Method and system for determining a faulty aircraft component

Abstract

This digital system makes it possible to determine at least one faulty component of an aircraft.It comprises: - a first module (MDS1) for determining at least one first symptom (S1i) of the aircraft from state data (DEk) collected under the wing; - a second module (MDS2) for determining at least one second symptom (S2i) of the aircraft from degradation data (DCk) representative of an abnormal variation in flight data; - a module (MDPD) for determining the probabilities of failure of a plurality of aircraft components from said at least one first and second symptoms, this module comprising a Bayesian network (BR) trained from a knowledge base (BC) including data produced by a natural language processing module (MNLP) configured to process aircraft maintenance reports; - a module (MSAC) configured to flag at least one failing component among said plurality of components based on said probabilities. Figure for the abbreviation: Fig. 1.

Description

Title of the invention: Method and system for determining a faulty component of an aircraft Previous technique

[0001] The present invention relates to the general field of aeronautics.

[0002] It relates more particularly to a method and a decision support system for identifying faulty components of an aircraft.

[0003] An aircraft can be considered as a complex system consisting of a number of hardware or software components (computer, sensors, ...) and that these are liable to fail.

[0004] When a failure is suspected, it is common for an operator to move near the aircraft (also called "under the wing") to perform tests in order to try to determine the faulty component(s) causing the failure.

[0005] In practice, the underwing operator follows a diagnostic scenario comprising a linear list of instructions, for example presented as a troubleshooting flowchart in a maintenance document in paper or digital format. These maintenance documents have a very long update cycle and are published approximately twice a year.

[0006] When the operator suspects that one or more components are faulty, he drops off these components to send them to a repair shop for further analysis, with industrial testing equipment.

[0007] In order to intervene quickly or due to insufficient training, it is common for the underwing operator to remove an excessive number of components. Furthermore, it may turn out, during the workshop analysis phase, that the components presumed to be faulty are not the actual cause of the failure and were therefore removed incorrectly.

[0008] For example, a wing operator may assume that a sensor is faulty, remove the sensor, and send it to the workshop for repair. During analysis of the supposedly faulty sensor in the workshop, it may be determined that the sensor is functioning normally and that the actual cause of the failure is in fact an electronic board configured to process the data measured by the sensor.*

[0009] A maintenance report is then drawn up, this report including, for example, warning messages or fault alarms, results of measurements carried out by the underwing operator or by the repair technician in the workshop, symptoms detected by the underwing operator or by the repair technician, the components determined to be faulty and the actual cause of the fault determined after analysis. This maintenance report may by example of being digital.

[0010] In the aeronautical industry in particular, there is a known reliability indicator called NFF (No Fault Found) which represents a rate of unplanned and unjustified removals calculated from the ratio between the mean time between unplanned removals MTBUR (Mean Time Between Unscheduled Removal) and the mean net time between unplanned, justified removals MTBF (Mean Time Between Failure). These indicators satisfy the following equality: MTBUR = (1 - NFF) * MTBF.

[0011] It is known that the cost of these unplanned NFF drops is very significant and that it strongly impacts the operating cost of airlines.

[0012] The invention aims to overcome at least some of these drawbacks by providing a fault determination system that allows the underwing operator to better identify the actual cause of an aircraft failure, and consequently the component(s) to be removed, so as to improve the NFF reliability indicator. Description of the invention

[0013] The present invention addresses this need in particular by proposing a digital system for determining at least one faulty component of an aircraft, this system comprising: - an aircraft failure information module; - a first module for determining at least one initial symptom of the aircraft from condition data collected under the wing; - a second module for determining at least one second symptom of the aircraft from degradation data representative of an abnormal variation in flight data; - a module for determining the probability of failure of a plurality of aircraft components based on said at least one first and second symptoms, this module comprising a Bayesian network trained from a knowledge base including data produced by a natural language processing module configured to process aircraft maintenance reports, - a module configured to signal at least one faulty component among said plurality of components according to said probabilities.

