A design method for intelligent actuator nodes of aero-engine axisymmetric tail nozzle under distributed architecture
By designing intelligent nodes under a distributed architecture and using DSP and FPGA chips to control the axisymmetric vector nozzle of an aero-engine, efficient isochronous coordinated control and fault diagnosis are achieved, solving the complexity problem of centralized controllers and reducing the cost of vector nozzle deflection control.
Patent Information
- Application Number
- CN202310177767.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-02-28
- Publication Date
- 2024-11-12
- Estimated Expiration
- 2043-02-28
AI Technical Summary
The centralized controller of existing axisymmetric vector nozzles for aero engines is burdensome, and the software development and maintenance are complex, making it difficult to meet the needs of distributed control. In addition, the verification cost of the deflection control algorithm of the vector nozzle is high.
The design of an axisymmetric vector nozzle for aero-engines adopts a distributed architecture, using a DSP chip as the host controller and an FPGA chip as the bus controller. By combining hardware and software design, isochronous coordinated control of three electric actuators is achieved, and fault diagnosis is performed through a Kalman filter.
It improves the performance of the aero-engine control system, reduces the verification cost of advanced tail nozzle control algorithms, and achieves effective fault isolation and stable control under partial sensor failure.
Smart Images

Figure CN116149229B_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the field of aero-engine control technology, and relates to a system design method for an intelligent node controller for an axisymmetric vector nozzle of an aero-engine, and specifically to an overall structure, an actuator structure, and a control method of an intelligent node controller for an axisymmetric vector nozzle under a distributed architecture. Background Art
[0002] The aircraft engine vector nozzle is a multifunctional exhaust nozzle used for flight attitude control of aircraft, such as pitch, yaw, roll and reverse thrust. This technology changes the direction of the engine tail jet by controlling the deflection of the vector nozzle, making the aircraft have super maneuverability, high agility, short take-off and landing performance, stealth performance and supersonic cruise capability, which can greatly improve the combat capability and survivability of fighter jets. The vector nozzles currently used are mainly divided into baffle vector nozzles, binary vector nozzles and axisymmetric vector nozzles.
[0003] The axisymmetric vector nozzle uses an actuator to drive the convergent section or the diffuser section of the nozzle to deflect in the axial range to generate vector thrust. Currently, it is mainly divided into universal head type and actuator ring type. The actuator ring type vector nozzle completes vector control through three vector adjustment actuators at 120 degrees on the actuator ring vector nozzle. Compared with the axisymmetric adjustable-expanding nozzle of the existing engine, the actuator ring type axisymmetric vector nozzle has little structural change and a certain degree of inheritance. The deflection mechanism is simple in design, the thrust loss is small and it is easy to close.
[0004] The current engine control uses a Full Authority Digital Engine Control (FADEC) to provide full authority for engine control to achieve stable and transient engine characteristics. The central processing unit needs to complete all tasks of aircraft engine control, including the deflection control of the vector nozzle. The central processing unit has heavy tasks, the development and iteration of software are difficult, and the subsequent maintenance is also very complicated. With the continuous development of aircraft engine technology, people's control requirements are constantly increasing, among which the development from centralized control to distributed control is a trend. A controller under a distributed architecture is designed separately for the engine's axisymmetric vector nozzle to meet the mathematical calculations of the vector azimuth angle, deflection angle and vector actuator when the nozzle is flipped, as well as small closed-loop control. Summary of the invention
[0005] The present invention is used to provide a design scheme for an intelligent node of an aero-engine axisymmetric vector tail nozzle under a distributed architecture, and the node is verified by isochronous coordinated control of three-way electric actuators. Its control logic can be used for actual aero-engine tail nozzle control, which can greatly improve the performance of aero-engine control systems and reduce the verification cost of advanced tail nozzle control algorithms.
[0006] The present invention specifically adopts the following technical solutions to solve the above technical problems:
[0007] A design method for an intelligent actuator node of an aero-engine axisymmetric tail nozzle under a distributed architecture includes an intelligent actuator node controller, an actuator simulation system, and a control system design.
