Method and device for detecting aircraft fault, flight control computer device, aircraft, computer program product and storage medium

Abstract

The invention relates to a method and a device for detecting and locating a failure of an input sensor (10a) of a flight control computer (1000) of an aircraft, a flight control computer device, an aircraft, a computer program product and a storage medium. The method comprises the step of comparing first information representative of the probability of occurrence of the fault transmitted by the neural network with second information representative of the probability of occurrence of the fault transmitted by a classifier trained to perform detection and localization of the fault. Thus, fault diagnosis and fault location of input sensors useful for determining flight control can be advantageously optimized.

Description

Technical Field

[0001] The present invention relates to a method for detecting faults in a system of an aircraft. The invention more precisely relates to a method for detecting and locating faults in sensors connected to the inputs of a flight control computer of an aircraft. Background Art

[0002] Current aircraft include many onboard devices and systems, including flight control computers. The flight control computer of an aircraft implements the control of the actuators of the aircraft, such as the guidance mechanisms of the aircraft (as examples, ailerons, elevators, steering devices). The flight commands issued by such a computer are defined based on control set points but also based on information provided by sensors associated with the control elements. For example, the flight control computer implements the control of the guidance mechanism actuators taking into account the detected position of the guidance mechanism, which is obtained by the position sensor of the guidance mechanism. In order to ensure the operational safety of the aircraft, the flight control computer implements the monitoring of various sensors to detect and signal potential failures of these sensors. In view of the fact that the flight control computer must exhibit a very high level of safety and a very high level of availability in its operation, the computing power of the flight control computer processor is usually dedicated to the control of the aircraft. Therefore, the monitoring of sensors by the flight control computer is limited to the detection and signaling of faults, without including the location or identification of the detected faults. Therefore, the maintenance personnel must perform a search for faults, for example, the fault may be related to the sensor itself or to the wiring between the sensor and the flight control computer. This search may be even longer since the sensors used are generally passive sensors which are not able to transmit information about potential faults.It is desirable to be able to identify and effectively locate faults in order to optimize the implementation of maintenance operations and to reduce the downtime of the aircraft under maintenance. Summary of the invention

[0003] It is an object of the present invention to provide an improved method for detecting and locating a fault in a sensor providing a notification signal to a flight control computer of an aircraft.

[0004] To this end, a method for detecting and locating a fault in an input sensor of a flight control computer of an aircraft is provided, the method being performed by a module for detecting and locating a fault, the module comprising an electronic circuit configured to perform the steps of the method, the method comprising:

[0005] - conditioning at least a first signal transmitted by the flight control computer to the sensor according to a predetermined format, and conditioning at least a second signal transmitted by the sensor to the computer,

[0006] - a step of analyzing the first conditioned signal and the second conditioned signal by means of a neural network trained to perform detection and localization of a fault using the first signal representing the conditioned signal transmitted by the flight control computer or the second signal representing the conditioned signal transmitted to the flight control computer,

[0007] - a step of comparing first information representative of the probability of occurrence of a fault delivered by the neural network with second information representative of the probability of occurrence of a fault delivered by a classifier trained to perform detection and localization of the fault using a signal representative of the conditioned signal, then,

[0008] - Using said first information and said second information to provide third information representing the probability of occurrence of a fault.

[0009] Therefore, fault diagnosis or fault localization estimates of input sensors useful in determining flight control may be advantageously optimized.

[0010] Furthermore, in the case where the module comprising electronic circuits for detecting and locating faults is independent of the main processor or processors of the flight control computer, the processing power of these processors remains dedicated to determining the flight control signals and the performance and safety of the flight control are accordingly enhanced.

[0011] The method for detecting and locating faults according to the present invention may also include the following features considered individually or in combination:

[0012] - The method further comprises comparing the third information with at least one predetermined threshold value.

[0013] The method further comprises filtering the third information and providing fourth information representative of the probability of occurrence of a fault based on the filtered third information.

