CN-121978957-A - Combustion instability feedforward control method and system based on modal early identification

CN121978957ACN 121978957 ACN121978957 ACN 121978957ACN-121978957-A

Abstract

The invention discloses a combustion instability feedforward control method and system based on modal early identification, which belong to the technical field of combustion control and comprise the steps of S1, collecting multiple physical field signals in a combustion chamber in real time, S2, identifying the current dominant mode and amplitude increasing trend of combustion instability in real time based on the multiple physical field signals, S3, carrying out double-condition judgment, triggering control before the oscillation amplitude reaches a preset safety threshold, S4, matching a corresponding open loop feedforward control strategy from a preset control strategy library, and S5, generating and applying a control instruction of a mode specific type to drive an executing mechanism to intervene. By adopting the combustion instability feedforward control method and system based on the modal early identification, the early warning and the modal specific control of the combustion instability are realized, the problems of phase mismatch and nonlinear inadaptation are effectively avoided, the thermal acoustic injury risk is obviously reduced, and the robustness, the instantaneity and the engineering applicability of the control system are improved.

Inventors

ZHANG ZHIHAO
YI LIN
CUI YANYU
DING QINGMIAO
GUO YAOWEN

Assignees

中国民航大学

Dates

Publication Date: 20260505
Application Date: 20260209

Claims (9)

1. The combustion instability feedforward control method based on the mode early identification is characterized by comprising the following steps of: S1, acquiring multiple physical field signals in a combustion chamber in real time; S2, based on multiple physical field signals, identifying the current dominant mode and amplitude increasing trend of combustion instability in real time; s3, performing double-condition judgment according to the type of the dominant mode and the amplitude increasing trend, and triggering control before the oscillation amplitude reaches a preset safety threshold; S4, based on a dominant mode, matching a corresponding feedforward control strategy from a preset control strategy library; S5, based on the matched feedforward control strategy, a control instruction of the mode special type is generated and applied to drive the executing mechanism to intervene.
2. The combustion instability feedforward control method based on modal early recognition according to claim 1, wherein in step S2, the real-time recognition specifically includes: A modal identification criterion constructed based on historical experimental data, wherein the modal identification criterion is composed of a plurality of quantization indexes reflecting the nonlinear dynamics characteristics of oscillation; Extracting quantization indexes by relying on high-frequency dynamic pressure signals acquired in real time; And (3) finishing early identification of the limit cycle mode, the quasi-cycle mode and the chaotic mode based on the quantization index by a nonlinear dynamics analysis method.
3. The combustion instability feedforward control method based on mode early identification according to claim 1, wherein in step S3, the dual condition determination is specifically that when a specific instability mode is identified and the real-time increment of the oscillation amplitude exceeds a preset increment threshold corresponding to the mode in unit time, the triggering condition is satisfied.
4. The combustion instability feedforward control method based on modal early recognition according to claim 2 is characterized in that feedforward control is open loop feedforward control in step S4, strategies and parameters in a control strategy library are calibrated and optimized in advance based on open loop control experiments respectively carried out on different oscillation modes, wherein a chaos destruction synchronization strategy is adopted for a limit cycle mode or a quasi-cycle mode, a phase synchronization or chaos suppression strategy is adopted for a chaos mode, and the strategy changes distribution of internal oscillation energy of a combustion system in frequency or phase dimension by applying external excitation to achieve instability suppression.
5. The combustion instability feedforward control method based on modal early recognition according to claim 1, wherein the parameters of the control command in S5 are adaptively adjusted according to the signal characteristics and the working condition changes detected in real time.
6. The combustion unstable feed forward control method based on early modal identification of claim 1 wherein in step S5, the step of generating and applying control commands of a modal specific type includes: Determining at least one target executing mechanism from the acoustic exciter, the fuel flow modulator and the plasma exciter according to the control requirement and the real-time working condition; the drive target actuator applies an excitation corresponding to the control command.
7. The combustion instability feedforward control method based on modal early recognition of claim 6, wherein the dominant modes include a limit cycle mode, a quasi-periodic mode, and a chaotic mode.
8. A combustion instability feedforward control system based on modal early recognition for executing the combustion instability feedforward control method based on modal early recognition of any one of claims 1-7, comprising: the system comprises a signal acquisition module, a combustion chamber, a combustion module and a control module, wherein the signal acquisition module synchronously acquires multiple physical field signals reflecting the combustion state through a sensor array arranged in the combustion chamber; The mode identification module is used for identifying dominant modes and amplitude increasing trends of the combustion instability in real time based on the multiple physical field signals; the control trigger module is used for performing double-condition judgment according to the type of the dominant mode and the amplitude increasing trend, and generating a control trigger instruction before the oscillation amplitude is obviously increased; The strategy matching module is used for matching a corresponding open loop feedforward control strategy from a preset control strategy library based on a dominant mode; the instruction generation and execution module is used for generating a control instruction of a mode special type based on the matched feedforward control strategy and driving the execution mechanism to intervene.
9. The combustion unstable feedforward control system based on modal early recognition of claim 8, wherein the modal recognition module is specifically configured to process only high-frequency dynamic pressure signals collected by the signal collection module based on a modal recognition criterion constructed by historical experimental data to complete early recognition of limit cycles, quasi-cycles and chaotic modes; the control trigger module is configured with a dynamic trigger threshold, and the judgment logic is used for generating a control trigger instruction when the mode identification module identifies a specific instability mode and the signal amplitude increment acquired by the signal acquisition module exceeds a preset increment threshold corresponding to the mode in unit time; the strategy and parameters in the control strategy library are calibrated and optimized in advance based on open loop control experiments respectively developed for different oscillation modes; The system adopts a mode of combining closed-loop real-time monitoring and open-loop feedforward control, and the instruction generation and execution module can adaptively adjust parameters of a control instruction according to the signal characteristics monitored in real time; The actuator includes at least one of an acoustic actuator, a fuel flow modulator, and a plasma actuator.

