A Servo Elastic Stabilization Method Based on Multi-Sensor Information Fusion

Abstract

This invention discloses a servo-elastic stabilization method based on multi-sensor information fusion, belonging to the field of flight control. Specifically, it involves: extracting and modeling the main load-bearing structure of the aircraft; calculating the modal frequencies, mode shapes, and mode shape slopes of the aircraft under unconstrained conditions; then combining the aircraft, a first sensor, servo motors, and the control system to form a closed-loop servo-elastic system; calculating the open-loop system transfer function from the control input to the feedback input of the control system, drawing an open-loop Bode plot, and determining the dangerous frequencies and corresponding dangerous natural modes; based on this, placing a second angular velocity sensor on the aircraft; then calculating and correcting the weighting coefficients of the two angular velocity sensor signals, and introducing the weighted signal into the control system instead of the original signal; comparing the system characteristics to ensure the servo-elastic stability of the aircraft without affecting the rigid body phase and amplitude.

Description

Technical Field

[0001] This invention belongs to the field of flight control, specifically relating to a servo elastic stabilization method based on multi-sensor information fusion. Background Technology

[0002] The aircraft's motion signals are input to the control system via sensors, which then generates control signals that are input to the control surfaces, causing them to deflect and generating control forces, which are ultimately fed back to the aircraft. Because the aircraft is an elastic body, the sensors collect not only the rigid body motion signals but also elastic vibration signals. As long as the servo control system's frequency band covers the aircraft's main natural frequencies, it will provide additional high-frequency control forces, thus affecting the aircraft's motion.

[0003] Under normal circumstances, an individual elastic aircraft is stable, and an individual servo control system is also stable. However, once a closed loop is formed, its stability changes significantly, even becoming servo-elastic instability. This servo-elastic effect can cause structural fatigue damage, reduce the performance of the control system, and even lead to serious structural failure. Therefore, the issue of servo-elastic stability must be taken seriously in aircraft design.

[0004] Currently, the main engineering methods for servo elastic stabilization are to add low-pass filters or structural notch filters to the control system to eliminate elastic vibration signals, thereby improving the system's stability.

[0005] For aircraft with elastic frequencies much higher than rigid body frequencies, adding low-pass filters or structural notch filters is effective. However, when the elastic frequencies are close to the rigid body frequencies, the phase lag introduced by low-pass filters or structural notch filters can significantly impact the aircraft's flight control, thereby affecting its flight performance. Furthermore, as aircraft become increasingly complex, it is impossible to accurately calculate elastic mode information using structural finite element analysis or ground-based experiments.

[0006] Therefore, there is an urgent need in engineering for a servo-elastic robust stabilization method that can take into account the severe coupling between rigid and elastic systems. Summary of the Invention

[0007] To address the problem of servo-elastic stability in rigid-elastic coupling with uncertain mode shapes, this invention proposes a servo-elastic stability enhancement method based on multi-sensor information fusion.

[0008] The servo elastic stabilization method based on multi-sensor information fusion comprises the following steps:

[0009] Step 1: Extract the main load-bearing structure of the aircraft, perform geometric modeling and finite element modeling, and calculate the modal frequencies, mode shapes and mode slopes of the aircraft under no constraints.

[0010] Specifically:

[0011] First, based on the modeling, the overall mass matrix M of the aircraft is obtained. s and the overall stiffness matrix K s ;

[0012] Matrix M s and K s All are ng×ng dimensional square matrices, where ng is the number of degrees of freedom of the finite element model.

[0013] Then, within the linear elastic range, the free vibration equation of the load-bearing structure is:

[0014]

[0015] In the formula, x = [x1, x2, x3, ..., x ng ] T It is a node displacement array of ng×1.

[0016] Furthermore, by introducing the harmonic oscillation condition into the free vibration equation, we obtain:

[0017] (K S -ω 2 M S )Φ=0

[0018] ω is the natural angular frequency of oscillation, and Φ is the natural mode shape of ng×1.

[0019] Finally, by solving the above equation, we obtain: ng eigenvalues As modal frequencies, and the corresponding ng eigenvectors Φ1, Φ2, Φ3, ..., Φ ng As mode shapes; and based on the nodal coordinates and the natural mode shapes, the slopes Φ1′, Φ′2, Φ′3, ..., Φ′ corresponding to the mode shapes are obtained. ng .

