Intake-constrained hypersonic vehicle flight-propulsion integrated protection control method

Abstract

The present application belongs to the technical field of hypersonic vehicle control, and particularly relates to a hypersonic vehicle flight-propulsion integrated protection control method considering air intake constraints. The present application introduces a dynamic adjustment coefficient varying with the flight Mach number on the basis of the existing current linear safety margin representation formula, constructs a nonlinear safety margin formula, so as to more accurately describe the safety margin redundancy under different Mach numbers. Further, the fuel equivalence ratio and the air intake non-starting safety margin are taken as additional state variables for dimension expansion, a dimension expansion modeling capable of describing the bidirectional coupling effect of flight and propulsion is constructed, and a constraint prediction control problem is established, so that the controller can adjust the control input in advance according to the change of the safety margin in the prediction time domain, and the active protection of the air intake non-starting risk is realized.

Description

Technical Field

[0001] This invention belongs to the field of hypersonic vehicle control technology, specifically relating to an integrated protection and control method for hypersonic vehicles that takes into account air intake constraints. Background Technology

[0002] With the development of aerospace technology, hypersonic vehicles have gradually become a focus of competition among countries due to their broad application prospects. Hypersonic vehicles employ an integrated design of a forebody compression inlet and propulsion system, resulting in strong bidirectional coupling between aerodynamic characteristics and the propulsion system. This leads to typical characteristics of "strong coupling, strong nonlinearity, rapid time-varying, and static instability." Simultaneously, the complex flow field within the scramjet engine inlet, along with shock wave-boundary layer interference and strong coupling between combustion and flow, easily causes shock train forward movement. When the back pressure in the combustion chamber exceeds a critical value, the shock train is pushed out of the inlet, causing inertia, a sudden drop in thrust, or even a flight accident. This is a core risk factor restricting flight safety.

[0003] The existing technology has the following main shortcomings: Firstly, in terms of modeling, traditional dynamic analysis usually treats the aerodynamic characteristics and propulsion system of an aircraft as independent systems and decouples them. However, since hypersonic aircraft adopt an integrated flight and propulsion configuration, the flight and propulsion systems of the aircraft will generate strong coupling phenomena in hypersonic state. Flight and attitude are closely related. Decoupling modeling ignores the interaction between flight and propulsion, which will lead to an inaccurate reflection of the aircraft's state and the model accuracy cannot be guaranteed.

[0004] Secondly, regarding control, PID control, due to its simple structure, strong robustness, and ease of implementation, has been widely used in flight control. This control method is very suitable for real-time control requirements compared to modern control methods, which is of great significance for resource-constrained flight control systems. However, when facing complex systems with strong coupling and strong nonlinear characteristics, such as hypersonic vehicles, the limitations of traditional PID control gradually become apparent. The traditional discrete PID control architecture designs each control loop independently, which cannot effectively handle the cross-coupling effect between flight and thrust, and is prone to oscillation or divergence under coupling, leading to a serious deterioration in control performance.

[0005] Meanwhile, in terms of intake protection, the assessment of the safety margin of intake not starting is mostly based on the linear relationship of shock train length. However, such methods ignore the nonlinear variation of shock train length with back pressure. At low Mach numbers, the sensitivity of shock train to back pressure will be significantly enhanced, and the linear formula will seriously underestimate the actual safety risks. Moreover, under the same shock train length ratio, there are significant differences in the actual safety redundancy under different Mach number conditions.

[0006] Existing related patent technologies also have certain limitations: Patent application CN202410451639.5 proposes a method for controlling the outlet flow of an aircraft inlet. This method primarily achieves precise control of the inlet outlet airflow by dynamically adjusting the inlet outlet pressure during CFD time-propagation calculations. The method focuses on outlet boundary pressure correction during numerical simulation, mainly serving inlet flow field calculations and flow matching. It does not establish an integrated flight-propulsion closed-loop protection control mechanism for the actual flight process of the aircraft, nor does it incorporate the inlet's safety margin from starting as a state variable into the aircraft's dynamic equations.

[0007] Patent application CN202110370886.9 proposes a design method for a two-dimensional adjustable air inlet, a variable geometry air inlet design method using the throat Mach number as the control variable. This method can be used for the profile design and performance prediction of two-dimensional adjustable air inlets under multiple operating conditions. However, this method focuses on air inlet geometry design and performance improvement at non-design points, belonging to the technical solution at the air inlet design level. It has not yet unified flight state, propulsion state, and air inlet safety state into an integrated flight-propulsion dynamics model, nor has it formed an online predictive protection control system for non-starting risks.

[0008] Patent application CN201711119283.1 proposes a multi-system correlation modeling method for hypersonic vehicle control system design. This method starts from the control system design requirements, establishing correlations between multiple subsystems such as kinematic equations, aerodynamic layout systems, propulsion systems, and structural elastic systems. It then uses correlation indicators to screen subsystems with significant impact on control design, thus providing a model foundation for hypersonic vehicle control law design and integrated optimization. While this method embodies the concept of multi-system coupled modeling for hypersonic vehicles, its focus is on multi-system correlation modeling and control performance analysis during the design phase. It struggles to achieve unified coordination between flight control, propulsion regulation, and inertial induction protection.

