An aircraft fault identification method based on extended-dimensional parallel filtering and hierarchical judgment

Abstract

A method for aircraft fault identification based on expanded-dimensional parallel filtering and hierarchical judgment. The purpose of the present invention is to solve the problems of slow speed and poor accuracy of existing aircraft fault identification. By studying a fast and accurate online fault identification method with low model / data dependence, the speed and accuracy of fault identification technology are improved. At the same time, the dependence on measurement information and the use of sensors can be reduced. The present invention adopts the method of expanded-dimensional parallel filtering, which can greatly reduce the amount of calculation and improve the identification speed. It does not require the addition of measurement sensors and can achieve fast and accurate state estimation only by relying on the information provided by the upper-level navigation system. A hierarchical judgment method of fault location / fault mode / fault degree is adopted to gradually cut off the invalid branches of the observer, reduce the amount of calculation, and improve the identification speed and accuracy.

Description

Technical Field

[0001] The present invention relates to a low-time, high-precision aircraft online fault identification method based on adaptive filtering technology and hierarchical determination technology, which can identify the fault location, fault mode and fault degree according to the current aircraft status. Background Art

[0002] The probability of failure of aircraft power system and actuator components increases dramatically under the influence of complex flight environments. Online fault identification is essential for effective response. Due to the difficulty in accurately modeling aircraft systems, a lack of prior knowledge about fault information, and limited measurement data quantity and quality, existing fault identification technologies require extensive sensor data, resulting in complex models, slow fault identification, and insufficient accuracy in identifying the severity of the fault. There is an urgent need to develop new, rapid, and accurate identification methods. Summary of the Invention

[0003] The purpose of the present invention is to solve the problems of slow speed and poor accuracy in existing aircraft fault identification, and to provide a method for rapid and accurate aircraft fault identification based on expanded-dimensional parallel adaptive filtering and hierarchical judgment technology. By studying a fast and accurate online fault identification method with low model / data dependence, the speed and accuracy of fault identification technology are improved, while the dependence on measurement information and the use of sensors can be reduced.

[0004] The purpose of the present invention is achieved through the following technical solutions:

[0005] A method for rapid and accurate identification of aircraft faults based on extended-dimensional parallel adaptive filtering and hierarchical determination technology, the method steps are as follows:

[0006] Step 1: Analyze the typical failure modes of the thrust system / servo mechanism according to the working status of the aircraft engine thrust system and servo mechanism, and establish a typical thrust failure model T iout =k i T i Typical servo mechanism model Where T i is the thrust corresponding to the i-th position, k i is the damage coefficient of the thrust corresponding to the i-th position, δ i (t) is the servo mechanism swing angle corresponding to the i-th position, σ i is the damage coefficient of the swing angle corresponding to the i-th position, δ ki is the angle of the swing angle corresponding to the i-th position, and w is the random value of the servo mechanism swing angle;

[0007] Step 2: Analyze the aircraft state before and after the failure: x-axis displacement x, y-axis displacement y, z-axis displacement z, x-axis speed v x , y-axis speed vy , z-axis speed v z , x-axis attitude angular velocity ω x , y-axis attitude angular velocity ω y , z-axis attitude angular velocity ω z , the i-th position corresponds to the thrust T i , the jth position corresponds to the servo mechanism swing angle θ j According to the changes of fault models, multiple unbiased dynamic models with faults are established, and the system state equation and measurement equation are established.

[0008] Step 3: Change the thrust of the main engine T i 、The swing angle θ of the servo mechanism of each engine j As the expanded state quantity, it is respectively expanded to the state quantity X of multiple low-dimensional filters. T =[ω x ,ω y ,ω z ,T i ], X θ =[ω x ,ω y ,ω z ,θ j ], where X T Indicates thrust T i The state quantity of the expanded dimension, X θ Indicates the servo mechanism swing angle θ j Expanded state quantity, multiple filters run simultaneously to obtain engine thrust estimation and the estimated servo mechanism swing angle

[0009] Step 4: Make the first level judgment based on the estimation result of the total thrust observer. If a thrust failure occurs, disable the servo mechanism observer. T =[ω x ,ω y ,ω z ,T i ] Carry out the second level judgment to obtain the fault location and fault extent. If there is no thrust fault, disable the thrust observer and enter the servo mechanism fault judgment;

[0010] Step 5: Servo mechanism fault judgment First, estimate the parameter error S according to the observer. i and covariance R i Determine the fault location, stop the servo mechanism observer at the fault location to perform the third level judgment, and expand the servo mechanism observer X θ =[ω x ,ω y ,ω z ,θj ]Get the estimated value of the swing angle Determine the failure mode and extent of the failure.