[0014] Another subject of the invention is a device for assisting in the detection and localization of a fault in an input sensor of a flight control computer of an aircraft, the device comprising a module for detecting and localizing a fault, the module comprising an electronic circuit configured to perform the following steps:

[0015] - conditioning at least a first signal transmitted by the flight control computer to the sensor according to a predetermined format, and conditioning at least a second signal transmitted by the sensor to the computer,

[0016] - a step of analyzing the first and second regulated signals by means of a neural network trained to perform detection and localization of faults based on the first and second signals representative of the regulated signals transmitted by or to the flight control computer,

[0017] - a step of comparing first information representative of the probability of occurrence of a fault delivered by the neural network with second information representative of the probability of occurrence of a fault delivered by a classifier trained to perform detection and localization of the fault based on a signal representative of the regulated signal, then,

[0018] - providing at least third information representing the probability of occurrence of a fault based on the first information and the second information.

[0019] The device for detecting and locating faults according to the present invention may also include the following features considered individually or in combination:

[0020] - The device further comprises an electronic circuit configured to perform a comparison of the third information with at least one predetermined threshold value.

[0021] The device further comprises an electronic circuit configured to perform filtering of the third information and to provide fourth information representative of the probability of occurrence of a fault based on the filtered third information.

[0022] Another subject of the invention is a flight control computer device comprising a first main processor configured to determine flight control signals and a device for the detection and location of auxiliary faults such as previously described, the device comprising a module for detecting and locating faults, the module comprising an electronic circuit that is independent of the first processor.

[0023] Another subject of the invention is an aircraft comprising a device for assisting the detection and localization of faults or a flight control computer such as previously described.

[0024] The invention finally relates to a computer program product comprising program code instructions for executing the steps of a method for detecting and locating faults such as previously described, and a storage medium comprising a computer program product such as previously described. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] The foregoing and other features of the present invention will become more apparent upon reading the following description of an exemplary embodiment, which is presented with respect to the accompanying drawings:

[0026] Figure 1 An aircraft comprising a device for assisting the detection and location of a failure of an input sensor of a flight control computer according to one embodiment of the invention is shown;

[0027] Figure 2 Schematically shows a device for assisting in the detection and location of a fault of an input sensor of a flight control computer according to one embodiment of the present invention;

[0028] Figure 3Schematically shows the Figure 2 A method for assisting the detection and location of a fault in an input sensor of a flight control computer, performed by the device for assisting and locating a fault as shown in the above;

[0029] Figure 4 Schematically shows Figure 2 The internal architecture of the device for assisting and locating faults has been shown;

[0030] Figure 5 Schematically shows that Figure 2 Internal details of the device for assisting and locating faults shown in , which includes a post-processing unit; and

[0031] Figure 6 Schematically shows the device for assisting and locating faults already in Figure 5 Details of the post-processing unit are shown in FIG. DETAILED DESCRIPTION

[0032] Figure 1 Schematically and symbolically, a device for detecting and locating a fault in an input sensor of a flight control computer of an auxiliary aircraft 1 is shown ( Figure 1 An aircraft 1 (not shown in detail) is provided on board of the aircraft 1. Advantageously, such a device on board the aircraft 1 allows optimizing the detection and localization of a given fault of a sensor connected to a flight control computer of the aircraft 1. More generally, the device 10 for assisting in the detection and localization of faults is configured to detect several types of faults and a plurality of faults, as well as to localize at least approximately (or more precisely) each of the detectable and detected faults of all sensors connected to one or more inputs of a flight control computer of the aircraft 1.

[0033] In the following part of the description, the connection link carrying the signal, the signal carried by the connection link, and the information encoded by the signal carried by the connection link are named indiscriminately. For example, the reference numeral "P" represents both the electrical signal P and the connection link P carrying the signal P, just as the information P transmitted (or encoded) by the signal P on the connection link carrying the signal P.

[0034] Figure 2 Schematically and symbolically, a flight control computer 1000 of an aircraft 1 is shown, the input of which is connected to sensors, such as position sensors for the guidance mechanism of the aircraft 1, and the output of which is connected to actuators, such as actuators of the guidance mechanism of the aircraft 1. For the sake of simplicity, Figure 2A single sensor 10a is shown in FIG. 1 , however, a plurality of sensors of various characteristics may be connected to the input of the flight control computer 1000. For the same reasons intended to simplify the description of the invention, Figure 2 A single control signal 1001 for a flight control actuator is shown in , however many actuators of various characteristics may be connected to the output of the flight control computer 1000 .

