An airport maintenance knowledge question and answer method, system, electronic device and storage medium

Abstract

This application provides a method, system, electronic device, and storage medium for airport maintenance knowledge question answering, belonging to the field of airport maintenance technology. The method involves: acquiring a user's query question; identifying entities and entity states within the query question, and constructing initial evidence based on the identified entities and entity states; using the initial evidence as the starting point for reasoning, performing structured reasoning within a pre-constructed airport maintenance knowledge graph according to an ontology constraint framework to obtain structured reasoning results; simultaneously extracting original text fragments associated with the entities in the structured reasoning results; and fusing the structured reasoning results and original text fragments to generate the answer to the corresponding query question. This invention establishes domain knowledge constraints based on an ontology constraint framework, combines structured domain constraints and an airport maintenance knowledge graph for structured reasoning, and incorporates original text fragments, effectively overcoming the shortcomings of traditional knowledge retrieval methods, such as insufficient relevance and lack of specificity, leading to difficulties in obtaining intended knowledge.

Description

Technical Field

[0001] This invention relates to the field of airport maintenance technology, and more specifically, to an airport maintenance knowledge Q&A method, system, electronic device, and storage medium. Background Technology

[0002] As airport operations and maintenance gradually transform towards digitalization and intelligence, the traditional methods of maintaining and repairing airport electromechanical systems (such as boarding bridges and baggage handling systems) by relying on scattered maintenance manuals, equipment manuals, and operation manuals are clearly no longer sufficient to meet the needs.

[0003] Existing technologies often employ document indexing systems to build maintenance knowledge bases, locate document fragments through keyword matching, and rely on manually preset rules to generate question-and-answer results for maintenance-related issues.

[0004] This approach only creates a shallow file index, exacerbating the fragmentation of maintenance and repair knowledge. Furthermore, the search results lack specificity and relevance, requiring users to perform secondary filtering of the answers, resulting in low efficiency and difficulty in obtaining the desired knowledge. Summary of the Invention

[0005] In view of this, the purpose of the present invention is to provide an airport maintenance knowledge Q&A method, system, electronic device and storage medium to overcome the problem that the above-mentioned search results lack specificity and relevance.

[0006] Firstly, a question-and-answer method for airport maintenance knowledge is provided, including:

[0007] Obtain the user's query question;

[0008] Identify the entities and entity states in the query question, and construct initial evidence based on the identified entities and entity states;

[0009] Using initial evidence as the starting point for reasoning, structured reasoning is performed in the pre-constructed airport maintenance knowledge graph according to the ontology constraint framework to obtain structured reasoning results; at the same time, the original text fragments associated with the entities in the structured reasoning results are extracted.

[0010] The structured reasoning results and original text fragments are combined to generate the answer to the corresponding query question.

[0011] Optionally, the initial evidence based on the identified entities and their states includes:

[0012] Determine the entity name based on the identified entity;

[0013] The Boolean value of the entity state is determined based on the identified entity state; where the entity state includes occurrence and non-occurrence;

[0014] The entity types corresponding to the identified entities are determined based on a pre-built ontology constraint framework. The ontology constraint framework includes four entity types and three entity relationships. The four entity types are troubleshooting, fault cause, fault event, and maintenance operation event. The three entity relationships are suggestion, cause, and belong. The entity relationship between troubleshooting and fault cause is suggestion; the entity relationship between fault cause and fault event is cause; and the entity relationship between fault event and maintenance operation event is belong.

[0015] The entity type, entity name, and entity status are encapsulated as Boolean values ​​to form initial evidence.

[0016] Optionally, using the initial evidence as the starting point for reasoning, structured reasoning is performed within the pre-constructed airport maintenance knowledge graph according to the ontology constraint framework, yielding structured reasoning results including:

[0017] Using the initial evidence as the starting point for reasoning, structured reasoning is performed in the pre-constructed airport maintenance knowledge graph according to the entity relationships in the ontology constraint framework to obtain all entities and entity relationships associated with the entities in the query question.

[0018] The entities and entity relationships obtained through reasoning are organized hierarchically according to the ontology constraint framework to obtain structured reasoning results.

