TBCC inlet restart method and device based on adaptive optimization algorithm, computer device and medium

Abstract

Embodiments of the present application provide a TBCC inlet restart method and device based on an adaptive optimization algorithm, computer equipment and a medium, relating to the technical field of aero-engine control, wherein the method comprises the following steps: constructing an engine component-level nonlinear model, determining an input parameter vector and an output parameter vector; inputting the input parameter vector into a black box model for calculation, taking the calculation result as the fitness evaluation data of the current particle, dynamically adjusting the independent weight and independent learning rate of each input parameter vector, generating the input parameter vector of the next iteration step, repeating the iteration until the stop condition is met, and taking the input parameter vector of the last iteration step as the restart target point data; using a linear interpolation method to generate continuous intermediate state points and constructing a restart control path. Due to the adaptive weight adjustment strategy, the stability and safety of the restart process are improved.

Description

Technical Field

[0001] This invention relates to the field of aero-engine control technology, and in particular to a restart method, apparatus, computer equipment, and medium for a TBCC inlet based on an adaptive optimization algorithm. Background Technology

[0002] The strategic direction for future aerospace is to develop reusable, horizontally operated, airport-grade runway-based air-to-space transportation systems. This is the core approach to significantly reduce the cost of accessing space and achieve rapid Earth-to-ground transportation. This requires the propulsion system to operate efficiently across a wide speed range from Mach 0 to Mach 5. Turbine-based combined cycle (TBCC) engines, by physically combining mature turbine engines (excelling at Mach 0-2) with ramjet engines (excelling at Mach 2-5), are considered the most realistic and promising propulsion solution for achieving this goal at present. Their success directly impacts the feasibility of next-generation aerospace vehicles. Compared to a single ramjet engine inlet, the TBCC inlet is several orders of magnitude more complex geometrically, aerodynamically, and in terms of control, directly leading to a higher risk of "failure to restart" and more stringent "restart" requirements. Especially during ramjet operation, the hypersonic flight environment is extremely complex and harsh, making the inlet highly susceptible to "failure to restart," meaning the normal shock wave structure within the inlet is disrupted, the captured airflow drops sharply, the airflow becomes turbulent, and the total pressure recovery coefficient decreases drastically. Inlet restart refers to the process of restoring the air intake to a normal, efficient restarting state through a series of active or passive measures after it fails to restart. If it fails to restart, the aircraft will rapidly decelerate and may even lose control and crash. Therefore, without reliable restart capability, the practical application of hypersonic ramjet engines is impossible.

[0003] Traditional intake manifold restart technology reduces fuel flow to decrease engine outlet back pressure, thereby allowing the positive shock wave to be "drawn" back into the intake manifold, achieving restart. However, while this traditional method achieves intake manifold restart and is simple to implement, it leads to a reduction in engine thrust, resulting in greater danger should the intake manifold restart fail. Summary of the Invention

[0004] In view of this, embodiments of the present invention provide a restart method for a TBCC inlet based on an adaptive optimization algorithm to solve the technical problem of low stability and safety during the restart process of the inlet in the prior art. The method includes:

[0005] A component-level nonlinear model of the TBCC engine is constructed, and the component-level nonlinear model of the TBCC engine is encapsulated as a black box model. The input parameter vector and output parameter vector of the black box model are determined. The input parameter vector includes the inlet area of ​​the air intake, the fuel flow rate and the throat area of ​​the tail nozzle, and the output parameter vector includes the position of the normal shock wave and the engine thrust. Based on the adaptive optimization algorithm, the input parameter vector of the current iteration step is input into the black box model for solution, and the output parameter vector obtained by solution is used as the fitness evaluation data of the current particle. According to the contribution of each input parameter vector to the output parameter vector in the fitness evaluation data, the independent weight and independent learning rate of each input parameter vector are dynamically adjusted, and the velocity and position of the particle are updated based on the independent weight and independent learning rate to generate the input parameter vector of the next iteration step. The iterative calculation is repeated until the iteration stopping condition is met, and the input parameter vector corresponding to the last iteration step is used as the restart target point data. Using the initial running data as the starting point of the path and the restart target point data as the ending point of the path, a continuous series of intermediate state points are generated using a linear interpolation method. Based on the starting point of the path, the intermediate state points, and the ending point of the path, a restart control path for the TBCC intake is constructed.

[0006] This invention also provides a TBCC inlet restart device based on an adaptive optimization algorithm to solve the technical problem of low stability and safety during the inlet restart process in the prior art. The device includes: The model building module is used to build a component-level nonlinear model of the TBCC engine, encapsulate the component-level nonlinear model of the TBCC engine into a black box model, and determine the input parameter vector and output parameter vector of the black box model. The input parameter vector includes the inlet area of ​​the air intake, the fuel flow rate and the throat area of ​​the tail nozzle, and the output parameter vector includes the position of the normal shock wave and the engine thrust. The iterative module is used to input the input parameter vector of the current iteration step into the black box model for solution based on an adaptive optimization algorithm, and use the solution output parameter vector as the fitness evaluation data of the current particle. According to the contribution of each input parameter vector to the output parameter vector in the fitness evaluation data, the independent weight and independent learning rate of each input parameter vector are dynamically adjusted, and the velocity and position of the particle are updated based on the independent weight and independent learning rate to generate the input parameter vector of the next iteration step. The iterative calculation is repeated until the iteration stopping condition is met, and the input parameter vector corresponding to the last iteration step is used as the restart target point data. A restart control path generation module is used to take the initial running data as the path start point and the restart target point data as the path end point, and use a linear interpolation method to generate continuous intermediate state points. Based on the path start point, the intermediate state points and the path end point, a restart control path for the TBCC intake is constructed.