[0009] Patent application CN202110861739.1 proposes an air intake and method based on a controllable venting slot switch. By incorporating a controllable venting slot in the air intake fairing, redundant flow under off-design conditions is discharged, thereby improving air intake start-up performance, reducing start-up Mach number, and widening the air intake's operating range. However, this method primarily relies on local structural adjustments within the air intake to achieve start-up and flow matching, and fails to establish a unified predictive control relationship between flight control, propulsion regulation, and the safety margin for air intake inactivity from the perspective of overall aircraft dynamics.

[0010] In summary, existing technologies mostly focus on inlet outlet flow regulation, variable geometry inlet design, and improvement of bleeder structures, which can improve inlet start-up performance or flow matching capability to a certain extent. However, they generally lack unified modeling of flight state, propulsion state, and inlet safety state, and do not embed the inlet non-start safety margin as an augmented state and predictive control constraint into the flight-propulsion integrated control framework. Therefore, this invention introduces fuel equivalence ratio and inlet non-start safety margin for dynamic expansion, and constrains the safety margin in the predictive time domain, thereby realizing the transformation from local structural adjustment or ex-post response to flight-propulsion integrated active protection. Summary of the Invention

[0011] The purpose of this invention is to provide an integrated flight-thrust protection and control method for hypersonic vehicles that considers inlet constraints. Building upon existing linear safety margin characterization formulas, a dynamic adjustment coefficient varying with the flight Mach number is introduced to construct a nonlinear safety margin formula, more accurately describing safety margin redundancy under different Mach numbers. Furthermore, fuel equivalence ratio and inlet non-start safety margin are added as additional state variables to construct a topological model that describes the two-way coupling effect between flight and thrust. A constrained predictive control problem is then established, enabling the controller to adjust the control input in advance based on changes in the safety margin in the prediction time domain, thus achieving proactive protection against inlet non-start risk.

[0012] To achieve the above objectives, the present invention adopts the following technical solution: A method for integrated flight and thrust protection and control of hypersonic vehicles considering air intake constraints, comprising the following specific steps: Step 1: Construct a longitudinal dynamic model of the hypersonic vehicle integrating flight and thrust; By introducing the fuel equivalence ratio state, a hypersonic vehicle integrated flight-propulsion longitudinal dynamics model is established to describe the coupling relationship between the longitudinal motion state of the vehicle and the propulsion system state. Furthermore, its small-disturbance linearized expression is obtained, providing a basic model for the state topology, local linearization, and constraint predictive control of the air intake safety margin.

[0013] In hypersonic vehicles, the influence of attitude on thrust in the flight-thrust coupling effect is mainly reflected in the influence of angle of attack on thrust. The influence of thrust on attitude manifests in two ways: first, it generates an eccentric torque that does not pass through the center of mass, causing the vehicle to rotate; second, the high-temperature, high-speed airflow expelled from the tail nozzle generates additional aerodynamic forces and torques on the lower surface of the vehicle's rear body. It can be seen that the flight-thrust coupling effect undoubtedly increases the complexity of controlling the vehicle's longitudinal motion and also makes it difficult for traditional longitudinal motion equations to adequately describe the flight state of hypersonic vehicles during longitudinal motion. Therefore, the fuel equivalence ratio is incorporated into the longitudinal motion equations for dynamic modeling; the expression for the fuel equivalence ratio is as follows: (1) in, Fuel equivalence ratio, For speed, For the angle of attack, This refers to the fuel equivalence ratio command. This refers to the physical relationship between fuel equivalence ratio and speed, angle of attack, and fuel equivalence ratio command.

[0014] Ignoring lateral parameters, the force analysis of the longitudinal channel of the hypersonic vehicle yields the following expanded longitudinal dynamic model of the hypersonic vehicle's integrated flight and thrust system: (2) Among them, the superscript " " represents the first derivative, It is the pitch angular velocity. It is the pitch moment of inertia. It is the pitching moment; It is the trajectory inclination angle; For the mass of the aircraft, and These represent thrust and lift, respectively. Represents gravitational acceleration; The pitch angle, Flight altitude; It is the pitching moment generated by the propulsion system. By introducing the fuel equivalence ratio as a state variable, the traditional equations are expanded. This state variable connects the flight state with the propulsion system, and can quantify the flight-propulsion coupling relationship to a certain extent.

[0015] Applying a small perturbation to equation (2) linearly: (3) in, For flight speed disturbance, Let be the trajectory tilt angle disturbance. The pitch angular velocity disturbance is... This represents the fuel equivalence ratio disturbance. This represents the elevator deflection angle disturbance. For the angle of attack disturbance, This represents the fuel equivalence ratio command disturbance. Let be the partial derivative of thrust with respect to velocity. Let be the partial derivative of thrust with respect to angle of attack. This is the partial derivative of thrust with respect to fuel equivalence ratio; Let be the partial derivative of the axial force with respect to the velocity. Let be the partial derivative of the axial force with respect to the angle of attack. Let be the partial derivative of axial force with respect to fuel equivalence ratio. This is the partial derivative of the axial force with respect to the elevator deflection angle; Let be the partial derivative of the normal force with respect to the velocity. Let be the partial derivative of the normal force with respect to the angle of attack. Let be the partial derivative of the normal force with respect to the fuel equivalence ratio. This is the partial derivative of the normal force with respect to the rudder deflection angle; The disturbance force is caused by the change in the projection of gravity along the axial direction. The disturbance force caused by the change in the projection of gravity onto the normal direction; This is the partial derivative of the aerodynamic pitching moment with respect to velocity. Let be the partial derivative of the thrust pitching moment with respect to velocity. Let be the partial derivative of the aerodynamic pitching moment with respect to the angle of attack. Let be the partial derivative of the thrust pitching moment with respect to the angle of attack. This is the partial derivative of the aerodynamic pitching moment with respect to the pitching angular velocity. Let be the partial derivative of the aerodynamic pitching moment with respect to the elevator deflection angle. Let be the partial derivative of the thrust pitching moment with respect to the fuel equivalence ratio. This is the additional eccentric torque disturbance term caused by gravity due to the shift of the center of mass or the change of the point of application of force.

