Hypersonic vehicle flight-propulsion integrated control method based on INDI

Abstract

The present application belongs to the technical field of hypersonic vehicle control, and particularly relates to a hypersonic vehicle flight-propulsion integrated control method based on INDI. The specific steps include: establishing a hypersonic vehicle flight-propulsion integrated longitudinal dynamics model in the form of a wall, designing a trajectory guide outer loop to realize height and speed tracking control, designing an attitude stabilization outer loop to realize angle of attack tracking control, constructing an attitude control inner loop to realize flight-propulsion control at the acceleration level, and designing an engine compensation layer to handle actuator dynamic delay. The present application actively utilizes the aerodynamic / propulsion coupling effect through the MIMO-INDI mechanism to realize collaborative distribution of elevators and engine fuel equivalence ratio, thereby improving the comprehensive control performance of the vehicle. The method is a flight / propulsion integrated control method, and has a wide application prospect.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of hypersonic vehicle control, and particularly relates to a hypersonic vehicle flight-propulsion integrated control method based on INDI (Incremental Nonlinear Dynamic Inversion), which realizes deep coupling control of aerodynamic surfaces and propulsion systems through Multiple-Input Multiple-Output (MIMO) control. BACKGROUND

[0002] Hypersonic vehicles are strategic high points in the future aerospace field. Their unique fuselage-propulsion integrated configuration determines their strong coupling characteristics between aerodynamics and propulsion. Changes in engine thrust will affect the flight attitude through eccentric moments; the lower surface of the rear body as part of the tail nozzle will change the air pressure distribution at this point when the thrust changes, generating additional moments. The lower surface of the front body as the pre-compression surface of the inlet will affect the inlet flow field when the flight attitude changes, affecting the working state of the propulsion system, forming strong two-way coupling.

[0003] In the prior art, the hypersonic vehicle control system mainly faces the following problems:

[0004] 1. Limitations of traditional decoupling control methods: The traditional method of separating flight control and engine control design attempts to suppress aerodynamic / propulsion coupling effects through decoupling compensation. This method cannot fully exploit the potential performance of the vehicle. And in the case of strong coupling, the error of decoupling compensation will lead to a decrease in control accuracy, and even cause system instability.

[0005] 2. Control challenges caused by strong nonlinearity: The aerodynamic characteristics and propulsion system characteristics of hypersonic vehicles change dramatically with flight state, showing strong nonlinearity. Traditional linear control methods are difficult to guarantee control performance within a wide flight envelope, and gain scheduling is required, but the design complexity of gain scheduling methods is high, and performance jumps may occur when the working conditions are switched.

[0006] 3. Challenges of model uncertainty: During hypersonic flight, there is a large uncertainty in aerodynamic parameters and propulsion parameters, including modeling errors, parameter perturbations, and external disturbances. Traditional control methods based on accurate models are sensitive to model uncertainty and lack robustness.

[0007] Existing related patent technologies also have certain deficiencies:

[0008] The patent "Hypersonic vehicle adaptive compensation control method based on neural network" (CN108375907B) solves the elevator failure and actuator saturation problem by introducing a radial basis function neural network to estimate the system nonlinearity function, but its control law is mainly designed for the longitudinal attitude subsystem, without considering the deep coupling of flight control and propulsion control, and the learning process of the neural network depends on the accurate fault model, which requires high computational resources in actual engineering applications.

[0009] The patent "Adaptive control method for hypersonic vehicle considering angle of attack constraint" (CN111158398A) proposes an angle of attack constraint scheme based on adaptive control, but the controller design handles speed control and attitude control independently, without actively utilizing the aerodynamic / propulsion coupling effect, and the compensation for engine dynamic delay is insufficient, making it difficult to guarantee control performance within a wide flight envelope.

[0010] The patent "Hypersonic vehicle expert intelligent control method and aircraft" (CN110244752B) realizes parameter adaptive adjustment of morphing aircraft through an expert system, improving the adaptability to morphing process and aerodynamic changes, but its control structure still treats the propulsion system as an independent module, without realizing true integration of flight and propulsion for collaborative distribution, and the knowledge base establishment process of the expert system is complex, with relatively insufficient robustness.

[0011] In summary, the main shortcomings of existing technologies are: lack of flight-propulsion integrated control mechanism, unable to fully utilize aerodynamic / propulsion coupling effect; insufficient adaptability to strong nonlinear systems, difficult to guarantee control performance within a wide flight envelope; unable to effectively handle the impact of engine dynamic delay on control performance; insufficient robustness to model uncertainty. Therefore, a new control method is needed that can actively utilize flight-propulsion coupling effect while having good nonlinear adaptability and robustness. SUMMARY

[0012] The present application provides a hypersonic vehicle flight-propulsion integrated control method based on INDI, aiming to actively utilize aerodynamic / propulsion coupling effect through MIMO-INDI mechanism, realize collaborative distribution of elevator and engine fuel equivalence ratio, and improve the overall control performance of the aircraft.

[0013] The technical solution of the present application is as follows:

[0014] The hypersonic vehicle flight-propulsion integrated control method based on INDI has the following steps:

[0015] Step (1) Establish a control-oriented hypersonic vehicle flight-propulsion integrated longitudinal dynamics model

[0016] A control-oriented longitudinal dynamics model of hypersonic vehicles is established, which can describe the strong coupling effect between aerodynamic characteristics and propulsion characteristics at high speed.

