Orbital capability assessment and trajectory reconstruction method based on local collocation method and neural network

Abstract

The present invention discloses a method for orbital capability assessment and trajectory reconstruction based on a local collocation method and a neural network. The method comprises the following steps: (1) using the local collocation method to optimize the trajectory and obtain the optimal orbital trajectory for a launch vehicle under fault-free conditions; (2) logically classifying power failures, using a neural network model to train and learn a large database obtained through offline simulation, and establishing a mapping relationship between power failures and mission capabilities; and (3) for any predetermined power failure mode, evaluating the current mission capability of the rocket based on the mapping relationship between the failure mode and the rocket's mission capability established above. The present invention implements orbital capability assessment and online trajectory reconstruction during thrust reduction of the launch vehicle, thereby improving the success rate of launch missions.

Description

Technical Field

[0001] The present invention belongs to the field of flight dynamics, and in particular relates to an orbital capability assessment and trajectory reconstruction method based on a local point matching method and a neural network. Background Art

[0002] Launch vehicles are fundamental to the development of space technology, the exploitation of space resources, and ensuring space safety. They are a key component of the space transportation system and a driving force behind the development of the space industry. A launch vehicle's control system encompasses navigation, guidance, attitude control, and system integration, serving as the vehicle's brain and nerve center. The technical level of the control system directly determines the success of a launch mission and plays a crucial role in maximizing launch capacity. Existing guidance and control systems rely on system redundancy and fault tolerance, as well as the amplitude and phase margins of the attitude control system, to accommodate minor deviations caused by failures within the rocket. These control methods are mature and reliable, meeting current mission requirements. While they offer a certain degree of stability within this small deviation range, ensuring successful payload orbital placement, they are less adaptable to larger, non-fatal failures (such as booster failures) during flight. Therefore, research is needed on new intelligent rocket control systems with increased reliability and resilience to failures.

[0003] Propulsion system failures are a common type of rocket failure. Existing research has explored guidance technologies for non-catastrophic propulsion system failures. Currently, my country's launch vehicles lack similar capabilities. Therefore, it is necessary to conduct research on ascent trajectory reconstruction techniques under propulsion failures to enhance the rocket's autonomous flight capabilities. Neural networks have excellent nonlinear fitting capabilities and can theoretically approximate arbitrary nonlinear functions. Once trained, the networks can be executed online, making them widely used for online identification of unknown functions or systems. Summary of the Invention

[0004] Purpose of the invention: The purpose of the present invention is to propose an orbital capability assessment and trajectory reconstruction method based on the local point matching method and neural network, so that the launch vehicle can quickly perform self-assessment after a power failure, determine whether it has mission capability, and re-plan a new trajectory while satisfying various constraints.

[0005] Technical Solution: The present invention provides a method for orbital capability assessment and trajectory reconstruction based on a local collocation method and a neural network, comprising the following steps:

[0006] (1) Establishing a launch vehicle ascent trajectory optimization problem, solving the trajectory optimization problem, and obtaining a reference trajectory;

[0007] (2) Conduct offline simulation analysis on launch vehicles with different faults, optimize trajectories using the local collocation method, and establish an offline trajectory database for launch vehicles entering target orbits and secondary orbits;

[0008] (3) A neural network is used to learn the trajectory database and establish a mapping relationship between thrust descent failure and orbital insertion capability. For any predetermined power failure mode, the current mission capability of the rocket is evaluated based on the mapping relationship. If the rocket still has mission capability, the target mission is readjusted within the mission capability range according to the preset preferences, and the ascent trajectory is replanned using an online planning method.

[0009] Furthermore, the implementation process of step (1) is as follows:

[0010] The objective function of the launch vehicle ascent trajectory optimization problem is:

[0011] min J=-m f

[0012] Among them, J is the objective function of trajectory optimization; m f For the launch vehicle at t f The quality of the moment; f The moment when the launch vehicle core stage is shut down;

[0013] The state equation of the launch vehicle is as follows:

[0014]

[0015] Where r is the distance between the center of mass of the rocket and the center of the earth (m); v is the speed of the spacecraft relative to the earth (m / s); m is the mass of the rocket (kg); θ is the flight range angle (rad); γ is the track angle (rad); α is the angle of attack (rad); I ep is the fuel specific impulse of the rocket engine (s); g0 is the acceleration of gravity at sea level (m / s 2 ); μ is the Earth's gravitational constant; G = -mμ / r 2 is the gravitational force on the launch vehicle; P is the total thrust from the engine on the launch vehicle during flight:

[0016]

[0017] Among them, P E is the thrust of the booster stage of the launch vehicle, P C is the thrust of the core stage rocket of the carrier rocket, t e The moment when the booster stage of the launch vehicle separates;

[0018] D and L are drag and lift respectively:

[0019]

[0020] Among them, C L and C D are the drag coefficient and lift coefficient, S ref is the reference area of ​​the rocket (m 2 ); ρ is the atmospheric density (kg / m 3 ), ρ0 is the atmospheric density at sea level, h is the flight altitude of the launch vehicle, and h0 is a constant;

[0021] The launch vehicle must meet initial constraints, terminal constraints, range constraints, and path constraints during ascent; the initial constraints are as follows:

[0022] r(t0)=r0, θ(t0)=θ0, v(t0)=v0, γ(t0)=γ0, m(t0)=m0, α(t0)=α0

[0023] Where t0 is the initial time; r is the flight altitude; m0 is the initial mass of the launch vehicle, m0 = 4m P +m C +m f +m l , where m P is the mass of the booster stage of the launch vehicle, m C is the mass of the core stage engine of the launch vehicle, m f is the mass of the launch vehicle fairing, m l is the mass of the launch vehicle payload; θ0, v0, γ0 and α0 are the initial values ​​of the corresponding state variables;

[0024] The launch vehicle flight should meet the requirements of maneuverability and structural bearing capacity. The control quantity and state quantity of the launch vehicle during flight should change within a certain range to ensure the normal flight of the launch vehicle. Therefore, the launch vehicle should meet the following range constraints:

[0025] α min ≤α≤α max ,γβ min ≤γ≤γ max , r min ≤r≤r max , v min ≤v≤v max ,

[0026] Among them, α min , α max , γ min , γ max , r min , r max , v min , v max are the upper and lower bounds of the corresponding control variables and state variables, is the rate of change of the angle of attack α with time, ω min With ω max for The upper and lower bounds of

[0027] The launch vehicle maintains vertical flight until it leaves the launch tower, and the following path constraints are imposed on the angle of attack:

[0028] α=0,t0≤t≤t1

[0029] Among them, α is the angle of attack, t0 is the time when the carrier rocket starts flying, and t1 is the time when the carrier rocket ends vertical flight;

[0030] The terminal constraints of the launch vehicle are:

[0031] r(t f )=r f ,v(t f )=v f ,γ(t f )=γ f

[0032] Among them, r f , v f , γ f It is the orbital altitude, speed and track angle when the rocket enters orbit.

[0033] Furthermore, the different launch vehicle failure modes described in step (2) are two modes: a proportional decrease in engine thrust and a linear continuous decrease in launch vehicle engine thrust, wherein:

[0034] a) The thrust reduction of the launch vehicle engine is proportional to the original thrust of the engine. The corresponding mathematical model is:

[0035]

[0036] Among them, P xt is the thrust of the core engine of the launch vehicle; P zt is the sum of the thrusts of the four booster stages; k x_k , k z_k They represent the ratio coefficients of the thrust output of the core stage engine and the booster stage engine to the rated thrust after a thrust ratio reduction failure occurs; P xt_k , P zt_k They are the thrust values ​​output by the core stage engine and the booster stage engine after a thrust drop failure occurs;

[0037] b) The thrust of the launch vehicle engine continues to decrease linearly to a certain thrust value. The corresponding mathematical model is:

[0038]

[0039] Among them, k x_l , k z_l They represent the ratio coefficients of the thrust output of the core stage engine and the booster stage engine to the rated thrust after a thrust ratio reduction failure occurs; P xt_l , P zt_l are the thrust values ​​output by the core stage engine and the booster stage engine after a thrust drop failure occurs, t m is the time of failure occurrence, t x_f , t z_f They represent the flameout time of the core stage engine and the booster stage engine respectively.

[0040] Furthermore, the database establishment process in step (2) is as follows:

[0041] According to the mission requirements of the launch vehicle, the secondary orbit is set before launch. The orbit parameters and number are set according to the mission requirements; the target orbit height r0 and the secondary orbit heights are r1, r2, r3…r n Establish an offline database; assume that the launch vehicle has a thrust ratio reduction failure at time t1, and the failure ratio coefficient is k x_k , then the thrust of the carrier rocket is P xt_k =k x_k P0, and the current state of the launch vehicle can be obtained by interpolating the nominal trajectory as x1; the local collocation method is used, x1 is used as the initial value constraint of trajectory optimization, the path constraint and the objective function remain unchanged, and the secondary orbit r1 is used as the terminal constraint for trajectory optimization; with the adjustment of the fault proportion coefficient k x_k The size of k can be found x_kmin Make k x_k >k x_kmin , then the carrier rocket can enter orbit; obtain a set of data pairs (t1,k x_kmin ,r1); then, the control variable method is used to change the fault occurrence time and secondary orbit one by one, and the above process is repeated to obtain multiple sets of data pairs of fault occurrence time, fault tolerance limit and secondary orbit height to form a trajectory database.

