A method for reconstructing and fusing an angle-of-attack signal

Abstract

The application relates to a method for reconstructing and fusing an angle-of-attack signal, which comprises the following steps: reconstructing an inertial angle-of-attack estimation value according to first inertial data; performing transient fault judgment and steady-state fault judgment by using an effective measured angle-of-attack signal and the inertial angle-of-attack estimation value, screening a non-fault measured angle-of-attack signal according to the transient fault judgment result and the steady-state fault judgment result, and performing primary monitoring voting by using the non-fault measured angle-of-attack signal to obtain a primary voting state and a primary voting value; if the primary voting state is a non-singular state, the primary voting value is taken as a first angle-of-attack voting output value; if the primary voting state is a singular state, secondary monitoring voting is performed by using the inertial angle-of-attack estimation value and the non-fault measured angle-of-attack signal to obtain a secondary voting state and a second angle-of-attack voting output value; and a precise angle-of-attack fusion signal is reconstructed according to any one of the first angle-of-attack voting output value and the second angle-of-attack voting output value and an inertial angle-of-attack rate signal, so as to be used by an aircraft.

Description

Technical Field

[0001] This application relates to the field of flight control technology, and in particular to a method for reconstructing and fusing angle-of-attack signals. Background Technology

[0002] Angle of attack (AOA) signals are crucial for aircraft flight safety across the entire flight envelope, as well as for control stabilization, envelope protection, and stall warning. Their reliability significantly impacts aircraft safety control. However, due to variations in installation location and operating principles, AOA sensors are highly susceptible to interference from aircraft maneuvers and atmospheric wind fields. For instance, encountering turbulent winds and wind shear can reduce the availability of AOA signals and degrade control quality. Furthermore, icing or jamming of AOA sensors can also lead to erroneous AOA signals, thus significantly impacting aircraft safety.

[0003] In existing technologies, estimated angle-of-attack signals are typically used to replace those with low availability for aircraft use. However, these estimated angle-of-attack signals often contain errors, which cannot guarantee aircraft flight safety.

[0004] Therefore, it is necessary to improve one or more of the problems existing in the above-mentioned related technical solutions.

[0005] It should be noted that the information disclosed in the background section above is only used to enhance the understanding of the background of this application, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0006] The purpose of this application is to provide a method for reconstructing and fusing angle-of-attack signals, which can reconstruct accurate angle-of-attack fused signals for use by aircraft, effectively ensuring flight safety.

[0007] According to an embodiment of this application, a method for reconstructing and fusing angle-of-attack signals is provided, including:

[0008] The estimated inertial angle of attack is reconstructed based on the first inertial data, which includes the aircraft's pitch attitude, roll attitude, track tilt attitude, and sideslip angle.

[0009] Acquire effective measured angle of attack signals from multiple angle of attack sensors, and use the effective measured angle of attack signals and the inertial angle of attack estimation value to perform transient fault judgment, thereby obtaining the transient fault judgment result of each angle of attack sensor;

[0010] The inertial angle of attack estimate and the effective measured angle of attack signal are filtered, and the filtered effective measured angle of attack signal and the inertial angle of attack estimate are used to determine steady-state faults, so as to obtain the steady-state fault determination results of each angle of attack sensor.

[0011] Based on the transient fault judgment result and the steady-state fault judgment result, non-faulty measured angle of attack signals are selected from the valid measured angle of attack signals of the plurality of angle of attack sensors;

[0012] The initial monitoring vote is performed using the non-faulty measured angle of attack signal to obtain the initial voting status and initial voting value;

[0013] If the initial voting state is a non-singular state, then the initial voting value is used as the first angle of attack voting output value, which is determined based on the non-fault measured angle of attack signal.

[0014] If the initial voting state is a singular state, a secondary monitoring vote is performed using the inertial angle of attack estimate and the non-faulty measured angle of attack signal to obtain the secondary voting state and the second angle of attack voting output value; wherein, the second angle of attack voting output value is determined based on the inertial angle of attack estimate, or the second angle of attack voting output value is determined based on the inertial angle of attack estimate and the non-faulty measured angle of attack signal;

[0015] The angle-of-attack fusion signal is reconstructed based on either the first angle-of-attack voting output value or the second angle-of-attack voting output value, as well as the inertial angle-of-attack rate of change signal.

[0016] In one embodiment of this application, the estimated inertial angle of attack is expressed as:

[0017] (1)

[0018] in, This represents the estimated angle of attack due to inertia. , , , , Indicates a tilting or bending posture. Indicates rolling posture, Indicates the tilt attitude of the flight path. Indicates the sideslip angle.

[0019] In one embodiment of this application, the step of using the effective measured angle of attack signal and the inertial angle of attack estimate to perform transient fault judgment and obtain the transient fault judgment results of each angle of attack sensor includes:

[0020] The effective measured angle of attack signal is subtracted from the inertial angle of attack estimate to calculate the first deviation value;

[0021] The first deviation value is compared with the preset transient fault judgment threshold:

[0022] If the first deviation value is greater than the preset transient fault judgment threshold, then the angle of attack sensor is determined to have a transient fault.

[0023] Otherwise, it is determined that the angle of attack sensor does not have a transient fault.

[0024] In one embodiment of this application, the step of using the filtered effective measured angle of attack signal and the inertial angle of attack estimate to perform steady-state fault judgment and obtain the steady-state fault judgment results of each angle of attack sensor includes:

[0025] The second deviation value is calculated by subtracting the filtered effective measured angle of attack signal from the filtered inertial angle of attack estimate.

[0026] The second deviation value is compared with the preset steady-state fault judgment threshold:

[0027] If the second deviation value is greater than the preset steady-state fault judgment threshold and continues for a first preset duration, then the angle of attack sensor is determined to have a steady-state fault.

[0028] Otherwise, it is determined that the angle of attack sensor does not have a steady-state fault.

[0029] In one embodiment of this application, the non-fault-prone measured angle of attack signal is an effective measured angle of attack signal from an angle of attack sensor that meets a preset screening condition, wherein the preset screening condition is that neither the transient fault nor the steady-state fault exists simultaneously.

[0030] In one embodiment of this application, when there are at least two non-faulty angle-of-attack measurement signals, the step of performing initial monitoring voting using the non-faulty angle-of-attack measurement signals to obtain the initial voting status and initial voting value includes:

[0031] The first absolute value of the difference is calculated based on the non-faulty measured angle of attack signal, and the first absolute value of the difference is processed based on a preset monitoring threshold to obtain the first current monitoring matrix;

[0032] Based on the first current monitoring matrix and the first preset conditions, each of the non-fault measurement angle of attack signals is filtered to obtain the non-fault measurement angle of attack signals used for the initial voting;

[0033] If the number of non-faulty angle-of-attack measurement signals used for the initial vote is odd, then the median of the non-faulty angle-of-attack measurement signals used for the initial vote is taken as the initial vote value, and the initial vote state is set to the non-singular state.

[0034] If the number of non-faulty angle-of-attack measurement signals used for the initial voting is even, the initial voting status and the initial voting value are determined based on the first current monitoring matrix and the second preset conditions.

[0035] In one embodiment of this application, determining the initial voting state and the initial voting value based on the first current monitoring matrix and the second preset conditions includes:

[0036] If the number of non-faulty angle-of-attack measurement signals used for the initial voting is even and the first current monitoring matrix satisfies the second preset condition, then the initial voting state is set to the singular state and the initial voting value is set to an invalid value.

