A variable cycle engine dual-loop closed-loop mode switching control method

Abstract

The application discloses a variable cycle engine double-loop closed-loop mode switching control method, which comprises the following steps: determining the tail nozzle network model parameters according to the dynamic thermodynamic mechanism analysis of a tail nozzle component, constructing a tail nozzle data-driven network model based on a BP neural network, and combining a component-level dynamic model of a main component in an upstream flow path to obtain a hybrid dynamic model of the variable cycle engine; according to the influence of an adjustable geometry component on performance parameters of the variable cycle engine, designing a double-loop closed-loop control scheme with variable speed and thrust feedback by taking thrust performance as an optimization target, and designing a mode switching control method of the model based on the hybrid dynamic model of the variable cycle engine. The tail nozzle model with a long time consumption in the variable cycle engine is simplified to reduce the time consumption of the whole machine dynamic model calculation, the double-loop closed-loop control scheme with variable speed and thrust and containing the mode switching process is proposed based on the model, and the stability of the variable cycle engine during mode switching can be effectively improved.

Description

Technical Field

[0001] This invention relates to the field of aero-engine modeling and control, and in particular to a dual-loop closed-loop mode switching control method for a variable cycle engine. Background Technology

[0002] A neural network is a mathematical model that simulates the structure of the human brain's neural network, learning how the brain processes complex information. It features high nonlinearity, high precision, and powerful information processing capabilities. A backpropagation (BP) neural network is a multi-layered network trained with weights on a nonlinear differentiable function. It includes an input layer, hidden layers, and an output layer. Neurons between layers are fully connected, while neurons within the same layer are not connected. Its most significant advantage is that it can achieve a highly nonlinear mapping from a space of pattern vectors p (composed of m input neurons) to a space of yn (where n is the number of output nodes) using only sample data, without needing to establish a systematic mathematical model. Therefore, this invention proposes to use a neural network to build a tailpipe model.

[0003] The process of variable cycle mode switching refers to the opening or closing of the mode switching valve, which is a transitional state of the engine. The mode switching process of a variable cycle engine is relatively complex. Since the engine contains many components, based on the performance requirements and working limits of the main engine components such as the compressor, combustion chamber and turbine components, the following principles should be followed: (1) Due to the limitations of rotor strength and stiffness, as well as the limitations of thermal strength and thermal stress of the combustion chamber and turbine components, parameters such as temperature and speed should change smoothly, and the rotor speed and turbine inlet temperature should not exceed the theoretical maximum value to avoid overheating, overspeeding, surge, etc.; (2) Meet the working state plan and formulate an engine adjustment plan according to the specific working state, such as constant thrust and reduced fuel consumption rate.

[0004] Before building a variable cycle engine control system, it is necessary to formulate the control law for the variable cycle engine, and ensure that the controlled parameters change according to the predetermined law by adjusting each control variable. Single bypass mode and double bypass mode are two typical operating modes of double bypass variable cycle engines. How to ensure that the engine switches safely between these two operating modes, and ensure that the engine does not over-rev, over-temperature, or surge during the switching process, while minimizing the fluctuation of parameters such as engine speed, is the main research content. Summary of the Invention

[0005] The technical problem this invention aims to solve is to address the shortcomings of the prior art by using a neural network nozzle model to replace the gas path calculation nozzle model, thereby providing a more real-time hybrid component-level model for variable cycle engines. Simultaneously, it proposes a dual-loop closed-loop mode-switching control method, employing an incremental PI controller to achieve closed-loop control of variable speed and thrust. This scheme simplifies the time-consuming nozzle model in variable cycle engines, reducing the computation time of the overall engine dynamic model. Based on this model, a dual-loop closed-loop control scheme for variable speed and thrust, including the mode-switching process, is proposed, effectively improving the stability of variable cycle engines during mode switching.

[0006] To solve the above-mentioned technical problems, the present invention adopts the following technical solution:

[0007] Step A) Evaluate the single-step time consumption of the main components in the flow path calculation of the variable cycle engine. For the tail nozzle component with the longest time consumption, determine the tail nozzle network model parameters based on aerodynamic and thermodynamic mechanism analysis, and construct a tail nozzle data-driven network model based on BP neural network. Combine the tail nozzle neural network model with the component-level dynamic model of the main components of the upstream flow path to obtain the hybrid dynamic model of the variable cycle engine.

