Identifying a fault in a sensor-based air data system

Abstract

System (1) for identifying a fault in a sensor-based air data system (3) of an aircraft, wherein the system (1) is configured to acquire data from an inertial measurement unit of the aircraft and / or data on control variables of the aircraft, each with respect to a period of dynamic excitation of the aircraft, and to determine from these data expected sensor output values ​​of the air data system (3) and to compare them with the actual sensor output values ​​of the air data system (3) for the period of dynamic excitation, and to identify a fault in the air data system (3) when the deviation between the expected and the actual sensor output values ​​of the air data system (3) exceeds a predetermined condition, wherein the system (1) is further configured toa) to effect the dynamic excitation of the aircraft by generating a setpoint profile of at least one control variable and specifying the setpoint profile to an actuator system (5) of the aircraft, and to generate the setpoint profile depending on the respective air data system (3) to be tested for a fault by the dynamic excitation of the aircraft and depending on a flight area boundary determined from the current flight state of the aircraft,or b) to check the suitability of a dynamic excitation of the aircraft that has already occurred due to manual or automated flight control of the aircraft or an external disturbance according to predefined criteria and, if suitability is found, to subsequently select a data recording of the data of the inertial measuring unit and / or the data on the control variable during the period of dynamic excitation with the suitability and to subsequently determine the expected sensor output values ​​of the air data system (3), whereby the system can execute option a) as well as option b).

Description

The invention relates to a system for identifying a fault in a sensor-based air data system of an aircraft, and to an aircraft with such a system. The following information does not necessarily derive from a specific prior art document, but rather from expert considerations in the field of modern aviation, where typically at least a significant portion of manual and automatic flight control relies on aircraft sensor data: An aircraft's aeronautical data system usually comprises one or more sensors mounted on the aircraft to determine atmospheric properties and aerodynamic data of the oncoming airflow conditions. A pressure sensor, for example, measures static and total atmospheric pressure, while a temperature probe measures the total temperature. Based on the readings from such aeronautical data sensors, aircraft states such as airspeed, altitude, and Mach number can be calculated.Furthermore, wind vanes or pressure sensors can be used to determine the aircraft's current angle of attack and sideslip. The sensor output values ​​of a given aeronautical data system are essential for flight guidance and navigation, but can also serve as input for flight control and automated assistance systems, which, for example, warn and protect the pilot before leaving the flight envelope. Therefore, the sensor data must meet high standards of accuracy and reliability. A failure of an aeronautical data system or faulty measurements must be detected quickly and reliably. Current fault detection systems are typically based on a voting algorithm. Redundantly acquired sensor output values ​​are compared for consistency. Aeronautical data systems in use are typically designed with triple or quadruple redundancy.If the sensor output value of one air data system deviates from those of two other air data systems, the first air data system is identified as faulty. The failure of a single sensor within an air data system can be easily detected using this method. In this case, the sensor output values ​​of the remaining air data systems can be used to continue operating flight control and automatic assistance functions safely. Commercial aircraft typically travel long distances along flight paths where aircraft conditions, including altitude and speed, change very slowly. Once cruising altitude is reached, altitude and speed remain constant for an extended period. While the complete failure of one aerodynamic speed system and one aerodynamic angle of attack system can still be reliably detected by comparing the results with the two remaining redundant systems operating within that flight path, this method fails if two mutually redundant systems fail simultaneously, or if one system fails during a cruise flight with constant or very slowly changing aircraft conditions and maintains a constant sensor output value due to the error. This can occur, for example, if...This can occur when a pitot tube or an aerodynamic angle-of-attack vane ices up. An incorrect measurement and display of the aerodynamic speed can thus remain undetected for a long leg of the flight. If, later in the flight, a commanded change to the flight parameters (altitude, speed) is made, an unsafe flight situation can arise due to the unreliable sensor information. For example, the simultaneous icing and freezing of two aerodynamic angle-of-attack sensors has been a cause of crashes in the past. Possible known solutions to this problem, such as monitoring frequency content or individual sensor output values ​​for exceeding defined limits, as well as model-based comparison methods that use inertial sensor output values ​​and control variables with a dynamic model of the aircraft to reconstruct aerodynamic data, allow deviations from normal behavior to be detected. For this