[0007] This invention also provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements any of the above-mentioned TBCC intake restart methods based on adaptive optimization algorithms, thereby solving the technical problem of low stability and safety in the intake restart process in the prior art.

[0008] This invention also provides a computer-readable storage medium storing a computer program that executes any of the above-described TBCC intake restart methods based on adaptive optimization algorithms, in order to solve the technical problem of low stability and safety in the intake restart process in the prior art.

[0009] Compared with the prior art, the beneficial effects that at least one technical solution adopted in the embodiments of this specification can achieve include at least: The inlet area of ​​the air intake and the throat area of ​​the tail nozzle are introduced to balance the engine thrust variation caused by changes in fuel flow. An adaptive weight adjustment strategy is introduced to set an independent learning rate and weight for each variable, which improves the safety and stability of the TBCC engine during flight. Attached Figure Description

[0010] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0011] Figure 1 This is a flowchart of a TBCC intake restart method based on an adaptive optimization algorithm provided in an embodiment of the present invention; Figure 2 This is an overall schematic diagram of the TBCC engine component level provided in an embodiment of the present invention; Figure 3 This is a flowchart of a restart method for the TBCC inlet based on the adaptive optimization algorithm provided in an embodiment of the present invention; Figure 4 This is a verification diagram of the effect of the first embodiment of the restart method provided by the present invention; Figure 5 This is a verification diagram of the effect of the second embodiment of the restart method provided in this invention; Figure 6 This is a verification diagram of the effect of the third embodiment of the restart method provided in this invention; Figure 7 This is a structural block diagram of a computer device provided in an embodiment of the present invention; Figure 8 This is a structural block diagram of a TBCC intake restart device based on an adaptive optimization algorithm provided in an embodiment of the present invention. Detailed Implementation

[0012] The embodiments of this application will now be described in detail with reference to the accompanying drawings.

[0013] The following specific examples illustrate the implementation of this application. Those skilled in the art can easily understand other advantages and effects of this application from the content disclosed in this specification. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. This application can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of this application. It should be noted that, in the absence of conflict, the following embodiments and features in the embodiments can be combined with each other. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0014] In this embodiment of the invention, a restart method for the TBCC inlet based on an adaptive optimization algorithm is provided, such as... Figure 1 As shown, the method includes: Step S101: Construct a nonlinear model at the component level of the TBCC engine, encapsulate the nonlinear model at the component level of the TBCC engine into a black box model, and determine the input parameter vector and output parameter vector of the black box model. The input parameter vector includes the inlet area of ​​the air intake, the fuel flow rate and the throat area of ​​the tail nozzle, and the output parameter vector includes the position of the normal shock wave and the engine thrust. Step S102: Based on the adaptive optimization algorithm, the input parameter vector of the current iteration step is input into the black box model for solution, and the output parameter vector obtained by solution is used as the fitness evaluation data of the current particle. According to the contribution of each input parameter vector to the output parameter vector in the fitness evaluation data, the independent weight and independent learning rate of each input parameter vector are dynamically adjusted, and the velocity and position of the particle are updated based on the independent weight and independent learning rate to generate the input parameter vector of the next iteration step. The iterative calculation is repeated until the iteration stopping condition is met, and the input parameter vector corresponding to the last iteration step is used as the restart target point data. Step S103: Using the initial running data as the starting point of the path and the restart target point data as the ending point of the path, a continuous intermediate state point is generated using a linear interpolation method. Based on the starting point of the path, the intermediate state point, and the ending point of the path, a restart control path for the TBCC intake is constructed.

[0015] In practice, the following steps are used to construct a component-level nonlinear model of the TBCC engine: Based on the geometric and aerodynamic parameter data of the TBCC engine, the first airflow parameter data after compression by three oblique shock waves is generated sequentially using the oblique shock wave compression relationship. Based on the first airflow parameter data and the geometric parameter data of the intake duct, the second airflow parameter data after compression by the normal shock wave is generated using the normal shock wave compression relationship. The first and second airflow parameter data are temporally correlated to generate a ramjet-based intake duct sub-model with a four-wave system design. Turbine-based component sub-models, ramjet combustion chamber sub-models, tail nozzle models, and modal control valve sub-models are constructed respectively. The turbine-based component sub-model, ramjet-based intake duct sub-model, ramjet combustion chamber sub-model, tail nozzle model, and modal control valve sub-model are integrated to generate the TBCC engine component-level nonlinear model, and the TBCC engine component-level nonlinear model is stored in a preset task environment.

[0016] In specific implementation, the following steps are used to input the input parameter vector of the current iteration step into the black box model for solution, and the resulting output parameter vector is used as the fitness evaluation data of the current particle: The position of the normal shock wave is output after the black box model is solved. and engine thrust ; Obtain the engine thrust when the intake is not running, as the baseline engine thrust. ; through the thrust of the engine and the reference engine thrust The relative error data of engine thrust were calculated. error,in, The position of the normal shock wave and the engine thrust relative error data error Perform data fusion and encapsulate it into fitness evaluation data for the current particle.