[0042] Furthermore, the implementation process of step (3) is as follows:

[0043] When training the neural network, the input layer selects the fault mode, fault occurrence time t and secondary track height, and the output layer fault tolerance limit k min ; Learn to obtain the launch vehicle failure time t and the secondary orbit height to the fault tolerance limit k min The mapping relationship;

[0044] According to the thrust drop failure mode, failure occurrence time and target orbit altitude of the launch vehicle, the corresponding fault tolerance limit k is obtained based on the established mapping relationship between the failure mode and the rocket's orbital insertion capability. min ; The failure ratio k and k when the launch vehicle fails min For comparison, if k>k min Then the carrier rocket can enter orbit, otherwise it cannot; the target orbit and secondary orbit are evaluated. If the rocket has the ability to enter orbit, then within the scope of the orbital insertion capability, the target mission is readjusted, and the ascent trajectory is replanned using the ascent trajectory online planning method to guide the rocket to fly autonomously.

[0045] Beneficial effects: Compared with the existing technology, the beneficial effects of the present invention are as follows: in order to address the problem that the increase in the number of launch vehicle engines leads to reduced reliability and the susceptibility to power failures, the launch vehicle power failures are logically classified, and a neural network is used to establish a mapping relationship between power failures and mission capabilities to evaluate the current mission capabilities of the launch vehicle; the present invention realizes the orbital insertion capability assessment and online trajectory reconstruction when the launch vehicle thrust decreases, thereby improving the reliability of the launch mission. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 is a flow chart of the present invention;

[0047] Figure 2 is a cross-sectional view of the flight trajectory of the carrier rocket of the present invention;

[0048] Figure 3 It is a flow chart of the local point matching method of the present invention;

[0049] Figure 4 It is the curve diagram of the nominal trajectory height change of the launch vehicle;

[0050] Figure 5 It is the curve diagram of the attack angle change of the nominal trajectory of the launch vehicle;

[0051] Figure 6 It is the velocity variation curve of the nominal trajectory of the launch vehicle;

[0052] Figure 7 It is the curve diagram of the range angle change of the nominal trajectory of the launch vehicle;

[0053] Figure 8 is the curve of the mass variation of the launch vehicle's nominal trajectory;

[0054] Figure 9 1. It is a schematic diagram of the thrust reduction of the carrier rocket according to the present invention;

[0055] Figure 10 is a fault tolerance limit diagram for the booster stage thrust descent launch vehicle according to the present invention;

[0056] Figure 11 This is the core stage engine thrust drop launch vehicle limit diagram of the present invention;

[0057] Figure 12 This is a diagram of the thrust drop failure tolerance limit of the booster stage when the carrier rocket of the present invention enters the secondary orbit;

[0058] Figure 13 The diagram of allowable limits for a core stage engine thrust ratio reduction failure into different secondary orbits according to the present invention;

[0059] Figure 14 The diagram of allowable limits for a core stage engine thrust linear descent failure into different secondary orbits according to the present invention;

[0060] Figure 15 This is a reconstruction diagram of the linear descent trajectory of the core-stage engine thrust according to the present invention;

[0061] Figure 16 This is a reconstruction diagram of the thrust ratio reduction trajectory of the core stage engine described in the present invention. DETAILED DESCRIPTION

[0062] The present invention will be further described in detail below with reference to the accompanying drawings.

[0063] This paper proposes a method for orbital capability assessment and trajectory reconstruction based on local point matching method and neural network. Figure 1 As shown, the following steps are included:

[0064] Step 1: Establish a launch vehicle ascent trajectory optimization problem, solve the trajectory optimization problem, and obtain a reference trajectory.

[0065] Based on the speed, altitude, heading angle, track angle and other state quantities of the carrier rocket at the initial moment of the ascent flight and the target speed, altitude, heading angle, track angle and other state quantity constraints at the orbital terminal, a reference trajectory is optimized while satisfying the flight constraints and control quantity constraints such as angle of attack, speed and altitude.

