A method for controlling unsteady flow in a compressor based on micro-injection of air into the casing wall

Abstract

The present invention proposes a method for controlling unsteady flow in a compressor based on micro-jets on the casing wall, which relates to the technical field of unsteady flow control. Micro-jets are introduced into the compressor through the casing wall, and the unsteady flow is regulated through closed-loop feedback. The disturbance generated by the unsteady micro-jets in the flow field is obtained, which has the same frequency and opposite phase as the original flow field fluctuations, thereby offsetting the unsteady flow. The present invention can effectively suppress unsteady flow.

Description

Technical Field

[0001] The present invention proposes a compressor unsteady flow control method based on casing wall micro-jet, which relates to the technical field of unsteady flow control, and specifically to the technical field of compressor unsteady flow control based on micro-jet. Background Art

[0002] The flow within a compressor exhibits significant unsteady characteristics, with typical phenomena including rotating stall, unsteady tip leakage flow, and the dynamic evolution of flow separation. Existing unsteady flow control technologies are primarily categorized as passive and active. Both traditional passive and active control technologies suffer from low control accuracy, weak performance, and instability, making it difficult to accurately and efficiently suppress unsteady flows. Summary of the Invention

[0003] The present invention provides a compressor unsteady flow control method based on casing wall micro-injection to solve the above problems:

[0004] The present invention proposes a method for controlling unsteady flow in a compressor based on casing wall micro-injection, the method comprising:

[0005] Microjets are introduced into the compressor through the casing wall, and the unsteady flow is regulated through closed-loop feedback. The disturbances generated by the unsteady microjets in the flow field are of the same frequency and opposite phase to the original flow field fluctuations, thereby offsetting the unsteady flow.

[0006] Obtain an unsteady flow region. Arrange n microjet nozzles on the casing in the unsteady flow region. Install a dynamic pressure sensor at a preset position corresponding to each microjet nozzle. Determine the parameters of the microjet nozzles using five variables. The total number of variables for the n microjet nozzle parameters is 5n.

[0007] The five variables include flow rate m, the time derivative of flow rate dm / dt, the angle between the jet direction and the normal of the casing wall, and the angle between the jet direction and the normal of the casing wall. , the projection of the jet direction on the casing wall and the axial angle , compressor operating flow coefficient .

[0008] Furthermore, the method of introducing microjets into the compressor through the wall of the casing and regulating the unsteady flow through closed-loop feedback obtains disturbances in the flow field generated by the unsteady microjets that are of the same frequency and opposite phase as the original flow field fluctuations, thereby offsetting the unsteady flow, including:

[0009] S1. Obtaining the coordinates and number data of given micro jets;

[0010] Determine the open and closed states of the micro-jet nozzles and obtain working state determination information of the micro-jet nozzles;

[0011] S2. Performing non-operation analysis and operation analysis based on the working status judgment information of the micro-jet nozzle to obtain non-operation analysis data and operation analysis data;

[0012] S3. Construct an unsteady micro-jet disturbance based on the non-operating analysis data and the operating analysis data to perform disturbance cancellation.

[0013] S4. Determine whether the micro-jet sufficiently suppresses the unsteady waves, and complete the establishment of flow control or perform reconstruction of dynamic modal decomposition based on the judgment information.

[0014] Furthermore, the S1 includes:

[0015] Get the switch status of each micro jet;

[0016] When the switch status information of all the micro-jet ports is in the open state, it is determined that the overall micro-jet ports are in the working state;

[0017] When the switch status information of all the micro-jet ports is in the closed state, it is determined that the overall micro-jet ports are in the non-working state;

[0018] When the switch status information of all the micro jets is both open and closed, the overall micro jets are determined to be in the working state. The switch status of the overall micro jets is the working state judgment information of the micro jets.

[0019] Furthermore, the S2 includes:

[0020] Obtaining information on the working status of the micro jet nozzle;

[0021] When the working state judgment information of the micro-jet nozzle indicates that the micro-jet nozzle is not working, the pressure time series is obtained, and dynamic mode decomposition is performed to obtain a reconstructed signal;

[0022] When the working state of the microjet nozzle is judged to be microjet working, online learning is performed through machine learning to establish a mapping relationship between the microjet parameters and the generated disturbance.

[0023] Furthermore, when the working state judgment information of the micro jet is that the micro jet is not working, obtaining a pressure time series, performing dynamic mode decomposition, and obtaining a reconstructed signal include:

[0024] Construct unsteady mathematical equations of flow field;

[0025] When the microjet is not working, the solid-wall casing structure is analyzed, and the correlation analysis between the disturbance measured by the pressure sensor and the jet is not performed;

[0026] Perform dynamic modal decomposition on the time series snapshot p(t) consisting of n pressure sensor signals p1 to pn fixed at different positions on the wall of the casing, perform bandpass filtering on the dynamic modal decomposition spectrum, and select the first r modes for reconstruction to obtain the reconstructed signal;

[0027] The calculation formula of the reconstructed signal is:

[0028]

[0029] in, is the angular frequency, is the phase, and ψ is the modal amplitude.

[0030] Furthermore, when the working state judgment information of the micro-jet port indicates that the micro-jet is working, online learning is performed through machine learning to establish a mapping relationship between the micro-jet parameters and the generated disturbance, including:

[0031] Establishing a mapping relationship between micro-jet parameters and generated disturbances, wherein the disturbance is determined by two variables: pressure p monitored by a dynamic pressure sensor and its time derivative dp / dt;

[0032] The mapping relationship is obtained by adjusting micro-jet parameters, establishing machine learning samples, and performing online learning using a neural network.