[0017] The following basic assumptions are adopted: the vehicle is regarded as a mass and moment of inertia constant rigid body, the elastic deformation of the body is ignored; only the longitudinal motion of the vehicle in the symmetry plane is considered; the air density is only regarded as a function of altitude .

[0018] (1.1) Longitudinal dynamics equation:

[0019] The longitudinal motion of the vehicle is described by the force balance equation of the mass center motion and the moment balance equation around the mass center:

[0020] (1)

[0021] where: the superscript "d" represents the first-order derivative; is the flight speed (m / s), is the ballistic angle (rad), is the pitch angle rate (rad / s), is the flight altitude (m), is the pitch angle (rad), is the angle of attack (rad), is the actual fuel equivalence ratio, is the fuel equivalence ratio command, is the engine response time constant (s), is the thrust (N), is the drag (N), is the lift (N), is the pitching moment (N·m), is the vehicle mass (kg), is the pitching moment of inertia (kg·m²), is the gravitational acceleration (m / s²). Equation The fuel equivalence ratio command is modeled as a first-order inertial link to the fuel equivalence ratio, which reflects the actual response characteristics of the engine, and its rate equation is included in the state equation.

[0022] (1.2) Force and moment model:

[0023] All aerodynamic forces, thrusts and moments are calculated by dimensionless coefficients to establish a unified force and moment expression:

[0024] (2) ​

[0025] where: P is the dynamic pressure (Pa), p is the air density (kg / m³), S is the reference area (m²), L is the reference length, i.e. the mean aerodynamic chord (m), l is the force arm of the thrust point relative to the center of mass (m), , , , CL, CD, CT, and Cm are the dimensionless coefficients of lift, drag, thrust, and pitching moment, respectively.

[0026] Pitching moment is composed of two parts, the first term is the eccentric moment of the thrust; the second term is the pure aerodynamic moment.

[0027] (1.3) Aerodynamic and propulsion coefficient fitting models:

[0028] To accurately describe the characteristics of the aircraft within a wide envelope, a high-precision multivariate nonlinear fitting formula is used.

[0029] Lift coefficient:

[0030] (3)

[0031] Drag coefficient:

[0032] (4)

[0033] Pitching moment coefficient:

[0034] (5)

[0035] Thrust coefficient:

[0036] (6)

[0037] where the reference value coefficient represents the aerodynamic coefficient value at a specific reference state, where C is the reference value of the lift coefficient, representing the lift coefficient at zero angle of attack and zero deflection; C is the reference value of the drag coefficient, representing the drag coefficient at zero angle of attack and zero deflection; C is the reference value of the pitching moment coefficient, representing the pitching moment coefficient at zero angle of attack and zero deflection; C is the reference value of the thrust coefficient, representing the thrust coefficient at zero angle of attack. Angle of attack related coefficients include: Cp, a is the first order angle of attack coefficient, representing the linear rate of change of the aerodynamic coefficient with respect to the angle of attack; Cp, a2 is the second order angle of attack coefficient, representing the quadratic non-linear rate of change of the aerodynamic coefficient with respect to the angle of attack; Cp, a3 is the third order angle of attack coefficient, representing the cubic non-linear rate of change of the aerodynamic coefficient with respect to the angle of attack. The control surface deflection dependent coefficients include: Cp, e is the first order elevator deflection coefficient, representing the linear rate of change of the aerodynamic coefficient with respect to the elevator deflection angle; Cp, e2 is the second order elevator deflection coefficient, representing the quadratic non-linear rate of change of the aerodynamic coefficient with respect to the elevator deflection angle. The fuel equivalence ratio dependent coefficients include: Cp, f is the fuel equivalence ratio coefficient, representing the rate of change of the aerodynamic coefficient with respect to the fuel equivalence ratio; Cp, ae is the angle of attack-fuel equivalence ratio cross coefficient, representing the influence of the coupling between the angle of attack and the fuel equivalence ratio on the aerodynamic coefficient. All coefficients are functions of velocity, and the expressions of these coefficients with respect to velocity can be obtained by fitting the results of mechanism models, CFD numerical calculations, wind tunnel test results or real flight data.

[0038] Cross terms in coefficients and are used to describe the aerodynamic / propulsive coupling effects. Cp, af represents the influence of the fuel equivalence ratio on the pitching moment under the influence of the angle of attack, embodying that the pitching moment is simultaneously affected by the angle of attack, the fuel equivalence ratio and the velocity; Cp, af represents the influence of the fuel equivalence ratio on the thrust under the influence of the angle of attack, embodying that the thrust is affected by the flight attitude (angle of attack), the fuel equivalence ratio and the internal flow shock system (velocity).

[0039] (1.4) Actuator dynamic modeling:

[0040] Elevator model: The elevator system is assumed to have extremely fast response speed, with bandwidth much higher than the dynamics of the vehicle body, so it is simplified as an ideal element:

[0041] (7)

[0042] where, is the actual elevator deflection angle, is the elevator command.

