completion suggestion data is generated, including: From the FMEA knowledge graph, extract the upstream and downstream node information that is related to the missing part of the causal logic chain to form the reasoning context; The completion suggestion data is generated by using a third language model based on the reasoning context.

7. The intelligent interaction and completion method for FMEA knowledge according to claim 1, characterized in that, When a missing causal logic chain is identified, corresponding knowledge node completion suggestion data is generated, which further includes: The fourth language model is used to predict the score value after the completion suggestion data is adopted based on historical data in the FMEA knowledge graph; The step of presenting the FMEA knowledge graph and the completion suggestion data to the user through an interactive interface also includes: The rating values ​​are presented.

8. An electronic device, characterized in that, The electronic device includes: One or more processors; and A memory storing computer program instructions, which, when executed, cause the processor to perform the steps of the intelligent interaction and completion method for FMEA knowledge as described in any one of claims 1 to 7.

9. A computer-readable medium having a computer program / instructions stored thereon, characterized in that, When the computer program / instruction is executed by the processor, it implements the steps of the intelligent interaction and completion method of FMEA knowledge as described in any one of claims 1 to 7.

10. A computer program product comprising a computer program / instructions, characterized in that, When the computer program / instruction is executed by the processor, it implements the steps of the intelligent interaction and completion method of FMEA knowledge as described in any one of claims 1 to 7.

Images