[0019] Optionally, the method further includes:

[0020] Obtain the entity attributes of each entity obtained through structured reasoning in the airport maintenance knowledge graph;

[0021] The rendering output method of each entity obtained through reasoning is determined based on the preset correspondence between entity attributes and rendering output methods.

[0022] Output the answer according to the rendering output method.

[0023] Optionally, entity attributes include basic attributes, link attributes, descriptive attributes, and attributes without extensions; the preset correspondence between entity attributes and rendering output methods includes:

[0024] When entity attributes are basic attributes, the rendering output method is to display the label text content;

[0025] When the entity attribute is a link attribute, the rendering output is red text + bound link, and the bound link is used to jump to external multimodal resources;

[0026] When the entity attribute is a descriptive attribute, the rendering output method is a floating text box;

[0027] When an entity has no extended attributes, the rendering output will display static orange text.

[0028] Optionally, the process of constructing an airport maintenance knowledge graph includes:

[0029] Construct an ontological constraint framework for airport maintenance;

[0030] Extracting entities and entity relationships from historical airport maintenance documents based on ontology constraint framework;

[0031] Based on the pre-defined entity normalization and encapsulation model, the extracted entities are mapped to entity nodes of the airport maintenance knowledge graph.

[0032] Based on a pre-defined relational normalization encapsulation model, the extracted entity relations are mapped to edges of the airport maintenance knowledge graph.

[0033] Optionally, the method further includes:

[0034] Configure an external database preloading interface for entity types in the ontology constraint framework; the external database should include at least a 3D maintenance plan library.

[0035] Bind instances from external databases to entity nodes in the airport maintenance knowledge graph and generate unique identifiers;

[0036] External links for entity nodes are generated based on the address and unique identifier of the preloaded interface.

[0037] Secondly, an airport maintenance knowledge Q&A system is provided, including:

[0038] The retrieval unit is used to retrieve the user's query question;

[0039] The identification unit is used to identify entities and entity states in the query question, and to construct initial evidence based on the identified entities and entity states.

[0040] The reasoning unit is used to perform structured reasoning in a pre-constructed airport maintenance knowledge graph according to the ontology constraint framework, starting from the initial evidence, and to obtain the structured reasoning result; at the same time, it extracts the original text fragments associated with the entities in the structured reasoning result.

[0041] The generation unit is used to merge the structured reasoning results and the original text fragments to generate the answer to the corresponding query question.

[0042] Thirdly, an electronic device is provided, including a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus;

[0043] The memory is used to store computer programs;

[0044] When the processor executes a program stored in the memory, it implements any of the methods described in the first aspect.

[0045] Fourthly, a computer-readable storage medium is provided, wherein a computer program is stored therein, and the computer program, when executed by a processor, implements any of the methods described in the first aspect.

[0046] This invention provides a method, system, electronic device, and storage medium for airport maintenance knowledge question answering. The method involves: acquiring a user's query question; identifying entities and their states within the query question, and constructing initial evidence based on these entities and states; using this initial evidence as the starting point for reasoning; performing structured reasoning within a pre-constructed airport maintenance knowledge graph according to an ontology constraint framework to obtain a structured reasoning result; simultaneously extracting original text fragments associated with the entities in the structured reasoning result; and fusing the structured reasoning result and the original text fragments to generate an answer to the corresponding query question. This invention addresses the problems of scattered knowledge and insufficient targeting in knowledge queries by establishing domain knowledge constraints based on an ontology constraint framework, combining structured domain constraints and an airport maintenance knowledge graph for structured reasoning, and integrating original text fragments. This effectively overcomes the shortcomings of traditional knowledge retrieval methods, such as insufficient relevance and lack of targeting, which makes it difficult to obtain intended knowledge.

[0047] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0048] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0049] Figure 1 The flowchart of an airport maintenance knowledge Q&A method provided by an embodiment of the present invention is shown;

[0050] Figure 2 This diagram illustrates the structure of the entity state reasoning encapsulation specification provided in an embodiment of the present invention.

[0051] Figure 3 A schematic diagram of the structure of the ontological constraint frame provided in an embodiment of the present invention is shown;

[0052] Figure 4 This diagram illustrates the structure of the entity normalization encapsulation model provided in an embodiment of the present invention.

[0053] Figure 5This diagram illustrates the structure of the relation normalization encapsulation model provided in an embodiment of the present invention.