purpose, actual sensor data and sensor data reconstructed from models can be compared. Estimating sensor data from air data systems is known in the prior art, for example, from the publication: F. Adhika Pradipta Lie (2014): Synthetic Air Data Estimation. A case study of model-aided estimation. PhD thesis. University of Minnesota.Furthermore, approaches for detecting sensor faults are known in the prior art; see the publication: Isermann, Rolf (2004): Model-Based Fault Detection and Diagnosis - Status and Applications. In: IFAC Proceedings Volumes 37 (6), pp. 49-60. DOI: 10.1016 / S1474-6670(17)32149-3. However, these methods require sufficient dynamic excitation of the aircraft or dynamic changes in the sensor signals being monitored to detect faults quickly and reliably. Extended flight segments in cruise flight without significant changes in altitude and speed are unsuitable for such methods. Ground-based testing of air data systems is usually very complex, as special tools and measuring devices are required, for example, to adjust the aerodynamic angle-of-attack sensors. Static ambient pressure sensors, by their very nature, can only be subjected to test pressures under purely static conditions. Therefore, preventive testing of air data systems is only carried out through special test flights as part of maintenance work. Consequently, the gradual degradation of individual sensors in air data systems is hardly noticeable during operation and is only discovered after a complete failure. The publication: Rolf Isermann, “Model-Based Fault Detection and Diagnosis - Status and Applications”, IFAC Proceedings Volumes, Volume 37, Issue 6, 2004, Pages 49-60, ISSN 1474-6670, https: / / doi.org / 10.1016 / S1474-6670(17)32149-3. (https: / / www.sciencedirect.com / science / article / pii / S1474667017321493) concerns model-based methods for fault detection using input and output signals and the application of dynamic process models. These methods are based, for example, on parameter estimation, parity equations, or state observers. Signal modeling approaches are also mentioned. The aim of this publication is to generate several symptoms that indicate the difference between nominal and faulty status. Based on various symptoms, fault diagnosis procedures follow, which determine the fault by applying classification or inference methods.After an introduction, some applications for an actuator, a passenger car and an internal combustion engine will be shown. In the dissertation “Synthetic Air Data Estimation - A case study of model-aided estimation” by F. Adhika Pradipta Lie, submitted to the faculty of the graduate school of the university of minnesota, 2014, a method for estimating airspeed, angle of attack and sideslip angle without the use of conventional pitot-static air data systems is presented. In this context, DE 10 2017 102 923 A1 also relates to a method for determining an error of a barometric pressure measuring system arranged on board an aircraft, comprising the following steps: Determining a current position POSGNSS(t) and an altitude zGNSS(t) of the aircraft in a geodetic reference system at time t using a GNSS satellite navigation system arranged on board the aircraft;Determine a static pressure pAC(t) and / or a pressure height zAC(t) for the position POSGNSS(t) in provided weather analysis data ANDAT or in provided weather forecast data PROGDAT of a numerical weather forecast model NWP using the pressure measurement system; determine a geopotential height zAN / PROG(t) associated with the static pressure pAC(t); and / or determine a static pressure pAN / PROG(t) associated with the height zGNSS(t) for the position POSGNSS(t) in the provided weather analysis data ANDAT or in the provided weather forecast data PROGDAT of the numerical weather forecast model NWP; determine the height deviation Δz(t) = zGNSS(t) - zAN / PROG(t); and / or determine (104) the pressure deviation Δp(t) = pAC(t) - pAN / PROG(t).Determining an altitude deviation Δz* averaged over a period Δt from Δz(t) = zGNSS(t) - zAN / PROG(t) and / or determining a pressure deviation Δp* averaged over the period Δt from Δp(t) = pAZ(t) - pAN / PROG(t), whereby a fault in the pressure measuring system is considered detected if the average altitude deviation |Δz*| is greater than or equal to a predefined limit value G1, or if the average pressure deviation |Δp*| is greater than or equal to a predefined limit value G2; and generating a warning signal if a fault is detected. DE 10 2008 036 638 A1 further relates to a method for determining the functional state of a pressure sensor for dynamic pressure measurement in a rotor blade of a helicopter, in which the harmonic flow pressure signal correlated to the rotor speed is filtered out from a pressure sensor signal and a sensor state signal is determined from it taking into account state parameters of the rotor blade and / or helicopter.US 2020 / 0201312 A1 further concerns a method for sensor fault detection and identification using residual fault pattern recognition. Specifically, the method for detecting and identifying sensor faults in a vehicle involves acquiring data from sensors located on the vehicle. The method further includes conducting a majority vote on the data for each data type to generate a single voted value for each data type. The method also includes generating estimated values ​​for each data type using some of the voted values. Finally, the method includes generating residuals by comparing the estimated values ​​with the voted values.Furthermore, the procedure includes analyzing a pattern of the residuals to determine which