[0008] The intelligent actuator node controller selects a digital signal processing (DSP) chip as the main control chip of the host controller according to the mathematical operation performance requirements of the aircraft engine control, and selects a field programmable gate array (FPGA) chip as the chip of the bus controller according to the requirements of the distributed communication bus. The intelligent actuator node controller includes hardware design and software design.
[0009] The hardware design is divided according to function, including a tail nozzle bus controller, a host controller and a conditioning drive circuit, which are used to complete the data processing of the intelligent node, the calculation of the control algorithm, the output of the control quantity and the transmission of the host computer signal; in the distributed control architecture, the intelligent node needs to be mounted to the entire control system through the bus, and the intelligent node needs to establish consistent communication according to the communication protocol.
[0010] The tail nozzle bus controller includes a container network interface (CNI) and a physical layer driver module. The container network interface is used to communicate with the host controller, and the physical layer driver module is used for data communication with the communication bus.
[0011] The host controller includes a 32-bit CPU timer, an enhanced analog-to-digital converter module (ADC), a serial communication interface module, an enhanced pulse width modulation (EPWM) output module, a digital interface and a shared pin function module; the 32-bit CPU timer is used to provide precise timing to meet the high real-time control system; the enhanced analog-to-digital converter module is used to collect the voltage data output by the displacement sensor; the serial communication interface module is used to communicate with the host computer to realize the display of displacement data and the debugging of various intermediate variables; the enhanced pulse width modulation output module is used to output PWM data, and the output of the drive PWM of the electric actuator drive module is adjusted by adjusting the duty cycle and output frequency to realize the speed control of the electric actuator. The digital interface and shared pin function module are used to control the driving direction of the motor and the communication connection of the bus controller.
[0012] The conditioning drive circuit is the signal conditioning part of the entire intelligent node, including the electric actuator drive module and the sensor signal conditioning module. The electric actuator drive module drives the high-voltage, high-power electric actuator through the control signal output by the host controller; the sensor signal conditioning module is used to condition the ±10V signal output by the displacement sensor into a 0-3V signal, which is then used for data acquisition by the ADC module on the DSP chip board.
[0013] The software design adopts bare metal programming, including foreground program and background program. The foreground program is the interrupt service program, including timer interrupt, AD acquisition interrupt, host interface interrupt and serial port interrupt receiving host computer instructions. The background program is the main process operation, including data reading, processing and coordinated control algorithm execution, control quantity output and SCI serial port data transmission.
[0014] The actuator simulation system includes three electric actuators and a displacement sensor;
[0015] The electric actuator is used to adjust the tail nozzle outlet area and deflection angle. Preferably, the electric actuator includes a tail nozzle actuator and a matching actuator driver, which receives the actuator displacement command sent by the controller and feeds back the working status of the actuator. The actuator needs to have high displacement accuracy (≤2mm) and strong synchronization (≤2mm), and a servo motor is selected as the actuator of the precision motion actuator.
[0016] The displacement sensor is used to realize displacement feedback of the electric actuator, and the two are rigidly connected by fasteners. Preferably, in order to realize closed-loop control of the electric actuator, a linear variable differential transformer (LVDT) sensor is used. Its main principle is similar to that of a transformer, and the output electromotive force is proportional to the position of the internal iron core, so the displacement measurement data can be converted into measurement data of a voltage signal.
[0017] The control system design includes cooperative control and fault diagnosis. The cooperative control allows multiple electric actuators to reach the expected value at the same time to ensure the stability of the actuator. The fault diagnosis is used to avoid fault transmission problems under cooperative control.
[0018] Control system scheme: The control scheme of this invention is used for the control of axisymmetric vector nozzles, and its motion mechanism can be realized by the control of three actuators. The control of the three actuators is divided into three steps according to the idea of independent control to coordinated control and finally to fault diagnosis:
[0019] S1. When the engine state remains unchanged, the areas of A8 (i.e., throat area adjustment loop) and A9 (i.e., steering control loop) are fixed values. The outer loop control of the axisymmetric vector nozzle is based on the vector azimuth angle θ and vector deflection angle δ output by the flight control system and the areas of A8 and A9. The given values of the displacements of the three actuators are solved by the control law, and then the actuator is driven through the inner loop (small closed loop) to achieve A9 area adjustment and deflection control. Since the vector deflection angle and azimuth angle cannot be directly measured by the sensor, A9 control adopts open-loop control, that is, the control law solves the given displacement value of the actuator according to the vector azimuth angle θ and vector deflection angle δ, and then the movement of the actuator is achieved through small closed-loop control.
