Radial double-V-groove flow vortex regulation and control flame stabilization method of jet propulsion device

Abstract

The invention relates to a radial double-V-groove flow vortex regulation and control flame stabilization method for a jet propulsion device. The method comprises the following steps: acquiring flow field data of an afterburner and a spray pipe; carrying out dimension reduction processing on the flow field data, and determining low-order modal feature data; determining target feature data from the low-order modal features according to a preset optimization method, wherein the preset optimization method is realized according to an adaptive multi-response surface model and a genetic algorithm; and inputting the target characteristic data into a flame stability prediction model, and determining the flame stability according to the output of the flame stability prediction model. By adopting the method, the flame stability can be accurately determined.

Description

Technical Field

[0001] The present application relates to the technical field of jet propulsion devices, and in particular to a method for regulating flame stabilization of a jet propulsion device using radial double V-grooves and flow vortexes. Background Art

[0002] As a key component of military aviation engines, the afterburner can significantly increase thrust in a short period of time, supporting fighters to achieve short-distance takeoff, air maneuvers and supersonic cruise. The radial double V-groove flame stabilizer is an important component of the afterburner, and its flame stability is affected by many factors, including oil-gas ratio, electrode position, parasitic capacitance and installation size. In addition, factors such as violent fluctuations in aerodynamic heat caused by sudden changes in engine operating conditions, structural deformation and electrical system abnormalities will also affect its stability.

[0003] Therefore, when determining the flame stability of the radial double V-groove, a method that can accurately determine the flame stability is urgently needed. Summary of the invention

[0004] Based on this, it is necessary to provide a radial double V-groove flow vortex control flame stabilization method for a jet propulsion device that can accurately determine the flame stability in order to address the above technical problems.

[0005] In a first aspect, the present application provides a method for controlling flame stabilization of a jet propulsion device by using radial double V-groove flow vortexes, comprising:

[0006] Obtain flow field data of afterburner and nozzle;

[0007] Perform dimension reduction processing on flow field data to determine low-order modal characteristic data;

[0008] Determining target feature data from low-order modal features according to a preset optimization method, wherein the preset optimization method is implemented according to an adaptive multi-response surface model and a genetic algorithm;

[0009] The target characteristic data is input into the flame stability prediction model, and the flame stability is determined according to the output of the flame stability prediction model.

[0010] In one embodiment, acquiring flow field data of an afterburner and a nozzle includes:

[0011] Acquire the geometric model of the afterburner and the nozzle; determine the flow field data of the afterburner and the nozzle according to the preset aerodynamic thermal coupling calculation method and the geometric model; wherein the preset aerodynamic thermal coupling calculation method is an aerodynamic thermal coupling calculation method corrected according to the experimental method, and the experimental method includes particle image velocimetry and thermocouple temperature measurement.

[0012] In one embodiment, the low-order mode feature data includes low-order mode basis coefficients and low-order mode basis vectors. Determining target feature data from the low-order mode features according to a preset optimization method includes:

[0013] Determining a predicted value corresponding to the low-order mode basis coefficient according to the low-order mode basis coefficient and the preset optimization method, where the predicted value corresponding to the low-order mode basis coefficient is a value within a preset numerical range; determining the target feature data according to the product of the predicted value of the low-order mode basis coefficient and the low-order mode basis vector.

[0014] In one embodiment, the genetic algorithm is the Non-dominated Sorting Genetic Algorithm II. Determining a predicted value corresponding to the low-order mode basis coefficient according to the low-order mode basis coefficient and the preset optimization method includes:

[0015] Determining the mapping relationship between the low-order mode basis coefficient and the optimization objective according to the adaptive multi-response surface model, where the optimization objective includes flow field data; using the low-order mode basis coefficient as the input variable, and determining the predicted value corresponding to the low-order mode basis coefficient according to the Non-dominated Sorting Genetic Algorithm II and the mapping relationship.

[0016] In one embodiment, determining the mapping relationship between the low-order mode basis coefficient and the optimization objective according to the adaptive multi-response surface model includes:

[0017] Sampling the low-order mode basis coefficient according to a preset sampling method to determine sample points, where the sampling method includes the Latin hypercube sampling method, the central composite sampling method, or the Box-Behnken sampling method; updating the sample points according to the distribution of the sample points and the adaptive grid refinement technique; determining the mapping relationship between the updated sample points and the corresponding optimization objective according to the adaptive multi-response surface model; determining the accuracy of the mapping relationship, and when the accuracy does not meet the preset requirements, continuing to sample the low-order mode basis coefficient according to the distribution of the sample points and the adaptive grid refinement technique until the accuracy of the mapping relationship between the low-order mode basis coefficient and the corresponding optimization objective meets the preset requirements.

[0018] In one embodiment, performing dimensionality reduction processing on the flow field data to determine low-order mode feature data includes:

[0019] Performing dimensionality reduction processing on the flow field data according to proper orthogonal decomposition to determine low-order mode feature data.

[0020] In one embodiment, the target feature data includes flow vorticity feature information in the radial double V-groove region. Inputting the target feature data into a flame stability prediction model, and determining the flame stability according to the output of the flame stability prediction model includes:

[0021] Obtain the flow-vortex feature information in the target feature data; input the flow-vortex feature information into the flame stability prediction model, and determine the flame stability according to the output of the flame stability prediction model.

[0022] In a second aspect, the present application also provides a radial double-V groove flow-vortex control flame stabilization device for a jet propulsion device, including:

[0023] An acquisition module, configured to acquire the flow field data of the afterburner and the nozzle;

[0024] A dimensionality reduction module, configured to perform dimensionality reduction processing on the flow field data to determine low-order modal feature data;

[0025] A first determination module, configured to determine the target feature data from the low-order modal features according to a preset optimization method, and the preset optimization method is implemented based on an adaptive multi-response surface model and a non-dominated sorting genetic algorithm II;

[0026] A second determination module, configured to input the target feature data into the flame stability prediction model, and determine the flame stability according to the output of the flame stability prediction model.