[0033] Furthermore, the S3 includes:

[0034] By reconstructing the signal and mapping relationship, an unsteady micro-jet disturbance is constructed, so that the pressure sensor receives the same frequency and inverse phase signal of the reconstructed signal.

[0035] The calculation formula of the same-frequency reverse signal is:

[0036]

[0037] Furthermore, the S4 includes:

[0038] To judge whether the microjet can adequately suppress the unsteady waves, an error function e is introduced;

[0039]

[0040] in,

[0041]

[0042] Determine whether the error function e meets the preset target, obtain the target judgment result, and perform adjustment operations based on the target judgment result.

[0043] Furthermore, the step of determining whether the error function e satisfies a preset target, obtaining a target determination result, and performing an adjustment operation according to the target determination result includes:

[0044] When the error function e meets the preset target, the establishment of flow control is completed;

[0045] When the error function e does not meet the preset target, the selected modal number r is increased and the dynamic modal analysis is corrected.

[0046] Furthermore, the control system includes:

[0047] A micro jet state analysis module is used to obtain the coordinates and number data of a given micro jet;

[0048] Determine the open and closed states of the micro-jet nozzles and obtain working state determination information of the micro-jet nozzles;

[0049] a classification analysis module, configured to perform non-operation analysis and operation analysis based on the working status judgment information of the micro-jet nozzle, and obtain non-operation analysis data and operation analysis data;

[0050] The disturbance cancellation module is used to construct unsteady micro-jet disturbances based on the non-working analysis data and the working analysis data to perform disturbance cancellation;

[0051] The judgment module is completed to judge whether the micro-jet can sufficiently suppress the unsteady wave, and the flow control is established or the dynamic modal decomposition is reconstructed according to the judgment information.

[0052] Beneficial effects of the invention: The present invention proposes an active control method for unsteady flow in a compressor based on casing wall micro-injection and closed-loop feedback. Compared with the existing technology, the present invention has the following significant advantages:

[0053] (1) Accurate matching of dynamic response and unsteady flow

[0054] Closed-loop feedback control: Dynamic pressure sensors are used to monitor flow field pressure fluctuations in real time, and combined with dynamic mode decomposition (DMD), the dominant frequency and phase characteristics of the unsteady flow are extracted to ensure that the micro-jet disturbance is in the same frequency and opposite phase as the original flow fluctuation, thereby actively offsetting unsteady disturbances (such as rotating stall and surge precursor oscillations).

[0055] Adaptive frequency domain matching: Through DMD spectrum bandpass filtering and reconstruction, key unsteady modes (such as low-frequency rotating stall and high-frequency tip vortex breakup) are automatically screened, avoiding control failures caused by frequency domain mismatch in traditional technologies (such as synthetic jets).

[0056] (2) Multi-parameter collaborative optimization and efficient energy utilization

[0057] Multi-degree-of-freedom jet parameter design: Coordinated control of the control variables of each micro-jet port breaks through the limitations of single parameters (such as steady jet) in existing technologies.

[0058] Machine learning-driven mapping relationship modeling: A neural network is used to establish a nonlinear mapping relationship between jet parameters and disturbance effects, dynamically optimizing the jet strategy to avoid energy waste or insufficient control caused by fixed jet parameters in traditional methods.

[0059] (3) High precision and low invasiveness design

[0060] Non-contact dynamic sensing: The wall-embedded pressure sensor is used to avoid the insufficient spatial resolution and clogging effect of traditional contact measurement (such as LDV), while the small size design of the micro-jet nozzle reduces additional interference to the flow field.

[0061] Lightweight and low-cost: No complex mechanical structure (such as variable guide vanes) is required, and only micro-injection and sensor integration are achieved through the casing wall, reducing system complexity and maintenance costs. It is suitable for extreme environments such as high temperature and high pressure. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] Figure 1 Schematic diagram of a compressor unsteady flow control method based on case wall micro-injection;

[0063] Figure 2 Schematic diagram of the main ideas;

[0064] Figure 3 Schematic diagram of the casing wall micro-injection control process based on online learning. DETAILED DESCRIPTION

[0065] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0066] In one embodiment of the present invention, a method for controlling unsteady flow in a compressor based on casing wall micro-injection is proposed, the method comprising:

[0067] Microjets are introduced into the compressor through the casing wall, and the unsteady flow is regulated through closed-loop feedback. The disturbances generated by the unsteady microjets in the flow field are of the same frequency and opposite phase to the original flow field fluctuations, thereby offsetting the unsteady flow.

[0068] Obtain an unsteady flow region, arrange n micro jets on the casing in the unsteady flow region, install a dynamic pressure sensor at a preset position corresponding to each micro jet, and determine the parameters of the micro jet using five variables. The total number of variables for the n micro jet parameters is 5n; the corresponding preset position is within a preset range centered on the micro jet and does not affect the installation of other micro jets and dynamic pressure sensors, such as Figure 2 As shown;

[0069] The five variables include flow rate m, the time derivative of flow rate dm / dt, the angle between the jet direction and the normal of the casing wall, and the angle between the jet direction and the normal of the casing wall. , the projection of the jet direction on the casing wall and the axial angle , compressor operating flow coefficient .

[0070] The working principle of the above technical solution is: the existing unsteady flow control technology is mainly divided into two categories: passive control and active control.

[0071] (1) Passive control technology

[0072] Casing treatment structures such as slots and circumferential grooves suppress the development of leakage vortices by changing the flow structure of the blade tip clearance;

[0073] The blade leading edge serration design delays separation by promoting boundary layer mixing.

