A helicopter fault problem collection method and system
By classifying the helicopter fault data step by step and extracting the fault keyword, establishing correspondence relationships and performing data splitting and weight allocation, the problems of low efficiency and high complexity of fault data processing in the existing technology are solved, and efficient and accurate fault data processing and diagnosis are achieved.
Patent Information
- Application Number
- CN202510036300.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-01-09
- Publication Date
- 2025-04-01
AI Technical Summary
The prior art has problems of inefficiency and high complexity in the classification, storage and extraction of critical information of helicopter failure data, especially in the processing of fault descriptions of many-to-many relationships.
By classifying the helicopter fault data step by step in dimensions such as model, working timing and configuration, and automatically extracting the fault keywords from the fault description, establishing the correspondence between the keywords and the fault description, and performing data splitting and weight allocation.
It significantly improves the efficiency and accuracy of fault data processing, reduces manual operation errors, ensures the accuracy of fault description and comprehensive extraction of key information, and improves the accuracy of fault diagnosis.
Smart Images
Figure CN119740083A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of helicopter configuration management, and in particular to a method and system for collecting helicopter fault problems. Background Art
[0002] With the widespread use of helicopters, the complexity of their systems continues to increase, and failures are inevitable. In order to ensure the flight safety of helicopters, it is crucial to detect and solve failures in a timely manner. Usually, a large amount of failure data is accumulated during the daily maintenance and overhaul of helicopters. These data include failure conditions of different models, working times, configurations, and working parts. However, due to the wide variety and complex structure of data, the analysis and processing of failure information becomes particularly difficult.
[0003] At present, many methods of helicopter fault data collection and processing have the following main problems:
[0004] 1. Data classification and storage are not systematic enough: Existing fault data is usually recorded manually, the data classification is not clear enough, and there is a lack of automated classification methods. This leads to low efficiency in information acquisition and analysis when the amount of fault data increases.
[0005] 2. Difficulty in extracting key fault information: Since fault descriptions usually contain a large amount of natural language information, existing technologies find it difficult to accurately extract keywords from fault descriptions, further affecting fault diagnosis and troubleshooting.
[0006] 3. The processing of fault descriptions with many-to-many relationships is complex: In some complex fault scenarios, there is a situation where one fault description corresponds to multiple fault keywords or multiple fault keywords correspond to multiple fault descriptions. The existing system is difficult to handle such complex many-to-many correspondence relationships, resulting in unreasonable fault data splitting and weight distribution.
[0007] Therefore, there is an urgent need for a method and system that can efficiently collect, classify and process helicopter fault problem data to improve the accuracy and efficiency of fault analysis and ensure the safe operation of helicopters. Summary of the invention
[0008] In order to solve the above problems, the present invention provides a helicopter fault problem collection system, which can effectively classify fault data, extract keywords in fault descriptions, and perform reasonable processing and data storage for different fault relationships, thereby significantly improving the efficiency and accuracy of fault data processing.
[0009] In order to achieve the above-mentioned invention object, the technical solution provided by the present invention includes:
[0010] A method for collecting helicopter fault problems, comprising the steps of:
[0011] Collecting a helicopter fault problem data set, wherein the helicopter fault problem data includes aircraft model, working time, configuration, working location and fault description;
[0012] Extracting fault keywords from the fault description and adding them to the helicopter fault problem data;
[0013] Classifying the helicopter fault problem data set into a plurality of first-level data sets according to the model category; classifying each of the first-level data sets into a plurality of second-level data sets according to the working opportunity category; and classifying each of the second-level data sets into a plurality of third-level data sets according to the configuration category;
[0014] Traverse all the helicopter failure problem data in the three-level data sets and perform the following operations:
[0015] Determine the corresponding relationship between the fault keyword and the fault description;
[0016] If there is a unique corresponding relationship between a fault keyword and a fault description, the fault description is associated with the fault keyword, and the weight of the helicopter fault problem data is marked as 100%, and saved in the problem library.
[0017] Preferably, if there is a correspondence between multiple fault keywords and one fault description, the fault description is associated with each fault keyword, thereby splitting the helicopter fault problem data into multiple associated helicopter fault problem data, the number of splits being equal to the number of corresponding fault keywords, and the weight of each helicopter fault problem data is marked as 100% and saved in the problem library.