[0016] To simplify the expression, the partial derivative terms are presented as dynamic coefficients. The dynamic coefficients are shown below: , , , , , , , , , , , , , , , , , , , , , , , , , , , Wherein, the first digit of the subscript is the equation number of formula (4); the second digit of the subscript is the sequence number of the small perturbation bias: Corresponding to serial number 1, Corresponding to serial number 2, Corresponding to serial number 3, Corresponding to serial number 4. The rate of change of fuel mass flow rate, The rate of change of air mass flow rate, For air mass flow rate, Fuel mass flow rate, For an appropriate oil-gas ratio, T This is the equivalent time constant.

[0017] Step 2: Define the safety margin for the intake manifold not to start; A dynamic adjustment coefficient is introduced based on the safety margin of the shock train length. Furthermore, by combining the characteristic Mach number of the intake duct, a safety margin for non-starting is constructed to quantitatively characterize the safety level of the intake duct from the non-starting state, providing an evaluation index for safety margin prediction and active protection.

[0018] The scramjet engine inlet typically consists of multiple stages of oblique shock waves and an isolator section. Its normal start-up depends on the dynamic balance between the shock wave system within the inlet and the combustion chamber back pressure. When combustion chamber heat release increases or the flight attitude changes significantly, the combustion chamber back pressure rises, causing the shock wave train within the isolator section to move upstream, resulting in a decrease in the inlet's effective flow capacity. When the leading edge of the shock wave train crosses the isolator inlet or the shock wave system cannot maintain a stable structure, the inlet fails to start, manifested as a sharp increase in total pressure loss, a sudden decrease in intake mass flow rate, and a significant drop in engine thrust. Therefore, the spatial location and physical length of the shock wave train within the isolator section are key physical quantities characterizing the inlet's operating state and the risk of failure to start. The length of the shock wave train... The empirical formula is: (4) in, The radius of the isolation segment, To determine the Mach number at the front end of the pre-burning shock train, The Mach number at the exit of the isolation section. This is the specific heat ratio of the gas within the isolation section.

[0019] The inlet non-starting safety margin is defined based on the length of the shock train. : (5) in, The length of the isolation segment. The coefficients are dynamically adjusted; its core design logic is based on the influence of flight Mach number on the stability of the flow field inside the inlet, dynamically adjusting the safety margin. shock train length The degree of sensitivity. It is mainly affected by the flight Mach number, because the flight Mach number directly determines the compression intensity of the incoming flow in the inlet and the Mach number distribution of the internal flow field. The lower the Mach number, the smaller the incoming kinetic energy, and the more easily the shock wave train moves forward due to the increase in back pressure, resulting in a higher risk of not starting and requiring a larger [massage / enhancing factor]. Enhance early warning sensitivity; the higher the Mach number, the stronger the incoming flow compression effect, the better the flow field stability, and the greater the potential for reduction. Balancing safety and performance.

[0020] Step 3: Establish a flight-thrust coupled augmented model that considers the safety margin of the inlet not starting and a reduced-order prediction model for roll optimization; To incorporate the inlet inactivity protection into the system, the inlet inactivity safety margin is introduced into the integrated flight-thrust longitudinal dynamics model obtained in step 1, and a flight-thrust coupling augmented model considering the inlet inactivity safety margin is established: (6) in, This represents the combustion chamber expansion ratio. (Function) The relationship between the safety margin for preventing the intake manifold from starting and speed, angle of attack, fuel equivalence ratio, and combustion chamber expansion ratio.

[0021] Because the complete flight-thrust coupled augmented model that considers the safety margin of the inlet not starting has a high order, directly using it for online prediction optimization would impose a large computational burden. Therefore, a reduced-order prediction model is used for rolling optimization. Prediction status as follows: (7) Corresponding control input Selected as: (8) Using the current flight state and control input as the operating point, the flight-thrust coupled augmented model considering the safety margin of inlet inactivity is linearized with small perturbations. After selecting the reduced-order prediction state and control input, the following local linear prediction relationship is obtained: (9) in, These are the partial derivatives of the inert safety margin with respect to speed, angle of attack, fuel equivalence ratio, and combustion chamber expansion ratio, respectively. The disturbance is the rate of change of pitch angular velocity. The disturbance is the rate of change of angle of attack. This represents the disturbance of the rate of change in fuel equivalence ratio. The disturbance is the rate of change of the safety margin for the intake manifold not starting. This represents the disturbance of the combustion chamber expansion ratio.

[0022] for The expressions for the dynamic coefficients are not explicitly presented; instead, a discrete prediction model under the current operating condition is obtained using online local linearization. Specifically, a reduced-order prediction state is selected, and small perturbations are applied to the state and control variables near the current state and control input, respectively. The difference is then calculated using the propagation results.