[0074] (2) Active control technology

[0075] Synthetic jet technology uses high-frequency oscillating jets (frequency can reach kHz level) to inject momentum into the flow field, disrupting the vortex structure in the separation zone;

[0076] Plasma actuators generate induced airflow by ionizing gas, changing the local flow direction, and are suitable for flow reattachment control in high-temperature environments;

[0077] Variable guide vane adjustment changes the airflow angle of attack by mechanically adjusting the blade angle to adapt to the flow requirements under different working conditions.

[0078] The present invention introduces microjets into the compressor through the wall of the casing, adjusts the unsteady flow through closed-loop feedback, and obtains disturbances in the flow field generated by the unsteady microjets that are of the same frequency and opposite phase as the original flow field fluctuations, thereby offsetting the unsteady flow and improving the defects of the existing technology.

[0079] The main defects of existing technologies include:

[0080] Passive control technologies, such as casing treatment structures (such as slots and circumferential grooves), suppress leakage vortex development by altering tip clearance flow. However, these structural modifications can result in significant aerodynamic efficiency losses, such as increased flow resistance or the induction of localized secondary flows. Furthermore, the optimization effects of such structures are typically limited to specific operating conditions; deviations from the design point can exacerbate flow separation or induce new instabilities. In engineering applications, wear issues associated with high-speed rotation and high temperatures can reduce casing structural reliability and even lead to crack propagation due to stress concentration, threatening the overall life of the engine.

[0081] The serrated design of the blade leading edge delays flow separation by promoting boundary layer mixing. However, the serrated geometry increases friction losses on the blade surface, potentially reducing stage efficiency, especially in high-speed flows. The complex serrated edges present precision challenges during machining, requiring costly specialized processes such as five-axis milling or 3D printing. Furthermore, the mechanical strength of the blade leading edge may be compromised by the weakened geometry, making it difficult to withstand high-frequency vibration loads, increasing the risk of fatigue fracture during long-term operation.

[0082] Active control technologies, including synthetic jets, rely on high-frequency oscillating jets to disrupt the vortex structure in the separation zone. However, their core drawback lies in their heavy reliance on external energy systems and complex control logic. The actuator requires continuous energy input to maintain the jet, significantly increasing system energy consumption. Furthermore, the precise matching of the kHz-level jet frequency with the flow field dynamics places stringent demands on sensors and controllers. Any phase deviation or response delay can exacerbate flow instability. Furthermore, piezoelectric materials or electromagnetic drive components are prone to thermal failure in high-temperature and high-pressure environments, leading to soaring maintenance costs.

[0083] Plasma actuators induce local flow direction changes by ionizing gas. However, the high-voltage ionization process can cause electromagnetic interference to onboard electronic equipment, and the rapid ablation of electrode materials at high temperatures can shorten service life. While this technology is suitable for high-temperature environments, its high ionization energy consumption and low efficiency make it difficult to achieve global flow control in large compressors, limiting its use to localized auxiliary measures.

[0084] Variable guide vane adjustment mechanically adjusts the blade angle to suit varying operating conditions. However, its mechanical drive system (such as connecting rods and servo mechanisms) significantly increases the weight and complexity of the compression components, running counter to the need for lightweight aircraft engines. Frequent guide vane angle adjustment can accelerate wear of bearings and transmission components. Thermal expansion of metal at high temperatures can also reduce adjustment accuracy. Dynamic response delays further hinder the ability to cope with rapid changes in transient operating conditions.

[0085] The technical effects of the above technical solution are as follows: Compared with the existing technology, this method has the following significant advantages:

[0086] (1) Accurate matching of dynamic response and unsteady flow

[0087] Closed-loop feedback control: Dynamic pressure sensors are used to monitor flow field pressure fluctuations in real time, and combined with dynamic mode decomposition (DMD), the dominant frequency and phase characteristics of the unsteady flow are extracted to ensure that the micro-jet disturbance is in the same frequency and opposite phase as the original flow fluctuation, thereby actively offsetting unsteady disturbances (such as rotating stall and surge precursor oscillations).

[0088] Adaptive frequency domain matching: Through DMD spectrum bandpass filtering and reconstruction, key unsteady modes (such as low-frequency rotating stall and high-frequency tip vortex breakup) are automatically screened, avoiding control failures caused by frequency domain mismatch in traditional technologies (such as synthetic jets).

[0089] (2) Multi-parameter collaborative optimization and efficient energy utilization

[0090] Multi-degree-of-freedom jet parameter design: Coordinated control of the control variables of each micro-jet port breaks through the limitations of single parameters (such as steady jet) in existing technologies.

[0091] Machine learning-driven mapping relationship modeling: A neural network is used to establish a nonlinear mapping relationship between jet parameters and disturbance effects, dynamically optimizing the jet strategy to avoid energy waste or insufficient control caused by fixed jet parameters in traditional methods.

[0092] (3) High precision and low invasiveness design

[0093] Non-contact dynamic sensing: The wall-embedded pressure sensor is used to avoid the insufficient spatial resolution and clogging effect of traditional contact measurement (such as LDV), while the small size design of the micro-jet nozzle reduces additional interference to the flow field.

[0094] Lightweight and low-cost: No complex mechanical structure (such as variable guide vanes) is required, and only micro-injection and sensor integration are achieved through the casing wall, reducing system complexity and maintenance costs. It is suitable for extreme environments such as high temperature and high pressure.

