CN-121981710-A - Aviation fault elimination method, device, electronic equipment and storage medium

CN121981710ACN 121981710 ACN121981710 ACN 121981710ACN-121981710-A

Abstract

The invention relates to the technical field of artificial intelligence and provides an aviation troubleshooting method, device, electronic equipment and storage medium, wherein the method comprises the steps of performing cross-language semantic mapping on a fault description text of an aircraft to be checked to obtain a fault description vector, searching a multi-source heterogeneous knowledge base based on the fault description text and the fault description vector to obtain candidate troubleshooting knowledge, and performing troubleshooting decision based on the candidate troubleshooting knowledge and an airworthiness rule constraint in the aviation field and an airworthiness stock state corresponding to the aircraft to be checked, so that the generated troubleshooting scheme is technically and logically derived from an authority manual and historical experience, accords with airworthiness release standard in regulation and matches with current airworthiness resource supply in execution, thereby remarkably reducing maintenance delay caused by missing parts, violating regulations or misunderstanding the manual, greatly improving the first repair success rate and airworthiness release efficiency of the aircraft, and providing strong intelligent guarantee for aviation maintenance safety.

Inventors

FAN LIANG
HUANG FAN
Zheng Daiyang
Hu menghai
TANG CHENG

Assignees

科大讯飞股份有限公司

Dates

Publication Date: 20260505
Application Date: 20260402

Claims (10)

