Self-adaptive backstepping fault-tolerant control method for aircraft wing surface faults

Abstract

The invention relates to a self-adaptive backstepping fault-tolerant control method for a transportation airfoil surface fault. The method comprises the following steps: establishing a six-degree-of-freedom equation of an airplane; in combination with a six-degree-of-freedom equation, determining a relationship between different types of faults and aerodynamic coefficients according to an airfoil fault mode; the wing surface fault mode comprises wing damage, control surface damage and control surface jamming; designing an angular velocity loop adaptive backstepping control law according to a nonlinear dynamic item about a state variable and a control input matrix about a control variable in a fault mode based on a relationship between different types of faults and aerodynamic coefficients; and performing control distribution on the angular velocity loop by using a weighted pseudo-inverse method, combining the adaptive backstepping control law of the angular velocity loop with the weighted pseudo-inverse method, and performing valuing on a weight matrix of the control distribution of the angular velocity loop according to an airfoil fault mode to obtain an adaptive backstepping fault-tolerant angular velocity control law with control distribution. According to the method, the fault-tolerant control efficiency after the airplane breaks down is optimized, and the fault-tolerant flight capability of the airplane is improved.

Description

Technical Field

[0001] The present invention belongs to the field of aviation technology, and particularly relates to an adaptive backstepping fault-tolerant control method for transport aircraft wing surface faults. Background Technique

[0002] Aircraft wing surfaces mainly include control surfaces operated by pilots and wing surfaces that provide lift for the aircraft. Compared with other types of faults, wing surface faults are often more destructive and uncertain, seriously endangering the structural integrity and aerodynamic performance of the aircraft itself, and affecting the controllability of the aircraft. There are many reasons for wing surface faults, such as accidental impacts, wear, and changes in structural characteristics. In addition, the flight speed of a transport aircraft may exceed the critical speed of flutter. At this time, the wing surface is no longer stable and divergent vibrations occur, resulting in damage or even failure of the wing surface; when the aircraft encounters strong wind shear during takeoff and approach landing phases, the forces on both wings are different, and the lift of the wing on the windward side increases rapidly compared with the other side, and accidents such as wing fracture and fuselage damage may also occur. How to effectively control the aircraft in the event of a wing surface fault is an important part of the research on fault-tolerant flight control systems.

[0003] With the continuous increase in the complexity of aircraft and the diversity of faults, it is difficult for flight controllers designed based on linear methods to achieve the desired dynamic performance throughout the flight envelope. Nonlinear control technology has become the fastest-developing control method. Among them, backstepping control, as an effective method for dealing with nonlinear problems, especially has great advantages in dealing with uncertain nonlinear flight control problems, and has been widely studied. The backstepping method can perform feedback linearization on nonlinear problems and improve the performance of the control system without changing the original characteristics of the system. For the problem of aircraft modeling uncertainty caused by faults, the backstepping method can estimate unknown dynamic characteristics through the idea of adaptive parameter estimation. In the problem of model uncertainty, flight control research dominated by backstepping is gradually showing its unique performance. Therefore, the backstepping control method is widely used in aircraft flight control and fault-tolerant control.

[0004] How to solve the fault-tolerant control of aircraft wing surface faults based on the backstepping method and improve the fault-tolerant ability of the aircraft has become an urgent problem to be solved. Summary of the Invention

[0005] In order to solve the above problems existing in the prior art, the present invention provides an adaptive backstepping fault-tolerant control method for transport aircraft wing surface faults. The technical problems to be solved by the present invention are realized through the following technical solutions:

[0006] An embodiment of the present invention provides an adaptive backstepping fault-tolerant control method for aircraft wing surface faults, including the steps:

[0007] Establish the six-degree-of-freedom equations of the aircraft;

[0008] Combined with the six-degree-of-freedom equations, determine the relationship between different types of faults and aerodynamic coefficients according to the wing surface fault modes; the wing surface fault modes include: wing damage, rudder surface damage, and rudder surface jamming;

[0009] Based on the relationship between different types of faults and aerodynamic coefficients, design an angular velocity loop adaptive backstepping control law according to the nonlinear dynamic term with respect to the state variable and the control input matrix with respect to the control variable in the fault mode;

[0010] Use the weighted pseudoinverse method to perform control allocation on the angular velocity loop, combine the angular velocity loop adaptive backstepping control law with the weighted pseudoinverse method, and obtain an adaptive backstepping fault-tolerant angular velocity control law with control allocation according to the wing surface fault mode to take values for the weight matrix of the angular velocity loop control allocation; wherein, the adaptive backstepping fault-tolerant angular velocity control law with control allocation is used to calculate the control input vector of different wing surfaces of the aircraft according to the weight matrix, the nonlinear dynamic term with respect to the state variable, and the control input matrix with respect to the control variable when the aircraft wing surface fails, so as to achieve adaptive backstepping fault-tolerant control.

[0011] In an embodiment of the present invention, the six-degree-of-freedom equations include:

[0012]

[0013]

[0014] Wherein, C D is the drag coefficient, C L is the lift coefficient, C Y is the side force coefficient, C l is the rolling moment coefficient, C m is the pitching moment coefficient, C n is the yaw moment coefficient, C D0 is the zero-lift drag coefficient, C L0 is the zero-lift coefficient, C m0 is the zero-pitching moment coefficient, C Dα is the drag derivative of the angle of attack, C Lα is the lift derivative of the angle of attack, C mα is the static stability derivative of the pitching moment, C Yβ is the side force derivative of the sideslip angle, C lβ is the static stability derivative of the roll, C nβ is the static stability derivative of the yaw, C Dq , C Lq, C mq are the derivatives of the drag D, lift L, and pitching moment m with respect to the pitch angular velocity q, respectively. C Yp , C lp , C np are the derivatives of the side force Y, rolling moment l, and yawing moment n with respect to the roll angular velocity p, respectively. C Yr , C lr , C nr are the derivatives of the side force, rolling moment, and yawing moment with respect to the yaw angular velocity r. is the control derivative of the drag with respect to each control surface δ j . is the control derivative of the lift with respect to each control surface δ j . is the control derivative of the side force with respect to each control surface δ j . is the control derivative of the rolling moment with respect to each control surface δ j . is the control derivative of the pitching moment with respect to each control surface δ j . is the control derivative of the yawing moment with respect to each control surface δ j , δ j = δ el , δ er , δ al , δ ar , δ r , δ el is the left elevator, δ er is the right elevator, δ al is the left aileron, δ ar is the right aileron, δ r is the rudder, α is the angle of attack, q is the pitch angular velocity, is the mean aerodynamic chord length, V a is the airspeed, β is the sideslip angle, p is the roll angular velocity, b is the wingspan, and r is the yaw angular velocity.