[0066] The carrier rocket is a 1.5-stage rocket, and its power system consists of a core stage rocket engine and four booster stage rocket engines. When performing a launch mission, the booster stage and the core stage work together to enable the carrier rocket to take off vertically and leave the launch tower. Then the booster stage and the fairing separate in sequence, and finally the core stage rocket puts the payload into orbit. The entire flight trajectory profile is as follows Figure 2 shown.

[0067] When optimizing the trajectory of a launch vehicle, the rotation of the Earth is ignored and only the motion of the center of mass of the launch vehicle is considered. The state equation of the center of mass of the launch vehicle in the ballistic plane is as follows:

[0068]

[0069] Where: r is the distance between the center of mass of the rocket and the center of the earth (m); v is the speed of the spacecraft relative to the earth (m / s); m is the mass of the rocket (kg); θ is the flight range angle (rad); γ is the track angle (rad); α is the angle of attack (rad); I ep is the fuel specific impulse of the rocket engine (s); g0 is the acceleration of gravity at sea level (m / s 2 ); μ is the Earth's gravitational constant; G = -mμ / r 2 is the gravitational force on the launch vehicle; P is the total thrust from the engine on the launch vehicle during flight, expressed as follows:

[0070]

[0071] Where: P E is the thrust of the booster stage of the launch vehicle, P C is the thrust of the core stage rocket of the carrier rocket, t e is the separation moment of the booster stage of the carrier rocket, t f The moment when the core stage of the carrier rocket is shut down.

[0072] D and L are the drag and lift forces, respectively, and are given by:

[0073]

[0074] Where: C L and C D are the drag coefficient and lift coefficient, S ref is the reference area of ​​the rocket (m 2 ); ρ is the atmospheric density (kg / m 3 ), ρ0 is the atmospheric density at sea level, h is the flight altitude of the launch vehicle, and h0 is a constant.

[0075] The launch vehicle must meet initial constraints, terminal constraints, range constraints, and path constraints during ascent. The initial constraints are described as follows:

[0076] The initial constraints of the launch vehicle ascent trajectory optimization problem are as follows:

[0077] r(t0)=r0, θ(t0)=θ0, v(t0)=v0, γ(t0)=γ0, m(t0)=m0, α(t0)=α0

[0078] Where: t0 is the initial time; r is the flight altitude; m0 is the initial mass of the launch vehicle, m0 = 4m P +m C +m f +m l , where m Pis the mass of the booster stage of the launch vehicle, m C is the mass of the core stage engine of the launch vehicle, m f is the mass of the launch vehicle fairing, m l is the mass of the launch vehicle payload; θ0, v0, γ0 and α0 are the initial values ​​of the corresponding state variables.

[0079] The objective function of the launch vehicle ascent trajectory optimization problem is as follows:

[0080] min J=-m f

[0081] Where: J is the performance index of trajectory optimization; m f For the launch vehicle at t f The quality of the moment; f The moment when the core stage of the carrier rocket is shut down.

[0082] The path constraints of the launch vehicle during flight are as follows:

[0083] α min ≤α≤α max , γ min ≤γ≤γ max , r min ≤r≤r max , v min ≤v≤v max ,

[0084] Where: α min , α max , γ min , γ max , r min , r max , v min , v max are the upper and lower bounds of the corresponding control variables and state variables, is the rate of change of the angle of attack α with time, ω min With ω max for The upper and lower bounds of .

[0085] The launch vehicle maintains vertical flight before leaving the launch tower, and the constraint on the angle of attack is:

[0086] α=0 t0≤t≤t1

[0087] Where: α is the angle of attack, t0 is the time when the carrier rocket starts flying, and t1 is the time when the carrier rocket ends vertical flight.

[0088] The terminal constraint during the launch vehicle's ascent is:

[0089] r(t f)=r f ,v(t f )=v f ,γ(t f )=γ f

[0090] Where: t f is the moment when the rocket enters orbit, r f , v f , γ f It is the orbital altitude, speed and track angle when the rocket enters orbit.

[0091] The local collocation method discretizes the optimal control problem into an NLP problem and solves it directly. It also ensures the calculation accuracy and efficiency through sparse difference and multi-resolution technology, thus meeting the real-time requirements of the launch vehicle online trajectory reconstruction. The general trajectory optimization problem can be described as: determine the control variables Minimize the following objective function:

[0092]

[0093] Where: x is the state variable, u is the control variable, t is the time, M is the Mayer term of the objective function, L is the Lagrange term of the objective function, t0 is the initial time, t f is the terminal time. State variable Initial time t0 and terminal time t f Satisfies the state equation:

[0094]

[0095] Where: f is the state equation function, is an m-dimensional real number space.

