Combustion state real-time regulation and control system and method for gas turbine combustion chamber test bed

Abstract

The invention discloses a gas turbine combustion chamber test bed combustion state real-time regulation and control system and method, and relates to the technical field of gas turbine combustion regulation and control, and the method comprises the steps: achieving the precise recognition of the combustion state through the analysis of a multi-field coupling mechanism by comprehensively capturing the internal multi-dimensional combustion parameters and flow field characteristic information of a combustion chamber; and the combustion state deviation degree and grade are determined in combination with deviation quantitative analysis. A self-adaptive regulation and control strategy is generated based on a combustion process time sequence evolution rule, the strategy is converted into regulation actions of fuel supply, an air inlet angle and ignition energy, and a closed-loop regulation and control link is formed through continuous data acquisition and feedback. The system and the method effectively solve the problems of incomplete combustion state identification and lagging regulation response in the prior art, improve the combustion state regulation precision and real-time performance, ensure that the combustion process is stably located in the optimal interval, adapt to multi-working-condition switching and transient load change requirements, and improve the accuracy and repeatability of test data of a test bed.

Description

Technical Field

[0001] This invention relates to the field of gas turbine combustion control technology, and in particular to a real-time control system and method for combustion state of a gas turbine combustion chamber test bench. Background Technology

[0002] As a core component of the power system, the combustion stability of the gas turbine combustor directly affects the testing accuracy and operational safety of the test bench, playing a crucial role in testing and verification in aerospace, power generation, and other fields. With the development of gas turbines towards higher parameters and higher efficiency, the combustion process inside the combustor exhibits complex characteristics of multi-field coupling and dynamic evolution. The interaction of temperature distribution, pressure fluctuations, flow field structure, and component concentration forms a highly nonlinear system, making traditional control methods difficult to adapt to the demands of test benches for multi-condition switching and transient load changes. During testing, the combustion state is easily affected by fuel characteristic fluctuations, changes in intake conditions, and combustor structural parameters. Real-time and precise control methods are needed to ensure the combustion process remains within the optimal range. Therefore, developing an integrated system and method that integrates multi-parameter sensing, coupling mechanism identification, and adaptive control has become a core requirement for solving the problem of precise combustion state control on test benches.

[0003] Existing technologies for combustion control in gas turbine combustor test benches have two major drawbacks: First, combustion state identification lacks consideration of multi-field coupling mechanisms, relying heavily on the independent monitoring and analysis of single or partial parameters. It fails to fully integrate the correlation characteristics of multi-dimensional data such as temperature, pressure, flow field, and composition, making it difficult to comprehensively analyze the coupling laws of heat and mass transfer and chemical reactions in the combustion field. This results in insufficient accuracy in characterizing the steady-state characteristics of combustion, failing to provide a comprehensive and reliable basis for control decisions. Second, the temporal adaptability and dynamic optimization capabilities of control strategies are insufficient. Existing control methods mostly employ fixed parameters or simple feedback logic, failing to dynamically adjust control parameters and strategies in conjunction with the temporal evolution of the combustion process. This makes it difficult to cope with transient changes and cumulative deviations in the combustion state, leading to delayed control response, insufficient accuracy, and an inability to achieve real-time dynamic optimization of the combustion state, affecting the accuracy and repeatability of test bench data. Summary of the Invention

[0004] In order to overcome the shortcomings and deficiencies of the existing technology, the present invention provides a real-time control system and method for combustion state of a gas turbine combustion chamber test bench.

[0005] The technical solution adopted in this invention is a real-time combustion state control system for a gas turbine combustor test bench, comprising: a multi-dimensional combustion parameter sensing module, a flow field characteristic dynamic acquisition module, a multi-field coupled combustion identification module, a combustion state deviation analysis module, a deep time-series adaptive control module, and a control command precision execution module; The combustion parameter multidimensional sensing module and the flow field characteristic dynamic acquisition module interact bidirectionally. The output of the flow field characteristic dynamic acquisition module is connected to the input of the multi-field coupled combustion identification module. The multi-field coupled combustion identification module establishes a feedback link with the deep time-series adaptive control module through the combustion state deviation analysis module. The output of the deep time-series adaptive control module is connected to the input of the control command precision execution module. The combustion parameter multidimensional sensing module captures the temperature distribution, pressure fluctuation, component concentration, and flow velocity vector information inside the gas turbine combustion chamber. The flow field characteristic dynamic acquisition module acquires data on the turbulence intensity, vortex structure evolution, and gas-solid mixing homogeneity of the combustion region. The multi-field coupled combustion identification module constructs a combustion steady-state identification model based on the aforementioned data and analyzes the heat and mass transfer law of the combustion field. The combustion state deviation analysis module compares the identification results with the preset combustion threshold range and quantifies the degree of deviation. The deep time-series adaptive control module triggers the deep time-series combustion adaptive control algorithm to generate a control strategy based on the deviation data. The control command precision execution module transforms the control strategy into adjustment actions for the burner fuel supply, air intake angle, and ignition energy intensity.

[0006] Furthermore, the multi-field coupled combustion identification module includes: a combustion field data fusion unit, a thermal parameter correlation analysis unit, a combustion steady-state feature extraction unit, and a multi-field coupled model construction unit. The combustion field data fusion unit receives heterogeneous data transmitted from the combustion parameter multi-dimensional sensing module and the flow field characteristic dynamic acquisition module, and integrates temperature, pressure, and flow field related data through a multi-source information complementary fusion method. The thermal parameter correlation analysis unit mines the nonlinear mapping relationship between parameters in different combustion stages. The combustion steady-state feature extraction unit separates combustion steady-state calibration feature quantities from the fused data and removes transient interference components. The multi-field coupled model construction unit establishes a combustion steady-state identification model based on the feature quantities, including the interaction mechanism of thermal field, flow field, and component field, to accurately characterize the combustion state in multiple dimensions.

[0007] Furthermore, the combustion state deviation analysis module includes: a preset threshold storage unit, a real-time data calibration unit, a deviation quantification calculation unit, and a deviation level determination unit. The preset threshold storage unit stores the optimal threshold range for combustion state under different operating conditions. The real-time data calibration unit performs system error correction on the identification results output by the multi-field coupled combustion identification module. The deviation quantification calculation unit calculates the deviation value between the real-time combustion state and the optimal threshold range through difference calculation and weight allocation. The deviation level determination unit divides the deviation level according to the deviation value and outputs the corresponding level identifier.