[0054] Figure 6 This diagram illustrates the structure of an airport maintenance knowledge Q&A system provided in an embodiment of the present invention.

[0055] Figure 7 A schematic diagram of the structure of an electronic device provided in an embodiment of the present invention is shown. Detailed Implementation

[0056] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0057] Considering that existing technologies mostly use document indexing systems to build maintenance knowledge bases, this approach locates document fragments through keyword matching and relies on manually preset rules to generate question-and-answer results for maintenance-related issues.

[0058] This approach only creates a shallow file index, exacerbating the fragmentation of maintenance and repair knowledge. Furthermore, the search results lack specificity and relevance, requiring users to perform secondary filtering of the answers, resulting in low efficiency and difficulty in obtaining the desired knowledge.

[0059] Based on this, embodiments of the present invention provide an airport maintenance knowledge Q&A method and apparatus, which are described below through embodiments.

[0060] This invention provides a method for answering airport maintenance knowledge questions, wherein the execution subject of this method is a server, such as... Figure 1 As shown, it includes the following steps:

[0061] Step S101: Obtain the user's query question.

[0062] Users can enter their questions on the airport maintenance-related mobile or desktop clients.

[0063] Step S102: Identify the entities and entity states in the query question, and construct initial evidence based on the identified entities and entity states.

[0064] After obtaining the user's query, pre-trained language models (such as BERT, ChatGLM, ERNIE, etc.) can be used to perform deep semantic analysis on the user's query, automatically identifying the entities and entity states in the question.

[0065] Based on the identified entity and entity state, the entity and entity state are encapsulated using a pre-configured entity state reasoning encapsulation specification to establish an initial state baseline for fault diagnosis and maintenance operation queries.

[0066] Specifically, the entity state reasoning encapsulation specification is as follows: Figure 2 As shown, the entity name, entity type, and entity status are encapsulated in JSON format, where the entity status indicates whether the entity has occurred.

[0067] In one feasible implementation, the initial evidence based on the identified entity and its state includes:

[0068] Step S102A: Determine the entity name based on the identified entity.

[0069] Step S102B: Determine the Boolean value of the entity state based on the identified entity state.

[0070] The entity status includes "occurred" and "not occurred". For example, the boolean value for "occurred" can be set to True, and the boolean value for "not occurred" can be set to False.

[0071] Step S102C: Determine the entity type corresponding to the identified entity based on the pre-built ontology constraint framework.

[0072] An ontology is a conceptual model abstracted from the objective world. This model contains basic terms and relationships between terms (or concepts and relationships between concepts) within a specific discipline. Therefore, the ontology constraint framework of this invention provides the basic terms and relationships constituting the vocabulary of the airport maintenance domain, as well as the definitions of rules governing the extension of these terms and relationships, specifically for the airport maintenance scenario.

[0073] like Figure 3 The diagram shows a schematic representation of the ontology constraint framework; a four-layer ontology framework is used to standardize and constrain knowledge in the airport maintenance domain. Ontology Constraint Framework It includes four entity types and three entity relationships. The entity types are divided into four core concepts: maintenance operation items, fault items, fault causes and troubleshooting, and three entity relationships are defined: "belongs to", "cause", and "recommend", forming a logical link for fault propagation and response.

[0074] Specifically, the entity relationship between troubleshooting and the cause of the fault is "suggestion"; the entity relationship between the cause of the fault and the fault event is "cause"; and the entity relationship between the fault event and the maintenance operation event is "belongs to".

[0075] This ontology constraint framework can be represented as follows: ;

[0076] Among them, entity set for: ;

[0077] In the formula: For maintenance and operation-related entities; For entities of the fault item type; For entities related to the cause of the failure; This is a troubleshooting / maintenance entity.

[0078] The set of relational concepts between different ontological concepts for: ;

[0079] In the formula: The malfunction falls under the category of maintenance and repair procedures. ;

[0080] The cause of the malfunction is the reason for the occurrence of the malfunction phenomenon. ;

[0081] Suggest possible causes for troubleshooting and repair. .

[0082] This invention constructs a four-layer maintenance ontology framework (MTO): through entity layering of maintenance operation items, fault items, fault causes, and troubleshooting, and a "belongs to - leads to - recommends" relationship chain, it builds a logical constraint framework for airport maintenance knowledge, laying the foundation for standardization.