[0014] Correspondingly, the invention relates to a method for determining at least one faulty component of an aircraft, this method implemented by computer comprising the following steps: - obtaining information about an aircraft malfunction; - determination of at least one initial symptom of the aircraft from condition data collected under the wing; - determination of at least one second symptom of the aircraft from degradation data representative of an abnormal variation in flight data; - determination of failure probabilities of a plurality of aircraft components from said at least one first and second symptoms, this determination using a Bayesian network trained from a knowledge base comprising data produced by a natural language processing module configured to process aircraft maintenance reports, - reporting of at least one faulty component among said plurality of components according to said probabilities.

[0015] Thus, and very advantageously, the digital system determines the failing components using both "state data" and "degradation data".

[0016] Status data represents the state of the aircraft. This data can be collected by an operator (or a robot), for example with measuring equipment, operating close to the aircraft, i.e., "under the wing." It is, for example, transmitted via a human-machine interface through alarms or fault messages. It can sometimes be directly determined or observed by the operator when the latter detects a noise or odor characteristic of a malfunction.

[0017] On the contrary, degradation data, or behavioral data, are data that represent a trend, or an abnormal variation in flight data. They can be determined, for example: - when flight data changes and exceeds a threshold; - when a signature of a series of flight data is representative of a degradation in the aircraft's operation; - by processing flight data continuously.

[0018] Unlike state data, this degradation data is difficult for the operator to access. However, it is very rich in information, and taking it into account makes it possible to overcome the limitations of knowledge in modeling the physical phenomena that cause failures.

[0019] Very advantageously, state data as well as behavioral data are used to determine symptoms (first and second symptom within the meaning of the invention) which are treated in the same way by the module for determining component failure probabilities.

[0020] Moreover, and very advantageously, the determination of the probabilities of failure of the aircraft components uses a Bayesian network trained from a knowledge base comprising data produced by a natural language processing module configured to process aircraft maintenance reports.

[0021] Taking into account behavioral data to determine the symptoms of the aircraft and feedback in the Bayesian network makes it possible to improve the modeling of physical phenomena and consequently the modeling of degradation phenomena.

[0022] The system and method of determination thus make it possible to improve the diagnosis leading to the identification of the components which are the real causes of failure and consequently to reduce the direct and indirect costs associated with a bad diagnosis of fault finding during the ground maintenance of the aircraft.

[0023] The invention thus makes it possible to improve the detectability of failures of systems and their equipment, to put the aircraft back into service more quickly and therefore to reduce the time of unplanned operational unavailability.

[0024] In one embodiment, said at least one second symptom is a signature determined from one or more indicators positioned when the second determination module detects behavioral data representative of a trend of degradation in the operation of at least one component of the aircraft.

[0025] For example, a signature can be determined from an indicator representing a degradation of a first component and an indicator representing a degradation of a second component, these components possibly being included in different subsystems of the aircraft.

[0026] This embodiment advantageously allows for the modeling of symptoms resulting from interactions between aircraft subsystems, even when these interactions are difficult to understand or model. For example, mechanical vibrations from a rotating assembly can generate vibrations and degrade equipment in another system that is not directly related to the rotating system.

[0027] The invention also relates to a method and device for driving a Bayesian network that can be used in a system according to the invention.

[0028] Thus, and according to a second aspect, the invention also relates to a method for training a Bayesian network intended to be implemented by a module configured to determine the probabilities of failure of a plurality of aircraft components from at least one first and at least one second aircraft symptom, said Bayesian network being trained from a knowledge base comprising data produced by a natural language processing module configured to process aircraft maintenance reports, in which: - said at least one initial symptom of the aircraft is determined from underwing condition data collected; and - said less a second symptom of the aircraft is determined from data degradation representative of an abnormal variation in flight data.

[0029] Correspondingly, the invention also relates to a Bayesian network training device intended to be implemented by a module configured to determine the probabilities of failure of a plurality of aircraft components from at least one first and second aircraft symptom, said Bayesian network being trained from a knowledge base comprising data produced by a natural language processing module configured to process aircraft maintenance reports, wherein: - said at least one initial symptom of the aircraft is determined from underwing condition data collected; and - said less a second symptom of the aircraft is determined from degradation data representative of an abnormal variation in flight data.