[0017] In specific implementation, the following steps are used to dynamically adjust the independent weights and independent learning rates of each input parameter vector based on its contribution to the output parameter vector in the fitness evaluation data, and update the particle's velocity and position based on the independent weights and independent learning rates to generate the input parameter vector for the next iteration step: Based on the engine thrust relative error data (error) and normal shock wave position data Generate independent weights for adjusting particle positions. Where d is the index of the input parameter vector, and t is the time step; based on the set base learning rate and the independent weights Generate the independent learning rate of the d-th variable for the i-th particle at time t. ,in, ,in, Let d be the independent weight of the d-th variable of the i-th particle at time t; obtain the velocity of the d-th input parameter vector of the i-th particle at time t. The position of the d-th input parameter vector of the i-th particle at time t Compare the current position of the particles. The individual's fitness and historical best fitness are compared to determine the individual's historical best position. The global optimal position is determined by comparing the individual historical best position of all particles with the global optimal position at the previous moment. A velocity update formula is constructed, and the velocity of the particle in the next iteration step is generated based on the velocity of the particle in the current iteration step using this formula. , For the first i The first particle d The input parameter vectors are in t The velocity at time +1 For the first d Variables in t Independent learning rate at any given moment For inertial weights, As a learning factor, A random number between 0 and 1 is used; a position update formula is constructed, and the position of the particle in the next iteration step is generated based on the position of the particle in the current iteration step using the position update formula, wherein, , For the first i The first particle d The input parameter vectors are in t The position at time +1 For time steps.

[0018] In practice, this is achieved through the following steps, based on the engine thrust relative error data (error) and the normal shock wave position data. Generate independent weights for adjusting particle positions. : Based on engine thrust relative error data error Generate the weight of the first particle. Based on normal shock wave location data Generate the weight of the second particle. ; obtain t Time-based trade-off coefficients ,in, ; through the aforementioned tradeoff coefficients The weight of the first particle and the weight of the second particle Calculate the independent weights ,in, .

[0019] In practice, the iteration stopping condition is met through the following steps: After each iteration, obtain the current iteration number. t The fitness of the global optimal position is determined; reaching a preset maximum iteration count threshold is used as a first termination condition; reaching a preset target fitness value is used as a second termination condition; and the fitness of the global optimal position is further determined by the number of consecutive iterations. k The third termination condition is that the change within the range is less than a preset change range. , For the preset range of change, Let be the fitness of the globally optimal position in the t-th iteration. For the first The fitness of the global optimal position after iteration; the iteration stops when any one of the first termination condition, the second termination condition, or the third termination condition is met.

[0020] In specific implementation, the following steps are used to take the initial running data as the starting point of the path, the restart target point data as the ending point of the path, and a continuous series of intermediate state points are generated using a linear interpolation method. Based on the starting point, the intermediate state points, and the ending point of the path, a restart control path for the TBCC intake is constructed: Based on the variable values ​​corresponding to the starting point and ending point of the path, the variable values ​​corresponding to each intermediate state point are calculated using a linear interpolation method. These variable values ​​include the relative values ​​of the intake manifold inlet area, the relative value of the fuel flow rate, and the relative value of the exhaust nozzle throat area. , For the first m The variable values ​​corresponding to each intermediate state point The variable value corresponding to the starting point of the path. The variable value corresponding to the endpoint of the path. M This represents the total number of intermediate state points. m From 1 to M Integer; the starting point of the path, M The intermediate state points and the path endpoints are combined in sequence to form a system containing... M +2 status points of the intake duct restart control path.

[0021] To overcome the shortcomings of existing technologies in the restart of the intake manifold of TBCC combined engines, a multivariate adaptive optimization algorithm is proposed, which extends the adjustable parameters to the intake manifold inlet area, fuel flow rate and exhaust nozzle throat area, thereby ensuring a smooth thrust transition during intake manifold restart.

[0022] To achieve the above objectives, the technical solution provided by this invention is as follows: First, a model-based task environment is constructed, namely, a TBCC engine component-level model. This task environment is treated as a black-box model, with three parameters related to the inlet area of ​​the air intake, fuel flow rate, and the throat area of ​​the exhaust nozzle as inputs, and the outputs being the position of the normal shock wave and the relative thrust error. A particle swarm optimization algorithm is used, and the weights and rates of the three independent variables are adaptively changed during the optimization process, thereby improving the efficiency of the optimization algorithm.

[0023] The concept and principle of this invention are as follows: like Figure 2 As shown, a component-level model of the TBCC combined engine is established to create the mission environment of this invention. Figure 3 As shown, this framework includes four modules: initialization module, convergence judgment module, particle swarm optimization (PSO) update module, and weight adaptive adjustment module. Through continuous iteration, the convergence judgment module is input with the changed weights and the new independent variables obtained from the re-exploration, and finally outputs a solution that meets the requirements.

[0024] like Figure 3 As shown, the restart method for the TBCC intake includes the following steps: Step 1: Construct a TBCC engine component-level model.

[0025] Step 1.1: Construct a turbine-based component-level model, including components such as the intake, compressor, combustion chamber, turbine, mixture, and exhaust nozzle; Step 1.2: Construct a stamping-based component-level model, including components such as the front body, air intake, combustion chamber, and tail nozzle; Step 1.3: Construct the TBCC whole machine component level model: including modal control valve, rear body and other components.

[0026] Since the main target of this invention in the entire engine is the intake duct component, only the modeling method of the intake duct component will be described in detail.