[0011] Furthermore, in step 2, the system state equation and measurement equation are specifically:

[0012] Step 21: Set the system state vector x = [ω x ω y ω z ] T The output of the system in Indicates the x-axis attitude angular velocity output, Indicates the y-axis attitude angular velocity output, Indicates the z-axis attitude angular velocity output;

[0013] Step 22: Establish the state equation and measurement equation based on the rocket's center-of-mass dynamics equation and kinematics equation:

[0014]

[0015] Among them, A represents the system matrix, B represents the control matrix of the system, C represents the output matrix of the system, d(t) represents the disturbance vector of the system, and E d represents the disturbance transformation matrix, x(t) represents the state quantity of the system, u(t) represents the control vector of the system, and y(t) represents the quantity measurement of the system.

[0016] Furthermore, in step 3, the dimension expansion steps for each filter observer are as follows:

[0017] Step 31: Design the state quantity X of engine thrust expansion T =[w x ,w y ,w z ,T j ];

[0018] Step 32: The system attitude dynamics model is expressed as follows:

[0019]

[0020] y=[I 0]x+v

[0021] Where x is the state vector, f(x) is the nonlinear function vector of the state vector, B is the control matrix of the system, T is the engine nozzle thrust vector, which is the input vector of the system, T=[T1 T2] T, w is the state disturbance vector of the system, y is the measurement vector of the system, v is the measurement noise vector of the system, I is the identity matrix, T1 is the thrust corresponding to the No. 1 engine, T2 is the thrust corresponding to the No. 2 engine, M1 is the torque generated by the thrust T1, and M2 is the torque generated by the thrust T2;

[0022] Step 3: Design the servo mechanism to expand the state variable: X θ =[w x ,w y ,w z ,θ j ], the design of the servo mechanism expanded state equation is the same as the thrust expansion method.

[0023] Furthermore, the step 4 is specifically as follows:

[0024] Step 41: At each moment, the system status and measurement updates are corrected using information input from the navigation system. Each filter observer operates normally. The total thrust observer estimates the total thrust of the aircraft and determines whether a thrust failure has occurred. If a thrust failure has occurred, proceed to Step 42; if not, proceed to Step 43.

[0025] Step 42: If a thrust failure occurs, disable the servo mechanism's extended-dimensional filter and estimate the thrust of each engine using the extended-dimensional single-engine thrust observer to determine the location and severity of the thrust failure.

[0026] Step 43: If there is no thrust fault, disable the thrust part expansion observer and enter the servo mechanism fault judgment.

[0027] Furthermore, the step five is specifically as follows:

[0028] Step 51: Servo mechanism fault judgment First, estimate the parameter error S according to the servo mechanism observer. i and covariance R i Determine the fault location, calculation method:

[0029]

[0030] Where Φ is the error solution matrix, Φ(i,:) is the vector of elements in the i-th row of Φ, satisfying Φ=αM p , where α is the matrix allocation coefficient, M p is the torque distribution matrix, It represents the torque caused by the change of the servo mechanism's swing angle, [δ c0 ,δ c1 ,δ c2 ,δ c3] are the x and y direction components of the swing angles of the two servo mechanisms corresponding to one engine, R x 、R yz Respectively represent the projection of the distance between the engine and the center of mass of the aircraft on the x-axis and y-axis / z-axis (the aircraft axis is symmetrical), P c Represents the thrust of a single engine, J x 、J yz Represents the moment of inertia of the aircraft about the x-axis and y-axis / z-axis respectively;

[0031] Step 52: Disable the servo mechanism observer at the non-faulty position and perform the third-level judgment to obtain the swing angle estimate through the servo swing angle expansion observer. Determine the servo mechanism failure mode and failure severity.