[0035] According to the described example, the sensor 10a is connected to the flight control computer 1000 by a first connection carrying the first signal 10b transmitted by the flight control computer 1000 to the sensor 10a, and by a second connection carrying the second signal 10c transmitted by the sensor 10a to the flight control computer 1000, when the first signal 10b has a predetermined form. The signal 10b is, for example, a signal for the power supply of the sensor 10a, or an electrical excitation signal for the nominal operation of the sensor 10a. For example, such an excitation signal allows, for example, a moving element to be powered, the movement of which causes a variation in the output signal of the sensor 10a, the output signal being, for example, the signal 10c. It goes without saying that, depending on the characteristics and type of the sensor, additional connections may be present, and several signals of one or more types may be transmitted by the computer 1000 to the sensor 10a, and vice versa.

[0036] Advantageously, the flight control computer 1000 comprises a device 10 for assisting the detection and localization of faults, in particular for the detection and localization of faults of the sensor 10a. To this end, the first signal 10b and the second signal 10c are connected to the device 10 for assisting the detection and localization of faults inside the flight control computer 1000. In other words, the connection between the flight control computer 1000 and the sensor 10a passes through the device 10 for assisting the detection and localization of faults. At the input of the flight control computer 1000, all connections between the flight control computer 1000 and the various sensors connected thereto are also substantially made in the same way. Thus, given that the device 10 for assisting the detection and localization of faults is included in the flight control computer 1000, its operation does not require the installation of specific wiring. On the one hand, this allows to maintain operational safety; on the other hand, this even allows such a device 10 for assisting the detection and localization of faults in the flight control computer of an existing aircraft to be implemented by simply replacing the flight control computer. The flight control computer 1000 also comprises a flight command control unit 100, also referred to herein as the "main processor 100" of the flight control computer. The main processor 100 is configured to determine the state of the flight control outputs to the flight control actuators based on the set point inputs and information and signals from various sensors connected to the flight control computer. The set point signals applied to the flight control computer may come from a set point device (e.g., a flight control stick or yoke in the cockpit of the aircraft 1), and / or from a flight assistance device (e.g., a system for correcting flight control or an automated navigation system).

[0037] According to one embodiment, the device 10 for assisting the detection and location of faults transmits to the main processor 100 input information 12 in the form of a signal representing the state of a signal 10c provided by a sensor 10a to an input of a flight control computer 1000, and information 11 (also in the form of a signal) representing the probability of a fault of a predetermined type of sensor 10a at a predetermined location of an onboard system of the aircraft 1. According to one embodiment, the information 11 carried by the coded signal is composite digital information, in other words, it comprises a plurality of fields each representing an information element relating to the probability of a fault associated with the type of fault of the sensor 10a. For example, an information field encodes the type of fault from a list of predetermined types of faults, another information field encodes the fault from a list of predetermined faults, and yet another field encodes the area or location of the fault within the aircraft system from a list of predetermined areas or locations. According to one embodiment, the processor 100 generates an information signal 1002 at at least one of its output terminals based on the internal signal 12, which information signal 1002 is intended to notify a third-party system of the presence of a fault on a flight control sensor, such as an alarm device inside the aircraft 1 or a ground station configured to monitor the status of the aircraft 1 and potential maintenance operations to be performed.

[0038] According to a variant embodiment, the information 11 is an analog or digital signal associated with two predetermined states, one of which encodes the detection of a fault and the other of which encodes the absence of a fault. According to this example, the same is true for the signal 1002 .

[0039] Ingeniously and advantageously, the device 10 for assisting in the detection and localization of faults in a flight control computer 1000 comprises a neural network NN and a classifier CLASS, as well as a regulation module PRE preconfigured to perform normalization of signals sent from the flight control computer 1000 to the sensors connected thereto and from these sensors to the flight control computer 1000.