[0020] S2. By using the logic of isochronous cooperative control, the design integrates the characteristics of deviation coupling control and master-slave control. This solves the problem that when the A9 steering vector is deflected, the displacement output of the A9 ring driving the actuator is unequal, causing the actuator to reach the expected value at different times.
[0021] The variable deviation cooperative control based on deviation ratio is that the three actuator control loops each calculate the deviation between the given signal and the feedback signal (hereinafter referred to as deviation), and the loop with the largest absolute value of the original deviation signal of the three loops is used as the main loop; the deviations of other loops are proportionally reduced according to the ratio of the maximum deviation, that is, comprehensive coupling correction is realized, and the corrected deviation is calculated by the same controller to obtain a DC motor voltage value proportional to the deviation. The three DC motors thus realize variable speed movement, and the adjustment time is equal and determined by the main loop.
[0022] S3. Variable Deviation Cooperative Control Based on Deviation Ratio In the dynamic control process, if the displacement sensor has an inherent deviation fault or a sensor drift fault, the deviation value of the faulty loop quickly becomes the maximum deviation, and the cooperative following of the slave loop at this time will cause the fault of the single loop to be transmitted to the entire control system. The fault diagnosis method based on the Kalman filter can determine whether the displacement sensor of the path is faulty by calculating the weighted sum-squared residual (WSSR), thereby preventing the propagation of the fault in other loops.
[0023] Beneficial effects of the present invention: The intelligent actuator node of the axisymmetric tail nozzle of an aircraft engine under a distributed architecture designed by the present invention can well complete the communication with other intelligent nodes under the distributed architecture, and realize the isochronous coordinated control of multiple actuators of the axisymmetric tail nozzle system, and can also achieve effective fault isolation under partial sensor failure. BRIEF DESCRIPTION OF THE DRAWINGS
[0024] Figure 1 It is an overall implementation scheme of the axisymmetric tail nozzle control node controller under the distributed architecture of the present invention.
[0025] Figure 2 Implement a flow chart for the host controller software functionality.
[0026] Figure 3 This is a schematic diagram of the connection between the actuator and the displacement sensor.
[0027] Figure 4 This is the intelligent node control architecture diagram of the axisymmetric tail nozzle.
[0028] Figure 5 Schematic diagram of the cooperative control deviation integrated coupler.
[0029] In the figure: 1. fixing block, 2. fixing screw, 3. LVDT displacement sensor probe, 4. electric actuator telescopic rod, 5. electric actuator housing, 6. LVDT displacement sensor plastic fixture, 7. LVDT displacement sensor tube body. DETAILED DESCRIPTION
[0030] The present invention is further described below with reference to the accompanying drawings. As shown in the figure, the aircraft engine distributed tail nozzle intelligent node system mainly includes five parts: the overall implementation scheme of the axisymmetric tail nozzle controller, the host controller software flow, the actuator simulation, the node control architecture, and the coordinated control deviation comprehensive coupling.
[0031] The controller comprehensive implementation scheme is shown in FIG. Figure 1 From left to right, it is divided into three parts: conditioning drive circuit, tail nozzle host controller and tail nozzle bus controller. The hardware module is mainly used to illustrate the relationship between the hardware requirements and data communication under the various requirements of the tail nozzle intelligent control node. The following is a further detailed description of each part:
[0032] (1) Conditioning the drive circuit
[0033] The conditioning drive circuit is the signal conditioning part of the entire intelligent node, which includes an actuator drive module, which drives the high-voltage, high-power actuator through the control signal output by the host controller; the sensor signal conditioning module conditions the ±10V signal output by the displacement sensor into a 0-3V signal, which is then used for data acquisition by the ADC module on the DSP chip board.