[0015] In one embodiment of the present invention, the relationship between different types of faults and aerodynamic coefficients is determined according to the wing surface fault mode, including:

[0016] The relationship between the wing damage and the aerodynamic coefficients is:

[0017] The wing surface damage is k1% to 100%, that is

[0018] The relationship between the control surface damage and the aerodynamic coefficients is:

[0019] The control surface damage is k2% to 100%, that is (i = δ el , δ er , δal , δ ar , δ r );

[0020] The relationship between the stuck control surface and the aerodynamic coefficients is as follows:

[0021] Stuck control surface That is (i = δ el , δ er , δ al , δ ar , δ r );

[0022] Where S W is the wing area, k1 is the wing damage rate, C D_WingDam is the drag coefficient after wing surface damage, C Y_WingDam is the side force coefficient after wing surface damage, C L_WingDam is the lift coefficient after wing surface damage, C l_WingDam is the rolling moment coefficient after wing surface damage, C m_WingDam is the pitching moment coefficient after wing surface damage, C n_WingDam is the yawing moment coefficient after wing surface damage, S i is the control surface area, k2 is the control surface damage rate, C L_DiDam is the lift coefficient after control surface damage, C D_DiDam is the drag coefficient after control surface damage, C m_DiDam is the pitching moment coefficient after control surface damage, δ el is the left elevator, δ er is the right elevator, δ al is the left aileron, δ ar is the right aileron, δ r is the rudder, is the stuck angle of the control surface, δ i is the deflection angle of the control surface, C Li is the lift coefficient after the control surface is stuck, C Di is the drag coefficient after the control surface is stuck, C mi is the pitching moment coefficient after the control surface is stuck.

[0023] In an embodiment of the present invention, based on the relationship between the different types of faults and the aerodynamic coefficients, an angular velocity loop adaptive backstepping control law is designed according to the non-linear dynamic terms of the state variables and the control input matrix of the control variables in the fault mode, including:

[0024] Based on the relationship between the different types of faults and the aerodynamic coefficients, construct the aircraft moment equation;

[0025] Obtain the matrix form of the angular velocity according to the aircraft moment equation, and obtain the non-linear dynamic term with respect to the state variables and the control input matrix with respect to the control variables;

[0026] Design the angular velocity loop adaptive backstepping control law according to the non-linear dynamic term with respect to the state variables and the control input matrix with respect to the control variables of the aircraft in the fault mode.

[0027] In an embodiment of the present invention, the matrix form of the angular velocity is:

[0028]

[0029] Wherein, is the derivative of the aircraft angular velocity, is the derivative of the roll angular velocity, is the derivative of the pitch angular velocity, is the derivative of the yaw angular velocity, X = [p q r] T is the angular velocity of the aircraft, p is the roll angular velocity, q is the pitch angular velocity, r is the yaw angular velocity, u = [δ el δ er δ al δ ar δ r T is the control input vector composed of the left elevator δ el , the right elevator δ er , the left aileron δ al , the right aileron δ ar , and the rudder δ r . F(X) is the non-linear dynamic term with respect to the state variables, G(X) is the control input matrix with respect to the control variables, ζ(X, u) is the change of the aircraft's aerodynamic parameters and unknown dynamic parameters when the angular velocity and the control input vector change due to faults, and y represents the output state quantity;

[0030] The non-linear dynamic term with respect to the state variables is:

[0031]

[0032] Wherein, J x is the inertia matrix, Q is the dynamic pressure, b is the wingspan, is the mean aerodynamic chord length, C lf is the set of roll moment coefficients independent of the rudder deflection, C mf is the set of pitch moment coefficients independent of the rudder deflection, C nf is the set of yaw moment coefficients independent of the rudder deflection;

[0033] The control input matrix with respect to the control variables is: ​

[0034]

[0035] Among them, S is the wing reference area, is the control derivative of the rolling moment with respect to each control surface δ j . is the control derivative of the pitching moment with respect to each control surface δ j . is the control derivative of the yawing moment with respect to each control surface δ j , δ j = δ el , δ er , δ al , δ ar , δ r , δ el is the left elevator, δ er is the right elevator, δ al is the left aileron, δ ar is the right aileron, δ r is the rudder.

[0036] In an embodiment of the present invention, the angular velocity loop adaptive backstepping control law is:

[0037]

[0038] Among them, K and Γ are diagonal matrices, X ref is the angular velocity desired command, X is the angular velocity of the aircraft, is the estimated value of the change in the aerodynamic parameters of the aircraft and the unknown dynamic parameter ζ(X,u) when the angular velocity and control input vector change due to a fault.

[0039] In an embodiment of the present invention, the weight matrix for the angular velocity loop control allocation is determined according to the wing surface fault mode, including:

[0040] When the wing surface fault mode is wing damage and the i-th control surface is lost, the weight matrix for the control allocation is:

[0041] W u = diag(k1, k2,..., 1,..., k M );

[0042] Among them, W u is the weight matrix for the control allocation, 0 < k1, k2,..., k M < 1, k1, k2,..., k M are the weights of different control surfaces, and M is the number of control surfaces;

[0043] When the wing surface fault mode is control surface damage, the weight matrix for the control allocation is:

[0044] W u = diag(1, 1, …, k i , …, 1)

[0045] where ξ is the damage size of the control surface, 0 < ξ < 1, k i is the weight of the damaged control surface i, k i > 1;

[0046] When the wing surface fault mode is control surface jamming, the weight matrix of the control allocation is:

[0047] W u = diag(k1, k2, …, 1, …, k M )

[0048] where 0 < k1, k2, …, k M < 1, k1, k2, ..., k M are the weights of different control surfaces, and M is the number of control surfaces.

[0049] In an embodiment of the present invention, using the weighted pseudo-inverse method to perform control allocation for the angular velocity loop includes:

[0050] Using the weighted pseudo-inverse method to formulate the control allocation problem of the angular velocity loop as:

[0051]

[0052] subject to B(x)u = v ref

[0053] where Ω is the solution space of the control input, W u is the weight allocation matrix and is invertible, u is the output value of the aircraft control surface, u c is the desired control input, B(x) is the mapping relationship between the virtual control quantity command v ref and the control input to the virtual control quantity, B(x): R M → R r (M > r), R M is the M-dimensional real space, R r is the r-dimensional real space, M is the number of control surfaces, and r is the dimension of the virtual control quantity.

[0054] In an embodiment of the present invention, the adaptive backstepping fault-tolerant angular velocity control law with control allocation is:

[0055]

[0056] where u is composed of the left elevator δ el , the right elevator δ er , the left aileron δal and the right aileron δ ar and the rudder δ r to form a control input vector u = [δ el δ er δ al δ ar δ r T , where W u is the control allocation weight matrix, G(X) is the control input matrix with respect to the control variables, K and Γ are diagonal matrices, X ref is the angular velocity desired command, X is the angular velocity, and F(X) is the nonlinear dynamic term with respect to the state variables. is the estimated value of ζ(X, u) which represents the changes in the aircraft's aerodynamic parameters and unknown dynamic parameters when the angular velocity and control input vector change due to faults. is 's Moore - Penrose inverse. Taking the matrix G(X) as an example, T(X) is the Moore - Penrose inverse of G(X), and T(X) = G(X) T (G(X)G(X) T ) -1 .