[0020] Step 2: Based on the mean body axis, natural modes, and Lagrange equations, establish the aircraft dynamics equations in modal generalized coordinates;

[0021] Step 3: The aircraft is equipped with a rigid body stabilization angular velocity sensor as the first sensor. The aircraft is combined with the servo motor, the first sensor and the control system to form a closed-loop servo elastic system.

[0022] First, the servo motor receives external control signals and feedback signals from the control system, and outputs the control surface deflection angle to the aircraft; then, the aircraft outputs the angular velocity signal from the first sensor. Give it to the first sensor;

[0023] angular velocity signal Represented as: In the formula, W1' is a 1×N vector, representing the slope of the modal shape at that location.

[0024] Furthermore, the first sensor sends the acquired angular velocity signal to the control system;

[0025] Finally, the control system outputs a feedback signal to the servo motor to achieve closed-loop operation.

[0026] Step 4: Calculate the time-domain state-space form and corresponding frequency-domain equations of each part of the closed-loop servo elastic system, and finally obtain the transfer function of each part.

[0027] The closed-loop servo elastic system comprises four parts: the aircraft, the first sensor, the servo motor, and the control system; the common time-domain state-space form of these four parts is as follows:

[0028]

[0029] In the formula, x is n x A state vector of size 1, u is n u The input vector is ×1, and y is n. y The output vector is ×1, and A is n. x ×n x The state matrix, B is n x ×n u The input matrix, C is n y ×n x The direct feed matrix, D is n y ×n c The output matrix.

[0030] Furthermore, the time-domain state-space equations are transformed into frequency-domain equations:

[0031]

[0032] In the formula, Where is the frequency in the frequency domain, and 1 is the frequency of n. x ×n x unit array, n y ×n c The transfer function matrix, where i is the imaginary number.

[0033] Step 5: Using the transfer functions of each part of the closed-loop servo elastic system, calculate the open-loop system transfer function from the manipulated variable to the feedback variable of the control system.

[0034] Step 6: Based on the open-loop system transfer function Draw the open-loop Bode plot and determine the danger frequency ω based on the highest amplitude peak. d Select the modal frequency and the danger frequency ωd The closest k-th intrinsic mode is identified as the dangerous intrinsic mode;

[0035] Step 7: Based on the slope Φ′ corresponding to the dangerous natural mode shape. k The second angular velocity sensor was placed on the aircraft.

[0036] The principle for selecting the placement position of the servo elastically stabilized angular velocity sensor as the second angular velocity sensor on the aircraft is that the slope of the dangerous natural mode corresponding to the placement position of the first angular velocity sensor and the slope of the dangerous natural mode corresponding to the placement position of the second angular velocity sensor must satisfy opposite signs and the absolute value error must be within the set threshold range.

[0037] Step 8: Calculate the weighting coefficients K1 and K2 based on the slopes of the dangerous natural mode shapes at the locations of the two angular velocity sensors.

[0038] K1 = 1 - K2

[0039]

[0040] In the formula, W1′ and W2′ are the slopes of the dangerous inherent mode shapes at the first and second angular velocity sensors, respectively.

[0041] Step 9: Adjust the weighting coefficients K1 and K2 to obtain the adaptively adjusted weighting coefficient K1. * K2 * and weighted signals and the weighted signal Replace the original signal Introduced into the control system;

[0042] The formula for calculating the correction amount is:

[0043]

[0044] In the formula, a > 0 is a proportional coefficient used to adjust the speed of adaptive adjustment. These are the signals collected by the first angular velocity sensor and the second angular velocity sensor, respectively.

[0045] The adaptively adjusted weighting coefficients are:

[0046] K1 * =K1+ΔK1

[0047] K2 * =K2+ΔK2

[0048] The weighted signal is:

[0049]

[0050] The initial value is

[0051] Step 10: Signal The output introduced into the closed-loop servo elastic system, compared with the original signal The output of the closed-loop servo elastic system is introduced into the time-domain simulation under external excitation. By comparing the response amplitude and convergence speed, the stability enhancement of the servo elasticity is verified.

