Carrier rocket distributed uncertainty multidisciplinary optimization design method and system

Abstract

The invention discloses a distributed uncertainty multidisciplinary optimization design method and system for a carrier rocket, and aims to process uncertain parameters, convert high-dimensional uncertain data into low-dimensional independent variables, reduce the calculation amount and improve the optimization design efficiency. When the data in the quantization parameter matrix is reconstructed, the optimization problem is solved based on the quantized parameters to obtain the probability density function, compared with a traditional uncertainty quantization means, the method can more accurately describe the probability distribution of the uncertain parameters, a reliable data basis is provided for subsequent optimization design, and the probability density function is obtained. The later calculation result is more accurate and does not need to depend on design experience. The constructed subject uncertainty and subject analysis agent model replaces high-cost high-fidelity subject analysis, and by constructing an offline agent model and combining with high-performance parallel calculation, frequent calling of a high-fidelity model in a UMDO solving process is avoided, and the calculation efficiency is remarkably improved.

Description

Technical Field

[0001] The present invention belongs to the field of refined design of launch vehicles, and relates to a distributed uncertainty multidisciplinary optimization design method and system for launch vehicles. Background Art

[0002] There are many uncertainties in the design, manufacture, and flight process of launch vehicles. Traditional engineering design methods consider uncertainties by reserving design margins to ensure system reliability. However, this method highly relies on design experience and is difficult to meet the future design requirements of rockets. To meet the requirements of high reliability and high safety for complex coupled systems, uncertainty-based multidisciplinary design optimization (UMDO) is an effective method to optimize the impact of these uncertainties.

[0003] In the engineering design of launch vehicles, not only the objectively existing random uncertainties in the design, manufacture, and use processes need to be considered, but also the cognitive uncertainties caused by insufficient subjective cognition or lack of information of people need to be considered, making the modeling of uncertainties in the UMDO process more complex. Moreover, in the process of solving UMDO problems, not only the propagation accuracy of uncertainties needs to be considered, but also uncertainty analysis needs to be repeatedly performed at each optimization search point to calculate the reliability and robustness of the system, resulting in high computational costs and slow convergence, which poses greater challenges for its engineering applications. Summary of the Invention

[0004] The purpose of the present invention is to solve the problems of large design margins, heavy dependence on design experience in traditional engineering design modes, high computational costs, and slow convergence in multidisciplinary uncertainty optimization design in the prior art, and to provide a distributed uncertainty multidisciplinary optimization design method and system for launch vehicles, so as to effectively improve the efficiency of refined design of launch vehicle engineering and reduce the design cost of launch vehicles.

[0005] To achieve the above object, the present invention adopts the following technical solutions:

[0006] A distributed uncertainty multidisciplinary optimization design method for launch vehicles, comprising the following steps:

[0007] Determine the uncertainty optimization design problem of the launch vehicle, input the uncertainty optimization design problem of the launch vehicle into the uncertainty and discipline analysis proxy model for analysis, and obtain the optimization analysis result;

[0008] The acquisition of the discipline uncertainty and discipline analysis proxy model includes:

[0009] Obtain the disciplinary parameters of the launch vehicle, quantify the uncertainty of the disciplinary parameters of the launch vehicle, and obtain a quantified parameter matrix;

[0010] Judge and process the certainty of the quantified parameter matrix to obtain a sample set of parameters. Judge and process the sample set of parameters to obtain a data set of uncertainty and certainty data;

[0011] Construct an uncertainty and disciplinary analysis surrogate model based on the data set of uncertainty and certainty data.

[0012] A further improvement of the present invention lies in:

[0013] The quantification of the uncertainty of the disciplinary parameters of the launch vehicle to obtain a quantified parameter matrix includes:

[0014] Determine the baseline scheme for the design of the launch vehicle. Based on the baseline scheme, analyze the inputs and outputs of the launch vehicle's power, the inputs and outputs of the launch vehicle's structure, the inputs and outputs of the launch vehicle's aerodynamics, the inputs and outputs of the launch vehicle's trajectory discipline, and the disciplinary transfer relationships of the launch vehicle to obtain all the disciplinary parameters and design space of the launch vehicle;

[0015] According to the disciplinary parameters and design space, combine the baseline scheme to determine the uncertainty optimization design problem of the launch vehicle. The uncertainty optimization design problem of the launch vehicle is used as the input of the uncertainty and disciplinary analysis surrogate model;

[0016] Judge and process based on the disciplinary parameters to obtain a sample set of s-dimensional independent vectors. Judge and process the sample set of s-dimensional independent vectors to obtain a quantified parameter matrix.