[0083] Step S102D: Encapsulate the Boolean values ​​of entity type, entity name, and entity status to form initial evidence.

[0084] For example, when an airport maintenance worker asks a question such as, "The brake system warning light came on when aircraft B-1234 was taxiing," the trained language model identifies the entity "brake system," and the entity name is determined to be "brake system." The status is judged as "warning light on," indicating a malfunction has occurred, so the entity status is "occurred." The associated entity type is "malfunction event."

[0085] Step S103: Using the initial evidence as the starting point for reasoning, perform structured reasoning in the pre-constructed airport maintenance knowledge graph according to the ontology constraint framework to obtain the structured reasoning result; at the same time, extract the original text fragments associated with the entities in the structured reasoning result.

[0086] In this embodiment of the invention, the process of constructing an airport maintenance knowledge graph includes the following steps:

[0087] Step 1: Construct the ontological constraint framework for airport maintenance.

[0088] The ontology constraint framework in this step is similar to the example above and will not be repeated here.

[0089] Step 2: Extract entities and entity relationships from historical airport maintenance documents based on ontology constraint framework.

[0090] In this step, historical airport maintenance documents include maintenance manuals, equipment manuals, and operation manuals.

[0091] In one example, named entity recognition technology can be used to identify entities such as proper nouns and meaningful quantity phrases in a document. Based on the entity types and extraction rule set defined in the ontology constraint framework, the text is matched to identify entities that satisfy the rules. For example, entities representing fault conditions such as "engine cannot start" and "brake vibration" can be identified.

[0092] Step 3: Based on the preset entity normalization and encapsulation model, the extracted entities are mapped to entity nodes of the airport maintenance knowledge graph.

[0093] like Figure 4 The diagram shown is a JSON-formatted structural diagram of the entity normalization encapsulation model. The entity normalization encapsulation model includes the entity's unique identifier UUID (Universally Unique Identifier), entity name, entity type (from MTO), and entity extended attributes.

[0094] According to this specification, entities identified in historical airport maintenance documents are encapsulated and stored as entity nodes in the airport maintenance knowledge graph.

[0095] Step 4: Based on the preset relation normalization encapsulation model, the extracted entity relations are mapped to the edges of the airport maintenance knowledge graph.

[0096] Figure 5 The diagram shows the structure of the relation normalization encapsulation model in JSON format. The relation normalization encapsulation model includes the relation's unique identifier UUID, target entity, source entity, relation type (from MTO), and relation extended attributes, ensuring the unified expression and standardized storage of knowledge elements.

[0097] According to this specification, the entity relationships identified in historical airport maintenance documents are encapsulated and stored as edges of the airport maintenance knowledge graph.

[0098] This invention integrates fragmented knowledge into an organic knowledge graph; it achieves structured reasoning through graph topology deduction, thereby improving the relevance and relevance of knowledge retrieval and reducing the workload of secondary screening by on-site personnel.

[0099] Existing question-and-answer solutions that rely on manually preset templates provide answers limited to text descriptions, unable to connect to multimedia resources such as 3D dynamic plans, and unable to meet the needs of spatial operation guidance such as equipment disassembly and assembly.

[0100] Therefore, this invention proposes a standardized linking mechanism. This mechanism achieves multimodal knowledge fusion through hierarchical mapping between entity ontology and external knowledge base, associating internal knowledge entities with external multimodal resources (external videos, manuals, images, 3D models, etc.) to realize multi-dimensional resource presentation. Specifically, this standardized linking mechanism includes the following steps:

[0101] Step 5: Configure the external database preloading interface for entity types in the ontology constraint framework.

[0102] External databases should include at least a 3D maintenance contingency plan library, PDF maintenance manuals, troubleshooting videos, 3D equipment disassembly diagrams, and technical standard documents.

[0103] In a knowledge graph, each specific object is a "node". For example, "landing gear cannot be retracted" is an entity node of the "fault item" type.

[0104] This invention employs a preloading mechanism to establish a connection between an external multimodal knowledge base and specific type entity nodes in advance, achieving pre-binding loading of entity nodes with the external knowledge base. The preloading mechanism means preparing resources in advance and storing them in memory or cache for rapid response, rather than waiting until a user queries resources.