[0008] Step B) Analyze the impact of different geometric parameters on engine performance, select appropriate control and controlled variables, and determine the multivariable control loop for the mode switching process; with thrust performance as the optimization objective, design a dual-loop closed-loop control scheme for variable speed and thrust, and design a mode switching control method based on the hybrid dynamic model of the variable cycle engine.

[0009] As a further optimization of the dual-loop closed-loop mode switching control method for a variable-cycle engine according to the present invention, the specific steps of step A) are as follows:

[0010] Step A1): In order to study how to reduce the calculation time of the main aerodynamic and thermodynamic components in the component-level model, the main components are run separately in the component-level model. Since the time consumed by a single run is very small, each component is run 10,000 times for comparison. The simulation results show that the four components with the most time consumption are the compressor, high-pressure turbine, low-pressure turbine and tail nozzle. Among them, the tail nozzle consumes the most time, accounting for 20% of the total time consumption in the entire gas path calculation.

[0011] Step A2) Determine the parameters of the nozzle network model based on the aerodynamic and thermodynamic mechanism analysis. The nozzle aerodynamic and thermodynamic model has six output parameters y: nozzle outlet flow rate W9, nozzle outlet pressure P9, nozzle outlet temperature T9, nozzle outlet airflow velocity V9, thrust Fn, and fuel consumption rate sfc. That is, the output variable y is:

[0012] y = [W9, P9, T9, V9, Fn, sfc]

[0013] In the aerodynamic and thermodynamic component-level model of the exhaust nozzle, the actual inlet airflow of the exhaust nozzle is obtained by mixing the exhaust gas from the afterburner and the exhaust gas from the bypass duct. Therefore, a mixed airflow flow rate W is required. cool7 External bypass outlet pressure P 16 Temperature T at the duct outlet 16 To obtain the outlet pressure P9, outlet airflow velocity V9, flow rate W9, thrust Fn, and fuel consumption rate sfc, the tail nozzle requires a fuel quantity W. fb The following parameters are used to determine the following parameters: throat area A8, nozzle outlet area A9, inlet static pressure Ps0, inlet air velocity v0, nozzle inlet flow rate W7, nozzle inlet pressure P7, and nozzle inlet temperature T7.

[0014] Therefore, the nozzle input quantity x is:

[0015] x = [W] fb ,A8,A9,Ps0,v0,W7,P7,T7,W cool7 ,P 16 ,T 16 ]

[0016] Step A3): Select inputs under different operating conditions and perform gas path calculations in the variable cycle engine component-level model to obtain corresponding dynamic outputs, thereby acquiring a large amount of sample data. This large amount of sample data is then allocated into training data, validation data, and test data, with proportions of 70%, 15%, and 15%, respectively. There are 11 inputs and 6 outputs. Based on the number of inputs and outputs, the number of hidden layer neurons is determined using an empirical formula, ranging from 5 to 15. For the training algorithm, the commonly used LM optimization algorithm is selected, which is a good training algorithm for medium-sized networks. Since the number of hidden layer neurons is uncertain, the method of increasing or decreasing the number of hidden layer neurons is chosen for retraining. The accuracy and reliability of the data are compared using performance graphs, training state graphs, error histograms, and regression graphs of the simulated data to determine the optimal number of neurons. After determining the number of hidden layer neurons to be 12, training continues until the maximum number of training iterations or the required accuracy is reached. Training is then completed, the data is saved, and an offline calculation model is constructed, thus establishing a tail nozzle data-driven network model based on a BP neural network.

[0017] Step A4) involves constructing a hybrid dynamic model of the variable cycle engine by combining the tail nozzle neural network model with the component-level dynamic models of the main upstream flow path components. The hybrid dynamic model sequentially calls each component model for calculation, with the flow path calculation order consistent with the component-level models. When calling the tail nozzle model for calculation, the tail nozzle aerodynamic-thermodynamic model is replaced with the tail nozzle neural network model. The input obtained from the upstream flow path calculation is processed by the offline neural network tail nozzle model to obtain the output. Based on the component constraints, a set of nonlinear common working equations that need to be solved is established.