[0017] The judgment and processing based on the disciplinary parameters to obtain a sample set of s-dimensional independent vectors, and the judgment and processing of the sample set of s-dimensional independent vectors to obtain a quantified parameter matrix includes:

[0018] Judge the disciplinary parameter z:

[0019] If z is high-dimensional data, use the proper orthogonal decomposition algorithm to convert the sample set Z = [z 1 , z 2 ,..., z m ∈ R n×m ; of the disciplinary parameter z into a sample set V = [v 1 , v 2 ,..., v m ∈ R s×m , otherwise, directly assign the value of z to v to obtain the sample set V, where m is the number of samples and n is the dimension of the variable z;

[0020] Judge the independent vector v. If the independent vector v is an uncertain quantity, calculate the statistical moment parameters of v and form a quantization parameter matrix μ. If the independent vector v is a definite quantity, the quantization parameter matrix μ can be directly obtained, where μ ∈ R s×k 。

[0021] Judge and process the certainty of the quantization parameter matrix to obtain a set of sample variables, including:

[0022] Judge the certainty of the quantization parameter matrix μ ∈ R s×k :

[0023] If k > 1, μ represents an uncertain quantity. Construct an analytical function of the quantization parameter matrix μ, optimize the probability density function PDF of the subject variables through the analytical function, and sample the probability density function PDF to obtain a set of uncertain subject variable samples;

[0024] If k < 1, the quantization parameter matrix μ directly represents a set of samples of the deterministic subject variables.

[0025] Judge and process the set of sample variables to obtain a data set of uncertainty and certainty data, including:

[0026] If s > 1, for the modal coefficient part in the high-dimensional sample set, obtain the corresponding uncertainty and certainty high-dimensional data through POD reconstruction. Otherwise, directly output the uncertainty and certainty values.

[0027] Construct an uncertainty and subject analysis surrogate model based on the data set of uncertainty and certainty data, including:

[0028] According to the data set of uncertainty and certainty data, combined with the certainty of each subject variable of the launch vehicle, perform uncertainty processing:

[0029] If each subject variable is uncertain, perform uncertainty analysis to obtain the subject output response;

[0030] If each subject variable is certain, perform subject analysis to obtain the subject output response;

[0031] Quantify the uncertainty of the output response to obtain a quantization parameter matrix of the output response;

[0032] Construct an uncertainty and subject analysis surrogate model based on the quantization parameter matrix of the output response and each subject variable of the launch vehicle.

[0033] Each subject variable of the launch vehicle includes a launch vehicle engine variable and a launch vehicle mass variable.

[0034] A distributed uncertainty multidisciplinary optimization design system for a launch vehicle, including:

[0035] An optimization module for optimizing the uncertain multidisciplinary problems of a launch vehicle through an uncertainty and discipline analysis surrogate model;

[0036] The acquisition of the discipline uncertainty and discipline analysis surrogate model includes:

[0037] A quantization parameter matrix construction module for obtaining the discipline parameter variables of the launch vehicle, quantifying the uncertainty of the discipline parameter variables of the launch vehicle to obtain a quantization parameter matrix;

[0038] A data set construction module for judging the certainty of the quantization parameter matrix to obtain a parameter variable sample set, and judging and processing the parameter variable sample set to obtain a data set of uncertainty and certainty data;

[0039] A model acquisition module for constructing an uncertainty and discipline analysis surrogate model based on the data set of uncertainty and certainty data.

[0040] A terminal device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of any method of the present invention are implemented.

[0041] A computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of any method of the present invention are implemented.

[0042] Compared with the prior art, the present invention has the following beneficial effects:

[0043] The present invention discloses a distributed uncertainty multidisciplinary optimization design method for a launch vehicle. For the treatment of uncertainty, high-dimensional uncertain data is transformed into low-dimensional independent data, reducing the calculation amount and improving the optimization design efficiency. When reconstructing the data in the quantization parameter matrix, an optimization problem is solved based on the quantized parameters to obtain a probability density function. Compared with traditional uncertainty quantization means, this method can more accurately describe the probability distribution of uncertain parameters, providing a reliable data basis for subsequent optimization design, with more accurate calculation results in the later stage, not relying on design experience. The constructed discipline uncertainty and discipline analysis surrogate model replaces high-cost high-fidelity discipline analysis. By constructing an offline surrogate model combined with high-performance parallel computing, frequent calls to high-fidelity models during the UMDO solution process are avoided, significantly improving the calculation efficiency, solving the problems of excessive calculation burden, large calculation cost, and slow convergence in the engineering application of traditional methods, and the calculation accuracy of the model is better. Description of the Drawings

[0044] To more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the accompanying drawings required for the embodiments. It should be understood that the following drawings only show some embodiments of the present invention, and therefore should not be regarded as a limitation of the scope. For those of ordinary skill in the art, without creative efforts, other related drawings can also be obtained based on these drawings.