[0008] Furthermore, the deep temporal adaptive control module includes: a control strategy generation unit, an algorithm parameter optimization unit, a control effect prediction unit, and a control command output unit. The control strategy generation unit triggers the deep temporal adaptive combustion control algorithm based on the combustion state deviation level. The algorithm parameter optimization unit dynamically adjusts the algorithm calibration parameters according to the temporal change law of the combustion process. The control effect prediction unit simulates the evolution trend of combustion state under different control strategies. The control command output unit selects the control strategy corresponding to the optimal evolution trend and converts it into a standardized control command.

[0009] Furthermore, the combustion steady-state identification model adopted by the multi-field coupled combustion identification module satisfies: ,in, For the temperature distribution in the combustion zone, For the pressure distribution in the combustion chamber, This refers to the mass fraction of fuel components. The velocity vector of the flow field. The density of the combustion products, For isobaric specific heat capacity, Thermal conductivity, For the first The rate of formation of each reaction component For the first Enthalpy values ​​of the components Here is the viscous dissipation coefficient. For dynamic viscosity, This represents the number of components participating in the reaction.

[0010] Furthermore, the optimization model for the control parameters of the deep temporal adaptive control module satisfies: , in, To regulate the objective function, For speed adjustment weighting coefficients, For pressure regulation weighting coefficient, For component regulation weighting coefficients, For reference temperature distribution, For reference pressure distribution, For reference component mass fraction, This is time-series temperature data. For time-series stress data, The mass fraction of the time-series components. This is the deviation weighting coefficient. To regulate time.

[0011] Furthermore, the combustion state deviation quantification model satisfies: , in, This is the overall deviation value. To monitor the number of parameters, For the first Real-time measured values ​​of each parameter For the first Optimal values ​​for each parameter For the first Standard deviation of each parameter The coefficient representing the influence of time-series changes. The parameter is the synergistic influence coefficient. For the first The rate of change of each parameter over time.

[0012] Furthermore, the flow field characteristic prediction model satisfies: , in, For the velocity of the three-dimensional spatiotemporal flow field, The initial flow field velocity distribution, Kinematic viscosity, Let be the effect function of temperature and composition. The vortex diffusion coefficient is... For spatial coordinates, For time.

[0013] Furthermore, the fuel supply regulation model satisfies: , in, For real-time fuel supply, As the baseline fuel supply, The deviation response coefficient, The temperature-pressure coupling coefficient is... The component rate synergy coefficient, This is the overall deviation value. The temperature gradient is in the x-direction. This represents the pressure gradient in the y-direction.

[0014] A method for real-time control of combustion state on a gas turbine combustor test bench, applied to a real-time control system for combustion state on a gas turbine combustor test bench, includes the following steps: S1, continuously capturing internal temperature, pressure, component concentration, and flow velocity data of the gas turbine combustor through a distributed sensing unit deployed in a multi-dimensional combustion parameter sensing module; simultaneously acquiring information related to turbulence intensity, vortex structure, and mixing uniformity of the combustion region flow field using a laser velocimetry component and a particle image velocimetry device in a dynamic flow field characteristic acquisition module; S2, transmitting the captured multi-dimensional parameters and flow field information to a multi-field coupled combustion identification module; performing feature extraction and coupling analysis on the data using a constructed multi-field coupled combustion steady-state identification model to analyze the coupling mechanism of heat and mass transfer and chemical reaction in the combustion field; S3, using a combustion state deviation analysis module to call a preset combustion threshold range to analyze the output of the identification module. After the combustion state data is calibrated for error, the deviation between the real-time combustion state and the optimal range is quantified and the deviation level is divided through difference calculation and weight allocation; S4, based on the deviation level, the deep temporal adaptive combustion control algorithm is triggered. In the deep temporal adaptive control module, the temporal window length and iteration step size of the algorithm are dynamically adjusted, and a multi-dimensional control strategy is generated by combining the temporal evolution law of the combustion process; S5, the generated control strategy is transmitted to the control command precision execution module. The module's built-in actuator drives the flow regulating valve of the fuel supply pipeline, the angle adjusting blade of the air intake duct, and the energy output component of the ignition system to complete the corresponding adjustment action; S6, the parameters and flow field information in the combustion chamber after adjustment are continuously captured. The parameter acquisition, coupling identification, deviation analysis, strategy generation, and command execution process are repeated to form a closed-loop control link for real-time dynamic optimization of the combustion state.

[0015] Beneficial Effects: This invention proposes a real-time combustion state control system and method for a gas turbine combustor test bench. By constructing a multi-module collaborative integrated control architecture and scientific control process, it achieves precise, real-time, and dynamic optimization of the combustion state of the gas turbine combustor test bench. This invention comprehensively captures combustion parameters and flow field characteristic data by setting up multi-dimensional sensing and dynamic acquisition modules. Combined with a multi-field coupled combustion identification module, it deeply integrates the correlation characteristics of multi-source data, fully analyzes the coupling law between heat and mass transfer and chemical reaction in the combustion field, and accurately characterizes the steady-state characteristics of combustion. This effectively compensates for the shortcomings of existing technologies, such as the lack of consideration for multi-field coupling mechanisms and incomplete identification of combustion states. This provides a comprehensive and reliable basis for control decisions. At the same time, the combustion state deviation analysis module quantifies the deviation level, triggering the deep time-series adaptive control module to dynamically adjust the control parameters and strategies according to the time-series evolution law of the combustion process, forming a closed-loop control link. This significantly improves the time-series adaptability and dynamic optimization capability of the control strategy, solves the problems of lagging control response and insufficient accuracy of existing technologies, and can quickly respond to transient changes and cumulative deviations in the combustion state, ensuring that the combustion process is always in the optimal range. This significantly improves the accuracy and repeatability of test data on the test bench, and adapts to the test requirements of high-parameter gas turbines with multi-condition switching and transient load changes. Attached Figure Description

[0016] Figure 1 This is a diagram showing the system module composition of the present invention; Figure 2 This is a flowchart of the method steps of the present invention. Detailed Implementation

[0017] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. The application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0018] like Figure 1 As shown, a real-time combustion state control system for a gas turbine combustor test bench includes: a multi-dimensional combustion parameter sensing module, a dynamic flow field characteristic acquisition module, a multi-field coupled combustion identification module, a combustion state deviation analysis module, a deep time-series adaptive control module, and a control command precision execution module. The combustion parameter multidimensional sensing module and the flow field characteristic dynamic acquisition module interact bidirectionally. The output of the flow field characteristic dynamic acquisition module is connected to the input of the multi-field coupled combustion identification module. The multi-field coupled combustion identification module establishes a feedback link with the deep time-series adaptive control module through the combustion state deviation analysis module. The output of the deep time-series adaptive control module is connected to the input of the control command precision execution module. The combustion parameter multidimensional sensing module captures the temperature distribution, pressure fluctuation, component concentration, and flow velocity vector information inside the gas turbine combustion chamber. The flow field characteristic dynamic acquisition module acquires data on the turbulence intensity, vortex structure evolution, and gas-solid mixing homogeneity of the combustion region. The multi-field coupled combustion identification module constructs a combustion steady-state identification model based on the aforementioned data and analyzes the heat and mass transfer law of the combustion field. The combustion state deviation analysis module compares the identification results with the preset combustion threshold range and quantifies the degree of deviation. The deep time-series adaptive control module triggers the deep time-series combustion adaptive control algorithm to generate a control strategy based on the deviation data. The control command precision execution module transforms the control strategy into adjustment actions for the burner fuel supply, air intake angle, and ignition energy intensity.