[0018] As a further optimization of the dual-loop closed-loop mode switching control method for a variable-cycle engine according to the present invention, the specific steps of step B) are as follows:

[0019] Step B1) Analyze the impact of each adjustable geometric component on the performance parameters of the variable cycle engine. By analyzing the specific effects of each adjustable parameter on the output parameters, analyze the influence of each adjustable parameter on the changing trend of the output parameters of the variable cycle engine component-level model and on the average performance of the output parameters, in order to select the control and controlled variables during the closed-loop control mode switching process;

[0020] Step B2) Design an incremental PI controller, using a closed-loop control scheme with nozzle throat area to control the low-pressure rotor speed, a closed-loop control scheme with main fuel flow to control engine thrust, and an open-loop control scheme for other adjustable parameters. Thrust is the optimization objective; based on a hybrid dynamic model of a variable-cycle engine, a dual-loop closed-loop control method with variable speed and thrust feedback is designed.

[0021] Furthermore, the specific steps of step B1) are as follows:

[0022] Step B1.1) The variable cycle engine has numerous geometrically adjustable parameters, and these parameters are strongly coupled, increasing the design difficulty of its mode switching control plan. To lay the foundation for the engine control plan, a component-level model of the variable cycle engine was simulated. All geometrically variable parameters were increased by 5% during a dynamic process where the main fuel flow rate increased by 1% near the design point within 50 seconds, to obtain the impact of changes in adjustable geometric parameters on engine performance.

[0023] Step B1.2): Fan speed represents the overall engine flow rate, and engine thrust represents the combustion gas's work capacity. Therefore, low-pressure rotor speed and engine thrust are selected as the controlled variables. During mode switching, thrust fluctuations should be minimized. Simulations show that the geometrically adjustable parameter with the greatest average impact on fan speed and engine thrust is the nozzle throat area A8, while the other parameter with the greatest impact in the dynamic process is the fuel flow rate W. fb Therefore, the nozzle throat area and fuel flow rate are selected as control variables.

[0024] Furthermore, the specific steps of step B2) are as follows:

[0025] Step B2.1), after determining the control variable and the controlled variable, determines which specific control parameter controls the low-pressure rotor speed n. L (Or engine thrust Fn), which needs to be analyzed. When the nozzle throat area increases, the engine low-pressure rotor speed and engine thrust also increase, and the nozzle throat area has a significant impact on the low-pressure rotor speed; when the main fuel flow rate increases, the engine low-pressure rotor speed and engine thrust also increase, and the main fuel flow rate has a significant impact on the engine thrust.

[0026] If the following dual-loop closed-loop control scheme is adopted: main fuel flow closed-loop control of engine low-pressure rotor speed and nozzle throat area closed-loop control of engine thrust. During mode switching, if engine thrust tends to decrease, the thrust closed-loop control loop will increase the nozzle throat area. This increase in nozzle throat area will simultaneously increase low-pressure rotor speed. However, since the nozzle throat area has a significant impact on low-pressure rotor speed, the low-pressure rotor speed closed-loop will significantly reduce fuel flow to suppress the rapid decrease in low-pressure rotor speed, causing the engine to enter the lean-fire shutdown boundary. Alternatively, if nozzle throat area closed-loop control of engine low-pressure rotor speed and main fuel flow closed-loop control of engine thrust are adopted, if engine thrust tends to increase during mode switching, the thrust closed-loop control loop will reduce main fuel flow. This decrease in main fuel flow will simultaneously decrease low-pressure rotor speed. In this case, the low-pressure rotor speed closed-loop can suppress the decrease in low-pressure rotor speed by slightly increasing the nozzle throat area. Therefore, nozzle throat area closed-loop control of low-pressure rotor speed and main fuel flow closed-loop control of engine thrust are chosen. Therefore, a dual-loop closed-loop mode switching control scheme is designed, which uses closed-loop control of the tail nozzle throat area to control the low-pressure rotor speed and closed-loop control of the main fuel flow to control the engine thrust.