Description

Combustion instability feedforward control method and system based on modal early identification Technical Field The invention relates to the technical field of combustion control, in particular to a combustion instability feedforward control method and system based on modal early identification. Background Currently, combustion instability control methods mainly surround passive control and active control. The passive control method has the advantages that the structure is simple, the implementation is easy, the control characteristic is fixed, the control effect is difficult to follow the operation requirements of the gas turbine on wide working conditions and variable loads, and the control effect is often remarkably attenuated under non-design working conditions. In terms of active control, the mainstream and mature technologies at the present stage rely on a closed-loop feedback framework of "real-time phase locking-inversion". The method collects pressure or heat release rate pulsation of the combustion chamber in real time through the sensor, and instantly generates fuel or air modulation disturbance opposite to the pressure or heat release rate pulsation, so as to inhibit thermoacoustic oscillation under wide working conditions. However, this technical path has inherent drawbacks. First, its control effect is highly sensitive to the phase difference between the excitation signal and the combustion oscillations. Microsecond time lags existing in the system from sensing, calculating to executing links can cause preset opposite phases to drift, so that the counteracting effect is weakened, and even constructive interference is converted when serious, so that the oscillation amplitude does not drop and rise reversely. Secondly, when the combustion system presents non-linear characteristics such as chaos and multi-mode coupling, the linear phase compensation strategy is difficult to match with a non-linear energy transfer path, so that the control effect is attenuated sharply. Most critical is that the method is essentially a post-hoc intervention, i.e. it must wait for the oscillation amplitude to start after exceeding the detection threshold. From the oscillation occurrence, the detection by the system, the validation of the control command and the action on the combustion process, there is an unavoidable detection-execution delay window, resulting in a combustion chamber still subjected to an impact of exponentially increasing amplitude, with a very high risk of thermoacoustic damage. The traditional strategy depends on a feedback mechanism, namely regulation and control are implemented after unstable oscillation is detected, and the method is influenced by system sensing, execution and calculation delays, has response lag, cannot intervene before the instability occurs, and causes that a combustion chamber still needs to bear periodic oscillation load, so that long-term reliable operation of equipment is not facilitated. Disclosure of Invention The invention aims to provide a combustion instability feedforward control method and a combustion instability feedforward control system based on modal early identification, which aim to overcome the defects of lag response and phase locking and phase inversion dependence of traditional feedback control, realize early identification and early intervention of combustion instability, prevent thermoacoustic oscillation and improve the operation reliability and the working range of power devices such as a gas turbine and the like. In order to achieve the above object, the present invention provides a combustion instability feedforward control method based on modal early recognition, comprising the steps of: S1, acquiring multiple physical field signals in a combustion chamber in real time; S2, based on multiple physical field signals, identifying the current dominant mode and amplitude increasing trend of combustion instability in real time; s3, performing double-condition judgment according to the type of the dominant mode and the amplitude increasing trend, and triggering control before the oscillation amplitude reaches a preset safety threshold; S4, based on a dominant mode, matching a corresponding feedforward control strategy from a preset control strategy library; S5, based on the matched feedforward control strategy, a control instruction of the mode special type is generated and applied to drive the executing mechanism to intervene. Preferably, in step S2, the real-time identification specifically includes: A modal identification criterion constructed based on historical experimental data, wherein the modal identification criterion is composed of a plurality of quantization indexes reflecting the nonlinear dynamics characteristics of oscillation; Extracting quantization indexes by relying on high-frequency dynamic pressure signals acquired in real time; And (3) finishing early identification of the limit cycle mode, the quasi-cycle mode and the chaotic mo