[0096] Boundary conditions:

[0097] Φ(x(t0),t0,x(t f ),t f )=0

[0098] Where: Φ is the boundary condition function.

[0099] Path Constraints:

[0100] C(x(t),u(t),t)≤0,t∈[t0,t f ]

[0101] Where: C is the path constraint function. In the formula, M, L, f, Φ and C are defined as follows:

[0102] M:

[0103] L:

[0104] f:

[0105] Φ:

[0106] C:

[0107] in: is the real number space, is an n-dimensional real number space, is a φ-dimensional real number space.

[0108] Assume that the N+1 discrete points on the unit interval [0,1] are:

[0109] G={τ i ∈[0,1],i=0,1,…,N; τ0=0,τ N =τ f =1; τ i <τ i+1 ,i=0,1,…,N-1}

[0110] Where: G is a set of discrete points, τ i is called a node or grid point, τ i It can be uniformly distributed or non-uniformly distributed on [0,1].

[0111] Brief Notes x i =x(τ i ),u i =u(τ i ),definition:

[0112] X={x0,x1,…,x N}; U={u0,u1,…,u N}

[0113]

[0114] Among them: X is the discrete state variable, U is the discrete control variable, is the set of discrete points that do not belong to G, q is the order of the discrete format used, j is an integer, For The set of state variables and control variables corresponding to the node.

[0115] The NLP obtained by discretizing the above continuous Bolza problem can be described as determining the discrete state variables X, discrete control variables U and , initial time t0, terminal time t f , so that the following objective function is minimized:

[0116]

[0117] Where: Δt is the time difference, h i is the length of the time interval, β j is the integral coefficient, L ij Is the interpolation polynomial. And satisfies the constraints:

[0118]

[0119] Where: i is the discrete residual of the state equation, is the node τ ij The corresponding path function.

[0120] The Hermite-Simpson (HS) format requires variables and function values ​​at the midpoint of the interval. To do this, the control variable at the midpoint of the interval needs to be used as the optimization variable, and a path constraint needs to be added at the midpoint of the interval. The midpoint of the interval is:

[0121]

[0122] The optimized variables of NLP obtained in HS format are (x0,x1,···,x N ;u0,u1,···,u N ; t0,t f ), and the generalized objective function is discretized as:

[0123]

[0124] The constraints are discretized as:

[0125]

[0126] C i =C(x i ,u i ,τ i ;t0,t f )≤0

[0127]

[0128] E(x0,t0,x f ,t f )=0

[0129] in,

[0130]

[0131] in: and for Corresponding state variables, state equations, and interpolation polynomials.

[0132] In numerical optimization, in order to make the problem have actual physical meaning, it is also necessary to add time difference constraints:

[0133] Δt=t f -t0>0

[0134] Therefore, the optimal control problem is transformed into a large-scale sparse NLP problem, which can be solved by the SQP algorithm. Mature solvers include SNOPT and IPOPT. The algorithm flow chart is as follows Figure 3 shown.

[0135] The simulation parameters for the launch vehicle trajectory optimization are: the total mass of the launch vehicle booster stage engine is m P =323028kg, the core engine mass is m x =88500kg, the mass of the fairing is m f =2700kg, load mass is m l =10000kg; the total thrust of the booster stage engine and the core stage engine of the carrier rocket are P zt =5400kN, P xt =650kN; lift coefficient C L =0.01, resistance coefficient C D =0.1, reference area S ref =12.57m 2 ; Booster stage separation time t2 = 173s, fairing separation time t3 = 190s; initial altitude h0 = 0, initial velocity v0 = 0, initial range angle θ0 = 0, initial track angle γ0 = π / 2, initial mass m0 = 424228kg; terminal altitude r f = 6900km (i.e. target orbit altitude), terminal velocity v f =7700km / s, terminal track angle γ f =0. Figure 4-Figure 8 This is the nominal trajectory diagram obtained by simulation using the local collocation method.

[0136] Step 2: Conduct offline simulation analysis on different faulty launch vehicles, use the local collocation method to optimize the trajectory, and establish an offline trajectory database for the launch vehicle entering the target orbit and secondary orbit.