[0040] Advantageously, the neural network NN is configured to be able to provide the probability of occurrence of a failure of a sensor, in particular among the sensors connected to the input of the flight control computer 1000, to supply information representative of the probability of occurrence of a failure of the sensor 10a. The failure of the sensor 10a may be a failure of the power supply of the sensor 10a, due to a failure in the connection carrying the signal 10b or in the connection at the output of the sensor 10a, in other words a failure of the link carrying the signal 10c, or a failure inside the sensor 10a, for example a failure of an internal transducer or any other internal element or component of the sensor 10a. In order to be able to provide information on the probability of occurrence of a failure, the neural network NN is pre-trained using predetermined signals corresponding to a plurality of scenarios or configurations related to normal operating states and to abnormal states, i.e. fault states. This corresponds to supervised learning, for example. According to one embodiment, the neural network NN is trained for use in a system of an aircraft before it is incorporated into the system. According to a variant, the neural network NN is used to detect and locate failures during specific periods of flight and ground taxiing of the aircraft 1, while still being used in a learning mode during other periods of flight and taxiing of the aircraft 1.

[0041] The operation for normalizing the signals applied at the input of the neural network NN allows the consistency of the format of these signals with the signals previously applied during the learning (especially supervised learning) phase of the neural network NN, before the neural network NN is used on board the aircraft. The classifier CLASS of the device 10 for assisting the detection and localization of faults of the flight control computer 1000 is implemented in a "post-processing" unit (or module) POST, which allows obtaining second information on the probability of failure of the sensor 10a and other sensors connected to the input of the flight control computer 1000 (not shown in the figure), using the same signal as that applied to the input of the neural network NN.

[0042] Advantageously and cleverly, it is then possible to verify the consistency of two pieces of information representative of the probability of a fault obtained by two different ways, on the one hand a neural network such as the FCNN (“Fully Connected Neural Network”) type and a classifier connected to a database of information about faults, which then provide third information about the probability of occurrence of a fault of the sensor connected to the input of the flight control computer 1000. This allows obtaining the desired performance of the device 10 for assisting in the detection and localization of faults using a less complex neural network than in the case of using a neural network alone (without a classifier). This further allows simplifying the training of the neural network, for example by using a more limited training data set, the classifier allowing mitigating unlikely scenarios for which the neural network has not been trained.

[0043] It should be noted that the structure of the one or more learning phases of the neural network NN and of the classifier CLASS associated with the database DB is such that the information about the probability of a fault includes or is associated with information about the localization of the fault. For example, the scenarios reserved for simulating or reproducing a fault of an input sensor of the flight control computer 1000, or a fault of a connecting circuit useful for the operation of the sensor, include the reproduction or simulation of faults of the same type at different locations. For example, in the case of a leakage current, the fault that causes this leakage current appears at several locations of the same equipotential link to obtain various configurations of variations of other signals related to the same sensor as the sensor simulating the fault.

[0044] According to one embodiment, the device 10 for detecting and locating faults of auxiliary sensors is independent of the main processor 100 in the flight control computer 1000. In other words, this means that, apart from the signals transmitted by the device 10 to the main processor 100, the two elements do not share their respective computing power to perform processing operations. In other words again, the detection and / or localization of faults is performed by the device 10 and is notified to the main processor 100, which can take these detection and / or localization into account in the determination of flight control, but the computing power of the main processor 100 is not available to the method for detecting and locating faults of input sensors of the flight control computer 1000, thereby optimizing safety in the determination of flight control of the aircraft 1.

[0045] Figure 3 Flow chart showing the steps of a method for providing information representing the probability of occurrence of a failure of a sensor 10a connected to the input of a flight control computer 1000 using two further intermediate information representing the probability of occurrence of a failure of the sensor 10a, the further two intermediate information being obtained respectively with the aid of a network of the neural network NN type of the device 10 for assisting in the detection and localization of failures and a classifier CLASS of the same device 10 for assisting in the detection and localization of failures.

[0046] The initial step S0 is a step for initializing all the systems described, at the end of which all the systems of the aircraft are correctly initialized, configured and capable of operating in a nominal manner. During a step S1, the device 10 for assisting the detection and localization of faults performs a conditioning, also called normalization or ensuring consistency regarding the signal of the sensor 10a, whether this is a signal coming from the sensor 10a or a signal going to the sensor 10a. The signal thus conditioned is applied, on the one hand, to the input of the neural network NN and, on the other hand, to the input of the classifier CLASS. Subsequently, during a step S2, the neural network NN determines and provides first information P1 representative of the probability of a fault, based on an analysis of the signals 10b and 10c conditioned and applied to the input of the neural network NN.