[0019] The multi-dimensional combustion parameter sensing module adopts a distributed array deployment, with a total of 32 high-precision sensing units deployed at key locations such as the inner wall of the combustion chamber flame tube, the central axis of the combustion zone, and the outlet section. The temperature sensing unit uses a combination of high-precision thermocouples and infrared temperature measuring elements, covering a temperature measurement range of 200 to 1800 degrees Celsius with a measurement resolution of 0.1 degrees Celsius and a sampling frequency of 100 Hz, enabling continuous capture of temperature distribution gradients in different areas. The pressure sensing unit uses miniature piezoelectric pressure sensors, arranged at eight evenly distributed measuring points on the circumferential side wall of the combustion chamber, with a measurement range of 0.1 to 10 MPa and a sampling frequency of 200 Hz, capturing real-time combustion process data. The pressure fluctuation signal is detected; the component concentration sensing unit uses laser absorption spectroscopy technology and sets up 6 detection channels for the main fuel components and combustion products, with a detection response time of less than 5 milliseconds, which can accurately obtain the mass fraction changes of each component; the flow velocity vector sensing unit uses the ultrasonic velocimetry principle and arranges 4 sets of dual-channel ultrasonic probes in the key section of the flow field in the combustion zone, with a measurement range of 0.5 to 30 meters per second and a measurement accuracy of 1%, simultaneously collecting information on the magnitude and direction of the flow velocity. All sensing units aggregate data through a high-speed data bus with a data transmission rate of 1 gigabits per second, ensuring the synchronization and integrity of multi-dimensional parameters, and providing comprehensive and accurate raw data support for subsequent combustion state analysis.

[0020] The dynamic flow field acquisition module integrates laser Doppler velocimetry and particle image velocimetry. Three laser emitting units and four high-speed imaging units are installed outside the combustion chamber. The laser emitting units output continuous laser light with a wavelength of 532 nm, focusing on 12 preset measurement planes within the combustion area. Each plane is divided into 256×256 measurement grids with a grid spacing of 2 mm, achieving high-resolution coverage of the flow field. The high-speed imaging units are set to a frame rate of 500 frames per second and an exposure time of 1 microsecond. Using inert tracer particles with a diameter of 1 micrometer, the module captures the particle trajectories in the flow field and analyzes particle displacement information through image matching algorithms. Simultaneously, it integrates turbulence... The flow intensity analysis module calculates the turbulence intensity at each measurement point based on the collected flow velocity data, with a measurement range of 0.01 to 0.3 and a resolution of 0.001. The vortex structure evolution analysis identifies the vortex core position, size, and movement speed through correlation calculations of continuous frame images, with a tracking frequency of 100 Hz. The gas-solid mixing homogeneity is calculated through the standard deviation of particle concentration distribution, with a sampling interval of 5 milliseconds, and outputs the mixing homogeneity index in real time. This module has a data acquisition cycle of 20 milliseconds and keeps synchronized with the combustion parameter multidimensional sensing module. Through the fusion of multiple technologies, it achieves dynamic and comprehensive capture of flow field characteristics, revealing the complex changing laws of the flow field during combustion.

[0021] The multi-field coupled combustion identification module receives synchronous data transmitted from the multi-dimensional combustion parameter sensing module and the dynamic acquisition module of flow field characteristics. First, it performs data alignment and noise filtering through a multi-source data preprocessing unit. A sliding window averaging method is used to process continuous signals such as temperature and pressure, with a window length set to 50 sampling points. Outlier data points are removed using median filtering. Then, in the data fusion stage, a weighted fusion algorithm assigns different weights to heterogeneous data: temperature and pressure data have a weight coefficient of 0.3, flow field characteristic data has a weight coefficient of 0.2, component concentration data has a weight coefficient of 0.2, and flow velocity vector data has a weight coefficient of 0.1. Weighted summation achieves the organic integration of multi-dimensional data. The thermal parameter correlation analysis is based on partial least squares regression, constructing 15... The nonlinear correlation model among key parameters was trained with 1000 iterations and achieved a convergence accuracy of 10 to the power of -5. Combustion steady-state feature extraction involved wavelet transform decomposition of the data, extracting 5 low-frequency and 8 high-frequency feature components, and selecting 12 core feature parameters characterizing combustion steady-state. A multi-field coupling model was constructed based on the finite volume method, dividing the data into over 100,000 computational grids. It considered the interactions of the thermal field, flow field, and component fields, introducing the radiative heat transfer coefficient and chemical reaction rate constant. The combustion field control equations were solved numerically iteratively with an iteration step of 1 millisecond. The convergence criterion was a residual less than 10 to the power of -4. The final output included the heat and mass transfer laws of the combustion field, the chemical reaction process, and the set of steady-state feature parameters, achieving accurate identification and quantitative characterization of the combustion state.

[0022] The combustion state deviation analysis module incorporates a database of optimal combustion threshold ranges under different operating conditions, including 18 typical test conditions. For each condition, it stores the optimal value ranges for 20 key parameters such as temperature, pressure, component concentration, and flow field characteristics. The threshold ranges are determined through statistical analysis of extensive benchmark test data, achieving a confidence level of 99%. The real-time data calibration unit employs a systematic error correction algorithm based on a pre-set sensor error model to correct deviations in the input combustion state identification data. Correction coefficients are obtained through periodic calibration tests, with an update cycle of 30 days. Deviation quantification calculations utilize the weighted Euclidean distance method, assigning different weights to parameters based on their importance, with core parameters such as main combustion zone temperature and combustion pressure receiving higher weights. The weight of the primary parameter is 0.2, and the weight of secondary parameters such as turbulence intensity and mixing uniformity is 0.05. The deviation value of a single parameter is obtained by calculating the weighted distance between the real-time data and the center value of the optimal threshold interval. Then, the deviation value is mapped to the interval of 0 to 1 through normalization. The deviation level is set with 5 levels: deviation value of 0 to 0.2 is level 1, 0.2 to 0.4 is level 2, 0.4 to 0.6 is level 3, 0.6 to 0.8 is level 4, and 0.8 to 1.0 is level 5. Each level corresponds to a clear description of the degree of deviation and the priority of subsequent control. The processing cycle of this module is 50 milliseconds, and it quickly outputs the deviation quantification results and level labels, providing a clear basis for the generation of control strategies.