[0023] After performing online local linearization and discretization near the current operating point, the reduced-order prediction model can be written as: (10) in, , These are the discrete state matrix and control matrix under the current operating conditions, respectively. The affine bias term is generated by the local linearization of the nonlinear model. For the first The reduced-order predicted state vector at each sampling time, For the first The control input vector at each sampling time.

[0024] Due to the strong nonlinearity and rapid time-varying characteristics of hypersonic vehicles, the matrix... , and bias terms All data are updated online according to the current flight status. The advantage is that it avoids directly calling the complete nonlinear flight-thrust coupling augmented model in the controller for multi-step optimization, while still reflecting the flight-thrust coupling dynamics near the current operating condition at each sampling time.

[0025] Step 4: Design an integrated protection and control method for the flight propulsion system that takes into account air intake constraints; The design of the constraint predictive controller realizes integrated flight and thrust protection control that takes into account air intake constraints; in terms of flight performance, it ensures that the aircraft's angle of attack can track the given command, so that the attitude response meets the mission requirements; in terms of air intake safety, it ensures that the inert safety margin is always kept above the given safety lower limit, and avoids the air intake from entering the inert state.

[0026] Therefore, the design goal of the constrained predictive controller is to minimize state tracking error and control input cost in the prediction time domain, while explicitly satisfying various constraints. When the inlet non-start safety margin in each prediction step within the prediction time domain is far from the lower safety limit, the constrained predictive controller mainly ensures angle-of-attack tracking performance. When the prediction results show that the safety margin is gradually approaching the lower safety margin, the constrained predictive controller should adjust control variables such as elevator deflection angle, fuel equivalence ratio command, and combustion chamber expansion ratio in advance to suppress further reduction of the safety margin.

[0027] Establish a rolling optimization problem with a constrained predictive controller: (11) in, For performance index functions, To predict the step size.

[0028] The performance index function is defined as follows: (12) in, The state weight matrix is... To control the input weight matrix, To control the incremental weight matrix, Let be the prediction step number in the prediction time domain, and ; Indicates based on the first The prediction obtained from the known state and control information at the sampling time is the first... The state vector at each sampling time. For the first The reference state vector corresponding to each sampling time; Indicates the first The prediction obtained at the sampling time, which affects the future... The control input vector at each sampling time; The change in control input at adjacent sampling times, i.e.: (13) That is, for the first term in the prediction time domain The prediction step, the control increment is from the first prediction step. The control input at the predicted sampling time and the first The difference between the control inputs at each prediction sampling time is determined.

[0029] During the optimization process, the predicted state must satisfy discrete dynamic constraints: (14) in, For the first The discrete state matrix is ​​obtained by online local linearization and discretization at each sampling time point based on the current operating point. For the first The discrete control matrix corresponding to each sampling time.

[0030] The control input constraints are: (15) in, , These are the lower and upper bounds of the control input, respectively. Input constraint shrinkage amount.

[0031] State constraints are represented as: (16) in, , These are the upper bound and the lower bound of the predicted state, respectively. This represents the state constraint contraction amount.

[0032] Inlet safety margin constraints are the core of active protection control. Because predictive measures are employed, the lower limit of the safety margin can be set lower than traditional protection methods to meet flight requirements. The inlet safety margin constraint in the predictive time domain can be expressed as: (17) in, For the first The future flight state and control information at the sampling time are used to predict the next sampling time. The safety margin for the air intake not to start at each sampling time. This is the minimum allowable safety margin for the intake manifold not to start.

[0033] The inlet safety margin constraint in the prediction time domain requires that the overall safety margin at each step in the future prediction time domain be no less than the lower limit of the safety boundary. Compared to constraining the margin only at the current moment, the safety margin constraint in the prediction time domain can identify the downward trend of the safety margin in advance and correct the control input before the safety actually enters the danger zone.

[0034] At each sampling time, the constrained predictive controller solves for the future. The optimal control sequence for the step: (18) in, For the first Future obtained at each sampling time Step-optimal control input sequence, For the first The optimal control input vector is obtained at each sampling time and applied to the current sampling period. For the first The prediction obtained at the sampling time is the first The optimal control input vector at each sampling time.

[0035] In actual execution, only the first control variable in the optimal control sequence is applied to the controlled object: (19) At the next sampling time, the constrained predictive controller re-performs online local linearization and discretization of the flight-thrust coupled augmented model established in step 3, obtaining the... The local discrete prediction model corresponding to each sampling time is used, and the optimization problem is solved again, thus forming a rolling time-domain closed-loop control. In this way, the constrained predictive controller can continuously modify the control input according to the real-time changes in flight status and safety margin, thereby actively suppressing the risk of inlet start-up.

[0036] The beneficial effects of this invention are: (1) In terms of modeling, the fuel equivalence ratio and the safety margin of the inlet not starting are introduced as topological state variables, and a hypersonic flight-propulsion integrated protection and control model considering inlet constraints is established. Compared with the traditional modeling method that only considers the flight state or simplifies the propulsion system to an external input, this invention can describe the coupling relationship between the flight state, propulsion state and inlet safety state under the same dynamic framework, which is more in line with the actual motion characteristics under the hypersonic vehicle flight-propulsion integrated configuration.

[0037] (2) The inlet non-starting safety margin is defined, and a nonlinear safety margin formula is established based on the relationship between the shock train length and the isolation section length. The nonlinear safety margin formula, by introducing a dynamic adjustment coefficient, can reflect the different impacts of shock train forward movement on the non-starting risk under different Mach number conditions, avoiding the problem of the traditional linear margin formula underestimating the safety risk under low Mach number or strong back pressure disturbance conditions, thereby improving the accuracy of inlet non-starting risk assessment.