[0057] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0058] The present invention combines the angular velocity loop adaptive backstepping control law with the weighted pseudoinverse method to obtain an adaptive backstepping fault - tolerant angular velocity control law with control allocation. The angular velocity loop adaptive backstepping control law can compensate for the changes in the aircraft's aerodynamic parameters caused by faults, and the weighted pseudoinverse method is used for angular velocity loop control allocation to compensate for the effectiveness of the faulty control surface. As a result, the aircraft can meet the requirements of the flight mission in the case of wing surface faults, realizing the flight of the transport aircraft in the wing surface fault environment, optimizing the fault - tolerant control efficiency after the aircraft fails, improving the fault - tolerant flight ability of the aircraft, and enhancing the reliability and survivability of the transport aircraft. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 is a schematic flow chart of an adaptive backstepping fault - tolerant control method for wing surface faults of a transport aircraft provided by an embodiment of the present invention;

[0060] Figure 2 is a structural block diagram of an adaptive backstepping fault - tolerant control method for wing surface faults of a transport aircraft provided by an embodiment of the present invention;

[0061] Figure 3 is a schematic diagram of the main control surfaces of a certain large - scale transport aircraft provided by an embodiment of the present invention;

[0062] Figure 4 ​It is the structural block diagram of the angular velocity loop adaptive backstepping controller provided by the embodiment of the present invention;

[0063] Figures 5a - 5h It is the simulation result diagram of the response of the aircraft with damaged left wing under the pitch angular velocity command;

[0064] Figures 6a - 6h It is the simulation result diagram of the response of the aircraft with a fault in the left elevator under the pitch angular velocity command. Specific embodiments

[0065] The following further describes the present invention in detail with reference to specific embodiments, but the implementation manners of the present invention are not limited thereto.

[0066] Embodiment 1

[0067] In this embodiment, a certain transport aircraft is taken as the research object to design an adaptive backstepping fault-tolerant control method for the wing surface faults of the transport aircraft. Refer to Figure 1 and Figure 2 , Figure 1 It is the flow schematic diagram of an adaptive backstepping fault-tolerant control method for the wing surface faults of a transport aircraft provided by the embodiment of the present invention, Figure 2 It is the structural block diagram of an adaptive backstepping fault-tolerant control method for the wing surface faults of a transport aircraft provided by the embodiment of the present invention. When designing the adaptive backstepping fault-tolerant control method, under normal circumstances, the control law controls the normal flight of the aircraft. When simulating the aircraft fault, the fault injection module triggers the wing surface fault of the aircraft, brings the aerodynamic parameters of the fault aircraft into the aircraft object, simulates the aircraft under the fault condition, and the aerodynamic parameters used in the control law design remain the parameters of the normal aircraft. Through the adaptive adjustment of the adaptive backstepping control law parameter adaptive module, the fault-tolerant flight control design analysis of the fault aircraft is completed.

[0068] The adaptive backstepping fault-tolerant control method for the wing surface faults of the transport aircraft in this embodiment specifically includes the following steps:

[0069] S1. Establish the six-degree-of-freedom equation of the aircraft.

[0070] Specifically, the aerodynamic coefficients of the transport aircraft include three force coefficients and three moment coefficients. The three force coefficients are respectively: drag coefficient C D , lift coefficient C L and side force coefficient C Y , and the force coefficients are given in the airflow coordinate system (wind axis system). The three moment coefficients are respectively roll moment coefficient C l , pitch moment coefficient C m , yaw moment coefficient C n . Starting from the relationship between the force coefficients and the control surfaces, and the moment coefficients and the control surfaces, the three force coefficients and the three moment coefficients can be disassembled to obtain the six-degree-of-freedom equation of the aircraft:

[0071]

[0072] Among them, C D is the drag coefficient, C L is the lift coefficient, C Y is the side force coefficient, C l is the rolling moment coefficient, C m is the pitching moment coefficient, C n is the yawing moment coefficient, C D0 is the zero-drag coefficient, C L0 is the zero-lift coefficient, C m0 is the zero-pitching moment coefficient, C Dα is the drag derivative with respect to the angle of attack, C Lα is the lift derivative with respect to the angle of attack, C mα is the pitching moment static stability derivative, C Yβ is the side force derivative with respect to the sideslip angle, C lβ is the rolling static stability derivative, C nβ is the yawing static stability derivative, C Dq 、C Lq 、C mq are the derivatives of the drag D, lift L, and pitching moment m with respect to the pitching angular velocity q respectively, C Yp 、C lp 、C np are the derivatives of the side force Y, rolling moment l, and yawing moment n with respect to the rolling angular velocity p respectively, C Yr 、C lr 、C nr are the derivatives of the side force, rolling moment, and yawing moment with respect to the yaw angular velocity r, is the control derivative of the drag with respect to each control surface δ j , is the control derivative of the lift with respect to each control surface δ j , is the control derivative of the side force with respect to each control surface δ j , is the control derivative of the rolling moment with respect to each control surface δ j , is the control derivative of the pitching moment with respect to each control surface δ j , is the control derivative of the yawing moment with respect to each control surface δ j ,δ j =δ el ,δ er ,δ al ,δ ar ,δ r ,δ el is the left elevator, δ er is the right elevator, δ al is the left aileron, δar is the right aileron, δ r is the rudder, α is the angle of attack, is the mean aerodynamic chord, b is the wingspan, V a is the airspeed, β is the sideslip angle, p is the roll angular velocity, q is the pitch angular velocity, r is the yaw angular velocity.

[0073] S2. Combine the six - degree - of - freedom equations to determine the relationship between different types of faults and aerodynamic coefficients according to the wing surface fault modes; the wing surface fault modes include: wing damage, control surface damage, and control surface jamming.

[0074] In this embodiment, the aircraft wing surface faults are divided into two categories: one is the fixed wing surface fault, that is, the wing surface represented by the wing is damaged; the other is the control surface damage and control surface jamming faults.

[0075] As the main component of the aircraft to generate aerodynamic forces and moments, the wing may cause varying degrees of changes in the wing shape when suffering damage faults, which in turn affect the air flow direction and flow characteristics. This influence is manifested as changes in aerodynamic parameters on the aircraft, affecting the aerodynamic performance, and at the same time causing changes in the aircraft mass and center of gravity. Ultimately, these factors will significantly affect the aircraft's motion characteristics, and in severe cases, may lead to the aircraft losing effective control and causing flight accidents.

[0076] Control surface faults are also one of the main inducements of flight accidents. In addition to directly affecting the aircraft's handling performance, it also changes the aircraft's flight dynamics characteristics, resulting in a decrease in aircraft stability and a reduction in the flight envelope. According to the fault characteristics, the control surface faults can be further divided into control surface jamming and control surface damage (missing) faults. Control surface jamming means that the actuator is stuck at a certain fixed position and loses control, resulting in no response to the input command and generating immutable forces and moments on the control surface. The jamming fault does not damage the aircraft's external shape structure, so the influence on the aircraft's aerodynamic characteristics can be ignored. Control surface damage will cause changes in the aircraft's external shape to a certain extent, thereby affecting the aircraft's aerodynamic force and moment coefficients, and the degree of reduction in control surface effectiveness depends on the degree of damage.

