A method and system for aero-engine based pipe model predictive control

Abstract

The application provides an aero-engine pipeline model predictive control method and system, and belongs to the field of high-performance control of aero-engines. The method comprises the following steps: at the moment when an event is triggered, a pipeline model predictive control is used to solve a control problem of an aero-engine, Np control inputs are obtained, and a counter q is set to zero; in the case that the counter q is less than Np, a sum of a qth control input in the Nq control inputs and a feedback control input is used as an actual control input to control the aero-engine; and in the case that the counter q is equal to or greater than Np, a linear quadratic regulator is used to calculate a regulator control input, and the aero-engine is controlled by using the regulator control input. The method effectively reduces the online calculation burden.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of high-performance control of aero-engines, and particularly relates to a pipeline model predictive control method and system based on an aero-engine. BACKGROUND

[0002] Aero-engines are key components in the field of aviation, directly affecting the performance, safety and reliability of aircraft, and play a crucial role in the field of aviation. The performance of an aircraft, including its speed, acceleration and fuel efficiency, is closely related to the ability of the aero-engine that powers it. The thrust generated by the engine determines the ability of the aircraft to climb, cruise at optimal speed and maintain stable flight characteristics; aviation safety is of great importance, and the reliability of the aero-engine is crucial to ensure the safe operation of the aircraft throughout its life; in addition, the reliability of the aero-engine directly affects the overall operational reliability of the aircraft, affecting its ability to adhere to flight schedules and meet passenger expectations. However, the dynamic system of the aero-engine is affected by external factors such as inlet duct wind speed, ambient temperature, air pressure, inlet duct conditions, etc., resulting in challenging aero-engine control problems. This challenge includes dealing with complex dynamics, uncertainties, and constraints of environmental changes and system limitations. Therefore, developing advanced and efficient aero-engine control methods is the key to solving these problems.

[0003] In the past, aero-engine control typically relied on rule-based strategies such as proportional-integral-derivative (PID) controllers and linear quadratic regulators (LQRs). While these methods can provide good performance in certain situations, they often struggle to effectively handle complex nonlinear dynamics, uncertainties, and multiple constraints. Model predictive control (MPC) has become a popular control method for aero-engines due to its ability to handle various constraints in complex multi-input multi-output systems. MPC uses a mathematical model of the system to predict its future behavior and computes the optimal control input that minimizes a predefined cost function. The computed control input is then applied to the system, and the process is repeated in a back-to-time domain manner, where the optimization problem is solved at each time step. This iterative process allows MPC to proactively consider system dynamics, constraints, and objectives, effectively controlling the system's behavior. Therefore, MPC has been widely applied in modern industrial fields, including smart grids, transportation systems, and aero-engines. However, due to the presence of model uncertainties, the consequences of using traditional MPC methods may include suboptimal performance and reduced robustness. To address model uncertainties, current research has developed robust stochastic model predictive control strategies. However, traditional robust model predictive control requires iterative online solutions to optimization problems, resulting in poor real-time performance, which limits their implementation in aero-engine control. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this application proposes a pipeline model predictive control method and system based on aero-engines. This method solves the problem of reducing the computational burden of robust MPC in the presence of noise, while ensuring safety and robustness. By employing an event-triggered mechanism, the number of robust MPC calculations is reduced, and the control input calculated by robust MPC is fully utilized, thus resolving the enormous computational burden caused by the traditional online iterative solution of optimization problems in robust model predictive control.

[0005] Firstly, this application proposes a pipeline model predictive control method based on aero-engines, comprising:

[0006] Step S1: Establish the nonlinear dynamic equations of the aero-engine, and linearize the nonlinear equations of the aero-engine at the steady state point to obtain the state-space model of the aero-engine. Based on the state-space model, obtain the control problem of the aero-engine.

[0007] Step S2: Initialize counter q to zero;

[0008] Step S3: At the current moment, determine whether the pre-constructed event has been triggered. The pre-constructed event is constructed from the control problem of the aero-engine.

[0009] Step S4: At the moment the event is triggered, use pipeline model predictive control to solve the control problem of the aero-engine, and obtain N. p One control input is given, and the counter q is set to zero. Proceed to step S6.