[0045] Figure 1 It is a flowchart of the distributed uncertainty multidisciplinary optimization design method for launch vehicles;

[0046] Figure 2 It is a flowchart for calculating uncertainty quantification parameters;

[0047] Figure 3 It is a flowchart of the inverse parameterization process;

[0048] Figure 4 They are the maximum, minimum, and optimal sizes of the launch vehicle;

[0049] Figure 5 It is a design matrix diagram of the optimization problem of the launch vehicle;

[0050] Figure 6 It is a flowchart of the hybrid optimization of the launch vehicle;

[0051] Figure 7 It is the system-level optimization iteration process of the launch vehicle;

[0052] Figure 8 They are the uncertainties of the ballistic parameters of the launch vehicle; (where a is the variation of velocity with time; b is the trajectory and uncertainty of the launch vehicle in the launch coordinate system);

[0053] Figure 9 They are the probability density distribution histograms of important parameters (where a is the uncertainty distribution of the booster separation time and the core stage working end time; b is the uncertainty distribution of the total mass flow rate and the vacuum thrust F P ; c is the uncertainty distribution of the oxygen-fuel ratio r P and the specific impulse I s ). Specific Embodiments

[0054] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. Usually, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0055] Accordingly, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the scope of protection of the present invention.

[0056] It should be noted that like reference numerals and letters denote like items in the following drawings, and thus, once an item is defined in one drawing, it is not necessary to further define and explain it in subsequent drawings.

[0057] In the description of the embodiments of the present invention, it should be noted that if terms such as "upper", "lower", "horizontal", "inner", etc. are used to indicate the orientation or positional relationship, it is based on the orientation or positional relationship shown in the drawings, or the orientation or positional relationship in which the product of the invention is usually placed during use. This is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus should not be construed as a limitation of the present invention. In addition, terms such as "first", "second", etc. are only used for distinguishing descriptions and cannot be understood as indicating or implying relative importance.

[0058] In addition, if the term "horizontal" appears, it does not mean that the component is required to be absolutely horizontal, but it can be slightly inclined. For example, "horizontal" only means that its direction is more horizontal relative to "vertical", and does not mean that the structure must be completely horizontal, but it can be slightly inclined.

[0059] In the description of the embodiments of the present invention, it should also be noted that unless otherwise clearly specified and limited, if terms such as "set", "installed", "connected", "coupled" are used, they should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be directly connected or indirectly connected through an intermediate medium, and it can be the communication inside two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific situations.

[0060] The present invention will be further described in detail below with reference to the accompanying drawings:

[0061] See Figures 1 to 3 , the embodiments of the present invention disclose a distributed uncertainty multidisciplinary optimization design method for launch vehicles, mainly aiming at the problems of large design margins, heavy dependence on design experience in the traditional engineering design mode of launch vehicles, and high computational cost and slow convergence in multidisciplinary uncertainty optimization design.

[0062] It includes the following steps:

[0063] Step 1: Quantify the uncertainties of the parametric variables of the launch vehicle to obtain a quantified parameter matrix;

[0064] Step 1.1: Determine the baseline design of the launch vehicle, and analyze the input-output and discipline transfer relationships of the disciplines of the launch vehicle's power, structure, aerodynamics, and trajectory;

[0065] Specifically include:

[0066] Step 1.2: Based on the discipline transfer relationships and input-outputs in Step 1.1, determine all parametric variables and design spaces, and clarify the uncertainty optimization design problem of the launch vehicle in combination with the baseline design of the launch vehicle;

[0067] Step 1.3: Judge the parametric variable z obtained in S1.2:

[0068] If z is high-dimensional data, use the Proper Orthogonal Decomposition (POD) algorithm to transform the sample set Z = [z 1 , z 2 ,..., z m ∈ R n×m (m is the number of samples, n is the dimension of the variable z) into a sample set V = [v 1 , v 2 ,..., v m ∈ R s×m of s-dimensional independent vectors v. Otherwise, directly assign the value of z to v to obtain the sample set V;

[0069] Judge the independent vector v:

[0070] If the independent vector v is an uncertain quantity, calculate the statistical moment parameters of v and form a quantified parameter matrix μ. If the independent vector v is a certain quantity, the matrix μ can be directly obtained, μ ∈ R s×k .

[0071] Step 2: Inverse parameterize the quantified parameter matrix to obtain a data set;

[0072] Specifically include:

[0073] Step 2.1: Judge the determinacy of the matrix μ:

[0074] If k > 1, then μ represents an uncertain quantity. Based on the Maximum Entropy (MaxEnt) principle, construct an analytical function of the quantified parameter matrix μ, optimize to obtain the probability density function (PDF) of the parametric variable, and finally, sample the PDF to obtain a sample set of uncertain parametric variables.

[0075] If k < 1, μ can directly represent the sample set of deterministic parameters without processing μ.