[0077] Please refer to Figure 3 , Figure 3 which is the schematic diagram of the main control surfaces of a certain large - transport aircraft provided by the embodiment of the present invention. Figure 3 The aerodynamic shape of the transport aircraft is designed and completed by the three - dimensional modeling software CATIA. The main control surfaces of the transport aircraft include two elevators, two ailerons, and one rudder. When calculating the aerodynamic derivatives and analyzing the influence of faults on the aircraft aerodynamics in the follow - up, three types of wing surface fault types including wing damage, control surface damage, and control surface jamming are considered, where the control surfaces include elevators, rudders, and ailerons.

[0078] Specifically, define the aircraft wing damage, that is, the wing areas S on the left and right sides of the aircraftW Damage. The aerodynamic coefficient and aerodynamic moment coefficient after wing damage are calculated by the computational fluid dynamics (CFD) numerical calculation and analysis software. According to the calculation and analysis of forces and moments in the aircraft model, assuming the wing damage rate is k1% - 100%, the relationship between wing damage and aerodynamic coefficients can be obtained as follows:

[0079] Wing surface damage is k1% - 100%, that is

[0080] where S W is the wing area, k1 is the wing damage rate, C D_WingDam is the drag coefficient after wing surface damage, C Y_WingDam is the side force coefficient after wing surface damage, C L_WingDam is the lift coefficient after wing surface damage, C l_WingDam is the rolling moment coefficient after wing surface damage, C m_WingDam is the pitching moment coefficient after wing surface damage, C n_WingDam is the yawing moment coefficient after wing surface damage.

[0081] Define the damage of the aircraft control surface, that is, the area damage S el of the aircraft control surfaces δ er 、δ al 、δ ar 、δ r 、δ i . The aerodynamic coefficient and aerodynamic moment coefficient after control surface damage are calculated by the CFD numerical calculation and analysis software. Control surface damage not only affects the force and moment system of the aircraft, but also affects the control effectiveness of the moment generated by the control surface. According to the analysis of forces and moments, assuming the control surface damage rate is k2% - 100%, the relationship between control surface damage and aerodynamic coefficients can be obtained as follows:

[0082] Control surface damage is k2% - 100%, that is

[0083] where S i is the control surface area, k2 is the control surface damage rate, C L_DiDam is the lift coefficient after control surface damage, C D_DiDam is the drag coefficient after control surface damage, C m_DiDam is the pitching moment coefficient after control surface damage, δ el is the left elevator, δ er is the right elevator, δ al is the left aileron, δ ar is the right aileron, δ r is the rudder.

[0084] Define the control surface jamming fault. Control surface jamming has no effect on the aerodynamic layout of the aircraft. The deflection angle of the control surface will remain constant due to jamming, that is:

[0085]

[0086] Among them, δ i is the rudder surface deflection angle, i = δ el , δ er , δ al , δ ar , δ r , const1 is a constant value, and k3 is the rudder surface jamming angle.

[0087] If the rudder surface is jammed at k3, the current effective rudder surface area is the projection of the actual rudder surface area on the horizontal plane. The current effective rudder surface area is:

[0088]

[0089] Among them, S new is the current effective rudder surface area, S is the actual rudder surface area, and const2 is a constant value.

[0090] It can be seen from the above formula that the essence of the rudder surface jamming is that the input signal corresponding to the jammed rudder surface and the effective area of the rudder surface are a certain constant value. According to the analysis, the relationship between the rudder surface jamming and the aerodynamic coefficient is as follows:

[0091] Rudder surface jamming That is

[0092] Among them, C Li is the lift coefficient after the rudder surface is jammed, C Di is the drag coefficient after the rudder surface is jammed, C mi is the pitching moment coefficient after the rudder surface is jammed, δ el is the left elevator, δ er is the right elevator, δ al is the left aileron, δ ar is the right aileron, δ r is the rudder.

[0093] S3. Based on the relationship between different types of faults and aerodynamic coefficients, design the angular velocity loop adaptive backstepping control law according to the nonlinear dynamic terms of the state variables and the control input matrix of the control variables in the fault mode.

[0094] Angular velocity control is a basic control law of the flight control system, and its performance directly determines the handling quality of the aircraft. An ideal angular velocity control law should ensure that the angular velocity of the aircraft can accurately and stably track the pilot's commands in the face of various uncertain disturbances, thereby ensuring the maneuverability and stability of the aircraft. Therefore, based on the changes in aerodynamic parameters under fault modes, this embodiment analyzes the disturbance range caused by faults according to the aerodynamic parameters calculated by CFD, and uses this as the design constraint of the control system to design an adaptive backstepping control law for the angular velocity loop, ensuring that when the aerodynamic parameters are disturbed by faults, the angular velocity of the aircraft can accurately and stably track the pilot's commands, thereby ensuring the maneuverability and stability of the aircraft.

[0095] Step S3 specifically includes:

[0096] S31. Based on the relationship between different types of faults and aerodynamic coefficients, construct the aircraft moment equation.

[0097] Specifically, in combination with the relationship between different types of faults and aerodynamic coefficients, use CFD numerical calculation and analysis software to solve the aerodynamic data of the aircraft under normal conditions and wing surface fault conditions.

[0098] Based on the solved aerodynamic data, construct the aircraft moment equation as follows:

[0099]

[0100] Where p is the roll angular velocity, q is the pitch angular velocity, r is the yaw angular velocity, l is the roll moment, m is the pitch moment, n is the yaw moment,

[0101] I x is the moment of inertia about the x-axis, I y is the moment of inertia about the y-axis, I z is the moment of inertia about the z-axis, I xz is the xz product of inertia.

[0102] S32. Obtain the angular velocity matrix form according to the aircraft moment equation, and obtain the nonlinear dynamic term about the state variables and the control input matrix about the control variables.

[0103] Specifically, under the condition of considering the influence of faults, the angular velocity matrix form obtained according to the aircraft moment equation is as follows:

[0104]

[0105] Where, is the derivative of the aircraft angular velocity, is the derivative of the roll angular velocity, is the derivative of the pitch angular velocity, is the derivative of yaw angular velocity, X = [p q r] T is the angular velocity of the aircraft, p is the roll angular velocity, q is the pitch angular velocity, r is the yaw angular velocity, u = [δ el δ er δ al δ ar δ r T is the control input vector composed of the left elevator δ el , the right elevator δ er , the left aileron δ al , the right aileron δ ar , and the rudder δ r . F(X) is the nonlinear dynamic term with respect to the state variables, G(X) is the control input matrix with respect to the control variables, K and Γ are diagonal matrices, ζ(X, u) is the change in the aerodynamic parameters of the aircraft and the unknown dynamic parameters when the angular velocity and the control input vector change due to a fault, and y represents the system output. It should be noted that all the aerodynamic parameters and mass parameters in the above equation are consistent with the real-time state of the aircraft.