[0037] If the number of non-faulty angle-of-attack measurement signals used for the initial vote is even and the first current monitoring matrix does not meet the second preset condition, then the median of the non-faulty angle-of-attack measurement signals used for the initial vote is taken as the initial vote value, and the initial vote state is set to the non-singular state.

[0038] In one embodiment of this application, in the step of filtering each of the non-faulty angle-of-attack measurement signals according to the first current monitoring matrix and the first preset conditions, if the number of non-faulty angle-of-attack measurement signals used for the initial voting is zero, then the initial voting state is set to a singular state and the initial voting value is set to an invalid value.

[0039] In one embodiment of this application, if the initial voting state is a singular state, then performing a secondary monitoring vote using the inertial angle of attack estimate and the non-faulty measured angle of attack signal to obtain the secondary voting state and the second angle of attack voting output value includes:

[0040] The second absolute value of the difference is calculated based on the estimated inertial angle of attack and the non-faulty measured angle of attack signal, and the second absolute value of the difference is processed based on the preset monitoring threshold to obtain the second current monitoring matrix;

[0041] Based on the second current monitoring matrix and the first preset condition, the inertial angle of attack estimate and the non-fault measured angle of attack signal are filtered to obtain the angle of attack signal for secondary voting;

[0042] If the number of angle-of-attack signals used for the second vote is greater than zero and is odd, then the median of the angle-of-attack signals used for the second vote is taken as the second angle-of-attack voting output value, and the second vote state is set to the non-singular state.

[0043] If the number of angle-of-attack signals used for secondary voting is greater than zero and is even, then the secondary voting state and the second angle-of-attack voting output value are determined based on the second current monitoring matrix and the second preset condition.

[0044] If the number of angle-of-attack signals used for the second vote is zero, then the estimated inertial angle of attack is used as the second angle-of-attack voting output value, and the second vote state is set to the singular state.

[0045] In one embodiment of this application, the inertial angle of attack rate of change signal is represented as:

[0046] (2)

[0047] in, This represents the rate of change of inertial angle of attack. Indicates pitch angular velocity, This indicates the first angle of attack voting output value or the second angle of attack voting output value. Indicates the sideslip angle. Indicates a tilting or bending posture. Indicates rolling posture, Indicates forward overload. Indicates normal overload, Represents gravitational acceleration. Indicates the roll rate. Indicates yaw rate, Indicates the vacuum velocity of the aircraft;

[0048] The angle-of-attack fusion signal is represented as:

[0049] (3)

[0050] in, Indicates angle-of-attack fusion signal, This indicates the angle of attack deviation value. Indicates the second-order filter frequency. This indicates second-order filter damping. Indicates time.

[0051] The technical solutions provided by the embodiments of this application may include the following beneficial effects:

[0052] In the embodiments of this application, a high-precision inertial angle of attack estimate is first reconstructed using first inertial data. This inertial angle of attack estimate is applicable to static estimation of angle of attack signals under various flight scenarios. Using the reconstructed inertial angle of attack estimate and the acquired effective measured angle of attack signal, transient and steady-state fault judgments are performed on the angle of attack sensor to filter out non-faulty measured angle of attack signals for monitoring and voting. This ensures that faults are isolated from subsequent monitoring and voting to the greatest extent possible. Then, the selected non-faulty measured angle of attack signals are used to perform initial monitoring and voting. This ensures that when the initial voting state is non-singular, a first angle of attack voting output value that conforms to the actual operating conditions of the aircraft and has high reliability is obtained. Furthermore, when the initial monitoring and voting is singular and cannot be effectively voted, the reconstructed inertial angle of attack estimate is introduced, and a second monitoring and voting is performed using the inertial angle of attack estimate and the non-faulty measured angle of attack signal. This ensures that even when the initial monitoring and voting is singular, a second angle of attack voting output value that conforms to the actual operating conditions of the aircraft and has high reliability can still be output through the second monitoring and voting. Finally, the reliable angle-of-attack voting output value obtained from the monitoring vote is fused with the inertial angle-of-attack rate of change signal to obtain an accurate angle-of-attack fused signal for the aircraft to use, effectively ensuring the aircraft's flight safety. Attached Figure Description

[0053] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application. It is obvious that the drawings described below are merely some embodiments of this application, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort.

[0054] Figure 1 A flowchart illustrating the steps of a method for reconstructing and fusing angle-of-attack signals in an exemplary embodiment of this application is shown.

[0055] Figure 2 A simulation illustration of Embodiment 1 in the exemplary embodiments of this application is shown. Figure 1 ;

[0056] Figure 3 A simulation illustration of Embodiment 1 in the exemplary embodiments of this application is shown. Figure 2 ;

[0057] Figure 4 A simulation illustration of Embodiment 1 in the exemplary embodiments of this application is shown. Figure 3 ;

[0058] Figure 5 A simulation illustration of Embodiment 2 in the exemplary embodiments of this application is shown. Figure 1 ;

[0059] Figure 6A simulation illustration of Embodiment 2 in the exemplary embodiments of this application is shown. Figure 2 ;

[0060] Figure 7 A simulation illustration of Embodiment 2 in the exemplary embodiments of this application is shown. Figure 3 . Detailed Implementation

[0061] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, they are provided to make this application more comprehensive and complete, and to fully convey the concept of the exemplary embodiments to those skilled in the art. The described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.

[0062] Furthermore, the accompanying drawings are merely illustrative diagrams of embodiments of this application and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities.

[0063] This example implementation provides a method for reconstructing and fusing angle-of-attack signals, referencing... Figure 1 As shown, the method may include steps S101 to S108.

[0064] Step S101: Reconstruct the estimated inertial angle of attack based on the first inertial data, which includes the aircraft's pitch attitude, roll attitude, track tilt attitude, and sideslip angle.

[0065] Step S102: Obtain the effective measured angle of attack signals from multiple angle of attack sensors, and use the effective measured angle of attack signals and the inertial angle of attack estimation value to perform transient fault judgment, and obtain the transient fault judgment results of each angle of attack sensor.

[0066] Step S103: Filter the inertial angle of attack estimate and the measured angle of attack signal, and use the filtered effective measured angle of attack signal and the inertial angle of attack estimate to determine steady-state faults and obtain the steady-state fault determination results of each angle of attack sensor.

[0067] Step S104: Based on the transient fault judgment results and the steady-state fault judgment results, select the non-faulty measured angle of attack signals from the valid measured angle of attack signals of multiple angle of attack sensors.

[0068] Step S105: Perform initial monitoring voting using the non-faulty angle of attack measurement signal to obtain the initial voting status and initial voting value.

[0069] Step S106: If the initial voting state is a non-singular state, the initial voting value is used as the first angle of attack voting output value, which is determined based on the non-fault measurement angle of attack signal.

[0070] Step S107: If the initial voting state is a singular state, then a secondary monitoring vote is performed using the inertial angle of attack estimate and the non-fault measurement angle of attack signal to obtain the secondary voting state and the second angle of attack voting output value; wherein, the second angle of attack voting output value is determined based on the inertial angle of attack estimate, or the second angle of attack voting output value is determined based on the inertial angle of attack estimate and the non-fault measurement angle of attack signal.