[0076] Step 2.2: Judge the dimension of the parameter sample set obtained in Step 2.1:

[0077] If s > 1, for the modal coefficient part in the high-dimensional sample set, corresponding uncertain and deterministic high-dimensional data are obtained through POD reconstruction; otherwise, the uncertain and deterministic values are directly output.

[0078] Step 3: Establish a disciplinary uncertainty surrogate model and a disciplinary analysis surrogate model based on the uncertainty quantification and inverse parameterization process;

[0079] Step 3.1: For the parameter sample set obtained in Step 2.2, different treatments are carried out according to the deterministic situations of disciplinary parameters such as the carrier rocket engine and mass (the specific variable determination situation can be judged based on Step 2). If the variable is uncertain, uncertainty analysis is performed to obtain the disciplinary output response; if the variable is determined, disciplinary analysis is performed to obtain the disciplinary output response.

[0080] Step 3.2: Perform uncertainty quantification on the output response to obtain the output quantization parameter matrix;

[0081] Step 3.3: Construct a surrogate model based on the input variables and the quantization parameter matrix of the corresponding output response.

[0082] Step 4: Based on the constructed disciplinary uncertainty surrogate model, solve the multidisciplinary uncertainty optimization problem of the carrier rocket until the optimization result converges.

[0083] Based on the uncertainty surrogate model obtained in Step 3 to replace the uncertainty propagation process, and the disciplinary analysis surrogate model to replace the disciplinary analysis process, solve the multidisciplinary uncertainty optimization problem of the carrier rocket in Step 2 until the optimization result converges.

[0084] Step 5: Evaluate the final optimization design result obtained in S4, and check whether indicators such as the strength of the structure and the orbital accuracy meet the design standards.

[0085] See Figures 4 to 9 , and the present invention also discloses a specific embodiment:

[0086] Step 1: Perform uncertainty quantification on the disciplinary parameters of the carrier rocket to obtain the quantization parameter matrix.

[0087] The specific steps are as follows:

[0088] Step 1.1: The launch vehicle in this embodiment is a single-stage semi-rocket with four boosters. Both the core stage and the boosters use 2 identical liquid engines, and their propellants are liquid oxygen (LOX) and methane (CH4). The mission of the launch vehicle is to deliver a 50,000 kg payload to an orbit with a semi-major axis of 6,629,387 m, an eccentricity of 0.012104, and an orbital inclination of 0.723962 rad. The diameters of the core stage and the boosters are 5 m and 3.35 m respectively, and each booster has a fixed wing of a fixed size. The disciplines involved in the launch vehicle include aerodynamics, power, structure, mass, trajectory, and guidance and control. The disciplinary analysis model of aerodynamics is used to calculate the drag coefficient, lift coefficient, and distributed pressure coefficient under different aerodynamic shapes and flight conditions. The aerodynamic shape parameters are the lengths of the core stage and the boosters, and their variation ranges are as Figure 4 shown. The disciplinary analysis model of power calculates the specific impulse I P according to the oxygen-fuel ratio r N , the nozzle expansion ratio r C , and the combustion chamber pressure p S . The engine thrust can be obtained from , where r E is the efficiency of the nozzle, and is the total mass flow rate of LOX and CH4. The disciplinary analysis model of structure is used to calculate the structural mass under a given load while ensuring structural reliability. The loads include aerodynamic loads, the axial force and normal force of the engine. The structural reliability of the i-th component of the launch vehicle is expressed as where σ i is the maximum stress of the i-th component, and is the allowable yield strength of the i-th component. Then the total reliability can be expressed as:

[0089]

[0090] where N is the number of components, and the maximum stress is calculated by the Finite Element Method (FEM); the lengths and diameters of the components are fixed parameters, and the thicknesses t s of the water tank, fairing, and interstage are design variables. The structural mass m s is obtained by solving the optimization problem:

[0091] min m s

[0092] w.r.t.t s (2)

[0093] s.t.P s (σ s <σ * )≥0.999

[0094] The mass discipline model calculates the propellant mass m P , the total mass m T and the centroid position X c . The liquid propellant height is a design variable, the length of the fuel tank is calculated based on the liquid propellant height, and the ellipsoid ratio of the fuel tank is a fixed parameter. The trajectory discipline model uses deterministic inputs to plan the standard trajectory. Then, the guidance and control disciplines track the standard trajectory and the orbit with uncertain inputs. The three-degree-of-freedom dynamics equation is used as the trajectory model, and then the pseudospectral method with the goal of minimizing the remaining propellant mass m R is used to optimize the trajectory. For the guidance discipline, perturbation guidance is applied before booster separation to track the standard trajectory, and iterative guidance is used after separation to track the orbit. Finally, PID is applied as the control model to calculate the nozzle angle. The reliability P M of the launch vehicle's orbital mission is the probability of successful orbit insertion, and the mission fails when all the propellant is consumed and the trajectory parameters cannot be met.

