A method for guiding and verifying a rocket return stage
Patent Information
- Application Number
- CN202510317974.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-03-17
- Publication Date
- 2026-09-25
AI Technical Summary
[0002]垂直起降火箭在再入大气后的飞行环境复杂多变，飞行空域广、速域大，火箭执行快速响应任务时缺乏准确返回空域大气和风场数据
Smart Images
Figure CN122819010A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of rocket technology, specifically, it is a method for guidance and verification of rocket reentry stage. Background Technology
[0002] The flight environment of vertical takeoff and landing (VTOL) rockets after reentry is complex and variable, with a wide flight airspace and a large velocity range. When rockets perform rapid response missions, there is a lack of accurate atmospheric and wind field data for the return airspace. When the rocket's first stage re-enters in the opposite direction, its irregular aerodynamic shape, the randomness of ground winds, fuel consumption during the rapid launch and recovery guidance phases, and structural deviations of the rocket body are all unmodeled dynamic and complex internal and external disturbances and strong uncertainties, which seriously affect the reliability of the rocket's return phase. Summary of the Invention
[0003] To address the shortcomings of existing technologies, this application provides a rocket reentry guidance and verification method, which includes: establishing statistical data on ground and high-altitude wind fields at the recovery point, including wind speed and direction at various altitudes; establishing an aerodynamic database for rocket reentry; and evaluating the reliability of the method based on the statistical data and the aerodynamic database.
[0004] In the above method, an optimal control problem model is established for the guidance problem of the reusable rocket's powered landing phase. The problem is then transformed into a second-order cone optimization problem through lossless convexity transformation and successive convexity transformation methods, and solved quickly using a custom solver based on the interior point method.
[0005] According to the above method, the reliability of the rocket's reliance on statistical data and aerodynamic databases to measure the method includes: the rocket controlling the engine's thrust vector based on statistical data and aerodynamic databases.
[0006] Based on the above method, the reliability of the rocket is measured by the target analysis. Attached Figure Description
[0007] Figure 1 According to some embodiments of this application, a schematic diagram of a rocket recovery coordinate system is shown;
[0008] Figure 2 According to some embodiments of this application, a flowchart of a rocket reentry stage guidance and verification method is shown. Detailed Implementation
[0009] The present application will now be described in detail with reference to specific embodiments. These embodiments will help those skilled in the art to further understand the present application, but do not limit the present application in any way. It should be noted that those skilled in the art can make several changes and improvements without departing from the concept of the present application. These all fall within the protection scope of the present application.
[0010] An ideal scenario for the final stage of the return journey is to keep the speed within a constraint range and the acceleration approximately zero within a dozen meters above the target point. In this way, the rocket will basically land in a slow-changing descent mode or maintain a certain near-Earth altitude. The setting of this buffer zone can effectively "absorb" various deviations.
[0011] This application first establishes statistical data on the ground and upper-air wind fields at the recovery point, records the ground and upper-air wind field data within the 0-2km height range of the recovery point, including the maximum wind speed and direction, minimum wind speed and direction, and average wind speed and direction at each height section, and stores the wind field data in a database.
[0012] Secondly, an aerodynamic database for reentry is established. Force and torque sets under air-rocket interaction at different angles of attack and Mach numbers are calculated. During the design phase, CFD numerical simulation is used to simulate the flow field during the launch vehicle's reentry process, calculating the rocket's six-component aerodynamic coefficients under different operating conditions and storing them in the database.
[0013] Among them, the axial force coefficient C of the launch vehicle a :
[0014]
[0015] The normal force coefficient C of the launch vehicle n :
[0016]
[0017] The lateral force coefficient C of the launch vehicle z :
[0018]
[0019] Among them, F X The axial force acting on the launch vehicle;
[0020] F Y The normal force acting on the launch vehicle;
[0021] F Z The lateral force acting on the launch vehicle;
[0022] For dynamic pressure;
[0023] S represents the reference area of the launch vehicle;
[0024] Rolling torque coefficient C mx :
[0025]
[0026] Yaw moment coefficient C my :
[0027]
[0028] Pitch moment coefficient C mz :
[0029]
[0030] Where l is the reference length of the launch vehicle;
[0031] M X This is the rolling torque;
[0032] M Y This is the yaw moment;
[0033] M Z This is the pitching moment.
