A cascade aerodynamic probe rapid measurement method based on online system identification

Abstract

The present application relates to the field of aero-engine, gas turbine aerodynamic thermal test, discloses a kind of cascade aerodynamic probe fast measurement method based on online system identification, by regarding aerodynamic measurement pipeline as a dynamic system, according to the time series pressure signal of first pressure measuring point, the ARX model and pressure prediction model of first pressure measuring point are constructed, then the ARX model parameter of last pressure measuring point of current pressure measuring point is used to construct the initial prediction model of steady pressure of current pressure measuring point;And according to the pressure measurement value of a small amount of sampling time of current pressure measuring point, the steady pressure prediction model of aerodynamic measurement pipeline at current pressure measuring point is fitted in real time, and then the final steady pressure prediction value is deduced. High-precision steady value can be obtained in the early stage of physical response of aerodynamic measurement pipeline, the single-point measurement efficiency is improved, the operation cost of test is greatly reduced, and the fast convergence and robustness of steady pressure prediction model of each pressure measuring point under severe working condition can also be guaranteed.

Description

Technical Field

[0001] This invention relates to the field of aerodynamic and thermodynamic testing of aero-engines and gas turbines, and discloses a rapid measurement method for blade aerodynamic probes based on online system identification. Background Technology

[0002] In the planar blade cascade tests of aero-engine compressors and turbines, in order to obtain detailed aerodynamic performance of the blade profile (such as total pressure loss coefficient, lag angle, work done, etc.), it is usually necessary to use a multi-hole probe to perform high-resolution point-by-point scanning measurements of its outlet flow field.

[0003] Traditional planar blade aerodynamic measurement processes typically follow a serial pattern of "probe relocation—passive waiting—data acquisition." Existing related technologies mainly focus on improving the accuracy of single-point measurements or correcting sensor errors. For example, Chinese patent CN114077775B discloses a dynamic pressure intelligent measurement method for aero-engines. This method addresses the difficulty of measuring the dynamic total pressure at the compressor outlet by proposing an intelligent compensation algorithm that focuses on correcting single-point dynamic pressure signals. Chinese patent CN102803659A discloses a correction method and apparatus for measuring the pressure of gas flow within an aero-engine, improving measurement accuracy by estimating and correcting sensor offset errors.

[0004] However, when faced with the complex flow field environment of cascade experiments, existing technologies still have the following significant shortcomings:

[0005] First, it cannot adapt to the complex dynamic characteristics of pipelines, resulting in low efficiency: The pneumatic measurement system, composed of the fine tubing inside the probe and the pressure scanning valve cavity, is essentially a damped-inertial system. Under transonic conditions, this system often exhibits second-order underdamped (damped oscillation) or second-order overdamped (slow ascent) characteristics. Existing technologies typically rely solely on statistical characteristics such as the mean and variance of the sliding window to determine "pipeline stability," failing to predict the final equilibrium position of oscillation convergence. It must passively wait for the physical oscillations to completely decay, leading to extremely long single-point measurement times (typically tens to hundreds of seconds).

[0006] Second, it is prone to misjudgment in unsteady flow field regions: In the cascade suction surface separation region or trailing edge wake region, the flow field itself has inherent physical unsteady fluctuations (such as Karman vortex streets and corner stall cluster shedding). Traditional variance-based steady-state judgment logic has difficulty distinguishing between "pipeline response instability" and "physical unsteadiness of the flow field," often misjudging physical fluctuations as pipeline instability, causing the measurement system to fall into a long dead loop and be unable to output measurement results.

[0007] Third, the lack of utilization of geometric prior information leads to slow convergence: Existing methods treat each measurement point as an independent measurement event, ignoring the strict geometric periodicity of the flow field in a planar blade cascade. The flow field characteristics (pressure level, flow angle) of adjacent blade channels at the same relative position are highly similar. Existing technologies do not utilize this valuable prior information to assist and accelerate the parameter identification process of the measurement system, resulting in the need to establish equilibrium from scratch for each measurement point. Summary of the Invention

[0008] The purpose of this invention is to provide a rapid measurement method for blade aerodynamic probes based on online system identification. This method can obtain high-precision steady-state values ​​in the early stages of the physical response of the aerodynamic measurement pipeline, improve single-point measurement efficiency, significantly reduce the operating cost of the test, and ensure the rapid convergence and robustness of the steady-state pressure prediction model for each pressure measurement point under harsh operating conditions.

[0009] To achieve the above-mentioned technical effects, the technical solution adopted by the present invention is as follows:

[0010] A rapid measurement method for blade cascade aerodynamic probes based on online system identification includes:

[0011] Step S1: Start collecting time-series pressure signals from the first pressure measurement point of the blade cascade test piece at the set sampling frequency, establish an ARX model describing the dynamic characteristics of the aerodynamic measurement pipeline at the first pressure measurement point, and construct a pressure prediction model for the first pressure measurement point of the blade cascade test piece based on the ARX model parameters.