[0018] Preferably, the method further includes the following steps: if a correspondence exists between a plurality of the fault keywords and a plurality of the fault descriptions, then according to the fault keywords, the helicopter fault problem data is split into a plurality of helicopter fault problem data, wherein one fault keyword corresponds to a plurality of the fault descriptions, and the number of splits is equal to the number of corresponding fault keywords.
[0019] Preferably, the method further includes the following steps: if there is a correspondence between one of the fault keywords and a plurality of the fault descriptions, a semantic analysis is performed on each of the fault descriptions to extract a verification fault keyword, the verification fault keywords are sorted according to the degree of overlap between the verification fault keywords and the corresponding fault keywords, and each piece of helicopter fault problem data is initially assigned a weight according to the sorting, and the data is saved in a cache area.
[0020] Preferably, the method further includes the steps of traversing all helicopter fault problem data in the cache area, deleting helicopter fault problem data whose initial weight is lower than a preset threshold, manually performing a final weight assignment on each remaining helicopter fault problem data, and using the verification fault keyword as the fault keyword of the helicopter fault problem data and saving it to the problem library.
[0021] The present invention also provides a helicopter fault problem collection system, comprising:
[0022] A data collection module is used to collect a helicopter fault problem data set, wherein the helicopter fault problem data includes the model, working time, configuration, working position and fault description;
[0023] A keyword extraction module, used to extract fault keywords from the fault description and add them to the helicopter fault problem data;
[0024] A data classification module is used to classify the helicopter fault problem data set into a plurality of first-level data sets according to the aircraft type category; classify each of the first-level data sets into a plurality of second-level data sets according to the working opportunity category; and classify each of the second-level data sets into a plurality of third-level data sets according to the configuration category;
[0025] The data processing module is used to traverse all helicopter failure problem data in the three-level data sets and perform the following operations:
[0026] Determine the corresponding relationship between the fault keyword and the fault description;
[0027] If there is a unique corresponding relationship between a fault keyword and a fault description, the fault description is associated with the fault keyword, and the weight of the helicopter fault problem data is marked as 100%, and saved in the problem library.
[0028] Preferably, the data processing module further comprises the following operations:
[0029] If there is a corresponding relationship between multiple fault keywords and one fault description, the fault description is associated with each fault keyword, so that the helicopter fault problem data is split into multiple associated helicopter fault problem data, and the number of splits is equal to the number of corresponding fault keywords. The weight of each helicopter fault problem data is marked as 100% and saved to the problem library.
[0030] Preferably, the data processing module further comprises the following operations:
[0031] If there is a correspondence between multiple fault keywords and multiple fault descriptions, then the helicopter fault problem data is split into several helicopter fault problem data in which one fault keyword corresponds to multiple fault descriptions according to the fault keywords, and the number of splits is equal to the number of corresponding fault keywords.
[0032] Preferably, the data processing module further comprises the following operations:
[0033] If a fault keyword corresponds to multiple fault descriptions, semantic analysis is performed on each fault description to extract the verification fault keyword, and the verification fault keywords are sorted according to the degree of overlap between the verification fault keywords and the corresponding fault keywords. Based on the sorting, initial weight allocation is performed on each helicopter fault problem data and the data is saved in the cache area.
[0034] Preferably, the data processing module also includes performing the following operations: traversing all helicopter fault problem data in the cache area, deleting helicopter fault problem data whose initial weight is lower than a preset threshold, manually performing final weight assignment on each remaining helicopter fault problem data, and using the verification fault keyword as the fault keyword of the helicopter fault problem data and saving it to the problem library.
[0035] Beneficial Effects
[0036] The present invention avoids the inefficiency of traditional manual classification by classifying different dimensions such as machine type, working time, and configuration step by step, greatly improving the efficiency of data management and query, and can quickly and accurately organize fault data, especially in the case of large amounts of data. By automatically extracting fault keywords from fault descriptions and associating them with fault data, errors and inefficiencies in manual operations are avoided, the accuracy of fault descriptions and the comprehensive extraction of key fault information are ensured, which helps to improve the accuracy of fault diagnosis and provides data support for subsequent fault prediction, analysis, and prevention. BRIEF DESCRIPTION OF THE DRAWINGS
[0037] Figure 1 It is a flow chart of a method for collecting helicopter fault problems in a preferred embodiment of the present invention;
[0038] Figure 2 It is a schematic diagram of classification results of a helicopter fault problem data set in another preferred embodiment of the present invention;
[0039] Figure 3 The figure is a schematic diagram of a helicopter fault problem collection system in another preferred embodiment of the present invention. DETAILED DESCRIPTION
[0040] In order to make the purpose, technical solution and advantages of the present invention clearer, the present invention is further described below in conjunction with the accompanying drawings. In the description of the present invention, it should be understood that the terms "upper", "lower", "front", "back", "left", "right", "top", "bottom", "inside", "outside" and the like indicate directions or positional relationships based on the directions or positional relationships shown in the accompanying drawings, which are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore cannot be understood as limiting the present invention.