[0032] The beneficial effects of the present invention relative to the prior art are:

[0033] (1) The use of extended-dimensional parallel filtering can greatly reduce the amount of calculation and improve the recognition speed. There is no need to add measurement sensors. Fast and accurate state estimation can be achieved by relying solely on the information provided by the upper-level navigation system.

[0034] (2) A hierarchical determination method based on fault location / fault mode / fault severity is adopted to gradually cut off invalid branches of the observer, reduce the amount of calculation, and improve the identification speed and accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 This is a diagram of the architecture of the dimension-expanding parallel observer of the present invention;

[0036] Figure 2 This is a flowchart of the hierarchical determination of the present invention. DETAILED DESCRIPTION

[0037] The technical solution of the present invention is further described below with reference to the accompanying drawings and embodiments, but is not limited thereto. Any modification or equivalent replacement of the technical solution of the present invention that does not depart from the spirit and scope of the technical solution of the present invention should be included in the scope of protection of the present invention.

[0038] Example 1:

[0039] A method for rapid and accurate identification of aircraft faults based on extended-dimensional parallel adaptive filtering and hierarchical determination technology is designed. The specific steps of the method are as follows:

[0040] Step 1: Analyze the typical failure modes of the thrust system / servo mechanism according to the working status of the aircraft engine thrust system and servo mechanism, and establish a typical thrust failure model T i =k i T i Typical servo mechanism model

[0041]

[0042] Step 2: Analyze the aircraft status before and after the failure: x-axis position x, y-axis position y, z-axis position z, x-axis speed v x , y-axis speed v y , z-axis speed v z , x-axis attitude angular velocity ω x , y-axis attitude angular velocity ω y , z-axis attitude angular velocity ω z , the thrust T of the engine at position i i , the swing angle θ of the jth servo mechanism j According to the changes of fault models, multiple unbiased dynamic models with faults are established, and the system state equation and measurement equation are established.

[0043]

[0044] Step 3: Set the main engine thrust T i 、The swing angle θ of the servo mechanism of each engine j As the dimension-expanded state quantity, the state quantity X=[ω x ,ω y ,ω z ,T i ],X=[ω x ,ω y ,ω z ,θ j ], multiple filters are run simultaneously to control the engine thrust and servo mechanism swing angle Make estimates;

[0045] Step 4: Perform the first-level judgment based on the estimation result of the total thrust observer. If a thrust failure occurs, disable the servo mechanism observer and use the extended dimension single-engine thrust observer X = [ω x ,ω y ,ω z ,T i ] Carry out the second level judgment to obtain the fault location and fault extent. If there is no thrust fault, disable the thrust observer and enter the servo mechanism fault judgment;

[0046] Step 5: Servo mechanism fault judgment First, estimate the parameter error S according to the observer i and covariance R i Determine the fault location, stop the servo mechanism observer at the fault location to perform the third level judgment, and use the augmented servo mechanism observer X = [ω x ,ω y ,ω z,θ j ]Get the estimated value of the swing angle Determine the failure mode and extent of the failure.

[0047] Furthermore, in step 2, the system state equation and measurement equation are established as follows:

[0048] Step 2.1: Set the system state vector x = [ω x ω y ω z ] T The output of the system

[0049] Step 2.2: Based on the dynamic equations and kinematic equations of the rocket orbiting the center of mass, establish the following state equations and measurement equations:

[0050]

[0051] Among them, A represents the system matrix, B represents the control matrix of the system, C represents the output matrix of the system, d(t) represents the disturbance vector of the system, and E d represents the perturbation transformation matrix.

[0052] Furthermore, in step 3, the dimension expansion steps for each filter observer are as follows:

[0053] Step 3.1: First, design the state quantity X = [ω x ,ω y ,ω z ,T j ];

[0054] Step 3.2: The system attitude dynamics model can be expressed as follows:

[0055]

[0056] y=[I 0]x+v

[0057] Where x is the state vector, in ψ,γ represent the pitch angle, yaw angle and roll angle respectively, f(x) is the nonlinear function vector of the state vector, B is the control matrix of the system, T is the engine nozzle thrust vector, which is the input vector of the system, T=[T1 T2] T , w is the state disturbance vector of the system, y is the measurement vector of the system, and v is the measurement noise vector of the system;

[0058] Step 3.3: Design the servo mechanism to expand the state variables: X = [ω x ,ω y ,ω z,θ j ], the design of the servo mechanism expansion state equation is the same as the thrust expansion method, repeat step 3.2.