[0105] For example, when the airport maintenance knowledge Q&A system is launched, the system knows that "hydraulic oil leakage" is a common fault. Therefore, when the system starts, it pre-loads the "Hydraulic System Maintenance Manual" PDF, an instructional video on "How to Replace Seals," and related 3D structural diagrams. When someone queries "hydraulic oil leakage," these resources can be accessed at any time, reducing loading time.

[0106] Step 6: Bind the instances in the external database to the entity nodes in the airport maintenance knowledge graph and generate unique identifiers.

[0107] Preloading only prepares the resource pool, but it doesn't bind which resource to which user. Therefore, this step uses link generation and binding technology, specifically dynamic retrieval and UUID binding, to map specific instances in the preloaded knowledge base to specific entity nodes in the graph. This completes the static contract binding of "entity type - knowledge source." This mechanism ensures that entity types acquire the foundation for cross-database knowledge expansion capabilities during the definition phase.

[0108] In order to accurately associate a specific external resource (such as a segment of a video) with a specific entity in the knowledge graph (such as a fault record), the system assigns a unique UUID number to each "instance" of the external resource, and then connects this number with the entity node in the knowledge graph to form a one-to-one "knowledge link".

[0109] For example, the preloaded 3D contingency plan library contains 100 troubleshooting videos, and the link generation technology determines that the entity node "engine overheating" is bound to the specific video with video ID inst-7a8b9c.

[0110] Step 7: Generate external links for entity nodes based on the address and unique identifier of the preloaded interface.

[0111] After identifying the target instance, an external link is generated using the standardized template [link] service prefix / instance access path#unique identifier. The service prefix is ​​inherited from the static binding relationship configured for the entity type; that is, all entities of the "fault item" type use the same knowledge base prefix. This ensures system consistency and facilitates unified migration or replacement of the resource library later.

[0112] The instance access path is the address of the preloaded interface, and the unique identifier corresponds to the permanent identifier of a specific instance in the knowledge base.

[0113] This link is stored as a core attribute in the entity node metadata. In the subsequent actual interaction stage, a dedicated parsing engine enables precise retrieval from the graph node to multimodal knowledge content (such as the action sequence of the 3D plan, the execution object, etc.).

[0114] In one feasible implementation, initial evidence is used as the starting point for reasoning. Structured reasoning is performed within a pre-constructed airport maintenance knowledge graph according to an ontology constraint framework, yielding the following structured reasoning results:

[0115] Step S103A: Using the initial evidence as the starting point for reasoning, perform structured reasoning in the pre-constructed airport maintenance knowledge graph according to the entity relationships in the ontology constraint framework to obtain all entities and entity relationships associated with the entities in the query question.

[0116] In this step, a pre-built airport maintenance knowledge graph is used to perform graph traversal and path reasoning. For example, starting from "engine cannot start", the system traces upwards or downwards along relationship chains such as "cause → cause of failure" and "belongs to → subsystem".

[0117] Activate the fault propagation model, which is a causal reasoning model that simulates how a fault gradually affects the top-level system from the bottom-level components.

[0118] Using this fault propagation model, all possible fault paths are evaluated. A three-dimensional inference diagnostic conclusion is obtained. Here, "three-dimensional" does not refer to spatial dimensions, but rather to three state classifications, constituting a diagnostic confidence dimension. Specifically, the three-dimensional diagnostic conclusion includes the identified fault group, the ruled-out fault group, and the pending-confirmation fault group. The identified fault group refers to a highly credible cause of the fault based on current evidence; the ruled-out fault group refers to cases where there is evidence indicating that the fault is unlikely to occur; and the pending-confirmation fault group refers to cases where the fault is possible, but there is currently no direct evidence to support it.

[0119] Step S103B: Organize the entities and entity relationships obtained through reasoning in a hierarchical manner according to the ontology constraint framework to obtain the structured reasoning results.

[0120] All reasoning results are organized according to predefined entity types, such as maintenance operation items, fault items, fault causes, and troubleshooting, forming a hierarchical structure that facilitates subsequent display and interaction.

[0121] Step S104: Merge the structured reasoning results and the original text fragments to generate the answer to the corresponding query question.