[0095] In one embodiment of the present invention, the microjets are introduced into the compressor through the wall of the casing, and the unsteady flow is regulated through closed-loop feedback to obtain disturbances in the flow field generated by the unsteady microjets that are of the same frequency and opposite phase as the original flow field fluctuations, thereby offsetting the unsteady flow, including:

[0096] S1. Obtaining the coordinates and number data of given micro jets;

[0097] Determine the open and closed states of the micro-jet nozzles and obtain working state determination information of the micro-jet nozzles;

[0098] S2. Performing non-operation analysis and operation analysis based on the working status judgment information of the micro-jet nozzle to obtain non-operation analysis data and operation analysis data;

[0099] S3. Construct an unsteady micro-jet disturbance based on the non-operating analysis data and the operating analysis data to perform disturbance cancellation.

[0100] S4. Determine whether the microjet has sufficiently suppressed the unsteady wave, and complete the establishment of flow control or the reconstruction of dynamic modal decomposition based on the judgment information, such as Figure 1 shown.

[0101] The working principle of the above technical solution is as follows: Figure 2 As shown in the figure, micro-jets are introduced on the casing wall. For the unsteady flow, closed-loop feedback regulation is used to make the unsteady micro-jets produce disturbances in the flow field that are in the same frequency and opposite phase as the original flow field fluctuations, thereby offsetting the unsteady flow.

[0102] like Figure 2 As shown in Figure 1, n micro jets are arranged on the casing near the unsteady flow region, and a dynamic pressure sensor is installed near each micro jet. The parameters of each micro jet are calculated by the flow rate m, the time derivative of the flow rate dm / dt, and the angle between the jet direction and the normal to the casing wall. , the projection of the jet direction on the casing wall and the axial angle , compressor operating flow coefficient A total of 5 variables are determined, and the total number of variables of n micro-jet nozzle parameters is 5n.

[0103] Step 1: Establish mathematical equations to describe the unsteadiness of the flow field. When the microjet is not working, the research object is a solid-wall casing structure. At this time, the disturbance measured by the pressure sensor is unrelated to the jet. The time series snapshot p(t) consisting of n pressure sensor signals p1, p2, ..., pn fixed at different positions on the wall is subjected to dynamic mode decomposition, and the dynamic mode decomposition (DMD) spectrum is band-pass filtered. The first r modes are selected for reconstruction to obtain the reconstructed signal .

[0104] Step 2: Establish a mapping relationship between microjet parameters and the generated disturbance. The disturbance is determined by two variables: pressure (p) monitored by a pressure sensor and its time derivative (dp / dt). By varying the jet parameters, machine learning samples are created. A neural network (such as a radial basis function neural network (RBFNN)) is used for online learning to establish this mapping relationship.

[0105] Step 3: Reconstruction in step 1 According to the mapping relationship established in step 2, an unsteady micro jet disturbance is constructed so that the pressure sensor receives The same frequency and opposite phase signal :

[0106] Step 4: Determine whether the microjet has sufficiently suppressed the unsteady waves. Introduce the error function e:

[0107] in,

[0108]

[0109] Determine whether the error e meets the established target. If not, return to step 1 and increase the selected modal number r. In addition, it is planned to study the influence of the coordinates and number of micro-jet nozzles on rotational instability. Through multiple rounds of iteration, the unsteady flow active control scheme suitable for the current research object and working conditions is finally determined, such as Figure 3 shown.

[0110] The technical effect of the above technical solution is: real-time monitoring of flow field pressure fluctuations through dynamic pressure sensors, and combined with dynamic mode decomposition (DMD) to extract the dominant frequency and phase characteristics of unsteady flow, ensuring that the micro-jet disturbance and the original flow fluctuation are at the same frequency and anti-phase, and actively offsetting unsteady disturbances (such as rotating stall and surge precursor oscillations).

[0111] Through DMD spectrum bandpass filtering and reconstruction, key unsteady modes (such as low-frequency rotating stall and high-frequency tip vortex breakup) are automatically screened to avoid control failure caused by frequency domain mismatch in traditional technologies (such as synthetic jets).

[0112] The coordinated control of the control variables of each micro-jet port breaks through the limitations of single parameters (such as steady jet) in existing technologies.

[0113] Machine learning-driven mapping relationship modeling: A neural network is used to establish a nonlinear mapping relationship between jet parameters and disturbance effects, dynamically optimizing the jet strategy to avoid energy waste or insufficient control caused by fixed jet parameters in traditional methods.

[0114] The use of wall-embedded pressure sensors avoids the problems of insufficient spatial resolution and clogging effect of traditional contact measurements (such as LDV), while the small size design of the micro-jet nozzle reduces additional interference to the flow field.

[0115] There is no need for complex mechanical structures (such as variable guide vanes). Only micro-jet injection and sensor integration on the casing wall are required to reduce system complexity and maintenance costs. It is suitable for extreme environments such as high temperature and high pressure.

[0116] In one embodiment of the present invention, S1 includes:

[0117] Get the switch status of each micro jet;

[0118] When the switch status information of all the micro-jet ports is in the open state, it is determined that the overall micro-jet ports are in the working state;

[0119] When the switch status information of all the micro-jet ports is in the closed state, it is determined that the overall micro-jet ports are in the non-working state;

[0120] When the switch status information of all the micro jets is both open and closed, the overall micro jets are determined to be in the working state. The switch status of the overall micro jets is the working state judgment information of the micro jets.

[0121] The working principle of the above technical solution is: obtaining the switch status information of each micro-jet outlet, which is detected and obtained through sensors or control signals, etc., and can reflect the current status of each micro-jet outlet in real time and accurately.