[0059] Furthermore, the thrust fault determination steps in step 4 are as follows:

[0060] Step 4.1: At each moment, the system status and measurement updates are corrected using information input from the navigation system. Each filter observer operates normally. The total thrust observer estimates the total thrust of the aircraft and determines whether a thrust failure has occurred. If a thrust failure has occurred, proceed to Step 4.2; if not, proceed to Step 4.3.

[0061] Step 4.2: If a thrust failure occurs, disable the servo mechanism part of the extended-dimensional filter and estimate the thrust of each engine through the extended-dimensional single-engine thrust observer to determine the location and extent of the thrust failure.

[0062] Step 4.3: If there is no thrust fault, disable the thrust part expansion observer and enter the servo mechanism fault judgment.

[0063] Furthermore, the servo mechanism fault determination steps in step 5 are as follows:

[0064] Step 5.1: Servo mechanism fault judgment First, estimate the parameter error S according to the servo mechanism observer. i and covariance R i Determine the fault location, calculation method:

[0065]

[0066] where Φ(i,:) satisfies Φ=αM p , It represents the torque caused by the change of the servo mechanism's swing angle, [δ c0 ,δ c1 ,δ c2 ,δ c3 ] are the x and y direction components of the swing angles of the two servo mechanisms corresponding to one engine, R x 、R yz Respectively represent the projection of the distance between the engine and the center of mass of the aircraft on the x-axis and y-axis / z-axis (the aircraft axis is symmetrical), P c Represents the thrust of a single engine, J x 、J yz Represents the moment of inertia of the aircraft about the x-axis and y-axis / z-axis respectively;

[0067] Step 5.2: Disable the servo mechanism observer at the non-faulty position and perform the third-level judgment to obtain the swing angle estimate through the servo swing angle expansion observer. Determine the servo mechanism failure mode and failure severity.

Claims

1. A method for rapid and accurate aircraft fault identification based on extended-dimensional parallel adaptive filtering and hierarchical determination technology, characterized by: The method steps are: Step 1: Analyze the typical failure modes of the thrust system / servo mechanism according to the working status of the aircraft engine thrust system and servo mechanism, and establish a typical thrust failure model T iout =k i T i Typical servo mechanism model Where T i is the thrust corresponding to the i-th position, k i is the damage coefficient of the thrust corresponding to the i-th position, δ i (t) is the servo mechanism swing angle corresponding to the i-th position, σ i is the damage coefficient of the swing angle corresponding to the i-th position, δ ki is the angle of the swing angle corresponding to the i-th position, and w is the random value of the servo mechanism swing angle; Step 2: Analyze the aircraft state before and after the failure: x-axis displacement x, y-axis displacement y, z-axis displacement z, x-axis speed v x , y-axis speed v y , z-axis speed v z , x-axis attitude angular velocity ω x , y-axis attitude angular velocity ω y , z-axis attitude angular velocity ω z , the i-th position corresponds to the thrust T i , the jth position corresponds to the servo mechanism swing angle θ j According to the changes of fault models, multiple unbiased dynamic models with faults are established, and the system state equation and measurement equation are established. Step 3: Change the thrust of the main engine T i 、The swing angle θ of the servo mechanism of each engine j As the expanded state quantity, it is respectively expanded to the state quantity X of multiple low-dimensional filters. T =[ω x ,ω y ,ω z ,T i ], X θ =[ω x ,ω y ,ω z ,θ j ], where X T Indicates thrust T i The state quantity of the expanded dimension, X θ Indicates the servo mechanism swing angle θ j Expanded state quantity, multiple filters run simultaneously to obtain engine thrust estimation and the estimated servo mechanism swing angle Step 4: Make the first level judgment based on the estimation result of the total thrust observer. If a thrust failure occurs, disable the servo mechanism observer. T =[ω x ,ω y ,ω z ,T i ] Carry out the second level judgment to obtain the fault location and fault extent. If there is no thrust fault, disable the thrust observer and enter the servo mechanism fault judgment; Step 5: Servo mechanism fault judgment First, estimate the parameter error S according to the observer. i and covariance R i Determine the fault location, stop the servo mechanism observer at the fault location to perform the third level judgment, and expand the servo mechanism observer X θ =[ω x ,ω y ,ω z ,θ j ]Get the estimated value of the swing angle Determine the failure mode and extent of the failure.