[0137] Based on the reference trajectory generated in step 1, the state of the launch vehicle at each time point is obtained, and the secondary orbits r1, r2, r3…r are designed according to the mission requirements or rescue capabilities. n . Select appropriate time points t1, t2, t3…t m,Assuming that the launch vehicle has thrust reduction failures of different modes at these moments, ,the target orbit and secondary orbit are used as the orbital entry targets, ,and the trajectory optimization is performed using the local collocation method to ,judge the orbital entry capability of the launch vehicle and establish an ,offline database.

[0138] When a carrier rocket has different power failures, the thrust of the carrier rocket engine decreases. There are two descent modes. The first is that the thrust of the carrier rocket engine decreases to a proportional value of the original thrust of the engine. The second is that the thrust of the carrier rocket engine continues to decrease linearly to a certain thrust value.

[0139] Schematic diagrams of two thrust failures are shown as follows Figure 9 As shown in the figure, the two power failure models are described as follows:

[0140] (1) The thrust reduction of the launch vehicle engine is proportional to the original thrust of the engine. The corresponding mathematical model is:

[0141]

[0142] Where: P xt is the thrust of the core engine of the launch vehicle; P zt is the sum of the thrusts of the four booster stages; k x_k , k z_k They represent the ratio coefficients of the thrust output of the core stage engine and the booster stage engine to the rated thrust after a thrust ratio reduction failure occurs; P xt_k , P zt_k They are the thrust values ​​output after the core stage engine and booster stage engine have thrust drop failures.

[0143] (2) The thrust of the launch vehicle engine continues to decrease linearly to a certain thrust value.

[0144]

[0145] Among them, k x_l , k z_l They represent the ratio coefficients of the thrust output of the core stage engine and the booster stage engine to the rated thrust after a thrust ratio reduction failure occurs; P xt_l , P zt_l are the thrust values ​​output by the core stage engine and the booster stage engine after a thrust drop failure occurs, t m is the time of failure occurrence, t x_f , t z_f They represent the flameout time of the core stage engine and the booster stage engine respectively.

[0146] The database of the target orbit and secondary orbits of the launch vehicle is generated offline for different propulsion failures. The secondary orbit is set before launch according to the mission requirements of the launch vehicle. The offline trajectory database is established based on the thrust drop failure mode of the launch vehicle and the current flight status parameters. The specific process is as follows:

[0147] According to the mission requirements of the launch vehicle, the secondary orbit is set before launch. The orbit parameters and number are set according to the mission requirements. The target orbit height r0 and the secondary orbit heights are r1, r2, r3…r n Establish an offline database. Assume that the launch vehicle has a thrust proportional reduction failure at time t1, and the failure proportional coefficient is k x_k , then the thrust of the carrier rocket is P xt_k =k x_k P0, and the current state of the launch vehicle can be obtained by interpolating the nominal trajectory as x1. Using the aforementioned local collocation method, x1 is used as the initial value constraint for trajectory optimization. The path constraint and the objective function remain unchanged, and the trajectory optimization is performed with the secondary orbit r1 as the terminal constraint. With the adjustment of the fault ratio coefficient k x_k The size of k can be found x_kmin Make k x_k >k x_kmin , then the launch vehicle can enter orbit. At this time, we get a set of data pairs (t1,k x_kmin ,r1). Then, the control variable method is used to change the fault occurrence time and secondary track one by one, and the above process is repeated to obtain multiple sets of data pairs of fault occurrence time, fault tolerance limit and secondary track height, thus forming a trajectory database.

[0148] Figure 10 and Figure 11 It is the allowable failure limit of the booster stage thrust descent launch vehicle and the allowable failure limit of the core stage engine thrust descent launch vehicle. Figure 12 The booster thrust of the launch vehicle descends into R f =6878km secondary track fault tolerance limit diagram. Figure 13 The thrust ratio of the core stage engine is reduced as the launch vehicle enters (1) (2) (3) Secondary track fault tolerance limit, Figure 14 Core stage engine thrust linearly decreases launch vehicle entry (1) (2) (3) Secondary track fault tolerance limit.

[0149] Step 3: Use a neural network to learn the trajectory database and establish a mapping relationship between thrust descent failure and orbital insertion capability; for any predetermined propulsion failure mode, evaluate the rocket's current mission capability based on the mapping relationship; if the rocket is still mission-capable, then within the mission capability range, readjust the target mission based on the preset preferences and use the online planning method to replan the ascent trajectory. The specific plan is as follows:

[0150] (1) Mapping relationship between thrust failure and rocket orbital capability

[0151] The study analyzes the laws of the rocket's orbital ability under different thrust failures based on the large database generated above. The database is learned using a neural network. When training the neural network, the input layer selects the fault mode, fault occurrence time t and secondary orbit height, and the output layer fault tolerance limit k min . Learn to get the launch vehicle failure time t and the secondary orbit height to the fault tolerance limit k min The mapping relationship.