[0012] Step S2: Construct the discrete characteristic equation based on the ARX model parameters of the first pressure measuring point, and solve for the two poles of the first pressure measuring point;

[0013] Step S3: After the probe is moved to the next pressure measurement point, the two poles of the first pressure measurement point are corrected based on the ratio between the initial pressure measurement value at the first sampling moment of the current pressure measurement point and the pressure prediction value at the last sampling moment of the pressure prediction model of the first pressure measurement point before the shift, so as to obtain the corrected two poles.

[0014] Step S4: Update the initial values ​​of the ARX model parameters with the corrected two poles to obtain the initial ARX model for the current pressure measurement point;

[0015] Step S5: Collect the pressure sequence of the current pressure measurement point, and use the recursive least squares method to update the initial ARX model parameters and the corresponding steady-state pressure prediction model of the current pressure measurement point in real time until the statistical variance of the predicted steady-state value of the steady-state pressure prediction model of the current pressure measurement point is less than the preset threshold. Then, terminate the pressure collection of the current pressure measurement point and output the predicted steady-state value.

[0016] Step S6: Move the probe to the next pressure measurement point, construct the discrete characteristic equation using the ARX model parameters of the previous pressure measurement point, and solve for the two poles of the corresponding pressure measurement point. Based on the ratio between the initial pressure measurement value at the first sampling moment of the current pressure measurement point and the pressure prediction value at the last sampling moment before the shift of the updated steady-state pressure prediction model of the previous pressure measurement point, correct the two poles of the previous pressure measurement point. Repeat steps S4 to S5 until the pressure measurement of all pressure measurement points is completed.

[0017] Furthermore, the ARX model is expressed in the form of a second-order difference equation as follows: ,in For the first The pressure measuring point is at the first Pressure measurement at the sampling time For the first The pressure measuring point is at the first Pressure measurement at the sampling time For the first The pressure measuring point is at the first Pressure measurement at the sampling time , The first The coefficients of the ARX model parameters at the sampling time. For the first The constant term of the ARX model parameters at the sampling time. For the first White noise at the sampling time.

[0018] Furthermore, the pressure prediction model constructed based on the ARX model parameters is... ,in For the first The first pressure measurement point The pressure prediction value of the pressure prediction model corresponding to the ARX model parameters at the sampling time.

[0019] Furthermore, in step S1, the pressure prediction model corresponding to the pressure measurement value fluctuation amplitude of the first pressure measuring point being less than the preset fluctuation threshold is taken as the steady-state pressure prediction model of the first pressure measuring point; in step S3, the two poles of the first pressure measuring point are corrected according to the ratio between the initial pressure measurement value at the first sampling moment of the current pressure measuring point and the pressure prediction value at the last sampling moment of the steady-state pressure prediction model of the first pressure measuring point before the shift.

[0020] Furthermore, the discrete characteristic equation constructed based on the ARX model parameters of the pressure measuring points is as follows: And solve for the two poles of the corresponding pressure measuring point. , ,in For the first The variables in the discrete characteristic equation corresponding to each pressure measurement point , For the first The coefficients in the steady-state pressure prediction model for each pressure measurement point.

[0021] Furthermore, the two poles after correction: , ,in , These are the corrected poles. , Two poles , The length of the mold, This is the pole scaling adjustment factor. , For the first The initial pressure measurement value at the first sampling time of each pressure measuring point. For the first The pressure prediction value of the steady-state pressure prediction model after updating the pressure measurement points at the last sampling time before the shift.

[0022] Furthermore, the updated initial values ​​for the ARX model parameters are as follows: , , ,in , The first The coefficients of the initial ARX model parameters after updating each pressure measurement point. For the first The constant term of the initial ARX model parameters after updating each pressure measurement point.

[0023] Furthermore, if the predicted steady-state value in step S5 continues to fluctuate and the statistical variance cannot meet the preset threshold, a Fourier transform is performed on the pressure sequence and the synchronously acquired airflow angle sequence. If the characteristic frequency of the main frequency peak in the spectrum after the Fourier transform is within the physical frequency band corresponding to the characteristic Strouhal number of the blade shedding vortex, and the signal-to-noise ratio of the amplitude to the mean of the broadband background noise exceeds the preset judgment threshold, it is judged as physically unsteady, and the statistical mean of the data window is output. If the spectrum is broadband noise, it is judged as the aerodynamic measurement pipeline is unstable, and the process returns to step S3 or step S6 to continue sampling.