[0023] The deep temporal adaptive control module, based on the deviation data and level identifiers output by the combustion state deviation analysis module, triggers the deep temporal adaptive combustion control algorithm. The algorithm first constructs a combustion process temporal prediction model using a long short-term memory network structure with three hidden layers, each containing 128 neurons. The input sequence length is 100 time steps, and the output is the predicted combustion state value for the next 50 time steps. The dynamic adjustment mechanism of the temporal window length is adaptively set according to the deviation level: 20 time steps for level 1 to 2 deviations, 10 time steps for level 3 to 4 deviations, and 5 time steps for level 5 deviations. The iteration step size is set to 0.01 to 0.1. The adaptive range allows for larger iteration steps as the deviation increases, accelerating the control response speed. The control strategy generation employs a multi-objective optimization algorithm, aiming to minimize combustion state deviation and maximize the smoothness of control actions. Constraints include a fuel supply adjustment range of 50% to 150% of the baseline value, an air intake angle adjustment range of 0 to 45 degrees, and an ignition energy intensity adjustment range of 80% to 120% of the baseline value. Control effect prediction uses numerical simulation to predict the evolution trend of combustion state under different control strategies, with a prediction time of 1 second, selecting the optimal control strategy. This module has a processing cycle of 100 milliseconds, ensuring the real-time performance and effectiveness of the control strategy, achieving precise adaptation to changes in combustion state.

[0024] The precise execution module for control commands receives the control strategy output by the deep timing adaptive control module. Through the command parsing unit, it transforms abstract control parameters into specific execution commands. For example, the fuel supply adjustment command is converted into an opening control signal for the flow control valve, output using pulse width modulation (PWM), with a control signal frequency of 100 Hz and a resolution of 0.1%. The air intake angle adjustment command is converted into a stepper motor control signal for the angle adjustment blades, with a step angle of 0.5 degrees, a positioning accuracy of 0.1 degrees, and a response time of less than 100 milliseconds. The ignition energy intensity adjustment command is achieved by adjusting the output voltage and pulse frequency of the ignition module, with a voltage adjustment range... The voltage range is 12 to 24 volts, and the pulse frequency adjustment range is 1 to 10 Hz. The module has a built-in actuator feedback unit that collects the actual opening degree of the flow regulating valve, the actual angle of the angle regulating blade, and the actual output value of the ignition energy in real time at a frequency of 200 Hz. Through closed-loop feedback control, the accuracy of command execution is calibrated, and the control error is less than 1%. At the same time, a fault diagnosis unit is set up to monitor the operating status of the actuator. When abnormal conditions such as jamming or overtravel occur, an alarm signal is immediately output and the system switches to the standby control mode to ensure the reliable execution of control commands. The control strategy is accurately translated into the actual adjustment action of the burner, thereby achieving effective control of the combustion state.

[0025] Preferably, the multi-field coupled combustion identification module includes: a combustion field data fusion unit, a thermal parameter correlation analysis unit, a combustion steady-state feature extraction unit, and a multi-field coupled model construction unit. The combustion field data fusion unit receives heterogeneous data transmitted from the combustion parameter multi-dimensional sensing module and the flow field characteristic dynamic acquisition module, and integrates temperature, pressure, and flow field related data through a multi-source information complementary fusion method. The thermal parameter correlation analysis unit mines the nonlinear mapping relationship between parameters in different combustion stages. The combustion steady-state feature extraction unit separates combustion steady-state calibration feature quantities from the fused data and removes transient interference components. The multi-field coupled model construction unit establishes a combustion steady-state identification model based on the feature quantities, including the interaction mechanism of thermal field, flow field, and component field, to perform multi-dimensional and accurate characterization of the combustion state.

[0026] Specifically, the multi-field coupled combustion identification module's combustion field data fusion unit receives 32 channels of sensor data from the multi-dimensional combustion parameter sensing module and 12 measurement plane data from the flow field characteristic dynamic acquisition module. It uses synchronous clock calibration technology to control the timestamp accuracy of both types of data to within 1 microsecond. Through data format standardization, heterogeneous data is uniformly converted into decimal floating-point data. Different fusion weights are then assigned based on data reliability grading to ensure complementary integration of multi-source data. The thermal parameter correlation analysis unit selects 15 key thermal parameters in the combustion process, divides the data into segments using a sliding time window method, with each window including 500 sampling points. It calculates the linear and nonlinear correlation coefficients between parameters using a correlation analysis algorithm, achieving a correlation coefficient calculation accuracy of one ten-thousandth, clarifying the dominant influence relationship between parameters at different combustion stages. The combustion steady state... The feature extraction unit sets the signal decomposition layer to 6 layers, and separates feature components in different frequency ranges through multi-scale analysis. Among them, low-frequency feature components correspond to the basic steady-state characteristics of combustion, and high-frequency feature components correspond to transient interference signals. The threshold screening method is used to retain feature components with a variance ratio of more than 0.01, and finally 12 core steady-state feature parameters are selected. The multi-field coupling model construction unit adopts structured mesh generation technology, dividing the combustion area into more than 100,000 hexahedral computational meshes, with a mesh quality compliance rate of no less than 95%. The control equations of thermal field, flow field and component field are transformed into a set of algebraic equations through numerical discretization method. The maximum number of iteration steps is set to 5,000 steps to ensure that the model can fully reflect the multi-field interaction mechanism in the combustion process, realize high-precision identification of combustion state, and provide accurate basic data for subsequent deviation analysis.

[0027] Preferably, the combustion state deviation analysis module includes: a preset threshold storage unit, a real-time data calibration unit, a deviation quantification calculation unit, and a deviation level determination unit. The preset threshold storage unit stores the optimal threshold range for combustion state under different operating conditions. The real-time data calibration unit performs system error correction on the identification results output by the multi-field coupled combustion identification module. The deviation quantification calculation unit calculates the deviation value between the real-time combustion state and the optimal threshold range through difference calculation and weight allocation. The deviation level determination unit divides the deviation level according to the deviation value and outputs the corresponding level identifier.