[0038] (3) This invention integrates the inlet safety margin constraint, control input constraint, state constraint, and control increment constraint into a rolling optimization problem, enabling the controller to explicitly satisfy the inlet non-starting safety constraint while ensuring basic flight performance such as angle of attack tracking. Compared with the traditional threshold-triggered protection method, this invention can correct the control input in advance before the safety margin falls below the danger threshold, realizing the transformation from ex-post response to predictive protection. Attached Figure Description

[0039] Figure 1 This is the overall system architecture flowchart; Figure 2 It is an angle-of-attack tracking comparison curve; Figure 3 It is a curve showing the change in the safety margin of the air intake; Figure 4 It is a graph showing the change in flight speed; Figure 5 It is a graph showing the change in elevator deflection angle; Figure 6 It is a curve showing the change in fuel equivalence ratio. Detailed Implementation

[0040] The specific embodiments of the present invention will be further described below with reference to the accompanying drawings and technical solutions.

[0041] The present invention provides an integrated protection and control method for hypersonic flight thrust considering inlet constraints, the basic process of which is as follows: Figure 1 As shown. The specific steps of this embodiment are as follows: Step (1) Derivation of small perturbation coefficients in the integrated flight and propulsion dynamics model When linearizing for small perturbations and introducing dynamic coefficients, each coefficient has a clear physical meaning. Some coefficients correspond to the aerodynamic derivatives in traditional longitudinal dynamic models, primarily describing the influence of flight state and control surface inputs on longitudinal motion; others are dynamically related to the fuel equivalence ratio, used to describe the impact of propulsion system state changes on aircraft motion. It is precisely related to the fuel equivalence ratio... The relevant coefficients enable the linearized model to reflect the flight-thrust coupling characteristics of hypersonic vehicles. This is to obtain the fuel equivalence ratio. A linearized expression of the change requires the fuel equivalence ratio. The dynamic equations are expanded with small perturbations. Due to the complex relationship between the actual fuel equivalence ratio and speed, angle of attack, and fuel equivalence ratio command, this invention still uses partial derivatives to describe its local variation.

[0042] The fuel equivalence ratio is defined as follows: (20) Differentiating both sides of the formula with respect to time yields the rate of change of fuel equivalence ratio: (twenty one) Therefore, the introduction of the fuel equivalence ratio results in two new dynamic coefficients. and The calculation formula is as follows: (twenty two) (twenty three) Fuel mass flow rate change rate It is only affected by changes in throttle opening, while the rate of change of air mass flow rate It is a function of velocity and angle of attack, and its partial derivatives can be taken with respect to these two variables. However, considering the extremely complex physical coupling relationships between the variables, a direct derivation was not provided. The analytical expression, but during simulation. , An initial value is assigned, and then the solution is calculated using the numerical difference method between adjacent time steps. and These represent the effects caused by velocity bias and angle of attack bias, respectively. The changes.

[0043] Step (2) System initialization and parameter setting In this invention, the typical flight conditions and model parameters of the hypersonic vehicle are first defined. The initial altitude of the vehicle is set to 21,000 m, the initial Mach number to 6, the mass to 33,000 kg, and the reference area to 60 m². 2 The reference length is taken as 7m, and the pitch moment of inertia is taken as 400,000 kg·m. 2 The simulation sampling time was set to 0.02 s.

[0044] After the program starts running, it first reads the nominal trajectory data, including information such as time series, reference angle of attack, and reference fuel equivalence ratio. Then, it establishes the controller parameter structure and sets parameters such as prediction step size, state weight matrix, control input weight matrix, control increment weight matrix, control input upper and lower limits, state upper and lower limits, and safety margin lower limit.

[0045] In this invention, the predicted state of the constraint prediction controller is selected as follows: (twenty four) The control input is selected as follows: (25) To ensure that the air intake is always within a safe operating range, this invention sets the lower limit of the safety margin as follows: (26) After system initialization, the aircraft's initial state is set. Complete state parameters include pitch rate, angle of attack, fuel equivalence ratio, safety margin, pitch angle, trajectory tilt angle, flight altitude, and flight speed. The initial angle of attack is given based on the reference angle of attack, the initial fuel equivalence ratio is given based on the nominal fuel equivalence ratio, and the initial safety margin is set within a safe range.

[0046] During the initial control input calculation, the program calls the trim function to calculate the initial elevator deflection angle, ensuring the aircraft achieves pitch moment balance near the initial operating conditions. The initial fuel equivalence ratio command is taken from the current fuel equivalence ratio, and the combustion chamber expansion ratio is taken from the nominal value. This yields the initial control input: (27) in, As the initial control input, The initial rudder deflection angle, The initial fuel equivalence ratio, This represents the initial combustion chamber expansion ratio.

[0047] Step (3) Online Local Linearization Implementation Process At each sampling moment, the program first reads the current state and reference instructions, and uses the current state and the control input of the previous moment as linearized operating points to establish a local prediction model under the current operating conditions.