[0010] Step S5: If the event is not triggered, proceed to step S6;

[0011] Step S6: When the counter q is less than N p In the case of N, q The sum of the q-th control input and the feedback control input is used as the actual control input to control the aero-engine, and q = q + 1. Return to step S3 until the counter q is equal to or greater than N. p In this case, proceed to step S7;

[0012] Step S7: Using a linear quadratic regulator, calculate the regulator control input, use the regulator control input to control the aero-engine, and q = q + 1, then return to step S3.

[0013] Based on the state-space model, the control problem of the aero-engine is obtained, and the calculation formula is as follows:

[0014]

[0015] ste(k) = x(k) - r(k)

[0016] Where r(k) is the target reference value of the control problem at time k, x(k) is the system state vector of the aero-engine at time k, e(k) is the difference between the system state vector and the target reference value of the control problem at time k, u(k) is the control input at time k, J is the output of the control problem of the aero-engine, Q is the weight of the first quadratic form, R is the weight of the second quadratic form, T is the matrix transpose, and Q≥0, R≥0.

[0017] The pre-constructed events are derived from the control problem of an aero-engine, and the calculation formula is as follows:

[0018] G={‖x(k)-r(k)‖2≥α∪‖r(k)-r(k-1)‖2>0}

[0019] Where G is a pre-constructed event, α is a preset threshold, r(k) is the target reference value of the control problem at time k, x(k) is the system state vector of the aero-engine at time k, and r(k-1) is the target reference value of the control problem at time k-1.

[0020] The calculation process for the feedback control input is as follows:

[0021] Calculate the difference between the system state vector of the aero-engine at time k and the predicted system state of the aero-engine at time k, wherein the predicted system state of the aero-engine at time k is obtained by pipeline model predictive control.

[0022] The product of the difference and the optimal gain is used as the feedback control input.

[0023] The optimal gain is calculated as follows:

[0024] K = -(B T PB+R) -1 B T PA

[0025] P = Q + A T PA-A T PB(B T PB+R) -1 B T PA

[0026] Where K is the optimal gain, Q is the first quadratic form weight, R is the second quadratic form weight, A is the system state matrix, B is the input matrix, and P is the solution to the discrete-time algebraic Riccati equation.

[0027] The linear quadratic regulator is used, and the regulator control input is calculated as follows:

[0028] u′(k)=K a x a (k)

[0029]

[0030] Where u′(k) is the regulator control input, K a For feedback control gain, A a To augment the system state matrix, B a To augment the system input matrix, P a To augment the solution of the discrete-time algebraic Riccati equation of the system, R is the weight of the second quadratic form, x a (k) represents the augmented system state at time k.

[0031] Secondly, this application proposes a pipeline model predictive control system based on an aero-engine, comprising:

[0032] The control problem establishment module is used to establish the nonlinear dynamic equations of the aero-engine, and linearize the nonlinear equations of the aero-engine at the steady state point to obtain the state-space model of the aero-engine. Based on the state-space model, the control problem of the aero-engine is obtained.

[0033] The initialization module is used to initialize the counter q to zero;

[0034] The event triggering module is used to determine whether a pre-constructed event has been triggered at the current moment. The pre-constructed event is derived from the control problem of the aero-engine. At the moment the event is triggered, predictive control using a pipeline model is employed to solve the control problem of the aero-engine, yielding N. p When a control input is received and the counter q is set to zero, the process switches to the first control module; if no event is triggered, the process switches to the first control module.

[0035] The first control module is used when the counter q is less than N. p In the case of N, q The sum of the q-th control input and the feedback control input is used as the actual control input to control the aero-engine, and q = q + 1. The process then returns to the event triggering module until the counter q equals or exceeds N. p In this case, switch to the second control module;

[0036] The second control module is used to calculate the controller control input using a linear quadratic controller, control the aero-engine using the controller control input, and set q = q + 1 before returning to the event triggering module.

[0037] Thirdly, this application proposes an electronic device comprising: one or more processors, and a memory for storing instructions that, when executed by the one or more processors, cause the one or more processors to execute the aforementioned pipeline model predictive control method based on an aero-engine.

[0038] Fourthly, this application proposes a computer-readable storage medium storing executable instructions that, when executed, cause a processor to perform the aforementioned pipeline model predictive control method based on an aero-engine.