[0028] Specifically, the preset threshold storage unit of the combustion state deviation analysis module adopts a distributed database architecture, storing the optimal threshold ranges for 20 key parameters under 18 typical test conditions. Each threshold range is determined through statistical analysis of 1000 sets of benchmark test data, with the confidence interval for the upper and lower limits set to 99%. The database supports 100 threshold call requests per second and has an automatic condition matching retrieval function with a response time of less than 10 milliseconds. The real-time data calibration unit has built-in error models for 20 sensors, including five types of error compensation terms such as zero-point drift and temperature drift. The correction coefficients obtained through periodic calibration tests are updated every 30 days. During the calibration process, a standard reference source is used to compare the sensor data, and the error of the corrected measurement data is reduced to less than 30% of the original error. The deviation quantification calculation unit sets separate parameters for core parameters and secondary parameters. The weighting coefficients are as follows: core parameters such as main combustion zone temperature and combustion pressure have a weighting coefficient of 0.2, while secondary parameters such as turbulence intensity and mixing uniformity have a weighting coefficient of 0.05. The deviation of each parameter from the center value of the threshold interval is obtained by distance calculation method, and then normalization is performed to map all parameter deviation values ​​to the interval between 0 and 1. The repeatability accuracy of deviation calculation reaches one-thousandth. The deviation level judgment unit sets a five-level deviation classification standard. The width of the numerical interval corresponding to each level of deviation is 0.2. Each deviation level is associated with a clear control priority coefficient. The priority coefficient of the first level deviation is 0.1, and the priority coefficient of the fifth level deviation is 1.0. The priority coefficient directly serves as an important basis for the generation of subsequent control strategies. The overall data processing cycle of this unit is 50 milliseconds, ensuring that the deviation analysis results can support control decisions in a timely manner and improve the pertinence and effectiveness of combustion state control.

[0029] Preferably, the deep temporal adaptive control module includes: a control strategy generation unit, an algorithm parameter optimization unit, a control effect prediction unit, and a control command output unit. The control strategy generation unit triggers the deep temporal adaptive combustion control algorithm based on the combustion state deviation level. The algorithm parameter optimization unit dynamically adjusts the algorithm calibration parameters according to the temporal change law of the combustion process. The control effect prediction unit simulates the evolution trend of combustion state under different control strategies. The control command output unit selects the control strategy corresponding to the optimal evolution trend and converts it into a standardized control command.

[0030] Specifically, the control strategy generation unit of the deep temporal adaptive control module triggers the corresponding deep temporal control algorithm based on the deviation level output by the combustion state deviation analysis module. The algorithm start time is less than 10 milliseconds. Initial control directions are preset for different deviation levels: level 1 to 2 deviations use a fine-tuning mode, level 3 to 4 deviations use a medium-tuning mode, and level 5 deviations use a rapid-tuning mode. Multiple candidate control schemes are generated by combining historical data sequences of the combustion process. The algorithm parameter optimization unit sets the adjustment step size of the temporal window length to 5 time steps. The window length is dynamically adjusted according to the deviation change rate. When the deviation change rate is greater than 0.02, the window length is shortened; when it is less than 0.005, the window length is extended. Simultaneously, the iteration step size is linearly correlated with the deviation value; for every 0.1 increase in the deviation value, the iteration step size increases by 0.01, ensuring that the algorithm parameters can adapt to combustion. The dynamic changes in the state; the control effect prediction unit uses numerical simulation to build a prediction model based on the current combustion state data. The prediction time is set to 1 second and the time step is 20 milliseconds. It simulates the change trajectory of combustion parameters under different control schemes and calculates the deviation convergence time and steady-state fluctuation amplitude corresponding to each scheme. The convergence time accuracy is controlled within 10 milliseconds. The control command output unit uses a multi-objective decision algorithm to evaluate candidate control schemes. The evaluation indicators include deviation convergence speed, control action amplitude, and steady-state stability. The weights of each indicator are set to 0.4, 0.3, and 0.3, respectively. The control scheme with the highest comprehensive score is selected and converted into standardized digital control commands. The command transmission rate reaches 1000 bits per second to ensure that the control strategy can be accurately and quickly transmitted to the execution module to achieve efficient correction of combustion state deviations.

[0031] Preferably, the combustion steady-state identification model adopted by the multi-field coupled combustion identification module satisfies: ,in, For the temperature distribution in the combustion zone, For the pressure distribution in the combustion chamber, This refers to the mass fraction of fuel components. The velocity vector of the flow field. The density of the combustion products, For isobaric specific heat capacity, Thermal conductivity, For the first The rate of formation of each reaction component For the first Enthalpy values ​​of the components Here is the viscous dissipation coefficient. For dynamic viscosity, This represents the number of components participating in the reaction.

[0032] Specifically, the multi-field coupled combustion steady-state identification model is based on the coupling mechanism of heat and mass transfer and chemical reaction in the combustion field. It is grounded in the laws of energy conservation, mass conservation, and momentum conservation. The model quantifies the interaction between temperature, pressure, component concentration, and flow velocity during combustion, and constructs core equations by incorporating multi-field coupling effects. Considering the variation of the thermophysical properties of combustion products with temperature and composition, the model introduces dynamic calculation relationships for key thermophysical parameters such as thermal conductivity and specific heat capacity at constant pressure. Then, through numerical discretization, the continuous combustion field control equations are transformed into solvable mathematical expressions. Simultaneously, the coupled effects of viscous dissipation and exothermic reactions are integrated to ensure the model comprehensively reflects the laws of multi-field interactions. The number of components participating in the reaction is set to eight, including the main fuel components and combustion products. The thermal conductivity ranges from 0.02 to 0.15, the specific heat capacity at constant pressure is dynamically adjusted according to the temperature range, and the viscous dissipation coefficient is set to a dynamic range of 0.001 to 0.01. During implementation, the model receives fused data from multi-field coupled combustion identification modules and achieves accurate characterization of combustion steady-state characteristics through 5000 iterative solutions. The establishment of this model solves the limitations of single physical field analysis, can accurately analyze the coupling mechanism of heat and mass transfer and chemical reaction in the combustion field, provides scientific mathematical support for combustion state identification, and improves the comprehensiveness and accuracy of identification results.