[0048] In its implementation, the program encapsulates the complete nonlinear fly-thrust coupled model into a one-step propulsion function. During online local linearization, the predicted state is processed separately. and control input Apply a small perturbation and compare the changes in the result of one step forward before and after the perturbation. Calculate the state matrix and control matrix under the current operating condition using the central difference method to obtain the discrete prediction model: (28) in, , and The data is updated online at each sampling time based on the current flight status. This implementation does not require directly calling the complete nonlinear model for multi-step prediction in the optimizer. Instead, it approximates the nonlinear flight-thrust coupling model as a locally discrete linear model near the current operating point. This reduces the computational cost of online optimization while still reflecting the changes in flight-thrust coupling characteristics under the current flight status.

[0049] Step (4) Rolling optimization controller implementation process At each sampling time, the program establishes a rolling optimization problem based on the prediction model obtained from online local linearization. The optimization objectives include three parts: first, reducing the deviation between state variables such as angle of attack, fuel equivalence ratio, and safety margin and reference values; second, suppressing excessive control input; and third, limiting the change in control input between adjacent sampling times to ensure smooth control.

[0050] The constrained predictive controller considers the following constraints during the solution process: amplitude constraints of elevator deflection angle, fuel equivalence ratio command and combustion chamber expansion ratio; control input rate of change constraint; predictive state constraint; and intake safety margin constraint.

[0051] The intake safety margin constraint is as follows: (29) The intake safety margin constraint means that the safety margin for each prediction step in the future prediction time domain must not be lower than 0.1. If the prediction results show that the safety margin may approach the lower limit at some point in the future, the optimizer will adjust the control input in advance, rather than waiting for the safety margin to actually fall below the threshold before taking protective action.

[0052] The elevator limit is: (30) The fuel equivalence ratio directive limit is: (31) The combustion chamber expansion ratio limit is: (32) Meanwhile, to prevent abrupt changes in control input between adjacent sampling times, the program also sets a control increment limit, with the control increment constraint between adjacent sampling times being: (33) After solving the optimization problem, the program obtains the optimal control sequence for a future period of time: (34) In actual execution, only the first control variable in the sequence is applied to the flight-thrust coupling system, and then the next sampling time is entered to reacquire the flight state, relinearize, and resolve the optimization problem, thus forming a rolling time-domain closed-loop control.

[0053] When the safety margin is high, the constrained predictive controller primarily ensures angle-of-attack tracking performance, enabling the aircraft to track the reference command effectively. In this state, elevator deflection, fuel equivalence ratio command, and combustor expansion ratio all vary within normal ranges, resulting in relatively smooth control actions.

[0054] When the prediction results indicate a decreasing trend in the safety margin, the constraint predictive controller will proactively adjust the control inputs. Specifically, the elevator deflection angle is used to adjust the angle of attack and inlet flow conditions; the fuel equivalence ratio command is used to adjust the combustion chamber heat release intensity and back pressure level; and the combustion chamber expansion ratio is used to improve the pressure matching relationship between the isolator and the combustion chamber. Through the coordinated adjustment of these three control variables, the shock train can be suppressed from moving further upstream, keeping the intake safety margin above the lower safety limit.

[0055] Compared to traditional threshold-triggered protection methods, this invention does not wait until the safety margin has fallen below the threshold before taking control action. Instead, it detects the decreasing trend of the safety margin in advance within the prediction time domain and intervenes early by optimizing the control input. Therefore, this invention can achieve proactive protection against the risk of the intake duct not starting.

[0056] In the actual solution process, if the solution fails due to abnormal model linearization, excessively tight constraints, or infeasibility of the optimization problem, the program sets up an emergency protection strategy. This strategy will no longer use infeasible optimization results, but will instead generate conservative control variables based on the current safety margin and angle of attack error.

[0057] The emergency protection strategy mainly includes the following aspects: First, appropriately adjusting the elevator deflection angle to suppress further increases in the angle of attack; second, reducing the fuel equivalence ratio command to decrease combustion chamber heat release and the upward trend of back pressure. Through this treatment, even if the optimizer is temporarily unavailable, the system can still prioritize ensuring that the intake safety margin does not continue to deteriorate.

[0058] Step (5) Simulation verification To verify the effectiveness of the hypersonic flight thrust-protection control method considering intake constraints described in this invention, a typical hypersonic cruise condition was selected for simulation analysis. During the simulation, the controller used pitch rate, angle of attack, fuel equivalence ratio, and intake inactivity safety margin as predicted states, and elevator deflection angle, fuel equivalence ratio command, and combustion chamber expansion ratio as control inputs. At each sampling time, online local linearization and rolling optimization were performed based on the current state. Limits were set for all control quantities in the program, with the elevator deflection angle limited to ±30 degrees, the fuel equivalence ratio command limited to ([0.30, 1.00]), and the combustion chamber expansion ratio limited to ([1.00, 1.50]). Control increment constraints were also set to avoid abrupt changes in the control inputs.

[0059] Depend on Figure 2 It can be seen that the actual angle of attack of the aircraft can track the changes in the reference angle of attack quite well, indicating that the designed controller still has good attitude tracking capability after introducing the safety margin constraint of the air intake. Since the controller is simultaneously affected by the safety margin constraint, the control input amplitude constraint, and the control increment constraint, the actual angle of attack is not necessarily completely equal to the reference command, but the overall tracking trend remains consistent, indicating that the method can maintain basic flight control performance while meeting the safety protection requirements.