[0052] The advantages of the present invention are:

[0053] 1. A servo elastic stabilization method based on multi-sensor information fusion, which determines the dangerous elastic modes based on the open-loop Bode plot of the aircraft, and determines the position of the second angular velocity sensor based on the position of the first angular velocity sensor. Then, it calculates the signal weighting coefficient based on the mode shape slope of the two positions, and introduces the weighted signal into the control system instead of the original signal to ensure the servo elastic stability of the aircraft without affecting the rigid body phase and amplitude.

[0054] 2. A servo elastic stabilization method based on multi-sensor information fusion, using weighting coefficients K1 and K2 obtained from theoretical mode shape slopes as initial values, and calculating corrections by integrating the product of the difference between the two sensor signals and the weighted signal, thereby adaptively adjusting the weighting coefficients to ensure the servo elastic stability of the aircraft when there is a certain uncertainty in the mode shape data. Attached Figure Description

[0055] Figure 1 This is a flowchart of a servo elastic stabilization method based on multi-sensor information fusion according to the present invention.

[0056] Figure 2 This is a schematic diagram of the closed-loop servo elastic system of the present invention;

[0057] Figure 3 This forms the framework of the multi-sensor information fusion and stabilization method of the present invention;

[0058] Figure 4 This is the framework of the multi-sensor information fusion adaptive stabilization method of the present invention;

[0059] Figure 5 This is a Bode diagram of an open-loop system according to an embodiment of the present invention;

[0060] Figure 6 This is a schematic diagram showing the position of the angular velocity sensor in an embodiment of the present invention;

[0061] Figure 7 This is a multi-sensor information fusion and stabilization system according to an embodiment of the present invention;

[0062] Figure 8 This is a comparison chart of the frequency characteristics of the stabilization system without a filter, with a traditional filter, and using multi-sensor information fusion in an embodiment of the present invention;

[0063] Figure 9 This is a comparison chart of the pitch angular velocity response of the system with and without the adaptive stabilization system enabled at 1 second, according to an embodiment of the present invention. Detailed Implementation

[0064] To facilitate understanding and implementation of the present invention by those skilled in the art, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Obviously, the described embodiments are merely some, not all, embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort should fall within the scope of protection of the present invention.

[0065] In general, the arrangement of multiple sensors in an aircraft is merely for redundancy, and the control system uses only the signal from one sensor as input. However, the method of this invention rationally weights the signals from two sensors and uses the weighted result as the input to the control system. The purpose of this weighting is to eliminate certain elastic mode information from the sensor signals, thereby ensuring the servo elastic stability of that mode. Furthermore, with appropriate weighting coefficients, it can achieve the effect of not affecting the rigid body amplitude and phase. In addition, the method of this invention relies entirely on modal shape data; therefore, when there is some uncertainty, this invention uses an integral network to adaptively adjust the weighting coefficients to ensure the effectiveness of the method.

[0066] The servo elastic stabilization method based on multi-sensor information fusion, such as Figure 1 As shown, the steps are as follows:

[0067] Step 1: Extract the main load-bearing structure of the aircraft, perform geometric modeling and finite element modeling, and use the modal analysis module of commercial software such as Nastran to calculate the modal frequencies, mode shapes and mode slopes of the aircraft under no constraints.

[0068] Specifically:

[0069] First, based on the modeling, the overall mass matrix M of the aircraft is obtained. s and the overall stiffness matrix K s ;

[0070] Matrix M s and K s All are ng×ng dimensional square matrices, where ng is the number of degrees of freedom of the finite element model.

[0071] Then, within the linear elastic range, the free vibration equation of the load-bearing structure is:

[0072]

[0073] In the formula, x = [x1, x2, x3, ..., x ng ] T It is a node displacement array of ng×1.

[0074] Furthermore, a harmonic oscillation condition is introduced into the free vibration equation:

[0075] x=[φ1,φ2,φ3,…,φ ng ] T exp(iωt)=Φexp(iωt) (2)

[0076] Substituting equation (2) into equation (1), we get:

[0077] (K S -ω 2 M S )Φ=0 (3)

[0078] ω is the natural angular frequency of oscillation, and Φ is the natural mode shape of ng×1.