[0027] In one embodiment, the acquisition module is specifically configured to acquire the geometric models of the afterburner and the nozzle; determine the flow field data of the afterburner and the nozzle according to a preset aerodynamic-thermal coupling calculation method and the geometric models; wherein, the preset aerodynamic-thermal coupling calculation method is an aerodynamic-thermal coupling calculation method corrected according to an experimental method, and the experimental method includes particle image velocimetry and thermocouple temperature measurement.

[0028] In one embodiment, the low-order modal feature data includes low-order modal basis coefficients and low-order modal basis vectors. The first determination module is specifically configured to determine the predicted value corresponding to the low-order modal basis coefficient according to the low-order modal basis coefficient and the preset optimization method, and the predicted value corresponding to the low-order modal basis coefficient is a value within a preset numerical range; determine the target feature data according to the product of the predicted value of the low-order modal basis coefficient and the low-order modal basis vector.

[0029] In one embodiment, the genetic algorithm is a non-dominated sorting genetic algorithm II. The first determination module is specifically configured to determine the mapping relationship between the low-order modal basis coefficient and the optimization objective according to the adaptive multi-response surface model, and the optimization objective includes the flow field data; use the low-order modal basis coefficient as the input variable, and determine the predicted value corresponding to the low-order modal basis coefficient according to the non-dominated sorting genetic algorithm II and the mapping relationship.

[0030] In one embodiment, the first determination module is specifically configured to sample the low-order mode basis coefficients according to a preset sampling method to determine sample points, where the sampling method includes the Latin hypercube sampling method, the central composite sampling method, or the Box-Behnken sampling method; update the sample points according to the distribution of the sample points and the adaptive grid refinement technique; determine the mapping relationship between the updated sample points and the corresponding optimization objectives according to the adaptive multi-response surface model; determine the accuracy of the mapping relationship, and when the accuracy does not meet the preset requirements, continue to sample the low-order mode basis coefficients according to the distribution of the sample points and the adaptive grid refinement technique until the accuracy of the mapping relationship between the low-order mode basis coefficients and the corresponding optimization objectives meets the preset requirements.

[0031] In one embodiment, the dimension reduction module is configured to perform dimension reduction processing on the flow field data to determine low-order mode feature data, including: performing dimension reduction processing on the flow field data according to the proper orthogonal decomposition to determine the low-order mode feature data.

[0032] In one embodiment, the second determination module is specifically configured to obtain the streamwise vortex feature information in the target feature data; input the streamwise vortex feature information into the flame stability prediction model, and determine the flame stability according to the output of the flame stability prediction model.

[0033] In a third aspect, the present application further provides a computer device, including a memory and a processor, where the memory stores a computer program, and when the processor executes the computer program, the method described in any one of the first aspects above is implemented.

[0034] In a fourth aspect, the present application further provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the method described in any one of the first aspects above is implemented.

[0035] In a fifth aspect, the present application further provides a computer program product, including a computer program, and when the computer program is executed by a processor, the method described in any one of the first aspects above is implemented.

[0036] The radial double-V groove flow vorticity control flame stabilization method for the above jet propulsion device obtains the flow field data of the afterburner and the nozzle, then performs dimensionality reduction processing on the flow field data to determine the low-order modal characteristic data, and then determines the target characteristic data from the low-order modal characteristics according to a preset optimization method. The preset optimization method is implemented based on an adaptive multi-response surface model and a genetic algorithm. Next, the target characteristic data is input into the flame stability prediction model, and the flame stability is determined according to the output of the flame stability prediction model. In this way, when determining the flame stability based on the flow field data of the afterburner and the nozzle, the flow field data can be subjected to dimensionality reduction processing and optimization to obtain the target characteristic data, which can optimize the low-order modal characteristic data, reduce the error caused by dimensionality reduction processing, and based on the target characteristic data and the preset flame stability prediction model, an accurate prediction result can be obtained. Based on this, the flame stability of the jet propulsion device under different working conditions can be accurately determined. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the following will briefly introduce the drawings required for the description in the embodiments of the present application or related technologies. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other related drawings can be obtained based on these drawings.

[0038] Figure 1 It is a schematic flow chart of the radial double-V groove flow vorticity control flame stabilization method for the jet propulsion device in one embodiment;

[0039] Figure 2 It is a schematic flow chart of the step of obtaining the flow field data of the afterburner and the nozzle in one embodiment;

[0040] Figure 3 It is a schematic flow chart of the step of determining the target characteristic data in one embodiment;

[0041] Figure 4 It is a schematic flow chart of the process of determining the predicted value corresponding to the low-order modal basis coefficient in one embodiment;

[0042] Figure 5 It is a schematic flow chart of the step of determining the mapping relationship between the low-order modal basis coefficient and the optimization target in one embodiment;

[0043] Figure 6 It is a schematic flow chart of the step of determining the flame stability in one embodiment;

[0044] Figure 7 It is a schematic flow chart of the radial double-V groove flow vorticity control flame stabilization method for the jet propulsion device in another embodiment;

[0045] Figure 8 A structural block diagram of a radial double V-groove flow vortex control flame stabilization device of a jet propulsion device in one embodiment;

[0046] Figure 9 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0047] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0048] As a key component of military aviation engines, the afterburner can significantly increase thrust in a short period of time, supporting fighters to achieve short-range takeoff, air maneuvers, and supersonic cruise. However, due to its extremely harsh working environment (such as ultra-high temperature, low oxygen content, large disturbances, strong oscillations, etc.), ensuring the stability of the combustion flame has become the main challenge to improving performance. Once the flame is extinguished, not only will the engine be unable to continue to provide additional thrust, it may also cause takeoff interruptions, air maneuver failures, and even serious flight safety accidents. Therefore, enhancing the reliability of the afterburner flame stability under extreme conditions is crucial to improving the overall performance of aircraft engines.

[0049] The flame stability of the traditional radial double V-groove flame stabilizer is affected by many factors, including oil-gas ratio, electrode position, parasitic capacitance and installation size. In addition, factors such as severe fluctuations in aerodynamic heat, structural deformation and electrical system abnormalities caused by sudden changes in engine operating conditions will also affect its stability, resulting in frequent failures in starting and connection. Such failures not only increase maintenance costs and time consumption, but more importantly, reduce the equipment's attendance rate.