[0043] Engine model: The process from receiving the command to actually generating the thrust by the engine has thermal and mechanical inertia, and cannot be completed instantaneously. The response process from the fuel equivalence ratio command to the actual fuel equivalence ratio is modeled as a first order inertia element:

[0044] (8)

[0045] where, is the response time constant of the engine, which is a key parameter to characterize the response speed of the engine. This modeling method explicitly incorporates the engine dynamics into the system state equation.

[0046] Step (2) Design trajectory guidance outer loop to realize height and speed tracking control

[0047] The trajectory guidance outer loop is the outermost layer of the control process, working in the slow time scale, responsible for height and speed tracking control. The trajectory guidance outer loop converts the task-level height command and speed command into intermediate control variables: angle of attack command and speed rate command .

[0048] (2.1) Height tracking loop:

[0049] The height tracking loop adopts a PI controller to convert the height error into the angle of attack command:

[0050] (9)

[0051] where, is the height tracking proportional gain, is the height tracking integral gain, is the desired height, is the actual height.

[0052] (2.2) Speed tracking loop:

[0053] The speed tracking loop also adopts a PI controller to convert the speed error into the speed rate command:

[0054] (10)

[0055] where, is the speed tracking proportional gain, is the speed tracking integral gain, is the desired speed, is the actual speed.

[0056] Step (3) Design attitude stabilization outer loop to realize angle of attack tracking control

[0057] The attitude stabilization outer loop is the middle layer of the control process, working in the fast time scale, responsible for the tracking control of the angle of attack. The attitude stabilization outer loop receives the angle of attack command output by the trajectory guidance outer loop , and converts it into the pitch angle rate command .

[0058] The angle of attack tracking control law is as follows:

[0059] The attitude stabilization outer loop adopts a proportional controller to convert the angle of attack error into a pitch rate command:

[0060] (11)

[0061] where, is the angle of attack tracking proportional gain, is the angle of attack command (from the trajectory guidance outer loop), is the actual angle of attack, is the pitch rate command.

[0062] Step (4) constructs the attitude control inner loop to realize the acceleration level flight control

[0063] The attitude control inner loop is the core of the entire control process, working at the fastest time scale, directly controlling the speed and acceleration and pitch angle acceleration . The attitude control inner loop adopts a MIMO-INDI controller to realize the deep collaborative allocation of elevator and engine fuel equivalence ratio at the acceleration level.

[0064] (4.1) Desired dynamics generation:

[0065] First, the pitch rate command output by the attitude stabilization outer loop is converted into the desired pitch angle acceleration by a PID controller:

[0066] (12)

[0067] where, is the pitch rate tracking proportional gain, is the integral gain, is the derivative gain, is the actual pitch rate.

[0068] The desired value of the speed and acceleration comes directly from the trajectory guidance outer loop:

[0069] (13)

[0070] The desired dynamic vector, i.e., the desired acceleration is composed of:

[0071] (14)

[0072] The physical meaning of the desired dynamics: the vector represents the desired acceleration state of the aircraft at the current time. The rate of change of velocity that is desired is described, The pitch acceleration that is desired is described. The goal of the INDI controller is to make the actual acceleration track this desired dynamics by adjusting the elevator and fuel equivalence ratio.

[0073] (4.2) Control effectiveness matrix online computation:

[0074] Control effectiveness matrix is the core of the INDI controller, describing the instantaneous effect of control inputs on acceleration. This matrix is obtained by taking the partial derivative of the system dynamics equation with respect to control inputs.

[0075] First, define the controlled output vector:

[0076] (15)

[0077] Ideal physical control quantities:

[0078] (16)

[0079] Here, is a state quantity, not the underlying control command .

[0080] The basis of the INDI method is the approximation of the system dynamics after a first order Taylor expansion. The desired acceleration is equal to the measured acceleration at the previous time instant, plus the change brought by the ideal physical control quantities increment .

[0081] (17)

[0082] where, is the control increment, is the fuel equivalence ratio increment, is the elevator deflection increment; is the measured acceleration, is the acceleration value returned by the sensor, is the pitch acceleration value returned by the sensor. Parameters in bold represent vectors.

[0083] The control effectiveness matrix is defined as:

[0084] (18)

[0085] The specific expression of each element is obtained by taking the partial derivative of the dynamics equation of step (1):

[0086] : partial derivative of velocity acceleration with respect to fuel equivalence ratio

[0087] (19)

[0088] This term describes the direct contribution of increasing fuel equivalence ratio to velocity acceleration. Increasing fuel equivalence ratio leads to an increase in thrust, and the horizontal component of thrust (multiplied by ) directly acts in the velocity direction, resulting in velocity acceleration. The presence of the cross term indicates that this effect is influenced by the angle of attack.

[0089] : partial derivative of velocity acceleration with respect to elevator deflection

[0090] (20)

[0091] This term describes the effect of elevator deflection on velocity acceleration. Elevator deflection changes the drag, which in turn affects velocity. Note the negative sign, which indicates that increasing elevator deflection typically leads to an increase in drag and a decrease in velocity acceleration. The quadratic coefficient captures the nonlinear effect of large elevator deflection.

[0092] : partial derivative of pitch acceleration with respect to fuel equivalence ratio

[0093] (21)

[0094] This term describes the effect of increasing fuel equivalence ratio on pitch acceleration, which is a key manifestation of the propeller coupling. The first term represents the pitch moment due to the increase in thrust (propeller force arm effect), and the second term represents the direct change in aerodynamic moment due to the change in fuel equivalence ratio (inlet effect).