[0039] Fifthly, this application proposes a computer program product, including a computer program or instructions, which, when executed by a processor, implements the aforementioned pipeline model predictive control method based on an aero-engine.

[0040] Beneficial effects:

[0041] This application proposes a pipeline model predictive control method and system based on aero-engines, effectively reducing the online computational burden. First, when a safety constraint violation is possible, an event mechanism is triggered, and pipeline model predictive control is executed to obtain control input. Second, unlike traditional methods that solve the optimization problem once per step and only use the first control input, this application's method fully utilizes all control inputs obtained from solving the optimization problem. Finally, to further reduce computational burden, LQR tracking control is applied when the control input calculated using pipeline model predictive control is exhausted. Compared to pipeline model predictive control, LQR tracking control significantly reduces the online computational burden. Attached Figure Description

[0042] Figure 1 A flowchart of a pipeline model predictive control method based on an aero-engine according to an embodiment of this application;

[0043] Figure 2 A schematic flowchart of a pipeline model predictive control method based on an aero-engine according to an embodiment of this application;

[0044] Figure 3 Comparison of target reference values ​​for high-voltage rotor speed tracking under different methods provided in the embodiments of this application;

[0045] Figure 4 Comparison of target reference values ​​for low-pressure rotor speed tracking under different methods in the embodiments of this application;

[0046] Figure 5 The consumption of engine fuel flow under different methods in the embodiments of this application;

[0047] Figure 6 The variation of engine nozzle area under different methods in the embodiments of this application. Detailed Implementation

[0048] The specific implementation methods of this application will be further described in detail below with reference to the accompanying drawings and embodiments.

[0049] This embodiment takes a twin-shaft turbojet engine as the research object and uses the efficient pipeline model predictive control method for safe operation of aero-engines proposed in this invention. The aim is to reduce computational burden and meet the real-time requirements of high-performance aero-engines. The method includes the design of an event-triggered mechanism and methods for tracking target reference values ​​under different conditions. The event-triggered mechanism reduces the use of computationally complex pipeline model predictive control. Simultaneously, it fully utilizes every control input obtained from the pipeline model predictive control solution. For example... Figure 1As shown, first initialize the parameters and the parameters required for calculating the optimization problem, and detect whether event G is triggered. If event G occurs, solve the control input and the nominal state through the pipeline model predictive control method to prepare for the subsequent calculation of the control input, and set q = 0. Next, determine which control input calculation method to use according to the value of q. If {q < N}_p, use the pipeline model predictive control to calculate the control input; otherwise, use the LQR tracking control method to calculate the control input. Finally, set q = q + 1 and re-determine whether event G occurs.

[0050] Example 1:

[0051] This example proposes a pipeline model predictive control method based on an aeroengine, as Figure 1 、 Figure 2 shown, including:

[0052] Step S1: Establish the nonlinear dynamic equation of the aeroengine, linearize the aeroengine nonlinear equation at the steady state point to obtain the state space model of the aeroengine, and obtain the control problem of the aeroengine according to the state space model;

[0053] Step S1.1: Take the twin-spool turbojet engine as the research object, and establish the nonlinear dynamic equation of the remaining torque of the high- and low-pressure rotors of the aeroengine according to the engine principle:

[0054] ΔM H =M TH -M CH -M fr,H =ΔM H (n H ,n L ,q m,f ,A8), (1)

[0055] ΔM L =M TL -M CL -M fr,L =ΔM L (n H ,n L ,q m,f ,A8), (2)

[0056] Among them, ΔM H represents the remaining torque of the high-pressure rotor in the engine, ΔM L represents the remaining torque of the low-pressure rotor, M TH 、M TL [[ID=##]]respectively represent the high- and low-pressure compressor torques, M CH 、M CL respectively represent the high- and low-pressure turbine torques, M fr，H ,M fr，Ln represents the frictional torque of the high-pressure and low-pressure rotors, respectively. H n L q represents the high-pressure and low-pressure rotor speeds, respectively. m，f A8 indicates the fuel supply amount, and A8 indicates the tail nozzle area.