[0079] Equation (3) represents the expression for M. s and K s The generalized eigenvalue problem can be solved by obtaining ng eigenvalues ​​from the above equation. As modal frequencies, and the corresponding ng eigenvectors Φ1, Φ2, Φ3, ..., Φ ng As mode shapes; and based on the nodal coordinates and the natural mode shapes, the slopes Φ1′, Φ′2, Φ′3, ..., Φ′ corresponding to the mode shapes are determined. ng .

[0080] Step 2: Based on the mean body axis, natural modes, and Lagrange equations, establish the aircraft dynamics equations in modal generalized coordinates:

[0081]

[0082] In the formula, M q C q K q Let M be the generalized mass, generalized damping, and generalized stiffness matrices of the N×N spacecraft, q be the N×1 generalized modal coordinates of the spacecraft, and M be the generalized coordinates of the N×1 spacecraft modal coordinates. qδ Let be the N×L coupled inertial mass matrix of the aircraft control surfaces, δ be the L×1 control surface deflection mode coordinates, N be the number of natural modes after mode truncation, and L be the number of control surface deflection modes; ρ be the atmospheric density, V be the flight speed, and A be the air velocity. qLet A be an N×N array of unsteady aerodynamic coefficients for the inherent modes of the aircraft. δ For N×L, there is an unsteady aerodynamic coefficient matrix for the rudder deflection mode.

[0083] Step 3: The aircraft is equipped with a rigid body stabilization angular velocity sensor as the first sensor. The aircraft is combined with the servo motor, the first sensor and the control system to form a closed-loop servo elastic system.

[0084] The function of the first sensor is to increase the damping of rigid body motion;

[0085] First, the servo motor receives external control signals and feedback signals from the control system, and outputs the control surface deflection angle to the aircraft; then, the aircraft outputs the angular velocity signal from the first sensor. Give it to the first sensor;

[0086] angular velocity signal Represented as:

[0087]

[0088] In the formula, W1' is a 1×N vector, representing the slope of the modal shape at that location.

[0089] Furthermore, the first sensor sends the angular velocity signal, after being filtered through limiting and low-pass filtering, to the control system;

[0090] Finally, the control system outputs a feedback signal to the servo motor to achieve closed-loop operation.

[0091] Step 4: Calculate the time-domain state-space form and corresponding frequency-domain equations of each part of the closed-loop servo elastic system, and finally obtain the transfer function of each part.

[0092] like Figure 2 As shown, the closed-loop servo elastic system comprises four parts: the aircraft, the first sensor, the servo motor, and the control system. Theoretical models of the servo motor and angular velocity sensor can be established based on the parameters provided by the manufacturer. A theoretical model of the aircraft can be established based on the equations from step two. A theoretical model of the control system can be established based on the control requirements. The common time-domain state-space form of the theoretical models for these four parts is as follows:

[0093]

[0094] In the formula, x is n x A state vector of size 1, u is n u The input vector is ×1, and y is n. y The output vector is ×1, and A is n. x ×n x The state matrix, B is n x ×n uThe input matrix, C is n y ×n x The direct feed matrix, D is n y ×n c The output matrix.

[0095] Furthermore, the time-domain state-space equations are transformed into frequency-domain equations:

[0096]

[0097] In the formula, Where is the frequency in the frequency domain, and 1 is the frequency of n. x ×n x unit array, n y ×n c The transfer function matrix, where i is the imaginary number.

[0098] Step 5: Using the transfer functions of each part of the closed-loop servo elastic system, calculate the open-loop system transfer function from the manipulated variable to the feedback variable of the control system.

[0099]

[0100] In the formula, For servo transfer functions, The transfer function of the aircraft. The transfer function of the first angular velocity sensor. The transfer function for the control system.

[0101] Step 6: Based on the open-loop system transfer function Draw the open-loop Bode plot and determine the danger frequency ω based on the highest amplitude peak. d Select the modal frequency and the danger frequency ω d The closest k-th intrinsic mode is identified as the dangerous intrinsic mode;

[0102] Step 7: Based on the slope Φ′ corresponding to the dangerous natural mode shape. k The second angular velocity sensor was placed on the aircraft.

[0103] The servo elastic stabilization angular velocity sensor, as a second angular velocity sensor, serves to improve servo elastic stability.