1. An aeronautical troubleshooting method, comprising: Determining a fault description text of the aircraft to be inspected; performing cross-language semantic mapping on the fault description text to obtain a fault description vector; searching a pre-constructed multi-source heterogeneous knowledge base based on the fault description text and the fault description vector to obtain candidate troubleshooting knowledge, wherein the candidate troubleshooting knowledge comprises at least one of fault isolation rules, maintenance manual paragraphs and fault handling cases; And carrying out an troubleshooting decision based on the candidate troubleshooting knowledge, the airworthiness compliance constraint in the aviation field and the airworthiness stock state corresponding to the aircraft to be troubleshooted to obtain an troubleshooting scheme, wherein the troubleshooting scheme comprises an troubleshooting operation path, airworthiness basis tracing marks and airworthiness resource allocation.
2. The method of aviation troubleshooting according to claim 1, wherein said cross-lingual semantic mapping of said fault description text to obtain a fault description vector comprises: Based on a cross-language semantic model, cross-language semantic mapping is carried out on the fault description text to obtain a fault description vector; The cross-language semantic model is obtained by training based on the similarity between sample semantic vectors corresponding to sample texts of different languages in the aviation field and the similarity between sample keyword vectors of sample keywords in the sample texts of different languages.
3. The method of claim 1, wherein the retrieving the pre-built multi-source heterogeneous knowledge base based on the fault description text and the fault description vector to obtain candidate troubleshooting knowledge comprises: Performing cross-language semantic mapping on the fault keywords in the fault description text to obtain keyword vectors; Based on the keyword vector, determining a foreign language standard term aligned with the semantics of the fault keyword and a corresponding foreign language term vector thereof; and searching the multi-source heterogeneous knowledge base based on the fault description text and the fault description vector, the fault keywords and the keyword vector, and the foreign language standard terms and the foreign language term vector to obtain candidate troubleshooting knowledge.
4. The method of claim 3, wherein said retrieving said multi-source heterogeneous knowledge base based on said fault description text and said fault description vector, said fault keywords and said keyword vectors, and said foreign language standard terms and said foreign language term vectors to obtain candidate fault knowledge comprises: performing term retrieval in a fault isolation library and an equipment list library in the multi-source heterogeneous knowledge library based on the keyword vector and the foreign language term vector to obtain the fault isolation rule, wherein the fault isolation rule comprises a fault isolation term and a navigable release rule; Based on the fault description vector, the fault keyword and the foreign language standard term, performing paragraph retrieval in a fault isolation library, an equipment list library and a maintenance knowledge library in the multi-source heterogeneous knowledge library to obtain a maintenance manual paragraph; based on the fault description text, carrying out case retrieval in a fault plan library in the multi-source heterogeneous knowledge base to obtain the fault handling case; Determining the candidate troubleshooting knowledge based on at least one of the fault isolation rules, the service manual paragraphs, and the fault handling cases.
5. The method according to any one of claims 1 to 4, wherein the obtaining an troubleshooting solution based on the candidate troubleshooting knowledge and an airworthiness compliance constraint in the aviation field and an airworthiness stock state corresponding to the aircraft to be troubleshooted includes: Analyzing fault handling cases in the candidate troubleshooting knowledge to obtain candidate associated components which cause faults corresponding to the fault description text, and generating a maintenance troubleshooting priority sequence based on the failure probability of each candidate associated component; Based on the model configuration information and the operation environment parameters of the aircraft to be inspected and the airworthiness release rules in the fault isolation rules in the candidate troubleshooting knowledge, carrying out compliance verification on the fault corresponding to the fault description text to obtain a verification result; And under the condition that the verification result indicates that the clearance limit exists, generating a navigable limit condition, and determining an troubleshooting scheme based on the navigable limit condition, the maintenance and investigation priority sequence, the navigable compliance constraint and the aviation material inventory state corresponding to the aircraft to be investigated.
6. The method of claim 5, wherein determining an troubleshooting solution based on the airworthiness constraints and a sequence of repair troubleshooting priorities, and the airworthiness compliance constraints and corresponding aircraft stock status of the aircraft to be troubleshooted, comprises: Sequentially checking the inventory available state of each candidate associated component in the maintenance and inspection priority sequence based on the aviation material inventory state; Determining the inventory availability state of the interchange part corresponding to any candidate associated part under the condition that the inventory availability state of any candidate associated part is unavailable; And generating an air material missing constraint signal under the condition that the stock available state of the interchangeable part is still unavailable, and generating the troubleshooting scheme based on the air material missing constraint signal, the navigable limit condition and the navigable compliance constraint.
7. The method according to any one of claims 1 to 4, wherein the step of obtaining an troubleshooting solution based on the candidate troubleshooting knowledge and an airworthiness constraint in the aviation field and an airworthiness inventory status corresponding to the aircraft to be troubleshooted further comprises: comparing the troubleshooting scheme with the fault handling case to obtain an operation difference item; Performing attribution analysis on the operation difference item to obtain a difference item attribution; determining a first repair rate improvement value and/or an average troubleshooting duration shortening value of the troubleshooting scheme relative to the fault handling case based on execution feedback of the troubleshooting scheme; And writing the troubleshooting scheme into a fault plan library in the multi-source heterogeneous knowledge base under the condition that the difference term is reasonable and the first repair rate lifting value and/or the average troubleshooting duration shortening value meet a preset threshold.
8. An aeronautical fault-removing device, comprising: The text determining unit is used for determining a fault description text of the aircraft to be checked; The semantic mapping unit is used for performing cross-language semantic mapping on the fault description text to obtain a fault description vector; the knowledge retrieval unit is used for retrieving a multi-source heterogeneous knowledge base constructed in advance based on the fault description text and the fault description vector to obtain candidate troubleshooting knowledge, wherein the candidate troubleshooting knowledge comprises at least one of a fault isolation rule, a maintenance manual paragraph and a fault handling case; And the troubleshooting decision unit is used for performing troubleshooting decision based on the candidate troubleshooting knowledge, the airworthiness compliance constraint in the aviation field and the airworthiness stock state corresponding to the aircraft to be checked to obtain an troubleshooting scheme, wherein the troubleshooting scheme comprises an troubleshooting operation path, airworthiness basis tracing marks and airworthiness resource allocation.
9. An electronic device comprising a memory, a processor and a computer program stored on the memory and running on the processor, wherein the processor implements the aeronautical troubleshooting method of any one of claims 1 to 7 when the computer program is executed by the processor.
10. A non-transitory computer readable storage medium having stored thereon a computer program, which when executed by a processor implements the aeronautical troubleshooting method of any one of claims 1 to 7.