[0034] (2) Nozzle main engine controller
[0035] The main functions of the tail nozzle host controller are divided into signal acquisition module, control calculation module and signal output module. The signal acquisition module includes: enhanced analog-to-digital converter module (ADC), which is used to collect voltage data output by the displacement sensor. The control calculation module specifically includes: 32-bit CPU timer, which is used to provide precise timing to meet the high real-time control system. The signal output module includes: serial communication interface module (SCI-A), which is used for communication with the host computer to realize the display of displacement data and the debugging of various intermediate variables; enhanced PWM output module (ePWM), which outputs PWM data, adjusts the output of the drive PWM of the actuator drive module by adjusting the duty cycle and output frequency, and realizes the speed control of the actuator. Digital IO and shared pin functions are used to control the driving direction of the motor.
[0036] (3) Nozzle bus controller
[0037] The tail nozzle bus controller includes CNI for communicating with the host controller, and a physical layer driver module for data communication with the communication bus.
[0038] Figure 2 This is a function flow chart of the host controller software. The host controller of this invention adopts bare metal programming, including foreground program and background program. The foreground program is an interrupt service program, including timer interrupt, AD acquisition terminal, host interface interrupt and serial port interrupt receiving host computer instructions. The background program is the main process operation, including data reading, processing and coordinated control algorithm execution, control quantity output and SCI serial port data transmission. The specific process is as follows:
[0039] S1: The processor enters a timer interrupt every 5ms to provide a time reference for the control system;
[0040] S2: Read serial port data to obtain control instructions;
[0041] S3: AD collects the voltage value of the displacement sensor to calculate the displacement data;
[0042] S4: Perform control calculation according to the control strategy;
[0043] S5: The numerical value obtained by the control calculation is converted into a PWM duty cycle to drive the displacement of the actuator;
[0044] S6: Use the SCI interface to transfer data to the host computer for data display and debugging, and enter the next control cycle.
[0045] Based on the above hardware and software foundation, the electric actuator is selected as the actuator of the tail nozzle. The movement of the three electric actuators completes the adjustment of the tail nozzle outlet area and the deflection angle. In order to achieve closed-loop control of the electric actuator, LVDT is used to achieve displacement feedback of the electric actuator, and the two are rigidly connected by fasteners. Figure 3 As shown, the electric actuator includes a tail nozzle actuator and a matching actuator driver, and the electric actuator telescopic rod 4 is connected to the LVDT displacement sensor probe 3 through a fixing screw 2 and a fixing block 1. The housing of the displacement sensor is fixed to the electric actuator housing 5 through a plastic clamp 6.
[0046] like Figure 4 The control scheme of the invention is used for the control of the axisymmetric vector nozzle, and its motion mechanism can be realized by the control of three actuators. The control of the three actuators is divided into three steps according to the idea of independent control to coordinated control and finally to fault diagnosis:
[0047] S1: When the engine state remains unchanged, the areas of A8 and A9 are constant values, and the AVEN outer loop control is based on the vector azimuth angle θ and vector deflection angle δ output by the flight control system and the areas of A8 and A9, such as Figure 4 The control law solving module calculates the given values of the displacement of the three actuators, and then drives the actuator through the inner loop (small closed loop) to achieve A9 area adjustment and deflection control. Since the vector deflection angle and azimuth angle cannot be directly measured by the sensor, A9 control usually adopts open-loop control, that is, the control law solves the given displacement value of the actuator according to the vector azimuth angle θ and the vector deflection angle δ, and then realizes the movement of the actuator through small closed-loop control.