[0034] Figure 1 A schematic diagram of a rocket recovery coordinate system is shown, such as... Figure 1 As shown, a station-centered rectangular coordinate system is established with the target landing point as the origin. The y-axis is opposite to the direction of gravity and perpendicular to the local horizontal plane, pointing upwards. The x-axis and z-axis are parallel to the local horizontal plane and point due north and east, respectively. The three axes form a right-handed coordinate system. Since the rocket's altitude above the ground is on the order of kilometers when the rocket's powered landing engine starts, the effect of gravity is simplified to a constant value in this stage. The maximum flight speed is less than 150 m / s, and non-inertial forces caused by the Earth's rotation are not considered. The dynamic equations for this stage can be expressed as:
[0035]
[0036] Where r and v represent the rocket's current position and velocity, respectively, g0 represents the gravitational acceleration at the landing point, m is the rocket's mass, and a d To mitigate drag interference caused by various aerodynamic deviations, I sp This represents the specific impulse of the rocket engine. The rocket's landing trajectory is constrained by the ground obstacle avoidance constraint angle β. The angular amplitude constraint of the thrust direction is denoted by γ. Furthermore, the constraints of the rocket landing phase also include a soft landing constraint, i.e., an acceleration of [0 m / s²] at 10 m above the target impact point. 2 0m / s 2 0m / s 2 The velocity is [0 m / s, -0.5 m / s, 0 m / s]. Considering minimizing fuel consumption during landing, the rocket's landing trajectory optimization can be described as follows:
[0037] minJ=-m(t f )
[0038] Among them, t f The state constraints are as follows, given the time required to reach a point 10 meters above the landing point:
[0039] r(t0)=r0
[0040] v(t0)=v0
[0041] m(t0)=m0
[0042] r(t f = [0, 10, 0]
[0043]
[0044] m(t f )≥m dry
[0045]
[0046] Where m dry This represents the dry weight of the rocket, that is, the mass of the rocket when all fuel is depleted. The control constraints are:
[0047] T min ≤||T||≤T max
[0048]
[0049] Since the landing speed is typically below 150 m / s, the drag term 'a' is used in the solution. d Ignoring this, the problem is included as a disturbance term in the Monte Carlo firing analysis, which improves the program's solution efficiency and meets the rocket's control frequency requirements during the reentry phase. The original problem is convexly optimized using variable substitution and relaxation techniques, while addressing the non-convex optimization constraint of the thrust amplitude. During convexification, time needs to be discretized, transforming it into linear equality constraints at discrete points. Due to the complexity of the rocket landing problem, the flight time is difficult to pre-set; therefore, the terminal moment is the unknown quantity. N discrete points are taken in the time domain, and the time interval between adjacent discrete points can be expressed as… Transforming the nonlinear dynamic equation into a discrete algebraic equation constraint and treating the time term as a separate optimization variable, and then linearizing the nonlinear equation, results in better convergence.
[0050] In this application, the specific steps of sequential convex optimization are as follows: First, let n = 0, and give initial values r0, v0, and control variable T for the state variable, control variable, and time variable. 0 And given the remaining time t f (During the first run, a predicted value can be given based on the standard trajectory, and iteratively updated in the algorithm). The constrained endpoint is defined as r(N) = [0, 10, 0], v(N) = [0, -0.5, 0]. Then at the reference point r (k) v (k) and t(k) Linearize the dynamic equations to obtain the SOCP subproblem. Then, solve the subproblem using the cvxpy general solver with a Python interface to obtain the state variable r. (k+1) v (k+1) Control variable T (k+1) This yields the state and control variables at each time point. Subsequent one-dimensional search updates the remaining time t. f This completes one round of online trajectory simulation, with the rocket based on the control variable T. (1) Control the engine's thrust vector.