[0071] It should be noted that if the initial voting state is singular, the obtained initial voting value will be set to an invalid value, for example, the initial voting value will be set to 0.

[0072] Step S108: Reconstruct the angle-of-attack fusion signal based on either the first angle-of-attack voting output value or the second angle-of-attack voting output value, and the inertial angle-of-attack change rate signal.

[0073] It should be noted that in step S108, it is necessary to determine whether to use the first angle-of-attack voting output value or the second angle-of-attack voting output value to reconstruct the angle-of-attack fusion signal based on the result of the initial monitoring vote. For example, if the initial voting state is a non-singular state, the angle-of-attack fusion signal is reconstructed using the first angle-of-attack voting output value and the inertial angle-of-attack rate of change signal; if the initial voting state is a singular state, the angle-of-attack fusion signal is reconstructed using the second angle-of-attack voting output value and the inertial angle-of-attack rate of change signal.

[0074] In the embodiments of this application, a high-precision inertial angle of attack estimate is first reconstructed using first inertial data. This inertial angle of attack estimate is applicable to static estimation of angle of attack signals under various flight scenarios. Using the reconstructed inertial angle of attack estimate and the acquired effective measured angle of attack signal, transient and steady-state fault judgments are performed on the angle of attack sensor to filter out non-faulty measured angle of attack signals for monitoring and voting. This ensures that faults are isolated from subsequent monitoring and voting to the greatest extent possible. Then, the selected non-faulty measured angle of attack signals are used to perform initial monitoring and voting. This ensures that when the initial voting state is non-singular, a first angle of attack voting output value that conforms to the actual operating conditions of the aircraft and has high reliability is obtained. Furthermore, when the initial monitoring and voting is singular and cannot be effectively voted, the reconstructed inertial angle of attack estimate is introduced, and a second monitoring and voting is performed using the inertial angle of attack estimate and the non-faulty measured angle of attack signal. This ensures that even when the initial monitoring and voting is singular, a second angle of attack voting output value that conforms to the actual operating conditions of the aircraft and has high reliability can still be output through the second monitoring and voting. Finally, the reliable angle-of-attack voting output value obtained from the monitoring vote is fused with the inertial angle-of-attack rate of change signal to obtain an accurate angle-of-attack fused signal for the aircraft to use, effectively ensuring the aircraft's flight safety.

[0075] The steps of the above-described angle-of-attack signal reconstruction and fusion method in this example embodiment will now be explained in more detail.

[0076] In one embodiment, in step S101, the estimated angle of attack is expressed as:

[0077] (1)

[0078] in, This represents the estimated angle of attack due to inertia. , , , , Indicates a tilting or bending posture. Indicates rolling posture, Indicates the tilt attitude of the flight path. Indicates the sideslip angle.

[0079] The above calculation of the inertial angle of attack estimate using formula (1) only requires numerical calculation based on the relationship between the track tilt attitude, pitch attitude, roll attitude and sideslip angle, without the need for matrix calculation. This not only reduces the amount of calculation but also provides high steady-state accuracy of the angle of attack estimate.

[0080] It should be noted that, in calculating the estimated inertial angle of attack using formula (1), the pitch attitude needs to be taken into account. Sideslip angle and rolling posture A [-80°, 80°] amplitude limiting process is applied to prevent division by zero. During aircraft flight, sideslip angle is taken into account. The value is relatively small, and its impact on angle of attack estimation is negligible. Therefore, if the aircraft cannot obtain the sideslip angle... The sideslip angle in formula (1) can be... Set to 0°.

[0081] It should also be noted that when calculating the estimated inertial angle of attack using formula (1), the following conditions must be met: the first inertial data is valid and the aircraft is in the air.

[0082] In one embodiment, before acquiring the effective angle-of-attack signals from the multiple angle-of-attack sensors in step S102, the following steps are further included:

[0083] The system acquires the measured angle of attack signals from multiple angle of attack sensors. Based on the fault reports from the multiple angle of attack sensors and the aircraft bus, it determines whether the measured angle of attack signals from the multiple angle of attack sensors are valid, in order to obtain the valid measured angle of attack signals from the multiple angle of attack sensors.

[0084] It should be noted that if the aircraft bus transmission verification is fault-free and a fault-free angle-of-attack sensor exists, then the measured angle-of-attack signal from the fault-free angle-of-attack sensor will be considered a valid measured angle-of-attack signal.

[0085] It should also be noted that in the event of multiple angle-of-attack sensors malfunctioning or a fault in the aircraft bus transmission verification, the angle-of-attack fused signal can be directly reconstructed using the inertial angle-of-attack estimate and the inertial angle-of-attack rate of change signal. For example, the inertial angle-of-attack estimate can be reconstructed using formula (1); and the value in formula (2) can be used to reconstruct the angle-of-attack fused signal. The value is taken as the estimated value of the inertial angle of attack. The inertial angle of attack change rate signal is calculated using formula (2). The calculated inertial angle of attack change rate signal is substituted into formula (3), and the angle of attack fusion signal is further determined using formula (3).

[0086] In one embodiment, step S102, which uses the effective measured angle of attack signal and the inertial angle of attack estimate to determine transient faults and obtain the transient fault determination results of each angle of attack sensor, includes the following steps S1021 and S1022.

[0087] Step S1021: Subtract the effective measured angle of attack signal from the inertial angle of attack estimate and calculate the first deviation value.

[0088] Step S1022: Compare the first deviation value with the preset transient fault judgment threshold:

[0089] If the first deviation value is greater than the preset transient fault judgment threshold, then the angle of attack sensor is determined to have a transient fault; otherwise, the angle of attack sensor is determined not to have a transient fault.

[0090] For example, based on the effective measured angle of attack signals obtained from multiple angle of attack sensors, in step S1021, the effective measured angle of attack signals from the multiple angle of attack sensors are subtracted from the estimated inertial angle of attack value to obtain the first deviation value corresponding to the multiple angle of attack sensors. Correspondingly, in step S1022, the multiple first deviation values ​​obtained in step S1021 are compared with a preset transient fault judgment threshold to realize transient fault judgment of the multiple angle of attack sensors. Specifically, angle of attack sensors with a first deviation value greater than the preset transient fault judgment threshold are considered to have transient faults, and angle of attack sensors with a first deviation value less than or equal to the preset transient fault judgment threshold are considered to not have transient faults.

[0091] It should be noted that the transient fault flag of an angle-of-attack sensor with a transient fault is set to TRUE, while the transient fault flag of an angle-of-attack sensor without a transient fault is set to FALSE. TRUE indicates true, and FALSE indicates false.

[0092] In one embodiment, step S103, which uses the filtered effective measured angle of attack signal and the inertial angle of attack estimate to determine steady-state faults and obtain the steady-state fault determination results of each angle of attack sensor, includes the following steps S1031 and S1032.

[0093] Step S1031: Subtract the filtered effective measured angle of attack signal from the filtered inertial angle of attack estimate to calculate the second deviation value.

[0094] Step S1032: Compare the second deviation value with the preset steady-state fault judgment threshold:

[0095] If the second deviation value is greater than the preset steady-state fault judgment threshold and continues for a first preset duration, then the angle of attack sensor is determined to have a steady-state fault; otherwise, the angle of attack sensor is determined not to have a steady-state fault.