[0050] In view of this, the present application provides a method for regulating flame stabilization of a jet propulsion device with radial double V-groove flow vortexes, which can accurately determine flame stability. The method for regulating flame stabilization of a jet propulsion device with radial double V-groove flow vortexes provided in the embodiment of the present application, the execution subject of which can be a radial double V-groove flow vortex regulating flame stabilization device of the jet propulsion device, the radial double V-groove flow vortex regulating flame stabilization device of the jet propulsion device can be implemented by software, hardware, or a combination of software and hardware, and can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in the computer device in the form of software. In the following method embodiments, the execution subject is taken as an example of a computer device, wherein the computer device can be a server or a desktop computer, and the embodiments of the present application do not limit the specific type of the computer device.

[0051] In an exemplary embodiment,Figure 1 As shown, a method for regulating the flame stability of the radial double V-groove flow in a jet propulsion device is provided, including the following steps 101 to 104. Among them:

[0052] Step 101, obtain the flow field data of the afterburner and the nozzle.

[0053] Optionally, the flow field data can be a set of various parameters describing the flow state of the fluid in space and time during flame combustion, and the flow field data can be used to characterize phenomena such as the motion characteristics, energy conversion, and heat transfer of the fluid.

[0054] Optionally, the flow field data can include velocity field data, pressure field data, temperature field data, turbulence parameters, etc., which are not limited in the embodiments of the present application.

[0055] Optionally, obtaining the flow field data of the afterburner and the nozzle can be to obtain the pre-stored flow field data from a database or a storage medium, or the flow field data obtained by solving according to the numerical simulation method. Or, the flow field data can also be determined through machine learning or a neural network model.

[0056] Exemplarily, the pre-stored flow field data can be obtained by experimental measurement. For example, the velocity field of the flow field can be calculated by photographing the movement of tiny particles in the flow field.

[0057] It can be understood that the determined flow field data is the flow field characteristics and flow behavior of the radial double V-groove region in the afterburner and the nozzle under the afterburning state.

[0058] Step 102, perform dimensionality reduction processing on the flow field data to determine the low-order modal feature data.

[0059] Optionally, after obtaining the flow field data, dimensionality reduction processing can be performed to reduce noise and redundant information. Compressing high-dimensional data to a lower dimension can improve the subsequent calculation efficiency. At the same time, after mapping the high-dimensional data to a low-dimensional space, the flow field data is easier to visualize and understand.

[0060] In a possible implementation manner, perform dimensionality reduction processing on the flow field data according to the proper orthogonal decomposition to determine the low-order modal feature data.

[0061] Optionally, when performing the proper orthogonal decomposition (POD), the flow field data can be preprocessed first, such as data normalization processing, noise removal, and data alignment, etc., to ensure that the format and quality of the flow field data are suitable for subsequent analysis.

[0062] Optionally, when performing proper orthogonal decomposition, a data matrix is constructed based on the flow field data; and the covariance matrix of the data matrix is calculated, where the covariance matrix is used to characterize the correlation between the features in the flow field data; then the covariance matrix is eigen-decomposed to solve its eigenvalues and eigenvectors, where the eigenvalues represent the variance magnitudes corresponding to each eigenvector, and the eigenvectors represent the main directions of the data. The larger the eigenvalue, the more important the corresponding eigenvector; Next, according to the magnitudes of the eigenvalues, the first k largest eigenvalues and their corresponding eigenvectors are selected. These eigenvectors are called proper orthogonal bases (POD bases), which form the basis of the low-dimensional space. Usually, the selected k is much smaller than the dimension m of the flow field data; Finally, the flow field data is projected into the low-dimensional space formed by the proper orthogonal bases, and the reduced-dimension data matrix can be obtained. Among them, the projection process can be realized through matrix multiplication.

[0063] Among them, the low-order mode feature data can include eigenvalues (basis coefficients) and corresponding eigenvectors (basis vectors).

[0064] Optionally, truncating high-order modes according to the POD method can retain key features while reducing the data dimension.

[0065] In another possible implementation, the flow field data can also be reduced in dimension according to singular value decomposition, and important components are selected according to certain metrics (such as the magnitudes of singular values), and secondary components are discarded, so as to achieve the purpose of dimension reduction.

[0066] Step 103, determining target feature data from the low-order mode features according to a preset optimization method, where the preset optimization method is implemented according to an adaptive multi-response surface model and a genetic algorithm.

[0067] Optionally, there are differences between the low-order mode feature data obtained after dimension reduction and the acquired flow field data, which may bring errors in subsequent calculations. Therefore, the basis coefficients can be adjusted according to a preset optimization algorithm to reduce the errors caused by the dimension reduction process.

[0068] Optionally, there are multiple groups of low-order mode features. One group of low-order mode features can include a basis coefficient and a corresponding basis vector.

[0069] Optionally, an adaptive multi-response surface model (AMRSM) can be used to characterize the mapping relationship between low-order mode features and the optimization target, and the target feature data can be determined from the low-order mode features according to the genetic algorithm and this mapping relationship.

[0070] Optionally, the target feature data can be the reduced-order feature data of the flow field data.

[0071] Optionally, the adaptive multi-response surface model can be trained through multiple response surface models and combined with a cross-validation strategy to ensure the prediction accuracy and generalization ability of the adaptive multi-response surface model. In the embodiments of the present application, the adaptive multi-response model can be pre-trained or obtained by training based on low-order modal features and corresponding target values. Among them, the target value can be determined according to the flow field data. The embodiments of the present application do not limit the establishment process of the adaptive multi-response surface model.

[0072] In a possible implementation manner, a preset adaptive multi-response surface model can be obtained, and then after encoding, initializing the population, fitness evaluation, selection operation, crossover operation, mutation operation, and iterative update according to the genetic algorithm, the individual with the highest fitness in the population is output as the optimal solution, and the target feature data is determined according to the optimal solution. Among them, the individual can be substituted into the adaptive multi-response surface model, and the individual fitness is determined according to the output of the adaptive multi-response surface model.