[0152] (2) Rocket Orbital Capability Assessment and Decision-Making under Thrust Failure

[0153] According to the mapping relationship between the thrust-down failure mode, failure occurrence time and target orbital altitude of the launch vehicle, the corresponding fault tolerance limit k can be obtained based on the aforementioned established relationship between the failure mode and the rocket's orbital insertion capability. kmin Compare the failure ratios k and k when the launch vehicle fails. kmin , if k>k kmin The launch vehicle can enter orbit; otherwise, it cannot. This method is used to evaluate the target orbit and secondary orbit. If the rocket has orbital insertion capability, the target mission is readjusted within the orbital insertion capability range, and the ascent trajectory is replanned using the aforementioned online ascent trajectory planning method to guide the rocket into autonomous flight.

[0154] Table 1 Core stage engine thrust reduction simulation parameters

[0155]

[0156] Table 1 is Figure 15 and Figure 16 Core stage engine thrust reduction simulation parameter table. Figure 15 and Figure 16 When the core stage engine of the launch vehicle experiences power failure at 50s, 150s, 250s and 350s respectively, the launch vehicle performs self-assessment and selects secondary orbits r1, r2, r3 and original orbit r1 respectively. f =6900km for orbit insertion and online planning of a new ascent trajectory.

Claims

1. A method for orbital capability assessment and trajectory reconstruction based on local collocation method and neural network, characterized in that: The following steps are involved: (1) Establishing a launch vehicle ascent trajectory optimization problem, solving the trajectory optimization problem, and obtaining a reference trajectory; (2) Conduct offline simulation analysis on launch vehicles with different power failures, optimize trajectories using the local collocation method, and establish an offline trajectory database for launch vehicles entering target orbits and secondary orbits; (3) Using a neural network to learn the trajectory database, a mapping relationship between power failure and orbital capability is established; for any predetermined power failure mode, the current mission capability of the rocket is evaluated based on the mapping relationship; if the rocket still has mission capability, the target mission is readjusted within the mission capability range according to the preset preferences, and the ascent trajectory is replanned using an online planning method; The database establishment process in step (2) is as follows: According to the mission requirements of the launch vehicle, the secondary orbit is set before launch. The orbit parameters and number are set according to the mission requirements; the target orbit height r0 and the secondary orbit heights are r1, r2, r3…r n Establish an offline database; assume that the launch vehicle has a thrust ratio reduction failure at time t1, and the failure ratio coefficient is k xk , then the thrust of the carrier rocket is P xt_k =k x_k P0, and the current state of the launch vehicle can be obtained by interpolating the nominal trajectory as x1; the local collocation method is used, x1 is used as the initial value constraint of trajectory optimization, the path constraint and the objective function remain unchanged, and the secondary orbit r1 is used as the terminal constraint for trajectory optimization; with the adjustment of the fault proportion coefficient k x_k The size of k can be found x_kmin Make k x_k >k x_kmin , then the carrier rocket can enter orbit; obtain a set of data pairs (t1,k x_kmin ,r1); then, the control variable method is used to change the fault occurrence time and secondary orbit one by one, and the above process is repeated to obtain multiple sets of data pairs of fault occurrence time, fault tolerance limit and secondary orbit height to form a trajectory database.