[0122] Based on the acquired switch status information, a logical judgment is performed: if all microjet outlets are in the open state, the entire microjet outlet is determined to be in an operational state. If all microjet outlets are in the closed state, the entire microjet outlet is determined to be in an inoperative state. If the microjet outlets are both in the open and closed states, the entire microjet outlet is also determined to be in an operational state. This achieves efficient judgment of the entire microjet outlet. The judgment result is output as information on the operational state of the microjet outlets.

[0123] The technical effect of the above technical solution is: by real-time monitoring of the on / off status of each micro-jet, timely detection and response to the state changes of the micro-jet, thereby improving the reliability and stability of the entire system.

[0124] The decision logic is clear and concise, making it easy to implement and program. The overall micro-jet operating status can be determined by performing a simple logical judgment based on the acquired switch status information, reducing the complexity of the control logic.

[0125] Different combinations of switch states are allowed between the micro-jet ports. As long as the judgment conditions are met, the overall micro-jet port can be considered to be in working state.

[0126] In one embodiment of the present invention, the S2 includes:

[0127] Obtaining information on the working status of the micro jet nozzle;

[0128] When the working state judgment information of the micro-jet nozzle indicates that the micro-jet nozzle is not working, the pressure time series is obtained, and dynamic mode decomposition is performed to obtain a reconstructed signal;

[0129] When the working state of the microjet nozzle is judged to be microjet working, online learning is performed through machine learning to establish a mapping relationship between the microjet parameters and the generated disturbance.

[0130] The working principle of the above technical solution is to obtain information on the working status of the micro-jet by real-time monitoring or receiving signals from sensors. Specifically, it is determined whether the micro-jet is currently in working state based on comprehensive judgment of parameters such as the opening and closing status of the micro-jet, flow rate, and pressure.

[0131] When the system determines that the microjet is not operating, it further acquires pressure time series data for analysis corresponding to the non-operational state. Dynamic Mode Decomposition (DMD) is performed on the acquired pressure time series. DMD is a data-driven analysis method that can extract the system's dynamic characteristics, such as modes, frequencies, and growth rates, from time series data.

[0132] Through DMD analysis, the reconstructed signal can be obtained.

[0133] When the system determines the microjet's operating status is "active," it initiates online learning with a machine learning algorithm. During this learning process, it collects microjet parameter data (such as jet volume, jet frequency, and jet angle) as well as corresponding disturbance data (such as airflow disturbances and pressure changes). This data is used to train the machine learning model, enabling it to learn the mapping between microjet parameters and generated disturbances.

[0134] Through online learning, the mapping relationship model is continuously updated and optimized.

[0135] The technical effect of the above technical solution is: by obtaining the working status judgment information of the micro-air jet, the working status of the micro-air jet can be monitored in real time.

[0136] When the microjet nozzle is not working, the reconstructed signal is obtained through dynamic mode decomposition, which can provide a deep understanding of the dynamic characteristics of the system.

[0137] When the microjet is operating, machine learning establishes a mapping relationship between microjet parameters and the resulting disturbance, allowing for more accurate prediction and control of the disturbance generated by the microjet. Mapping accuracy can be improved by adjusting the microjet's operating parameters.

[0138] The online learning mechanism enables the system to continuously adapt to various data, automatically update and optimize model parameters, and improve the accuracy of mapping relationships.

[0139] In one embodiment of the present invention, when the working state judgment information of the microjet nozzle indicates that the microjet nozzle is not working, obtaining a pressure time series, performing dynamic mode decomposition, and obtaining a reconstructed signal includes:

[0140] Construct unsteady mathematical equations of flow field;

[0141] When the microjet is not working, the solid-wall casing structure is analyzed, and the correlation analysis between the disturbance measured by the pressure sensor and the jet is not performed;

[0142] Perform dynamic modal decomposition on the time series snapshot p(t) consisting of n pressure sensor signals p1 to pn fixed at different positions on the wall of the casing, perform bandpass filtering on the dynamic modal decomposition spectrum, and select the first r modes for reconstruction to obtain the reconstructed signal;

[0143] The calculation formula of the reconstructed signal is:

[0144]

[0145] in, is the angular frequency, is the phase, and ψ is the modal amplitude.

[0146] The working principle of the above technical solution is as follows: Before processing the pressure time series when the microjet is not operating, it is necessary to first construct a mathematical equation describing the unsteady nature of the flow field. When the microjet is not operating, the solid-walled casing structure is analyzed to obtain the inherent dynamic characteristics of the casing structure in the absence of air jets. During this time, the correlation analysis between the disturbance measured by the pressure sensor and the air jet is not performed because the microjet is not operating, so the air jet has no effect on the disturbance of the pressure sensor.

[0147] Dynamic mode decomposition (DMD) is performed on a time series snapshot p(t) consisting of signals from n pressure sensors p1 to pn fixed at different locations on the casing wall. DMD is a data-driven analysis method that can extract the system's dynamic modes from time series data, including their frequency, growth rate, and mode shape.

[0148] The DMD spectrum is band-pass filtered to remove noise and unnecessary frequency components and retain the frequency components related to the dynamic characteristics of the flow field.

[0149] The first r modes are selected for reconstruction to obtain the reconstructed signal.

[0150] The technical effect of the above technical solution is that by constructing the unsteady mathematical equations of the flow field and performing dynamic modal decomposition, we can deeply understand the dynamic characteristics of the flow field when the micro-jet nozzle is not working. We can also accurately identify the natural frequency, modal shape, and growth rate of the flow field.