2. The method for rapid and accurate aircraft fault identification based on dimensional expansion parallel adaptive filtering and hierarchical determination technology according to claim 1, characterized in that: In step 2, the system state equation and measurement equation are specifically: Step 21: Set the system state vector x = [ω x ω y ω z ] T The output of the system in Indicates the x-axis attitude angular velocity output, Indicates the y-axis attitude angular velocity output, Indicates the z-axis attitude angular velocity output; Step 22: Establish the state equation and measurement equation based on the rocket's center-of-mass dynamics equation and kinematics equation: Among them, A represents the system matrix, B represents the control matrix of the system, C represents the output matrix of the system, d(t) represents the disturbance vector of the system, and E d represents the disturbance transformation matrix, x(t) represents the state quantity of the system, u(t) represents the control vector of the system, and y(t) represents the quantity measurement of the system.

3. The method for rapid and accurate aircraft fault identification based on dimensional expansion parallel adaptive filtering and hierarchical determination technology according to claim 1 is characterized by: In step 3, the dimensionality augmentation steps for each filter observer are as follows: Step 31: Design the state quantity X of engine thrust expansion T =[ω x ,ω y ,ω z ,T i ]; Step 32: The system attitude dynamics model is expressed as follows: y=[I 0]x+v Where x is the state vector, f(x) is the nonlinear function vector of the state vector, B is the control matrix of the system, T is the engine nozzle thrust vector, which is the input vector of the system, T=[T1T2] T , w is the state disturbance vector of the system, y is the measurement vector of the system, v is the measurement noise vector of the system, I is the identity matrix, T1 is the thrust corresponding to the No. 1 engine, T2 is the thrust corresponding to the No. 2 engine, M1 is the torque generated by the thrust T1, and M2 is the torque generated by the thrust T2; Step 3: Design the servo mechanism to expand the state variable: X θ =[w x ,w y ,w z ,θ j ], the design of the servo mechanism expanded state equation is the same as the thrust expansion method.

4. The method for rapid and accurate aircraft fault identification based on dimensional expansion parallel adaptive filtering and hierarchical determination technology according to claim 1, characterized in that: The step 4 is specifically as follows: Step 41: At each moment, the system status and measurement updates are corrected using information input from the navigation system. Each filter observer operates normally. The total thrust observer estimates the total thrust of the aircraft and determines whether a thrust failure has occurred. If a thrust failure has occurred, proceed to Step 42; if not, proceed to Step 43. Step 42: If a thrust failure occurs, disable the servo mechanism's extended-dimensional filter and estimate the thrust of each engine using the extended-dimensional single-engine thrust observer to determine the location and severity of the thrust failure. Step 43: If there is no thrust fault, disable the thrust part expansion observer and enter the servo mechanism fault judgment.

5. The method for rapid and accurate aircraft fault identification based on dimensional expansion parallel adaptive filtering and hierarchical determination technology according to claim 1 is characterized by: The step five is specifically as follows: Step 51: Servo mechanism fault judgment First, estimate the parameter error S according to the servo mechanism observer. i and covariance R i Determine the fault location, calculation method: Where Φ is the error solution matrix, Φ(i,:) is the vector of elements in the i-th row of Φ, satisfying Where α is the matrix allocation coefficient, M p is the torque distribution matrix, It represents the torque caused by the change of the servo mechanism's swing angle, [δ c0 ,δ c1 ,δ c2 ,δ c3 ] are the x and y direction components of the swing angles of the two servo mechanisms corresponding to one engine, R x 、R yz Respectively represent the projection of the distance between the engine and the center of mass of the aircraft on the x-axis and y-axis / z-axis, P c Represents the thrust of a single engine, J x 、J yz Represents the moment of inertia of the aircraft about the x-axis and y-axis / z-axis respectively; Step 52: Disable the servo mechanism observer at the non-faulty position and perform the third-level judgment to obtain the swing angle estimate through the servo swing angle expansion observer. Determine the servo mechanism failure mode and failure severity.

Images