[0073] In another possible implementation manner, the adaptive multi-response surface model can be obtained by first training based on low-order modal features and corresponding target values. The process of determining the target feature data from the low-order modal features according to the genetic algorithm and the mapping relationship is the same as the process in the above implementation manner and will not be elaborated here.

[0074] Step 104: Input the target feature data into the flame stability prediction model, and determine the flame stability according to the output of the flame stability prediction model.

[0075] Optionally, the flame stability prediction model can be trained based on a machine learning method. During the training process, multi-condition combustion data and corresponding stability data are used as training parameters, and then the initial flame stability prediction model is trained by extracting features. Through cross-validation and optimizing the model parameters, when the prediction accuracy meets the preset requirements, the flame stability prediction model is obtained to ensure the model accuracy and generalization ability, so as to improve the accuracy and reliability of the flame stability prediction.

[0076] Exemplarily, the flame stability prediction model can be a classification model, which is trained based on a decision tree method, a random forest method, or a neural network.

[0077] Optionally, when the determined flame stability is poor, the geometric parameters of the radial double V-grooves can be optimized to improve the flame stability in the afterburner.

[0078] The above-mentioned method for regulating flame stability by radial double-V groove flow vortices in a jet propulsion device obtains the flow field data of the afterburner and the nozzle, then reduces the dimension of the flow field data to determine the low-order mode characteristic data, and then determines the target characteristic data from the low-order mode characteristics according to a preset optimization method. The preset optimization method is implemented based on an adaptive multi-response surface model and a genetic algorithm. Then, the target characteristic data is input into the flame stability prediction model, and the flame stability is determined according to the output of the flame stability prediction model. In this way, when determining the flame stability based on the flow field data of the afterburner and the nozzle, the flow field data can be dimensionally reduced and optimized to obtain the target characteristic data, which can optimize the low-order mode characteristic data, reduce the error caused by dimensional reduction, and based on the target characteristic data and the preset flame stability prediction model, an accurate prediction result can be obtained. Based on this, the flame stability of the jet propulsion device under different working conditions can be accurately determined.

[0079] In an exemplary embodiment, as Figure 2 shown, optionally, obtaining the flow field data of the afterburner and the nozzle includes the following steps 201 to 202. Among them:

[0080] Step 201, obtain the geometric models of the afterburner and the nozzle.

[0081] In a possible implementation manner, the geometric models of the afterburner and the nozzle can be pre-stored in a storage medium, and the identification information of the geometric model input by the user can be queried and obtained from the storage medium.

[0082] In another possible implementation manner, the key dimensions and shape parameters of the afterburner and the nozzle can also be determined by a parametric design method. By calling the relevant interfaces of the modeling software, a geometric model is generated according to the key dimensions and shape parameters. When generating the geometric model, Boolean operations are used to achieve the precise combination between components, and the generated geometric model is inspected and optimized to obtain the geometric models of the afterburner and the nozzle.

[0083] Step 202, determine the flow field data of the afterburner and the nozzle according to the preset aerodynamic-thermal coupling calculation method and the geometric model.

[0084] Among them, the preset aerodynamic-thermal coupling calculation method is an aerodynamic-thermal coupling calculation method corrected according to the experimental method. The experimental methods include particle image velocimetry and thermocouple temperature measurement.

[0085] Optionally, the aerodynamic-thermal coupling calculation method can be implemented based on the Navier-Stokes equations. The accuracy of flow behavior prediction can be improved by setting the turbulent intensity distribution, the simulation accuracy of temperature can be enhanced by adjusting parameters such as the flame propagation speed, the convergence can be accelerated and the solution process can be stabilized by adjusting the relaxation factor, and the iterative convergence criterion can be optimized by adjusting the physical quantity change threshold. The specific expression of the Navier-Stokes equations is as follows:

[0086]

[0087]

[0088]

[0089] where is the fluid density, t is the time, which is used to measure the sequence and duration of physical processes, is the Hamiltonian operator, is the fluid velocity vector, p is the fluid pressure, is the dynamic viscosity, is the second viscosity coefficient, is the strain rate tensor, is the external force vector acting on the fluid per unit mass, h is the specific enthalpy, is the dissipation function, is the internal heat source intensity per unit mass of the fluid, is the heat flux vector, is the term related to dissipation, is the thermal conductivity, and T represents the fluid temperature.

[0090] Optionally, the numerical values of key physical parameters such as velocity and temperature in the radial double-V groove region can be obtained through experimental methods, and the aerodynamic-thermal coupling calculation method can be corrected according to the numerical values of the key physical parameters to improve the calculation accuracy of the aerodynamic-thermal coupling calculation method.

[0091] Exemplarily, the velocity field in the radial double-V groove region can be measured by Particle Image Velocimetry (PIV), the temperature distribution can be obtained by thermocouples, the experimental data of the flow field and temperature field in the radial double-V groove region can be compared with the calculation data obtained by the aerodynamic-thermal coupling calculation method to identify the deviation, and the turbulent parameters and combustion performance influence parameters in the aerodynamic-thermal coupling calculation method can be adjusted based on the experimental data to correct the aerodynamic-thermal coupling calculation method. When comparing and identifying the deviation, the Root Mean Square Error (RMSE) is used for evaluation, where:

[0092]

[0093] Among them, is the experimental value, is the calculated value, and n is the number of samples.

[0094] The geometric models of the afterburner and the nozzle are obtained above. According to the preset aerodynamic-thermal coupling calculation method and the geometric models, the flow field data of the afterburner and the nozzle are determined. Among them, the preset aerodynamic-thermal coupling calculation method is the aerodynamic-thermal coupling calculation method corrected according to the experimental method. The experimental method includes particle image velocimetry and thermocouple temperature measurement. In this way, according to the corrected aerodynamic-thermal coupling calculation method and the obtained geometric models, the flow field data of the radial double V-groove region can be accurately calculated, so as to accurately determine the flame stability.

[0095] In an exemplary embodiment, as Figure 3 shown, optionally, the low-order mode feature data includes low-order mode basis coefficients and low-order mode basis vectors. Determining the target feature data from the low-order mode features according to the preset optimization method includes the following steps 301 to 302. Among them:

[0096] Step 301, determine the predicted value corresponding to the low-order mode basis coefficient according to the low-order mode basis coefficient and the preset optimization method.