[0106] The nonlinear dynamic term F(X) with respect to the state variables is as follows:

[0107]

[0108] where J x is the inertia matrix, Q is the dynamic pressure, b is the wingspan, is the mean aerodynamic chord length, C lf is the set of roll moment coefficients independent of the rudder deflection, C mf is the set of pitch moment coefficients independent of the rudder deflection, C nf is the set of yaw moment coefficients independent of the rudder deflection.

[0109] The control input matrix G(X) with respect to the control variables is as follows:

[0110]

[0111] where S is the wing area, is the control derivative of the roll moment with respect to each rudder δ j , is the control derivative of the pitch moment with respect to each rudder δ j , is the control derivative of the yaw moment with respect to each rudder δ j , δ j = δ el , δ er , δ al , δ ar , δ​r , δ el is the left elevator, δ er is the right elevator, δ al is the left aileron, δ ar is the right aileron, δ r is the rudder.

[0112] S33. Design an adaptive backstepping control law for the angular velocity loop based on the nonlinear dynamic terms of the state variables and the control input matrix of the control variables in the fault mode of the aircraft, and form an adaptive backstepping controller for the angular velocity loop.

[0113] Specifically, the design of the adaptive backstepping control law for the angular velocity loop is as follows:

[0114]

[0115] Specifically expanded as:

[0116]

[0117] Among them, K and Γ are diagonal matrices, K = diag(150, 150, 150), Γ = diag(50, 50, 50), X ref is the desired command of the angular velocity, X is the angular velocity of the aircraft, is the estimated value of the aerodynamic parameter change of the aircraft and the unknown dynamic parameter ζ(X, u) when the angular velocity and the control input vector change due to the fault, is the estimated value of the change rate of ζ(X, u), S is the wing reference area, is the control derivative of the rolling moment with respect to each control surface δ j of, is the control derivative of the pitching moment with respect to each control surface δ j of, is the control derivative of the yawing moment with respect to each control surface δ j of, δ j = δ el , δ er , δ al , δ ar , δ r , δ el is the left elevator, δ er is the right elevator, δ al is the left aileron, δ ar is the right aileron, δ r is the rudder, J x is the inertia matrix, Q is the dynamic pressure, b is the wingspan, is the mean aerodynamic chord length, C lf is the set of rolling moment coefficients independent of the control surface deflection, C mfis the set of pitch moment coefficients independent of aileron deflection, C nf is the set of yaw moment coefficients independent of aileron deflection, p is the roll angular velocity, q is the pitch angular velocity, r is the yaw angular velocity, p ref is the desired command of roll angular velocity, q ref is the desired command of pitch angular velocity, r ref is the desired command of yaw angular velocity.

[0118] Please refer to Figure 4 , Figure 4 which is the structural block diagram of the angular velocity loop adaptive backstepping controller provided by the embodiment of the present invention. The structural diagram includes command signals, command filtering, and the angular velocity loop adaptive backstepping control law formed by the adaptation law and the control law. Specifically, after receiving the command signal X = [p q r] T , the angular velocity loop adaptive backstepping controller performs command filtering to obtain the filtered signal X ref = [p q r] T ; and based on the command signal X = [p q r] T , it obtains F(X), G(X), and Then, according to the angular velocity loop adaptive backstepping control law of formula (16), signal calculation is performed. After the calculated signal passes through the servo, the output is distributed to the control input vector u = [δ el δ er δ al δ ar δ r T , and then the five groups of ailerons drive the aircraft to change its state.

[0119] S4. Use the weighted pseudoinverse method to perform control allocation for the angular velocity loop, combine the angular velocity loop adaptive backstepping control law with the weighted pseudoinverse method, and obtain the adaptive backstepping fault-tolerant angular velocity control law with control allocation according to the wing surface fault mode to take values for the weight matrix of the angular velocity loop control allocation.

[0120] When a fault occurs in the aircraft's actuator, the control allocation module can reasonably allocate control commands according to the fault information output by the fault diagnosis module, that is, allocate the deflection command of the failed aileron to other ailerons, so as to ensure the normal flight of the aircraft.

[0121] Considering the controlled object and the influence of faults, in this embodiment, the weighted pseudoinverse method is used to perform control allocation for the angular velocity loop of the aircraft. The control allocation problem of the angular velocity loop of the flight control system is described as: assuming that the output value u of the aircraft aileron belongs to R M , for the given virtual control quantity command v ref and the mapping relationship B(x) from the control input to the virtual control quantity: R​M →R r (M > r), solve the indefinite equation Bu = v ref and make u satisfy the desired performance index, where R M represents the M-dimensional real space, and R r represents the r-dimensional real space, M represents the number of control surfaces, and r represents the dimension of the virtual control quantity. The entire control allocation problem should satisfy the form:

[0122] B(x)u = v ref ; (18)

[0123] The weighted pseudoinverse method summarizes the control allocation problem of the angular velocity loop in the following form with an optimization idea:

[0124]

[0125] where Ω is the solution space of the control input, W u is the weight distribution matrix and is invertible, u is the output value of the aircraft control surface, u c is the desired input value, B(x) is the mapping relationship between the virtual control quantity command v ref and the control input to the virtual control quantity, R M represents the M-dimensional real space, and R r represents the r-dimensional real space, M is the number of control surfaces, and r is the dimension of the virtual control quantity.

[0126] Let u c = 0, and the allocation solution of the control allocation problem is:

[0127]

[0128] where u is the control input vector composed of the left elevator δ el , right elevator δ er , left aileron δ al , right aileron δ ar , and rudder δ r , u = [δ el δ er δ al δ ar δ r T , W u is the control allocation weight matrix, and G(X) is the control input matrix with respect to the control variables.

[0129] Equation (20) is the allocation solution of the weighted pseudoinverse method. It can be seen that each control surface of the aircraft participates in the control allocation strategy, and the final allocation result will be affected by the weight matrix W u and G(X).

[0130] The weight matrix W​u is a diagonal matrix of order M, and the values on the diagonal can be selected with different weights according to the uses and usage frequencies of each control surface, as well as the results output by the fault diagnosis module.

[0131] Take the state vector X of the angular velocity loop as X = [p q r] T , u = [δ el δ er δ al δ ar δ r T , and the specific form of the virtual control quantity instruction v ref is the manipulation moment deviation of the aircraft's three-axis channels, and G(X) is a control effectiveness matrix with a rank equal to 3.

