Adaptive Intelligent Control System and Method for Aero-engines Based on Multimodal Perception

Abstract

This invention discloses an adaptive intelligent control system and method for aero-engines based on multimodal perception, including a command generator, an adaptive controller, a time disturbance observer, a multi-model sensor, a parameter estimator, and a data cloud platform. The command generator is used to generate dynamic command signals that meet the target requirements according to the system's set control objectives. The dynamic command signals have preset speed tracking trajectory characteristics to guide the engine speed adjustment. The adaptive controller is used to input the output control quantity of the adaptive specified disturbance cancellation composite controller, the state feedback value characterizing the aero-engine shaft speed, and the system output quantity. This invention can quickly and accurately describe and compensate for internal and external disturbance factors of the aero-engine system. Compared with traditional observers, it has a definite convergence time and better dynamic performance, thereby ensuring that the system can still maintain stable operation when facing complex disturbances.

Description

Technical Field

[0001] This invention relates to the field of adaptive intelligent control system technology, and in particular to an adaptive intelligent control system and method for aero-engines based on multimodal perception. Background Technology

[0002] Existing domestic and international aero-engine correction technologies mainly focus on component characteristic correction based on steady-state operating points. That is, for a given steady-state operating point, various solution methods are used to adjust the correction parameters. This type of correction process is, to some extent, an optimization process of steady-state model parameters. However, for the transient state of aero-engines, the commonly used methods are still based on steady-state operating point models and use interpolation algorithms to approximate the transient process. Due to the limited number of characteristic steady-state operating points and the large interpolation error, the dynamic response error of the model is large. The above-mentioned model correction methods are difficult to solve the accuracy problem of key output parameters in the transient process within the full envelope. In addition, the number of health parameters involved in aero-engine models is far greater than the number of key measurable parameters of aero-engines. In actual engineering applications, only health parameters with a number equivalent to the number of key measurable parameters are often selected for adaptive adjustment.

[0003] In Chinese patent 201910522225.6, only the faults of aero-engines are monitored. This approach ignores the positive effects that certain disturbances may have under specific conditions. When aero-engines are running under different operating conditions, certain external disturbances may have a positive effect on the closed-loop performance of the system, especially in terms of improving system response speed and reducing fuel consumption. Therefore, this invention proposes an adaptive intelligent control system and method for aero-engines based on multimodal perception to solve the problems existing in the prior art. Summary of the Invention

[0004] To address the aforementioned problems, the present invention aims to propose an adaptive intelligent control system and method for aero-engines based on multimodal perception. This adaptive intelligent control system and method for aero-engines based on multimodal perception can quickly and accurately describe and compensate for internal and external disturbances in the aero-engine system. Compared with traditional observers, it has a definite convergence time and better dynamic performance, thereby ensuring that the system can maintain stable operation when facing complex disturbances. It utilizes multiple sensors to collect multi-dimensional data of the engine, providing comprehensive engine status awareness.

[0005] To achieve the objectives of this invention, the invention is implemented through the following technical solution: an adaptive intelligent control system and method for aero-engines based on multimodal perception, including a command generator, an adaptive controller, a time disturbance observer, a multi-model sensor, a parameter estimator, and a data cloud platform;

[0006] The command generator is used to generate dynamic command signals that meet the target requirements according to the control objectives set by the system. The dynamic command signals have preset speed tracking trajectory characteristics to guide the speed adjustment of the engine.

[0007] The adaptive controller is used to input the output control quantity of the adaptive specified disturbance cancellation composite controller, the state feedback value characterizing the shaft speed of the aero-engine, and the system output quantity. Based on the state feedback information of the aero-engine and the output control command of the controller, the lumped disturbance is estimated in real time within a fixed time.

[0008] The time-perturbation observer achieves real-time observation and estimation of lumped disturbances over a fixed time period by extending state observation.

[0009] Multi-model sensors, including temperature sensors, pressure sensors, vibration sensors, and sound sensors, are used to monitor various physical and chemical parameters of the engine in real time;

[0010] The parameter estimator is used to observe the model parameters. The parameter estimator is equipped with three input terminals that are connected to the system input port, system output port and time disturbance observer of the aero-engine respectively. Its output terminal is connected to the adaptive controller to output the estimated values ​​of the model parameters.