[0057] Step S1.2: Linearize the nonlinear dynamic equations of the high and low pressure residual torque of the aero-engine to obtain the following linearized model:

[0058]

[0059] Substituting equations (3) and (4) into the following engine rotor torque equation:

[0060]

[0061] A linearized model of the aero-engine is obtained:

[0062]

[0063] in, The derivative of the high-voltage rotor speed. This is the derivative of the low-pressure rotor speed.

[0064] Step S1.3: By rearranging formulas (6) and (7), the state-space model of the aero-engine can be obtained as follows:

[0065]

[0066] make Given the system state vector, the discrete-time state-space model of an aero-engine can be expressed as:

[0067] x(k+1)=Ax(k)+Bu(k)+w(k) (9)

[0068] in, w(k) = [w1w2] T ∈W represents a bounded perturbation including the origin. The system is subject to the following state and control constraints:

[0069] Fx(k)+Gu(k)≤1, (10)

[0070] Step S1.4: The control objective of the aero-engine is to track the system state to the target reference value while satisfying system safety constraints. Therefore, the control problem can be expressed as follows:

[0071]

[0072] Where r(k) is the target reference value of the control problem at time k, x(k) is the system state vector of the aero-engine at time k, e(k) is the difference between the system state vector and the target reference value of the control problem at time k, u(k) is the control input at time k, J is the output of the control problem of the aero-engine, Q is the weight of the first quadratic form, R is the weight of the second quadratic form, T is the matrix transpose, and Q≥0, R≥0.

[0073] Step S2: Initialize counter q to zero;

[0074] Step S3: At the current moment, determine whether the pre-constructed event has been triggered. The pre-constructed event is constructed from the control problem of the aero-engine.

[0075] The pre-constructed events are derived from the control problem of an aero-engine, and the calculation formula is as follows:

[0076] G={‖x(k)-r(k)‖2≥α∪‖r(k)-r(k-1)‖2>0}

[0077] Where G represents a pre-constructed event, and α is a preset threshold, which is generally relatively small.

[0078] To alleviate the computational burden, an event-triggered mechanism is designed. If event G occurs, it indicates that the system state deviates significantly from the target reference value (exceeding the set threshold α) or that the target reference value has changed.

[0079] Step S4: At the moment the event is triggered, use pipeline model predictive control to solve the control problem of the aero-engine, and obtain N. p One control input is given, and the counter q is set to zero. Proceed to step S6.

[0080] Step S5: If the event is not triggered, proceed to step S6;

[0081] In this embodiment, at the moment the event is triggered, a pipeline model predictive control is used to solve the control problem of the aero-engine, resulting in N. p One control input, using pipeline model predictive control to calculate N. p There are one control input, and the calculation process is as follows:

[0082] First, define a set. As s→∞, F s →Z (Z is the smallest robust positive invariant set). At t=t i At time t, determine whether event G has occurred. If event G has occurred, solve the following optimization problem, while setting the counter value q = 0:

[0083]

[0084] Where, N p To control the time domain, For tightening state constraints, define as follows: (X represents the original state constraints). For tightening input constraints, define as follows: (U represents the original input constraints). V f For terminal costs, X f Let N be the terminal constraint set. After solving the optimization problem (13), N can be obtained. p One control input and N p Nominal status t i For time i, Let i be the nominal system state at time i. This is the nominal initial state of the system. The nominal system tracking error, For the (i+N)th p The nominal system state at any given moment.

[0085] Step S6: When the counter q is less than N p In the case of N, q The sum of the q-th control input and the feedback control input is used as the actual control input to control the aero-engine, and q = q + 1. Return to step S3 until the counter q is equal to or greater than N. p In this case, proceed to step S7;

[0086] In this embodiment, different methods are used to calculate the control input based on the value of the counter q. Pipeline model predictive control is used to calculate the control input. To improve the computational efficiency of pipeline model predictive control, a simple feedback control is introduced to reduce the computational burden. Before designing the feedback control, the optimal gain K is first defined by the algebraic Riccati equation (ARE):

[0087] P = Q + A T PA-A T PB(B T PB+R) -1 B T PA, (14)

[0088] K = -(B T PB+R) -1 B T PA.

[0089] Where K is the optimal gain, Q is the first quadratic form weight, R is the second quadratic form weight, A is the system state matrix, B is the input matrix, and P is the solution to the discrete-time algebraic Riccati equation.