[0096] For example, based on the effective measured angle of attack signals from multiple angle-of-attack sensors, in step S1031, the filtered effective measured angle-of-attack signals from the multiple angle-of-attack sensors are subtracted from the filtered inertial angle-of-attack estimate to obtain the second deviation values ​​corresponding to the multiple angle-of-attack sensors. Correspondingly, in step S1032, the multiple second deviation values ​​obtained in step S1031 are compared with a preset steady-state fault judgment threshold to perform steady-state fault judgment on the multiple angle-of-attack sensors. Specifically, angle-of-attack sensors with second deviation values ​​greater than the preset steady-state fault judgment threshold are considered to have a steady-state fault, while angle-of-attack sensors with first deviation values ​​less than or equal to the preset steady-state fault judgment threshold are considered to not have a steady-state fault.

[0097] It should be noted that the steady-state fault flag of an angle-of-attack sensor with a steady-state fault is set to TRUE, while the steady-state fault flag of an angle-of-attack sensor without a steady-state fault is set to FALSE.

[0098] In one embodiment, the non-faulty measured angle of attack signal in step S104 is a valid measured angle of attack signal of an angle of attack sensor that meets a preset screening condition, which is that there are no transient faults and steady-state faults at the same time.

[0099] By selecting valid angle-of-attack signals from angle-of-attack sensors that meet preset screening criteria as non-faulty angle-of-attack signals, the interference of faults on subsequent monitoring and voting can be isolated to the greatest extent.

[0100] Specifically, in step S104, based on the transient fault judgment result and the steady-state fault judgment result, non-faulty angle-of-attack signals are selected from the valid angle-of-attack signals of multiple angle-of-attack sensors. The implementation method can be as follows: performing an OR operation on the transient fault flag bit and the steady-state fault flag bit to obtain the final fault flag of the angle-of-attack sensor; and selecting non-faulty angle-of-attack signals based on the final fault flag of the angle-of-attack sensor.

[0101] It should be explained that when either the transient fault flag or the steady-state fault flag is TRUE, the final fault flag is TRUE; when both the transient fault flag and the steady-state fault flag are FALSE, the final fault flag is FALSE.

[0102] It should also be explained that for angle-of-attack sensors with a final fault flag of FALSE, the corresponding valid measured angle-of-attack signals are filtered out as non-faulty measured angle-of-attack signals.

[0103] By setting transient fault judgment, the following angle-of-attack signals can be avoided: angle-of-attack signals collected by angle-of-attack sensors that malfunction due to byte abnormalities or impacts. By setting steady-state fault judgment, the following angle-of-attack signals can be avoided: angle-of-attack signals collected by angle-of-attack sensors that malfunction due to static errors, dynamic errors, icing, jamming, etc. Furthermore, by combining the transient fault judgment results and the steady-state fault judgment results, the interference of faults on subsequent monitoring and voting can be isolated to the greatest extent.

[0104] In one embodiment, when there are at least two non-faulty angle-of-attack measurement signals, step S105 involves performing initial monitoring voting using the non-faulty angle-of-attack measurement signals to obtain the initial voting status and initial voting value, including the following steps S1051, S1052, S1053A, and S1053B.

[0105] Step S1051: Calculate the absolute value of the first difference based on the non-faulty angle of attack signal, and process the absolute value of the first difference based on the preset monitoring threshold to obtain the first current monitoring matrix.

[0106] Step S1052: Filter each non-faulty angle of attack measurement signal according to the first current monitoring matrix and the first preset conditions to obtain the non-faulty angle of attack measurement signal for the initial voting.

[0107] Step S1053A: If the number of non-faulty angle-of-attack measurement signals used for the initial voting is odd, then the median of the non-faulty angle-of-attack measurement signals used for the initial voting is taken as the initial voting value, and the initial voting state is set to a non-singular state.

[0108] Step S1053B: If the number of non-faulty angle-of-attack measurement signals used for the initial voting is even, the initial voting status and initial voting value are determined based on the first current monitoring matrix and the second preset conditions.

[0109] Specifically, in step S1051, pairwise subtraction is performed on all non-faulty measured angle of attack signals to calculate multiple absolute values ​​of the first difference, wherein the absolute value of the first difference is... The calculation formula is expressed as:

[0110] (4)

[0111] in, Indicates non-fault measurement angle of attack signal Measuring angle of attack signal with or without fault The absolute value of the first difference between them This indicates the total number of angle-of-attack sensors corresponding to non-faulty angle-of-attack measurement signals. Indicates the first The non-faulty measured angle of attack signal corresponding to each angle of attack sensor. Indicates the first The non-faulty measured angle of attack signal corresponds to each angle of attack sensor.

[0112] Specifically, in step S1051, the first current monitoring matrix is ​​represented as:

[0113] (5)

[0114] in, This represents the first current monitoring matrix. The first current monitoring matrix is ​​a matrix composed of... OK A matrix composed of column elements, with For example, This represents the element located in the 1st row and 2nd column of the first current monitoring matrix.

[0115] It should be noted that the first current monitoring matrix is ​​constructed by processing the absolute values ​​of each first difference based on a preset monitoring threshold. This is done to obtain the elements of the first current monitoring matrix. For example, the absolute values ​​of the first differences... For example, if the absolute value of the first difference If the value exceeds the preset monitoring threshold and persists for a second preset duration, then the corresponding element in the first current monitoring matrix will be... Set to 1, otherwise set element Set to 0; element Located in the first current monitoring matrix Line 1 Column, 1≤ ≤ ,1≤ ≤ .

[0116] It should also be noted that the first current monitoring matrix The Middle The elements of the row are related to the first row. The voting process parameters are related to the angle of attack sensor.

[0117] Specifically, in step S1052, the first preset condition is expressed as follows:

[0118] (6)

[0119] in, This indicates that the column elements of each row in the first current monitoring matrix are summed, and the result of the column element summation is judged.

[0120] It should be noted that in step S1052, when the first current monitoring matrix contains the first... The sum of the column elements of the row is greater than At that time, the first Non-faulty angle-of-attack measurement signals corresponding to each angle-of-attack sensor are discarded. If we assume that m non-faulty angle-of-attack measurement signals corresponding to each angle-of-attack sensor are discarded in step S1052, then nm non-faulty angle-of-attack measurement signals for initial voting will be obtained, where n ≥ m.

[0121] Specifically, in step S1053A, if the number of non-faulty angle-of-attack measurement signals used for the initial voting is odd, i.e. and If the number of angle-of-attack (AOA) signals is odd, the median of the non-faulty AOA signals used for the initial voting is taken as the initial voting value, and the initial voting state is set to a non-singular state. In this case, taking the median of the non-faulty AOA signals used for the initial voting as the initial voting value should be understood as follows: arranging the odd number of non-faulty AOA signals in ascending order, the middle non-faulty AOA signal is the initial voting value.

[0122] Specifically, in step S1053B, if the number of non-faulty angle-of-attack measurement signals used for the initial voting is even, i.e. and If the number is even, the initial voting status and initial voting value are determined based on the first current monitoring matrix and the second preset conditions.

[0123] Furthermore, the determination of the initial voting status and initial voting value based on the first current monitoring matrix and the second preset conditions in step S1053B includes the following steps S1053BM and S1053BN.

[0124] Step S1053BM: If the number of non-faulty angle-of-attack measurement signals used for the initial voting is even and the first current monitoring matrix meets the second preset condition, then the initial voting state is set to a singular state, and the initial voting value is set to an invalid value. For example, the initial voting value is set to 0.