types of data are faulty in order to detect and identify a fault occurring at least one of the sensors on the vehicle. US 2020 / 0319238 A1 concerns an aircraft sensor fault detection system. The system comprises a sensor system with a sensor device for measuring an environmental parameter and a measuring circuit coupled to the sensor device. The system also includes an AC power source connected to the sensor system. Furthermore, the system includes an AC current measuring system that measures the AC current flowing through the sensor system and indicates a fault when a threshold, based on a change in the impedance of the sensor device, is exceeded. The object of the invention is to monitor an aircraft's air data system in such a way that, in particular, sensor output values ​​that remain constant erroneously can be identified as a fault. The invention is defined by the features of the independent claims. Advantageous further developments and embodiments are the subject of the dependent claims. A first aspect of the invention relates to a system for identifying a fault in a sensor-based air data system of an aircraft, wherein the system is designed to acquire data from an inertial measuring unit of the aircraft and / or data on control variables of the aircraft, each with respect to a period of dynamic excitation of the aircraft, and to determine from these data expected sensor output values ​​of the air data system and to compare them with the actual sensor output values ​​of the air data system for the period of dynamic excitation, and to identify a fault in the air data system when the deviation between the expected and the actual sensor output values ​​of the air data system exceeds a predetermined condition, wherein the system is further designed toa) to effect the dynamic excitation of the aircraft by generating a setpoint profile of at least one control variable and specifying the setpoint profile to an actuator system of the aircraft, and to generate the setpoint profile depending on the respective air data system to be tested for a fault by the dynamic excitation of the aircraft and depending on a flight range boundary determined from the current flight state of the aircraft, or b) to check the suitability of a dynamic excitation of the aircraft that has already occurred due to manual or automated flight control of the aircraft or an external disturbance according to predefined criteria and, if suitability is found, to subsequently select a data recording of the data from the inertial measuring unit and / or the data about the control variable during the period of dynamic excitation with the suitability and to subsequently determine the expected sensor output values ​​of the air data system,where the system can execute option a) as well as option b). Preferably, the system includes an input unit designed to capture an input from the aircraft pilot and, upon receiving the input, to execute option a). In an alternative configuration, the system checks whether, after each recurring time interval of a predetermined length, option b) could be executed due to the suitability of the dynamic excitation, and if not, whether option a) could be executed. Accordingly, the system continuously checks for the regular execution of option b) in order to execute option a) in any time interval where option b) was not possible. During flight segments with many maneuvers, it can therefore be expected that option b) can be executed continuously, while during longer (quasi-)stationary cruise flight, option a) can be used. Options a) and b) can therefore also be executed alternately. The aircraft can be a fixed-wing aircraft or a rotary-wing aircraft such as a helicopter, or have another configuration, and can be manned or unmanned. Preferably, the aircraft is a commercial aircraft for cargo and / or passengers, since such commercial aircraft, in particular, experience (quasi-)stationary flight states over extended periods, such as long cruise flights, during which the flight conditions do not change or change only slowly, with the exception of position, if position is considered a flight condition. The aircraft has at least one sensor-based air data system, which includes, for example, an aerodynamic angle-of-attack sensor, an aerodynamic sideslip sensor, a pitot tube, or a barometric altimeter. All these air data systems have in common the determination of static or dynamic parameters of the air in the vicinity of the aircraft.For example, a barometric altimeter directly determines a static ambient air pressure, which can be corrected in a further function of the associated aeronautical data system for altitude measurement and can take into account nonlinearities, temperatures, humidity, etc.; a pitot tube also determines an air pressure, which, however, does not correspond to the static ambient air pressure, but has an increased pressure due to the airflow and the impact of the incoming air on a surface of the aircraft and, using known functions in the associated aeronautical data system for speed measurement, can provide an aerodynamic speed of the aircraft relative to the surrounding air. The term "sensor output values" encompasses not only direct measurements from a sensor of an air data system, but also processed sensor data, particularly filtered, estimated from multiple sensor readings, combined, and otherwise modified quantities obtained as results of the air data system. These are used, for example, for display in the cockpit or as actual values ​​in a computer-aided flight control system. Sensor output values ​​can also be obtained by applying an