[0060] Depend on Figure 3 It can be seen that the safety margin for the intake duct not starting remained above the lower safety limit of 0.1 throughout the simulation process, without falling below the constraint boundary. The results show that by treating the safety margin as an explicit constraint in the predictive control optimization problem, the constrained predictive controller can identify the decreasing trend of the safety margin in advance in the prediction time domain, and suppress the intake duct state from continuing to evolve towards the inactivity boundary by coordinating and adjusting the elevator deflection angle, fuel equivalence ratio command, and combustion chamber expansion ratio, thereby achieving active protection against the risk of intake duct not starting.

[0061] Depend on Figures 4 to 6 As can be seen, the flight speed changes continuously and smoothly, without significant drops or divergences due to the protective control action; the elevator deflection angle remains within the physical limits without sustained saturation or severe oscillations; the fuel equivalence ratio changes smoothly within the constraints and can participate in propulsion adjustment according to the trend of safety margin changes. The comprehensive simulation results show that this invention can maintain the inlet safety margin within a safe range while ensuring angle-of-attack tracking performance, and ensure that the control input meets engineering constraints, verifying the effectiveness of the proposed integrated flight-thrust active protection control method.

Claims

1. A method for integrated flight and thrust protection and control of a hypersonic vehicle considering air intake constraints, characterized in that, The specific steps are as follows: Step 1: Construct a longitudinal dynamic model of the hypersonic vehicle integrating flight and thrust; By introducing the fuel equivalence ratio state, a hypersonic vehicle integrated flight-propulsion longitudinal dynamics model is established to describe the coupling relationship between the longitudinal motion state of the vehicle and the propulsion system state. Furthermore, its small disturbance linearization expression is obtained, providing a basic model for the state topology, local linearization and constraint predictive control of the air intake safety margin. Step 2: Define the safety margin for the intake manifold not to start; A dynamic adjustment coefficient is introduced based on the safety margin of the shock train length. Furthermore, by combining the characteristic Mach number of the intake duct, a safety margin for non-starting is constructed to quantitatively characterize the safety level of the intake duct from the non-starting state, providing evaluation indicators for safety margin prediction and active protection. Step 3: Establish a flight-thrust coupled augmented model that considers the safety margin of the inlet not starting and a reduced-order prediction model for roll optimization; Step 4: Design an integrated protection and control method for the flight propulsion system that takes into account air intake constraints; The design of the constraint predictive controller realizes integrated flight and thrust protection control that takes into account air intake constraints; in terms of flight performance, it ensures that the aircraft's angle of attack can track the given command, so that the attitude response meets the mission requirements; in terms of air intake safety, it ensures that the inert safety margin is always kept above the given safety lower limit, and avoids the air intake from entering the inert state.

2. The integrated flight and thrust protection and control method for hypersonic vehicles considering air intake constraints according to claim 1, characterized in that, Step 1 is as follows: The fuel equivalence ratio is incorporated into the longitudinal motion equation for dynamic modeling; the expression for the fuel equivalence ratio is as follows: (1) in, Fuel equivalence ratio, For speed, For the angle of attack, For fuel equivalence ratio command; This describes the physical relationship between fuel equivalence ratio and speed, angle of attack, and fuel equivalence ratio commands. Ignoring lateral parameters, the force analysis of the longitudinal channel of the hypersonic vehicle yields the following expanded longitudinal dynamic model of the hypersonic vehicle's integrated flight and thrust system: (2) Among them, the superscript " " represents the first derivative, It is the pitch angular velocity. It is the pitch moment of inertia. It is the pitching moment; It is the trajectory inclination angle; For the mass of the aircraft, and These represent thrust and lift, respectively. Represents gravitational acceleration; The pitch angle, Flight altitude; It is the pitching moment generated by the propulsion system; Applying a small perturbation to equation (2) linearly: (3) in, For flight speed disturbance, Let be the trajectory tilt angle disturbance. The pitch angular velocity disturbance is... This represents the fuel equivalence ratio disturbance. This represents the elevator deflection angle disturbance. For the angle of attack disturbance, This represents the fuel equivalence ratio command disturbance. Let be the partial derivative of thrust with respect to velocity. Let be the partial derivative of thrust with respect to angle of attack. This is the partial derivative of thrust with respect to fuel equivalence ratio; Let be the partial derivative of the axial force with respect to the velocity. Let be the partial derivative of the axial force with respect to the angle of attack. Let be the partial derivative of the axial force with respect to the fuel equivalence ratio. This is the partial derivative of the axial force with respect to the elevator deflection angle; Let be the partial derivative of the normal force with respect to the velocity. Let be the partial derivative of the normal force with respect to the angle of attack. Let be the partial derivative of the normal force with respect to the fuel equivalence ratio. This is the partial derivative of the normal force with respect to the rudder deflection angle; The disturbance force is caused by the change in the projection of gravity along the axial direction. The disturbance force caused by the change in the projection of gravity onto the normal direction; This is the partial derivative of the aerodynamic pitching moment with respect to velocity. Let be the partial derivative of the thrust pitching moment with respect to velocity. Let be the partial derivative of the aerodynamic pitching moment with respect to the angle of attack. Let be the partial derivative of the thrust pitching moment with respect to the angle of attack. This is the partial derivative of the aerodynamic pitching moment with respect to the pitching angular velocity. Let be the partial derivative of the aerodynamic pitching moment with respect to the elevator deflection angle. Let be the partial derivative of the thrust pitching moment with respect to the fuel equivalence ratio. This is the additional eccentric torque disturbance term caused by gravity due to the shift of the center of mass or the change of the point of application of force.