[0027] The precursor geometry is conical, and its walls are often designed with three different turning angles. Its main functions are to pre-compress the incoming gas, reduce the gas velocity, and increase the total gas temperature and pressure. When the supersonic flow approaches the tip of the cone, a conical shock wave is formed, which consists of three oblique shock waves. After passing through the conical shock wave, the airflow enters the intake and then passes through a normal shock wave.

[0028] Therefore, the incoming flow enters the intake duct after undergoing adiabatic compression and directional reversal through three oblique shock waves in the forebody. The final-state parameters of the airflow after compression by the previous shock wave are the initial parameters for the compression process of the next shock wave. Based on the characteristics of shock wave compression, the relationship between the airflow parameters before and after shock wave compression can be expressed by the following formula:

[0029]

[0030]

[0031]

[0032] In the formula, subscript 1 represents the incoming flow, and subscript 2 represents the post-wave flow. Indicates a turning angle. Indicates the shock angle. M Represents the Mach number. P Represents static pressure. Represents the total temperature.

[0033] The incoming airflow remains supersonic after compression by the forebody. The inlet is a gradually expanding type, with a normal shock wave in the expansion section. After passing through the normal shock wave, the airflow velocity changes from supersonic to subsonic and gradually levels off. The relationship between the Mach number and total temperature and total pressure before and after the normal shock wave can be obtained using the formula for the relationship between airflow parameters before and after shock wave compression in this step. The back cross-sectional parameters of the shock wave are calculated in real time.

[0034] Step 2: The particle swarm optimization algorithm is used to obtain a restart design scheme.

[0035] Step 2.1: Population initialization phase; Set optimization goals:

[0036] Where error is the error value output by the model. This indicates the location of the normal shock wave.

[0037] Parameters such as particle number, population dimension, initial inertia weight, acceleration constant, maximum number of iterations, and iteration precision are set, and each particle is assigned a random initial position and initial exploration speed within the search range.

[0038] Step 2.2: Iterative Evaluation and Update Phase; Perform a fitness assessment on each particle: , in, This represents the position fitness of the d-th variable of the i-th particle at time t. This represents the position of the d-th variable of the i-th particle at time t. Indicates the position as The fitness of the corresponding particle variable.

[0039] Individual optimal position update determination:

[0040] in, and They represent The individual historical best position of the d-th variable of the i-th particle at time t and time t. When i=1, the initial position is taken as the individual historical best position.

[0041] Individual optimal fitness update determination:

[0042] in, Indicates the position as The fitness of the corresponding particle variable is the optimal fitness of the individual.

[0043] Global optimal position update determination: , in, Let represent the global optimal position of the d-th variable of the i-th particle at time t+1. Let represent the global optimal position of the d-th variable of the i-th particle at time t. This represents the individual optimal position of the m-th particle at time t for the n-th variable. In other words, the global optimal position is the optimal position selected from the individual optimal positions of all particles and the global optimal position at the previous time step.

[0044] Step 2.3: Velocity and Position Update Phase; For the d-th variable of the i-th particle, the velocity update formula is:

[0045] in, Let d represent the velocity of the i-th particle at time t+1. Indicates inertia weight, Indicates the first d Variables in t Independent learning rate at any given moment The learning factor is 2. A random number between 0 and 1. It is called the inertial term, and its function is to maintain the original direction of motion and exploration speed of the particle, and to avoid sudden changes. It is called the cognitive term, and its function is to drive the particle to move toward the best position it has ever discovered, reflecting individual experience; This is called the social term, and its function is to drive particles to move towards the best position found by the entire population, reflecting collective intelligence. A search speed boundary is set here to prevent particles from exploring too quickly, causing them to skip the optimal solution or diverge, or from exploring too slowly, reducing search efficiency and stabilizing the search process.

[0046] For the d-th variable of the i-th particle, the position update formula is:

[0047] in, This indicates the time step. Search boundaries need to be set here to ensure stable and reliable search results.

[0048] Step 2.4: Termination Phase; Termination conditions include: (1) Termination condition for maximum number of iterations,

[0049] Where u represents the current iteration number. This indicates the preset maximum number of iterations. When the time step is as described in step 2.3... When the time is 1 second, the number of iterations u is equal to the time t.

[0050] (2) The target value reaches the termination condition.

[0051] in, This represents the preset target fitness value.

[0052] (3) Termination conditions based on fitness convergence,

[0053] in, This indicates the magnitude of change, and its function is to consider the system to have converged if the global optimal fitness does not improve significantly over k consecutive generations.

[0054] Step 3: Adaptive adjustment of variable weights.

[0055] Each variable is assigned an independent weight and learning rate, which are dynamically adjusted based on its contribution to the objective function and constraints.

[0056] Step 3.1: Adjust the independent learning rate; Set an independent learning rate for each variable:

[0057] in, Based on the learning rate, Let be the weight of the d-th variable of the i-th particle at time t. Let d be the optimized weight of the d-th variable of the i-th particle at time t.

[0058] Based on the above formula, the improved exploration speed update formula is obtained:

[0059] Step 3.2: Constraint-oriented weight adjustment; Consider two optimization objectives: thrust error and normal shock wave position.

[0060] in, This represents the particle weights determined by the error constraint. This represents the particle weight determined by the normal shock wave position ratio. This represents the tradeoff coefficient at time t.