[0030] In a particular embodiment, the natural language processing module used in the invention is a module of the training device.

[0031] In one embodiment, this natural language processing module takes as input texts extracted from aircraft maintenance reports and produces as output labels which constitute structured information used by an enrichment module to enrich the knowledge base.

[0032] In one embodiment of the invention, the knowledge base comprises records that associate symptoms with failures. The natural language processing module produces: (i) at least one initial label representing at least one piece of information from among flight data, status data, behavioral data and / or symptoms; and (ii) at least one second label representing a defect.

[0033] In one embodiment, the knowledge base enrichment module is configured to enrich the knowledge base with at least one record containing: (i) at least one symptom identified from a first label, associated with (ii) a defect identified from a second label.

[0034] In a particular embodiment, the different steps of the process for determining faulty components and the main steps of the process for training the Bayesian network are determined by computer program instructions.

[0035] Consequently, the invention also relates to a computer program on an information medium, this program being capable of being implemented in a computer, this program comprising instructions adapted to the implementation of the steps of a process as described above.

[0036] This program can use any programming language, and be under the form of source code, object code, or intermediate code between source code and object code, such as in a partially compiled form, or in any other desirable form.

[0037] The invention also relates to a computer-readable information carrier, comprising instructions for a computer program as mentioned above.

[0038] The information medium can be any entity or device capable of storing the program. For example, the medium can include a storage means, such as a ROM, for example a CD ROM or a microelectronic circuit ROM, or a magnetic recording means, for example a hard disk drive.

[0039] On the other hand, the information medium can be a transmissible medium such as an electrical or optical signal, which can be transmitted via an electrical or optical cable, by radio, or by other means. The program according to the invention can, in particular, be downloaded onto an Internet-type network.

[0040] Alternatively, the information carrier may be an integrated circuit in which the program is incorporated, the circuit being adapted to execute or to be used in the execution of the process in question. Brief description of the drawings

[0041] Other features and advantages of the present invention will become apparent from the description below, with reference to the accompanying drawings, which illustrate an example of an embodiment without being limiting in any way. In the figures:

[0042] [Fig-1] [Fig.1] represents a system for determining faulty components in accordance with the invention;

[0043] [Fig.2] [Fig.2] represents the hardware architecture of a terminal conforming to a particular embodiment of the invention;

[0044] [Fig.3] [Fig.3] represents the material architecture of a drive device conforming to a particular embodiment of the invention;

[0045] [Fig.4] [Fig.4] represents, in flowchart form, the main steps of a training method according to the invention; and

[0046] [Fig.5] [Fig.5] represents, in the form of a flowchart, the main steps of a method for determining faulty components according to the invention. Detailed description of implementation methods

[0047] Fig. 1 represents a digital SYS system for determining faulty components according to an embodiment of the invention.

[0048] This system includes a first MDSi module for determining initial symptoms.

[0049] When a problem or malfunction of the aircraft is suspected, an operator moves near the aircraft (or "under the wing") and performs a number of tests, to obtain DEk status data from the aircraft, and deduce the first S1 symptoms; using this first MDSi symptom determination module.

[0050] In practice, the first MDSi symptom determination module can be a module integrated into an operator's TRM terminal. In the embodiment of [Fig. 1], the TRM terminal is configured to interface with a BDT database of test scenarios, which includes tests T1, T2, T3 to be performed for a predetermined SDP fault detection scenario.

[0051] In the embodiment described here, each test T1, T2, T3 allows the determination (for example by calculation or by measurement) of one or more state data DEk.

[0052] The first symptoms S1; can be obtained from a BDS database that associates at least one first symptom S1; with a combination of at least one state data DEk. In the example of [Fig. 1]: - DEb DE2 state data are representative of a first symptom S'i; - DEb DE3 state data are representative of a first symptom S*2; and - DE2 state data are representative of a first symptom S*3.