[0051] The following is combined Figure 2 This paper introduces the guidance and verification methods for the rocket's return phase in this application. For example... Figure 2 As shown, the method includes:
[0052] S101: Establish a refined deviation model.
[0053] Establishing a refined surface wind model typically refers to improving the accuracy and reliability of the model by identifying and correcting biases during the modeling or prediction process.
[0054] S102: Number of shots k = 1.
[0055] S103: Select the deviations sequentially.
[0056] Bias are randomly generated based on a bias generator. A bias generator is a tool or method used in modeling or data analysis to artificially introduce or simulate biases. Bias generators can be implemented through data sampling, data perturbation, feature engineering, etc. The selection of biases starts from 1 and continues until all biases are selected.
[0057] S104: Sampling of the deviation distribution model.
[0058] For example, multiple samples are randomly drawn with replacement from the original training data. These samples are called bootstrap samples. In this way, we can simulate different training data drawn from the same distribution. Wind speed and direction are selected at various height sections.
[0059] S105: Determine whether all deviations have been selected.
[0060] If all deviations have been selected, proceed to step S106; otherwise, proceed to step S103.
[0061] S106: Perform a round of convex optimization to output the thrust vector of the rocket for the next step.
[0062] S107: Calculate the current rocket's angle of attack and Mach number.
[0063] Based on the rocket's current altitude and velocity, the angle of attack and Mach number of the rocket under the selected wind field are calculated. By querying the aerodynamic database, the aerodynamic forces acting on the rocket at this moment are interpolated. Within this time period (100ms), the impact of aerodynamic forces on the rocket's state is calculated using aerodynamic forces and mass, obtaining the acceleration effect 'a' caused by the deviation. d At the same time, update the state vectors r0 and v0 for the next step.
[0064] S108: Determine if r y ≤10.
[0065] If the condition is not met, return to S106 for the next step of convex optimization until the judgment condition r is reached. y ≤10, meaning the rocket reaches a point 10m above the predetermined point.
[0066] S109: Record the rocket landing trajectory and deviation selection in the database.
[0067] S110: Determine if k is equal to n.
[0068] Determine whether the number of shots k is equal to the preset threshold n. If they are equal, proceed to step S112; otherwise, proceed to step S111.
[0069] S111: K = K + 1.
[0070] S112: Sensitivity analysis.
[0071] Sensitivity analysis is a method used to evaluate how sensitive a model's output is to changes in input parameters. Through local sensitivity analysis and global sensitivity analysis, the impact of parameters can be comprehensively assessed.
[0072] Probabilistic design methods can be summarized as using the failure probability of a system as a reliability metric to measure its reliability. Through probabilistic design, the system's limiting states can be clearly defined. When a portion of the system's limiting states exceed a specific state and no longer meet a certain conjugate energy requirement specified in the design, this specific state is considered the limiting state. Target analysis can determine and evaluate the system's limiting states, thereby measuring the reliability of the guidance method.
[0073] For the guidance problem of the reusable rocket's powered landing phase, an optimal control problem model was established. The problem was transformed into a second-order cone optimization problem by using lossless convexity transformation and successive convexity transformation methods, and then solved quickly using a custom solver based on the interior point method.
[0074] The specific embodiments of this application have been described above. It should be understood that this application is not limited to the specific embodiments described above, and those skilled in the art can make various changes or modifications within the scope of the claims, which do not affect the substantive content of this application. Unless otherwise specified, the embodiments and features described in the embodiments of this application can be arbitrarily combined with each other.
Claims
1. A rocket reentry phase guidance and verification method, applied to rockets, characterized in that, include: Establish statistical data on the ground and upper-level wind fields at the recovery point, including wind speed and direction at each height section; Establish an aerodynamic database for rocket reentry; The rocket measures the reliability of the method based on the statistical data and aerodynamic database.
2. The method according to claim 1, characterized in that, The rocket's assessment of the method's reliability based on the statistical data and aerodynamic database includes: the rocket controlling the engine's thrust vector according to the statistical data and the aerodynamic database.
3. The method according to claim 1, characterized in that, The reliability of the method is measured based on the rocket's performance in firing analysis.