[0027] Step B2.2) employs a dual-loop control scheme: closed-loop control of the nozzle throat area to regulate the low-pressure rotor speed, and closed-loop control of the main fuel flow to regulate engine thrust. Other adjustable parameters are controlled in an open-loop manner, linearly transitioning from the current mode design point to the target mode design point. To optimize thrust performance, the dual-loop closed-loop control system does not simply set the target speed and thrust constant. Analysis shows that if the control speed remains constant during mode switching, the engine thrust changes significantly. When switching from a single-bypass to a double-bypass system, the thrust decreases dramatically. If the low-pressure rotor speed remains constant, the thrust fluctuates greatly during the decrease. Therefore, it is necessary to appropriately increase the commanded low-pressure rotor speed. At this time, the nozzle throat area changes more, and the thrust increases slightly. Only a small adjustment to the fuel flow is needed to maintain constant thrust. When switching from a double-bypass to a single-bypass system, the thrust increases slightly. Therefore, only a slight decrease in the commanded low-pressure rotor speed is needed to maintain small thrust fluctuations. This method takes thrust performance as the optimization target. When switching from a single bypass to a double bypass, it designs a control law that keeps the engine thrust constant and increases the low-pressure rotor speed. When switching from a double bypass to a single bypass, it designs a control law that keeps the engine thrust constant and decreases the low-pressure rotor speed.

[0028] Step B2.3): At a ground point with H = 0 km and Ma = 0, the simulation demonstrates the switching between two modes of the variable cycle engine: single-bypass, double-bypass, and single-bypass. During the mode switching process, all other adjustable parameters linearly and gradually change from the current mode design point to the target mode design point over a time interval of 3 seconds, and all are open-loop inputs. The simulation results show that using the variable speed-thrust dual-loop closed-loop mode switching control method, the engine thrust fluctuation during the mode switching process is small, the fuel consumption rate does not increase, and the surge margin does not decrease to the critical value.

[0029] Compared with existing solutions, the present invention, employing the above technical solution, has the following technical advantages:

[0030] (1) The hybrid dynamic model of the variable cycle engine used in this invention is constructed by combining the tail nozzle data driven network model based on BP neural network with the component-level dynamic model of the main components of the upstream flow path. Compared with the conventional variable cycle engine component-level model, the time spent on the tail nozzle, which originally took the most time, is greatly reduced, saving calculation time, while ensuring the calculation accuracy of each section.

[0031] (2) The variable speed-tail nozzle throat area and thrust-fuel flow dual-loop closed-loop mode switching control method proposed in this invention uses an incremental PI controller to realize the closed-loop control of variable speed and thrust, which can ensure that the thrust fluctuation of the variable cycle engine is small during the mode switching process and the switching process remains stable, thus meeting the requirements of engineering application. Attached Figure Description

[0032] Figure 1This is a flowchart of the dual-loop closed-loop mode switching control method for the variable cycle engine of the present invention;

[0033] Figure 2 This refers to the time consumed by each component in the engine component-level model;

[0034] Figure 3 It is a neural network structure model of the tail nozzle;

[0035] Figure 4 This is the process of building a neural network model for the exhaust nozzle;

[0036] Figure 5 This is a diagram of a closed-loop control structure for dual-loop mode switching;

[0037] Figure 6 (a) Comparison of the time consumption of each component in the component-level model after replacing the original gas path calculation tail nozzle model with neural network modeling tail nozzle; Figure 6 (b) Comparison of the time consumption for calculating the tail nozzle using the original component-level model and the neural network model of the tail nozzle;

[0038] Figure 7 Fuel flow rate W when switching from a single bypass to a double bypass fb And the change in tail nozzle area A8;

[0039] Figure 8 This refers to the change in output parameters when switching from a single-function to a double-function architecture;

[0040] Figure 9 Fuel flow rate W when switching from a dual-shunt to a single-shunt switch fb And the change in tail nozzle area A8;

[0041] Figure 10 This refers to the change in output parameters when switching from a double-function to a single-function operation. Detailed Implementation

[0042] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings.

[0043] The present invention aims to evaluate the single-step time of major components in the flow path calculation based on the component-level dynamic model of a variable cycle engine. For the nozzle component with the longest time consumption, the parameters of the nozzle network model are determined based on its aerodynamic and thermodynamic mechanisms. A data-driven network model of the nozzle based on a BP neural network is constructed. Combined with the component-level dynamic model of the main upstream flow path components, a hybrid dynamic model of the variable cycle engine is obtained. Based on the influence of adjustable geometric components on the performance parameters of the variable cycle engine, and with thrust performance as the optimization objective, a dual-loop closed-loop control scheme with variable speed and thrust feedback is designed. A mode-switching control method is designed based on the hybrid dynamic model of the variable cycle engine.

