A data-based evidence-based aircraft continuous airworthiness auxiliary management system and method

Through the data-based evidence-based aircraft continuous airworthiness assistance management system, combined with fault events and flight QAR data, intelligent monitoring and closed-loop management are realized, solving the problems of lag in reporting and low supervision efficiency in the existing technology, and improving fleet health Management and secure operation capabilities.

CN114819207BActive Publication Date: 2024-11-05CHINA ACAD OF CIVIL AVIATION SCI & TECH
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202210371422.4
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Priority Date
2022-03-22
Filing Date
2022-04-08
Publication Date
2024-11-05
Estimated Expiration
2042-04-08

AI Technical Summary

Technical Problem

The existing continuous airworthiness management system has lag in reporting, difficulty in unifying unstructured data, lack of timely closed-loop supervision management methods, and insufficient fleet health status monitoring capabilities, resulting in lagging regulatory response and low maintenance management efficiency.

Method used

A continuous airworthiness assistance management system based on data evidence-based aircraft is designed, combining the fault event case database, aircraft usage information database, flight on-board QAR database and QAR parameter list database, combined with the fault event case management module and flight QAR data analysis Modules and fleet health management modules realize intelligent monitoring of failure cases, data batch processing and performance trend monitoring, and use big data models and unsupervised machine learning to identify and alert abnormal states.

Benefits of technology

By deeply integrating the actual operation data of the aircraft and failure cases, a more timely and powerful regulatory response has been achieved, the convenience of incident investigation and closed-loop supervision management capabilities have been improved, the refined management of fleet health status has been promoted, and civil aviation safety has been improved. Operational level.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114819207B_ABST
    Figure CN114819207B_ABST
Patent Text Reader

Abstract

The present invention provides an aircraft continuous airworthiness auxiliary management system and method based on data evidence, belonging to the field of civil aviation management. The system includes: a database layer and an application layer; the database layer includes: a fault event case database, an aircraft usage information database, a flight onboard QAR database and a QAR parameter list database; the application layer includes: a fault event case management module, a flight QAR data analysis module and a fleet health management module. The present invention enables continuous airworthiness management through the deep integration of actual aircraft / engine operation data and fault cases, which can provide more powerful data evidence-based means for supervision, and enhance the initiative and convenience of event investigation; by clustering reported fault cases, intelligent alarms for high-frequency similar faults or anomalies are realized, guiding supervision to intervene earlier, and conducting investigations and tracking within the industry based on fault characteristics and QAR data, which helps to move the safety window forward.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the field of civil aviation management, and in particular relates to a data-based evidence-based aircraft continuous airworthiness auxiliary management system and method. Background Art

[0002] The Quick Access Recorder (QAR) is a device on an aircraft used to monitor and record flight data. It records many parameters during the flight, including pilot control, aircraft attitude, operating status of onboard systems, and external environmental conditions. The flight data it records are all derived from the onboard bus, which is the same as the parameters recorded by the Flight Data Recorder (FDR), also known as the black box. By decoding the data, the entire flight process of the aircraft can be restored.

[0003] The most mature application of QAR data is as the core data of flight quality monitoring, which effectively ensures flight safety and prevents the occurrence of flight accidents. The Civil Aviation Administration of China uses QAR data for safety supervision to improve the level of flight safety management. So far, all transport aircraft registered and operated in China are equipped with QAR or equivalent equipment. QAR data also provides an effective basis for flight quality analysis and unsafe incident investigation.

[0004] With the popularization of QAR, how to apply massive airborne data to the fields of maintenance and health management, explore failure modes and potential operational risks, strengthen the role of data evidence and data intelligence in continuous airworthiness management, track and ensure the continuous airworthiness of aircraft, and thus better promote the refined management of civil aviation safety has become one of the focuses of industry attention.

[0005] At present, the industry's continuous airworthiness management uses the Service Difficulty Report (SDR) system as an important medium. Figure 1As shown. In addition to reporting incidents within the scope of regulations, air operators should also report incidents involving product design defects, repair quality issues, frequent failures, typical failures, and undetermined causes as much as possible to provide assistance for improving products, improving maintenance management, and exchanging experiences. In accordance with the provisions of ICAO and the requirements for continued airworthiness of aircraft, air operators shall submit use difficulty reports to the national civil aviation authorities and aircraft manufacturers of their country in order to improve the design of aircraft products and improve the reliability of safe aircraft operations. In order to avoid repeated failures and improve maintenance levels, air operators shall identify faulty parts or failed systems by establishing an investigation procedure for use difficulty reports. For some major or typical incidents, the authority will participate in the investigation. For unsafe conditions confirmed by comprehensive evaluation, the airworthiness authority will issue a corresponding airworthiness directive (AD).