[0151] Bandpass filtering and mode selection steps remove noise and unwanted frequency components, retaining signals related to the flow field dynamics. This improves the accuracy of signal processing and enables the reconstructed signal to better approximate the original pressure time series.

[0152] By reconstructing the signal, the flow field state can be reconstructed and predicted.

[0153] Analysis of the reconstructed signal can identify abnormal behavior or failure modes in the flow field. Furthermore, based on the analysis of the reconstructed signal, the casing structure or operating parameters can be adjusted to optimize flow field performance.

[0154] In one embodiment of the present invention, when the microjet operating state judgment information indicates that the microjet is operating, online learning is performed through machine learning to establish a mapping relationship between microjet parameters and generated disturbances, including:

[0155] Establishing a mapping relationship between micro-jet parameters and generated disturbances, wherein the disturbance is determined by two variables: pressure p monitored by a dynamic pressure sensor and its time derivative dp / dt;

[0156] The mapping relationship is obtained by adjusting micro-jet parameters, establishing machine learning samples, and performing online learning using a neural network.

[0157] The working principle of this technical solution is as follows: the disturbance is determined by two variables: pressure p and its time derivative dp / dt, as monitored by a dynamic pressure sensor. Pressure p reflects the current pressure state of the flow field, while the time derivative dp / dt reflects the rate of change of pressure over time, i.e., the dynamic characteristics of pressure. These two variables together constitute the characteristic vector that describes the disturbance.

[0158] To establish a mapping between microjet parameters and the resulting disturbance, it is necessary to adjust these parameters. These parameters may include jet volume, jet frequency, jet angle, and so on. By varying these parameters, it is possible to observe changes in the flow field disturbance under different parameter settings.

[0159] During the microjet parameter adjustment process, a dynamic pressure sensor is used to record the pressure p and its time derivative dp / dt, forming a series of sample data. Each sample data contains a set of microjet parameter settings and the corresponding disturbance feature vector.

[0160] A neural network was used for machine learning, using the sample data as a training set. The neural network's input layer received the microjet parameters, and the output layer predicted the disturbance feature vector (i.e., pressure p and time derivative dp / dt).

[0161] By training the neural network, it learns the complex mapping relationship between microjet parameters and disturbances. During the training process, the neural network continuously adjusts its internal weights to minimize prediction errors.

[0162] The online learning mechanism allows the neural network to continuously receive new sample data during actual operation and update its internal weights in real time to adapt to changes in flow field conditions and adjustments to microjet parameters.

[0163] The technical effect of the above technical solution is that by establishing a mapping relationship between microjet parameters and generated disturbances, the disturbance of the flow field under different microjet parameter settings can be predicted more accurately.

[0164] The online learning mechanism enables the neural network to adapt to changes in flow field conditions and adjustments to microjet parameters in real time, enhancing the system's adaptive capabilities and enabling it to maintain stable performance under different operating conditions.

[0165] By predicting the disturbance conditions, the setting of microjet parameters can be optimized to minimize the adverse effects on the flow field or maximize the desired disturbance effect, which can improve the overall performance and efficiency of the system.

[0166] In one embodiment of the present invention, S3 includes:

[0167] By reconstructing the signal and mapping relationship, an unsteady micro-jet disturbance is constructed, so that the pressure sensor receives the same frequency and inverse phase signal of the reconstructed signal.

[0168] The calculation formula of the same-frequency reverse signal is:

[0169]

[0170] The working principle of the above technical solution is: the reconstructed signal is extracted and reconstructed from the pressure time series through methods such as dynamic mode decomposition (DMD), reflecting the dynamic characteristics of the flow field under specific conditions.

[0171] The mapping relationship is the association between microjet parameters and the generated disturbances established through machine learning (such as neural networks), which is used to predict the disturbance conditions under different microjet parameter settings.

[0172] Based on the reconstructed signal and mapping relationship, the unsteady microjet disturbance can be constructed.

[0173] In order for the pressure sensor to receive a signal with the same frequency and opposite phase as the reconstructed signal, the reconstructed signal needs to be phase-inverted.

[0174] The generated signals of same frequency and opposite phase are used in the control of the microjet system as feedback signals to adjust the microjet parameters to achieve precise control of the flow field disturbance.

[0175] The technical effect of the above technical solution is that by constructing unsteady micro-jet disturbances and generating signals with the same frequency and opposite phase, it is possible to achieve precise control of flow field disturbances, reduce undesirable disturbances, and improve system stability and performance.

[0176] By actively generating and applying signals of the same frequency and opposite phase, active suppression or enhancement of flow field disturbances can be achieved.

[0177] In experiments or simulations, the use of reconstructed signals and signals with the same frequency and opposite phases can more quickly evaluate the effects of different control strategies and reduce trial-and-error costs and time.

[0178] In one embodiment of the present invention, the S4 includes:

[0179] To judge whether the microjet can adequately suppress the unsteady waves, an error function e is introduced;

[0180]

[0181] in,

[0182]

[0183] Determine whether the error function e meets the preset target, obtain the target judgment result, and perform adjustment operations based on the target judgment result.

[0184] The step of determining whether the error function e satisfies a preset target, obtaining a target determination result, and performing an adjustment operation according to the target determination result includes:

[0185] When the error function e meets the preset target, the establishment of flow control is completed;

[0186] When the error function e does not meet the preset target, the selected modal number r is increased and the dynamic modal analysis is corrected.

[0187] The working principle of the above technical solution is to introduce an error function e as a quantitative indicator in the process of judging whether the micro-jet has adequately suppressed unsteady waves. The error function e is used to measure the difference between the actual flow state and the expected flow state.