3. The integrated flight and thrust protection and control method for a hypersonic vehicle considering air intake constraints according to claim 2, characterized in that, To simplify the expression, the partial derivative terms are organized into dynamic coefficients; the dynamic coefficients are shown below: 、 、 、 、 、 、 、 、 、 、 、 、 、 、 、 、 、 、 、 、 、 、 、 、 、 、 、 Wherein, the first digit of the subscript is the equation number of formula (4); the second digit of the subscript is the sequence number of the small perturbation bias: Corresponding to serial number 1, Corresponding to serial number 2, Corresponding to serial number 3, Corresponding to serial number 4; The rate of change of fuel mass flow rate, The rate of change of air mass flow rate, For air mass flow rate, Fuel mass flow rate, For an appropriate oil-gas ratio, T This is the equivalent time constant.

4. The integrated flight and thrust protection and control method for hypersonic vehicles considering air intake constraints according to claim 1, characterized in that, Step 2 is as follows: Length of shock train The empirical formula is: (4) in, The radius of the isolation segment, To determine the Mach number at the front end of the pre-burning shock train, The Mach number at the exit of the isolation section. The specific heat ratio of the gas within the isolation section; The inlet non-starting safety margin is defined based on the length of the shock train. : (5) in, The length of the isolation segment. This is a dynamically adjusted coefficient.

5. The integrated flight and thrust protection and control method for a hypersonic vehicle considering air intake constraints according to claim 1, characterized in that, Step 3 is as follows: By incorporating the inlet-free safety margin into the longitudinal dynamics model of the integrated flight and thrust system obtained in step 1, a flight-thrust coupling augmented model considering the inlet-free safety margin is established: (6) in, The combustion chamber expansion ratio; function The relationship between the safety margin for the intake manifold not to start and speed, angle of attack, fuel equivalence ratio and combustion chamber expansion ratio; A reduced-order prediction model is used for rolling optimization; prediction state as follows: (7) Corresponding control input Selected as: (8) Using the current flight state and control input as the operating point, the flight-thrust coupled augmented model considering the safety margin of inlet inactivity is linearized with small perturbations. After selecting the reduced-order prediction state and control input, the following local linear prediction relationship is obtained: (9) in, These are the partial derivatives of the inert safety margin with respect to speed, angle of attack, fuel equivalence ratio, and combustion chamber expansion ratio, respectively. The disturbance is the rate of change of pitch angular velocity. The disturbance is the rate of change of angle of attack. This represents the disturbance of the rate of change in fuel equivalence ratio. The disturbance is the rate of change of the safety margin for the intake manifold not starting. This refers to the disturbance of the combustion chamber expansion ratio; After performing online local linearization and discretization near the current operating point, the reduced-order prediction model can be written as follows: (10) in, , These are the discrete state matrix and control matrix under the current operating condition, respectively. The affine bias term is generated by the local linearization of the nonlinear model. For the first The reduced-order predicted state vector at each sampling time, For the first The control input vector at each sampling time.

6. The integrated flight and thrust protection and control method for a hypersonic vehicle considering air intake constraints according to claim 1, characterized in that, Step 4 is as follows: The design goal of the constrained predictive controller is to minimize state tracking error and control input cost in the prediction time domain, while explicitly satisfying various constraints. Establish a rolling optimization problem with a constrained predictive controller: (11) in, For performance index functions, To predict the step size; The performance index function is defined as follows: (12) in, The state weight matrix is... To control the input weight matrix, To control the incremental weight matrix, Let be the prediction step number in the prediction time domain, and ; Indicates based on the first The prediction obtained from the known state and control information at the sampling time is the first... The state vector at each sampling time. For the first The reference state vector corresponding to each sampling time; Indicates the first The prediction obtained at the sampling time, which affects the future... The control input vector at each sampling time; The change in control input at adjacent sampling times, i.e.: (13) That is, for the first term in the prediction time domain The prediction step, the control increment is from the first prediction step. The control input at the predicted sampling time and the first The difference between the control inputs at each predicted sampling time is determined; During the optimization process, the predicted state must satisfy discrete dynamic constraints: (14) in, For the first The discrete state matrix is ​​obtained by online local linearization and discretization at each sampling time point based on the current operating point. For the first The discrete control matrix corresponding to each sampling time; The control input constraints are: (15) in, , These are the lower and upper bounds of the control input, respectively. Input constraint shrinkage amount; State constraints are represented as: (16) in, , These are the upper bound and the lower bound of the predicted state, respectively. This represents the state-constrained contraction amount. The inlet safety margin constraint in the prediction time domain is expressed as: (17) in, For the first The future flight state and control information at the sampling time are used to predict the next sampling time. The safety margin for the air intake not to start at each sampling time. This is the minimum allowable safety margin for the intake manifold not to start; At each sampling time, the constrained predictive controller solves for the future. The optimal control sequence for the step: (18) in, For the first The future obtained at each sampling time Step-optimal control input sequence, For the first The optimal control input vector is obtained at each sampling time and applied to the current sampling period. For the first The prediction obtained at the sampling time is the first The optimal control input vector at each sampling time; In actual execution, only the first control variable in the optimal control sequence is applied to the controlled object: (19) At the next sampling time, the constrained predictive controller re-performs online local linearization and discretization of the flight-thrust coupled augmented model established in step 3, obtaining the... The local discrete prediction model corresponding to each sampling time is used, and the optimization problem is solved again, thus forming a rolling time-domain closed-loop control.

Images