[0048] S2: By using the logic of isochronous cooperative control, the design integrates the characteristics of deviation coupling control and master-slave control. It solves the problem that when the A9 steering vector is deflected, the displacement output of the A9 ring driving the actuator is unequal, resulting in the actuator inequality reaching the expected value. Variable deviation cooperative control based on deviation ratio Figure 5 The three actuator control loops calculate the deviation between the given signal and the feedback signal (hereinafter referred to as deviation) to obtain l1, l2, and l3. The loop with the largest absolute value of the original deviation signal of the three loops is taken as the main loop to obtain l max ; The deviations of other loops are proportionally corrected according to the proportional values k1, k2, k3 with the maximum deviation to finally obtain the corrected l1', l2', l3', thereby finally realizing the comprehensive coupling correction. After the corrected deviation is calculated by the same controller, the DC motor voltage value proportional to the deviation is obtained. The three DC motors thus realize variable speed movement, and the adjustment time is equal and determined by the main loop.
[0049] S3: Variable Deviation Cooperative Control Based on Deviation Proportion During the dynamic control process, if the displacement sensor has an inherent deviation fault or a sensor drift fault, the deviation value of the faulty loop quickly becomes the maximum deviation. The cooperative following of the slave loop at this time will cause the single loop fault to be transmitted to all control loops. The principle of the sensor fault diagnosis method based on the Kalman filter group is as follows: Figure 4 As shown in the figure, there are 3 sensors in total, and 3 Kalman filters are designed. Take the i-th Kalman filter as an example, its input y i The error includes the deviation of the remaining two measurable outputs except the ith sensor. Therefore, if the ith sensor fails, as long as the estimation result obtained by the ith Kalman filter is correct, the estimation results of the remaining Kalman filters will deviate from the actual situation to varying degrees because they use the information of the failed sensor.
[0050] The difference between the input and filtered output of the Kalman filter is called the residual. The residual e of the i-th Kalman filter is:
[0051]
[0052] Where: Δy i is the measurable output deviation, which is the input of the filter; is the measurable output deviation after filtering.
[0053] The weighted sum of squares of the residuals calculated from this residual is:
[0054] WSSR i =(e i ) T ·[diag(σ i ) 2 ] -1 ·e i
[0055] Where: i is the standard deviation of each sensor, which is a measure of the characteristics of the sensor itself and the measurement noise; the order is the number of samples collected. When a displacement sensor fails, the characteristics of the residual between the measurement parameters and the filter estimation output will change accordingly, and WSSRi will change significantly. Therefore, by monitoring the fault indicator variable WSSR i The changes in the sensor can be used to determine whether the corresponding sensor is faulty.
Claims
1. A design method for intelligent actuator nodes of an aero-engine axisymmetric tail nozzle under a distributed architecture, characterized in that: Including intelligent actuator node controller, actuator simulation system, and control system design; The intelligent actuator node controller selects a digital signal processing DSP chip as the main control chip of the host controller according to the mathematical operation performance requirements of the aircraft engine control, and selects a field programmable gate array FPGA chip as the chip of the bus controller according to the requirements of the distributed communication bus; the intelligent actuator node controller includes hardware design and software design; The hardware design is divided according to functions, including tail nozzle bus controller, host controller and conditioning drive circuit, which are used to complete the data processing of intelligent nodes, the calculation of control algorithms, the output of control quantities and the transmission of host computer signals; in the distributed control architecture, the intelligent nodes need to be mounted to the entire control system through the bus, and the intelligent nodes need to establish consistent communication according to the communication protocol; The tail nozzle bus controller includes a container network interface CNI and a physical layer driver module. The container network interface is used to communicate with the host controller, and the physical layer driver module is used for data communication with the communication bus. The host controller includes a 32-bit CPU timer, an enhanced analog-to-digital converter module ADC, a serial communication interface module, an enhanced pulse width modulation EPWM output module, a digital interface and a shared pin function module; the 32-bit CPU timer is used to provide precise timing to meet the needs of a high-real-time control system; the enhanced analog-to-digital converter module is used to collect voltage data output by the displacement sensor; the serial communication interface module is used to communicate with the host computer to display displacement data and debug various intermediate variables; the enhanced pulse width modulation output module is used to output PWM data, and the output of the drive PWM of the electric actuator drive module is adjusted by adjusting the duty cycle and the output frequency to achieve speed control of the