[0006] The existing use difficulty reporting, investigation and management procedures effectively support the safety management work in the field of continuing airworthiness. The use difficulty reporting system also serves as an industry information sharing platform, gathering the experience of all parties and improving the maintenance management level and operational efficiency. However, continuing airworthiness management still faces the following problems:

[0007] (1) The reporting itself has a lag, which to some extent also leads to a lag in regulatory response and intervention when facing major or typical events;

[0008] (2) The report data contains a large amount of descriptive unstructured text data, which is difficult to standardize and unify, and lacks detailed data support on the actual operation of aircraft / engines, which increases the difficulty of supervision and slows down the confirmation and response speed for similar sudden high-frequency events;

[0009] (3) There is a lack of more timely and effective means to achieve closed-loop management of supervision for incidents that are required to be reported by regulations;

[0010] (4) The capabilities and roles of fleet health status monitoring and continuous airworthiness management need to be further enhanced. Summary of the invention

[0011] The purpose of the present invention is to solve the problems existing in the above-mentioned prior art, and to provide an aircraft continuous airworthiness auxiliary management system and method based on data evidence, which integrates aircraft / engine failure cases and airborne QAR data, and strengthens the role of data evidence and data intelligence in continuous airworthiness management. The QAR data of the actual operation of the aircraft can be used as direct evidence for failure case analysis and investigation, and the statistical laws obtained through historical data and the trained model algorithms can be used as indirect evidence for fleet health management, which can serve both the supervision of the authority and the fleet health management of airlines and manufacturers, and help all parties improve their ability to manage continuous airworthiness, timely discover design and manufacturing defects during the operation of the aircraft, and continuously track and ensure its airworthiness status.

[0012] The present invention is achieved through the following technical solutions:

[0013] A first aspect of the present invention provides an aircraft continuous airworthiness auxiliary management system based on data evidence, the system comprising: a database layer and an application layer;

[0014] The database layer includes: a fault event case database, an aircraft use information database, a flight onboard QAR database and a QAR parameter list database;

[0015] The application layer includes: a fault event case management module, a flight QAR data analysis module and a fleet health management module.

[0016] A further improvement of the present invention is:

[0017] The fault event case database is used to store fault event information;

[0018] The aircraft usage information database is used to store aircraft operation flight information;

[0019] The flight airborne QAR database is used to store the decoded engineering value files of QAR data by flight segment;

[0020] The QAR parameter list database is used to store parameter documents of the decoding library corresponding to each model.

[0021] A further improvement of the present invention is:

[0022] The fault event case database and the aircraft use information database both adopt structured storage;

[0023] The flight onboard QAR database and QAR parameter list database are both stored in files.

[0024] A further improvement of the present invention is:

[0025] The fault event case management module includes:

[0026] Case addition, deletion, modification and query submodule: connected to the fault event case database; adding, batch importing, modifying and deleting case data one by one;

[0027] Case multi-dimensional statistical analysis submodule: connected to the fault event case database; interactive multi-dimensional statistical analysis and visualization of case data;

[0028] Case analysis submodule: connected to the fault event case database and the aircraft use information database respectively; according to the case occurrence date, aircraft registration number and flight number, the aircraft's flight information and QAR data file name at the time of the case occurrence and before the case occurrence are obtained from the aircraft use information database, and sorted by take-off time;

[0029] Case intelligent monitoring submodule: connected to the fault event case database; analyzes the similarity between cases, realizes automatic classification of similar fault cases, and issues alarms for high-frequency and high-similarity reported cases within the monitoring period according to the set threshold, and continuously tracks the marked key cases, monitors and pushes the reporting status of similar cases.

[0030] A further improvement of the present invention is:

[0031] The flight QAR data analysis module includes:

[0032] Parameter profile management submodule: connects to the QAR parameter list database; checks the parameter list of the aircraft decoding library according to the decoding library version number corresponding to the QAR data, provides parameter profiles, and provides parameter selection items;

[0033] Data batch processing and calling submodule: connect with the fault event case database, aircraft use information database and flight onboard QAR database respectively; obtain the flight information of the aircraft at the time of the case occurrence and the previous k times from the aircraft use information database according to the case occurrence date, aircraft registration number and flight number in the fault event case set; extract and export QAR data in batches from the flight onboard QAR database;

[0034] Case data analysis submodule: Analyze the QAR data of the specific case and the previous flights.

[0035] A further improvement of the present invention is:

[0036] The fleet health management module includes:

[0037] Big data-based status monitoring submodule: Identify and warn of abnormal status of the fleet through big data model algorithms;

[0038] Fault and abnormality directional screening submodule: traverses the flight QAR database to screen faults and abnormal conditions;

[0039] Performance trend monitoring submodule: Sort by flight departure time, select parameters and data compression methods, and view the changing patterns of QAR data-related performance parameters; monitor and track aircraft systems, engine performance changes, and degradation patterns, and evaluate and manage the performance and health status of the fleet;

[0040] Data algorithm management platform sub-module: manages data sets and algorithms, realizes the division, storage, import and export of data sets, and realizes the development, import and verification of model algorithms.

[0041] A further improvement of the present invention is:

[0042] The three sub-modules of big data-based status monitoring, fault anomaly targeted screening, and performance trend monitoring are all connected to the aircraft usage information database and the flight onboard QAR database, and call the algorithms in the data algorithm management platform sub-module to achieve user fleet health management.