[0044] The specific implementation of this invention takes the closed-loop mode switching control of a certain type of variable cycle engine as an example. This invention describes a dual-loop mode switching control method for a hybrid variable cycle engine, specifically including the following steps:

[0045] Step A) Evaluate the single-step time consumption of the main components in the flow path calculation of the variable cycle engine. For the tail nozzle component with the longest time consumption, determine the tail nozzle network model parameters based on aerodynamic and thermodynamic mechanism analysis, and construct a tail nozzle data-driven network model based on BP neural network. Combine the tail nozzle neural network model with the component-level dynamic model of the main components of the upstream flow path to obtain the hybrid dynamic model of the variable cycle engine.

[0046] Step A1): To investigate how to reduce the computation time of major aerodynamic and thermodynamic components in the component-level model, the major components were run individually in the component-level model. Since the time consumed per run is extremely short, each component was run 10,000 times for comparison. The time consumption results are as follows: Figure 2 As shown in the simulation results, the four components that consume the most time are the compressor, high-pressure turbine, low-pressure turbine, and tail nozzle. Among them, the tail nozzle consumes the most time, accounting for 20% of the total time in the gas path calculation.

[0047] Step A2) Determine the parameters of the nozzle network model based on the aerodynamic and thermodynamic mechanism analysis. The nozzle aerodynamic and thermodynamic model has six output parameters y: nozzle outlet flow rate W9, nozzle outlet pressure P9, nozzle outlet temperature T9, nozzle outlet airflow velocity V9, thrust Fn, and fuel consumption rate sfc. That is, the output variable y is:

[0048] y = [W9, P9, T9, V9, Fn, sfc]

[0049] In the aerodynamic and thermodynamic component-level model of the exhaust nozzle, the actual inlet airflow of the exhaust nozzle is obtained by mixing the exhaust gas from the afterburner and the exhaust gas from the bypass duct. Therefore, a mixed airflow flow rate W is required. cool7 External bypass outlet pressure P 16 Temperature T at the duct outlet 16 To obtain the outlet pressure P9, outlet airflow velocity V9, flow rate W9, thrust Fn, and fuel consumption rate sfc, the tail nozzle requires a fuel quantity W. fb The following parameters are used to determine the following parameters: throat area A8, nozzle outlet area A9, inlet static pressure Ps0, inlet air velocity v0, nozzle inlet flow rate W7, nozzle inlet pressure P7, and nozzle inlet temperature T7.

[0050] Therefore, the nozzle input quantity x is:

[0051] x = [W] fb ,A8,A9,Ps0,v0,W7,P7,T7,W cool7 ,P16 ,T 16 ]

[0052] After determining the input and output quantities, the neural network structure model of the tail nozzle is established as shown in the figure below. Figure 3 As shown.

[0053] Step A3): Select inputs under different operating conditions and perform gas path calculations in the variable cycle engine component-level model to obtain corresponding dynamic outputs, thereby acquiring a large amount of sample data. This large amount of sample data is then allocated into training data, validation data, and test data, with proportions of 70%, 15%, and 15%, respectively. There are 11 inputs and 6 outputs. Based on the number of inputs and outputs, the number of hidden layer neurons is determined using an empirical formula, ranging from 5 to 15. For the training algorithm, the commonly used LM optimization algorithm is selected, which is a good training algorithm for medium-sized networks. Since the number of hidden layer neurons is uncertain, the number of hidden layer neurons is increased or decreased during retraining. The accuracy and reliability of the data are compared using performance graphs, training state graphs, error histograms, and regression graphs of the simulated data to determine the optimal number of neurons. After determining the number of hidden layer neurons to be 12, training continues until the maximum number of training iterations or the required accuracy is reached. Training is then completed and the data is saved. An offline calculation model is constructed, thus establishing a tail nozzle data-driven network model based on a BP neural network. The tail nozzle neural network model construction process is as follows: Figure 4 As shown;

[0054] Step A4) involves constructing a hybrid dynamic model of the variable cycle engine by combining the tail nozzle neural network model with the component-level dynamic models of the main upstream flow path components. The hybrid dynamic model sequentially calls each component model for calculation, with the flow path calculation order consistent with the component-level models. When calling the tail nozzle model for calculation, the tail nozzle aerodynamic-thermodynamic model is replaced with the tail nozzle neural network model. The input obtained from the upstream flow path calculation is processed by the offline neural network tail nozzle model to obtain the output. Finally, a set of nonlinear common working equations to be solved is established based on the component constraints.