estimator, especially a Kalman filter. The sensor output values ​​are typically regenerated at rapid intervals (e.g., 50 Hz, i.e., every 0.02 seconds) to provide a continuously updated sequence of sensor output values ​​in a (quasi-)continuous signal from the respective air data system. Thus, time series are obtained for both the determined expected sensor output values ​​and the actual sensor output values ​​as obtained from the respective air data system under consideration, i.e., signals that can be compared. Such a comparison of these signals is preferably performed by integrating the respective magnitude of the difference values ​​for each of the time points or corresponding time points (in the case of discretely obtained actual sensor output values ​​and continuously determined expected sensor output values). The expected sensor output values ​​are therefore reconstructed values, in other words, estimated values ​​that can serve as a reference for comparison with the sensor output values ​​actually obtained from the respective air data system. Aircraft control variables, in the case of a fixed-wing aircraft, primarily concern aerodynamic control surfaces such as elevators, ailerons, rudders, and thrust lever positions. Which control variable is relevant for determining the expected sensor output values ​​over time during the dynamic excitation of the aircraft, as well as which control variable is considered for generating a target value profile and specifying this profile to an aircraft actuator system in case a), depends largely on the aeronautical data system used for testing to identify potential fault conditions.Accordingly, to test for a fault in a barometric altimeter, a corresponding change in the aircraft's altitude must be carried out; to test for a fault in an aerodynamic angle of attack sensor, a change in the aircraft's aerodynamic angle of attack must be carried out; to test for a fault in a pitot tube, a change in the aircraft's airspeed must be carried out, etc. In case a), the system generates, in particular, temporal sequences of values ​​for at least one control variable that produce a sufficiently large dynamic excitation of the aircraft to obtain amplitudes in the actual sensor output values ​​that are sufficiently above a signal-to-noise ratio to allow for the utilization of sufficient information in the actual sensor output values. The flight envelope limits of the aircraft are also taken into account to ensure that these limits are not exceeded. Flight envelope limits are advantageously determined depending on the current flight condition, since, for example, the aerodynamic angle of attack of a fixed-wing aircraft may have different limits at low speeds than at high speeds, and the maximum permissible speed in the aircraft's direction of flight depends on the air density and temperature, and thus on the altitude.Furthermore, load factors and maximum flight altitudes are typically limited in fixed-wing aircraft. Advantageously, the dynamic excitations caused by the respective control variable are kept so small that they have no negative impact on passenger comfort, cargo, or the aircraft structure. The sequence of values ​​for at least one control variable can be represented as a sinusoidal signal with a frequency that changes over time (for example, a so-called "frequency sweep" or "chirp" signal), or alternatively as rectangular signals, such as a "3211" signal. Other alternatives are possible and must be selected according to their amplitude and spectrum, depending on the aviation data system under consideration. A dynamic excitation of the aircraft, as described in case b), resulting from manual or automated flight control or an external disturbance, occurs, for example, when a human pilot commands the aircraft's moment dynamics by inputting the control stick, such as transitioning from a straight climb to a horizontal turn. Flight maneuvers can also be initiated by autopilots. External disturbances to the aircraft primarily concern air currents such as turbulence, which can also excite the aircraft's inherent dynamics. The suitability of such a dynamic excitation of the aircraft, which may have already occurred due to manual or automated flight control or an external disturbance, is determined in particular by characterizing the excitation, for example, amplitudes, durations, frequencies / spectrum of the excitation, power spectral density, or similar parameters. Instead of the excitation itself, suitability can be assessed using indirect parameters, such as continuously determined expected sensor output values ​​and their changes – especially if these changes exceed a certain threshold. The system is preferably configured to execute both option a) and option b). Which of the two options is executed depends, in the case of option b), on whether the required suitability is met, but can also be generally enabled or disabled via a corresponding setting, for example, by the pilot. Option a) is preferably executed automatically either when a potential error in the air data system is suspected and / or upon input from the pilot. Options a) and b) can therefore, in principle, be executed independently of each other, but can also be combined. An advantageous aspect of the invention is that an aircraft's air data system is monitored in such a way that, in particular, sensor output values ​​that erroneously remain constant during flight can be identified as a fault by a system on board the aircraft. This is achieved, in particular, by the system either using existing dynamic excitations of the aircraft, after verifying their suitability, to