[0043] A second aspect of the present invention provides a data-based evidence-based aircraft continuous airworthiness auxiliary management method, the method is implemented using the above system, and the method includes:

[0044] The case intelligent monitoring submodule in the fault event case management module issues an alarm after monitoring the high-frequency and high-similarity fault case reports in the system over the past period of time;

[0045] After confirming that the faults all point to the A system of the same aircraft model, the parameter profile management submodule in the flight QAR data analysis module selects the relevant parameter profiles of the A system of the aircraft model. At the same time, the data batch processing and call submodule extracts the QAR data of the current and previous flights when the fault occurs. The case data analysis submodule uses the data representation of the evidence-based fault occurrence.

[0046] The data algorithm management platform submodule in the fleet health management module screens the occurrence of this fault on all flights in the industry in the flight onboard QAR database, and verifies it with the query and statistics of the fault reporting in the fault event case management module to form a closed loop of supervision.

[0047] The present invention also provides a second data-based evidence-based aircraft continuous airworthiness auxiliary management method, which is implemented using the above system and includes:

[0048] The data batch processing and calling submodule in the flight QAR data analysis module extracts the QAR data without abnormality / fault labels of the aircraft model to be analyzed;

[0049] Developing the M system anomaly detection algorithm through unsupervised machine learning in the data algorithm management platform submodule of the fleet health management module;

[0050] The case addition, deletion, modification and query submodule in the fault event case management module queries the fault anomaly cases of the M system of the model to be analyzed;

[0051] The data batch processing and calling submodule in the flight QAR data analysis module extracts the QAR data of the preceding flight of the fault case to verify the M system anomaly detection algorithm, obtains the verified algorithm, and sends the verified algorithm to the big data-based status monitoring submodule in the fleet health management module;

[0052] The big data-based state monitoring submodule uses the verified algorithm to realize the prediction and early warning of abnormal state of M system.

[0053] The present invention also provides a third data-based evidence-based aircraft continuous airworthiness auxiliary management method, which is implemented using the above system and includes:

[0054] The data batch processing and calling submodule in the flight QAR data analysis module accesses the aircraft usage information database to extract the full life cycle QAR data of multiple engines of the same model;

[0055] The data algorithm management platform submodule in the fleet health management module studies the degradation law and prediction algorithm of this type of engine, while the performance trend monitoring submodule monitors and tracks the performance of this type of engine.

[0056] Compared with the prior art, the present invention has the following beneficial effects:

[0057] The present invention enables continuous airworthiness management through the deep integration of actual aircraft / engine operation data and fault cases, which can provide more powerful data-based evidence means for supervision and enhance the initiative and convenience of event investigation; by clustering reported fault cases, intelligent alarms for high-frequency similar faults or anomalies are realized, guiding supervision to intervene earlier, and conducting investigations and tracking within the industry based on fault characteristics and QAR data, which helps to move the safety window forward; by screening events required to be reported by regulations through data-based evidence means, closed-loop management of supervision can be achieved and safety culture can be promoted; based on a large amount of historical flight operation data, model algorithms are studied to transform flight QAR data resources into data intelligence for fleet health status monitoring, evaluation and management, which can provide strong support for the continuous safe operation of civil aviation. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 Use the Difficulty Reporting System Relationship Diagram;

[0059] Figure 2 Schematic diagram of the composition structure of the system of the present invention. DETAILED DESCRIPTION

[0060] The present invention is further described in detail below in conjunction with the accompanying drawings:

[0061] As a major civil aviation country, my country has installed QAR equipment on all registered aircraft and continuously accumulated massive amounts of airborne data during continuous operation. QAR data can continuously and completely record the status of each aircraft / engine system in actual operation, which makes it possible to empower continuous airworthiness management through QAR data. Continuous airworthiness management in the era of big data should strengthen the role of data evidence and data intelligence within the existing system framework and achieve breakthroughs in the following aspects:

[0062] (1) Linking the onboard data of the flight on the day the failure occurred and the previous flight to strengthen the evidence-based role of data in failure case analysis and investigation, and enhance the means and convenience of investigation;

[0063] (2) Through algorithms, similar fault cases can be quickly retrieved and associated, and the same fault modes can be mined through data to achieve intelligent warning, troubleshooting, and continuous tracking and monitoring of high-frequency faults, prompting regulators to intervene earlier and move the safety window forward. This is particularly important in the early stages of the commissioning of newly introduced aircraft and domestically produced civil aircraft.

[0064] (3) For events that are required to be reported by regulations, the fleet’s historical operating data is screened through fault characteristics / algorithms to achieve closed-loop management of supervision;

[0065] (4) Monitor and manage the health status of the fleet based on big data to achieve early screening and early warning of abnormalities and failures;

[0066] (5) Develop fault diagnosis, performance evaluation, and trend prediction algorithms to improve fleet health management capabilities.