[0125] Step S1053BN: If the number of non-faulty angle-of-attack measurement signals used for the initial voting is even and the first current monitoring matrix does not meet the second preset condition, then the median of the non-faulty angle-of-attack measurement signals used for the initial voting is taken as the initial voting value, and the initial voting state is set to a non-singular state. In this case, taking the median of the non-faulty angle-of-attack measurement signals used for the initial voting as the initial voting value should be understood as follows: arranging the even number of non-faulty angle-of-attack measurement signals in ascending order, the average of the two middle non-faulty angle-of-attack measurement signals is the initial voting value.

[0126] Specifically, the second preset condition in step S1053B is expressed as follows:

[0127] (7)

[0128] in, This indicates that the column elements of all rows in the first current monitoring matrix are summed, and the summation result is evaluated; if the summation result of the column elements of all rows in the first current monitoring matrix is ​​equal to... If the sum of the column elements in any row of the first current monitoring matrix is ​​not equal to the second preset condition, then the first current monitoring matrix is ​​considered to satisfy the second preset condition. If the current monitoring matrix does not meet the second preset condition, then it is considered that the first current monitoring matrix does not meet the second preset condition.

[0129] Furthermore, regarding step S1052, in the step of filtering each non-faulty measured angle of attack signal according to the first current monitoring matrix and the first preset condition, if the number of non-faulty measured angle of attack signals used for the initial voting is zero, i.e. If the initial voting state is set to a singular state, the initial voting value will be set to an invalid value. For example, the initial voting value can be set to 0.

[0130] It should be noted that, for cases where the initial voting state is non-singular during the initial monitoring vote, such as in steps S1053A and S1053BN, the initial voting value obtained in step S106 is then used as the first angle of attack voting output value. As can be seen from steps S1053A and S1053BN, the first angle of attack voting output value obtained in both steps is determined based on the non-faulty measured angle of attack signal. This ensures that the voting prioritizes the use of the measured angle of attack signal that conforms to the actual operating conditions of the aircraft.

[0131] It should be noted that, in the case of an initial voting status that is singular during the initial monitoring vote, step S107 needs to be executed subsequently.

[0132] In one embodiment, if the initial voting state in step S107 is a singular state, then a secondary monitoring vote is performed using the inertial angle of attack estimate and the non-fault measurement angle of attack signal to obtain the secondary voting state and the second angle of attack voting output value, including the following steps S1071, S1072, S1073A, S1073B and S1073C.

[0133] Step S1071: Calculate the absolute value of the second difference based on the estimated inertial angle of attack and the non-fault measured angle of attack signal, and process the absolute value of the second difference based on the preset monitoring threshold to obtain the second current monitoring matrix.

[0134] It should be noted that the non-faulty angle of attack signal used in step S1071 is the same as the non-faulty angle of attack signal used in step S1051.

[0135] Step S1072: Based on the second current monitoring matrix and the first preset conditions, the inertial angle of attack estimate and the non-fault measurement angle of attack signal are filtered to obtain the angle of attack signal for secondary voting.

[0136] It should be noted that the first preset condition used in step S1072 is the same as the first preset condition used in step S1052.

[0137] Step S1073A: If the number of angle-of-attack signals used for the second voting is greater than zero and is odd, then the median of the angle-of-attack signals used for the second voting is taken as the second angle-of-attack voting output value, and the second voting state is set to a non-singular state.

[0138] Step S1073B: If the number of angle-of-attack signals used for secondary voting is greater than zero and is even, then determine the secondary voting status and the second angle-of-attack voting output value based on the second current monitoring matrix and the second preset conditions.

[0139] It should be noted that the second preset condition used in step S1073B is the same as the second preset condition used in step S1053B.

[0140] Step S1073C: If the number of angle of attack signals used for the second voting is zero, then the inertial angle of attack estimate is used as the second angle of attack voting output value, and the second voting state is set to a singular state.

[0141] It should be noted that in step S1071, pairwise subtraction is performed on the estimated inertial angle of attack and all non-fault-measured angle of attack signals to calculate multiple second absolute differences, which can be referred to the calculation method in step S1051. Furthermore, the method of processing the second absolute differences based on a preset monitoring threshold in step S1071 can be referenced in the method of processing the first absolute differences based on a preset monitoring threshold in step S1051. The representation of the second current monitoring matrix in step S1071 can refer to the representation of the first current monitoring matrix in step S1051, with the difference being that: the elements in one row of the second current monitoring matrix are voting process parameters related to the estimated inertial angle of attack, while the elements in the other rows of the second current monitoring matrix are voting process parameters related to the corresponding angle of attack sensors.

[0142] It should be noted that in step S1072, when the second current monitoring matrix contains the first... The sum of the column elements of the row is greater than At that time, the first The corresponding inertial angle of attack estimates or non-faulty measured angle of attack signals are discarded. If we assume that in step S1072, h non-faulty measured angle of attack signals and k inertial angle of attack estimates corresponding to angle of attack sensors are discarded, then nhk angle of attack signals are obtained for secondary voting, where n-1≥h≥0 and 1≥k≥0.

[0143] It should be explained that, since the initial voting state is a singular state, in the secondary monitoring and voting, in addition to using the non-fault-measured angle of attack signal used in the initial monitoring and voting, the key point is to introduce a steady-state, high-accuracy inertial angle of attack estimate for the secondary monitoring and voting. Based on this, for step S1072, in the second current monitoring matrix, the summation result of the column elements corresponding to the inertial angle of attack estimate should satisfy a condition less than or equal to... The inertial angle of attack estimate will not be eliminated. The angle of attack signal obtained in step S1072 for secondary voting should include the inertial angle of attack estimate.

[0144] Specifically, in step S1073A, if the number of angle-of-attack signals used for the second voting is greater than zero and odd (i.e., nhk ≥ 1 and nhk is odd), then the median of the angle-of-attack signals used for the second voting is taken as the second angle-of-attack voting output value, and the second voting state is set to a non-singular state. In this case, taking the median of the angle-of-attack signals used for the second voting as the second angle-of-attack voting output value should be understood as follows: arranging the odd number of angle-of-attack signals used for the second voting in ascending order, the middle angle-of-attack signal is the second angle-of-attack voting output value.

[0145] Specifically, for step S1073A, the specific values ​​of the second angle of attack voting output value include the following cases: when the number of angle of attack signals used for the second voting is 1, and the angle of attack signal used for the second voting is the inertial angle of attack estimate, the second angle of attack voting output value is the inertial angle of attack estimate; when the number of angle of attack signals used for the second voting is ≥3, and the angle of attack signal used for the second voting consists of the inertial angle of attack estimate and several non-fault measured angle of attack signals, the second angle of attack voting output value is: the inertial angle of attack estimate or one of the non-fault measured angle of attack signals.

[0146] Specifically, in step S1073B, if the number of angle-of-attack signals used for secondary voting is greater than zero and is even, i.e., nhk≥2 and nhk is even, then the secondary voting state and the second angle-of-attack voting output value are determined based on the second current monitoring matrix and the second preset conditions.

[0147] Specifically, the determination of the secondary voting status and the second angle of attack voting output value based on the second current monitoring matrix and the second preset conditions includes the following steps S1073BM and S1073BN.