[0061]

[0062] Step 4: Obtain the intake restart path using linear interpolation: After steps 1-3 above, the values ​​corresponding to the intake manifold inlet area, fuel flow rate, and exhaust nozzle throat area when the intake manifold restarts are obtained, which is the final solution. Using linear interpolation, nine restart process points are obtained from the starting point (non-start point) to the ending point (restart point), for a total of eleven points. The engine states represented by these eleven points constitute the complete intake manifold restart path.

[0063] Taking one of the variables 'a' as an example, the formula for calculating the complete path is as follows:

[0064] in, The index represents the value of 'a' corresponding to the m-th point. Indicates the starting point (beginning point), subscript Indicates the restart point (end point). The calculation formulas for b and c are the same as those for a.

[0065] In summary, this invention, by introducing a particle swarm optimization approach with multivariate adaptive optimization, achieves a stable and reliable design method for the TBCC engine inlet restart scheme.

[0066] The following verification will be performed: The TBCC engine model constructed adopts a dual design point, and the engine design point parameters under the ramjet base state are shown in Table 1: Table 1 Design Point Parameters

[0067] The invention was first verified by selecting three engine intake failure states caused by three factors: intake manifold inlet area, fuel flow rate, and exhaust nozzle throat area. These states were determined separately. The physical meanings of the three optimized parameters a, b, and c in this invention are as follows:

[0068] in, This indicates the area of ​​the air intake. Indicates fuel flow rate. Indicates the area of ​​the nozzle throat, subscript Indicates the current state, subscript Indicates the design point status.

[0069] There are two optimization objectives:

[0070] in, This indicates the optimized engine thrust. This indicates the engine thrust before optimization.

[0071] First embodiment: With the intake duct inlet area being 1.1 times the design intake duct inlet area, and other parameters remaining unchanged, the engine parameters are shown in Table 2: Table 2 Parameter table of the first embodiment

[0072] It can be seen that at this time The shock wave is blown out of the air intake, which is in a non-activated state. The search limits for a, b, and c are set to [0.9, 1.3], and the output parameters are... At this point, the shock wave position is 1.081, the thrust is 47274.3 N, and the maximum thrust error during the entire restart process is... The intake manifold successfully restarted with minimal thrust change. The intake manifold inlet area, fuel flow rate, nozzle throat area, thrust, normal shock wave position, and thrust error throughout the entire restart process are as follows: Figure 4 As shown.

[0073] Second embodiment: With the fuel flow rate set to 1.1 times the design point fuel flow rate and all other parameters remaining unchanged, the engine parameters are shown in Table 3. Table 3 Parameter table of the second embodiment

[0074] It can be seen that at this time The shock wave is blown out of the air intake, and the engine remains off. The search limits for a, b, and c are set to [0.9, 1.3], and the output parameters are... At this point, the shock wave position is 1.022, the thrust is 46133.9 N, and the maximum thrust error during the entire restart process is... The intake manifold successfully restarted with minimal thrust change. The intake inlet area, fuel flow rate, nozzle throat area, thrust, normal shock wave position, and thrust error throughout the restart process are as follows: Figure 5 As shown.

[0075] Third embodiment: With the nozzle throat area set to 0.9 times the design nozzle throat area and all other parameters remaining unchanged, the engine parameters are shown in Table 4. Table 4 Parameter table of the third embodiment

[0076] It can be seen that at this time The shock wave is blown out of the air intake, and the engine remains off. The search limits for a, b, and c are set to [0.9, 1.3], and the output parameters are... At this point, the position of the normal shock wave is 1.309, the thrust is 44704.8 N, and the maximum thrust error during the entire startup process is [missing value]. The intake manifold successfully restarted with minimal thrust change. The intake manifold inlet area, fuel flow rate, nozzle throat area, thrust, normal shock wave position, and thrust error throughout the entire restart process are as follows: Figure 6 As shown.

[0077] In this embodiment, a computer device is provided, such as... Figure 7 As shown, it includes a memory 701, a processor 702, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements any of the above-mentioned restart methods for the TBCC intake based on an adaptive optimization algorithm.

[0078] Specifically, the computer device can be a computer terminal, a server, or a similar computing device.

[0079] In this embodiment, a computer-readable storage medium is provided, which stores a computer program that executes any of the above-described TBCC intake restart methods based on adaptive optimization algorithms.

[0080] Specifically, computer-readable storage media include both permanent and non-permanent, removable and non-removable media, which can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer-readable storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable storage media do not include transient media, such as modulated data signals and carrier waves.

[0081] Based on the same inventive concept, this invention also provides a restart device for a TBCC inlet based on an adaptive optimization algorithm, as described in the following embodiments. Since the principle of the restart device for a TBCC inlet based on an adaptive optimization algorithm is similar to that of the restart method for a TBCC inlet based on an adaptive optimization algorithm, the implementation of the restart device for a TBCC inlet based on an adaptive optimization algorithm can refer to the implementation of the restart method for a TBCC inlet based on an adaptive optimization algorithm; repeated details will not be elaborated further. As used below, the terms "unit" or "module" can refer to a combination of software and / or hardware that performs a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0082] Figure 8 This is a structural block diagram of a TBCC intake restart device based on an adaptive optimization algorithm according to an embodiment of the present invention, such as... Figure 8 As shown, it includes: model building module 801, iteration module 802 and generation restart control path module 803. The structure is described below.