[0122] In one example, continuing from the previous one, starting from "engine cannot start", reasoning along the relationship chain of "cause → cause of failure" and "suggestion → maintenance measures" in the airport maintenance knowledge graph will yield the following structured reasoning result:

[0123] Fault: The engine cannot be started;

[0124] Possible causes: Damaged fuel pump, blocked fuel supply line;

[0125] Recommended measures: Replace the fuel pump and clean the fuel filter;

[0126] The original text snippet can be extracted from the associated service manuals and historical work orders. Specifically, text slicing techniques (such as sliding windows and paragraph splitting) can be used to divide the document into several contextual segments, and then the segments containing key entities (such as "fuel pump failure") can be selected. For example, the original text snippet could be: "When the fuel pump output pressure is below 15 psi, the ECU will trigger a FUEL_PUMP_FAIL alarm. It is recommended to check if the pump body is stuck and measure whether the current exceeds the rated value."

[0127] Finally, by inputting the above structured reasoning results and the original text fragment into the large language model, a natural language response is generated: "[Response text] The engine cannot start, and the initial judgment is that the fuel pump is damaged;

[0128] [Interactive Inquiry] Has the fuel pressure test been completed? [Yes] [No]

[0129] [Technical Basis (Click to Expand)];

[0130] CFM56 Maintenance Manual, Chapter 7: When the fuel pump fails, the pressure should be below 15 psi…;

[0131] Historical Case F2024-032: Similar failures were caused by pump body jamming…”.

[0132] The final result is a structured response system that includes identifying fault chains, interactive question groups, and collapsible original text panels.

[0133] The embodiments of the present invention integrate the structured reasoning results and the original text fragments, and the two verify each other to form a complete closed loop of "reasoning + evidence".

[0134] This invention addresses the problems of scattered knowledge and insufficient targeting in knowledge retrieval. It establishes domain knowledge constraints based on an ontology constraint framework, performs structured reasoning by combining structured domain constraints and airport maintenance knowledge graphs, and incorporates original text fragments. This effectively overcomes the shortcomings of traditional knowledge retrieval, such as insufficient relevance and lack of targeting, which makes it difficult to obtain the intended knowledge.

[0135] The answers provided by existing technologies are usually just piles of raw text, relying on plain text output. Key parameters and operation steps lack visual highlighting, and the question-and-answer results lack keyword targeting, resulting in a high barrier to interaction.

[0136] Therefore, based on the above embodiments, the method further includes:

[0137] Step S105: Obtain the entity attributes of each entity obtained through structured reasoning in the airport maintenance knowledge graph.

[0138] Step S106: Determine the rendering output method of each entity obtained by reasoning based on the preset correspondence between entity attributes and rendering output methods.

[0139] The entity attributes include basic attributes, link attributes, descriptive attributes, and attributes without extensions. The default correspondence between entity attributes and rendering output methods is shown in the table below:

[0140]

[0141] When the `properties` attribute of an entity contains a `link` key, its value is the complete URL used to access the multimodal external resource anchored by the URL. When the `description` key is present, its value is the original descriptive text. As shown in the table above, based on this data structure, if the key-value pair in `properties` exists and contains the `link` key, a red interactive element with click event binding is generated; if it contains the `description` key, a floating tooltip based on the entity is generated; when there are no extended attributes, orange static text is output by default.

[0142] This specification ensures that any entity node can automatically obtain adaptive visualization rendering capabilities based on entity attributes through strong coupling between data structure contracts and rendering rules, thereby achieving adaptive conversion from data structures to visualization components.

[0143] This invention addresses the problem of interactive cognitive load by constructing a dynamic rendering contract driven by entity attributes. Based on entity attributes, the coloring scheme is automatically adapted to highlight key entity nodes, reducing the cognitive difficulty of on-site personnel in answering questions. This enables the generated answers to have visual interactive capabilities, improving user perception and comprehension efficiency.

[0144] Step S107: Output the answer according to the rendering output method.

[0145] This rendering output method allows for different display formats for different entity attributes. For example, the system detects whether an entity node is bound to a 3D contingency plan link. If it is, the entity is rendered as a red interactive element and a click event is injected using the above rendering output method. During user interaction, the scheduling engine parses the service identifier and unique code in the link template, locates the dedicated display interface of the 3D contingency plan library, and dynamically loads the associated multimodal contingency plan content (including 3D model demonstrations, action sequence visualizations, and other video components). Unbound entities retain orange static text, forming a visually differentiated response structure and achieving a seamless transition from text-based diagnostic conclusions to dynamic contingency plan demonstrations.