2. The orbital capability assessment and trajectory reconstruction method based on the local collocation method and neural network according to claim 1 is characterized in that: The implementation process of step (1) is as follows: The objective function of the launch vehicle ascent trajectory optimization problem is: min J=-m f Among them, J is the objective function of trajectory optimization; m f For the launch vehicle at t f The quality of the moment; f The moment when the launch vehicle core stage is shut down; The state equation of the launch vehicle is as follows: Where r is the distance between the center of mass of the rocket and the center of the earth, in meters; v is the speed of the vehicle relative to the earth, in meters per second; m is the mass of the rocket, in kilograms; θ is the flight range angle, in rad; γ is the track angle, in rad; α is the angle of attack, in rad; I ep is the fuel specific impulse of the rocket engine, in s; g0 is the acceleration of gravity at sea level, in m / s 2 ; μ is the Earth's gravitational constant; G = -mμ / r 2 is the gravitational force on the launch vehicle; P is the total thrust from the engine on the launch vehicle during flight: Among them, P E is the thrust of the booster stage of the launch vehicle, P C is the thrust of the core stage rocket of the carrier rocket, t e The moment when the booster stage of the launch vehicle separates; D and L are drag and lift respectively: Among them, C L and C D are the drag coefficient and lift coefficient, S ref is the reference area of ​​the rocket, in m 2 ; ρ is the atmospheric density, unit is kg / m 3 ;ρ0 is the atmospheric density at sea level, h is the flight altitude of the launch vehicle, and h0 is a constant; The launch vehicle must meet initial constraints, terminal constraints, range constraints, and path constraints during ascent; the initial constraints are as follows: r(t0)=r0, θ(t0)=θ0, v(t0)=v0, γ(t0)=γ0, m(t0)=m0, α(t0)=α0 Where t0 is the initial time; r is the flight altitude; m0 is the initial mass of the launch vehicle, m0 = 4m P +m C +m f +m l , where m P is the mass of the booster stage of the launch vehicle, m C is the mass of the core stage engine of the launch vehicle, m f is the mass of the launch vehicle fairing, m l is the mass of the launch vehicle payload; θ0, v0, γ0 and α0 are the initial values ​​of the corresponding state variables; The launch vehicle flight should meet the requirements of maneuverability and structural bearing capacity. The control quantity and state quantity of the launch vehicle during flight should change within a certain range to ensure the normal flight of the launch vehicle. Therefore, the launch vehicle should meet the following range constraints: a min ≤α≤α max ,c min ≤γ≤γ max ,r min ≤r≤r max ,v min ≤v≤v max , Among them, α min , α max , γ min , γ max , r min , r max , v min , v max are the upper and lower bounds of the corresponding control variables and state variables, is the rate of change of the angle of attack α with time, ω min With ω max for The upper and lower bounds of The launch vehicle maintains vertical flight until it leaves the launch tower, and the following path constraints are imposed on the angle of attack: α=0,t0≤t≤t1 Among them, α is the angle of attack, t0 is the time when the carrier rocket starts flying, and t1 is the time when the carrier rocket ends vertical flight; The terminal constraints of the launch vehicle are: r(t f )=r f ,v(t f )=v f ,γ(t f )=γ f Among them, r f , v f , γ f It is the orbital altitude, speed and track angle when the rocket enters orbit.

3. The orbital capability assessment and trajectory reconstruction method based on the local collocation method and neural network according to claim 1 is characterized in that: The different power failure carrier rockets described in step (2) include two types: a proportional decrease in engine thrust and a linear continuous decrease in carrier rocket engine thrust, wherein: a) The thrust reduction of the launch vehicle engine is proportional to the original thrust of the engine. The corresponding mathematical model is: Among them, P xt is the thrust of the core engine of the launch vehicle; P zt is the sum of the thrusts of the four booster stages; k x_k , k z_k They represent the ratio coefficients of the thrust output of the core stage engine and the booster stage engine to the rated thrust after a thrust ratio reduction failure occurs; P xt_k , P zt_k They are the thrust values ​​output by the core stage engine and the booster stage engine after a thrust drop failure occurs; b) The thrust of the launch vehicle engine continues to decrease linearly to a certain thrust value. The corresponding mathematical model is: Among them, k x_l , k z_l They represent the ratio coefficients of the thrust output of the core stage engine and the booster stage engine to the rated thrust after a thrust ratio reduction failure occurs; P xt_l , P zt_l are the thrust values ​​output by the core stage engine and the booster stage engine after a thrust drop failure occurs, t m is the time of failure occurrence, t x_f , t z_f They represent the flameout time of the core stage engine and the booster stage engine respectively.

4. The orbital capability assessment and trajectory reconstruction method based on the local collocation method and neural network according to claim 1 is characterized in that: The implementation process of step (3) is as follows: When training the neural network, the input layer selects the fault mode, fault occurrence time t and secondary track height, and the output layer fault tolerance limit k min ; Learn the launch vehicle failure time t and the secondary orbit height to the failure tolerance limit k min The mapping relationship; According to the thrust drop failure mode, failure occurrence time and target orbit altitude of the launch vehicle, the corresponding fault tolerance limit k is obtained based on the established mapping relationship between the failure mode and the rocket's orbital insertion capability. min ; The failure ratio k and k when the launch vehicle fails min For comparison, if k>k min Then the carrier rocket can enter orbit, otherwise it cannot; the target orbit and secondary orbit are evaluated. If the rocket has the ability to enter orbit, then within the scope of the orbital insertion capability, the target mission is readjusted, and the ascent trajectory is replanned using the ascent trajectory online planning method to guide the rocket to fly autonomously.

Images