[0104] The selection principle for the placement of the second angular velocity sensor on the aircraft is that the slope of the dangerous natural mode corresponding to the placement position of the first angular velocity sensor and the slope of the dangerous natural mode corresponding to the placement position of the second angular velocity sensor must satisfy opposite signs and the absolute value error must be within the set threshold range.

[0105] Step 8: Calculate the weighting coefficients K1 and K2 based on the slopes of the dangerous natural mode shapes at the locations of the two angular velocity sensors.

[0106]

[0107] In the formula, W1′ and W2′ are the slopes of the dangerous inherent mode shapes at the first and second angular velocity sensors, respectively.

[0108] Step 9: Adjust the weighting coefficients K1 and K2 to obtain the adaptively adjusted weighting coefficient K1. * K2 * and weighted signals and the weighted signal Replace the original signal Introduced into the control system;

[0109] Considering the uncertainty of modal vibration data, using weighting coefficients K1 and K2 as initial values, the correction amounts ΔK1 and ΔK2 are calculated by integrating the product of the difference between the two sensor signals and the weighted signal. The formula for calculating the correction amounts is as follows:

[0110]

[0111] In the formula, a > 0 is a proportional coefficient used to adjust the speed of adaptive adjustment. These are the angular velocity signals collected by the first and second angular velocity sensors, respectively. When determining the specific value of 'a', it should be adjusted from small to large until the system has a relatively fast adjustment speed while maintaining stability.

[0112] The adaptively adjusted weighting coefficients are:

[0113]

[0114] The weighted signal is:

[0115]

[0116] The initial value is

[0117] Step 10: Signal The output introduced into the closed-loop servo elastic system, compared with the original signal The output of the closed-loop servo elastic system is introduced into the time-domain simulation under external excitation. By comparing the response amplitude and convergence speed, the stability enhancement of the servo elasticity is verified.

[0118] For linear systems based on multi-sensor information fusion stabilization methods, such as Figure 3As shown, the servo elastic stability is verified by comparing the gain margin and phase margin using an open-loop Bode plot. Specifically:

[0119] After calculating the weighting coefficients K1 and K2 based on the slopes of the dangerous natural mode shapes at the locations of the two angular velocity sensors, the signals collected by the two sensors are weighted using the following formula:

[0120]

[0121] In the formula, The angular velocity signal is acquired by the rigid body-stabilized angular velocity sensor. The angular velocity signal is acquired by a servo elastically stabilized angular velocity sensor. for

[0122] The weighted angular velocity signal.

[0123] The weighted signal Replace the original signal The open-loop system transfer function is introduced into the control system and resynthesized.

[0124]

[0125] In the formula, G b2 (ω) represents the angular velocity signal from the aircraft rudder deflection signal to the servo elastic stabilization signal. The transfer function.

[0126] According to the new open-loop transfer function Draw the open-loop Bode plot and compare it with the original transfer function. The open-loop Bode plot is drawn to compare the gain margin and phase margin, thereby verifying the servo elasticity enhancement.

[0127] For nonlinear systems based on multi-sensor information fusion adaptive stabilization methods, such as Figure 4 As shown, the adaptive adjustment process of weighting coefficients K1 and K2 and the change of system time-domain response are studied through time-domain simulation under the condition of erroneous initial values, so as to verify the servo elastic stability enhancement.

[0128] Example:

[0129] Step 1: For a closed-loop servo-elastic system, it generally includes four parts: servo motor, aircraft, angular velocity sensor, and control system. The transfer functions of each part can be obtained through theoretical calculations or experiments, namely the servo motor transfer function G. a Aircraft transfer function G b Angular velocity sensor transfer function and control system transfer function G cFurthermore, the open-loop system transfer function G from the manipulated variable to the feedback variable of the control system is obtained. open Draw the open-loop Bode plot based on the open-loop system transfer function, and determine the critical elastic mode based on the frequency corresponding to the highest peak amplitude.

[0130] Step 2, for the angular velocity signal The slopes of the mode shapes of the hazardous elastic modes are extracted to determine the placement of a second angular velocity sensor on the aircraft. The principle for selecting the location is that the slopes of the mode shapes of the hazardous elastic modes at the first and second sensors have opposite signs and similar absolute values. After the location is determined, weighting coefficients K1 and K2 are calculated based on the slopes of the mode shapes of the hazardous elastic modes at the two locations, and the signals collected by the two sensors are weighted.