[0097] Among them, the predicted value corresponding to the low-order mode basis coefficient is a value within the preset numerical range.

[0098] Exemplarily, taking the low-order mode basis coefficient as k for example, for each mode basis coefficient, a value can be determined from the preset numerical range according to the preset optimization method as the predicted value of the low-order mode basis coefficient.

[0099] It can be understood that a predicted value can be determined from the preset numerical range according to the genetic algorithm. This predicted value is the optimal solution, that is, the error between the data reconstructed according to this predicted value and the low-order mode basis vectors and the flow field data is the smallest.

[0100] Optionally, the genetic algorithm can be selected according to the number of optimization objectives. For example, when the optimization objective is one, the standard genetic algorithm can be selected. In multi-objective optimization problems, there may be conflicts between the objectives, and a multi-objective genetic algorithm can be used, such as the multi-objective evolutionary algorithm based on decomposition (Multi-Objective Evolutionary Algorithm based on Decomposition, MOEA / D), etc.

[0101] In a possible implementation manner, the genetic algorithm is the Non-dominated Sorting Genetic Algorithm II (NSGA-II), asFigure 4 As shown, determining the predicted value corresponding to the low-order modal basis coefficient according to the low-order modal basis coefficient and the preset optimization method includes the following steps 401 to 402. Wherein:

[0102] Step 401, determine the mapping relationship between the low-order modal basis coefficient and the optimization objective according to the adaptive multi-response surface model.

[0103] Among them, the optimization objective includes the flow field data, which may include the position, intensity, and evolution information of the streamwise vortices in the flow field, etc.

[0104] Optionally, the adaptive multi-response surface model AMRSM can sample the input data, train different response surface models, and randomly select a certain proportion of samples as the validation set to test the prediction errors of each model. Subsequently, for different optimization objectives, select the model with the smallest average relative error for the final response surface construction.

[0105] Optionally, AMRSM can be selected but is not limited to Gaussian process response surface, neural network response surface, Kriging response surface, radial basis function response surface, etc.

[0106] Optionally, the mapping relationship between the low-order modal basis coefficient and the optimization objective can be determined through AMRSM.

[0107] Step 402, take the low-order modal basis coefficient as the input variable, and determine the predicted value corresponding to the low-order modal basis coefficient according to the non-dominated sorting genetic algorithm II and the mapping relationship.

[0108] Optionally, for a low-order modal basis coefficient, determining the predicted value corresponding to the low-order modal basis coefficient according to the non-dominated sorting genetic algorithm II and the mapping relationship may include the following steps:

[0109] The first step is to initialize the population. A random initial population P can be generated within the range of this low-order modal basis coefficient, and the population size is N;

[0110] The second step is to bring each individual into the mapping relationship and calculate the fitness of each individual;

[0111] The third step is to divide the population into different non-dominated levels. The individuals in the first level are not dominated by other individuals, and the subsequent level individuals are sequentially dominated by the individuals in the previous level;

[0112] The fourth step is to calculate the crowding distance. For each individual within each non-dominated level, calculate its crowding distance to measure the density of the individual in the objective space;

[0113] Step 5: Selection operation. Use methods such as binary tournament selection to select parental individuals for crossover. Compare the non-dominated levels of individuals. Those with lower levels are preferred; if the levels are the same, compare the crowding distances, and those with larger distances are preferred.

[0114] Step 6: Crossover and mutation. Perform crossover operations (exchanging part of the chromosomes to generate new individuals) and mutation operations (introducing small random changes in the chromosomes) on the selected parental individuals.

[0115] Step 7: Generate the next generation. Combine the parental and offspring populations to form an intermediate population, perform non-dominated sorting on it, and select excellent individuals based on the non-dominated rank and crowding distance to form the next generation population.

[0116] Step 8: Termination judgment. Continuously repeat the processes of selection, crossover, mutation, and population update until the termination conditions are met (such as reaching the specified number of generations, convergence threshold, or exhausting computing resources, etc.). The final population contains the Pareto optimal solutions representing the optimal trade-off among the objectives.

[0117] Next, the concepts involved in the above steps are explained:

[0118] Non-dominated sorting is a technique for dividing individuals in a population into different levels (Pareto front levels). If there are no other individuals in the population that are superior to individual A in all objectives, then individual A is a non-dominated individual. Classify the non-dominated individuals into the first-level front; after removing the first-level front individuals, find the non-dominated individuals among the remaining individuals and classify them into the second-level front, and so on. This sorting can identify and retain high-quality individuals and is crucial for multi-objective optimization.

[0119] Crowding distance can be used to measure the distribution density of individuals within the Pareto front to maintain population diversity. When calculating, first sort the individuals under each objective. The crowding distance of the boundary individuals is set to infinity, and the crowding distance of the intermediate individuals is calculated through the normalized difference of the objective values of adjacent individuals. Individuals with larger crowding distances are more favored in selection, which can make the individuals evenly distributed in the objective space and avoid aggregation.

[0120] In step 302, determine the target feature data according to the product of the predicted value of the low-order modal basis coefficient and the low-order modal basis vector.

[0121] Optionally, for a low-order modal basis coefficient, multiply the predicted value of the low-order modal basis coefficient by the corresponding low-order modal basis vector, and use the product as the target feature data.

[0122] The predicted values corresponding to the low-order modal basis coefficients are determined based on the low-order modal basis coefficients and the preset optimization method. The target feature data is determined according to the product of the predicted values of the low-order modal basis coefficients and the low-order modal basis vectors. By processing the flow field data, the target feature data is obtained. In this way, by adjusting the low-order modal basis coefficients, the error between the obtained target feature data and the flow field data is reduced, that is, the target feature data can accurately represent the flow field data. Thus, the flame stability performance determined according to the target feature data is more accurate. At the same time, since the prediction of flame stability involves the rated interaction of multiple physical quantities, these variables can be comprehensively considered through the preset optimization method, and the influence of each variable on flame stability can be fully reflected, which can effectively improve the accuracy of flame stability prediction.