[0095] Step 1.2: Based on the discipline transfer relationships and inputs / outputs in Step 1.1, determine all discipline parameters and design spaces, as shown in Table 1.

[0096] Table 1 Design Spaces of Discipline Parameters

[0097]

[0098]

[0099] Combined with the baseline scheme of the launch vehicle design in Step 1.1, the uncertainty optimization design problem of the launch vehicle can be clarified. The uncertainty optimization design of the launch vehicle in this embodiment includes three sub-optimization problems: structural mass optimization, trajectory optimization, and overall system optimization.

[0100] The system optimization aims to minimize the total mass of the launch vehicle. In system optimization, the total mass of the launch vehicle is taken as the objective function, and design variables such as structural thickness and engine parameters are optimized considering constraint conditions such as structural strength and orbit accuracy; in trajectory optimization, the remaining propellant mass is minimized as the goal to optimize the flight trajectory of the launch vehicle; in structural optimization, the structural mass is optimized on the premise of ensuring structural reliability. The uncertainty optimization problem of the launch vehicle in this embodiment is specifically described as follows:

[0101]

[0102] where ε Orbit is the relative error of the orbit. The design structure matrix of the optimization problem is as Figure 5 shown, where q is the dynamic pressure and h is the current liquid propellant height.

[0103] Step 1.3: Quantify the uncertainty of the disciplinary parameters obtained in Step 1.2.

[0104] Among them, the distributed pressure coefficient C in the aerodynamics discipline P and other parametric variables are high-dimensional data. Using the Proper Orthogonal Decomposition (POD) algorithm, the data sample set of C P is converted into a sample set V = [v 1 , v 2 ,..., v m of s-dimensional independent uncertainty vectors. For other disciplinary variables such as lift and drag coefficients and structural mass, which are numerical data, the values of such data are directly assigned to v to form a data sample set v; then, the uncertainty of the disciplinary parameters is judged. For disciplinary parameters with uncertainty such as the shape parameters of the launch vehicle such as the structural thickness t S and environmental parameters such as the gravitational acceleration g, calculate the statistical moment parameters of the independent uncertainty vector v obtained after POD reduction, and form a quantization parameter matrix μ ∈ R s×k . For deterministic disciplinary parameters, the matrix μ ∈ R s×k can be directly obtained.

[0105] Specifically, calculate the statistical moment parameters of the independent uncertainty vector v. If v is an uncertain vector, the statistical moment of v will be calculated as a quantization parameter. Let the mathematical expectation of v be μ = E(v), then the k-th central moment μ k (k = 2, 3,...) can be obtained from the sample set v:

[0106]

[0107] The vector moment is expressed as μ = [μ, μ2, μ... μ k ∈ R k . Since v consists of s independent variables, μ1 to μ k are calculated independently.

[0108] Construct the parameter matrix μ = [μ1,... μ s T ∈ R s×k , specifically as follows:

[0109]

[0110] The first column represents the expected vector E(v), and the remaining columns represent the moments of each order. If z is a definite value, then s = k = 1, μ = [μ1]. If z is an uncertain value, then s = 1, k > 1, and μ = [μ1, μ 1，2 ,... μ 1，k. If z is a determined high-dimensional data, then s > 1, k = 1, and μ = [μ1, μ 1，2 ,... μ s T . If z is an undetermined high-dimensional data, then s > 1 and k > 1.

[0111] Step 2: Inverse parameterize the quantization parameter matrix to obtain a data set. The specific steps are as follows:

[0112] Step 2.1: Judge the matrix μ. If k > 1, then μ represents an uncertainty quantity. Based on the Maximum Entropy (MaxEnt) principle, construct an analytical function of the quantization parameter matrix μ, optimize to obtain the probability density function (PDF) of the disciplinary parameter variable, and finally, sample the PDF to obtain a sample set of the uncertainty parameter variable; if k < 1, μ can directly represent the sample set of the deterministic parameter variable, and no processing is performed on μ.

[0113] Based on the Maximum Entropy (MaxEnt) principle, construct an analytical function of the quantization parameter matrix for the given parameter variable. For each row of the given undetermined quantization parameter matrix μ, calculate the probability density function (PDF) p(x) through the Maximum Entropy (MaxEnt) principle. The analytical function constructed by the maximum entropy is:

[0114]

[0115] where λ i (i = 1, 2,..., k) are the coefficients of the analytical function, and k is the central moment order of the quantization parameter matrix. Then, solve an optimization problem to obtain λ = [λ1,... λ k T :

[0116]

[0117] Adopt the Differential Evolution (DE) algorithm to search for appropriate initial values of λ, and then use the sequential quadratic programming algorithm to solve the above optimization problem. After obtaining the optimal coefficients, calculate p(x) according to the formula.