[0053] DEk state data are for example the result of operator observations (odor, part inspection, ...), results of measurements carried out by the operator under the wing, or data obtained by the operator, for example data provided from an aircraft CALC computer to the TRM terminal via an input / output I / O interface.

[0054] In the described embodiment, the SYS system includes a second module MDS2 for determining symptoms S2;, from DCk behavior data representative of a tendency of degradation of the operation of at least one component of the aircraft.

[0055] In the embodiment described here, this second MDS2 module is a module of the aircraft's CALC computer.

[0056] In the embodiment described here, at least one symptom S2; determined by the second symptom determination module MDS2 is constituted by a SIG failure signature calculated from the DCk behavior data.

[0057] More specifically, in the embodiment described here, the second symptom determination module MDS2 comprises at least one MT processing module configured to process DV flight data made available by sensors or by electronic systems of the engines or aircraft and to position INDi indicators when they detect DCk behavior data representative of a trend of degradation in the operation of at least one component.

[0058] Such an MT processing module is configured for example to follow a trend in flight data over a given period or to detect a shape in a representation of this flight data.

[0059] For example, an MT processing module; positions a first indicator IND; when it detects a DCk data representative of an increase in pressure or temperature in an aircraft component, or more generally the exceeding of a normality threshold by an operating variable of an aircraft component.

[0060] For example, an MTj processing module positions a second indicator INDj when it detects DCk data representative of a particular vibrational phenomenon.

[0061] In the embodiment described here, the second symptom determination module MDS2 generates a SIG signature from these IND indicators; and this SIG signature is treated as an S2 symptom; in the same way as the S1 symptoms; determined by the first symptom determination module MDSi from the DEk state data.

[0062] For example, the second symptom determination MDS2 module includes LOG logic configured to generate: - a first SIG2i signature when the first indicator IND; is positioned but not the second indicator INDj; - a second SIG22 signature when the second INDj indicator is positioned but not the first IND indicator; ; - a third SIG23 signature when both indicators IND; and INDj are positioned.

[0063] The SYS system includes an MDPD module for determining the probability of failure from: - of the first S1 symptoms; determined from DEk status data; and - of the second symptoms S2; determined from the behavioral data DCk.

[0064] In the embodiment described here, the MDPD module for determining probability of failure uses a Bayesian network RB.

[0065] Depending on the presence or absence of first and / or second symptoms, the Bayesian network model RB dynamically updates the probabilities associated with each branch of a decision tree, and thus directs the diagnosis towards the identification of the root cause of the failure.

[0066] In the embodiment described here, the MDPD module for determining probability of failure is integrated into the operator's TRM terminal.

[0067] This Bayesian network RB is configured to determine, from the first and second symptoms, the probability of a Dj failure.

[0068] For example, a Dj failure can be: - a sensor malfunction; or - a torn harness; or - a software problem.

[0069] In the embodiment described here, this Bayesian network RB is trained from a knowledge base BC which includes records, each associating, to at least one symptom S, a failure D. This knowledge base BC is, for example, initialized from theoretical data from the FMEA (Failure Mode and Effects Analysis) of the different components.

[0070] This knowledge base BC can also be enriched by feedback from the use of the SYS digital system. For example, if it is detected that particular combinations of first symptoms S1; and second symptoms S2; determined by the CAL computer from behavioral data DCk are representative of a fault Dk, the knowledge base BC can be updated with this new information to retrain the Bayesian network RB.

[0071] In the embodiment described here, this knowledge base BC includes records from feedback determined by a natural language analysis of documents dealing with real causes of failure.

[0072] In the embodiment described here, the various available documents dealing with the actual causes of aircraft failure are analyzed by a Natural Language Processing (NLP) module. In the embodiment described here, the SYS system includes a DENT training module that incorporates the NLP module.

[0073] These documents include, for example, maintenance or troubleshooting reports provided by the airlines. In practice, for a given airline, the number of these reports can be very high, for example, several thousand per month. They are most often written in plain language: the operator describes a situation and details the corrective actions taken.

[0074] The invention advantageously allows these reports to be processed automatically to enrich the knowledge base BC of the Bayesian network RB.