[0095] : partial derivative of pitch acceleration with respect to elevator deflection

[0096] (22)

[0097] This term describes the direct control effect of elevator deflection on pitch acceleration. This is the primary means of traditional attitude control. The quadratic coefficient reflects the nonlinear moment characteristics of large elevator deflection.

[0098] (4.3) Incremental control law solution:

[0099] The core idea of the INDI control law is to use the measurement information from the previous time step to compensate for model uncertainty in an incremental manner, achieving the desired acceleration dynamics.

[0100] The control law equation is designed in the form of equation (17), and the control increment is obtained by matrix inversion:

[0101] (23)

[0102] Ideal physical quantity update:

[0103] (24)

[0104] (25)

[0105] where, represents the expected fuel equivalence ratio of the next period, represents the expected elevator deflection of the next period. represents the control period sequence number.

[0106] Step (5) designs an engine compensation layer to handle the actuator dynamic delay

[0107] Since the engine has a significant response delay, directly using the ideal fuel equivalence ratio output by the attitude control inner loop as a command will cause the actual fuel equivalence ratio to lag, thereby reducing control performance. The engine compensation layer converts the ideal fuel equivalence ratio command into the actual fuel equivalence ratio command by using a feedforward + feedback strategy, and actively compensates for the engine dynamic delay.

[0108] (5.1) Feedforward compensator design:

[0109] The feedforward compensator compensates for the dynamic delay predictively based on a first-order inertia model of the engine. The engine dynamic equation is:

[0110] (26)

[0111] Reorganize:

[0112] (27)

[0113] This shows that, in order to make the actual fuel equivalence ratio reach the expected value and change at the expected rate , the fuel equivalence ratio command should be:

[0114] (28)

[0115] where, is the feedforward compensation term; is the expected fuel equivalence ratio, given by INDI; The desired rate of change of fuel equivalence ratio; The engine time constant is obtained through system identification.

[0116] However, in practical implementation, the current actual fuel equivalence ratio is usually used as the reference, because The ideal value may have deviations. Therefore, the feedforward compensation formula is modified as:

[0117] (29)

[0118] (5.2) Desired rate of change calculation:

[0119] It can be obtained by numerical differentiation:

[0120] (30)

[0121] where, is the control cycle number, is the control cycle.

[0122] (5.3) Feedback compensator design:

[0123] Although the feedforward compensator can predictively compensate for the delay, due to the uncertainty of the engine time constant and other unmodeled dynamics, simple feedforward compensation cannot completely eliminate tracking errors. Therefore, a feedback compensator is needed to further improve tracking accuracy.

[0124] The feedback compensator uses a PI controller to compensate based on the error between the ideal fuel equivalence ratio and the actual fuel equivalence ratio:

[0125] (31)

[0126] where, is the feedback term, is the proportional gain, is the integral gain.

[0127] The role of feedback compensation: the proportional term quickly responds to tracking errors, and the integral term eliminates steady-state errors. When the feedforward compensation deviates due to inaccurate parameters or external disturbances, the feedback compensator can automatically adjust to ensure that the final tracking .

[0128] (5.4) Total command synthesis:

[0129] The final fuel equivalence ratio command is composed of feedforward and feedback:

[0130] ​ (32)

[0131] Unfolded in full form:

[0132] (33)

[0133] Advantages of the present application:

[0134] 1、The present application realizes the active utilization of the aerodynamic and propulsion coupling effect in theory through the MIMO-INDI controller. The traditional decoupling control regards the aerodynamic and propulsion coupling as an interference that needs to be suppressed, while the present application regards it as a controllable resource. The non-diagonal elements of the control effectiveness matrix describe the influence of the elevator on the velocity and the influence of the fuel equivalence ratio on the attitude, and the optimal control combination that satisfies the velocity and attitude double targets is automatically calculated through the matrix inversion operation. This design enables the system to find the optimal solution in the global range, rather than independently controlling each subsystem locally. When a certain actuator is close to saturation or limited, the control law will automatically adjust the output of another actuator to compensate, realizing the true sense of flight-propulsion integrated collaborative control, and fundamentally breaking through the performance limitations of the traditional separate design method.

[0135] 2、The present application adopts the incremental form of nonlinear dynamic inverse control, which has natural robustness advantage in theoretical design. Unlike the traditional nonlinear dynamic inverse method which requires a complete and accurate dynamic model, INDI only relies on the control effectiveness matrix, i.e. the first-order partial derivative information, greatly reducing the requirement for model accuracy. The core of the incremental control law is to use the control input at the previous moment and the actual measured acceleration information to implicitly compensate the influence of all unmodeled dynamics, parameter perturbations and external disturbances in incremental form. This means that even if there is modeling error in the aerodynamic coefficients, or the engine characteristics deviate greatly from the nominal model, as long as the calculation of the control effectiveness matrix is basically accurate, the system can maintain good control performance. This design theoretically solves the fundamental problem of model uncertainty in hypersonic flight that is difficult to accurately model, providing a theoretical basis for robust control.