[0047] The signals 10b and 10c are also applied to the input of the classifier class, and the second information P2 representative of the probability of failure is determined by the classifier CLASS connected to its database DB using the signals 10b and 10c conditioned and then applied to the input of the classifier CLASS, so as to compare the obtained first information P1 representative of the probability of failure and the second information P2 representative of the probability of failure in step S3. Then, during step S3, third information P3 representative of the probability of occurrence of a failure of the sensor 10a (this third information comes from the comparison between the first information P1 representative of the probability of failure and the second information P2 representative of the probability of failure) is determined to implement the classification step, and then this third information is provided at the output of the device for assisting the detection and location of failures during step S4.

[0048] According to one embodiment, the information signal 11 transmitted by the device 10 for assisting the detection and location of faults is a signal carrying or encoding information P3. According to one embodiment, an additional step for comparing the information P3 with one or more thresholds T is performed to achieve an improved decision step for detecting a fault of the sensor 10a. According to one embodiment, a final time filtering is performed at the output of the device for assisting the detection and location of faults to eliminate accidental transients and to confirm over time the state of the final information representative of the probability of occurrence of a fault provided at the output of the device 10.

[0049] Figure 5Schematically shown are the internal details of a device 10 for assisting the detection and localization of faults, connected to a sensor 10a by a connection link carrying signals 10b and 10c. The pre-processing module PRE is configured to perform conditioning or normalization of the signals 10b and 10c and to deliver at its output as many normalized signals as applied to its input. The set of output signals applied to the input of the neural network NN is called I and is also applied to the input of a post-processing module POST which performs a processing operation after determining the information P1, namely on Figure 3 The operations performed in step S3 and step S4 of the described method. The information signals 11 and 12 delivered at the output of the device 10 come from the post-processing module POST.

[0050] Figure 6 The internal electronic circuit of the post-processing module POST and the optional comparator circuit C2 and the time filtering circuit F (which is also optional) are schematically and symbolically shown. The internal electronic circuit is configured to perform the aforementioned steps S3 and S4. In steps S3 and S4, the information P1 and P2 are compared by the comparator circuit C1 to generate the information signal P3, the comparator circuit C2 is used to compare the information P3 with one or more thresholds T, and the time filtering circuit F is used to perform time filtering of the output signal where appropriate.

[0051] Thus, the information P1 corresponds to the probability of occurrence of a fault of the sensor 10a supplied, for example, by the neural network NN, the classifier CLASS determines the statistical attribution P2 of faults belonging to a certain fault type from predefined fault types stored in a database DB of predefined faults, and the comparator circuit C2 and the time filtering circuit F define, jointly or individually, the final decision module when they are implemented.

[0052] Figure 4 An example of the internal architecture of the device 10 for assisting in the detection and location of faults is schematically shown. As an illustration, Figure 4 The internal arrangement of a device 10 for assisting in the detection and location of faults, such as on an aircraft 1, is shown. It should be noted that Figure 4 An example of the hardware architecture of the fault detection and location module within the apparatus 10 for assisting in the detection and location of faults may also be schematically shown.

[0053] according to Figure 4As an example of the hardware architecture shown, the device 10 for detecting and locating faults thus comprises the following components connected via a communication bus 10-9: a processor or CPU (central processing unit) 10-1; a volatile memory RAM (random access memory) 10-2; a non-volatile memory ROM (read-only memory) 10-3; a storage unit, such as a hard disk (or a storage medium reader, such as an SD (Secure Digital) card reader) 10-4; a communication interface module 10-5, which allows the system of the device 10 for assisting in the detection and location of faults to communicate with remote devices (for example, other onboard systems of the aircraft 1, including elements related to the classifier CLASS and the neural network NN used).

[0054] The processor 10-1 of the apparatus 10 for assisting the detection and location of a fault can execute instructions loaded into the RAM 10-2 from the ROM 10-3, from an external memory (not shown), from a storage medium (e.g., an SD card), or from a communication network. When the apparatus 10 for assisting the detection and location of a fault is powered on, the processor 10-1 can read instructions from the RAM 10-2 and execute the instructions. These instructions form a computer program that enables the processor 10-1 to implement the instructions related to the detection and location of a fault. Figure 3 All or part of the described method for assisting in the detection and location of faults, or a variation of the described method.