[0067] The present invention is mainly aimed at industry supervision, and also has a certain reference and reference role in the continuous airworthiness management of airlines and manufacturers.

[0068] like Figure 2 As shown, the data-based evidence-based aircraft continuous airworthiness auxiliary management system provided by the present invention includes: a database layer and an application layer, which are as follows:

[0069] The database layer includes: fault event case database, aircraft usage information database, flight onboard QAR database and QAR parameter list database, and organizes and manages the underlying data in a scientific and effective manner to facilitate the application layer to quickly retrieve and call data (the organization and management of the underlying data mainly refers to the field information contained in each database described below and the data storage format (structured storage / file storage), which is the basis for the association between databases and the call of data by each functional module.).

[0070] The fault event case database is used to store fault event information in a structured manner, and the fault event information includes but is not limited to events within the scope of the use difficulty reporting regulations, abnormal events that lead to important consequences (flight return, alternate landing, taxiing back, aborted takeoff, ground parking, air parking) and delays of more than 2 hours, events involving product design defects, repair quality issues, frequent faults, typical faults, and events where the cause of the fault cannot be determined, etc. Specifically, the fault event case database uses a structured table to store cases, and the fields include: case number, reporting unit, date of occurrence, flight number, flight phase, take-off airport, landing airport, operator, aircraft registration number, aircraft model, engine model, engine serial number, faulty component / faulty location, fault mode, ATA chapter corresponding to the fault, fault phenomenon, fault consequence, fault cause, treatment measures, and a complete description of the problem.

[0071] The aircraft usage information database is used to structuredly store aircraft operation flight information, which includes: flight number, airline, take-off and landing airports, take-off time (14-bit time format), aircraft model, engine model, aircraft registration number, engine serial number, abnormal / fault label, QAR data file name, and decoding library version number. The aircraft usage information database stores the daily flight information of the fleet aircraft.

[0072] The flight airborne QAR database is used to store the decoded engineering value files (csv format) of QAR data by flight segment. The flight airborne QAR database stores the airborne QAR data of each flight in the fleet.

[0073] The QAR parameter list database is used to store the parameter documents of the corresponding decoding library of each model. The parameter documents include: parameter name, meaning, corresponding ATA chapter, data type, unit, value range, precision, data source and other information, so that users can query and select relevant parameters of the aircraft / engine system during the data extraction process and establish parameter profiles.

[0074] Structured storage refers to the use of a two-dimensional table structure to store data. Among the four databases, the fault event case database and the aircraft use information database both use structured storage, and the flight onboard QAR database and the QAR parameter list database both use file storage.

[0075] The application layer includes three modules: fault event case management module, flight QAR data analysis module and fleet health management module. The functions of these three modules and their sub-modules all rely on the support of the underlying data. The functional design of the three modules is as follows:

[0076] Fault incident case management module:

[0077] This module is used to support the maintenance of the fault event case database, including the addition, batch import, modification and deletion of cases; it supports interactive multi-dimensional statistical analysis and visualization, including case category, aircraft model, corresponding ATA chapter, flight phase, time and other dimensions, so that users can quickly grasp the characteristics of case distribution, so as to carry out deeper data analysis and mining work in a targeted manner.

[0078] This module is used to support users to quickly search for cases with specific tags or fields containing specific keywords. Users can view and filter the records of interest one by one through the returned results, mark and create new groups, and create and save customized case sets. For specific cases, users can refer to the QAR data of the time and previous flights based on the case data, and evaluate the pilot's handling and decision-making based on the underlying causes and signs of the failure. Since some fields of the case data correspond to unstructured text data, the module construction algorithm measures the case similarity. For the fault cases retrieved or selected by the user, the system will automatically push high-similarity cases for reference.

[0079] This module can cluster cases through algorithms (for case entries stored in the fault event case database, existing clustering analysis, text similarity judgment and matching algorithms can be used to measure the similarity between cases and realize automatic classification of similar fault cases). It can issue alarms for high-frequency and high-similarity reported cases during the monitoring period according to threshold settings, and continuously track key cases marked by users to monitor the reporting of similar cases.

[0080] The fault event case management module includes 4 submodules, namely: case addition, deletion, modification and query submodule, case multidimensional statistical analysis submodule, case analysis submodule, and case intelligent monitoring submodule. The following describes in detail the functions of each submodule and the functional connection and data interaction between each submodule and the underlying database:

[0081] 1. Case addition, deletion, modification and query submodule:

[0082] This submodule is connected to the fault event case database to support users' access, maintenance and management of the fault event case database, including adding, batch importing, modifying and deleting case data one by one.

[0083] In terms of case query, this submodule supports users to quickly search for cases with specific tags or fields containing specific keywords, such as filtering engine mid-air shutdown fault cases based on the fault consequence field, filtering fault cases of a certain model of air conditioning system (corresponding to ATA chapter number 21) based on the aircraft model and the ATA chapter field corresponding to the fault, and filtering fault cases containing the keyword "crack" in the complete description of the problem. Users can click on the case entry of interest based on the returned filtering results. After clicking, the case analysis submodule will be launched to view detailed case data.