[0033] Preferably, the optimization model for the control parameters of the deep temporal adaptive control module satisfies: ,in, To regulate the objective function, For speed adjustment weighting coefficients, For pressure regulation weighting coefficient, For component regulation weighting coefficients, For reference temperature distribution, For reference pressure distribution, For reference component mass fraction, This is time-series temperature data. For time-series stress data, The mass fraction of the time-series components. This is the deviation weighting coefficient. To regulate time.

[0034] Specifically, the control parameter optimization model takes minimizing combustion state deviation as its core objective. Combining the smoothness requirements of control actions, a multi-objective optimization function is constructed. Based on statistical analysis of time-series combustion data and sensitivity analysis of control parameters, the correlation between deviation quantification and control parameters is established. The squared terms of deviations in temperature, pressure, and component concentration are integrated using a weighted summation method, assigning weight coefficients to different deviation terms to reflect parameter importance. A spatial gradient term for the control parameters is then introduced to constrain drastic changes in the control parameters and ensure the stability of the control process. The deviation weight coefficients for temperature, pressure, and component concentration are set to 0.4, 0.3, and 0.3, respectively, while the control weight coefficient ranges from 0.1 to 0.8, dynamically adjusted according to combustion conditions. The reference value range is determined based on 1000 sets of benchmark test data. During implementation, the model receives deviation data from the combustion state deviation analysis module and, by dynamically adjusting the time window length and iteration step size, completes the optimization calculation within 1 second, generating the optimal combination of control parameters. The establishment of this model realizes multi-objective optimization of the control strategy, which ensures that the combustion state converges quickly to the optimal range, while avoiding combustion instability caused by overly drastic control actions, thereby improving the accuracy and stability of the control process.

[0035] Preferably, the combustion state deviation quantification model satisfies:

[0036] in, This is the overall deviation value. To monitor the number of parameters, For the first Real-time measured values ​​of each parameter For the first Optimal values ​​for each parameter For the first Standard deviation of each parameter The coefficient representing the influence of time-series changes. The parameter is the synergistic influence coefficient. For the first The rate of change of each parameter over time.

[0037] Specifically, the combustion state deviation quantification model is based on statistical analysis and time-series variation characteristics, combined with parameter distribution patterns and synergistic mechanisms, to construct a comprehensive deviation quantification expression. The root mean square method is used to calculate the degree of deviation of each parameter from its optimal value. The standard deviation of the parameters is introduced for normalization to eliminate the influence of differences in the dimensions of different parameters. An exponential term representing the time-varying rate of change of the parameters is then incorporated to reflect the dynamic evolution characteristics of the deviation. Finally, a synergistic influence term is added to quantify the contribution of the interaction between multiple parameters to the overall deviation. The number of monitored parameters is set to 20, including key parameters such as temperature, pressure, flow field, and composition. The time-series variation influence coefficient is set to 0.01 to 0.05, and the synergistic influence coefficient is set to 0.005 to 0.02. The standard deviation of each parameter is obtained through statistical analysis of 100 sets of steady-state test data. During implementation, the model receives calibrated real-time combustion data and preset optimal values, and completes comprehensive deviation quantification and level classification through 200 calculation steps, with a processing cycle of 50 milliseconds. The establishment of this model solves the problem that traditional deviation quantification methods ignore the dynamic changes and synergistic effects of parameters. It can comprehensively and accurately reflect the degree of deviation of combustion state and provide a clear basis for the generation of control strategies.

[0038] Preferably, the flow field characteristic prediction model satisfies: ,in, For the velocity of the three-dimensional spatiotemporal flow field, The initial flow field velocity distribution, Kinematic viscosity, Let be the effect function of temperature and composition. The vortex diffusion coefficient is... For spatial coordinates, For time.

[0039] Specifically, the flow field characteristic prediction model is based on the fundamental equations of fluid mechanics, combined with the influence of temperature and composition on the convective flow field during combustion. Through theoretical analysis and fitting of experimental data, a spatiotemporal evolution equation for the flow field velocity is constructed. Incorporating core fluid mechanics factors such as viscous diffusion, convective transport, and pressure gradients, the influence functions of temperature and component concentration are introduced to quantify the effects of thermal effects and component changes on the flow field. Finally, a vortex diffusion term is added to characterize the evolution characteristics of the flow field vortex structure. The kinematic viscosity is set to a dynamic range of 0.0001 to 0.001, and the vortex diffusion coefficient is set to 0.005 to 0.02. The initial flow field velocity distribution is determined based on unloaded test data from the test bench, and the temperature and component influence functions are obtained by fitting 500 sets of flow field test data under different operating conditions. During implementation, the model receives initial flow field data and real-time parameters from the dynamic acquisition module for flow field characteristics, and predicts the flow field velocity distribution within the next second through integral calculations, with a prediction step size of 20 milliseconds. The establishment of this model enables accurate prediction of flow field characteristics, allows for early prediction of flow field change trends, provides support for the early formulation of control strategies, and enhances the foresight and effectiveness of control.

[0040] Preferably, the fuel supply regulation model satisfies: , in, For real-time fuel supply, As the baseline fuel supply, The deviation response coefficient, The temperature-pressure coupling coefficient is... The component rate synergy coefficient, This is the overall deviation value. The temperature gradient is in the x-direction. This represents the pressure gradient in the y-direction.

[0041] Specifically, the fuel supply regulation model is based on the coupling relationship between fuel supply and combustion state, combined with deviation response characteristics and temporal evolution patterns, to construct a dynamic regulation equation for fuel supply. In the derivation process, an S-shaped function is used to characterize the response characteristics of fuel supply to deviations, ensuring a smooth transition in the regulation process. Integral terms of temperature and pressure gradients are then introduced to quantify the impact of the combustion field spatial distribution on fuel supply. Finally, a synergistic term for composition and velocity is added to reflect the effect of multi-parameter coupling on fuel demand. The baseline fuel supply is determined based on rated operating conditions, with deviation response coefficients set to 0.5 to 2.0, temperature and pressure coupling coefficients ranging from 0.01 to 0.05, and composition-velocity synergistic coefficients ranging from 0.005 to 0.015. During implementation, the model receives combustion state deviation data and real-time parameter gradient information, dynamically calculating and adjusting the fuel supply. The adjustment response time is less than 100 milliseconds, and the adjustment accuracy is controlled within 1%. This model achieves dynamic adaptive regulation of fuel supply, accurately matching the changing demands of combustion state, ensuring the optimal fuel-air ratio, improving combustion efficiency and stability, while avoiding fuel waste and pollutant generation.