[0011] The data cloud platform uploads the collected data to the cloud platform for storage and backup, performs remote analysis, and utilizes cloud computing resources for big data analysis and processing to collect and use engine information.

[0012] A further improvement is that the command generator transforms the decisions in the system into actual control actions, which are used to adjust fuel flow, change the angle of turbine blades, and regulate the speed of the compressor. The command generator dynamically adjusts the control parameters according to the real-time status of the engine and changes in the external environment.

[0013] The further improvement lies in the following: The adaptive controller combines the dynamic command aggregate disturbance estimate from the command generator with the state feedback information of the aero-engine, adjusts the control quantity in real time based on the backstepping method and the dynamic inverse method, and combines the adaptive specified disturbance cancellation method to generate adaptive control commands and output them to the aero-engine through real-time estimation and feedback compensation of aggregate disturbance.

[0014] A further improvement is that the time disturbance observer estimates and compensates for time-varying disturbances and uncertainties in the system. The disturbances include changes in the external environment, fluctuations in internal system parameters, and sensor noise. The time disturbance observer estimates the time-varying disturbance signal in the system in real time and feeds the estimated disturbance signal back to the control system, enabling the controller to dynamically adjust the control commands to compensate for the impact of the disturbance.

[0015] Further improvements include: the multi-model sensor provides comprehensive data input, enabling the system to monitor and analyze the engine status more accurately; data from different sensors can be cross-validated, improving the system's confidence in judging the engine status and reducing false alarms and missed alarms; different sensors have different sensitivities to environmental changes, and the combination of multi-model sensors can effectively monitor in different environments.

[0016] A further improvement is that the parameter estimator analyzes sensor data to calibrate the system model, ensuring that the model accurately reflects the engine's operating state. During system operation, the parameter estimator updates parameter values ​​in real time.

[0017] A further improvement is that the data cloud platform provides a central repository for storing various sensor data from aero engines, including temperature, pressure, vibration, and sound. Multimodal data enables secure storage, management, and retrieval of data. The data cloud platform processes and analyzes large amounts of sensor data in real time, providing immediate insights and feedback.

[0018] A further improvement is that the output of the instruction generator is connected to the input of the adaptive controller.

[0019] A further improvement is that the input of the fixed-time disturbance observer is connected to the output of the adaptive controller, the state feedback terminal of the downstream aero-engine, and the system output terminal, respectively.

[0020] Further improvements include the following steps:

[0021] Step 1: System modeling of the aero-engine is performed, input and output data are recorded, and a perturbation-based affine nonlinear data model is constructed using system identification methods.

[0022] Step 2: Generate instruction tracking through the model, adjust and arrange process steps to generate instructions that meet preset control indicators, and calculate the derivative of the instruction with respect to time to ensure that the dynamic response of the system under different working conditions can maintain good tracking performance.

[0023] Step 3: After the time differentiation is completed, fixed-time disturbance observation is performed. The lumped disturbance is observed and estimated in real time within a fixed time period to obtain the estimated value of the lumped disturbance.

[0024] Step 4: Perform adaptive specified disturbance elimination composite control calculations based on the estimated values, design virtual control quantities using the nonlinear backstepping method, and perform system control based on the system error feedback and dynamic command signals using the nonlinear backstepping method.

[0025] Step 5: Analyze the system control using closed-loop control. Determine if a control termination command has been received. If yes, terminate control. If no, jump to the command tracking stage for cyclic execution, continuing tracking control and disturbance estimation.

[0026] The beneficial effects of this invention are as follows: By using a combination of a command generator, an adaptive controller, a time disturbance observer, a multi-model sensor, a parameter estimator, and a data cloud platform, this invention can quickly and accurately describe and compensate for internal and external disturbances in aero-engine systems. Compared with traditional observers, it has a definite convergence time and better dynamic performance, thus ensuring that the system can maintain stable operation even when facing complex disturbances. By collecting multi-dimensional engine data using multiple sensors, it provides comprehensive engine status awareness, which helps to more accurately monitor and diagnose the engine's health status, monitor key engine parameters in real time, and analyze data through intelligent algorithms to promptly detect anomalies and potential faults, perform fault diagnosis and early warning, and improve the robustness and reliability of the control system. It balances dynamic and static performance with robustness, thus exhibiting good control performance. During dynamic changes, the system can quickly adjust and adapt to disturbances, thereby reducing steady-state errors and improving control stability and accuracy. It exhibits good dynamic and static performance under different operating conditions, especially when facing complex and nonlinear disturbances, ensuring efficient and accurate control performance. Attached Figure Description