[0084] Case data contains both label data and unstructured text data. The module can apply algorithms to measure case similarity. For fault cases retrieved or selected by users, the system will automatically push highly similar cases for reference. This submodule also supports users to mark and group cases, create, export, save and access customized case sets. When saving, only the case number needs to be saved, and when reading, the case data can be associated through the fault event case database.

[0085] 2. Case multidimensional statistical analysis submodule:

[0086] This submodule is connected to the fault event case database, supporting users to realize interactive multi-dimensional statistical analysis and visualization of fault event case data, including case category, aircraft model, corresponding ATA chapter, flight phase, time and other dimensions, so that users can quickly grasp the characteristics of case distribution, and thus carry out deeper data analysis and mining work in a targeted manner.

[0087] 3. Case analysis submodule:

[0088] This submodule supports users to view detailed case data of specific cases. This submodule is connected to the fault event case database and the aircraft use information database respectively. It can obtain the flight information of the aircraft at the time of the case and the previous flight information and QAR data file name from the aircraft use information database according to the case occurrence date, aircraft registration number and flight number, and sort them according to the take-off time (14-bit time format). Users can jump to the data batch processing and call submodule in the flight QAR data analysis module to extract, process and call case QAR data, and use the case analysis submodule in the flight QAR data analysis module to evaluate the pilot's handling and decision-making based on the deep causes and signs of the fault.

[0089] 4. Case intelligent monitoring submodule:

[0090] This submodule is connected to the fault event case database. For the stored case entries, existing algorithms such as cluster analysis, text similarity judgment and matching are used to measure the similarity between cases, realize automatic classification of similar fault cases, and issue alarms for high-frequency and high-similarity reported cases within the monitoring period according to threshold settings (time period, frequency, similarity). It also continuously tracks key cases marked by users, monitors and pushes the reporting status of similar cases.

[0091] Flight QAR data analysis module:

[0092] This module supports users to batch associate flight QAR data (including the time when the fault / incident occurred and the specified number of previous flights) according to customized case sets (the case addition, deletion, modification and query submodule in the fault event case management module supports users to mark cases and create new groups, create, export, save and access customized case sets. The flight QAR data analysis module can access the case sets stored by users.) (This module can associate the aircraft usage information database according to the fault event case occurrence date, aircraft registration number and flight number, return the aircraft's flight information and QAR data file name at the time of the case occurrence and the previous k times (a specified number) and sort by take-off time (14-bit time format).), thereby creating a research dataset for fault diagnosis and prediction algorithms (this algorithm means that users may need to develop fault diagnosis and prediction algorithms for specific fault modes, and need to screen corresponding fault cases and obtain QAR data to create an algorithm research dataset.) (Batch extract and export QAR data from the flight onboard QAR database according to the batch-associated QAR data file name.).

[0093] This module supports batch extraction and export of QAR data. For parameter profile selection during data extraction, the module can provide commonly used typical parameter profiles for users to choose from based on the parameter list of the decoding library corresponding to the QAR data to be extracted, and guide users to select parameter items according to the ATA chapter corresponding to the aircraft / engine system through the guidance interface. During the parameter selection process, the parameter name, meaning, corresponding ATA chapter, data type, unit, value range, precision, and data source are all information that must be provided to users for reference to ensure the accuracy and effectiveness of data use. The module supports users to save and manage the created parameter profiles.

[0094] This module has built-in and supports user-developed algorithms (this module can call the system's built-in Python-based algorithm development platform, supporting users to select system default algorithms or develop their own algorithms to perform batch preprocessing, feature extraction, and statistical analysis on QAR data.) It performs batch preprocessing, feature extraction, and statistical analysis on QAR data.

[0095] For specific case analysis, this module supports QAR data visualization and exploratory data analysis functions, including but not limited to parallel visualization output of multiple parameters according to time series, drawing scatter plot matrices reflecting parameter correlations, and drawing performance parameter evolution trend charts.

[0096] The flight QAR data analysis module includes three submodules, namely: parameter profile management submodule, data batch processing and call submodule, and case data analysis submodule. The following is a relatively detailed description of the functions of each submodule, as well as the functional connection and data interaction between each submodule and the underlying database:

[0097] 1. Parameter profile management submodule:

[0098] This submodule is connected to the QAR parameter list database, supports users to access it during the QAR data extraction process, and checks the parameter list of the aircraft decoding library according to the decoding library version number corresponding to the QAR data. It provides commonly used typical parameter profiles for users to choose from, and provides parameter selection items. Through the guidance interface, users are guided to select parameters according to the ATA chapter (Aircraft System Component Common Code) corresponding to each system / subsystem / component of the aircraft / engine. During the parameter selection process, the parameter name, meaning, corresponding ATA chapter, data type, unit, value range, precision, and data source are all information that must be provided to users for reference to ensure the accuracy and effectiveness of data use. The module supports users to save and manage custom created parameter profiles.