[0055] Step B) Analyze the impact of different geometric parameters on engine performance, select appropriate control and controlled variables, and determine the multivariable control loop for the mode switching process; with thrust performance as the optimization objective, design a dual-loop closed-loop control scheme for variable speed and thrust, and design a mode switching control method based on the hybrid dynamic model of the variable cycle engine.

[0056] Step B1) Analyze the impact of each adjustable geometric component on the performance parameters of the variable cycle engine. By analyzing the specific effects of each adjustable parameter on the output parameters, analyze the influence of each adjustable parameter on the changing trend of the output parameters of the variable cycle engine component-level model and the influence of each adjustable parameter on the average performance of the output parameters, in order to select the control and controlled variables during the closed-loop control mode switching process;

[0057] Step B1.1) The variable cycle engine has numerous geometrically adjustable parameters, and these parameters are strongly coupled, which increases the design difficulty of its mode switching control plan. To lay the foundation for the engine control plan, a component-level model of the variable cycle engine was simulated. All geometrically variable parameters were increased by 5% during a dynamic process in which the main fuel flow rate increased by 1% near the design point within 50 seconds. The simulation results are shown in Table 1.

[0058] Table 1. Numerical values ​​of the impact of adjustable parameter variations on the dynamic performance of the variable cycle engine.

[0059]

[0060] In Table 1, the average impact is:

[0061]

[0062] e i For the adjustable parameter variation in single-function and double-function modes, the effect on n L n H The magnitude of the influence of W2, T4, Fn, and bpr, n L n H W2, T4, Fn, and bpr represent the low-pressure rotor speed, high-pressure rotor speed, fan flow rate, combustion chamber outlet temperature, engine thrust, and engine bypass ratio, respectively. The adjustable geometry components mainly include the nozzle throat area A8 and the rear adjustable bypass ejector outlet area A. 163 Front fan guide vane angle α Fan CDFS guide vane angle α CDFS Compressor guide vane angle α Comp Low-pressure turbine guide angle α LT ;

[0063] Step B1.2): Fan speed represents the overall engine flow rate, and engine thrust represents the combustion gas's work capacity. Therefore, low-pressure rotor speed and engine thrust are selected as the controlled variables. During mode switching, thrust fluctuations should be minimized. Simulations show that the geometrically adjustable parameter with the greatest average impact on fan speed and engine thrust is the nozzle throat area A8, while the other parameter with the greatest impact in the dynamic process is the fuel flow rate W. fb Therefore, the nozzle throat area and fuel flow rate are selected as control variables.

[0064] Step B2) Design an incremental PI controller with a closed-loop control scheme using the nozzle throat area to control the low-pressure rotor speed, a closed-loop control of the main fuel flow to control the engine thrust, and an open-loop control scheme for other adjustable parameters. Thrust is the optimization objective. Based on a hybrid dynamic model of a variable-cycle engine, a dual-loop closed-loop control method with variable speed and thrust feedback is designed.