[0131] Step 3, weight the signal φ * Instead of introducing the original signal into the control system, a stabilization method framework based on multi-sensor information fusion is established. After introducing the above stabilization method, the open-loop system transfer function from the manipulated variable to the feedback variable of the control system can be re-synthesized. The open-loop Bode plot is drawn based on the new open-loop transfer function and compared with the open-loop Bode plot of the original system to confirm the effectiveness of the method of the present invention.

[0132] Step 4: Considering the inherent uncertainty in the modal shape data, the weighting coefficients K1 and K2 calculated in Step 2 are used as initial values. The correction is calculated by integrating the product of the difference between the two sensor signals and the weighted signal, thus establishing a framework for an adaptive stabilization method based on multi-sensor information fusion. After introducing the above adaptive stabilization method, the system transforms from a linear system to a nonlinear system. At this point, time-domain simulation is used to study the adaptive adjustment process of the weighting coefficients K1 and K2 and the changes in the system's time-domain response under incorrect initial values ​​to confirm the effectiveness of the method.

[0133] In step 2, the weighting coefficients are calculated based on the composition of the components in the angular velocity signal and the slope of the dangerous elastic mode. In step 3, a stabilization method framework for multi-sensor information fusion is constructed by weighting the signals from multiple sensors. In step 4, considering the uncertainty of the mode shape data, an adaptive stabilization method framework for multi-sensor information fusion is constructed through an integral network to achieve adaptive adjustment of the weighting coefficients.

[0134] The object of this invention is an aircraft with a rigid-elastic coupling servo-elasticity problem, and its open-loop system Bode diagram is as follows. Figure 5 As shown. The position of the first angular velocity sensor has now been selected. The position of the second angular velocity sensor is selected using the mode shape slope data, as shown below. Figure 6As shown. The theoretical values ​​of K1 and K2 are calculated based on the mode slope data at the two sensor locations, and a stabilization system based on multi-sensor information fusion is constructed, as follows. Figure 7 As shown, the open-loop system transfer function is calculated using the aerodynamic servo elastic frequency domain equation for three cases: no filter, a system with a traditional filter, and a system using multi-sensor information fusion for stabilization. Open-loop Bode plots are drawn, and the stabilization effect is analyzed. Figure 8 As shown. Considering the uncertainty of modal vibration data, an adaptive stabilization system based on multi-sensor information fusion was built. A step excitation signal was input to the elevator, and the time-domain response was calculated with and without the adaptive stabilization system activated at 1 second. The stabilization effect was then analyzed. Figure 9 As shown in the figure. The above embodiments illustrate the applicability of the method of the present invention to the problem of rigid-elastic coupling servo elasticity.

[0135] The open-loop Bode diagram of an embodiment of the present invention is shown below. Figure 5 As shown, a dangerous amplitude peak of 2.70 dB appears at 2.8 Hz (corresponding to the first elastic mode) in the amplitude-frequency response, indicating that the first elastic mode is the dangerous elastic mode. For traditional aerodynamic servo-elastic filters, the low-pass filter is ineffective due to the low frequency of the first elastic mode, while the low center frequency of the structural notch filter will affect the rigid body phase, leading to insufficient system phase margin. In this case, a servo-elastic stabilization method based on multi-sensor information fusion can be used to improve the aerodynamic servo-elastic stability under rigid-elastic coupling conditions.

[0136] According to the arrangement principle of the second angular velocity sensor, the second angular velocity sensor in this embodiment of the invention is arranged near the wingtip, such as... Figure 6 As shown. The slopes of the dangerous elastic mode shapes at the two sensor locations are as follows:

[0137] W1′=-0.0088

[0138] W2′=0.0128

[0139] Calculate the weighting coefficients based on the slopes of the dangerous elastic mode shapes at the two sensor locations:

[0140] K1 = 0.5926

[0141] K2 = 0.4074

[0142] The weighted signal is introduced into the control system instead of the original signal to establish a stabilization-enhancing system based on multi-sensor information fusion, such as... Figure 7 As shown, Bode plot results for three cases—without a filter, with a traditional filter, and using a multi-sensor information fusion stabilization system—were calculated using a pneumatic servo elastic frequency domain model. Figure 8 As shown in the results.