[0027] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0028] Figure 1 This is a schematic diagram of the system of the present invention;

[0029] Figure 2 This is a flowchart of the steps of the present invention. Detailed Implementation

[0030] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0031] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms installation, connection, and linking should be interpreted broadly. For example, they can refer to fixed connections, detachable connections, or integral connections; they can refer to mechanical connections or electrical connections; they can refer to direct connections or indirect connections through an intermediate medium; and they can refer to the internal communication between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0032] In document 201910522225.6, a comprehensive assessment of the state of an aero-engine is conducted. Compared to the previous single vibration signal analysis, this method adds a step of collecting acoustic emission data and monitors early fault signals of the aero-engine based on this data. This intelligent monitoring method enhances perception capabilities, monitors a wider range of fault signal frequencies, and strengthens the system's noise immunity. However, this application breaks away from the indiscriminate handling of interference by traditional control methods. By dynamically observing interference, it fully utilizes the beneficial effect of lumped interference in the aero-engine system on tracking control, improving the convergence speed of the closed-loop system. This not only improves the system's tracking accuracy of target commands but also takes into account static characteristics, indirectly improving the design and parameter tuning effects of the engine control system.

[0033] according to Figure 1 , Figure 2 As shown, this embodiment provides an adaptive intelligent control system and method for aero-engines based on multimodal perception, including a command generator, an adaptive controller, a time disturbance observer, a multi-model sensor, a parameter estimator, and a data cloud platform;

[0034] The command generator is used to generate dynamic command signals that meet the target requirements according to the control objectives set by the system. The dynamic command signals have preset speed tracking trajectory characteristics to guide the engine speed adjustment.

[0035] The adaptive controller is used to input the output control quantity of the adaptive specified disturbance cancellation composite controller, the state feedback value characterizing the shaft speed of the aero-engine, and the system output quantity. Based on the state feedback information of the aero-engine and the output control command of the controller, the lumped disturbance is estimated in real time within a fixed time.

[0036] The time-perturbation observer achieves real-time observation and estimation of lumped disturbances over a fixed time period by extending state observation.

[0037] Multi-model sensors, including temperature sensors, pressure sensors, vibration sensors, and sound sensors, are used to monitor various physical and chemical parameters of the engine in real time;

[0038] The parameter estimator is used to observe the model parameters. The parameter estimator is equipped with three input terminals that are connected to the system input port, system output port and time disturbance observer of the aero-engine respectively. Its output terminal is connected to the adaptive controller to output the estimated values ​​of the model parameters.

[0039] The data cloud platform uploads the collected data to the cloud platform for storage and backup, performs remote analysis, and utilizes cloud computing resources for big data analysis and processing to collect and use engine information.

[0040] The command generator translates system decisions into actual control actions, such as adjusting fuel flow, changing turbine blade angles, and regulating compressor speed. Based on the engine's real-time status and changes in the external environment, the command generator dynamically adjusts control parameters to ensure the engine operates in optimal condition. When the system detects an abnormal fault, the command generator generates emergency control commands, including reducing power and switching to a backup system.

[0041] The adaptive controller combines the dynamic command aggregate disturbance estimate from the command generator with the state feedback information of the aero-engine. Based on the backstepping method and the dynamic inverse method, it adjusts the control quantity in real time. Combined with the adaptive specified disturbance cancellation method, it generates adaptive control commands and outputs them to the aero-engine through real-time estimation and feedback compensation of aggregate disturbances, thereby realizing real-time tracking control of the aero-engine speed and adaptive cancellation of specified disturbances.

[0042] The time-period disturbance observer estimates and compensates for time-varying disturbances and uncertainties in a system. These disturbances include changes in the external environment, fluctuations in internal system parameters, and sensor noise. By estimating the time-varying disturbance signal in the system in real time, the time-period disturbance observer feeds the estimated disturbance signal back to the control system, enabling the controller to dynamically adjust control commands to compensate for the impact of the disturbance. Through disturbance estimation and compensation, the time-period disturbance observer improves the robustness of the system, enabling the system to maintain stable performance in the face of uncertainties and disturbances.