determine the dynamic excitations as suitable reference values ​​for calculating the expected sensor output values, or by the system itself proactively generating suitable dynamic excitations. It is therefore advantageously ensured that appropriate dynamic excitations of the aircraft are present, which also advantageously take into account the aircraft's flight envelope limits.Furthermore, the system advantageously generates dynamic excitations in such a way that they are barely perceptible or not at all to the aircraft's passengers, or are not recognized as such, and are ideally interpreted as natural aerodynamic disturbances such as turbulence. The system can be used in all avionics systems for both commercial and military aircraft with one or more air data systems. With the future introduction of single-crew cockpit concepts, the system can play a crucial role in fault detection and diagnosis, but it also offers a significant increase in safety for conventional two-crew cockpits. The automated detection of maneuvers for verification and fault detection enhances the reliability of the air data system.Automated triggering of test maneuvers and subsequent automated detection and diagnosis of errors relieves the pilot and leads to greater safety in situations where the reliability of the determined sensor output values ​​is uncertain. According to an advantageous embodiment, the system is designed to determine the expected sensor output values ​​of the air data system from the data of the inertial measurement unit and / or from the data on the control variable by at least one of the following methods: application of a dynamic model of the aircraft, wherein the dynamic model comprises a plurality of flight states of the aircraft and takes into account current and previous values ​​of the control variables, and in particular, data from the inertial measurement unit and / or data on the control variable are respective input data of the dynamic model; application of an estimator, in particular a Kalman filter; application of a purely kinematic relationship; application of a machine learning model with parameters detached from intuitively comprehensible relationships of flight physics and trained and abstracted by a machine learning process, in particular an artificial neural network;The purely kinematic relationship describes, for example, the integration of an aircraft's rate of rotation with respect to an angle. The dynamic model is capable of determining the expected sensor output values ​​of the respective air data system based on current and past values ​​of control variables as well as data from the inertial measurement unit. The dynamic model differs from an algebraic model in that it is based on the concept of differential equations, i.e., it determines the state of such a dynamic aircraft depending on current and past system input variables as well as the current state of the aircraft. According to a further advantageous embodiment, the system for identifying a fault case is designed to use one of the following sensor-based air data systems: Aerodynamic angle of attack sensor, aerodynamic sideslip sensor, aerodynamic dynamic pressure sensor, barometric altimeter. According to a further advantageous embodiment, the system is designed to determine continuously expected sensor output values ​​in the case of application of b) and, to test the suitability of the dynamic excitation, to check the expected sensor output values ​​for continuously exceeding a minimum amplitude above a predetermined limit value or the average amplitude of the expected sensor output values ​​for exceeding the predetermined limit value and only to determine the required suitability if the predetermined limit value is exceeded. Taking into account the amplitude of the expected sensor output values ​​is due to the fact that a non-zero noise component is to be expected in actual sensor output values, which implies that a minimum amplitude should be present in the actual and expected course of the sensor output values ​​in order to ensure a favorable signal-to-noise ratio. According to a further advantageous embodiment, the system is designed to determine a current probability of a fault in the air data system during a flight of the aircraft, and if the probability exceeds a predetermined threshold, to generate a dynamic excitation according to case a) and to check for the presence of a fault by comparing the expected sensor output values ​​with the actual sensor output values ​​of the air data system during the dynamic excitation. According to a further advantageous embodiment, the system is designed to determine the current probability of a fault in the air data system based on dynamic excitations of the aircraft due to manual or automated flight control of the aircraft or due to an external disturbance, the respective expected sensor output values ​​of the air data system, and to continuously compare these with actual sensor output values ​​of the air data system, and to determine the probability of a fault based on the comparison. Advantageously, the system can also apply options a) and b) in combination by executing option b) if a probability of a fault is exceeded, which can then be further investigated by executing option a). The system is advantageously configured to continuously check the suitability of dynamic excitations of the aircraft according to the procedure in option b), to determine a probability of a fault by first comparing the expected and actual sensor output values ​​according to option b), and then to execute option a) if the probability exceeds the