[0136] 3、The application realizes self-adaptation to wide flight envelope from design principle by calculating control effectiveness matrix in real time. Each element of the control effectiveness matrix is a function of current state, and is automatically updated with the change of state variables such as flight speed, height, angle of attack, etc. This design is essentially a self-scheduling mechanism of nonlinear control, and the control law can automatically perceive the change of system characteristics and adjust the control gain and distribution strategy accordingly. Unlike the traditional gain scheduling method which needs to design multiple controllers for working points offline and interpolate, the self-adaptation of the application is continuous and online, which theoretically avoids the performance jump or discontinuity problem that may occur at the switching point of gain scheduling. The quadratic term and cross term contained in the aerodynamic and propulsion coefficient fitting model can accurately describe the strong nonlinear characteristics, especially the cross term explicitly reflects the coupling relationship between the angle of attack and the fuel equivalence ratio, which provides a theoretical guarantee for the accurate calculation of the control effectiveness matrix. BRIEF DESCRIPTION OF DRAWINGS

[0137] Figure 1 is a high-speed hypersonic vehicle flight-propulsion integrated control structure block diagram based on INDI;

[0138] Figure 2 is a control method flowchart;

[0139] Figure 3 is a height tracking simulation curve;

[0140] Figure 4 is a Mach number tracking simulation curve;

[0141] Figure 5 is an angle of attack tracking simulation curve;

[0142] Figure 6 is a pitch angle velocity tracking simulation curve;

[0143] Figure 7 is a rudder deflection angle response curve;

[0144] Figure 8 is a fuel equivalence ratio response curve. DETAILED DESCRIPTION

[0145] The specific embodiments of the application will be further described below in combination with the drawings and technical solutions.

[0146] The overall framework of the high-speed hypersonic vehicle flight-propulsion integrated control method based on INDI of the embodiment is shown in Figure 1 .

[0147] 1、Task scenario and vehicle basic parameters

[0148] The embodiment takes a typical high-speed hypersonic vehicle climbing acceleration task as an example, and the vehicle needs to climb from the initial state (initial height = 25.0 km, initial velocity = 5.0 Ma) to target state (altitude command = 25.5 km, target velocity = 5.5 Ma) with 100 s mission duration. Vehicle basic parameters are set as: mass = 2500 kg, pitch moment of inertia = 18000 kg-m2, reference area = 1.0 m2, average aerodynamic chord = 3.5 m, thrust arm = 0.45 m, engine time constant = 0.8 s.

[0149] As Figure 2 shown, the control of the present application adopts a hierarchical architecture, which is executed in the following order: the control-oriented flight-propulsion integrated longitudinal dynamics model is the basis of the whole control, which provides system state equations and aerodynamic / propulsion characteristics calculation for each control layer. The model receives current state , , , , , ) and control input , ), and outputs state derivatives , , , , , ), which provides dynamic information for control law calculation.

[0150] 2. Foundation: control-oriented flight-propulsion integrated longitudinal dynamics model establishment

[0151] The longitudinal motion equations include velocity equation , flight-path angle equation , pitch rate equation , altitude equation , pitch angle equation , angle of attack equation , and engine dynamic equation .

[0152] Aerodynamic force and thrust are calculated through dimensionless coefficients: lift , drag , thrust , pitching moment , where dynamic pressure . Each dimensionless coefficient adopts multivariable nonlinear fitting, and thrust coefficient and pitching moment coefficient The paper introduces cross terms to describe the aerodynamic / propulsion coupling effect. This coupling effect is manifested in the following ways: changes in fuel equivalence ratio not only affect thrust, but also affect pitching moment through eccentric moment and changes in rear body pressure distribution; changes in angle of attack not only affect aerodynamic forces, but also affect thrust through changes in intake conditions.

[0153] 3. First layer: Implementation process of trajectory-guided outer ring

[0154] The outer loop, responsible for tracking and controlling altitude and velocity, updates every 0.1 seconds. This layer receives altitude commands. =25.5km, speed command =5.5Ma and current state =25.0km =5.0Ma. Altitude tracking uses a PI controller, and the control law is... ,in =0.00015 rad / m, =0.000005rad / (m·s). Speed ​​tracking also uses a PI controller, with the control law being... ,in =0.08(m / s²) / (m / s), =0.002(m / s²) / (m·s). The outer ring outputs the angle of attack command. and velocity change rate command Go to the next floor.

[0155] like Figure 3 As shown, the altitude steadily increased from 25.0 km to 25.5 km, with the response curve exhibiting an S-shaped characteristic. The first 50 seconds saw a slow ascent, followed by a main ascent phase from 50-80 seconds with a rate of ascent of approximately 8 m / s. After 80 seconds, the altitude stabilized at the target value, with a steady-state error of less than 20 m and no overshoot. Figure 4 As shown, the Mach number smoothly increases from 5.0 Mach to 5.5 Mach. Acceleration is slow for the first 30 seconds, with an average acceleration of approximately 3.2 m / s² during the main acceleration phase from 30 to 70 seconds. After 70 seconds, it gradually converges to the target value, with a steady-state error of less than 0.02 Mach. The velocity curve and altitude curve are highly synchronized, demonstrating multivariate coordinated control capabilities.

[0156] The attitude stabilization outer loop is responsible for tracking and controlling the angle of attack, with an update cycle of 0.02 seconds. This layer receives angle of attack commands from the trajectory guidance layer. and the current actual angle of attack A proportional controller is used, and the control law is: proportional gain =2.5 (rad / s) / rad. This controller converts the angle-of-attack error into a pitch rate command. The output is sent to the next layer, establishing a connection between outer loop trajectory tracking and inner loop attitude control.