[0099] 2. Data batch processing and calling submodules:

[0100] This submodule is connected to the fault event case database, aircraft usage information database and flight airborne QAR database respectively, and is used to obtain the aircraft's flight information at the time of the case occurrence and the previous k times (a specified number) from the aircraft usage information database according to the case occurrence date, aircraft registration number and flight number in the fault event case set (which can be the case queried in the case addition, deletion, modification and query submodule in the fault event case management module, or the customized case set saved by the user), batch extract and export QAR data from the flight airborne QAR database, and create a custom research data set for the development needs of fault diagnosis and prediction algorithms for specific fault modes. This submodule also supports users to directly access the aircraft usage information database, query relevant flight information, and extract QAR data.

[0101] This submodule can call the system's built-in algorithm development platform based on Python language, supporting users to select the system default algorithm or develop their own algorithm to achieve batch preprocessing, feature extraction and statistical analysis of QAR data.

[0102] The extracted and processed data sets can be saved to the Fleet Health Management Module-Data Algorithm Management Platform submodule for the development of fleet health management model algorithms.

[0103] 3. Case data analysis submodule:

[0104] This submodule supports the analysis of QAR data of the current and previous flights of a specific case. The module supports QAR data visualization and exploratory data analysis functions, including but not limited to parallel visualization output of multiple parameters according to time series, drawing scatter plot matrices reflecting parameter correlations, and drawing performance parameter evolution trend graphs.

[0105] Fleet Health Management Module:

[0106] This module includes: a big data-based status monitoring submodule, a fault anomaly directional screening submodule, a performance trend monitoring submodule and a data algorithm management platform submodule. The data algorithm management platform submodule mainly implements user data set management and algorithm development and management. The three submodules of big data-based status monitoring, fault anomaly directional screening and performance trend monitoring are connected with the aircraft usage information database and the flight onboard QAR database, and the algorithms in the data algorithm management platform submodule are called to realize the relevant requirements of user fleet health management, as follows:

[0107] 1. The state monitoring submodule based on big data uses the existing big data model algorithm (for example, conventional anomaly detection algorithms such as one-class SVM, DBSCAN, Isolation Forest, etc. can be used to meet relevant requirements.) to identify and warn the abnormal state of the fleet. For the target aircraft models and target systems to be monitored, based on a large amount of historical flight data in the system, the anomaly detection model is trained through supervised or unsupervised learning, and the model effect is verified through the airborne QAR data of the current and previous flights (with labels) when the fault case occurs, and the model algorithm is developed to achieve prediction and early warning of abnormal aircraft states (the flight airborne QAR database stores the QAR data of the fleet's daily flights, and the anomaly detection algorithm can identify abnormal aircraft states through flight QAR data.).

[0108] 2. The fault anomaly targeted screening submodule supports users to traverse the flight QAR database according to features, fault diagnosis and prediction algorithms (the algorithm here refers to the algorithm developed by the user according to needs, which is used for screening, diagnosis and prediction of specific anomalies / faults.) and screen faults and abnormal conditions in a targeted manner. On the one hand, it is used to screen the operator's reporting of events required to be reported by regulations to achieve closed-loop management of supervision; on the other hand, for abnormal situations whose causes cannot be determined temporarily, it can also be promptly investigated according to data features and all parties can be mobilized to conduct investigations; in addition, it can also predict a specific fault (the flight onboard QAR database stores the QAR data of the fleet's daily flights, and the fault prediction algorithm can predict and warn of the target fault.), promote refined safety management, and improve fleet operating efficiency.

[0109] 3. The performance trend monitoring submodule supports users to select aircraft / engines through the aircraft usage information database, sort by flight departure time, select parameters and data compression methods (day, month), and view the change patterns of performance parameters related to QAR data. The module also supports the use of algorithms (developed by users according to their needs) to monitor and track aircraft system / engine performance changes / degradation patterns, and evaluate and manage the performance and health status of the fleet.

[0110] 4. The data algorithm management platform submodule supports users to manage data sets and algorithms, realize the division, storage, import and export of data sets, and realize the development, import and verification of model algorithms. This submodule can call the system's built-in algorithm development platform based on Python language, and supports users to develop related algorithms based on big data status monitoring, fault anomaly directional screening, and performance trend monitoring according to their needs.

[0111] Each module of the above-mentioned system is implemented by computer programming, and the statistics, analysis, early warning, etc. of the data can be realized by using a variety of existing algorithms. The application of the above-mentioned system and its support and auxiliary role in continuous airworthiness management are very flexible in methods and means. The above-mentioned system deeply integrates fault event case data, aircraft usage information data and aircraft / engine actual operation airborne QAR data to enable continuous airworthiness management. The core of the so-called data-based evidence is that the QAR data of the actual operation of the aircraft can be used as direct evidence for fault case analysis and investigation, and the statistical laws obtained through historical data and the trained model algorithms can be used as indirect evidence for fleet health management.