[0041] Embodiment 1
[0042] like Figure 1 As shown, this embodiment provides a method for collecting helicopter fault problems, including the steps of:
[0043] S1. Collect a data set of helicopter fault problems, wherein the helicopter fault problem data includes the model, working time, configuration, working part and fault description. Among them, the working actually includes three categories: landing, take-off and landing. The configuration can be divided into several configuration modules according to the common characteristics of the models involved, as shown in the following table. It should be understood that the present invention does not make further limitations on the source of the helicopter fault problem data, which can be obtained by accessing the database of the maintenance unit or by collecting historical data.
[0044]
[0045]
[0046] S2. Extract fault keywords from the fault description and add them to the helicopter fault problem data. The extraction of the fault keywords can be done by manual annotation, or by using a word frequency statistics method commonly used in the art, an extraction method based on deep learning, etc. It should be understood that in order to make the data set more intuitive and clear, in some preferred embodiments, it is considered to organize the helicopter fault problem data in the following table.
[0047]
[0048] S3, such as Figure 2As shown, the helicopter fault problem data set is classified into several first-level data sets according to the model category; each of the first-level data sets is classified into several second-level data sets according to the working opportunity category; and each of the second-level data sets is classified into several third-level data sets according to the configuration category. The technical considerations of this step are as follows: The fault data of the helicopter may have multi-dimensional and multi-level characteristics. Through classification, complex fault data can be layered according to different dimensions, making it easier to analyze specific dimensions, especially for in-depth mining and processing of a certain type of fault problem. After classification, the system can apply different algorithms or processing strategies to different categories of data as needed. This step can structure the management of large and complex fault data by classifying the data (according to categories such as model, working opportunity, configuration, etc.). This can greatly improve the orderliness of data storage, and can locate relevant fault information more quickly and accurately in subsequent data query, retrieval and analysis.
[0049] S4. Traverse all helicopter failure problem data in the three-level data set and perform the following operations:
[0050] S41. Determine the correspondence between the fault keyword and the fault description. The correspondence specifically includes one-to-one, one-to-many and many-to-many. The specific determination method may be manual labeling by staff, or may be based on a string matching method or word vector similarity based on an embedding model (such as Word2Vec, GloVe, BERT, etc.), which is not further limited in the present invention.
[0051] S421. If there is a unique correspondence between a fault keyword and a fault description, the fault description is associated with the fault keyword, and the weight of the helicopter fault problem data is marked as 100%, and saved in the problem library.
[0052] When there is a unique and clear correspondence between a fault keyword and a fault description, it means that there is a direct and single match between the fault keyword and the fault description, without other interfering information or complex many-to-many or many-to-one situations. After determining the unique correspondence, the fault description is associated with the fault keyword, that is, the fault description clearly corresponds to the specific fault keyword. This step ensures the accuracy of data matching and helps with subsequent analysis and query. Since this correspondence is unique and clear, the system can be sure of the correctness and completeness of the data, so the weight of the helicopter fault problem data is marked as 100%. The "weight marked as 100%" here means that the data has the highest credibility and priority in the entire fault analysis process, indicating that the data has a complete and accurate correspondence. Finally, this verified and associated fault problem data is saved in the problem library and stored as a high-credibility fault record for subsequent fault retrieval, analysis and use.
[0053] In some preferred embodiments, the many-to-one correspondence is further discussed.
[0054] S422. If there is a corresponding relationship between multiple fault keywords and one fault description, the fault description is associated with each fault keyword, thereby splitting the helicopter fault problem data into multiple associated helicopter fault problem data, the number of splits is equal to the number of corresponding fault keywords, and the weight of each helicopter fault problem data is marked as 100%, and saved to the problem library.