[0143] 1) The peak amplitude of the first elastic mode can be reduced by adding a structural notch filter and introducing a second sensor. By properly designing the parameters of the structural notch filter, the peak amplitude can be made much smaller than the peak amplitude after introducing the second sensor. However, considering that the goal of introducing the second sensor is to eliminate the vibration signal of the first elastic mode, although the peak amplitude is not reduced much by this method, as long as the rigid body motion is stable, the stability of the first elastic mode can still be guaranteed after introducing the signal from the second sensor.

[0144] 2) For rigid body phase, since the center frequency of the structural filter is low, it has a greater impact on the rigid body phase. However, the introduction of a second sensor does not affect the rigid body phase at all. Therefore, in the case of severe rigid-elastic coupling, the stabilization method of multi-sensor information fusion is better.

[0145] Since the calculation of parameters such as K1 and K2 depends entirely on the mode shape, the effectiveness of the multi-sensor information fusion stabilization method will be significantly affected when there is a difference between the actual and theoretical mode shapes. In this embodiment, an adaptive stabilization system based on multi-sensor information fusion is introduced. The value of 'a' is adjusted from small to large. When 'a' = 0.03, the system remains stable and has a relatively fast adjustment speed. After adding the adaptive stabilization system, the original linear system has been transformed into a nonlinear system. Therefore, in this embodiment, the effectiveness of this method is illustrated through time-domain simulation. A step excitation signal is input to the elevator, and the time-domain response is calculated using an aerodynamic servo elastic time-domain model with and without the adaptive stabilization system activated at 1 second. Figure 9 As shown, it can be observed that after activating the adaptive stabilization system, the convergence speed of the system's pitch angular velocity is significantly accelerated.

Claims

1. A servo elastic stabilization method based on multi-sensor information fusion, characterized in that, The specific steps are as follows: Step 1: Extract the main load-bearing structure of the aircraft, perform geometric modeling and finite element modeling, and calculate the modal frequencies, mode shapes and mode slopes of the aircraft under no constraints. Step 2: Based on the mean body axis, the aircraft's natural modes, and the Lagrange equations, establish the aircraft dynamics equations in the modal generalized coordinate system. Step 3: The aircraft is equipped with a rigid body stabilization angular velocity sensor as the first sensor. The aircraft is combined with the servo motor, the first sensor and the control system to form a closed-loop servo elastic system. Step 4: Calculate the time-domain state-space form and corresponding frequency-domain equations of each part of the closed-loop servo elastic system, and finally obtain the transfer function of each part. Step 5: Using the transfer functions of each part of the closed-loop servo elastic system, calculate the open-loop system transfer function from the manipulated variable to the feedback variable of the control system. Step 6: Based on the open-loop system transfer function Draw the open-loop Bode plot and determine the danger frequency ω based on the highest amplitude peak. d Select the modal frequency and the danger frequency ω d The closest k-th intrinsic mode is identified as the dangerous intrinsic mode; Step 7: Based on the slope Φ′ corresponding to the dangerous natural mode shape. k The second angular velocity sensor was placed on the aircraft. Step 8: Calculate the weighting coefficients K1 and K2 based on the slopes of the dangerous natural mode shapes at the locations of the two angular velocity sensors; Step 9: Adjust the weighting coefficients K1 and K2 to obtain the adaptively adjusted weighting coefficient K1. * K2 * and weighted signals and the weighted signal Replace the original signal Introduced into the control system; Step 10: Signal The output introduced into the closed-loop servo elastic system, compared with the original signal The output of the closed-loop servo elastic system is introduced into the time-domain simulation under external excitation. By comparing the response amplitude and convergence speed, the stability enhancement of the servo elasticity is verified.

2. The servo elastic stabilization method based on multi-sensor information fusion as described in claim 1, characterized in that, Step one specifically involves: First, based on the modeling, the overall mass matrix M of the aircraft is obtained. s and the overall stiffness matrix K s ; Matrix M s and K s All are ng×ng dimensional square matrices, where ng is the number of degrees of freedom of the finite element model; Then, within the linear elastic range, the free vibration equation of the load-bearing structure is: In the formula, x = [x1, x2, x3, ..., x ng ] T It is an ng×1 node displacement array; Furthermore, by introducing the harmonic oscillation condition into the free vibration equation, we obtain: (K S -oh 2 M S )Φ=0 ω is the natural angular frequency, and Φ is the natural mode shape of ng×1; Finally, by solving the above equation, we obtain: ng eigenvalues As modal frequencies, and the corresponding ng eigenvectors Φ1, Φ2, Φ3, ..., Φ ng As mode shapes; and based on the nodal coordinates and the natural mode shapes, the slopes Φ1′, Φ′2, Φ′3, ..., Φ′ corresponding to the mode shapes are obtained. ng .