[0118] Step 2.2: Judge the dimension of the sample set of the parameter variable obtained in Step 2.1:

[0119] If s > 1, for the modal coefficient part in the high-dimensional sample set, obtain the corresponding high-dimensional data sample set through POD reconstruction; otherwise, directly output the sample set of the deterministic and uncertainty values. ​​

[0120] Step 3: Establish the uncertainty surrogate model and the disciplinary analysis surrogate model for each discipline according to the quantization parameter matrix. The specific steps are as follows:

[0121] Step 3.1: For the set of parametric variable samples obtained in Step 2.2, different treatments are performed according to the certainty of the disciplinary parametric variables such as the launch vehicle engine and mass (Step 2 can be used as the judgment basis). If the variable is uncertain, uncertainty analysis is performed to obtain the disciplinary output response; if the variable is certain, disciplinary analysis is performed to obtain the disciplinary output response;

[0122] Step 3.2: Perform uncertainty quantization on the output response to obtain the output quantization parameter matrix;

[0123] Step 3.3: Construct a surrogate model based on the quantization parameter matrix of the input variables and the corresponding output response. For the aerodynamic discipline, a surrogate model is constructed by combining the Back Propagation Neural Network (BPNN) with the POD method to predict aerodynamic forces and pressure distributions; for the structural discipline, a Kriging model is constructed as the surrogate model to predict structural stresses and mass; for the propulsion discipline, a BPNN model is constructed as the surrogate model to predict engine performance parameters; for the trajectory discipline, a surrogate model is constructed by combining the pseudospectral method with the BPNN model to predict the rocket's trajectory and orbit injection accuracy.

[0124] Step 4: Based on the disciplinary uncertainty surrogate model obtained in Step 2 to replace the uncertain analysis process of the discipline, and combined with the data set obtained in Step 3, solve the multidisciplinary uncertainty optimization problem of the launch vehicle until the optimization result converges. Take Figure 4 the shown multidisciplinary hybrid optimization process to solve the multidisciplinary uncertainty optimization problem of the launch vehicle. The system optimization is solved by the differential genetic algorithm, the trajectory optimization is solved by the sparse nonlinear optimizer algorithm, and the structural optimization is solved by the sequential quadratic programming algorithm.

[0125] Step 5: After the optimization in Step 4, the convergence of the system optimization objective function and the reliability constraints are as Figure 7 shown. At the beginning of the optimization, due to the conflict with P M the objective function is very large. After the 19th iteration, the constraints are fully satisfied, and the optimization direction is to reduce the total mass m T of the launch vehicle. From the 19th to the 100th iteration, the total mass m T of the launch vehicle is reduced by 98,379 kg. Table 2 lists the important optimal design variables, state variables, and responses. The payload weight ratio of the launch vehicle is 0.0285, and the launch thrust-to-weight ratio is 1.277. To obtain a greater launch thrust, the mass flow rate of the propellant Close to the upper limit. Since the engine will shut down when either LOX or CH4 is exhausted, the ratio of the propellant masses of LOX and CH4 is close. Finally, the optimal outer dimensions of the optimized launch vehicle are in the middle of the allowable dimension range, as Figure 6 shown.

[0126] Table 2 Parameters of the optimal scheme

[0127]

[0128] The uncertainty of the parameters along the trajectory is as Figure 8 (a) shown. At 173.7 s after launch, the booster separates at an altitude of 60 km. The uncertainty of the rocket flight speed is compensated by increasing the flight time to the final orbital point. Therefore, to meet the constraints of P M , the remaining propellant mass m R is 6758 kg. According to Figure 8 (b), it can be seen that the deviation of the rocket trajectory is the largest at the highest point of the trajectory and then gradually decreases due to iterative guidance. By quantifying the uncertainty based on 10,000 samples, the probability density distribution function of the parameters can be obtained, and the oxygen-fuel ratio r P , specific impulse I s , booster separation time, core stage working end time, total mass flow rate and vacuum thrust F P of the uncertainty distributions are respectively as Figure 9 shown.

[0129] The results show that the optimized design scheme of the launch vehicle can significantly improve the payload while meeting the requirements of the orbital injection accuracy, and has high reliability. Therefore, the proposed multi-disciplinary optimization design method for distributed uncertainty of launch vehicles is verified through a specific implementation case of the liquid launch vehicle optimization problem, and the proposed optimization design method can effectively obtain the optimal solution.