[0148] Step S1073BM: If the number of angle of attack signals used for secondary voting is even and the second current monitoring matrix satisfies the second preset condition, then the inertial angle of attack estimate is used as the second angle of attack voting output value, and the secondary voting state is set to a singular state.

[0149] Step S1073BN: If the number of angle-of-attack signals used for the second voting is even and the second current monitoring matrix does not meet the second preset condition, then the median of the angle-of-attack signals used for the second voting is taken as the second angle-of-attack voting output value, and the second voting state is set to a non-singular state. In this case, taking the median of the angle-of-attack signals used for the second voting as the second angle-of-attack voting output value should be understood as follows: arranging the even number of angle-of-attack signals used for the second voting in ascending order, the average of the two middle angle-of-attack signals is the second angle-of-attack voting output value.

[0150] It should be noted that, for step S1073BN, the specific values ​​of the second angle of attack voting output value include the following cases: when the number of angle of attack signals used for the second voting is 2, and the angle of attack signal used for the second voting consists of an inertial angle of attack estimate and a non-fault measured angle of attack signal, the second angle of attack voting output value is: the average value of the inertial angle of attack estimate and the non-fault measured angle of attack signal; when the number of angle of attack signals used for the second voting is ≥ 4, and the angle of attack signal used for the second voting consists of an inertial angle of attack estimate and several non-fault measured angle of attack signals, the second angle of attack voting output value is: the average value of the inertial angle of attack estimate and one of the non-fault measured angle of attack signals, or the average value of two of the non-fault measured angle of attack signals.

[0151] It should also be noted that the second preset condition used in step S1073B is the same as the second preset condition used in step S1053B. If the sum of the column elements of all rows in the second current monitoring matrix is ​​equal to... If the sum of the column elements in any row of the second current monitoring matrix is ​​not equal to the second preset condition, then the second current monitoring matrix is ​​considered to satisfy the second preset condition. If the second current monitoring matrix does not meet the second preset condition, then it is considered that the second current monitoring matrix does not meet the second preset condition.

[0152] Specifically, in step S1073C, if the number of angle of attack signals used for the second voting is zero, i.e., nhk=0, then the inertial angle of attack estimate is used as the second angle of attack voting output value, and the second voting state is set to a singular state.

[0153] It should be noted that, for the case where the initial voting state is non-singular, in step S108, the angle of attack fusion signal is reconstructed using the first angle of attack voting output value and the inertial angle of attack change rate signal. However, for the case where the initial voting state is singular, step S107 needs to be executed, and in step S108, the angle of attack fusion signal is reconstructed using the second angle of attack voting output value and the inertial angle of attack change rate signal.

[0154] It should be noted that if the number of non-faulty angle-of-attack signals selected in step S104 is only one, then the monitoring and voting will be performed according to the specific steps of step S107 using the inertial angle-of-attack estimate and the one non-faulty angle-of-attack signal. The specific voting process will not be described in detail here.

[0155] It should also be noted that if the number of non-faulty angle-of-attack signals selected in step S104 is zero, the angle-of-attack fusion signal is directly reconstructed using the inertial angle-of-attack estimate and the inertial angle-of-attack rate of change signal. For example, the inertial angle-of-attack estimate is reconstructed using formula (1); and the value in formula (2) is... The value is taken as the estimated value of the inertial angle of attack. The inertial angle of attack change rate signal is calculated using formula (2). The calculated inertial angle of attack change rate signal is substituted into formula (3), and the angle of attack fusion signal is further determined using formula (3).

[0156] In one embodiment, in step S108, the inertial angle of attack rate of change signal is represented as:

[0157] (2)

[0158] in, This represents the rate of change of inertial angle of attack. Indicates pitch angular velocity, This indicates the first angle of attack voting output value or the second angle of attack voting output value. Indicates the sideslip angle. Indicates a tilting or bending posture. Indicates rolling posture, Indicates forward overload. Indicates normal overload, Represents gravitational acceleration. Indicates the roll rate. Indicates yaw rate, This indicates the vacuum velocity of the aircraft.

[0159] In step S108, the angle-of-attack fusion signal is represented as:

[0160] (3)

[0161] in, Indicates angle-of-attack fusion signal, This indicates the angle of attack deviation value. Indicates the second-order filter frequency. This indicates second-order filter damping. This indicates time. It should be explained that the angle of attack deviation value is the difference between the current angle of attack voting output value and the angle of attack fusion signal value from the previous moment.

[0162] To more clearly describe the method proposed in this application, the following illustrative examples, namely, Embodiment 1 and Embodiment 2, illustrate the method proposed in this application.

[0163] Example 1:

[0164] In Example 1, the aircraft is equipped with a triplet angle-of-attack sensor configuration and flies at high angles of attack in the air, and the angle-of-attack measurement signals of all angle-of-attack sensors are valid angle-of-attack measurement signals.

[0165] If the aircraft satisfies the first inertial data validity and the aircraft is in the air, then the estimated value of the inertial angle of attack is calculated using formula (1).

[0166] By determining that none of the aircraft's triple angle-of-attack sensors issued fault flags and that the aircraft's bus transmission verification was fault-free, the measured angle-of-attack signals of the three angle-of-attack sensors that did not issue fault flags were taken as valid measured angle-of-attack signals, thus obtaining the valid measured angle-of-attack signals of the three angle-of-attack sensors.

[0167] According to step S102, the effective measured angle of attack signals of the three angle of attack sensors are subtracted from the estimated inertial angle of attack to calculate three first deviation values. If the absolute value of each of the three first deviation values ​​is less than 20°, the transient fault flag of the three angle of attack sensors is set to FALSE. The preset transient fault judgment threshold is set to 20°.

[0168] The effective measured angle of attack signals and inertial angle of attack estimates from the three angle of attack sensors are passed through a first-order low-pass filter with a cutoff frequency of 0.05 rad / s to obtain the three filtered effective measured angle of attack signals and the filtered inertial angle of attack estimates. According to step S103, the three filtered effective measured angle of attack signals are subtracted from the filtered inertial angle of attack estimates to calculate three second deviation values. If the absolute value of these three second deviation values ​​does not exceed 5° for two consecutive seconds, the steady-state fault flag for the three angle of attack sensors is set to FALSE. The preset steady-state fault judgment threshold is set to 5°, the first preset duration is set to 2 seconds, and rad / s represents radians per second.

[0169] According to step S104, based on the transient fault flag and steady-state fault flag of the three angle-of-attack sensors obtained above, the final fault flag of the three angle-of-attack sensors is FALSE. Then, the effective measured angle-of-attack signals of the three angle-of-attack sensors are filtered into non-fault measured angle-of-attack signals, that is, three non-fault measured angle-of-attack signals are obtained for subsequent monitoring and voting.

[0170] Using formula (4), the pairwise subtraction of the three non-faulty angle-of-attack signals is performed to calculate six first absolute differences. Based on the preset monitoring threshold, the six first absolute differences are processed to obtain the first current monitoring matrix. The number of rows and columns of the first current monitoring matrix are both 3. The preset monitoring threshold in this embodiment is set to 3°. According to step S1052, the column elements of the three rows in the first current monitoring matrix are summed. The three sums are all less than 1.5. Therefore, the three non-faulty angle-of-attack signals are retained for the initial monitoring vote. According to step S1053A, the median of the three non-faulty angle-of-attack signals is used as the initial voting value. This initial voting value is the first angle-of-attack voting output value. The initial voting state is set to a non-singular state. In this case, step S108 is executed directly.