[0146] Based on the same inventive concept, embodiments of the present invention provide an airport maintenance knowledge question-and-answer system, such as... Figure 6 As shown, it includes:

[0147] Unit 601 is used to retrieve the user's query question;

[0148] The identification unit 602 is used to identify entities and entity states in the query question, and to construct initial evidence based on the identified entities and entity states.

[0149] Reasoning unit 603 is used to perform structured reasoning in the pre-constructed airport maintenance knowledge graph according to the ontology constraint framework, using initial evidence as the starting point for reasoning, to obtain structured reasoning results; at the same time, it extracts the original text fragments associated with the entities in the structured reasoning results.

[0150] The generation unit 604 is used to merge the structured reasoning results and the original text fragments to generate the answer to the corresponding query question.

[0151] Based on the same technical concept, embodiments of the present invention also provide an electronic device, such as... Figure 7 As shown, it includes a processor 701, a communication interface 702, a memory 703, and a communication bus 704, wherein the processor 701, the communication interface 702, and the memory 703 communicate with each other through the communication bus 704.

[0152] Memory 703 is used to store computer programs;

[0153] The processor 701 is used to execute the program stored in the memory 703 to implement the steps of the airport maintenance knowledge question and answer method.

[0154] The communication bus mentioned in the above electronic devices can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. This communication bus can be divided into address bus, data bus, control bus, etc. For ease of illustration, only one thick line is used to represent it in the diagram, but this does not mean that there is only one bus or one type of bus.

[0155] The communication interface is used for communication between the aforementioned electronic devices and other devices.

[0156] The memory may include random access memory (RAM) or non-volatile memory (NVM), such as at least one disk storage device. Optionally, the memory may also be at least one storage device located remotely from the aforementioned processor.

[0157] The processors mentioned above can be general-purpose processors, including central processing units (CPUs), network processors (NPs), etc.; they can also be digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0158] The computer program product for the airport maintenance knowledge Q&A method provided in this embodiment of the invention includes a computer-readable storage medium storing program code. The instructions included in the program code can be used to execute the methods described in the preceding method embodiments. For specific implementation details, please refer to the method embodiments, which will not be repeated here.

[0159] The airport maintenance knowledge Q&A method system provided in this embodiment of the invention can be specific hardware on a device or software or firmware installed on the device. The system provided in this embodiment of the invention has the same implementation principle and technical effects as the aforementioned method embodiments. For the sake of brevity, any parts not mentioned in the system embodiments can be referred to the corresponding content in the aforementioned method embodiments. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, units, and processes described above can all be referred to the corresponding processes in the above method embodiments, and will not be repeated here.

[0160] In the embodiments provided by this invention, it should be understood that the disclosed systems and methods can be implemented in other ways. The system embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. Furthermore, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Additionally, the shown or discussed mutual couplings, direct couplings, or communication connections may be through some communication interfaces. Indirect couplings or communication connections between systems or units may be electrical, mechanical, or other forms.