[0112] The following is an introduction to the data-based evidence-based aircraft continuous airworthiness auxiliary management method based on the application cases of the system in specific scenarios:

[0113] [Example 1]

[0114] The case intelligent monitoring submodule in the fault event case management module will issue an alarm after monitoring the high-frequency and high-similarity fault case reports in the system in the past period. The user will check the detailed case data one by one in conjunction with the case analysis submodule in the fault event case management module. After confirming that the faults all point to the A system of the same model, the parameter profile management submodule in the flight QAR data analysis module selects the relevant parameter profiles of the A system of the model. At the same time, the data batch processing and call submodule extracts the QAR data of the current and previous flights when the fault occurred. The data representation of the fault occurrence is based on the case data analysis submodule, and the cause of the fault is confirmed and investigated through multiple parties (airlines, manufacturers, etc.) to explore the continued airworthiness of the aircraft. Through fault feature analysis, characteristic parameters that can reflect the occurrence and symptoms of faults are confirmed in the QAR data. The data algorithm management platform submodule in the fleet health management module uses a fault-oriented screening (diagnosis) algorithm to screen the occurrence of this fault in all flights in the industry in the flight onboard QAR database, and verify it with the query and statistics of the fault report in the fault event case management module to form a closed loop of supervision. Since there are obvious fault signs in the QAR data before the fault occurs, that is, the fault is predictable, the fault prediction algorithm is developed to predict and warn of the fault, thereby improving the safety and efficiency of fleet operations.

[0115] In the above process, the QAR data of the actual operation of the faulty flight serves as direct evidence for fault case analysis and investigation, and the fault-oriented screening (diagnosis) algorithm and fault prediction algorithm can serve as indirect evidence for fleet health management, both of which are based on the actual flight QAR data of the flight.

[0116] [Example 2]

[0117] The sensor measurement parameters of the M system of a certain aircraft model (the model to be analyzed) are numerous, and a large amount of QAR data of its "healthy" flight is stored in the system. The data batch processing and calling submodule in the flight QAR data analysis module extracts the QAR data without abnormality / fault labels of the model to be analyzed, and the M system anomaly detection algorithm is developed through unsupervised machine learning in the data algorithm management platform submodule in the fleet health management module; the case addition, deletion, modification and query submodule in the fault event case management module queries the fault abnormality cases of the M system of the model to be analyzed, and the data batch processing and calling submodule in the flight QAR data analysis module extracts the QAR data of the preceding flight of the fault case to verify the algorithm (various existing methods can be used for verification, which will not be repeated here.), obtains the verified algorithm, and sends the verified algorithm to the big data-based state monitoring submodule in the fleet health management module. The big data-based state monitoring submodule uses the verified algorithm to realize the prediction and early warning of the abnormal state of the M system, thereby assisting the fleet health management.

[0118] [Example 3]

[0119] Engine exhaust temperature (EGT) is an important indicator reflecting the decline in engine performance. The data batch processing and calling submodule in the flight QAR data analysis module accesses the aircraft usage information database and extracts the full life cycle QAR data of multiple engines of the same model. The data algorithm management platform submodule in the fleet health management module studies the decline law and prediction algorithm of this type of engine. At the same time, the performance trend monitoring submodule monitors and tracks the performance of this type of engine.

[0120] Finally, it should be explained that the above technical solution is only one implementation method of the present invention. For those skilled in the art, it is easy to make various types of improvements or modifications based on the application methods and principles disclosed in the present invention, and it is not limited to the method described in the above specific implementation method of the present invention. Therefore, the method described above is only preferred and does not have a restrictive meaning.