[0171] The inertial angle of attack rate of change signal is calculated using formula (2), and the first angle of attack voting output value is fused with the inertial angle of attack rate of change signal using formula (3) to obtain the angle of attack fused signal.

[0172] refer to Figures 2 to 4 As shown, Figure 2 The effective measured angle of attack signals of the three angle of attack sensors are shown. Figure 3 The estimated angle of attack is shown. Figure 4 The initial voting values ​​and angle-of-attack fusion signal are shown. (Comparison is possible.) Figure 2 and Figure 3 , Figure 3 The estimated inertial angle of attack and Figure 2The effective angle-of-attack signals measured by angle-of-attack sensors 1, 2, and 3 are basically consistent, which proves that the inertial angle-of-attack estimate reconstructed by formula (1) is accurate enough to basically reflect the actual operating conditions of the aircraft. Figure 4 As can be seen from the results, after angle-of-attack fusion processing, the obtained angle-of-attack fusion signal is consistent with the initial voting value in terms of both steady-state and large-range dynamic fluctuations. This proves that the angle-of-attack signal reconstruction and fusion method proposed in this application can compensate for dynamic changes in angle of attack and has high precision and high reliability.

[0173] Example 2:

[0174] In Example 2, the aircraft is equipped with a triplet angle-of-attack sensor. When it is disturbed by turbulent wind field in the air, the triplet angle-of-attack sensor monitoring votes briefly enter a 1:1:1 singular state. The angle-of-attack signal reconstruction and fusion method proposed in this invention will effectively improve the robustness and availability of the angle-of-attack signal.

[0175] If the aircraft satisfies the first inertial data validity and the aircraft is in the air, then the estimated value of the inertial angle of attack is calculated using formula (1).

[0176] By determining that none of the aircraft's triple angle-of-attack sensors issued fault flags and that the aircraft's bus transmission verification was fault-free, the measured angle-of-attack signals of the three angle-of-attack sensors that did not issue fault flags were taken as valid measured angle-of-attack signals, thus obtaining the valid measured angle-of-attack signals of the three angle-of-attack sensors.

[0177] According to step S102, the effective measured angle of attack signals of the three angle of attack sensors are subtracted from the estimated inertial angle of attack to calculate three first deviation values. If the absolute value of each of the three first deviation values ​​is less than 20°, the transient fault flag of the three angle of attack sensors is set to FALSE. The preset transient fault judgment threshold is set to 20°.

[0178] The effective measured angle of attack signals and inertial angle of attack estimates from the three angle of attack sensors are passed through a first-order low-pass filter with a cutoff frequency of 0.05 rad / s to obtain the three filtered effective measured angle of attack signals and the filtered inertial angle of attack estimates. According to step S103, the three filtered effective measured angle of attack signals are subtracted from the filtered inertial angle of attack estimates to calculate three second deviation values. If the absolute value of these three second deviation values ​​does not exceed 5° for two consecutive seconds, the steady-state fault flag for the three angle of attack sensors is set to FALSE. The preset steady-state fault judgment threshold is set to 5°, the first preset duration is set to 2 seconds, and rad / s represents radians per second.

[0179] According to step S104, based on the transient fault flag and steady-state fault flag of the three angle-of-attack sensors obtained above, the final fault flag of the three angle-of-attack sensors is FALSE. Then, the effective measured angle-of-attack signals of the three angle-of-attack sensors are filtered into non-fault measured angle-of-attack signals, that is, three non-fault measured angle-of-attack signals are obtained for subsequent monitoring and voting.

[0180] Using formula (4), pairwise subtraction is performed on the three non-faulty angle-of-attack signals to calculate six first absolute differences. Based on a preset monitoring threshold, the six first absolute differences are processed to obtain a first current monitoring matrix, which has 3 rows and 3 columns. In step S1051, since some of the first absolute differences meet the condition of being greater than 3° and lasting for 0.3 seconds, some elements in the first current monitoring matrix are set to 1. The preset monitoring threshold is set to 3°, and the second preset duration is set to 0.3 seconds.

[0181] According to step S1052, the column elements of all three rows in the first current monitoring matrix are summed. If the sums are all greater than 1.5, the three non-fault angle of attack signals are eliminated. The number of non-fault angle of attack signals used for the initial voting is zero. The initial voting state is set to a singular state, and the initial voting value is set to an invalid value. In this case, steps S107 and S108 need to be executed sequentially.

[0182] According to step S1071, the inertial angle of attack estimate and three non-fault-measured angle of attack signals are used as four angle of attack signals for screening. The pairwise subtraction of these four angle of attack signals is performed to calculate twelve absolute values ​​of the second difference. These twelve absolute values ​​of the second difference are then processed based on a preset monitoring threshold to obtain a second current monitoring matrix, which has four rows and four columns. According to step S1072, based on the obtained second current monitoring matrix, the column elements of all four rows in the second current monitoring matrix are summed, resulting in two rows in the second current monitoring matrix that contain... If the sum of the column elements is greater than 2, then the angle of attack signals corresponding to the two rows are removed, and the number of angle of attack signals used for secondary voting is 2. According to step S1073B, the secondary voting state and the second angle of attack voting output value are determined based on the second current monitoring matrix and the second preset condition expressed by formula (7). If it is determined that the second current monitoring matrix does not meet the second preset condition, then according to step S1073BN, the inertial angle of attack estimate value used for secondary voting and a non-fault measurement angle of attack signal are averaged to obtain the secondary voting value, which is the second angle of attack voting output value.

[0183] The inertial angle of attack rate of change signal is calculated using formula (2), and the second angle of attack voting output value is fused with the inertial angle of attack rate of change signal using formula (3) to obtain the angle of attack fused signal.

[0184] refer to Figures 5 to 7 As shown, Figure 5 The effective measured angle of attack signals of the three angle of attack sensors are shown. Figure 6 The estimated angle of attack is shown. Figure 7 The initial voting value, the second voting value, and the angle-of-attack fusion signal are shown. Figures 5 to 7 In the middle, the position of the dotted line corresponds to the moment when the initial monitoring vote shows an unusual state, from... Figure 5 As can be seen, although the measured angle of attack signals from the three angle of attack sensors are valid, during a certain time period to the right of the dashed line, the significant fluctuations in the valid measured angle of attack signals from angle of attack sensors 1, 2, and 3 lead to an unusual state in the initial monitoring vote. Furthermore, combined with... Figures 5 to 7 It can be seen that when the initial monitoring vote is in a singular state, by using the inertial angle of attack estimate and the non-fault measured angle of attack signal together for the secondary monitoring vote, a second angle of attack vote output value that is close to the actual measured angle of attack signal collected by the angle of attack sensor can be obtained. Based on the second angle of attack vote output value, an accurate and reliable angle of attack fusion signal can be obtained through further fusion.

[0185] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein.