[0123] In an exemplary embodiment, as Figure 5 shown, optionally, according to the adaptive multi-response surface model, the mapping relationship between the low-order modal basis coefficients and the optimization target is determined, including the following steps 501 to 504. Among them:

[0124] Step 501: Sample the low-order modal basis coefficients according to the preset sampling method to determine the sample points.

[0125] Among them, the sampling method includes the Latin hypercube sampling method, the central composite sampling method, or the Box-Behnken sampling method.

[0126] Optionally, for a low-order modal basis coefficient, sampling can be performed within a preset numerical range according to the preset sampling method to obtain sample points.

[0127] Exemplarily, the Latin hypercube sampling method can be to divide the preset numerical range into several intervals with equal probabilities, and then randomly select a sample point within each interval, and it is necessary to ensure that each hyperplane perpendicular to the axis contains at most one sample point.

[0128] Step 502: Update the sample points according to the distribution of the sample points and the adaptive grid refinement technique.

[0129] Optionally, after sampling the low-order modal basis coefficients to determine the sample points, the adaptive grid refinement technique can dynamically refine the sampling interval for the key area according to the distribution of the sample points and increase the sample points.

[0130] Optionally, when the accuracy of the mapping relationship does not meet the preset requirements, the sampling interval can also be dynamically refined in the area with a large difference from the actual value to increase the sample points.

[0131] Step 503: Determine the mapping relationship between the updated sample points and the corresponding optimization target according to the adaptive multi-response surface model.

[0132] Optionally, after determining the sample points, the flow field data can be used to determine the target value corresponding to the sample points, and this target value is the actual target value.

[0133] Exemplarily, taking the response surface function as a quadratic polynomial response surface function as an example, with the input being the sample points and the output being the target values, the coefficients in the response surface function can be estimated by methods such as the least squares method, so that the response surface function can fit the sample points and the target values as much as possible.

[0134] Step 504: Determine the accuracy of the mapping relationship. When the accuracy does not meet the preset requirements, according to the distribution of the sample points and the adaptive grid refinement technology, continue to sample the low-order modal basis coefficients until the accuracy of the mapping relationship between the low-order modal basis coefficients and the corresponding optimization objectives meets the preset requirements.

[0135] Optionally, when determining the mapping relationship between the updated sample points and the corresponding optimization objectives, the updated sample points can be divided into two groups according to a preset ratio, one group is the training samples and the other group is the test samples. After determining the mapping relationship based on the training samples, the mapping relationship can be tested according to the test samples to determine the accuracy of the mapping relationship.

[0136] Exemplarily, for each test sample point in the test samples, the error between the actual target value and the predicted target value determined by the adaptive multi-response surface model can be calculated. Commonly used error metrics include the mean square error (MSE), the mean absolute error (MAE), etc. When the error exceeds the preset error threshold, it can be considered that the accuracy of the mapping relationship does not meet the preset requirements.

[0137] Optionally, new sampling points can also be obtained, and whether the accuracy of the mapping relationship meets the preset requirements can be determined according to the error between the actual target value and the predicted target value of the new sampling points.

[0138] Optionally, when the accuracy meets the preset requirements, the above steps 502 to 504 can be repeated until a mapping relationship that meets the preset requirements is obtained.

[0139] The low-order modal basis coefficients are sampled according to the preset sampling method to determine sample points. Based on the distribution of the sample points and the adaptive grid refinement technique, the sample points are updated. According to the adaptive multi-response surface model, the mapping relationship between the updated sample points and the corresponding optimization objectives is determined, and the accuracy of the mapping relationship is determined. When the accuracy does not meet the preset requirements, the low-order modal basis coefficients are continuously sampled according to the distribution of the sample points and the adaptive grid refinement technique until the accuracy of the mapping relationship between the low-order modal basis coefficients and the corresponding optimization objectives meets the preset requirements. An approximate mapping relationship of the complex flame physical process can be constructed based on less data. By reasonably selecting sample points and using the response surface function to approximate the mapping relationship, the mapping efficiency can be effectively improved. At the same time, the mapping relationship can be further refined by the distribution of the sample points and the adaptive grid refinement technique to obtain a mapping relationship that can accurately describe the low-order modal basis coefficients and the corresponding optimization objectives under different working conditions.

[0140] In an exemplary embodiment, as Figure 6 shown, optionally, the target feature data includes the flow vortex feature information of the radial double V-groove region. The target feature data is input into the flame stability prediction model, and the flame stability is determined according to the output of the flame stability prediction model, including the following steps 601 to step 602. Wherein:

[0141] Step 601, obtain the flow vortex feature information in the target feature data.

[0142] Optionally, the target feature data includes the flow vortex feature information of the radial double V-groove region. Therefore, after determining the target feature data, the flow vortex feature information can be extracted from the target feature data.

[0143] Step 602, input the flow vortex feature information into the flame stability prediction model, and determine the flame stability according to the output of the flame stability prediction model.

[0144] Optionally, the flow vortex feature information can be input into a pre-trained flame stability prediction model. The output result of the flame stability prediction model can include flame stable or flame unstable, and can also include the stability level of the flame stability. The flame stability is determined according to the stability level of the flame stability.

[0145] Exemplarily, when the output result of the flame stability prediction model includes the stability level of the flame stability, the higher the stability level value, the more stable the predicted flame stability.

[0146] The above-mentioned method for obtaining the flow-vortex characteristic information in the target characteristic data, inputting the flow-vortex characteristic information into the flame stability prediction model, and determining the flame stability according to the output of the flame stability prediction model can determine the flame stability based on the flow-vortex characteristic information in the target data obtained after dimensionality reduction processing, adaptive multi-response surface model, and non-dominated sorting genetic algorithm II processing, which can effectively improve the determination accuracy.

[0147] As an alternative implementation, as Figure 7 shown, the method for regulating the flame stability by the radial double-V groove flow-vortex in the jet propulsion device provided in the embodiment of the present application may include the following specific steps:

[0148] Step 701, obtain the geometric models of the afterburner and the nozzle.

[0149] Step 702, determine the flow field data of the afterburner and the nozzle according to the preset aerodynamic-thermal coupling calculation method and the geometric models.