[0043] Multi-model sensors provide comprehensive data input, enabling the system to more accurately monitor and analyze engine status. Data from different sensors can be cross-validated, improving the system's confidence in judging engine status and reducing false alarms and missed alarms. Different sensors have different sensitivities to environmental changes, and the combination of multi-model sensors allows for effective monitoring under various environments. Temperature sensors measure internal and external temperatures of the engine, including the combustion chamber, turbine, and compressor. Pressure sensors measure pressure in the engine's intake, exhaust, fuel, and hydraulic systems, ensuring that the pressure in each part of the engine is within the normal range and preventing performance degradation due to abnormal pressure. Vibration sensors detect engine vibration levels, helping to monitor the engine's structural integrity and mechanical health. Abnormal vibration patterns indicate engine component wear and imbalance, and vibration sensors also provide early warning functions. Sound sensors measure the noise level generated by the engine, helping to assess the engine's operating status and potential mechanical problems.

[0044] The parameter estimator analyzes sensor data to calibrate the system model, ensuring that the model accurately reflects the engine's operating state. During system operation, the parameter estimator updates parameter values ​​in real time to adapt to changes in engine state and external environment. Accurate parameter estimation helps improve the precision of the control system, enabling the controller to adjust the engine's operating parameters more effectively.

[0045] The data cloud platform provides a central repository for storing various sensor data from aircraft engines, including temperature, pressure, vibration, and sound. Multimodal data enables secure storage, management, and retrieval. The platform processes and analyzes massive amounts of sensor data in real time, providing immediate insights and feedback, which is crucial for monitoring engine performance and responding promptly to potential problems. Leveraging cloud computing resources, the platform can perform complex computational tasks, including training and inference of machine learning models, without requiring additional hardware on the aircraft. The platform facilitates data sharing and collaboration, enabling maintenance personnel to access and analyze data across geographical locations and work together to optimize engine performance and maintenance plans.

[0046] The output of the instruction generator is connected to the input of the adaptive controller.

[0047] The input of the fixed-time disturbance observer is connected to the output of the adaptive controller, the state feedback terminal of the downstream aero-engine, and the system output terminal, respectively.

[0048] Includes the following steps:

[0049] Step 1: System modeling of the aero-engine is performed, input and output data are recorded, and a perturbation-based affine nonlinear data model is constructed using system identification methods.

[0050] Step 2: Generate instruction tracking through the model, adjust and arrange process steps to generate instructions that meet preset control indicators, and calculate the derivative of the instruction with respect to time to ensure that the dynamic response of the system under different working conditions can maintain good tracking performance.

[0051] Step 3: After the time differentiation is completed, fixed-time disturbance observation is performed. The lumped disturbance is observed and estimated in real time within a fixed time period to obtain the estimated value of the lumped disturbance.

[0052] Step 4: Perform adaptive specified disturbance elimination composite control calculations based on the estimated values, design virtual control quantities using the nonlinear backstepping method, and perform system control based on the system error feedback and dynamic command signals using the nonlinear backstepping method.

[0053] Step 5: Analyze the system control using closed-loop control. Determine if a control termination command has been received. If yes, terminate control. If no, jump to the command tracking stage for cyclic execution, continuing tracking control and disturbance estimation to ensure stable engine performance control under complex operating conditions.

[0054] This adaptive intelligent control system and method for aero-engines based on multimodal perception utilizes the beneficial effect of lumped disturbances in the aero-engine system on tracking control, improves the convergence speed of the closed-loop system, and takes into account static characteristics, including static error and control ripple. By benefiting from beneficial disturbances and compensating for adverse disturbances, it achieves superior speed tracking control performance of aero-engines and realizes adaptive intelligent control.