[0055] There is a situation where multiple fault keywords correspond to the same fault description, which means that the fault description may involve multiple different fault keywords instead of corresponding to a single keyword. In this case, the system will associate the fault description with each fault keyword respectively. This ensures that each fault keyword can be correctly associated with the fault description, avoiding omissions or association errors between keywords and descriptions. Since multiple keywords correspond to the same fault description, in order to facilitate further analysis and processing, the system splits the fault problem data into several pieces of data, each of which corresponds to a fault keyword. The number of split data is equal to the number of corresponding fault keywords. This splitting process ensures that each fault keyword has an independent fault description, which is convenient for subsequent processing. Similar to step S42, although there are multiple keywords corresponding to the same description, the system believes that these data are still highly accurate. Therefore, the weight of each split data is also marked as 100%, indicating that the credibility and priority of each data in fault analysis are also high.
[0056] In other preferred embodiments, the many-to-many correspondence is further discussed.
[0057] S423. If a plurality of the fault keywords correspond to a plurality of the fault descriptions, then the helicopter fault problem data is split into a plurality of helicopter fault problem data, one fault keyword corresponding to a plurality of the fault descriptions, according to the fault keywords, and the number of splits is equal to the number of corresponding fault keywords.
[0058] The situation where multiple fault keywords correspond to multiple fault descriptions means that the complexity of the fault data increases, and some fault keywords may correspond to multiple different fault descriptions, and vice versa. In this case, the system first splits the data according to the fault keywords. Specifically, the system splits the original fault problem data into multiple data entries, each of which contains a fault keyword and its corresponding multiple fault descriptions. This processing method simplifies the complex many-to-many relationship into multiple "one-to-many" data relationships, which is convenient for subsequent analysis and processing. The number of split data entries is determined according to the number of fault keywords. Each fault keyword corresponds to a new data entry, which contains all the fault descriptions associated with the keyword. Therefore, the number of split data entries is consistent with the number of fault keywords. This ensures that the fault keywords can be processed independently under the complex many-to-many correspondence relationship. By splitting the complex many-to-many relationship into multiple one-to-many relationships, these fault problem data can be processed and stored more efficiently. This structured data splitting method helps to improve the efficiency of the system in processing complex fault data, while ensuring that the correspondence between all fault keywords and fault descriptions is fully preserved.
[0059] Furthermore, further discussion on the one-to-many correspondence is given in other preferred embodiments.
[0060] S424. If a fault keyword corresponds to multiple fault descriptions, semantic analysis is performed on each fault description to extract the verification fault keyword, and the verification fault keywords are sorted according to the degree of overlap between the verification fault keywords and the corresponding fault keywords. Based on the sorting, initial weight allocation is performed on each helicopter fault problem data and the data is saved in the cache area.
[0061] The processing method when a fault keyword corresponds to multiple different fault descriptions, which means that the same fault keyword may involve multiple different fault descriptions, and the fault descriptions are complex and diverse. To further determine the accuracy and relevance of the fault description, the system performs semantic analysis on each fault description. Semantic analysis is a natural language processing technology that aims to extract "verification fault keywords" related to the fault by analyzing the language meaning in the description. These verification fault keywords can serve as a verification and supplement to the initial fault keywords, helping to more accurately determine the type of fault described. According to the degree of match (overlap) between the verification fault keyword and the initial fault keyword, the system sorts multiple fault descriptions. The greater the overlap, the higher the relevance of the fault description to the fault keyword. The purpose of sorting is to identify the most relevant fault description so that it can be prioritized in subsequent steps. Based on the above sorting results, an initial weight is assigned to each helicopter fault problem data. The weight reflects the relevance or credibility of the data. Fault descriptions with high overlap will be assigned a higher initial weight, and descriptions with low overlap will receive a lower weight. The weight assignment provides a reference for subsequent analysis. After the initial weight assignment is completed, all processed fault problem data will be saved in the cache area. The function of the buffer is to temporarily store data so that the final weights and other information of these data can be further processed or confirmed in subsequent steps. At this time, the helicopter fault problem data stored in the buffer after the initial weight allocation can be used directly, that is, directly transferred and stored in the problem library, and further weight allocation confirmation or redistribution can also be performed. Specifically:
[0062] S425, traverse all helicopter fault problem data in the cache area, delete the helicopter fault problem data whose initial weight is lower than the preset threshold, manually perform final weight allocation on each remaining helicopter fault problem data, and use the verification fault keyword as the fault keyword of the helicopter fault problem data and save it to the problem library.