[0090] Design Feedback Control Implement control input

[0091] in, For N q The q-th control input in the set of control inputs, u1(k), is the feedback control input. The feedback control input calculation process includes: calculating the difference between the system state vector of the aero-engine at time k and the predicted system state of the aero-engine at time k, wherein the predicted system state of the aero-engine at time k is obtained by pipeline model predictive control.

[0092] The product of the difference and the optimal gain is used as the feedback control input.

[0093] Step S7: Using a linear quadratic regulator, calculate the regulator control input, use the regulator control input to control the aero-engine, and q = q + 1, then return to step S3.

[0094] In this embodiment, a linear quadratic regulator (LQR) tracking control method is used to calculate the control input.

[0095] When the target reference value is constant, r(k+1) = Ir(k), x(k+1) = Ax(k) + Bu(k). Combining the above equations, we get:

[0096]

[0097] It can be abbreviated as x a (k+1)=A a x a (k)+B a u(k). The difference between the actual state and the target reference value can be expressed as follows:

[0098]

[0099] therefore, Use C T QC replaces Q, A a Replace A and B a Replacing B yields the feedback control gain. Implement control input u′(k)=K a x a (k). Where u′(k) is the regulator control input, K aFor feedback control gain, A a To augment the system state matrix, B a To augment the system input matrix, P a To augment the solution of the discrete-time algebraic Riccati equation for the system, I is the identity matrix, C is the output matrix, and x a (k) represents the augmented system state at time k.

[0100] To verify the effectiveness of the method of this invention, four target reference values ​​were designed for the system to track. The four selected target reference values ​​are close to the linearization point, resulting in negligible changes in the linearization model. The entire simulation time is 124 time steps, with each time step corresponding to 0.01 seconds. To verify the computational efficiency and tracking capability of the method of this application, four simulation experiments were conducted using LQR tracking control, pipeline model predictive control, and the method of this invention, respectively.

[0101] exist Figure 3 , Figure 4 In the simulation, the tracking behavior of two components of the system state, the high-pressure rotor speed and the low-pressure rotor speed, can be observed, respectively, for the three methods mentioned above. Clearly, all three methods successfully completed the objective task and tracked the target reference value. However, the tracking performance differs slightly. When tracking the reference value, LQR exhibits a small steady-state error. This means that although LQR can reach the target reference value, it may not match the precise expected value. On the other hand, both pipeline model predictive control and the method of this invention improve tracking accuracy by effectively reducing steady-state error. These methods show better performance, bringing the system closer to the target reference value throughout the simulation.

[0102] Figure 5 , Figure 6 Engine fuel flow and engine nozzle area were compared under four conditions. It is noteworthy that the control input generated by the LQR controller becomes excessively large at certain points, exceeding the defined constraint range. This makes the LQR controller unsafe for tracking the target reference value. On the other hand, the input curves of the method of this invention and conventional pipe model predictive control are almost indistinguishable; both the method of this invention and conventional pipe model predictive controllers exhibit bounded inputs that satisfy the constraints.

[0103] As shown in Table 1, LQR has a short computation time but cannot handle constraints, making the system unsafe. Traditional pipeline model predictive control has a longer computation time, reaching 2.39 seconds. In contrast, the entire process only takes 1.24 seconds (the discrete time of a typical engine is 0.01s; the simulation experiment in this application ran 124 time steps, each step being 0.01s, representing a total process time of 1.24 seconds). This demonstrates that the computation time of traditional pipeline model predictive control is too long and unsuitable for the experimental subject. However, the method in this application achieves a trade-off between LQR and traditional pipeline model predictive control, with a shorter computation time than traditional pipeline model predictive control, and the ability to handle constraints, ensuring the safety of the system.

[0104] Table 1: Comparison of Control Results of Various Methods

[0105]

[0106] Secondly, this application proposes a pipeline model predictive control system based on aero-engines, including: a control problem establishment module, an initialization module, an event triggering module, a first control module, and a second control module;

[0107] The control problem establishment module is connected to the initialization module, the initialization module is connected to the event triggering module, the event triggering module is connected to the first control module, and the first control module is connected to the second control module;

[0108] The control problem establishment module is used to establish the nonlinear dynamic equations of the aero-engine, and linearize the nonlinear equations of the aero-engine at the steady state point to obtain the state-space model of the aero-engine. Based on the state-space model, the control problem of the aero-engine is obtained.