[0055] about Figure 4 All or part of the described method or its described variants may be implemented in software by executing an instruction set by a programmable machine such as a DSP (digital signal processor) or a microcontroller, or in hardware by a machine or a dedicated component such as an FPGA (field programmable gate array) or an ASIC (application-specific integrated circuit). Typically, the device 10 for assisting the detection and location of a fault of a sensor comprises an electronic circuit configured to implement the described method in relation thereto. It goes without saying that the device 10 for assisting the detection and location of a fault also comprises all elements typically present in a system comprising a control unit and its peripherals, such as a power supply circuit, a power supply supervision circuit, one or more clock circuits, a reset circuit, an input-output port, an interrupt input, a bus driver, this list being non-exhaustive.

[0056] According to one embodiment, the signal processed by the neural network corresponds to an analog signal. According to another embodiment, the signal processed by the neural network corresponds to a digital signal. According to yet another embodiment, the signal processed by the neural network corresponds to a combination of an analog signal and a digital signal.

Claims

1. A method for detecting and locating a fault in an input sensor (10a) of a flight control computer (1000) of an aircraft (1), the method being performed by a module (10) for detecting and locating a fault, the module (10) comprising an electronic circuit configured to perform the steps of the method, the method comprising: - conditioning (S1) at least a first signal (10b) transmitted by the flight control computer (1000) to the sensor (10a) and conditioning at least a second signal (10c) transmitted by the sensor (10a) to the flight control computer (1000) according to a predetermined format, - analyzing (S2) the conditioned first signal (10b) and the conditioned second signal (10c) by means of a neural network, the neural network being trained to perform detection and localization of the fault using the first signal (10b) representing the conditioned signal transmitted by the flight control computer (1000) and the second signal (10c) representing the conditioned signal transmitted to the flight control computer (1000), - comparing (S3) first information representative of the probability of occurrence of said fault delivered by said neural network with second information representative of the probability of occurrence of said fault delivered by a classifier, said classifier being trained to perform detection and localization of said fault using a signal representative of the conditioned signal (10b, 10c), and - providing (S4) at least third information (11) representing the probability of occurrence of said fault based on said first information and said second information.

2. The method for detecting and locating faults according to claim 1, further comprising comparing the third information (11) with at least one predetermined threshold value (T).

3. The method for detecting and locating faults according to claim 1 or 2, further comprising filtering the third information (S4a), and providing fourth information representing the probability of occurrence of the fault based on the filtered third information.

4. A device (10) for assisting in the detection and location of a fault in an input sensor of a flight control computer of an aircraft, the device comprising a module for detecting and locating a fault, the module comprising an electronic circuit configured to: - conditioning at least a first signal transmitted by the flight control computer to the sensor and conditioning at least a second signal transmitted by the sensor to the flight control computer according to a predetermined format, - analyzing the conditioned first signal and the conditioned second signal by means of a neural network trained to perform detection and localization of the fault using the first signal representative of the conditioned signal transmitted by the flight control computer and the second signal representative of the conditioned signal transmitted to the flight control computer, - comparing first information representative of the probability of occurrence of said fault delivered by said neural network with second information representative of the probability of occurrence of said fault delivered by a classifier trained to perform detection and localization of said fault using a signal representative of the conditioned signal, and - providing at least third information representative of the probability of occurrence of said fault based on said first information and said second information.

5. The device (10) for assisting the detection and location of faults according to claim 4, further comprising an electronic circuit configured to perform filtering of the third information and to provide fourth information representing the probability of occurrence of the fault based on the filtered third information.

6. A flight control computer device (1000), comprising a first main processor (100) configured to determine flight control signals, and a device (10) for detecting and locating auxiliary faults according to claim 4 or 5, wherein the device (10) comprises a module for detecting and locating faults, the module comprising the electronic circuit, and the electronic circuit is independent of the first main processor.

7. An aircraft comprising a device for assisting the detection and location of faults according to claim 4 or 5, or a flight control computer device according to claim 6.

8. A computer program product comprising program code instructions which, when executed by a processor of an apparatus for assisting the detection and location of faults, perform the steps of the method according to one of claims 1 to 3.

9. A storage medium comprising the computer program product according to claim 8.

Images