[0132] Furthermore, from the angular velocity loop adaptive backstepping control law of formula (16), it can be obtained that:

[0133]

[0134] where K and Γ are diagonal matrices, X ref is the angular velocity desired instruction, X is the angular velocity, F(X) is the nonlinear dynamic term with respect to the state variables, is the estimated value of ζ(X, u) of the aircraft's aerodynamic parameter changes and unknown dynamic parameters when the angular velocity and control input vector change due to faults. Then formula (21) is transformed into:

[0135]

[0136] Comparing formula (22) and formula (18), it can be seen that the angular velocity loop adaptive backstepping control law has the same form as the control allocation problem after deformation. Thus, it can be seen that adaptive Backstepping and control allocation can be closely combined, that is, various methods and conclusions in the control allocation theory are still applicable in the design of the adaptive Backstepping control law. Combining the weighted pseudoinverse method with the angular velocity loop adaptive Backstepping, an adaptive backstepping fault-tolerant angular rate control law with control allocation is obtained, thereby forming an adaptive backstepping fault-tolerant angular rate controller with control allocation, as shown in Figure 1 the left block diagram in. Among them, the adaptive backstepping fault-tolerant angular rate control law with control allocation is:

[0137]

[0138]

[0139] where u is composed of the left elevator δ el , the right elevator δ er , the left aileron δal , the right aileron δ ar , the rudder δ r to form a control input vector, u = [δ el δ er δ al δ ar δ r T , W u is the control allocation weight matrix, G(X) is the control input matrix with respect to the control variables, K and Γ are diagonal matrices, X ref is the angular velocity desired command, X is the angular velocity, F(X) is the non-linear dynamic term with respect to the state variables, is the estimated value of ζ(X, u) of the aircraft's aerodynamic parameter changes and unknown dynamic parameters when the angular velocity and control input vector change due to faults, is 's Moore-Penrose inverse. Taking the matrix G(X) as an example, T(X) is the Moore-Penrose inverse of G(X), and there is T(X) = G(X) T (G(X)G(X) T ) -1 .

[0140] Furthermore, the adaptive backstepping fault-tolerant angular rate control law with control allocation is affected by the weight matrix W u . In this embodiment, the weight matrix W u = diag(k el , k er , k al , k ar , k r ) is the identity matrix when the aircraft is normal, that is, the coefficient of each control surface is 1; after a fault occurs, the coefficients of the weight matrix are re-valued according to the results of fault diagnosis, specifically as follows:

[0141] 1) The type of wing surface fault is wing damage.

[0142] In addition to causing changes in F(X) and G(X), wing damage may also cause the control surfaces on the wing to be lost together with the wing. When the fault diagnosis module outputs that the aircraft has a wing damage fault and the i-th control surface is lost, the control input of the aircraft becomes u = [δ1 δ2... 0... δ M T , δ1, δ2,..., δ M are the M control surfaces of the aircraft. At this time, the control allocation weight matrix becomes according to the fault information:

[0143] W u = diag(k1, k2,..., 1,..., k M ); (25)​​

[0144] Among them, k1, k2, ..., k M are the weights of different control surfaces, M is the number of aircraft control surfaces, and 0 < k1, k2, …, k M < 1.

[0145] It can be understood that when the wing is damaged and the ith control surface is lost, while keeping the coefficient of the lost control surface unchanged, the coefficient corresponding to the normal control surface is reduced, so as to increase the deflection of the normal control surface, achieving the purpose of increasing the usage frequency of the normal control surface and increasing the weight of the normal control surface.

[0146] 2) The type of wing surface failure is control surface damage.

[0147] Control surface damage will lead to a decrease in control efficiency. After the ith control surface is damaged, the fault diagnosis module outputs diagnostic information: damaged control surface i and the size of the control surface damage ξ (0 < ξ < 1). At this time, the control input of the aircraft becomes u = [δ1 δ2 … ξδ i …δ M T , δ i is the ith control surface of the aircraft. Using this information, the value of the weight matrix is changed to:

[0148] W u = diag(1, 1, …, k i , …, 1); (26)

[0149] Among them, ξ is the size of the control surface damage, 0 < ξ < 1, k i is the weight of the damaged control surface i, and k i > 1.

[0150] It can be understood that when the control surface is damaged, while keeping the coefficient of the normal control surface unchanged, the coefficient corresponding to the faulty control surface is increased, thereby reducing the weight of the damaged control surface and reducing the use of the damaged control surface.

[0151] 3) The type of wing surface failure is control surface jamming.

[0152] When a certain control surface has a jamming fault, it cannot produce the expected control result and will also generate negative additional forces and torques. After the ith control surface jams, the fault diagnosis module outputs diagnostic information: jammed control surface i and the size of the jam δ μ , and at this time, the control input of the aircraft becomes u = [δ1 δ2 … δ μ …δ M T , and using this information, the value of the weight matrix is changed to:

[0153] W u = diag(k1, k2, …, 1, …, k M ); (27)​​

[0154] where 0 < k1, k2, …, k M < 1, and k1, k2, …, k M are the weight values of M control surfaces of the aircraft, and M is the number of control surfaces of the aircraft.

[0155] It can be understood that when a control surface is jammed, the distribution control principle is the same as that when the wing is damaged and the control surface is lost, that is, the usage frequency of normal control surfaces is increased and the weights of normal control surfaces are increased.

[0156] Furthermore, when an aircraft wing surface fails, according to the weight matrix W u , the nonlinear dynamic term F(X) with respect to the state variables and the control input matrix G(X) with respect to the control variables, the control input vectors of different wing surfaces of the aircraft are calculated by using the adaptive backstepping fault-tolerant angular velocity control law with control allocation in formulas (24) and (25), so as to achieve adaptive backstepping fault-tolerant control.

[0157] In order to verify the effectiveness of the above model and algorithm, in this embodiment, the control performance and robustness of the adaptive backstepping fault-tolerant angular velocity controller with control allocation are verified under three types of typical faults (wing damage, control surface damage, control surface jamming).

[0158] Please combine with Figure 2 , first conduct a simulation analysis of the normal aircraft control algorithm, and then substitute the aerodynamic parameters of the faulty aircraft into the aircraft object to simulate the aircraft under actual fault conditions. The aerodynamic parameters used in the control law design remain the parameters of the normal aircraft, and the fault-tolerant flight control design analysis of the faulty aircraft is completed through the adaptive adjustment of the backstepping control law parameters.

[0159] 1) Simulation of wing damage fault-tolerant control.

[0160] The simulation time is 20 seconds. After 11 seconds from the start of the simulation, a left wing damage fault of 40% magnitude is injected into the aircraft. At this time, the aileron on the left wing completely falls off and lasts until the end of the simulation. A 5° pitch angular velocity square wave signal command is given, and the adaptive backstepping fault-tolerant angular velocity controller with control allocation is used to control different wing surfaces of the aircraft. Among them, the weight matrix of the control allocation is:

[0161]

[0162] Please refer to Figures 5a - 5h , Figures 5a - 5h which is the simulation result diagram of the response of the aircraft with left wing damage under the pitch angular velocity command. Among them, ABS represents adaptive Backstepping control, Figure 5a is the pitch angular velocity response diagram, Figure 5b is the angle of attack response diagram, Figure 5c is the pitch angle response diagram,Figure 5d is the left elevator response diagram, Figure 5e is the right elevator response diagram, Figure 5f is the left aileron response diagram, Figure 5g is the right aileron response diagram, Figure 5h is the rudder response diagram. It can be seen from Figures 5a - 5h that the adaptive backstepping control law can achieve the reconstruction of wing damage, the angular velocity of the aircraft can be restored to a stable state, and the expected dynamic performance can still be achieved under wing damage. Under the action of control allocation, the deflection of the normal control surfaces except the detached control surface increases, thereby offsetting the moment change caused by the left wing failure, achieving the effects of improving control performance, faster fault recovery and tracking commands.