[0109] The event triggering module is used to determine whether a pre-constructed event has been triggered at the current moment. The pre-constructed event is derived from the control problem of the aero-engine. At the moment the event is triggered, predictive control using a pipeline model is employed to solve the control problem of the aero-engine, yielding N. p When a control input is received and the counter q is set to zero, the process switches to the first control module; if no event is triggered, the process switches to the first control module.

[0110] The first control module is used when the counter q is less than N. p In the case of N, q The sum of the q-th control input and the feedback control input is used as the actual control input to control the aero-engine, and q = q + 1. The process then returns to the event triggering module until the counter q equals or exceeds N. p In this case, switch to the second control module;

[0111] The second control module is used to calculate the controller control input using a linear quadratic controller, control the aero-engine using the controller control input, and set q = q + 1 before returning to the event triggering module.

[0112] Example 3:

[0113] This embodiment proposes an electronic device, including: one or more processors, and a memory, wherein the memory is used to store instructions, and when the instructions are executed by the one or more processors, the one or more processors execute the aforementioned pipeline model predictive control method based on aero-engines.

[0114] The electronic device may be a mobile phone, computer, or tablet computer, etc., and includes a memory and a processor. The memory stores a computer program, which, when executed by the processor, implements a pipeline model predictive control method based on an aero-engine as described in the embodiments. It is understood that the electronic device may also include an input / output (I / O) interface and communication components.

[0115] The processor is used to execute all or part of the steps in the pipeline model predictive control method based on an aero-engine as described in the above embodiments. The memory is used to store various types of data, which may include, for example, instructions for any application or method in the electronic device, as well as application-related data.

[0116] The processor can be implemented as an Application Specific Integrated Circuit (ASIC), Digital Signal Processor (DSP), Programmable Logic Device (PLD), Field Programmable Gate Array (FPGA), controller, microcontroller, microprocessor, or other electronic components, and is used to execute the pipeline model predictive control method based on aero-engine described in the above embodiments.

[0117] Example 4:

[0118] This embodiment proposes a computer-readable storage medium that stores executable instructions. When these instructions are executed, if they are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium.

[0119] The computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the pipeline model predictive control method based on an aero-engine described in various embodiments of this application.

[0120] The aforementioned storage media include: flash memory, hard disk, multimedia card, card-type memory (e.g., SD (Secure Digital Memory Card) or DX (Memory Data Register, MDR) memory, random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, disk, optical disk, server, APP (Application) application store, and other media capable of storing program verification codes. These media store computer programs, which, when executed by a processor, can implement the various steps of the aforementioned pipeline model predictive control method based on aero-engines.

[0121] Example 5:

[0122] This embodiment proposes a computer program product, including a computer program or instructions, which, when executed by a processor, implements the aforementioned pipeline model predictive control method based on an aero-engine.

[0123] Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or part of the technical solution, can be embodied in the form of a computer program product.

[0124] The various embodiments in this application are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0125] The scope of protection of this application is not limited to the embodiments described above. Obviously, those skilled in the art can make various modifications and variations to this disclosure without departing from the scope and spirit of this disclosure. If such modifications and variations fall within the scope of the claims of this disclosure and their equivalents, then the intent of this disclosure also includes such modifications and variations.

Claims

1. A pipeline model predictive control method based on aero-engines, characterized in that, include: Step S1: Establish the nonlinear dynamic equations of the aero-engine, and linearize the nonlinear equations of the aero-engine at the steady state point to obtain the state-space model of the aero-engine. Based on the state-space model, obtain the control problem of the aero-engine. Step S2: Initialize counter q to zero; Step S3: At the current moment, determine whether the pre-constructed event has been triggered. The pre-constructed event is constructed from the control problem of the aero-engine. Step S4: At the moment the event is triggered, use pipeline model predictive control to solve the control problem of the aero-engine, and obtain N. p One control input is given, and the counter q is set to zero. Proceed to step S6. Step S5: If the event is not triggered, proceed to step S6; Step S6: When the counter q is less than N p In the case of N, q The sum of the q-th control input and the feedback control input is used as the actual control input to control the aero-engine, and q = q + 1. Return to step S3 until the counter q is equal to or greater than N. p In this case, proceed to step S7; Step S7: Using a linear quadratic regulator, calculate the regulator control input, use the regulator control input to control the aero-engine, and q = q + 1, then return to step S3.