[0075] In the embodiment described here, the MNLP natural language processing module is configured to receive unstructured DNS documents as input and to provide relevant structured information as output in the form of LABj labels used to feed the feedback data REX of the BC knowledge base used for training the Bayesian network RB.

[0076] In the embodiment described here, the MNLP natural language processing module uses an NLP learning model. As is known, the development of an NLP model essentially consists of: - build a learning base starting with data cleaning input and resulting in the labeling of input data; - implement the NLP model; and - validate the NLP model from a statistical and business perspective.

[0077] In the context of the invention, the input data of the MNLP module are maintenance report texts and the LABj labels are, for each report: - at least one first label representing at least one piece of information from among flight data, status data, behavior data and / or symptoms; and - at least one second label representing a defect.

[0078] In the embodiment described herein, the DENT training device of the SYS system includes a MEBC module for enriching the knowledge base with at least one record comprising: (i) at least one symptom identified from a first label, associated with: (ii) a defect identified from a second label.

[0079] In the embodiment described here, the MEBC knowledge base enrichment module uses: - the BDS symptom database to identify early symptoms from early LABj labels representative of state data; - the second MDS2 module for determining second symptoms to identify second symptoms from first LABj labels representing flight or behavioral data.

[0080] For example, with reference to [Fig. 1], if the MNLP natural language processing module produces a first label "DE2" (representing a state data point) and a second label "D5" (representing a failure) from the unstructured information contained in the maintenance reports, the MEBC knowledge base enrichment module: - determines from the third record in the BDS database that the DE2 status data is representative of an S3 symptom; and - enriches the BC knowledge base with a record that associates symptom S3 and fault D5.

[0081] The development of an NLP model includes, for example: - the construction of a training database which includes the cleaning of input data (management of punctuation, accents, synonyms and antonyms, removal of multiple spaces, transformation of uppercase letters into lowercase, ...); - the extraction of information (motifs or patterns) corresponding to regular expressions; - lemmatization of extracted information; - tokenization (or in English tokenization); - spelling correction; - removal of non-discriminatory words (in English, stop words).

[0082] The SYS system includes an MSAC module configured to signal at least one failing component among the plurality of aircraft components based on the probabilities of failure calculated by the MDPD module, and to propose an appropriate corrective action.

[0083] The operator can carry out a corrective action Q proposed by the digital system to be performed under the wing, for example the removal of a component to be sent to the workshop for analysis and repair

[0084] In the embodiment described here, the MSAC module for reporting faulty components is integrated into the operator's TRM terminal.

[0085] For example, if the failure probability determination module MDPD determines that the most probable failure is sensor wear, the MSAC module can signal to the operator that the sensor is faulty and propose a corrective action Ci consisting of replacing the sensor.

[0086] In one embodiment, the system for determining at least one faulty aircraft component is implemented in a TRM terminal.

[0087] In the embodiment described here, and as shown in [Fig.2], this TRM terminal has the hardware architecture of a computer. It comprises a processor 10A, a random access memory 10B, a read-only memory 10C, a non-volatile flash memory 10D, input / output means 10E, and communication means 10F.

[0088] These communication means 10F allow in particular the TRM 10 terminal to obtain DEk status data from the aircraft's CALC computer, or from externally configured measuring devices to perform under-wing tests.

[0089] In one embodiment these under-wing tests can be carried out by dedicated modules integrated into the TRM terminal.

[0090] In one embodiment, the second MDS2 module for determining at least one second symptom S2 is integrated into the TRM terminal. This module is configured to obtain DV flight data from the aircraft and to determine at least one SIG fault signature from DCk degradation data representative of an abnormal variation in this flight data.

[0091] The read-only memory 10C constitutes a recording medium according to the invention, readable by the processor 10A and on which is recorded a computer program PROGT according to the invention, comprising instructions for the execution of the steps of a method for determining at least one faulty aircraft component according to the invention and which will be described with reference to [Fig.5].