[0083] The model building module 801 is used to build a nonlinear model at the component level of the TBCC engine, encapsulate the nonlinear model at the component level of the TBCC engine into a black box model, and determine the input parameter vector and output parameter vector of the black box model. The input parameter vector includes the inlet area of ​​the air intake, the fuel flow rate and the throat area of ​​the tail nozzle, and the output parameter vector includes the position of the normal shock wave and the engine thrust. The iteration module 802 is used to input the input parameter vector of the current iteration step into the black box model for solution based on an adaptive optimization algorithm, and use the solution output parameter vector as the fitness evaluation data of the current particle. According to the contribution of each input parameter vector to the output parameter vector in the fitness evaluation data, the independent weight and independent learning rate of each input parameter vector are dynamically adjusted, and the velocity and position of the particle are updated based on the independent weight and independent learning rate to generate the input parameter vector of the next iteration step. The iterative calculation is repeated until the iteration stopping condition is met, and the input parameter vector corresponding to the last iteration step is used as the restart target point data. The restart control path generation module 803 is used to take the initial running data as the path start point and the restart target point data as the path end point, and use a linear interpolation method to generate continuous intermediate state points. Based on the path start point, the intermediate state points and the path end point, the restart control path of the TBCC intake is constructed.

[0084] In one embodiment, the model building module includes: The first airflow parameter data unit is used to generate the first airflow parameter data after compression by three oblique shock waves based on the geometric and aerodynamic parameter data of the TBCC engine and the oblique shock wave compression relationship. The second airflow parameter data unit is used to generate the second airflow parameter data after normal shock wave compression based on the first airflow parameter data and the geometric parameter data of the air intake, through the normal shock wave compression formula. A ramjet-based inlet sub-model unit is constructed to perform time-series correlation between the first airflow parameter data and the second airflow parameter data, thereby generating a ramjet-based inlet model with a four-wave system design. Other sub-model building units are used to build turbine base component sub-models, ramjet combustion chamber sub-models, tail nozzle models, and modal control valve sub-models, respectively; A component-level nonlinear model unit is constructed to integrate the turbine-based component sub-model, the ramjet-based intake sub-model, the ramjet combustion chamber sub-model, the tail nozzle model, and the modal control valve sub-model to generate the TBCC engine component-level nonlinear model, and the TBCC engine component-level nonlinear model is stored in a preset task environment.

[0085] In one embodiment, the iteration module includes: The data input unit is used to receive the normal shock wave position output after the black box model is solved. and engine thrust ; The reference thrust acquisition unit is used to acquire the engine thrust when the inlet is not started, as the reference engine thrust. ; The relative error data unit is used to calculate the engine thrust. and the reference engine thrust The relative error data of engine thrust were calculated. error ,in, ; Generate fitness evaluation data units for determining the position of the normal shock wave. and the engine thrust relative error data error Perform data fusion and encapsulate it into fitness evaluation data for the current particle.

[0086] In one embodiment, the iteration module further includes: Independent weighting calculation unit, used to calculate the engine thrust relative error data (error) and normal shock wave position data. Generate independent weights for adjusting particle positions. ,in, d The number of the input parameter vector.t For a specific moment; The learning rate calculation unit is used to calculate the learning rate based on the set base learning rate. and the independent weights Generate the independent learning rate of the d-th variable for the i-th particle at time t. ,in, ,in, Let d be the independent weight of the d-th variable of the i-th particle at time t; Acquire velocity and position units to obtain the first... i The first particle d The input parameter vectors are in t Speed ​​of time and the i The first particle d The input parameter vectors are in t Location at any moment ; Determine the best historical position unit for each individual particle, used to compare the current particle's position. The individual's fitness and historical best fitness are compared to determine the individual's historical best position. ; The global optimal position is determined by comparing the individual historical best positions of all particles with the global optimal position at the previous time step, and the global optimal position is determined based on the comparison result. ; A velocity update formula unit is constructed to generate a velocity update formula. This formula generates the velocity of the particle in the next iteration step based on the velocity of the particle in the current iteration step. , For the first i The first particle d The input parameter vectors are in t The velocity at time +1 For the first d Variables in t Independent learning rate at any given moment For inertial weights, As a learning factor, A random number between 0 and 1; A position update formula unit is constructed to generate a position update formula. This formula generates the position of the particle in the next iteration step based on the current particle position. , For the first i The first particle d The input parameter vectors are in t The position at time +1 For time steps.

[0087] In one embodiment, the independent weight calculation unit is further configured to calculate the engine thrust relative error data. error Generate the weight of the first particle. Based on normal shock wave location data Generate the weight of the second particle. ; obtain t Time-based trade-off coefficients ,in, ; through the aforementioned tradeoff coefficients The weight of the first particle and the weight of the second particle Calculate the independent weights ,in, .

[0088] In one embodiment, the iteration module further includes: The data acquisition unit is used to obtain the current iteration number after each iteration calculation. t Fitness at the global optimal position; A first termination condition unit is defined, which is used to take the number of iterations reaching a preset maximum number of iterations threshold as the first termination condition. The second termination condition unit is used to determine the second termination condition by setting the fitness of the global optimal position to a preset target fitness value. The third termination condition unit is determined to adjust the fitness of the global optimal position over a number of consecutive iterations. k The third termination condition is that the change within the range is less than a preset change range. , For the preset range of change, Let be the fitness of the globally optimal position in the t-th iteration. For the first The fitness of the global optimal position after iteration number; The iteration stopping unit is used to stop the iteration when any one of the first termination condition, the second termination condition, or the third termination condition is met.