[0161] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0162] In addition, the functional units in the embodiments provided by the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0163] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, essentially, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0164] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. In addition, the terms "first", "second", "third", etc. are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0165] Finally, it should be noted that the above-described embodiments are merely specific implementations of the present invention, used to illustrate the technical solutions of the present invention, and not to limit it. The scope of protection of the present invention is not limited thereto. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can still modify or easily conceive of changes to the technical solutions described in the foregoing embodiments within the technical scope disclosed in the present invention, or make equivalent substitutions for some of the technical features; and these modifications, changes, or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention. All should be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A question-and-answer method for airport maintenance knowledge, characterized in that, include: Obtain the user's query question; Identify the entities and their states in the query question; The entity name is determined based on the identified entity; The Boolean value of the entity state is determined based on the identified entity state; wherein the entity state includes occurrence and non-occurrence; The entity type corresponding to the identified entity is determined based on a pre-constructed ontology constraint framework; the ontology constraint framework includes four entity types and three entity relationships; the four entity types are troubleshooting, fault cause, fault event, and maintenance operation event; the three entity relationships include suggestion, cause, and belong; the entity relationship between troubleshooting and fault cause is suggestion; The entity relationship between the cause of the fault and the event of the fault is that it leads to; The entity relationship between the fault item and the maintenance operation item is that they belong to each other; The entity type, entity name, and entity status are encapsulated as Boolean values ​​to form initial evidence; Using the initial evidence as the starting point for reasoning, structured reasoning is performed in the pre-constructed airport maintenance knowledge graph according to the ontology constraint framework to obtain structured reasoning results; at the same time, the original text fragments associated with the entities in the structured reasoning results are extracted; wherein, during reasoning, the fault propagation model is activated, and the fault propagation model evaluates all possible fault paths, and simulates the propagation process of the fault from the bottom component to the top system for each fault path, to obtain reasoning results in three dimensions: identified fault group, eliminated fault group, and unconfirmed fault group; The structured reasoning results and the original text fragments are fused to generate a structured answer corresponding to the query question, which includes a defined fault chain, interactive query groups, and a collapsible original text panel.

2. The method according to claim 1, characterized in that, Using the initial evidence as the starting point for reasoning, structured reasoning is performed within the pre-constructed airport maintenance knowledge graph according to the ontology constraint framework, resulting in the following structured reasoning results: Using the initial evidence as the starting point of reasoning, structured reasoning is performed in the pre-constructed airport maintenance knowledge graph according to the entity relationships in the ontology constraint framework to obtain all entities and entity relationships associated with the entities in the query question; The entities and entity relationships obtained through reasoning are organized hierarchically according to the ontology constraint framework to obtain structured reasoning results.

3. The method according to claim 2, characterized in that, The method further includes: Obtain the entity attributes of each entity obtained through structured reasoning in the airport maintenance knowledge graph; The rendering output method of each entity obtained through reasoning is determined based on the preset correspondence between entity attributes and rendering output methods. Output the answer according to the rendering output method described.

4. The method according to claim 3, characterized in that, The entity attributes include basic attributes, link attributes, description attributes, and attributes without extensions; The preset correspondence between entity attributes and rendering output methods includes: When the entity attribute is a basic attribute, the rendering output method is to display the label text content; When the entity attribute is a link attribute, the rendering output is red text + bound link, and the bound link is used to jump to external multimodal resources; When the entity attribute is a descriptive attribute, the rendering output method is a floating text box; When the entity attribute has no extended attributes, the rendering output method is to display orange static text.

5. The method according to claim 1, characterized in that, The process of constructing the airport maintenance knowledge graph includes: Construct an ontological constraint framework for airport maintenance; Based on the ontology constraint framework, entities and entity relationships are extracted from historical airport maintenance documents. Based on the pre-defined entity normalization and encapsulation model, the extracted entities are mapped to entity nodes of the airport maintenance knowledge graph. Based on a pre-defined relational normalization encapsulation model, the extracted entity relations are mapped to the edges of the airport maintenance knowledge graph.

6. The method according to claim 5, characterized in that, The method further includes: Configure an external database preloading interface for entity types in the ontology constraint framework; the external database includes at least a 3D maintenance plan library. The instances in the external database are bound to the entity nodes in the airport maintenance knowledge graph, and a unique identifier is generated. External links for entity nodes are generated based on the address of the preloaded interface and the unique identifier.

7. An airport maintenance knowledge question-and-answer system employing the airport maintenance knowledge question-and-answer method according to any one of claims 1-6, characterized in that, include: The retrieval unit is used to retrieve the user's query question; An identification unit is used to identify entities and entity states in the query question, and to construct initial evidence based on the identified entities and entity states. The reasoning unit is used to take the initial evidence as the starting point of reasoning, perform structured reasoning in the pre-constructed airport maintenance knowledge graph according to the ontology constraint framework, and obtain structured reasoning results; at the same time, it extracts the original text fragments associated with the entities in the structured reasoning results. The generation unit is used to fuse the structured reasoning results and the original text fragments to generate an answer to the query question.

8. An electronic device, characterized in that, It includes a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus; The memory is used to store computer programs; When the processor executes the program stored in the memory, it implements the method described in any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the method described in any one of claims 1-6.

Images