Claims

1. A data-based evidence-based aircraft continuous airworthiness support management system. Features: The system comprises: a database layer and an application layer; The database layer includes: a fault event case database, an aircraft use information database, a flight onboard QAR database and a QAR parameter list database; The application layer includes: a fault event case management module, a flight QAR data analysis module and a fleet health management module; The fault event case management module includes: Case addition, deletion, modification and query submodule: connected to the fault event case database; adding, batch importing, modifying and deleting case data one by one; Case multi-dimensional statistical analysis submodule: connected to the fault event case database; interactive multi-dimensional statistical analysis and visualization of case data; Case analysis submodule: connected to the fault event case database and the aircraft use information database respectively; according to the case occurrence date, aircraft registration number and flight number, the aircraft's flight information and QAR data file name at the time of the case occurrence and before the case occurrence are obtained from the aircraft use information database, and sorted by take-off time; Case intelligent monitoring submodule: connected to the fault event case database; analyzes the similarity between cases, realizes automatic classification of similar fault cases, and issues alarms for high-frequency and high-similarity reported cases within the monitoring period according to the set threshold, and continuously tracks the marked key cases, monitors and pushes the reporting status of similar cases; The flight QAR data analysis module includes: Parameter profile management submodule: connects to the QAR parameter list database; checks the parameter list of the aircraft decoding library according to the decoding library version number corresponding to the QAR data, provides parameter profiles, and provides parameter selection items; Data batch processing and calling submodule: connect with the fault event case database, aircraft use information database and flight onboard QAR database respectively; obtain the flight information of the aircraft at the time of the case occurrence and the previous k times from the aircraft use information database according to the case occurrence date, aircraft registration number and flight number in the fault event case set; extract and export QAR data in batches from the flight onboard QAR database; Case data analysis submodule: analyzes the QAR data of the specific case and the previous flights; The fleet health management module includes: Big data-based status monitoring submodule: Identify and warn of abnormal status of the fleet through big data model algorithms; Fault and abnormality directional screening submodule: traverses the flight QAR database to screen faults and abnormal conditions; Performance trend monitoring submodule: Sort by flight departure time, select parameters and data compression methods, and view the changing patterns of QAR data-related performance parameters; monitor and track aircraft systems, engine performance changes, and degradation patterns, and evaluate and manage the performance and health status of the fleet; Data algorithm management platform submodule: manages data sets and algorithms, realizes the division, storage, import and export of data sets, and realizes the development, import and verification of model algorithms; The case intelligent monitoring submodule in the fault event case management module issues an alarm after monitoring the high-frequency and high-similarity fault case reports in the system over the past period of time; After confirming that the faults all point to the A system of the same aircraft model, the parameter profile management submodule in the flight QAR data analysis module selects the relevant parameter profiles of the A system of the aircraft model. At the same time, the data batch processing and call submodule extracts the QAR data of the current and previous flights when the fault occurs. The case data analysis submodule uses the data representation of the evidence-based fault occurrence. The data algorithm management platform submodule in the fleet health management module screens the occurrence of this fault on all flights in the industry in the flight onboard QAR database, and verifies it with the query and statistics of the fault reporting in the fault event case management module to form a closed loop of supervision.

2. According to claim 1, the data-based evidence-based aircraft continuous airworthiness auxiliary management system, Features: The fault event case database is used to store fault event information; The aircraft usage information database is used to store aircraft operation flight information; The flight airborne QAR database is used to store the decoded engineering value files of QAR data by flight segment; The QAR parameter list database is used to store parameter documents of the decoding library corresponding to each model.

3. The data-based evidence-based aircraft continuous airworthiness support management system according to claim 2, Features: The fault event case database and the aircraft use information database both adopt structured storage; The flight onboard QAR database and QAR parameter list database are both stored in files.

4. According to claim 3, the data-based evidence-based aircraft continuous airworthiness auxiliary management system, Features: The three sub-modules of big data-based status monitoring, fault anomaly targeted screening, and performance trend monitoring are all connected to the aircraft usage information database and the flight onboard QAR database, and call the algorithms in the data algorithm management platform sub-module to achieve user fleet health management.

5. A data-based evidence-based auxiliary management method for aircraft continuous airworthiness. Features: The method is implemented by using the system according to any one of claims 1 to 4, and the method comprises: The case intelligent monitoring submodule in the fault event case management module issues an alarm after monitoring the high-frequency and high-similarity fault case reports in the system over the past period of time; After confirming that the faults all point to the A system of the same aircraft model, the parameter profile management submodule in the flight QAR data analysis module selects the relevant parameter profiles of the A system of the aircraft model. At the same time, the data batch processing and call submodule extracts the QAR data of the current and previous flights when the fault occurs. The case data analysis submodule uses the data representation of the evidence-based fault occurrence. The data algorithm management platform submodule in the fleet health management module screens the occurrence of this fault on all flights in the industry in the flight onboard QAR database, and verifies it with the query and statistics of the fault reporting in the fault event case management module to form a closed loop of supervision.

6. A data-based evidence-based auxiliary management method for aircraft continuous airworthiness. Features: The method is implemented by using the system according to any one of claims 1 to 4, and the method comprises: The data batch processing and calling submodule in the flight QAR data analysis module extracts the QAR data without abnormality / fault labels of the aircraft model to be analyzed; Developing the M system anomaly detection algorithm through unsupervised machine learning in the data algorithm management platform submodule of the fleet health management module; The case addition, deletion, modification and query submodule in the fault event case management module queries the fault anomaly cases of the M system of the model to be analyzed; The data batch processing and calling submodule in the flight QAR data analysis module extracts the QAR data of the preceding flight of the fault case to verify the M system anomaly detection algorithm, obtains the verified algorithm, and sends the verified algorithm to the big data-based status monitoring submodule in the fleet health management module; The big data-based state monitoring submodule uses the verified algorithm to realize the prediction and early warning of abnormal state of M system.

7. A data-based evidence-based auxiliary management method for aircraft continuous airworthiness. Features: The method is implemented by using the system according to any one of claims 1 to 4, and the method comprises: The data batch processing and calling submodule in the flight QAR data analysis module accesses the aircraft usage information database to extract the full life cycle QAR data of multiple engines of the same model; The data algorithm management platform submodule in the fleet health management module studies the degradation law and prediction algorithm of this type of engine, while the performance trend monitoring submodule monitors and tracks the performance of this type of engine.