[0130] The present invention provides a multi-disciplinary optimization design method for distributed uncertainty of launch vehicles, which can reduce the expensive computational cost for solving the UMDO problem based on high-fidelity disciplinary analysis. This method uses a surrogate model to replace the high-cost high-fidelity disciplinary analysis. By constructing an offline surrogate model combined with high-performance parallel computing, it avoids frequent calls to the high-fidelity model during the UMDO solution process, significantly improves the computational efficiency, and solves the problem of excessive computational burden in the engineering application of traditional methods.

[0131] For the handling of uncertainties, this method uses the POD algorithm to transform complex high-dimensional uncertain data into low-dimensional independent variables, and then calculates the statistical moments of the modal coefficients to parameterize the variables, which can significantly reduce the computational load and improve the efficiency of the optimization design. By the principle of maximum entropy (MaxEnt), the optimization problem is solved based on the quantified parameters to obtain the probability density function. Compared with traditional uncertainty quantification methods, this method can more accurately describe the probability distribution of uncertain parameters and provide a reliable data basis for subsequent optimization design. In addition, this method provides an uncertainty propagation method based on surrogate models for approximating the disciplinary uncertainty analysis in the UMDO framework, which can transform the UMDO problem into an MDO problem and then effectively solve the problem through a mature MDO framework.

[0132] A distributed uncertainty multidisciplinary optimization design system for a launch vehicle, characterized by comprising:

[0133] An optimization module for optimizing the uncertainty multidisciplinary problems of a launch vehicle through an uncertainty and disciplinary analysis surrogate model;

[0134] The acquisition of the disciplinary uncertainty and disciplinary analysis surrogate model includes:

[0135] A quantified parameter matrix construction module for obtaining the disciplinary parameter variables of a launch vehicle, performing uncertainty quantification on the disciplinary parameter variables of the launch vehicle, and obtaining a quantified parameter matrix;

[0136] A data set construction module for judging the determinacy of the quantified parameter matrix to obtain a sample set of parameter variables, and performing judgment processing on the sample set of parameter variables to obtain a data set of uncertainty and determinacy data;

[0137] A model acquisition module for constructing an uncertainty and disciplinary analysis surrogate model based on the data set of uncertainty and determinacy data.

[0138] The schematic diagram of the terminal device provided by an embodiment of the present invention. The terminal device of this embodiment includes: a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps in the above-mentioned method embodiments are implemented. Alternatively, when the processor executes the computer program, the functions of each module / unit in the above-mentioned device embodiments are implemented.

[0139] The computer program can be divided into one or more modules / units, and the one or more modules / units are stored in the memory and executed by the processor to complete the present invention.

[0140] The terminal device may be a computing device such as a desktop computer, a notebook, a palm computer, and a cloud server. The terminal device may include, but is not limited to, a processor and a memory.

[0141] The processor may be a central processing unit (CPU), or may also be other general-purpose processors, digital signal processors (DSPs), application specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.

[0142] The memory may be used to store the computer program and / or module. The processor realizes various functions of the terminal device by running or executing the computer program and / or module stored in the memory, and by calling the data stored in the memory.

[0143] If the modules / units integrated in the terminal device are implemented in the form of software functional units and sold or used as independent products, they may be stored in a computer-readable storage medium. Based on this understanding, to implement all or part of the processes in the above method embodiments of the present invention, it may also be completed by a computer program instructing relevant hardware. The computer program may be stored in a computer-readable storage medium. When the computer program is executed by the processor, the steps of the above method embodiments may be implemented. Among them, the computer program includes computer program code, and the computer program code may be in the form of source code, object code, executable file, or some intermediate form, etc. The computer-readable medium may include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a mobile hard disk, a magnetic disk, an optical disc, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc. It should be noted that the content included in the computer-readable medium may be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, the computer-readable medium does not include electrical carrier signals and telecommunication signals.

[0144] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A distributed uncertainty multidisciplinary optimization design method for a launch vehicle, characterized in that It includes the following steps: Determine the uncertainty optimization design problem of the launch vehicle, input the uncertainty optimization design problem of the launch vehicle into the uncertainty and discipline analysis surrogate model for analysis, and obtain the optimization analysis result; The acquisition of the discipline uncertainty and discipline analysis surrogate model includes: Obtain the discipline parameter variables of the launch vehicle, quantify the uncertainty of the discipline parameter variables of the launch vehicle, and obtain a quantified parameter matrix; Judge and process the certainty of the quantified parameter matrix to obtain a sample set of parameter variables, and judge and process the sample set of parameter variables to obtain a data set of uncertainty and certainty data; Construct an uncertainty and discipline analysis surrogate model based on the data set of uncertainty and certainty data.