[0188] The error function e is compared with a preset target. The preset target is a pre-set threshold or range used to determine whether the flow control has achieved the desired effect. If the error function e meets the preset target, the microjet is considered to have adequately suppressed the unsteady waves.

[0189] The target judgment result is obtained based on the comparison result of the error function e and the preset target.

[0190] When the target judgment result is "yes", it means that the flow control has been successfully established, and the flow control establishment process can be completed at this time.

[0191] If the target judgment result is "no," it means that the flow control has not achieved the desired effect. In this case, it is necessary to increase the selected mode number r and modify the dynamic mode decomposition (DMD). Increasing the mode number r means that more dynamic characteristics are included in the reconstructed signal, which may more accurately reflect the actual state of the flow field and thus more effectively suppress unsteady waves.

[0192] The technical effect of the above technical solution is: by introducing the error function e and performing quantitative judgment, the effect of flow control can be evaluated more accurately, thereby adjusting the control strategy in a timely manner and improving the accuracy of flow control.

[0193] When the error function e fails to meet the preset target, the system automatically increases the modal number r and modifies the DMD to adapt to changes in flow conditions and the complexity of unsteady waves. This enhances the system's adaptive capabilities, enabling it to maintain stable performance under different operating conditions.

[0194] By continuously adjusting the modal number r and correcting the DMD, the control strategy can be optimized to more effectively suppress unsteady waves. This helps reduce energy losses, improve system efficiency, and potentially extend the life of the equipment.

[0195] Through the steps of error function judgment, target judgment result acquisition and adjustment operation, a closed-loop control system is formed to reduce manual intervention and improve control efficiency.

[0196] In one embodiment of the present invention, the control system includes:

[0197] A micro jet state analysis module is used to obtain the coordinates and number data of a given micro jet;

[0198] Determine the open and closed states of the micro-jet nozzles and obtain working state determination information of the micro-jet nozzles;

[0199] a classification analysis module, configured to perform non-operation analysis and operation analysis based on the working status judgment information of the micro-jet nozzle, and obtain non-operation analysis data and operation analysis data;

[0200] The disturbance cancellation module is used to construct unsteady micro-jet disturbances based on the non-working analysis data and the working analysis data to perform disturbance cancellation;

[0201] The judgment module is completed to judge whether the micro-jet can sufficiently suppress the unsteady wave, and the flow control is established or the dynamic modal decomposition is reconstructed according to the judgment information.

[0202] The working principle of the above technical solution is: micro-jets are introduced into the wall of the casing. For unsteady flow, closed-loop feedback regulation is used to make the unsteady micro-jets produce disturbances in the flow field with the same frequency and opposite phase as the original flow field fluctuations, thereby offsetting the unsteady flow.

[0203] In the case near the unsteady flow region, n micro jets are arranged, and a dynamic pressure sensor is installed near each micro jet. The parameters of each micro jet are calculated by the flow rate m, the time derivative of the flow rate dm / dt, the angle between the jet direction and the normal of the case wall, and the pressure difference between the jet direction and the normal of the case wall. , the projection of the jet direction on the casing wall and the axial angle , compressor operating flow coefficient A total of 5 variables are determined, and the total number of variables of n micro-jet nozzle parameters is 5n.

[0204] The technical effect of the above technical solution is: real-time monitoring of flow field pressure fluctuations through dynamic pressure sensors, and combined with dynamic mode decomposition (DMD) to extract the dominant frequency and phase characteristics of unsteady flow, ensuring that the micro-jet disturbance and the original flow fluctuation are at the same frequency and anti-phase, and actively offsetting unsteady disturbances (such as rotating stall and surge precursor oscillations).

[0205] Through DMD spectrum bandpass filtering and reconstruction, key unsteady modes (such as low-frequency rotating stall and high-frequency tip vortex breakup) are automatically screened to avoid control failure caused by frequency domain mismatch in traditional technologies (such as synthetic jets).

[0206] The coordinated control of the control variables of each micro-jet port breaks through the limitations of single parameters (such as steady jet) in existing technologies.

[0207] Machine learning-driven mapping relationship modeling: A neural network is used to establish a nonlinear mapping relationship between jet parameters and disturbance effects, dynamically optimizing the jet strategy to avoid energy waste or insufficient control caused by fixed jet parameters in traditional methods.

[0208] The use of wall-embedded pressure sensors avoids the problems of insufficient spatial resolution and clogging effect of traditional contact measurements (such as LDV), while the small size design of the micro-jet nozzle reduces additional interference to the flow field.

[0209] There is no need for complex mechanical structures (such as variable guide vanes). Only micro-jet injection and sensor integration on the casing wall are required to reduce system complexity and maintenance costs. It is suitable for extreme environments such as high temperature and high pressure.

[0210] Obviously, those skilled in the art may make various modifications and variations to the present invention without departing from the spirit and scope of the present invention. Thus, if such modifications and variations fall within the scope of the claims and their equivalents, the present invention is intended to include such modifications and variations.