[0163] 2) Elevator fault-tolerant control simulation.

[0164] After 2 seconds of the start of the simulation, a left elevator damage fault with a magnitude of 40% is injected into the aircraft for 8 s; after 14 s, a left elevator jamming fault is injected, and the jamming position is 8°, lasting until the end of the simulation. A 5° pitch angular velocity square wave signal command is given, and the adaptive backstepping fault-tolerant angular velocity controller with control allocation is used to control different wing surfaces of the aircraft. Among them, the weight matrix of control allocation is:

[0165]

[0166] Please refer to Figures 6a - 6h , Figures 6a - 6h is the simulation result diagram of the response of the aircraft with a left elevator fault under the pitch angular velocity command. Among them, Figure 6a is the pitch angular velocity response diagram, Figure 6b is the angle of attack response diagram, Figure 6c is the pitch angle response diagram, Figure 6d is the left elevator response diagram, Figure 6e is the right elevator response diagram, Figure 6f is the left aileron response diagram, Figure 6g is the right aileron response diagram, Figure 6h is the rudder response diagram. It can be seen from Figures 6a - 6h that the adaptive backstepping control method can achieve the reconstruction of elevator faults. Under its action, when the elevator fails, the q that is most affected by the elevator has a large tracking error in control in a short time, but it does not diverge. After a period of transition, it returns to the normal control state. The deflection of the control surface under the controller is smooth without oscillation. Under the action of control allocation, the aircraft uses the faulty left elevator less and instead increases the deflection of the normal control surfaces to offset the moment change caused by the left elevator fault, achieving the effects of improving control performance, faster fault recovery and tracking commands.

[0167] In this embodiment, an adaptive backstepping fault-tolerant angular velocity control law with control allocation is obtained by combining the angular velocity loop adaptive backstepping control law and the weighted pseudoinverse method. The angular velocity loop adaptive backstepping control law can compensate for the changes in the aircraft's aerodynamic parameters caused by faults, and the weighted pseudoinverse method is used for control allocation in the angular velocity loop to compensate for the effectiveness of the faulty control surface. Thus, the aircraft can meet the requirements of the flight mission in the case of wing surface faults, realizing the flight of the transport aircraft in the wing surface fault environment, optimizing the fault-tolerant control efficiency after the aircraft fails, improving the fault-tolerant flight ability of the aircraft, and enhancing the reliability and survivability of the transport aircraft.

[0168] The above content is a further detailed description of the present invention in combination with specific preferred embodiments, and it cannot be determined that the specific implementation of the present invention is only limited to these descriptions. For those of ordinary skill in the technical field to which the present invention pertains, without departing from the concept of the present invention, several simple deductions or substitutions can be made, and all should be regarded as belonging to the protection scope of the present invention.

Claims

1. An adaptive backstepping fault-tolerant control method for transport aircraft wing surface faults, characterized in that: Includes steps: Establish the six-degree-of-freedom equations of the aircraft; In combination with the six-degree-of-freedom equation, the relationship between different types of failures and aerodynamic coefficients is determined according to the airfoil failure mode; The wing surface failure modes include: wing damage, control surface damage and control surface jamming; Based on the relationship between the different types of faults and aerodynamic coefficients, an angular velocity loop adaptive backstepping control law is designed according to the nonlinear dynamic terms of the state variables and the control input matrix of the control variables under the fault mode; The weighted pseudo-inverse method is used to control the angular velocity loop, the angular velocity loop adaptive backstepping control law is combined with the weighted pseudo-inverse method, and the weight matrix of the angular velocity loop control allocation is valued according to the wing surface failure mode to obtain an adaptive backstepping fault-tolerant angular velocity control law with control allocation; wherein the adaptive backstepping fault-tolerant angular velocity control law with control allocation is used to calculate the control input vectors of different wing surfaces of the aircraft according to the weight matrix, the nonlinear dynamic terms about the state variables and the control input matrix about the control variables when the aircraft wing surface fails, so as to realize adaptive backstepping fault-tolerant control.

2. The adaptive backstepping fault-tolerant control method for transport aircraft wing surface faults according to claim 1, characterized in that: The six-degree-of-freedom equations include: Among them, C D is the drag coefficient, C L is the lift coefficient, C Y is the lateral force coefficient, C l is the rolling moment coefficient, C m is the pitch moment coefficient, C n is the yaw moment coefficient, C D0 is the zero lift drag coefficient, C L0 is the zero lift coefficient, C m0 is the zero pitch moment coefficient, C Dα is the angle of attack drag derivative, C Lα is the lift derivative of the angle of attack, C mα is the static stability derivative of the pitching moment, C Yβ is the side force derivative of the sideslip angle, C lβ is the rolling static stability derivative, C nβ is the yaw static stability derivative, C Dq , C Lq , C mq They are respectively the derivatives of the drag force D, lift force L, and pitch moment m with respect to the pitch angular velocity q, C Yp , C lp , C np They are the derivatives of the lateral force Y, rolling moment l, and yaw moment n with respect to the rolling angular velocity p, respectively. Yr , C lr , C nr is the derivative of the side force, rolling moment and yaw moment with respect to the yaw angular velocity r, is the resistance about each rudder surface δ j The manipulated derivative of is the lift with respect to each control surface δ j The manipulated derivative of is the side force about each rudder surface δ j The manipulated derivative of is the rolling moment about each rudder surface δ j The manipulated derivative of is the pitch moment about each rudder surface δ j The manipulated derivative of is the yaw moment about each rudder surface δ j The manipulated derivative, δ j =δ el ,δ er ,δ al ,δ ar ,δ r , δ el is the left elevator, δ er is the right elevator, δ al is the left aileron, δ ar is the right aileron, δ r is the rudder, α is the angle of attack, q is the pitch angular velocity, is the average aerodynamic chord length, V a is the airspeed, β is the sideslip angle, p is the roll angular velocity, b is the wingspan, and r is the yaw angular velocity.