Claims

1. A method for reconstructing and fusing angle-of-attack signals, characterized in that, include: The estimated inertial angle of attack is reconstructed based on the first inertial data, which includes the aircraft's pitch attitude, roll attitude, track tilt attitude, and sideslip angle. Acquire effective measured angle of attack signals from multiple angle of attack sensors, and use the effective measured angle of attack signals and the inertial angle of attack estimation value to perform transient fault judgment, thereby obtaining the transient fault judgment result of each angle of attack sensor; The inertial angle of attack estimate and the effective measured angle of attack signal are filtered, and the filtered effective measured angle of attack signal and the inertial angle of attack estimate are used to determine steady-state faults, so as to obtain the steady-state fault determination results of each angle of attack sensor. Based on the transient fault judgment result and the steady-state fault judgment result, non-faulty measured angle of attack signals are selected from the valid measured angle of attack signals of the plurality of angle of attack sensors; The initial monitoring vote is performed using the non-faulty measured angle of attack signal to obtain the initial voting status and initial voting value; If the initial voting state is a non-singular state, then the initial voting value is used as the first angle of attack voting output value, which is determined based on the non-fault measured angle of attack signal. If the initial voting state is a singular state, a secondary monitoring vote is performed using the inertial angle of attack estimate and the non-faulty measured angle of attack signal to obtain the secondary voting state and the second angle of attack voting output value; wherein, the second angle of attack voting output value is determined based on the inertial angle of attack estimate, or the second angle of attack voting output value is determined based on the inertial angle of attack estimate and the non-faulty measured angle of attack signal; The angle-of-attack fusion signal is reconstructed based on either the first angle-of-attack voting output value or the second angle-of-attack voting output value, as well as the inertial angle-of-attack rate of change signal.

2. The method for reconstructing and fusing angle-of-attack signals according to claim 1, characterized in that, The estimated inertial angle of attack is expressed as: （1） in, This represents the estimated angle of attack due to inertia. , , , , Indicates a tilting or bending posture. Indicates rolling posture, Indicates the tilt attitude of the flight path. Indicates the sideslip angle.

3. The method for reconstructing and fusing angle-of-attack signals according to claim 1, characterized in that, The transient fault judgment is performed using the effective measured angle of attack signal and the inertial angle of attack estimate to obtain the transient fault judgment results of each angle of attack sensor, including: The effective measured angle of attack signal is subtracted from the inertial angle of attack estimate to calculate the first deviation value; The first deviation value is compared with the preset transient fault judgment threshold: If the first deviation value is greater than the preset transient fault judgment threshold, then the angle of attack sensor is determined to have a transient fault. Otherwise, it is determined that the angle of attack sensor does not have a transient fault.

4. The method for reconstructing and fusing angle-of-attack signals according to claim 3, characterized in that, The steady-state fault judgment is performed using the filtered effective measured angle of attack signal and the inertial angle of attack estimate to obtain the steady-state fault judgment results for each angle of attack sensor, including: The second deviation value is calculated by subtracting the filtered effective measured angle of attack signal from the filtered inertial angle of attack estimate. The second deviation value is compared with the preset steady-state fault judgment threshold: If the second deviation value is greater than the preset steady-state fault judgment threshold and continues for a first preset duration, then the angle of attack sensor is determined to have a steady-state fault. Otherwise, it is determined that the angle of attack sensor does not have a steady-state fault.

5. The method for reconstructing and fusing angle-of-attack signals according to claim 4, characterized in that, The non-fault-prone measured angle of attack signal is an effective measured angle of attack signal from an angle of attack sensor that meets preset screening conditions, wherein the preset screening conditions are that neither the transient fault nor the steady-state fault exists simultaneously.

6. The method for reconstructing and fusing angle-of-attack signals according to claim 1, characterized in that, When there are at least two non-faulty angle-of-attack measurement signals, the initial monitoring vote is performed using the non-faulty angle-of-attack measurement signals to obtain the initial voting status and initial voting value, including: The first absolute value of the difference is calculated based on the non-faulty measured angle of attack signal, and the first absolute value of the difference is processed based on a preset monitoring threshold to obtain the first current monitoring matrix; Based on the first current monitoring matrix and the first preset conditions, each of the non-fault measurement angle of attack signals is filtered to obtain the non-fault measurement angle of attack signals used for the initial voting; If the number of non-faulty angle-of-attack measurement signals used for the initial vote is odd, then the median of the non-faulty angle-of-attack measurement signals used for the initial vote is taken as the initial vote value, and the initial vote state is set to the non-singular state. If the number of non-faulty angle-of-attack measurement signals used for the initial voting is even, the initial voting status and the initial voting value are determined based on the first current monitoring matrix and the second preset conditions.

7. The method for reconstructing and fusing angle-of-attack signals according to claim 6, characterized in that, The step of determining the initial voting status and the initial voting value based on the first current monitoring matrix and the second preset conditions includes: If the number of non-faulty angle-of-attack measurement signals used for the initial voting is even and the first current monitoring matrix satisfies the second preset condition, then the initial voting state is set to the singular state and the initial voting value is set to an invalid value. If the number of non-faulty angle-of-attack measurement signals used for the initial vote is even and the first current monitoring matrix does not meet the second preset condition, then the median of the non-faulty angle-of-attack measurement signals used for the initial vote is taken as the initial vote value, and the initial vote state is set to the non-singular state.

8. The method for reconstructing and fusing angle-of-attack signals according to claim 6, characterized in that, In the step of filtering each non-faulty angle-of-attack measurement signal according to the first current monitoring matrix and the first preset condition, if the number of non-faulty angle-of-attack measurement signals used for the initial voting is zero, the initial voting state is set to a singular state and the initial voting value is set to an invalid value.

9. The method for reconstructing and fusing angle-of-attack signals according to claim 7 or 8, characterized in that, If the initial voting state is a singular state, then a secondary monitoring vote is performed using the inertial angle of attack estimate and the non-faulty measured angle of attack signal to obtain the secondary voting state and the second angle of attack voting output value, including: The second absolute value of the difference is calculated based on the estimated inertial angle of attack and the non-fault-measured angle of attack signal, and the second absolute value of the difference is processed based on the preset monitoring threshold to obtain the second current monitoring matrix; Based on the second current monitoring matrix and the first preset condition, the inertial angle of attack estimate and the non-fault measured angle of attack signal are filtered to obtain the angle of attack signal for secondary voting; If the number of angle-of-attack signals used for the second vote is greater than zero and is odd, then the median of the angle-of-attack signals used for the second vote is taken as the second angle-of-attack voting output value, and the second vote state is set to the non-singular state. If the number of angle-of-attack signals used for secondary voting is greater than zero and is even, then the secondary voting state and the second angle-of-attack voting output value are determined based on the second current monitoring matrix and the second preset condition. If the number of angle-of-attack signals used for the second vote is zero, then the estimated inertial angle of attack is used as the second angle-of-attack voting output value, and the second vote state is set to the singular state.

10. The method for reconstructing and fusing angle-of-attack signals according to claim 1, characterized in that, The inertial angle of attack change rate signal is represented as: （2） in, This represents the rate of change of inertial angle of attack. Indicates pitch angular velocity, This indicates the first angle of attack voting output value or the second angle of attack voting output value. Indicates the sideslip angle. Indicates a tilting or bending posture. Indicates rolling posture, Indicates forward overload. Indicates normal overload, Represents gravitational acceleration. Indicates the roll rate. Indicates yaw rate, Indicates the vacuum velocity of the aircraft; The angle-of-attack fusion signal is represented as: （3） in, Indicates angle-of-attack fusion signal, This indicates the angle of attack deviation value. Indicates the second-order filter frequency. This indicates second-order filter damping. Indicates time.

Images