Claims

1. A method for controlling unsteady flow in a compressor based on casing wall micro-injection, characterized in that: The method comprises: Microjets are introduced into the compressor through the casing wall, and the unsteady flow is regulated through closed-loop feedback. The disturbances generated by the unsteady microjets in the flow field are of the same frequency and opposite phase to the original flow field fluctuations, thereby offsetting the unsteady flow. The method of introducing microjets into the compressor through the wall of the casing and regulating the unsteady flow through closed-loop feedback is used to obtain disturbances in the flow field generated by the unsteady microjets that are of the same frequency and opposite phase as the original flow field fluctuations, thereby offsetting the unsteady flow. The method includes: S1. Obtaining the coordinates and number data of given micro jets; Determine the open and closed states of the micro-jet nozzles and obtain working state determination information of the micro-jet nozzles; S2. Performing non-operation analysis and operation analysis based on the working status judgment information of the micro-jet nozzle to obtain non-operation analysis data and operation analysis data; S3. Construct an unsteady micro-jet disturbance based on the non-operating analysis data and the operating analysis data to perform disturbance cancellation. S4. Determine whether the micro-jet sufficiently suppresses the unsteady waves, and establish flow control or reconstruct the dynamic modal decomposition based on the determination information; Obtain an unsteady flow region. Arrange n microjet nozzles on the casing in the unsteady flow region. Install a dynamic pressure sensor at a preset position corresponding to each microjet nozzle. Determine the parameters of the microjet nozzles using five variables. The total number of variables for the n microjet nozzle parameters is 5n. The five variables include flow rate m, the time derivative of flow rate dm / dt, the angle between the jet direction and the normal of the casing wall, and the angle between the jet direction and the normal of the casing wall. , the projection of the jet direction on the casing wall and the axial angle , compressor operating flow coefficient .

2. The compressor unsteady flow control method based on casing wall micro-jet according to claim 1 is characterized in that: Said S1 comprises: Get the switch status of each micro jet; When the switch status information of all the micro-jet ports is in the open state, it is determined that the overall micro-jet ports are in the working state; When the switch status information of all the micro-jet ports is in the closed state, it is determined that the overall micro-jet ports are in the non-working state; When the switch status information of all the micro jets is both open and closed, the overall micro jets are determined to be in the working state. The switch status of the overall micro jets is the working state judgment information of the micro jets.

3. The compressor unsteady flow control method based on casing wall micro-jet according to claim 1 is characterized in that: The S2 includes: Obtaining information on the working status of the micro jet nozzle; When the working state judgment information of the micro-jet nozzle indicates that the micro-jet nozzle is not working, the pressure time series is obtained, and dynamic mode decomposition is performed to obtain a reconstructed signal; When the working state of the microjet nozzle is judged to be microjet working, online learning is performed through machine learning to establish a mapping relationship between the microjet parameters and the generated disturbance.

4. The compressor unsteady flow control method based on casing wall micro-jet according to claim 3 is characterized in that: When the working state judgment information of the micro-jet nozzle indicates that the micro-jet nozzle is not working, obtaining a pressure time series, performing dynamic mode decomposition, and obtaining a reconstructed signal include: Construct unsteady mathematical equations of flow field; When the microjet is not working, the solid-wall casing structure is analyzed, and the correlation analysis between the disturbance measured by the pressure sensor and the jet is not performed; Perform dynamic modal decomposition on the time series snapshot p(t) consisting of n pressure sensor signals p1 to pn fixed at different positions on the wall of the casing, perform bandpass filtering on the dynamic modal decomposition spectrum, and select the first r modes for reconstruction to obtain the reconstructed signal; The calculation formula of the reconstructed signal is: in, is the angular frequency, is the phase, and ψ is the modal amplitude.

5. The compressor unsteady flow control method based on casing wall micro-jet according to claim 3 is characterized in that: When the working state judgment information of the micro-jet nozzle indicates that the micro-jet nozzle is working, online learning is performed through machine learning to establish a mapping relationship between the micro-jet parameters and the generated disturbance, including: Establishing a mapping relationship between micro-jet parameters and generated disturbances, wherein the disturbance is determined by two variables: pressure p monitored by a dynamic pressure sensor and its time derivative dp / dt; The mapping relationship is obtained by adjusting micro-jet parameters, establishing machine learning samples, and performing online learning using a neural network.

6. The compressor unsteady flow control method based on casing wall micro-jet according to claim 1 is characterized in that: The S3 includes: By reconstructing the signal and mapping relationship, an unsteady micro-jet disturbance is constructed, so that the pressure sensor receives the same frequency and inverse phase signal of the reconstructed signal. The calculation formula of the same-frequency reverse signal is: 。 7. The compressor unsteady flow control method based on casing wall micro-jet according to claim 1, characterized in that: The S4 includes: To judge whether the microjet can adequately suppress the unsteady waves, an error function e is introduced; in, Determine whether the error function e meets the preset target, obtain the target judgment result, and perform adjustment operations based on the target judgment result.

8. The compressor unsteady flow control method based on casing wall micro-jet according to claim 7, characterized in that: The step of determining whether the error function e satisfies a preset target, obtaining a target determination result, and performing an adjustment operation according to the target determination result includes: When the error function e meets the preset target, the establishment of flow control is completed; When the error function e does not meet the preset target, the selected modal number r is increased and the dynamic modal analysis is corrected.

9. A control system for implementing the compressor unsteady flow control method based on casing wall micro-injection according to claim 1, characterized in that: The control system includes: A micro jet state analysis module is used to obtain the coordinates and number data of a given micro jet; Determine the open and closed states of the micro-jet nozzles and obtain working state determination information of the micro-jet nozzles; a classification analysis module, configured to perform non-operation analysis and operation analysis based on the working status judgment information of the micro-jet nozzle, and obtain non-operation analysis data and operation analysis data; The disturbance cancellation module is used to construct unsteady micro-jet disturbances based on the non-working analysis data and the working analysis data to perform disturbance cancellation; The judgment module is completed to judge whether the micro-jet can sufficiently suppress the unsteady wave, and the flow control is established or the dynamic modal decomposition is reconstructed according to the judgment information.

Images