[0092] This computer program PROGT defines functional and software modules configured herein to implement the steps of the determination process according to the invention. These functional modules rely on or control the hardware components 10A to 10D of the TRM terminal. They include, in particular: - a first MDSi module for determining at least one initial symptom S1; of the aircraft from DEk state data collected under the wing; - the second MDS2 module for determining at least one second S2 symptom; of the aircraft; - an MDPD module for determining the probabilities of failure of a plurality of aircraft components from at least one first and second symptom, this module comprising a Bayesian network RB trained from a knowledge base BC comprising data produced by a natural language processing module configured to process aircraft maintenance reports, and - an MSAC module configured to report at least one failing component among said plurality of components according to said probabilities.

[0093] Figure 3 represents the hardware architecture of a DENT drive device according to the invention. This device has the hardware architecture of a computer and comprises a processor 20A, a random access memory 20B, a read-only memory 20C, a non-volatile flash memory 20D, input / output means 20E and communication means 20F.

[0094] The read-only memory 20C of the drive device constitutes a recording medium according to the invention, readable by the processor 20A and on which is recorded a computer program PROGE according to the invention, comprising instructions for the execution of the steps of a drive process according to the invention and which will be described with reference to [Fig.4].

[0095] This DENT device is configured to enrich the BC knowledge base used to train the Bayesian network RB used by the MDPD module to determine the probabilities of failure of a plurality of aircraft components.

[0096] The PROGE computer program defines functional and software modules configured herein to implement the steps of the drive process according to the invention. These functional modules rely on or control the hardware elements 20A to 20E of the DENT drive device.

[0097] They include in particular a natural language processing module MNLP configured to process aircraft maintenance reports DNS and produce relevant structured information IS which feeds the feedback data REX of the knowledge base BC used to train the Bayesian network RB.

[0098] In one embodiment of the invention, this structured information consists of: (i) initial labels representing information from flight data, status data, behavioral data and / or symptoms; and (ii) second labels representing failures.

[0099] In the embodiment described herein, the functional modules of the DENT device include an enrichment MEBC module configured to enrich the BC knowledge base with at least one record comprising: (i) at least one symptom identified from a first label; associated with (ii) a defect identified from a second label.

[0100] In the embodiment described here, the MEBC module determines at least some of these symptoms using the BDS symptom database to identify first symptoms from first LABj labels representing state data and / or the second MDS2 module for determining second symptom to identify second symptoms from first LABj labels representing flight or behavior data.

[0101] Fig. 4 represents the main steps of a training method according to the invention.

[0102] It includes a step E10 of analyzing aircraft maintenance reports by a natural language processing method to produce structured data in the form of LABj labels.

[0103] In the embodiment described here, these LABj labels represent: - information including flight data, status data, behavioral data, and / or symptoms; and - failures.

[0104] In the embodiment described here, the training process includes a step E20 to enrich the knowledge base BC of the Bayesian network RB from: - initial symptoms determined from LABj labels representing state data and the BDS symptom database; - second symptoms determined from LABj labels representing flight or behavior data and the second MDS2 module for determining second symptom; - LABj labels representing failures.

[0105] In the embodiment described here, the training method includes a step E30 of training the Bayesian network RB from this knowledge base so as to update the probabilities of this network to improve the diagnosis of root cause identification of failure thanks to this structured information obtained by NLP processing of the unstructured information contained in the maintenance and troubleshooting reports.

[0106] Figure 5 represents the main steps of a method for determining faulty components according to the invention.

[0107] It comprises: - a step F10 of determining at least one first symptom S1; of the aircraft from DEk state data collected under the wing; - a step F20 of determining at least a second symptom S2; of the aircraft from DCk degradation data representative of an abnormal variation in flight data; - a step F30 of determining probabilities of failure of a plurality of aircraft components from at least one first and second symptom, this determination step using a Bayesian network RB trained from a knowledge base BC comprising data produced by a natural language processing module configured to process aircraft maintenance reports, - a step F40 of reporting at least one failing component among said plurality of components according to said probabilities.

[0108] This signal allows the operator to take corrective action with a high probability that it will be appropriate to the actual cause of the aircraft failure.