predefined limit. According to a further advantageous embodiment, the system is designed to determine the expected sensor output values ​​of the air data system and to compare them with the actual sensor output values ​​of the air data system over the period of dynamic excitation by determining new parameters of a sensor model of a sensor of the air data system from the expected sensor output values ​​and comparing them for agreement with previously stored parameters of the sensor model. According to another advantageous embodiment, the system is designed to recalibrate the sensor of the air data system with the new parameters of the sensor model. According to a further advantageous embodiment, the system is designed to check the suitability of the already occurring dynamic excitation of the aircraft itself in the case of application of b) by analyzing the data of the inertial measuring unit and / or the data on the control variables, and / or by analyzing a satellite-determined altitude change of the aircraft in the case of a barometric altimeter as an air data system to be tested. Another aspect of the invention relates to an aircraft with a system as described above and below. Advantages and preferred further developments of the proposed aircraft result from an analogous and substantive transfer of the above statements made in connection with the proposed system. Further advantages, features, and details will become apparent from the following description, in which – possibly with reference to the drawing – at least one embodiment is described in detail. Identical, similar, and / or functionally equivalent parts are identified by the same reference numerals. It shows: Fig. 1 : An aircraft with a system for identifying a fault in a sensor-based air data system of the aircraft according to an embodiment of the invention. The representations in the figure are schematic and not to scale. Fig. 1 shows a fixed-wing aircraft equipped with a system 1 for identifying a fault in a sensor-based air data system 3 of the aircraft. The system 1 sequentially examines an angle-of-attack sensor and a dynamic pressure sensor of the aircraft as the respective air data system 3 to be tested for a fault. With respect to the angle-of-attack sensor, the system 1 determines the current airspeed of the aircraft, as well as pitch rates and / or pitch angles of the aircraft, obtained by an inertial measurement unit of the aircraft, and data on the control input profiles of the aircraft's elevator, each with reference to a period of time during which the aircraft is dynamically excited.From this, system 1, implemented in a computing unit of the aircraft, determines the expected sensor output values ​​for this period, i.e., expected aerodynamic angles of attack, in order to compare these with the actual sensor output values ​​of the air data system 3, i.e., the actual aerodynamic angles of attack obtained from a measurement for the period of dynamic excitation. If the integral of the absolute differences between the actual and expected angles of attack exceeds a predefined error threshold, an error in the air data system 3 is identified.System 1 performs an analogous procedure for the dynamic pressure sensor as a further air data system 3 of the aircraft to be tested. This involves determining expected sensor output values ​​in the form of velocity by integrating accelerations from the aircraft's inertial measurement unit, projected onto the aircraft's kinematic velocity vector, and from changes in the throttle lever position, based on changes in the aircraft's velocity. This is then compared with the velocity change result of the dynamic pressure sensor. If the deviation exceeds a further error threshold, a fault in this dynamic pressure air data system 3 is identified.These dynamic excitations of the aircraft (changes in the aerodynamic angle of attack and changes in airspeed) are, in the case of the aerodynamic angle of attack, implemented by system 1 itself through corresponding inputs to the actuators of actuator system 5 associated with the elevator, as well as through a change in the thrust lever position for the thrust control of the engines, which is another part of actuator system 5. Alternatively, corresponding excitations generated by normal flight operations are checked for sufficiently high amplitudes, thus enabling the identification of a fault condition imperceptibly through normal flight operations. Although the invention has been further illustrated and explained in detail by means of preferred embodiments, the invention is not limited by the disclosed examples, and other variations can be derived from them by a person skilled in the art without departing from the scope of protection of the invention. It is therefore clear that a multitude of possible variations exist. It is also clear that the embodiments mentioned as examples are truly only examples and are not to be understood in any way as limiting, for example, the scope of protection, the possible applications, or the configuration of the invention.Rather, the preceding description and the description of the figures enable the person skilled in the art to implement the exemplary embodiments in concrete terms, whereby the person skilled in the art, with knowledge of the disclosed inventive concept, can make various changes, for example with regard to the function or the arrangement of individual elements mentioned in an exemplary embodiment, without leaving the scope of protection defined by the claims and their legal equivalents, such as further explanations in the description. Reference symbol list 1 System 3 Air data system 5 Actuator system