2. The pipeline model predictive control method based on aero-engines according to claim 1, characterized in that, Based on the state-space model, the control problem of the aero-engine is obtained, and the calculation formula is as follows: ste(k) = x(k) - r(k) Where r(k) is the target reference value of the control problem at time k, x(k) is the system state vector of the aero-engine at time k, e(k) is the difference between the system state vector and the target reference value of the control problem at time k, u(k) is the control input at time k, J is the output of the control problem of the aero-engine, Q is the weight of the first quadratic form, R is the weight of the second quadratic form, T is the matrix transpose, and Q≥0, R≥0.

3. The pipeline model predictive control method based on aero-engines according to claim 1, characterized in that, The pre-constructed events are derived from the control problem of an aero-engine, and the calculation formula is as follows: G={‖x(k)-r(k)‖2≥α∪‖r(k)-r(k-1)‖2>0} Where G is a pre-constructed event, α is a preset threshold, r(k) is the target reference value of the control problem at time k, x(k) is the system state vector of the aero-engine at time k, and r(k-1) is the target reference value of the control problem at time k-1.

4. The pipeline model predictive control method based on aero-engines according to claim 1, characterized in that, The calculation process for the feedback control input is as follows: Calculate the difference between the system state vector of the aero-engine at time k and the predicted system state of the aero-engine at time k, wherein the predicted system state of the aero-engine at time k is obtained by pipeline model predictive control. The product of the difference and the optimal gain is used as the feedback control input.

5. The pipeline model predictive control method based on aero-engines according to claim 4, characterized in that, The optimal gain is calculated as follows: K=-(B T PB+R) -1 B T PA P=Q+A T PA-A T PB(B T PB+R) -1 B T PA Where K is the optimal gain, Q is the first quadratic form weight, R is the second quadratic form weight, A is the system state matrix, B is the input matrix, and P is the solution to the discrete-time algebraic Riccati equation.

6. The pipeline model predictive control method based on aero-engines according to claim 1, characterized in that, The linear quadratic regulator is used, and the regulator control input is calculated as follows: u′(k)=K a x a (k) Where u′(k) is the regulator control input, K a For feedback control gain, A a To augment the system state matrix, B a To augment the system input matrix, P a To augment the solution of the discrete-time algebraic Riccati equation of the system, R is the weight of the second quadratic form, x a (k) represents the augmented system state at time k.

7. A pipeline model predictive control system based on an aero-engine, characterized in that, include: The control problem establishment module is used to establish the nonlinear dynamic equations of the aero-engine, and linearize the nonlinear equations of the aero-engine at the steady state point to obtain the state-space model of the aero-engine. Based on the state-space model, the control problem of the aero-engine is obtained. The initialization module is used to initialize the counter q to zero; The event triggering module is used to determine at the current moment whether a pre-constructed event has been triggered. The pre-constructed event is constructed from the control problem of the aero-engine. At the moment the event is triggered, predictive control using a pipeline model is employed to solve the control problem of the aero-engine, yielding N. p When a control input is received and the counter q is set to zero, the process switches to the first control module; if no event is triggered, the process switches to the first control module. The first control module is used when the counter q is less than N. p In the case of N, q The sum of the q-th control input and the feedback control input is used as the actual control input to control the aero-engine, and q = q + 1. The process then returns to the event triggering module until the counter q equals or exceeds N. p In this case, switch to the second control module; The second control module is used to calculate the controller control input using a linear quadratic controller, control the aero-engine using the controller control input, and set q = q + 1 before returning to the event triggering module.

8. An electronic device, characterized in that, include: One or more processors, and a memory for storing instructions that, when executed by the one or more processors, cause the one or more processors to perform a pipeline model predictive control method based on an aero-engine as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, It stores executable instructions that, when executed, cause the processor to perform the pipeline model predictive control method based on an aero-engine as described in any one of claims 1 to 6.

10. A computer program product, comprising a computer program or instructions, characterized in that, When the computer program or instructions are executed by the processor, they implement the pipeline model predictive control method based on any one of claims 1 to 6.

Images