3. The adaptive backstepping fault-tolerant control method for transport aircraft wing surface faults according to claim 1, characterized in that: According to the airfoil failure mode, the relationship between different types of failures and aerodynamic coefficients is determined, including: The relationship between the wing damage and the aerodynamic coefficient is: Wing surface damage k1%~100%, that is The relationship between the rudder surface damage and the aerodynamic coefficient is: Rudder surface damage k2%~100%, that is The relationship between the rudder surface stuck and the aerodynamic coefficient is: The rudder is stuck at k3, i.e. Among them, S W is the wing area, k1 is the wing damage rate, C D_WingDam is the drag coefficient after the wing is damaged, C Y_WingDam is the side force coefficient after the wing is damaged, C L_WingDam is the lift coefficient after the wing is damaged, C l_WingDam is the rolling moment coefficient after the wing is damaged, C m_WingDam is the pitching moment coefficient after the wing is damaged, C n_WingDam is the yaw moment coefficient after the airfoil is damaged, S i is the rudder surface area, k2 is the rudder surface damage rate, C L_DiDam is the lift coefficient after the rudder surface is damaged, C D_DiDam is the drag coefficient after the rudder is damaged, C m_DiDam is the pitching moment coefficient after the rudder is damaged, δ el is the left elevator, δ er is the right elevator, δ al is the left aileron, δ ar is the right aileron, δ r is the rudder, k3° is the rudder stuck angle, δ i is the rudder deflection angle, C Li is the lift coefficient after the rudder surface is stuck, C Di is the drag coefficient after the rudder is stuck, C mi is the pitching moment coefficient after the rudder surface is stuck.

4. The adaptive backstepping fault-tolerant control method for transport aircraft wing surface faults according to claim 1, characterized in that: Based on the relationship between the different types of faults and aerodynamic coefficients, an angular velocity loop adaptive backstepping control law is designed according to the nonlinear dynamic terms of the state variables and the control input matrix of the control variables under the fault mode, including: Based on the relationship between the different types of faults and aerodynamic coefficients, construct an aircraft moment equation; Obtaining the matrix form of the angular velocity according to the aircraft moment equation, and obtaining the nonlinear dynamic terms about the state variables and the control input matrix about the control variables; The adaptive backstepping control law of the angular velocity loop is designed according to the nonlinear dynamic terms of the state variables and the control input matrix of the control variables of the aircraft under fault mode.

5. The adaptive backstepping fault-tolerant control method for transport aircraft wing surface faults according to claim 4 is characterized in that: The matrix form of the angular velocity is: in, is the derivative of the aircraft angular velocity, is the roll angular velocity derivative, is the pitch angular velocity derivative, r is the yaw angular velocity derivative, X = [pqr] T is the angular velocity of the aircraft, p is the rolling angular velocity, q is the pitch angular velocity, r is the yaw angular velocity, u=[δ el δ er δ al δ ar δ r ] T The left elevator δ el 、Right elevatorδ er , left aileron δ al 、Right aileron δ ar 、Rudder δ r The control input vector is composed of: F(X) is the nonlinear dynamic term about the state variable; G(X) is the control input matrix about the control variable; ζ(X,u) is the change of aerodynamic parameters of the aircraft and unknown dynamic parameters when the angular velocity and the control input vector change due to a fault; y represents the output state quantity; The nonlinear dynamic term about the state variable is: Among them, J x is the moment of inertia matrix, Q is the dynamic pressure, b is the wing span, is the average aerodynamic chord length, C lf is the set of rolling moment coefficients that are independent of the deflection of the rudder surface, C mf is the set of pitching moment coefficients that are independent of the rudder deflection, C nf is the set of yaw moment coefficients that are independent of the deflection of the rudder surface; The control input matrix of the control variables is: Where S is the wing reference area, is the rolling moment about each rudder surface δ j The manipulated derivative of is the pitch moment about each rudder surface δ j The manipulated derivative of is the yaw moment about each rudder surface δ j The manipulated derivative, δ j =δ el ,δ er ,δ al ,δ ar ,δ r , δ el is the left elevator, δ er is the right elevator, δ al is the left aileron, δ ar is the right aileron, δ r For the rudder.

6. The adaptive backstepping fault-tolerant control method for transport aircraft wing surface faults according to claim 4, characterized in that: The angular velocity loop adaptive backstepping control law is: Among them, K and Γ are diagonal matrices, X ref is the desired angular velocity instruction, X is the angular velocity of the aircraft, is the change in the aircraft's aerodynamic parameters and the estimated value of the unknown dynamic parameters ζ(X,u) when the angular velocity and control input vector change due to a fault.

7. The adaptive backstepping fault-tolerant control method for transport aircraft wing surface faults according to claim 1, characterized in that: The weight matrix assigned to the angular velocity loop control is selected according to the airfoil failure mode, including: When the wing failure mode is wing damage and the i-th control surface is lost, the weight matrix of the control allocation is: IN u =diag(k1,k2,…,1,…,k M ); Among them, W u is the weight matrix for controlling the allocation, 0<k1,k2,…,k M <1, k1, k2, ..., k M is the weight of different rudder surfaces, M is the number of rudder surfaces; When the wing failure mode is rudder damage, the weight matrix of the control allocation is: W u =diag(1,1,…,k i ,…,1); Where ξ is the damage size of the rudder surface, 0<ξ<1, k i is the weight of damaged rudder surface i, k i >1; When the wing failure mode is that the rudder is stuck, the weight matrix of the control allocation is: IN u =diag(k1,k2,…,1,…,k M ); Among them, 0<k1,k2,…,k M <1, k1, k2, ..., k M is the weight of different rudder surfaces, and M is the number of rudder surfaces.

8. The adaptive backstepping fault-tolerant control method for transport aircraft wing surface faults according to claim 1, characterized in that: The control allocation of the angular velocity loop using the weighted pseudo-inverse method includes: The weighted pseudo-inverse method is used to formulate the control allocation problem of the angular velocity loop as follows: Among them, Ω is the solution space of control input, W u is a weight distribution matrix and is reversible, u is the output value of the aircraft control surface, u c is the desired control input, B(x) is the virtual control quantity instruction v ref and the mapping relationship from control input to virtual control quantity, B(x):R M →R r (M>r), R M is an M-dimensional real number space, R r is an r-dimensional real number space, M is the number of control surfaces, and r is the dimension of the virtual control quantity.

9. The adaptive backstepping fault-tolerant control method for transport aircraft wing surface faults according to claim 1, characterized in that: The adaptive backstepping fault-tolerant angular velocity control law with control allocation is: Where u is the left elevator delta el 、Right elevatorδ er , left aileron δ al 、Right aileron δ ar 、Rudder δ r The control input vector composed of u=[δ el δ er δ al δ ar δ r ] T , W u is the control allocation weight matrix, G(X) is the control input matrix for the control variables, K and Γ are diagonal matrices, X ref is the desired angular velocity command, X is the angular velocity, F(X) is the nonlinear dynamic term about the state variable, is the change in the aircraft's aerodynamic parameters and the estimated value of the unknown dynamic parameters ζ(X,u) when the angular velocity and control input vector change due to a fault, for Moore-Penrose inverse of G(X). Taking matrix G(X) as an example, T(X) is the Moore-Penrose inverse of G(X), and T(X) = G(X) T (G(X)G(X) T ) -1 .

Images