[0150] Among them, the preset aerodynamic-thermal coupling calculation method is the aerodynamic-thermal coupling calculation method corrected according to the experimental method, and the experimental method includes particle image velocimetry and thermocouple temperature measurement.

[0151] Step 703, perform dimensionality reduction processing on the flow field data according to proper orthogonal decomposition to determine the low-order modal characteristic data.

[0152] Among them, the low-order modal characteristic data includes low-order modal basis coefficients and low-order modal basis vectors.

[0153] Step 704, sample the low-order modal basis coefficients according to the preset sampling method to determine the first sample points.

[0154] Among them, the sampling method includes Latin hypercube sampling method, central composite sampling method or Box-Behnken sampling method.

[0155] Step 705, determine the sample points according to the distribution of the first sample points and the adaptive grid refinement technology.

[0156] Step 706, determine the mapping relationship between the sample points and the corresponding optimization objectives according to the adaptive multi-response surface model.

[0157] Step 707, determine whether the accuracy of the mapping relationship meets the preset requirements. When the accuracy of the mapping relationship does not meet the preset requirements, execute Step 705.

[0158] Step 708, when the accuracy meets the preset requirements, use the low-order modal basis coefficients as input variables, and determine the predicted values corresponding to the low-order modal basis coefficients according to the non-dominated sorting genetic algorithm II and the mapping relationship.

[0159] Among them, the predicted value corresponding to the low-order modal basis coefficient is a value within a preset numerical range.

[0160] Step 709: Determine the target feature data according to the product of the predicted value of the low-order modal basis coefficient and the low-order modal basis vector.

[0161] Step 710: Obtain the flow-vortex feature information in the target feature data.

[0162] Step 711: Input the flow-vortex feature information into the flame stability prediction model, and determine the flame stability according to the output of the flame stability prediction model.

[0163] It should be understood that although the steps in the flowcharts involved in the above-described embodiments are displayed in sequence according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise clearly stated in this article, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-described embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be executed alternately or in turn with at least a part of other steps or steps or stages in other steps.

[0164] Based on the same inventive concept, an embodiment of the present application further provides a radial double-V groove flow-vortex regulation flame stabilization device for a jet propulsion device for implementing the above-mentioned radial double-V groove flow-vortex regulation flame stabilization method for a jet propulsion device. The solution provided by this device to solve the problem is similar to the solution described in the above method. Therefore, the specific limitations in one or more embodiments of the following radial double-V groove flow-vortex regulation flame stabilization device for a jet propulsion device can refer to the limitations on the radial double-V groove flow-vortex regulation flame stabilization device method for a jet propulsion device in the above text, and will not be repeated here.

[0165] In an exemplary embodiment, as Figure 8 shown, a radial double-V groove flow-vortex regulation flame stabilization device 800 for a jet propulsion device is provided, including: an acquisition module 801, a dimensionality reduction module 802, a first determination module 803, and a second determination module 804, where:

[0166] The acquisition module 801 is configured to acquire the flow field data of the afterburner and the nozzle;

[0167] The dimensionality reduction module 802 is configured to perform dimensionality reduction processing on the flow field data to determine the low-order modal feature data;

[0168] The first determination module 803 is configured to determine target feature data from low-order modal features according to a preset optimization method, where the preset optimization method is implemented based on an adaptive multi-response surface model and a non-dominated sorting genetic algorithm II;

[0169] The second determination module 804 is configured to input the target feature data into a flame stability prediction model and determine the flame stability according to the output of the flame stability prediction model.

[0170] In one embodiment, the acquisition module 801 is specifically configured to acquire the geometric models of the afterburner and the nozzle; determine the flow field data of the afterburner and the nozzle according to a preset aerodynamic-thermal coupling calculation method and the geometric models; where the preset aerodynamic-thermal coupling calculation method is an aerodynamic-thermal coupling calculation method corrected according to an experimental method, and the experimental method includes particle image velocimetry and thermocouple temperature measurement.

[0171] In one embodiment, the low-order modal feature data includes low-order modal basis coefficients and low-order modal basis vectors. The first determination module 803 is specifically configured to determine the predicted value corresponding to the low-order modal basis coefficient according to the low-order modal basis coefficient and the preset optimization method, and the predicted value corresponding to the low-order modal basis coefficient is a value within a preset numerical range; determine the target feature data according to the product of the predicted value of the low-order modal basis coefficient and the low-order modal basis vector.

[0172] In one embodiment, the genetic algorithm is a non-dominated sorting genetic algorithm II. The first determination module 803 is specifically configured to determine the mapping relationship between the low-order modal basis coefficient and the optimization objective according to the adaptive multi-response surface model, where the optimization objective includes flow field data; use the low-order modal basis coefficient as the input variable, and determine the predicted value corresponding to the low-order modal basis coefficient according to the non-dominated sorting genetic algorithm II and the mapping relationship.

[0173] In one embodiment, the first determination module 803 is specifically configured to sample the low-order modal basis coefficient according to a preset sampling method to determine the first sample points, and the sampling method includes a Latin hypercube sampling method, a central composite sampling method, or a Box-Behnken sampling method; update the sample points according to the distribution of the sample points and the adaptive grid refinement technique; determine the mapping relationship between the updated sample points and the corresponding optimization objective according to the adaptive multi-response surface model; determine the accuracy of the mapping relationship, and when the accuracy does not meet the preset requirements, continue to sample the low-order modal basis coefficient according to the distribution of the sample points and the adaptive grid refinement technique until the accuracy of the mapping relationship between the low-order modal basis coefficient and the corresponding optimization objective meets the preset requirements.

[0174] In one embodiment, the dimensionality reduction module 802 is configured to perform dimensionality reduction processing on the flow field data to determine low-order mode feature data, including: performing dimensionality reduction processing on the flow field data according to proper orthogonal decomposition to determine low-order mode feature data.

[0175] In one embodiment, the second determination module 804 is specifically configured to obtain the streamwise vortex feature information in the target feature data; input the streamwise vortex feature information into the flame stability prediction model, and determine the flame stability according to the output of the flame stability prediction model.