[0157] 4. Second layer: Implementation process of attitude stabilization outer ring

[0158] like Figure 5 As shown, the angle of attack changes from an initial 3.0° to a final value of approximately 1.9°. It rises slightly to a peak of 3.18° in the first 15 seconds, then gradually decreases with increasing velocity and dynamic pressure. The change is smooth and continuous, with a tracking error of less than 0.1°, perfectly demonstrating the evolution of the aerodynamic characteristics of a hypersonic vehicle with varying flight conditions.

[0159] The attitude control inner loop is the core, employing a MIMO-INDI controller to coordinate the allocation of elevator and fuel equivalence ratio at the acceleration level, with an update cycle of 0.01 seconds. This layer receives velocity change rate commands from the trajectory guidance layer. Pitch rate command from attitude stabilization layer and the current measurement status ( , , , , ).

[0160] The expected dynamic generation module will convert the pitch rate command into the desired pitch acceleration via a PID controller. ,in =8.0, =0.5, =2.0. The expected value of velocity acceleration comes directly from the trajectory guidance layer: The desired dynamic vector is formed. .

[0161] 5. Third layer: Implementation process of attitude control inner loop

[0162] Control effectiveness matrix The online calculation module calculates matrix elements in real time based on the current flight status. ,in This indicates the effect of fuel equivalence ratio on velocity acceleration. This indicates the effect of the elevator on speed and acceleration. This indicates the effect of fuel equivalence ratio on pitch acceleration. This represents the effect of the elevator on pitch acceleration. The off-diagonal elements of the matrix... and This demonstrates the coupling relationship between the elevator and the fuel equivalence ratio, which is key to achieving coordinated control.

[0163] The INDI incremental control law calculation module calculates the formula based on... Calculate the control input increment, where , To measure the acceleration vector, the controller automatically calculates the optimal cooperative allocation scheme by inverting the matrix. The control command is updated as follows: , .in, The elevator offset corresponding to the current flight condition. This represents the trim fuel equivalence ratio corresponding to the current flight condition. This layer outputs elevator commands. Directly acting on the servo motor, ideal fuel equivalence ratio It is transferred to the engine compensation layer.

[0164] like Figure 6 As shown, the pitch rate response adjusts rapidly in the first 10 seconds, then exhibits small, periodic oscillations that are bounded and converge. The actual pitch rate maintains a very high degree of consistency with the commanded value. Figure 7 As shown, the elevator deflection angle changed from an initial approximately 4.7° to a final approximately 2.8°, ranging from 2.3° to 4.8°. The elevator exhibited a periodic adjustment characteristic, highly synchronized with the pitch rate fluctuations. The actual elevator deflection almost perfectly matched the commanded elevator deflection.

[0165] 6. Fourth Layer: Implementation Process of Engine Compensation Layer

[0166] The engine compensation layer handles the engine's dynamic latency, ensuring that the actual fuel equivalence ratio quickly and accurately tracks the ideal value. This layer receives the desired fuel equivalence ratio from the attitude control inner loop. Compared to current actual fuel equivalence A feedforward plus feedback strategy is adopted, with the feedforward compensation term... Predictive consideration of engine dynamics, feedforward gain =1.0. Feedback Compensation Item Eliminate tracking error, feedback gain =2.0. The final fuel equivalence ratio command is The output is sent to the engine controller. This strategy effectively compensates for the engine's 0.8-second time constant response delay.

[0167] like Figure 8 As shown, the fuel equivalence ratio initially decreases from 0.455 to 0.420 (for the first 20 seconds to ensure intake stability), then gradually increases to a final value of 0.505 (from 20 to 100 seconds to provide acceleration thrust). The fuel equivalence ratio curve exhibits slight periodic fluctuations, consistent with the elevator oscillation frequency, reflecting the coordinated control characteristics of the MIMO-INDI controller. The tracking error between the actual and commanded values ​​is less than 0.008 (relative error less than 2%), verifying the effectiveness of the compensation strategy.

[0168] 7. Fifth layer: Overall performance verification

[0169] Through the coordinated operation of the four control layers mentioned above, excellent overall performance is achieved. (Comparison) Figure 7 elevator and Figure 8 The response curves of the fuel equivalence ratio show a clear synergistic relationship between the two in the 50-100 second range. When the elevator deflection increases, the fuel equivalence ratio increases accordingly; when the elevator deflection decreases, the fuel equivalence ratio decreases accordingly, with highly consistent frequency and phase. This verifies that the MIMO-INDI controller automatically senses coupling effects and performs optimal synergistic allocation by controlling the off-diagonal elements of the effectiveness matrix. The physical mechanism is as follows: elevator deflection changes drag, affecting velocity acceleration; the controller automatically increases the fuel equivalence ratio through matrix inversion to compensate for drag loss and maintain the desired rate of velocity change. This is precisely the fundamental advantage of MIMO control over multiple SISO controllers in parallel.