[0042] like Figure 2As shown, a method for real-time control of combustion state on a gas turbine combustor test bench is described. This method is applied to a real-time control system for combustion state on a gas turbine combustor test bench and includes the following steps: S1, continuously capturing data on temperature, pressure, component concentration, and flow velocity inside the gas turbine combustor using distributed sensing units deployed in the multi-dimensional combustion parameter sensing module, and simultaneously acquiring information on turbulence intensity, vortex structure, and mixing uniformity in the combustion area using the laser velocimetry component and particle image velocimetry device of the flow field characteristic dynamic acquisition module; S2, transmitting the captured multi-dimensional parameters and flow field information to the multi-field coupled combustion identification module, and performing feature extraction and coupling analysis on the data using the constructed multi-field coupled combustion steady-state identification model to analyze the coupling mechanism of heat and mass transfer and chemical reaction in the combustion field; S3, using the combustion state deviation analysis module to call a preset combustion threshold range, and processing the data input to the identification module. After error calibration of the combustion state data, the deviation of the real-time combustion state from the optimal range is quantified and the deviation level is divided by difference calculation and weight allocation; S4, based on the deviation level, the deep time-series adaptive combustion control algorithm is triggered. In the deep time-series adaptive control module, the time window length and iteration step size of the algorithm are dynamically adjusted, and a multi-dimensional control strategy is generated by combining the time-series evolution law of the combustion process; S5, the generated control strategy is transmitted to the control command precision execution module. The module's built-in actuator drives the flow regulating valve of the fuel supply pipeline, the angle adjusting blade of the air intake duct, and the energy output component of the ignition system to complete the corresponding adjustment action; S6, the parameters and flow field information in the combustion chamber after adjustment are continuously captured. The parameter acquisition, coupling identification, deviation analysis, strategy generation, and command execution process are repeated to form a closed-loop control link for real-time dynamic optimization of the combustion state.

[0043] The formula in this invention integrates different scalar and vector parameters for unified calculation. Based on the laws of physical conservation and multi-field coupling mechanisms, it establishes the essential relationship between parameters and eliminates calculation conflicts caused by differences in parameter types through standardization, dimensional adaptation, and coupling term construction. For scalar parameters such as temperature, pressure, and component concentration, the formula maps their values ​​to a unified order of magnitude range through statistical normalization, and then reflects the importance ratio of different scalars in combustion control through weight allocation. For vector parameters such as flow field velocity and gradient distribution, the formula extracts scalar features such as magnitude, directional components, or spatial gradients, transforming vector information into quantitative indicators that can be calculated in conjunction with scalars. At the same time, coupling terms are introduced to quantify the interaction between scalars and vectors. For example, in the combustion steady-state identification model, the integration of scalars such as temperature and pressure with the flow field velocity vector is achieved by establishing the relationship between heat and mass transfer and momentum change through the energy conservation equation. The convective transport effect of the velocity vector is transformed into a scalar energy transfer term, and then coupled with coupling coefficients such as viscous dissipation, mathematical compatibility of different types of parameters is achieved. The fuel supply regulation model, on the other hand, transforms spatial distribution information into a scalar correction term through the integral operation of vectors such as temperature and pressure gradients. This correction term, along with scalar parameters such as the deviation response coefficient, participates in the dynamic calculation of fuel supply. This design is based on the physical nature of the interdependence and synergistic effect of various parameters during combustion. Through mathematical methods, it achieves the adaptation and integration of parameter types, ensuring that the formula can comprehensively reflect the multi-field coupling characteristics and provide a scientific and unified calculation basis for combustion state regulation.

[0044] A real-time combustion state control system and method for a gas turbine combustor test bench, through a specially designed multi-dimensional combustion parameter sensing module and a dynamic flow field characteristic acquisition module, achieves comprehensive capture of multi-dimensional data such as internal combustion chamber temperature, pressure, component concentration, flow velocity vector, flow field turbulence intensity, and vortex structure, avoiding the limitations of single-parameter monitoring. A multi-field coupled combustion identification module deeply integrates these heterogeneous data, systematically analyzes the coupling law between heat and mass transfer and chemical reactions in the combustion field, and accurately extracts steady-state combustion characteristics, providing a comprehensive and reliable basis for control decisions. This completely changes the current situation where existing technologies rely on independent analysis of only a few parameters and cannot characterize the complex characteristics of combustion.

[0045] This system and method possess deep temporal adaptive control capabilities and a closed-loop optimization mechanism, successfully solving the problems of delayed response and insufficient accuracy in existing technologies. The combustion state deviation analysis module provides a clear direction for control by quantifying the degree of deviation and classifying deviation levels; the deep temporal adaptive control module dynamically adjusts control parameters and strategies based on the temporal evolution of the combustion process, ensuring that control actions accurately adapt to changes in combustion state; the precise execution module of control commands efficiently converts control strategies, forming a complete closed-loop control chain in conjunction with continuous data acquisition and feedback. This design can quickly respond to transient changes and cumulative deviations in combustion state, ensuring that the combustion process is always within the optimal range, significantly improving the accuracy and repeatability of test data on the test bench, and perfectly adapting to the testing requirements of high-parameter gas turbines with multi-condition switching and transient load changes.

[0046] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "set," "install," "connect," "link," and "fix" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal communication between two components. Those skilled in the art will understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0047] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various equivalent changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A real-time combustion state control system for a gas turbine combustion chamber test bench, characterized in that, include: Combustion parameter multidimensional sensing module, flow field characteristic dynamic acquisition module, multi-field coupled combustion identification module, combustion state deviation analysis module, deep temporal adaptive control module, and control command precise execution module; The combustion parameter multidimensional sensing module and the flow field characteristic dynamic acquisition module interact bidirectionally. The output of the flow field characteristic dynamic acquisition module is connected to the input of the multi-field coupled combustion identification module. The multi-field coupled combustion identification module establishes a feedback link with the deep time-series adaptive control module through the combustion state deviation analysis module. The output of the deep time-series adaptive control module is connected to the input of the control command precision execution module. The combustion parameter multidimensional sensing module captures the temperature distribution, pressure fluctuation, component concentration, and flow velocity vector information inside the gas turbine combustion chamber. The flow field characteristic dynamic acquisition module acquires data on the turbulence intensity, vortex structure evolution, and gas-solid mixing homogeneity of the combustion region. The multi-field coupled combustion identification module constructs a combustion steady-state identification model based on the aforementioned data and analyzes the heat and mass transfer law of the combustion field. The combustion state deviation analysis module compares the identification results with the preset combustion threshold range and quantifies the degree of deviation. The deep time-series adaptive control module triggers the deep time-series combustion adaptive control algorithm to generate a control strategy based on the deviation data. The control command precision execution module transforms the control strategy into adjustment actions for the burner fuel supply, air intake angle, and ignition energy intensity.