[0176] Each module in the above-mentioned radial double-V groove streamwise vortex control flame stabilization device of the jet propulsion device can be implemented in whole or in part by software, hardware, and their combination. The above-mentioned modules can be embedded in the processor of the computer device in hardware form or independent of it, or stored in the memory of the computer device in software form, so that the processor can call and execute the operations corresponding to the above-mentioned modules.

[0177] In an exemplary embodiment, a computer device is provided. The computer device can be a server, and its internal structure diagram can be as Figure 9 shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, abbreviated as I / O), and a communication interface. Among them, the processor, the memory, and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store data. The input / output interface of the computer device is used to exchange information between the processor and external devices. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, it implements a method for controlling the flame stability of a radial double-V groove streamwise vortex of a jet propulsion device.

[0178] Those skilled in the art can understand that Figure 9 the structure shown in

[0179] In an exemplary embodiment, a computer device is provided, including a memory and a processor. A computer program is stored in the memory, and when the processor executes the computer program, the steps described in any of the above method embodiments are implemented.

[0180] In an embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps described in any of the above method embodiments are implemented.

[0181] In an embodiment, a computer program product is provided, including a computer program. When the computer program is executed by a processor, the steps described in any of the above method embodiments are implemented.

[0182] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, database, or other medium used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The databases involved in the embodiments provided in this application can include at least one of relational databases and non-relational databases. Non-relational databases can include distributed databases based on blockchain, etc., without limitation. The processors involved in the embodiments provided in this application can be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, data processing logics based on quantum computing, artificial intelligence (AI) processors, etc., without limitation.

[0183] The technical features of the above embodiments can be combined arbitrarily. For the sake of concise description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope recorded in this application.

[0184] The above-described embodiments merely represent several implementation manners of the present application. The description thereof is relatively specific and detailed, but it should not be construed as a limitation on the patent scope of the present application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all fall within the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the appended claims.

Claims

1. A method for controlling flame stabilization by using radial double V-groove flow vortexes for jet propulsion, characterized in that: The method comprises: Obtain flow field data of afterburner and nozzle; Performing dimensionality reduction processing on the flow field data to determine low-order modal characteristic data; Determining target feature data from the low-order modal features according to a preset optimization method, wherein the preset optimization method is implemented according to an adaptive multi-response surface model and a genetic algorithm; The target characteristic data is input into a flame stability prediction model, and the flame stability is determined according to an output of the flame stability prediction model.

2. The method according to claim 1, characterized in that The step of obtaining flow field data of the afterburner chamber and the nozzle comprises: Obtaining a geometric model of the afterburner and the nozzle; Determining flow field data of the afterburner and the nozzle according to a preset aerodynamic thermal coupling calculation method and the geometric model; The preset aerodynamic thermal coupling calculation method is a modified aerodynamic thermal coupling calculation method according to an experimental method, wherein the experimental method includes particle image velocimetry and thermocouple temperature measurement.

3. The method according to claim 1, characterized in that The low-order modal feature data includes low-order modal basis coefficients and low-order modal basis vectors, and determining target feature data from the low-order modal features according to a preset optimization method includes: Determining a predicted value corresponding to the low-order modal base coefficient according to the low-order modal base coefficient and the preset optimization method, wherein the predicted value corresponding to the low-order modal base coefficient is a value within a preset value range; The target feature data is determined according to the product of the predicted value of the low-order modal basis coefficient and the low-order modal basis vector.

4. The method according to claim 3, characterized in that The genetic algorithm is a non-dominated sorting genetic algorithm II, and the method of determining the predicted value corresponding to the low-order modal base coefficient according to the low-order modal base coefficient and the preset optimization method includes: Determining a mapping relationship between the low-order modal base coefficients and an optimization target according to the adaptive multi-response surface model, wherein the optimization target includes the flow field data; The low-order modal base coefficients are used as input variables, and the predicted values ​​corresponding to the low-order modal base coefficients are determined according to the non-dominated sorting genetic algorithm II and the mapping relationship.

5. The method according to claim 4, characterized in that Determining the mapping relationship between the low-order modal basis coefficients and the optimization target according to the adaptive multi-response surface model includes: Sampling the low-order modal basis coefficients according to a preset sampling method to determine sample points, wherein the sampling method includes a Latin hypercube sampling method, a central combination sampling method, or a Box-Behnken sampling method; updating the sample points according to the distribution of the sample points and the adaptive grid refinement technique; Determining a mapping relationship between updated sample points and corresponding optimization objectives according to the adaptive multi-response surface model; Determine the accuracy of the mapping relationship. When the accuracy does not meet the preset requirements, continue to sample the low-order modal basis coefficients according to the distribution of the sample points and the adaptive grid refinement technology until the accuracy of the mapping relationship between the low-order modal basis coefficients and the corresponding optimization objectives meets the preset requirements.

6. The method according to claim 1, characterized in that The step of performing dimensionality reduction processing on the flow field data to determine low-order modal feature data includes: The flow field data is subjected to dimensionality reduction processing according to the intrinsic orthogonal decomposition to determine the low-order modal characteristic data.

7. The method according to claim 1, characterized in that The target characteristic data includes the flow vortex characteristic information of the radial double V-groove area, the target characteristic data is input into a flame stability prediction model, and the flame stability is determined according to the output of the flame stability prediction model, including: Acquiring stream vortex characteristic information in the target characteristic data; The streamwise vortex characteristic information is input into a flame stability prediction model, and the flame stability is determined according to an output of the flame stability prediction model.

8. A radial double V-groove flow vortex control flame stabilization device for a jet propulsion device, characterized in that: The device comprises: An acquisition module, used to acquire flow field data of the afterburner and nozzle; A dimension reduction module, used for performing dimension reduction processing on the flow field data to determine low-order modal feature data; A first determination module is used to determine target feature data from the low-order modal features according to a preset optimization method, wherein the preset optimization method is implemented according to an adaptive multi-response surface model and a non-dominated sorting genetic algorithm II; The second determination module is used to input the target characteristic data into a flame stability prediction model, and determine the flame stability according to the output of the flame stability prediction model.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

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

Images