[0170] A hierarchical, multi-timescale design is employed. The trajectory guidance layer (0.1s cycle) handles slow dynamics. Figure 3 Height and Figure 4 The velocity curve exhibits a smooth, long-period variation characteristic. The attitude stabilization layer (0.02s period) handles medium-speed dynamics. Figure 5 The angle of attack change frequency falls between the altitude / velocity and pitch rate. The attitude control inner loop (0.01s cycle) handles rapid dynamics. Figure 6 Pitch rate and Figure 7 The elevator exhibits a rapid response characteristic. Command transmission between all levels is smooth, with no coupling issues between levels. The high-frequency operation (100Hz) of the inner loop does not generate high-frequency oscillations on the variable response curve of the outer loop, verifying the success of the bandwidth separation design.

[0171] The performance indicators are summarized as follows: In terms of tracking accuracy, the steady-state altitude error is less than 20m (less than 0.08%), the steady-state speed error is less than 0.02Ma (less than 0.36%), the angle-of-attack tracking error is less than 0.1°, the pitch rate tracking error is less than 0.005° / s (less than 10%), and the fuel equivalence ratio tracking error is less than 0.008 (less than 2%). Regarding response performance, the altitude adjustment time is approximately 80 seconds, the speed adjustment time is approximately 85 seconds, all variables show no overshoot, and stability is excellent. In terms of control smoothness, the elevator variation range of 2.3°~4.8° is far from the saturation limit, with a maximum change rate of approximately 0.15° / s, and the fuel equivalence ratio variation range is 0.420~0.505. All control variables change smoothly and continuously. It maintains good performance throughout the entire flight envelope and exhibits excellent robustness to model uncertainties and actuator delays.

[0172] In summary, this invention achieves deep coordinated allocation of elevator and fuel equivalence ratio at the acceleration level through a MIMO-INDI controller, actively utilizing aerodynamic / propulsion coupling effects and overcoming the limitations of traditional separate design methods. The incremental control law implicitly compensates for model uncertainties, the online calculation of the control effectiveness matrix automatically adapts to changes in flight state, the engine compensation layer effectively handles actuator delays, and the hierarchical architecture enables multi-timescale control. This successfully solves key technical challenges in hypersonic vehicle control, such as aerodynamic / propulsion coupling, strong nonlinearity, model uncertainty, and actuator delays.

Claims

1. A hypersonic vehicle flight-thrust integrated control method based on INDI, characterized in that, The steps are as follows: Step (1) Establish a longitudinal dynamic model of hypersonic aircraft with integrated flight and thrust for control. (1.1) Longitudinal dynamic equation: The longitudinal motion of the aircraft is described by the force balance equations for the motion of the center of mass and the moment balance equations for the motion of the center of mass: (1) ； Among them: superscript " " represents the first derivative; For flight speed, For the trajectory inclination angle, For pitch rate, For flight altitude, The pitch angle, For the angle of attack, This is the actual fuel equivalence ratio. For fuel equivalence ratio command, The engine response time constant. For thrust, As resistance, For lift, For pitching moment, For the mass of the aircraft, For pitch rotation inertia, It is the acceleration due to gravity; (1.2) Force and torque model: All aerodynamic forces, thrust, and torque are calculated using dimensionless coefficients, establishing unified expressions for force and torque: (2) ； in: For dynamic pressure, air density, For reference area, The reference length is the average aerodynamic chord length. The lever arm of the thrust relative to the center of mass. , , , These are dimensionless coefficients for lift, drag, thrust, and pitching moment, respectively. (1.3) Fitting model for aerodynamic and propulsion coefficients: Lift coefficient: (3) ； Drag coefficient: (4) ； Pitch moment coefficient: (5) ； Thrust coefficient: (6) ； Among them, the benchmark coefficient This represents the aerodynamic coefficient value under a specific reference state, where This is the baseline value for the lift coefficient; This is the baseline value for the drag coefficient; This is the reference value for the pitch moment coefficient; This is the baseline value for the thrust coefficient; It is the first-order angle of attack coefficient; It is the second-order angle of attack coefficient; It is the third-order angle of attack coefficient; This is the first-order rudder deflection coefficient; This is the second-order rudder deflection coefficient; This refers to the fuel equivalence ratio coefficient. Angle of attack - fuel equivalence ratio cross coefficient; coefficient This indicates the effect of fuel equivalence ratio on pitching moment under the influence of angle of attack; coefficient This indicates the effect of fuel equivalence ratio on thrust under the influence of angle of attack; (1.4) Dynamic modeling of actuators: Elevator model: (7) ； in, This is the actual elevator deflection angle. Elevator command; Engine model: Fuel equivalence ratio command To actual fuel equivalence ratio The response process is modeled as a first-order inertial element: (8) ； in, It is the engine's response time constant; Step (2) Design the trajectory guidance outer loop to achieve height and speed tracking control The trajectory guidance outer loop is the outermost layer of the control process, responsible for altitude and speed tracking control; the trajectory guidance outer loop transmits mission-level altitude commands. and speed command Converted to intermediate control variable: Angle of attack command and velocity change rate command ; Step (3) Design an attitude stabilization outer loop to achieve angle-of-attack tracking control The attitude stabilization outer loop is the intermediate layer of the control process, responsible for tracking the angle of attack; the attitude stabilization outer loop receives the angle of attack command output by the trajectory guidance outer loop. and convert it into pitch rate command. ; Step (4) Construct the attitude control inner loop to achieve acceleration-level flight thrust control. The attitude control inner loop directly controls velocity and acceleration. and pitch acceleration The attitude control inner loop uses a MIMO-INDI controller to achieve elevator control at the acceleration level. Compared to engine fuel equivalence ratio Deep collaborative allocation; Step (5) Design the engine compensation layer to handle the dynamic delay of the actuator. The engine compensation layer uses a feedforward + feedback strategy to optimize the ideal... Convert to actual fuel equivalence ratio command Actively compensates for engine dynamic delay.