Claims

System (1) for identifying a fault in a sensor-based air data system (3) of an aircraft, wherein the system (1) is configured to acquire data from an inertial measurement unit of the aircraft and / or data on control variables of the aircraft, each with respect to a period of dynamic excitation of the aircraft, and to determine from these data expected sensor output values ​​of the air data system (3) and to compare them with the actual sensor output values ​​of the air data system (3) for the period of dynamic excitation, and to identify a fault in the air data system (3) when the deviation between the expected and the actual sensor output values ​​of the air data system (3) exceeds a predetermined condition, wherein the system (1) is further configured toa) to effect the dynamic excitation of the aircraft by generating a setpoint profile of at least one control variable and specifying the setpoint profile to an actuator system (5) of the aircraft, and to generate the setpoint profile depending on the respective air data system (3) to be tested for a fault by the dynamic excitation of the aircraft and depending on a flight area boundary determined from the current flight state of the aircraft,or b) to check the suitability of a dynamic excitation of the aircraft that has already occurred due to manual or automated flight control of the aircraft or an external disturbance according to predefined criteria and, if suitability is found, to subsequently select a data recording of the data of the inertial measuring unit and / or the data on the control variable during the period of dynamic excitation with the suitability and to subsequently determine the expected sensor output values ​​of the air data system (3), whereby the system can execute option a) as well as option b). System (1) according to claim 1, wherein the system (3) is configured to determine the expected sensor output values ​​of the air data system (3) from the data of the inertial measuring unit and / or from the data on the control variable by at least one of the following methods: application of a dynamic model of the aircraft, wherein the dynamic model comprises a plurality of flight states of the aircraft and takes into account current and previous values ​​of the control variables; application of an estimator, in particular a Kalman filter; application of a purely kinematic relationship; application of a machine learning model with parameters detached from intuitively comprehensible relationships of flight physics and learned and abstracted by a machine learning process; System (1) according to one of the preceding claims, wherein the system (1) is designed to identify a fault case of one of the following sensor-based air data systems (3): Aerodynamic angle of attack sensor, aerodynamic sideslip sensor, aerodynamic dynamic pressure sensor, barometric altimeter. System (1) according to one of the preceding claims, wherein the system (1) is designed to determine continuously expected sensor output values ​​in the case of application of b) and to check the suitability of the dynamic excitation for continuously exceeding a minimum amplitude above a predetermined limit value or the average amplitude of the expected sensor output values ​​for exceeding the predetermined limit value and only then determining the required suitability if the predetermined limit value is exceeded. System (1) according to one of the preceding claims, wherein the system (1) is configured to determine a current probability of a fault in the air data system (3) during a flight of the aircraft, and if the probability exceeds a predetermined threshold, to generate a dynamic excitation according to case a) and to check for the presence of a fault by comparing the expected sensor output values ​​with the actual sensor output values ​​of the air data system (3) during the dynamic excitation. System (1) according to claim 5, wherein the system (1) is configured to determine the current probability of a fault in the air data system (3) based on dynamic excitations of the aircraft due to manual or automated flight control of the aircraft or due to an external disturbance, the respective expected sensor output values ​​of the air data system (3), and to continuously compare these with actual sensor output values ​​of the air data system (3), and to determine the probability of a fault on the basis of the comparison. System (1) according to one of the preceding claims, wherein the system (1) is configured to determine the expected sensor output values ​​of the air data system (3) and to compare them with the actual sensor output values ​​of the air data system (3) over the period of dynamic excitation by determining new parameters of a sensor model of a sensor of the air data system (3) from the expected sensor output values ​​and comparing them for agreement with previously stored parameters of the sensor model. System (1) according to claim 7, wherein the system (1) is configured to recalibrate the sensor of the air data system (3) with the new parameters of the sensor model. System (1) according to one of the preceding claims, wherein the system (1) is designed to check the suitability of the already occurring dynamic excitation of the aircraft itself by analyzing the data of the inertial measuring unit and / or the data on the control variables, and / or by analyzing a satellite-based determined altitude change of the aircraft in the case of a barometric altimeter as an air data system (3) to be tested. Aircraft with a system (1) according to any of the preceding claims.

Images