[0089] In one embodiment, the restart control path module includes: The intermediate state point calculation unit is used to calculate the variable value corresponding to each intermediate state point based on the variable value corresponding to the starting point of the path and the variable value corresponding to the ending point of the path, using a linear interpolation method. The variable values ​​include the relative value of the intake manifold inlet area, the relative value of the fuel flow rate, and the relative value of the exhaust nozzle throat area. , For the first mThe variable values ​​corresponding to each intermediate state point The variable value corresponding to the starting point of the path. The variable value corresponding to the endpoint of the path. M This represents the total number of intermediate state points. m From 1 to M Integers; Construct a control path unit to connect the path start point, M The intermediate state points and the path endpoints are combined in sequence to form a system containing... M +2 status points of the intake duct restart control path.

[0090] The embodiments of the present invention achieve the following technical effects: To overcome the problem of engine thrust variation caused by traditional solutions, this invention introduces the inlet area of ​​the air intake and the throat area of ​​the tail nozzle to balance the engine thrust variation caused by changes in fuel flow. At the same time, since the three variables have different degrees of influence on the air intake performance, an adaptive weight adjustment strategy is introduced to set independent learning rates and weights for each variable. Under the condition of satisfying a smooth transition of engine thrust, the air intake restart scheme design is completed, which has positive significance for improving the safety and stability of TBCC engines during flight.

[0091] This invention introduces multivariate restart path design, overcoming the shortcomings of traditional methods that only change fuel flow, thus altering engine thrust. For multivariate optimization, an adaptive variable weight adjustment strategy is employed, adjusting variable weights and exploration rate in real time during iterative optimization, improving exploration speed and facilitating rapid response to engine start-up problems. This invention introduces intelligent algorithm-based multivariate adaptive optimization into inlet restart design, providing a new approach to solving this problem and holding significant importance for advancing intelligent control of aero-engines and improving engine performance and reliability.

[0092] Obviously, those skilled in the art should understand that the modules or steps of the above-described embodiments of the present invention can be implemented using general-purpose computing devices. They can be centralized on a single computing device or distributed across a network of multiple computing devices. Optionally, they can be implemented using computer-executable program code, thereby storing them in a storage device for execution by a computing device. In some cases, the steps shown or described can be performed in a different order than those presented here, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. Thus, the embodiments of the present invention are not limited to any particular hardware and software combination.

[0093] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, various modifications and variations can be made to the embodiments of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method of restart of a TBCC inlet based on an adaptive optimization algorithm, characterized in that, include: A component-level nonlinear model of the TBCC engine is constructed, and the component-level nonlinear model of the TBCC engine is encapsulated as a black box model. The input parameter vector and output parameter vector of the black box model are determined. The input parameter vector includes the inlet area of ​​the air intake, the fuel flow rate and the throat area of ​​the tail nozzle, and the output parameter vector includes the position of the normal shock wave and the engine thrust. Based on the adaptive optimization algorithm, the input parameter vector of the current iteration step is input into the black box model for solution, and the output parameter vector obtained by solution is used as the fitness evaluation data of the current particle. According to the contribution of each input parameter vector to the output parameter vector in the fitness evaluation data, the independent weight and independent learning rate of each input parameter vector are dynamically adjusted, and the velocity and position of the particle are updated based on the independent weight and independent learning rate to generate the input parameter vector of the next iteration step. The iterative calculation is repeated until the iteration stopping condition is met, and the input parameter vector corresponding to the last iteration step is used as the restart target point data. Using the initial running data as the starting point of the path and the restart target point data as the ending point of the path, a continuous series of intermediate state points are generated using a linear interpolation method. Based on the starting point of the path, the intermediate state points, and the ending point of the path, a restart control path for the TBCC intake is constructed.

2. The method of restart of a TBCC inlet based on an adaptive optimization algorithm as recited in claim 1, wherein, Constructing a component-level nonlinear model for the TBCC engine, including: Based on the geometric and aerodynamic parameter data of the TBCC engine, the first airflow parameter data after compression by three oblique shock waves is generated sequentially through the oblique shock wave compression formula. Based on the first airflow parameter data and the geometric parameter data of the intake duct, the second airflow parameter data after normal shock wave compression is generated through the normal shock wave compression formula. The first airflow parameter data and the second airflow parameter data are correlated in time to generate a ramjet base inlet sub-model with a four-wave system design. Sub-models for turbine base components, ramjet combustion chamber, tail nozzle, and modal control valve are constructed respectively. The turbine-based component sub-model, the ramjet-based intake sub-model, the ramjet combustion chamber sub-model, the tail nozzle model, and the modal control valve sub-model are integrated to generate the TBCC engine component-level nonlinear model, and the TBCC engine component-level nonlinear model is stored in a preset task environment.

3. The method of claim 1, wherein the adaptive optimization algorithm-based restart of a TBCC inlet comprises: The input parameter vector of the current iteration step is input into the black box model for solution, and the output parameter vector obtained from the solution is used as the fitness evaluation data of the current particle, including: receiving the output of the black-box model solution and engine thrust ; Obtaining engine thrust at intake port non-activation as reference engine thrust ; by said engine thrust and said reference engine thrust , calculating engine thrust relative error data error wherein ; said shock wave position and said engine thrust relative error data error perform data fusion, encapsulated as fitness evaluation data for the current particle.