2. A distributed uncertainty multidisciplinary optimization design method for a launch vehicle according to claim 1, characterized in that The quantification of the uncertainty of the discipline parameter variables of the launch vehicle to obtain a quantified parameter matrix includes: Determine the benchmark scheme of the launch vehicle design, analyze the input and output of the launch vehicle power, the input and output of the launch vehicle structure, the input and output of the launch vehicle aerodynamics, the input and output of the launch vehicle trajectory discipline, and the discipline transfer relationship of the launch vehicle based on the benchmark scheme, and obtain all the discipline parameter variables and design space of the launch vehicle; According to the discipline parameter variables and design space, combine the benchmark scheme to determine the uncertainty optimization design problem of the launch vehicle, and use the uncertainty optimization design problem of the launch vehicle as the input of the uncertainty and discipline analysis surrogate model; Judge and process based on the discipline parameter variables to obtain a sample set of s-dimensional independent vectors, and judge and process the sample set of s-dimensional independent vectors to obtain a quantified parameter matrix.

3. A method for distributed uncertainty multidisciplinary optimization design of a launch vehicle according to claim 2, characterized in that The judgment and processing based on the discipline parameter variables to obtain a sample set of s-dimensional independent vectors, and the judgment and processing of the sample set of s-dimensional independent vectors to obtain a quantified parameter matrix includes: Judge the discipline parameter variable z: If z is high-dimensional data, the sample set Z = [z 1 , z 2 ,..., z m ∈ R n ×m ; of the parameter variable z is converted into the sample set V = [v 1 , v 2 ,..., v m ∈ R s×m of s-dimensional independent vectors v. Otherwise, the value of z is directly assigned to v to obtain the sample set V, where m is the number of samples and n is the dimension of the variable z; Judge the independent vector v. If the independent vector v is an uncertain quantity, calculate the statistical moment parameters of v and form a quantization parameter matrix μ. If the independent vector v is a definite quantity, the quantization parameter matrix μ can be directly obtained, where μ ∈ R s×k 。 4. A distributed uncertainty multidisciplinary optimization design method for a launch vehicle according to claim 1, characterized in that, The judgment and processing of the certainty of the quantified parameter matrix to obtain a sample set of parameter variables includes: Determine the certainty of the quantization parameter matrix μ ∈ R s×k : If k > 1, then μ represents the uncertainty quantity, construct an analytical function of the quantified parameter matrix μ, optimize the probability density function PDF of the discipline parameter variable through the analytical function, and sample the probability density function PDF to obtain a sample set of uncertainty parameter variables; If k < 1, the quantified parameter matrix μ directly represents the sample set of deterministic parameter variables.

5. A multidisciplinary optimization design method for distributed uncertainties of a launch vehicle according to claim 4, characterized in that The judgment and processing of the sample set of parameter variables to obtain a data set of uncertainty and certainty data includes: If s > 1, for the modal coefficient part in the high-dimensional sample set, obtain the corresponding uncertainty and deterministic high-dimensional data through POD reconstruction, otherwise, directly output the uncertainty and deterministic values.

6. The distributed uncertainty multidisciplinary optimization design method for a launch vehicle according to claim 1, characterized in that The construction of the uncertainty and discipline analysis surrogate model based on the data set of uncertainty and certainty data includes: According to the data set of uncertainty and certainty data, combined with the certainty of each discipline parameter variable of the launch vehicle, perform uncertainty processing: If each discipline parameter variable is uncertain, perform uncertainty analysis to obtain the discipline output response; If each discipline parameter variable is certain, perform discipline analysis to obtain the discipline output response; Quantify the uncertainty of the output response to obtain a quantified parameter matrix of the output response; Construct an uncertainty and disciplinary analysis surrogate model based on the quantization parameter matrix of the output response and the various disciplinary parameters of the launch vehicle.

7. A distributed uncertainty multidisciplinary optimization design method for a launch vehicle according to claim 1, characterized in that The various disciplinary parameters of the launch vehicle include the launch vehicle engine parameters and the launch vehicle mass parameters.

8. A distributed uncertainty multidisciplinary optimization design system for a launch vehicle, characterized in that, It includes: An optimization module for optimizing the multi-disciplinary uncertainty problem of the launch vehicle through the uncertainty and disciplinary analysis surrogate model; The acquisition of the disciplinary uncertainty and disciplinary analysis surrogate model includes: A quantization parameter matrix construction module for obtaining the disciplinary parameters of the launch vehicle, quantifying the uncertainty of the launch vehicle disciplinary parameters, and obtaining a quantization parameter matrix; A data set construction module for judging the certainty of the quantization parameter matrix to obtain a parameter sample set, and performing judgment processing on the parameter sample set to obtain a data set of uncertainty and certainty data; A model acquisition module for constructing an uncertainty and disciplinary analysis surrogate model based on the data set of uncertainty and certainty data.

9. A terminal device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, the steps of the method according to any one of claims 1-7 are implemented.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, the steps of the method according to any one of claims 1-7 are implemented.

Images