3. The servo elastic stabilization method based on multi-sensor information fusion as described in claim 1, characterized in that, In step two, the aircraft dynamics equations are: In the formula, M q C q K q Let M be the generalized mass, generalized damping, and generalized stiffness matrix of the N×N spacecraft. qδ Let be the N×L coupled inertial mass matrix of the aircraft control surfaces, δ be the L×1 control surface deflection mode coordinates, N be the number of natural modes after mode truncation, and L be the number of control surface deflection modes; ρ be the atmospheric density, V be the flight speed, and A be the air velocity. q Let A be an N×N array of unsteady aerodynamic coefficients for the inherent modes of the aircraft. δ For N×L, there is an unsteady aerodynamic coefficient matrix for the rudder deflection mode.

4. The servo elastic stabilization method based on multi-sensor information fusion as described in claim 1, characterized in that, In step three, firstly, the servo motor receives the external input control signal and the feedback signal output by the control system, and outputs the control surface deflection angle to the aircraft; then, the aircraft outputs the angular velocity signal φ1 at the first sensor to the first sensor. angular velocity signal Expressed as: In the formula, W1' is a 1×N vector representing the mode shape slope at that location; q is an N×1 generalized coordinate of the aircraft mode. Furthermore, the first sensor sends the acquired angular velocity signal to the control system; Finally, the control system outputs a feedback signal to the servo motor to achieve closed-loop operation.

5. The servo elastic stabilization method based on multi-sensor information fusion as described in claim 1, characterized in that, In step four, the closed-loop servo elastic system comprises four parts: the aircraft, the first sensor, the servo motor, and the control system; the common time-domain state-space form of these four parts is as follows: In the formula, x is n x A state vector of size 1, u is n u The input vector is ×1, and y is n. y The output vector is ×1, and A is n. x ×n x The state matrix, B is n x ×n u The input matrix, C is n y ×n x The direct feed matrix, D is n y ×n c The output matrix; Furthermore, the time-domain state-space equations are transformed into frequency-domain equations: Where, Let 1 be the frequency in the frequency domain, and 1 be the frequency of n. x ×n x unit array, For n y ×n c The transfer function matrix, where i is the imaginary number.

6. The servo elastic stabilization method based on multi-sensor information fusion as described in claim 1, characterized in that, In step five, the formula for calculating the open-loop system transfer function is: Where, For servo transfer functions, The transfer function of the aircraft. The transfer function of the first angular velocity sensor. The transfer function for the control system.

7. The servo elastic stabilization method based on multi-sensor information fusion as described in claim 1, characterized in that, Step seven specifically involves the following: The servo elastically stabilized angular velocity sensor is used as the second angular velocity sensor. The selection principle for its placement on the aircraft is that the slope of the dangerous natural mode corresponding to the placement position of the first angular velocity sensor and the slope of the dangerous natural mode corresponding to the placement position of the second angular velocity sensor must satisfy opposite signs and the absolute value error must be within a set threshold range.

8. The servo elastic stabilization method based on multi-sensor information fusion as described in claim 1, characterized in that, In step eight, the formula for calculating the weighting coefficients is as follows: K1 = 1 - K2 In the formula, W1′ and W2′ are the slopes of the dangerous inherent mode shapes at the first and second angular velocity sensors, respectively.

9. The servo elastic stabilization method based on multi-sensor information fusion as described in claim 1, characterized in that, In step nine, the formula for calculating the correction amount is: In the formula, a > 0 is a proportional coefficient used to adjust the speed of adaptive adjustment; These are the angular velocity signals collected by the first angular velocity sensor and the second angular velocity sensor, respectively. The adaptively adjusted weighting coefficients are: K1 * =K1+ΔK1 K2 * =K2+ΔK2 The weighted signal is: The initial value is

Images