2. The real-time combustion state control system for a gas turbine combustor test bench according to claim 1, characterized in that, The multi-field coupled combustion identification module includes: a combustion field data fusion unit, a thermal parameter correlation analysis unit, a combustion steady-state feature extraction unit, and a multi-field coupled model construction unit. The combustion field data fusion unit receives heterogeneous data transmitted from the combustion parameter multi-dimensional sensing module and the flow field characteristic dynamic acquisition module, and integrates temperature, pressure, and flow field related data through a multi-source information complementary fusion method. The thermal parameter correlation analysis unit mines the nonlinear mapping relationship between parameters in different combustion stages. The combustion steady-state feature extraction unit separates combustion steady-state calibration feature quantities from the fused data and removes transient interference components. The multi-field coupled model construction unit establishes a combustion steady-state identification model based on the feature quantities, including the interaction mechanism of thermal field, flow field, and component field, to accurately characterize the combustion state in multiple dimensions.

3. The real-time combustion state control system for a gas turbine combustor test bench according to claim 1, characterized in that, The combustion state deviation analysis module includes: a preset threshold storage unit, a real-time data calibration unit, a deviation quantification calculation unit, and a deviation level determination unit. The preset threshold storage unit stores the optimal threshold range for combustion state under different operating conditions. The real-time data calibration unit performs system error correction on the identification results output by the multi-field coupled combustion identification module. The deviation quantification calculation unit calculates the deviation value between the real-time combustion state and the optimal threshold range through difference calculation and weight allocation. The deviation level determination unit divides the deviation level according to the deviation value and outputs the corresponding level label.

4. The real-time combustion state control system for a gas turbine combustion chamber test bench according to claim 1, characterized in that, The deep temporal adaptive control module includes: a control strategy generation unit, an algorithm parameter optimization unit, a control effect prediction unit, and a control command output unit. The control strategy generation unit triggers the deep temporal adaptive combustion control algorithm based on the combustion state deviation level. The algorithm parameter optimization unit dynamically adjusts the algorithm calibration parameters according to the temporal change law of the combustion process. The control effect prediction unit simulates the evolution trend of combustion state under different control strategies. The control command output unit selects the control strategy corresponding to the optimal evolution trend and converts it into a standardized control command.

5. The real-time combustion state control system for a gas turbine combustor test bench according to claim 1, characterized in that, The combustion steady-state identification model used by the multi-field coupled combustion identification module satisfies: ， in, For the temperature distribution in the combustion zone, For the pressure distribution in the combustion chamber, This refers to the mass fraction of fuel components. The velocity vector of the flow field. The density of the combustion products, For isobaric specific heat capacity, Thermal conductivity, For the first The rate of formation of each reaction component For the first Enthalpy values ​​of the components Here is the viscous dissipation coefficient. For dynamic viscosity, This represents the number of components participating in the reaction.

6. The real-time combustion state control system for a gas turbine combustor test bench according to claim 1, characterized in that, The optimization model for the control parameters of the deep temporal adaptive control module satisfies: ,in, To regulate the objective function, For speed control weighting coefficients, For pressure regulation weighting coefficient, For component regulation weighting coefficients, For reference temperature distribution, For reference pressure distribution, For reference component mass fraction, This is time-series temperature data. For time-series stress data, The mass fraction of the time-series components. This is the deviation weighting coefficient. To regulate time.

7. The real-time combustion state control system for a gas turbine combustor test bench according to claim 1, characterized in that, The combustion state deviation quantification model satisfies: ， in, This is the overall deviation value. To monitor the number of parameters, For the first Real-time measured values ​​of each parameter For the first Optimal values ​​for each parameter For the first Standard deviation of each parameter The coefficient representing the influence of time-series changes. The parameter is the synergistic influence coefficient. For the first The rate of change of each parameter over time.

8. The real-time combustion state control system for a gas turbine combustion chamber test bench according to claim 1, characterized in that, The flow field characteristic prediction model satisfies: ,in, For the velocity of the three-dimensional spatiotemporal flow field, The initial flow field velocity distribution, Kinematic viscosity, Let be the effect function of temperature and composition. The vortex diffusion coefficient is... For spatial coordinates, For time.

9. The real-time combustion state control system for a gas turbine combustor test bench according to claim 1, characterized in that, The fuel supply regulation model satisfies: ， in, For real-time fuel supply, As the baseline fuel supply, The deviation response coefficient, The temperature-pressure coupling coefficient is... The component rate synergy coefficient, This is the overall deviation value. The temperature gradient is in the x-direction. This represents the pressure gradient in the y-direction.

10. A method for real-time control of combustion state on a gas turbine combustion chamber test bench, characterized in that, This method is applied to a real-time combustion state control system for a gas turbine combustor test bench as described in claim 1, comprising the following steps: S1, continuously capturing internal temperature, pressure, component concentration, and flow velocity data of the gas turbine combustor through a distributed sensing unit deployed in the multi-dimensional combustion parameter sensing module, and simultaneously acquiring information related to turbulence intensity, vortex structure, and mixing uniformity of the combustion area flow field using the laser velocimetry component and particle image velocimetry device of the flow field characteristic dynamic acquisition module; S2, transmitting the captured multi-dimensional parameters and flow field information to the multi-field coupled combustion identification module, and performing feature extraction and coupling analysis on the data through the constructed multi-field coupled combustion steady-state identification model to analyze the coupling mechanism of heat and mass transfer and chemical reaction in the combustion field; S3, using the combustion state deviation analysis module to call a preset combustion threshold range to process the combustion state data output by the identification module. After line error calibration, the deviation between the real-time combustion state and the optimal range is quantified and the deviation level is divided by difference calculation and weight allocation; S4, based on the deviation level, the deep time-series adaptive combustion control algorithm is triggered. In the deep time-series adaptive control module, the time window length and iteration step size of the algorithm are dynamically adjusted, and a multi-dimensional control strategy is generated by combining the time-series evolution law of the combustion process; S5, the generated control strategy is transmitted to the control command precision execution module. The module's built-in actuator drives the flow regulating valve of the fuel supply pipeline, the angle adjusting blade of the air intake duct, and the energy output component of the ignition system to complete the corresponding adjustment action; S6, the parameters and flow field information in the combustion chamber after adjustment are continuously captured. The parameter acquisition, coupling identification, deviation analysis, strategy generation, and command execution process are repeated to form a closed-loop control link for real-time dynamic optimization of the combustion state.

Images