[0024] Compared with the prior art, the beneficial effects of the present invention are: the present invention not only eliminates the problem of large test error caused by the use of a large-range scanning valve when the reference pressure is fixed, but also enhances the adaptability of the measurement system to complex flow environments; and significantly improves the flow measurement accuracy under complex flow fields by eliminating abnormal data, providing a strong guarantee for the accurate measurement of air flow at the inlet of the test piece. Attached Figure Description

[0025] Figure 1 This is a flowchart of the rapid measurement method for blade cascade aerodynamic probes based on online system identification in Example 1;

[0026] Figure 2 This is a flowchart of the rapid measurement method for blade cascade aerodynamic probes based on online system identification in Example 2. Detailed Implementation

[0027] The present invention will now be described in further detail with reference to the embodiments and accompanying drawings. However, this should not be construed as limiting the scope of the above-described subject matter of the present invention to the following embodiments; all technologies implemented based on the content of the present invention fall within the scope of the present invention.

[0028] Example 1

[0029] See Figure 1 A rapid measurement method for blade cascade aerodynamic probes based on online system identification includes:

[0030] Step S1: Start collecting time-series pressure signals from the first pressure measurement point of the blade cascade test piece at the set sampling frequency, establish an ARX model describing the dynamic characteristics of the aerodynamic measurement pipeline at the first pressure measurement point, and construct a pressure prediction model for the first pressure measurement point of the blade cascade test piece based on the ARX model parameters.

[0031] Step S2: Construct the discrete characteristic equation based on the ARX model parameters of the first pressure measuring point, and solve for the two poles of the first pressure measuring point;

[0032] Step S3: After the probe is moved to the next pressure measurement point, the two poles of the first pressure measurement point are corrected based on the ratio between the initial pressure measurement value at the first sampling moment of the current pressure measurement point and the pressure prediction value at the last sampling moment of the pressure prediction model of the first pressure measurement point before the shift, so as to obtain the corrected two poles.

[0033] Step S4: Update the initial values ​​of the ARX model parameters with the corrected two poles to obtain the initial ARX model for the current pressure measurement point;

[0034] Step S5: Collect the pressure sequence of the current pressure measurement point, and use the recursive least squares method to update the initial ARX model parameters and the corresponding steady-state pressure prediction model of the current pressure measurement point in real time until the statistical variance of the predicted steady-state value of the steady-state pressure prediction model of the current pressure measurement point is less than the preset threshold. Then, terminate the pressure collection of the current pressure measurement point and output the predicted steady-state value.

[0035] Step S6: Move the probe to the next pressure measurement point, construct the discrete characteristic equation using the ARX model parameters of the previous pressure measurement point, and solve for the two poles of the corresponding pressure measurement point. Based on the ratio between the initial pressure measurement value at the first sampling moment of the current pressure measurement point and the pressure prediction value at the last sampling moment before the shift of the updated steady-state pressure prediction model of the previous pressure measurement point, correct the two poles of the previous pressure measurement point. Repeat steps S4 to S5 until the pressure measurement of all pressure measurement points is completed.

[0036] In this embodiment, combining the spatial correlation, geometric periodicity, and aerodynamic physical laws of the cascade flow field, the aerodynamic measurement pipeline is regarded as a dynamic system. Based on the time-series pressure signal of the first pressure measurement point, an ARX model and pressure prediction model for the first pressure measurement point are constructed. Then, the initial steady-state pressure prediction model for the current pressure measurement point is constructed using the ARX model parameters of the previous pressure measurement point. Based on the pressure measurement values ​​at a small number of sampling times at the current pressure measurement point, the steady-state pressure prediction model of the aerodynamic measurement pipeline at the current pressure measurement point is fitted in real time using an online recursive least squares algorithm, thereby deriving the final steady-state pressure prediction value. In particular, high-precision steady-state values ​​can be obtained in the early stage of the physical response of the aerodynamic measurement pipeline, improving single-point measurement efficiency and significantly reducing the operating cost of the experiment. In addition, it can effectively cope with the measurement challenges in the region of drastic back pressure changes in the transonic flow field, ensuring the rapid convergence and robustness of the system under harsh operating conditions.

[0037] In this embodiment, the ARX model is expressed in the form of a second-order difference equation. ,in For the first The pressure measuring point is at the first Pressure measurement at the sampling time For the first The pressure measuring point is at the first Pressure measurement at the sampling time For the first The pressure measuring point is at the first Pressure measurement at the sampling time , The first The coefficients of the ARX model parameters at the sampling time. For the first The constant term of the ARX model parameters at the sampling time. For the first White noise at the sampling time. White noise represents the sum of unpredictable external disturbances, sensor electronic thermal noise, and unmodelable high-frequency system errors. Those skilled in the art know its statistical prior properties (e.g., it usually follows a Gaussian distribution with a mean of 0). Based on statistical properties and measured data, white noise can be estimated and is considered as known data.