[0063] Fault problem data with weights lower than the preset threshold are deleted, which means that these low-weight data have a poor match with the fault keyword or are not credible, and are therefore not suitable for being saved as valid fault data in the final question library. This step helps to screen out high-quality, highly credible data and exclude invalid or less relevant fault descriptions. For the remaining fault problem data that have been screened and whose initial weights are higher than the preset threshold, the system is handed over to manual labor for final weight assignment. The purpose of manual intervention is to confirm the final weight of the data based on actual experience and further understanding of the fault problem to improve the accuracy and reliability of the data. This method combines the advantages of automated processing and manual judgment to ensure data accuracy. In the process of manual weight assignment, the verification fault keyword is used as the formal fault keyword for each fault problem data to confirm the relationship between the keyword and the fault description. The verification fault keyword is extracted and verified in the previous semantic analysis process to ensure that it can accurately describe the characteristics of the fault problem.
[0064] Embodiment 2
[0065] like Figure 3 As shown, this embodiment provides a helicopter fault problem collection system, including:
[0066] A data collection module is used to collect a helicopter fault problem data set, wherein the helicopter fault problem data includes the model, working time, configuration, working position and fault description;
[0067] A keyword extraction module, used to extract fault keywords from the fault description and add them to the helicopter fault problem data;
[0068] A data classification module is used to classify the helicopter fault problem data set into a plurality of first-level data sets according to the aircraft type category; classify each of the first-level data sets into a plurality of second-level data sets according to the working opportunity category; and classify each of the second-level data sets into a plurality of third-level data sets according to the configuration category;
[0069] The data processing module is used to traverse all helicopter failure problem data in the three-level data sets and perform the following operations:
[0070] Determine the corresponding relationship between the fault keyword and the fault description;
[0071] If there is a unique corresponding relationship between a fault keyword and a fault description, the fault description is associated with the fault keyword, and the weight of the helicopter fault problem data is marked as 100%, and saved in the problem library.
[0072] In some preferred embodiments, the data processing module further comprises the following operations:
[0073] If there is a corresponding relationship between multiple fault keywords and one fault description, the fault description is associated with each fault keyword, so that the helicopter fault problem data is split into multiple associated helicopter fault problem data, and the number of splits is equal to the number of corresponding fault keywords. The weight of each helicopter fault problem data is marked as 100% and saved to the problem library.
[0074] In some preferred embodiments, the data processing module further comprises the following operations:
[0075] If there is a correspondence between multiple fault keywords and multiple fault descriptions, then the helicopter fault problem data is split into several helicopter fault problem data in which one fault keyword corresponds to multiple fault descriptions according to the fault keywords, and the number of splits is equal to the number of corresponding fault keywords.
[0076] In some preferred embodiments, the data processing module further comprises the following operations:
[0077] If a fault keyword corresponds to multiple fault descriptions, semantic analysis is performed on each fault description to extract the verification fault keyword, and the verification fault keywords are sorted according to the degree of overlap between the verification fault keywords and the corresponding fault keywords. Based on the sorting, initial weight allocation is performed on each helicopter fault problem data and the data is saved in the cache area.
[0078] In some preferred embodiments, the data processing module also includes performing the following operations: traversing all helicopter fault problem data in the cache area, deleting helicopter fault problem data whose initial weight is lower than a preset threshold, manually performing a final weight assignment on each remaining helicopter fault problem data, and using the verification fault keyword as the fault keyword of the helicopter fault problem data and saving it to the problem library.
[0079] The above shows and describes the basic principles, main features and advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments, and the above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.
Claims
1. A method for collecting helicopter fault problems, characterized in that: Includes steps: Collecting a helicopter fault problem data set, wherein the helicopter fault problem data includes aircraft model, working time, configuration, working location and fault description; Extracting fault keywords from the fault description and adding them to the helicopter fault problem data; Classifying the helicopter fault problem data set into a plurality of first-level data sets according to the model category; classifying each of the first-level data sets into a plurality of second-level data sets according to the working opportunity category; and classifying each of the second-level data sets into a plurality of third-level data sets according to the configuration category; Traverse all the helicopter failure problem data in the three-level data sets and perform the following operations: Determine the corresponding relationship between the fault keyword and the fault description; If there is a unique corresponding relationship between a fault keyword and a fault description, the fault description is associated with the fault keyword, and the weight of the helicopter fault problem data is marked as 100%, and saved in the problem library.