2. The integrated flight and thrust control method for hypersonic vehicles based on INDI according to claim 1, characterized in that, Step (2) is as follows: (2.1) Altitude tracking loop: The altitude tracking loop uses a PI controller to convert altitude error into angle of attack command. (9) ； in, To highly track the proportional gain, To highly track the integral gain, For the desired height, This refers to the actual height. (2.2) Speed ​​tracking loop: The speed tracking loop also uses a PI controller to convert speed error into speed change rate commands: (10) ； in, For speed tracking proportional gain, For speed tracking integral gain, For the desired speed, This refers to the actual speed.

3. The integrated flight and thrust control method for hypersonic vehicles based on INDI according to claim 1, characterized in that, In step (3), the angle-of-attack tracking control law is as follows: The attitude stabilization outer loop uses a proportional controller to convert the angle of attack error into a pitch rate command: (11) ； in, For angle-of-attack tracking proportional gain, Angle of attack command, originating from the outer loop of the trajectory guidance. This is the actual angle of attack. This is the pitch rate command.

4. The integrated flight and thrust control method for hypersonic vehicles based on INDI according to claim 1, characterized in that, Step (4) is as follows: (4.1) Expected dynamic generation: First, the pitch rate command output by the attitude stabilization outer loop is... The desired pitch angle acceleration is converted by a PID controller. : (12) ； in, To track the proportional gain for pitch rate, For integral gain, For differential gain, This refers to the actual pitch rate; Expected value of velocity acceleration Directly from the outer loop of the trajectory guidance: (13) ； This constitutes the desired dynamic vector, i.e., the desired acceleration. : (14) ； The physical meaning of expected dynamics: vector This represents the desired acceleration state of the aircraft at the current moment; Describe the desired rate of change of velocity. Describe the desired pitch angle acceleration; the goal of the INDI controller is to make the actual acceleration track this desired dynamic by adjusting the elevator and fuel equivalence ratio. (4.2) Control effectiveness matrix Online calculation: Validity Matrix It describes the instantaneous effect of the control input on the acceleration; this matrix is ​​obtained by taking the partial derivatives of the system dynamics equations with respect to the control input. First, define the controlled output vector: (15) ； Ideal physical control quantity: (16) ； Expected acceleration It is equal to the acceleration measured at the previous moment. Plus the increment of the ideal physical control quantity The changes brought about; (17) ； in, To control the increment, For fuel equivalence ratio increment, For elevator deflection increment; For the measured acceleration, The acceleration value returned by the sensor. The pitch acceleration value returned by the sensor; The control effectiveness matrix is ​​defined as follows: (18) ； The specific expressions for each element are obtained by taking the partial derivative of the dynamic equation in step (1): Partial derivative of velocity acceleration with respect to fuel equivalence ratio (19) ； Partial derivative of velocity acceleration with respect to elevator (20) ； The partial derivative of pitch acceleration with respect to fuel equivalence ratio (21) ； The partial derivative of pitch acceleration with respect to the elevator (22) ； (4.3) Solving the incremental control law: By using the measurement information from the previous moment, the model uncertainty is compensated for in an incremental manner to achieve the desired acceleration dynamics; The control law equation is designed in the form of equation (17), and the control increment is obtained by matrix inversion: (23) ； Ideal physical quantity update: (24) ； (25) ； in, This indicates the expected fuel equivalence ratio for the next cycle. This indicates the expected elevator deflection for the next cycle; Indicates the control cycle number.

5. The integrated flight and thrust control method for hypersonic vehicles based on INDI according to claim 1, characterized in that, Step (5) is as follows: (5.1) Design of feedforward compensator: The feedforward compensator, based on the engine's first-order inertial model, predictively compensates for dynamic delays; the engine's dynamic equations are: (26) ； Reorganized: (27) ； To achieve the actual fuel equivalence ratio Achieve the expected value and at the expected rate of change The change, the fuel equivalence ratio instruction is: (28) ； in, This is a feedforward compensation term; The desired fuel equivalence ratio; The expected rate of change of fuel equivalence ratio; The engine time constant; Using the current actual fuel equivalence ratio Instead As a benchmark, because This is an ideal value and contains deviations; therefore, the feedforward compensation formula is revised as follows: (29) ； (5.2) Calculation of expected rate of change: Obtained through numerical differentiation: (30) ； in, To control the cycle number, To control the cycle; (5.3) Feedback compensator design: Add a feedback compensator to further improve tracking accuracy; The feedback compensator uses a PI controller to compensate for the error between the ideal fuel equivalence ratio and the actual fuel equivalence ratio. (31) ； in, For feedback items, For proportional gain, For integral gain; (5.4) Overall instruction synthesis: The final fuel equivalence ratio command consists of two parts: feedforward and feedback. (32) ； Expand to full form: (33)。

Images