Description

Aviation fault elimination method, device, electronic equipment and storage medium Technical Field The invention relates to the technical field of artificial intelligence, in particular to an aviation accident-eliminating method, an aviation accident-eliminating device, electronic equipment and a storage medium. Background With the rapid development of aviation technology, the complexity of modern aircraft systems is increased, and maintenance and guarantee works face extremely high technical requirements and safety standards. In the maintenance and troubleshooting link of the aircraft, the fault is accurately and efficiently positioned and solved, and the method is very important for guaranteeing the flight safety, reducing the flight delay and reducing the operation cost. At present, aviation troubleshooting mainly depends on English technical documents provided by former factories, such as fault isolation manuals, however, the former factories have high language barriers for first-line crews, reading and understanding are difficult, and the fault coverage of the documents is limited, and in some cases, the documents still need to be judged by personal experience, so that the personal experience is difficult to precipitate. In addition, current maintenance decisions often are disjointed from the actual ones, which is poorly executable. Disclosure of Invention The invention provides an aviation troubleshooting method, an aviation troubleshooting device, electronic equipment and a storage medium, which are used for solving the problems of language barriers, information islands and actual disjointing in aviation maintenance in the prior art, and realizing the generation of an troubleshooting scheme which meets the requirement of airworthiness and has the guarantee of airworthiness. The invention provides an aviation accident-eliminating method, which comprises the following steps: Determining a fault description text of the aircraft to be inspected; performing cross-language semantic mapping on the fault description text to obtain a fault description vector; searching a pre-constructed multi-source heterogeneous knowledge base based on the fault description text and the fault description vector to obtain candidate troubleshooting knowledge, wherein the candidate troubleshooting knowledge comprises at least one of fault isolation rules, maintenance manual paragraphs and fault handling cases; And carrying out an troubleshooting decision based on the candidate troubleshooting knowledge, the airworthiness compliance constraint in the aviation field and the airworthiness stock state corresponding to the aircraft to be troubleshooted to obtain an troubleshooting scheme, wherein the troubleshooting scheme comprises an troubleshooting operation path, airworthiness basis tracing marks and airworthiness resource allocation. According to the aviation troubleshooting method provided by the invention, the cross-language semantic mapping is carried out on the fault description text to obtain the fault description vector, and the method comprises the following steps: Based on a cross-language semantic model, cross-language semantic mapping is carried out on the fault description text to obtain a fault description vector; The cross-language semantic model is obtained by training based on the similarity between sample semantic vectors corresponding to sample texts of different languages in the aviation field and the similarity between sample keyword vectors of sample keywords in the sample texts of different languages. According to the aviation troubleshooting method provided by the invention, the searching of the pre-constructed multi-source heterogeneous knowledge base based on the fault description text and the fault description vector to obtain candidate troubleshooting knowledge comprises the following steps: Performing cross-language semantic mapping on the fault keywords in the fault description text to obtain keyword vectors; Based on the keyword vector, determining a foreign language standard term aligned with the semantics of the fault keyword and a corresponding foreign language term vector thereof; and searching the multi-source heterogeneous knowledge base based on the fault description text and the fault description vector, the fault keywords and the keyword vector, and the foreign language standard terms and the foreign language term vector to obtain candidate troubleshooting knowledge. According to the aviation troubleshooting method provided by the invention, the multi-source heterogeneous knowledge base is searched based on the fault description text and the fault description vector, the fault keyword and the keyword vector, and the foreign language standard term and the foreign language term vector to obtain candidate troubleshooting knowledge, and the method comprises the following steps: performing term retrieval in a fault isolation library and an equipment list library in the multi-source heterogeneous