[0065] Step B2.1), after determining the control variable and the controlled variable, determines which specific control parameter controls the low-pressure rotor speed n. L (Or engine thrust Fn), which needs to be analyzed. When the nozzle throat area increases, the engine low-pressure rotor speed and engine thrust also increase, and the nozzle throat area has a significant impact on the low-pressure rotor speed. When the main fuel flow rate increases, the engine low-pressure rotor speed and engine thrust also increase, and the main fuel flow rate has a significant impact on the engine thrust. If the following dual-loop closed-loop control scheme is adopted: the main fuel flow rate closed-loop controls the engine low-pressure rotor speed, and the nozzle throat area closed-loop controls the engine thrust. When the engine thrust tends to decrease during mode switching, the thrust closed-loop control loop will increase the nozzle throat area. The increase in the nozzle throat area will also increase the low-pressure rotor speed. However, since the nozzle throat area has a huge impact on the low-pressure rotor speed, the low-pressure rotor speed closed-loop loop will significantly reduce the fuel flow rate to suppress the rapid decrease in low-pressure rotor speed, causing the engine to enter the lean-fuel shutdown boundary. If the nozzle throat area closed-loop controls the engine low-pressure rotor speed, and the main fuel flow closed-loop controls the engine thrust. When engine thrust tends to increase during mode switching, the thrust closed-loop control loop will reduce the main fuel flow. This reduction in main fuel flow will simultaneously decrease the low-pressure rotor speed. At this point, the low-pressure rotor speed closed-loop control loop can suppress this decrease by slightly increasing the nozzle throat area. Therefore, a dual-loop closed-loop mode switching control scheme is designed, using nozzle throat area closed-loop control for low-pressure rotor speed and main fuel flow closed-loop control for engine thrust. The control scheme is as follows: Figure 5 As shown.

[0066] Step B2.2) employs a dual-loop control scheme: closed-loop control of the nozzle throat area to regulate the low-pressure rotor speed, and closed-loop control of the main fuel flow to regulate engine thrust. Other adjustable parameters are controlled in an open-loop manner, linearly transitioning from the current mode design point to the target mode design point. To optimize thrust performance, the dual-loop closed-loop control system does not simply set the target speed and thrust constant. Analysis shows that if the control speed remains constant during mode switching, the engine thrust changes significantly. When switching from a single-bypass to a double-bypass system, the thrust decreases dramatically. If the low-pressure rotor speed remains constant, the thrust fluctuates greatly during the decrease. Therefore, it is necessary to appropriately increase the commanded low-pressure rotor speed. At this time, the nozzle throat area changes more, and the thrust increases slightly. Only a small adjustment to the fuel flow is needed to maintain constant thrust. When switching from a double-bypass to a single-bypass system, the thrust increases slightly. Therefore, only a slight decrease in the commanded low-pressure rotor speed is needed to maintain small thrust fluctuations. This method takes thrust performance as the optimization target. When switching from a single bypass to a double bypass, it designs a control law that keeps the engine thrust constant and increases the low-pressure rotor speed. When switching from a double bypass to a single bypass, it designs a control law that keeps the engine thrust constant and decreases the low-pressure rotor speed.

[0067] Step B2.3): At a ground point with H = 0 km and Ma = 0, the simulation demonstrates the switching between two modes of the variable cycle engine: single-bypass, double-bypass, and single-bypass. During the mode switching process, all other adjustable parameters linearly and gradually change from the current mode design point to the target mode design point over a time interval of 3 seconds, and all are open-loop inputs. The simulation results show that using the variable speed-thrust dual-loop closed-loop mode switching control method, the engine thrust fluctuation during the mode switching process is small, the fuel consumption rate does not increase, and the surge margin does not decrease to the critical value.

[0068] To verify the effectiveness of the variable cycle engine closed-loop mode switching control method designed in this invention, a digital simulation experiment of variable cycle engine mode switching control was designed in the MATLAB environment.

[0069] First, after constructing the tailpipe neural network model, the engine hybrid dynamics model was tested, with each component running 10,000 times. The time consumption results are as follows: Figure 6 As shown. In Figure 6 In (a), the time consumed by the tailpipe, which was originally the longest among all components, has been reduced to only 6.17% of the total calculated time of the air path, after the intake manifold and combustion chamber. This represents a 14.8% reduction in the total calculated time of the air path compared to the original calculation. Figure 6 In (b), the tail nozzle is remodeled using a neural network. The aerodynamic thermodynamic tail nozzle model takes about 0.746s to calculate, while the neural network tail nozzle model takes about 0.195s, saving 73.86% of the time. This method has a significant effect on reducing the time consumption of the tail nozzle.

[0070] At a ground point with H=0km and Ma=0, the simulation demonstrates the switching between two modes of the variable cycle engine: single-pass, double-pass, and single-pass. During the mode switching process, all other adjustable parameters change linearly from the current mode design point to the target mode design point in 3s, and all are open-loop inputs.