Claims

Demands

1. A digital system for determining at least one faulty component of an aircraft, this system comprising: - a first module (MDSi) for determining at least one first symptom (S1;) of the aircraft from state data (DEk) collected under the wing; - a second module (MDS2) for determining at least one second symptom (S2i) of the aircraft from degradation data (DCk) representative of an abnormal variation in flight data;- a module (MDPD) for determining the probabilities of failure of a plurality of aircraft components, this module comprising a Bayesian network (RB) configured to perform said probability determination from said at least one first and second symptoms, said Bayesian network (RB) being trained from a knowledge base (BC) comprising data produced by a natural language processing module (MNLP) configured to process aircraft maintenance reports, - a module (MSAC) configured to report at least one failing component among said plurality of components according to said probabilities.

2. A determination system according to claim 1, characterized in that said natural language processing module (MNLP) takes as input texts extracted from said aircraft maintenance reports and produces as output labels (LABj) representing: - information from flight data, state data, behavior data and / or symptoms; and - failures.

3. A determination system according to claim 1 or 2, characterized in that said at least one second symptom (S2;) is a signature determined from at least one indicator (IND;) positioned when the second determination module (MDS2) detects behavioral data (DCk) representative of a tendency of degradation of the operation of at least one component of the aircraft.

4. A method for determining at least one defective component of an aircraft, this method being implemented by computer and comprising the following steps:

5.

6. - determination (F10) of at least one first symptom (S1;) of the aircraft from state data (DEk) collected under the wing; - determination (F20) of at least one second symptom (S2;) of the aircraft from degradation data (DCk) representative of an abnormal variation in flight data; - determination (F30) of probabilities of failure of a plurality of aircraft components from said at least one first and second symptoms, this determination using a Bayesian network (RB) trained from a knowledge base (BC) comprising data produced by a natural language processing module (MNLP) configured to process aircraft maintenance reports, - reporting (F40) of at least one failing component among said plurality of components according to said probabilities. A method for training a Bayesian network (BN) intended to be implemented by a module (MDPD) configured to determine failure probabilities of a plurality of aircraft components, said determination being performed by said Bayesian network (BN) from at least one first and at least one second symptom of the aircraft, said Bayesian network (BN) being trained (E30) from a knowledge base (KB) comprising data produced by a natural language processing module (NLP) configured to process (E10, E20) aircraft maintenance reports, wherein: - said at least one first symptom (S1;) of the aircraft is determined from underwing state data (DEk); and - said less a second symptom (S2;) of the aircraft is determined from degradation data (DCk) representative of an abnormal variation in flight data. Device (DENT) for training a Bayesian network (BN) intended to be implemented by a module (MDPD) configured to determine the probabilities of failure of a plurality of aircraft components from at least one first and at least one second aircraft symptom, said determination being performed by said Bayesian network (BN) from at least one first and at least one second aircraft symptom, said Bayesian network (BN) being trained from a knowledge base (KB) comprising data produced by a natural language processing module (NLP) configured to process aircraft maintenance reports, in which: - said at least one first symptom (S1;) of the aircraft is determined from state data (DEk) collected under the wing; and - said at least one second symptom (S2;) of the aircraft is determined from degradation data (DCk) representative of an abnormal variation in flight data.

7. Training device (DENT) according to claim 6 characterized in that said natural language processing module (MNLP) takes as input texts extracted from said aircraft maintenance reports and produces as output: (i) at least one first label (LABj) representing information from flight data, state data, behavior data and / or symptoms; and (ii) at least one second label (LABj) representing a failure, the training device (DENT) comprising a knowledge base (BC) enrichment module (MEBC) with: (i) at least one symptom (Sh S2i) identified from said first label, associated with: (ii) a failure identified from said second label.

8. Computer program (PROGT) comprising instructions for performing the steps of the method for determining at least one defective component of an aircraft according to claim 4 when said program is executed by a computer.

9. Computer program (PROGE) comprising instructions for carrying out the steps of the training process according to claim 5 when said program is executed by a computer.

10. Support (10C, 20C) comprising a computer program according to claim 8 or according to claim 9.

Images