2. The helicopter fault problem collection method according to claim 1, characterized in that: It also includes the following steps: if there is a correspondence between multiple fault keywords and one fault description, then the fault description is associated with each fault keyword, thereby splitting the helicopter fault problem data into multiple associated helicopter fault problem data, the number of splits is equal to the number of corresponding fault keywords, and the weight of each helicopter fault problem data is marked as 100%, and saved to the problem library.
3. The helicopter fault problem collection method according to claim 1, characterized in that: It also includes the following steps: if there is a correspondence between multiple fault keywords and multiple fault descriptions, then according to the fault keywords, the helicopter fault problem data is split into a number of helicopter fault problem data, one fault keyword corresponding to multiple fault descriptions, and the number of splits is equal to the number of corresponding fault keywords.
4. The helicopter fault problem collection method according to claim 1 or 3, characterized in that: The method also includes the following steps: if a fault keyword corresponds to a plurality of fault descriptions, semantic analysis is performed on each fault description to extract a verification fault keyword, the verification fault keywords are sorted according to the degree of overlap between the verification fault keywords and the corresponding fault keywords, and initial weights are assigned to each piece of helicopter fault problem data according to the sorting, and the data are saved in a cache area.
5. The helicopter fault problem collection method according to claim 4, characterized in that: The method also includes the steps of traversing all helicopter fault problem data in the cache area, deleting helicopter fault problem data whose initial weight is lower than a preset threshold, manually performing a final weight assignment on each remaining helicopter fault problem data, and using the verification fault keyword as the fault keyword of the helicopter fault problem data and saving it to the problem library.
6. A helicopter fault problem collection system, characterized in that: include: A data collection module is used to collect a helicopter fault problem data set, wherein the helicopter fault problem data includes the model, working time, configuration, working position and fault description; A keyword extraction module, used to extract fault keywords from the fault description and add them to the helicopter fault problem data; A data classification module is used to classify the helicopter fault problem data set into a plurality of first-level data sets according to the aircraft type category; classify each of the first-level data sets into a plurality of second-level data sets according to the working opportunity category; and classify each of the second-level data sets into a plurality of third-level data sets according to the configuration category; The data processing module is used to traverse all helicopter failure problem data in the three-level data sets and perform the following operations: Determine the corresponding relationship between the fault keyword and the fault description; If there is a unique corresponding relationship between a fault keyword and a fault description, the fault description is associated with the fault keyword, and the weight of the helicopter fault problem data is marked as 100%, and saved in the problem library.
7. The helicopter fault problem collection system according to claim 6, characterized in that: The data processing module further comprises performing the following operations: If there is a corresponding relationship between multiple fault keywords and one fault description, the fault description is associated with each fault keyword, so that the helicopter fault problem data is split into multiple associated helicopter fault problem data, and the number of splits is equal to the number of corresponding fault keywords. The weight of each helicopter fault problem data is marked as 100% and saved to the problem library.
8. The helicopter fault problem collection system according to claim 6, characterized in that: The data processing module further comprises performing the following operations: If there is a correspondence between multiple fault keywords and multiple fault descriptions, then the helicopter fault problem data is split into several helicopter fault problem data in which one fault keyword corresponds to multiple fault descriptions according to the fault keywords, and the number of splits is equal to the number of corresponding fault keywords.
9. The helicopter fault problem collection system according to claim 6 or 8, characterized in that: The data processing module further comprises performing the following operations: If a fault keyword corresponds to multiple fault descriptions, semantic analysis is performed on each fault description to extract the verification fault keyword, and the verification fault keywords are sorted according to the degree of overlap between the verification fault keywords and the corresponding fault keywords. Based on the sorting, initial weight allocation is performed on each helicopter fault problem data and the data is saved in the cache area.
10. The helicopter fault problem collection system according to claim 9, characterized in that: The data processing module also includes the following operations: traversing all helicopter fault problem data in the cache area, deleting helicopter fault problem data whose initial weight is lower than a preset threshold, manually performing a final weight assignment on each remaining helicopter fault problem data, and using the verification fault keyword as the fault keyword of the helicopter fault problem data and saving it to the problem library.