4. The TBCC intake restart method based on adaptive optimization algorithm as described in claim 1, characterized in that, Based on the contribution of each input parameter vector to the output parameter vector in the fitness evaluation data, the independent weights and independent learning rates of each input parameter vector are dynamically adjusted, and the particle velocity and position are updated based on the independent weights and independent learning rates to generate the input parameter vector for the next iteration step, including: Based on the engine thrust relative error data (error) and normal shock wave position data Generate independent weights for adjusting particle positions. ,in, d The number of the input parameter vector. t For a specific moment; Based on the set base learning rate and the independent weights , generate the first i Individual particles t The first moment d Independent learning rates of each variable ,in, ,in, Let d be the independent weight of the d-th variable of the i-th particle at time t; Get the i The first particle d The input parameter vectors are in t Speed ​​of time and the i The first particle d The input parameter vectors are in t Location at any moment ; Compare the current position of the particles The individual's fitness and historical best fitness are compared to determine the individual's historical best position. ; The global optimal position is determined by comparing the individual historical best positions of all particles with the global optimal position at the previous time step. ; A velocity update formula is constructed, and the velocity of the particle in the next iteration step is generated based on the velocity of the particle in the current iteration step using this formula. , For the first i The first particle d The input parameter vectors are in t The velocity at time +1 For the first d Variables in t Independent learning rate at any given moment For inertial weights, As a learning factor, A random number between 0 and 1; A position update formula is constructed, and the position of the particle in the next iteration step is generated based on the position of the particle in the current iteration step using this formula. , For the first i The first particle d The input parameter vectors are in t The position at time +1 For time steps.

5. The TBCC inlet restart method based on adaptive optimization algorithm as described in claim 4, characterized in that, Based on the engine thrust relative error data (error) and normal shock wave position data Generate independent weights for adjusting particle positions. ,include: Based on engine thrust relative error data error Generate the weight of the first particle. ; Based on the normal shock wave location data Generate the weight of the second particle. ; Get t Time-based trade-off coefficients ,in, ; Through the aforementioned trade-off coefficients The weight of the first particle and the weight of the second particle Calculate the independent weights ,in, .

6. The TBCC inlet restart method based on adaptive optimization algorithm as described in claim 1, characterized in that, The iteration stopping conditions are met, including: After each iteration, obtain the current iteration number. t Fitness at the global optimal position; The first termination condition is to reach a preset maximum iteration count threshold. The second termination condition is to make the fitness of the global optimal position reach a preset target fitness value. The fitness of the global optimal position is determined in consecutive iterations. k The third termination condition is that the change within the range is less than a preset change range. , For the preset range of change, Let be the fitness of the globally optimal position in the t-th iteration. For the first The fitness of the global optimal position after iteration number; The iteration stops when any one of the first termination condition, the second termination condition, or the third termination condition is met.

7. The TBCC inlet restart method based on an adaptive optimization algorithm as described in any one of claims 1 to 6, characterized in that, Using the initial running data as the starting point and the restart target point data as the ending point, a continuous series of intermediate state points are generated using a linear interpolation method. Based on the starting point, the intermediate state points, and the ending point, a restart control path for the TBCC inlet is constructed, including: Based on the variable values ​​corresponding to the starting point and ending point of the path, the variable values ​​corresponding to each intermediate state point are calculated using a linear interpolation method. These variable values ​​include the relative values ​​of the intake manifold inlet area, the relative value of the fuel flow rate, and the relative value of the exhaust nozzle throat area. , For the first m The variable values ​​corresponding to each intermediate state point The variable value corresponding to the starting point of the path. The variable value corresponding to the endpoint of the path. M This represents the total number of intermediate state points. m From 1 to M Integers; The starting point of the path, M The intermediate state points and the path endpoints are combined in sequence to form a system containing... M +2 status points of the intake duct restart control path.

8. A restart device for a TBCC inlet based on an adaptive optimization algorithm, characterized in that, include: The model building module is used to build a component-level nonlinear model of the TBCC engine, encapsulate the component-level nonlinear model of the TBCC engine into a black box model, and determine the input parameter vector and output parameter vector of the black box model. The input parameter vector includes the inlet area of ​​the air intake, the fuel flow rate and the throat area of ​​the tail nozzle, and the output parameter vector includes the position of the normal shock wave and the engine thrust. The iterative module is used to input the input parameter vector of the current iteration step into the black box model for solution based on an adaptive optimization algorithm, and use the solution output parameter vector as the fitness evaluation data of the current particle. According to the contribution of each input parameter vector to the output parameter vector in the fitness evaluation data, the independent weight and independent learning rate of each input parameter vector are dynamically adjusted, and the velocity and position of the particle are updated based on the independent weight and independent learning rate to generate the input parameter vector of the next iteration step. The iterative calculation is repeated until the iteration stopping condition is met, and the input parameter vector corresponding to the last iteration step is used as the restart target point data. A restart control path generation module is used to take the initial running data as the path start point and the restart target point data as the path end point, and use a linear interpolation method to generate continuous intermediate state points. Based on the path start point, the intermediate state points and the path end point, a restart control path for the TBCC intake is constructed.

9. A computer 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, it implements the TBCC intake restart method based on the adaptive optimization algorithm as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that performs the restart method for the TBCC inlet based on an adaptive optimization algorithm as described in any one of claims 1 to 7.

Images