[0071] W during the single-culvert to double-culvert switching process fb Changes with A8, such as Figure 7 As shown, the simulation results are as follows: Figure 8 As shown, to minimize engine thrust fluctuations, the target speed of the low-pressure rotor is altered and increased during the switching process. During this process, the fuel flow rate initially decreases sharply and then gradually increases back to its initial value until stabilization. To ensure the speed doesn't decrease, the nozzle throat area initially increases and then gradually decreases until stabilization. However, because the overall target speed of the low-pressure rotor increases, the nozzle throat area generally still shows an increasing trend. After the switching is complete, the engine fuel consumption rate decreases, and the surge margin meets the requirements.

[0072] W during the switching from a double culvert to a single culvert fb Changes with A8, such as Figure 9 The simulation results are as follows Figure 10 As shown, the target speed of the low-pressure rotor is changed to slightly decrease during the switching process. During the switching process, the fuel flow gradually decreases to a stable level. In order to ensure that the speed remains at the commanded speed, the nozzle throat area first increases and then gradually decreases to a stable level. Overall, the nozzle throat area still has an increasing trend, and the thrust fluctuation is small throughout the entire mode switching process.

[0073] The simulation analysis above shows that during the mode switching process, the engine thrust fluctuates relatively little in the closed-loop control loop, the fuel consumption rate does not increase, and the surge margin does not decrease to the critical value. The variable speed-thrust dual-loop closed-loop control used in this paper exhibits small parameter fluctuations during the variable cycle engine mode switching, and its performance meets the expected goals.

[0074] It should be noted that the above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations and substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A dual-loop closed-loop mode switching control method for a variable cycle engine, characterized in that, Includes the following steps: Step A) Based on the aero-thermodynamic mechanism analysis, determine the parameters of the tail nozzle network model and construct a tail nozzle data-driven network model based on BP neural network; combine the tail nozzle neural network model with the component-level dynamic model of the main components of the upstream flow path to obtain a hybrid dynamic model of the variable cycle engine. Step B) With thrust performance as the optimization objective, design a dual-loop closed-loop control scheme for variable speed and thrust, and design a mode switching control method for the model based on the hybrid dynamic model of the variable cycle engine. The parameters of the exhaust nozzle network model include the following output parameters: exhaust nozzle outlet flow rate W9, exhaust nozzle outlet pressure P9, exhaust nozzle outlet temperature T9, exhaust nozzle outlet airflow velocity V9, thrust Fn, and fuel consumption rate sfc. The input parameters include: fuel quantity W. fb Throat area A8, nozzle exit area A9, inlet static pressure Ps0, inlet air velocity v0, nozzle inlet flow rate W7, nozzle inlet pressure P7 and nozzle inlet temperature T7, airflow rate W cool7 External bypass outlet pressure P 16 Temperature T at the duct outlet 16 ; The data-driven network model for the tail nozzle based on the BP neural network was trained using the LM optimization algorithm with 12 hidden neurons. An offline computational model for the tail nozzle was constructed by combining the training data. The hybrid dynamic model sequentially calls each component model for calculation. The flow path calculation order is consistent with the component-level model. When calling the tail nozzle model for calculation, the tail nozzle aerodynamic thermodynamic model is replaced with the tail nozzle neural network model. The input obtained from the upstream flow path calculation is used to obtain the output through the offline neural network tail nozzle model. The nonlinear common working equations that need to be solved are established according to the constraints of the components. When switching from a single-shunting to a double-shunting system, the control law is designed to keep the target thrust of the engine constant and increase the target speed of the low-pressure rotor. When switching from a double-shunting to a single-shunting system, the control law is designed to keep the target thrust of the engine constant and decrease the target speed of the low-pressure rotor.

2. The dual-loop closed-loop mode switching control method for a variable cycle engine as described in claim 1, characterized in that, During the closed-loop control mode switching process, the controlled variables are the nozzle throat area and fuel flow, while the controlled variables are the low-pressure rotor speed and engine thrust.

3. The dual-loop closed-loop mode switching control method for a variable cycle engine as described in claim 1, characterized in that, The design incorporates a control scheme that uses closed-loop control of the exhaust nozzle throat area to control the low-pressure rotor speed, closed-loop control of the main fuel flow to control the engine thrust, and open-loop control of other adjustable parameters.

Images