[0055] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. An aero-engine adaptive intelligent control system based on multi-modal perception, characterized in that: The system comprises an instruction generator, an adaptive controller, a time disturbance observer, a multi-model sensor, a parameter estimator and a data cloud platform. The instruction generator is configured to generate a dynamic instruction signal in accordance with a control target set by the system, the dynamic instruction signal having a preset speed tracking trajectory characteristic to guide the speed regulation of the engine. The adaptive controller is configured to input an output control quantity of the adaptive designated disturbance elimination compound controller, a state feedback value representing the shaft speed of the aero-engine and a system output quantity, and estimate the lumped disturbance in real time within a fixed time according to the state feedback information of the aero-engine and the output control instruction of the controller. The time disturbance observer is configured to realize the observation and real-time estimation of the lumped disturbance within a fixed time through the extension of state observation. The multi-model sensor comprises a temperature sensor, a pressure sensor, a vibration sensor and a sound sensor, and is configured to monitor various physical and chemical parameters of the engine in real time. The parameter estimator is configured to observe model parameters, and has three input ends connected with the system input port, the system output port of the aero-engine and the time disturbance observer respectively, and an output end connected with the adaptive controller to output the estimated value of the model parameters. The data cloud platform is configured to upload the collected data to the cloud platform for storage and backup, remote analysis, big data analysis and processing by using cloud computing resources, and collection and summarization of the information of the engine.

2. The multi-modal perception based adaptive intelligent control system for a gas turbine engine as recited in claim 1, wherein: The instruction generator converts decisions in the system into actual control actions for adjusting the fuel flow and changing the angle of the turbine blades and adjusting the speed of the compressor.

3. The multi-modal perception based adaptive intelligent control system for a gas turbine engine as recited in claim 1, wherein: The adaptive controller adjusts the control quantity in real time based on the backstepping method and the dynamic inverse method in combination with the dynamic instruction lumped disturbance estimation value of the instruction generator and the state feedback information of the aero-engine.

4. The multi-modal perception based adaptive intelligent control system for a gas turbine engine as recited in claim 1, wherein: The time disturbance observer estimates and compensates for the disturbances and uncertainties in the system that change over time, including external environmental changes, internal parameter fluctuations and sensor noise, and estimates the disturbance signals that change over time in the system in real time.

5. The multi-modal perception based adaptive intelligent control system for a gas turbine engine as recited in claim 1, wherein: The multi-model sensor provides comprehensive data input, enabling the system to more accurately monitor and analyze the state of the engine, and the data from different sensors can be verified with each other to improve the confidence of the system in judging the state of the engine, reduce false positives and false negatives, and different sensors have different sensitivities to environmental changes.

6. The multi-modal perception based adaptive intelligent control system for a gas turbine engine as recited in claim 1, wherein: The parameter estimator calibrates the system model by analyzing the data of the sensors to ensure that the model accurately reflects the working state of the engine, and updates the parameter values in real time during the operation of the system.

7. The multi-modal perception based adaptive intelligent control system for a gas turbine engine as recited in claim 1, wherein: The data cloud platform provides a central repository for storing various sensor data from the aero-engine, including temperature, pressure, vibration and sound, and the multi-modal data enables the data to be safely stored, managed and retrieved, and the data cloud platform processes and analyzes a large amount of sensor data in real time to provide immediate insights and feedback.

8. The multi-modal perception based adaptive intelligent control system for a gas turbine engine as recited in claim 1, wherein: The output end of the instruction generator is connected with the input end of the adaptive controller.

9. The multi-modal perception based adaptive intelligent control system for a gas turbine engine as recited in claim 1, wherein: The input end of the fixed time disturbance observer is connected with the output end of the adaptive controller, the state feedback end of the downstream aero-engine and the system output end.

10. The method applied to the aero-engine adaptive intelligent control system based on multi-modal perception according to any one of claims 1-9, characterized in that: Comprise the following steps: Step one: through the system of aero-engine system modeling, input and output data, through the system identification method, the construction of affine form nonlinear data model with disturbance; Step two: through the model to generate command tracking, through the adjustment and arrangement process link, produce in line with the preset control index, and the command to time differential, ensure that the dynamic response of the system under different working conditions can maintain good tracking performance; Step three: after the time differential is completed, the fixed time disturbance observation is carried out, the observation and real-time estimation of the lumped disturbance are carried out in the fixed time, and the estimated value of the lumped disturbance is obtained; Step four: adaptive designated disturbance elimination composite control calculation is carried out through the estimated value, virtual control quantity is designed by using nonlinear backstepping method, and system control is carried out based on error feedback and dynamic instruction signal of the system by using nonlinear backstepping method; Step five: the system control is judged, the closed loop control is used to judge whether the end control instruction is received, if yes, the control is ended, if not, the instruction tracking link is jumped to and is executed circularly, and the tracking control and disturbance estimation are continued.

Images