electric actuator; the digital interface and shared pin function module are used to control the driving direction of the motor and the communication connection of the bus controller; The conditioning drive circuit is the signal conditioning part of the entire intelligent node, including the electric actuator drive module and the sensor signal conditioning module. The electric actuator drive module drives the high-voltage and high-power electric actuator through the control signal output by the host controller; the sensor signal conditioning module is used to convert the displacement sensor output The 10V signal is conditioned into a 0-3V signal for data acquisition by the ADC module on the DSP chip; The software design adopts bare metal programming, including foreground program and background program; the foreground program is the interrupt service program, including timer interrupt, AD acquisition interrupt, host interface interrupt and serial port interrupt receiving host computer instruction; the background program is the main process operation, including data reading, processing and coordinated control algorithm execution, control quantity output and SCI serial port data transmission; The actuator simulation system includes three electric actuators and a displacement sensor; The electric actuator is used to adjust the outlet area of the tail nozzle and the deflection angle; The displacement sensor is used to realize displacement feedback of the electric actuator, and the two are rigidly connected by fasteners; The control system design includes cooperative control and fault diagnosis. The cooperative control allows multiple electric actuators to reach the expected value at the same time to ensure the stability of the actuator. The fault diagnosis is used to avoid the fault transmission problem under the cooperative control. The control system scheme is as follows: S1. When the engine state remains unchanged, the areas of the throat area adjustment loop A8 and the steering control loop A9 are fixed values. The outer loop control of the axisymmetric vector nozzle calculates the given values of the displacements of the three actuators according to the vector azimuth angle θ and the vector deflection angle δ output by the flight control system and the areas of A8 and A9 by the control law, and then drives the actuator through the inner loop to achieve A9 area adjustment and deflection control; since the vector deflection angle and azimuth angle cannot be directly measured by the sensor, A9 control adopts open-loop control, that is, the control law calculates the given displacement value of the actuator according to the vector azimuth angle θ and the vector deflection angle δ, and then realizes the movement of the actuator through a small closed-loop control; S2. By using the logic of isochronous cooperative control, the design integrates the characteristics of deviation coupling control and master-slave control; it solves the problem that when the A9 steering vector is deflected, the displacement output of the A9 ring driving the actuator is unequal, resulting in the actuator reaching the expected value at different times; The variable deviation cooperative control based on deviation ratio is that the three actuator control loops calculate the deviation between the given signal and the feedback signal respectively, and the loop with the largest absolute value of the original deviation signal of the three loops is used as the main loop; the deviations of other loops are proportionally reduced according to the ratio of the maximum deviation, that is, comprehensive coupling correction is realized, and the corrected deviation is calculated by the same controller to obtain the DC motor voltage value proportional to the deviation, so that the three DC motors can realize variable speed movement, and the adjustment time is equal and determined by the main loop; S3. Variable deviation cooperative control based on deviation ratio. During the dynamic control process, if the displacement sensor has an inherent deviation fault or a sensor drift fault, the deviation value of the faulty loop will quickly become the maximum deviation. The cooperative following of the slave loop at this time will cause the single-loop fault to be transmitted to the entire control system. The fault diagnosis method based on the Kalman filter can determine whether the displacement sensor of this path has a fault by calculating the weighted sum of square residuals, thereby preventing the fault from propagating to other loops.
2. The design method of an intelligent actuator node of an aero-engine axisymmetric tail nozzle under a distributed architecture according to claim 1 is characterized in that: The electric actuator includes a tail nozzle actuator and a matching actuator driver, which receives the actuator displacement command sent by the controller and provides feedback on the working status of the actuator.
3. The design method of an intelligent actuator node of an aero-engine axisymmetric tail nozzle under a distributed architecture according to claim 1 is characterized in that: In order to realize the closed-loop control of the electric actuator, a linear variable differential transformer LVDT sensor is used. Its principle is similar to that of a transformer. The output electromotive force is proportional to the position of the internal iron core, so the displacement measurement data can be converted into voltage signal measurement data.
Citation Information
Patent Citations
Hardware-in-the-loop simulation platform and method for distributed fault diagnosis system of aero-engine
CN114779735A
Aeroengine h∞ control method based on improved smith predictor
WO2021097738A1