[0038] The pressure prediction model built based on ARX model parameters is as follows: ,in For the first The first pressure measurement point The pressure prediction value of the pressure prediction model corresponding to the ARX model parameters at the sampling time.

[0039] Example 2

[0040] See Figure 2 This embodiment uses a high-load transonic planar blade cascade wind tunnel test as an example to describe in detail the rapid measurement method of the blade cascade aerodynamic probe based on online system identification of the present invention. The test incoming flow Mach number was 0.8, and a five-hole probe was used to perform a full-coverage scan of the flow field at the blade cascade exit, with a sampling frequency of... The frequency was set to 100Hz. The test method is as follows:

[0041] Step S1: Start collecting time-series pressure signals from the first pressure measurement point of the blade cascade test piece at the set sampling frequency, establish an ARX model describing the dynamic characteristics of the aerodynamic measurement pipeline at the first pressure measurement point, and construct a pressure prediction model for the first pressure measurement point of the blade cascade test piece based on the ARX model parameters.

[0042] In this embodiment, a second-order ARX model is established for the pneumatic measurement pipeline of a single pressure measurement orifice of the five-hole probe to describe its pressure response characteristics. The ARX model is expressed in the form of a second-order difference equation. ,in For the first The pressure measuring point is at the first Pressure measurement at the sampling time For the first The pressure measuring point is at the first Pressure measurement at the sampling time For the first The pressure measuring point is at the first Pressure measurement at the sampling time , The first The coefficients of the ARX model parameters at the sampling time. For the first The constant term of the ARX model parameters at the sampling time. For the first White noise at the sampling time.

[0043] Therefore, the pressure prediction model constructed based on the ARX model parameters is ,in For the first The first pressure measurement point The pressure prediction value of the pressure prediction model corresponding to the ARX model parameters at the sampling time.

[0044] Step S2: Construct the discrete characteristic equation based on the ARX model parameters of the first pressure measuring point, and solve for the two poles of the first pressure measuring point;

[0045] In this embodiment, the discrete characteristic equation constructed based on the ARX model parameters of the pressure measuring points is: And solve for the two poles of the first pressure measuring point. , ,in For the first pressure measurement point, the variables in the discrete characteristic equation are... , This refers to the coefficients in the steady-state pressure prediction model for the first pressure measurement point.

[0046] Step S3: After the probe is moved to the next pressure measurement point, the two poles of the first pressure measurement point are corrected based on the ratio between the initial pressure measurement value at the first sampling moment of the current pressure measurement point and the pressure prediction value at the last sampling moment of the pressure prediction model of the first pressure measurement point before the shift, so as to obtain the corrected two poles.

[0047] It should be noted that the present invention uses the ARX model parameters of the previous pressure measurement point as a benchmark to construct the initial steady-state pressure prediction model of the next pressure measurement point as the initial value of the ARX model parameters of the next pressure measurement point. In order to improve the accuracy of the initial steady-state pressure prediction model of the next pressure measurement point, in step S1 of this embodiment, the pressure prediction model corresponding to the pressure measurement value fluctuation amplitude of the first pressure measurement point being less than the preset fluctuation threshold should be used as the steady-state pressure prediction model of the first pressure measurement point, so as to output the pressure prediction value at the last sampling time of the first pressure measurement point.

[0048] In this embodiment, the two poles of the first pressure measuring point after correction are: , ,in , These are the corrected poles. , Two poles , The length of the mold, This is the pole scaling adjustment factor. , This represents the initial pressure measurement value at the first sampling time of the second pressure measuring point. This represents the pressure prediction value of the steady-state pressure prediction model at the last sampling moment before the shift at the first pressure measurement point.

[0049] Step S4: Update the initial values ​​of the ARX model parameters with the corrected two poles to obtain the initial ARX model for the current pressure measurement point;

[0050] In this embodiment, the updated initial values ​​of the ARX model parameters are: , , ,in , These are the coefficient terms of the initial ARX model parameters after the second pressure measurement point is updated. This is a constant term for the initial ARX model parameters after the second pressure measurement point is updated.

[0051] Step S5: Collect the pressure sequence of the current pressure measurement point, and use the recursive least squares method to update the initial ARX model parameters and the corresponding steady-state pressure prediction model of the current pressure measurement point in real time until the statistical variance of the predicted steady-state value of the steady-state pressure prediction model of the current pressure measurement point is less than the preset threshold. Then, terminate the pressure collection of the current pressure measurement point and output the predicted steady-state value.

[0052] In this embodiment, the pressure probe uses a fixed sampling frequency. High-frequency pressure sampling is continuously performed at 100Hz, and the model parameters are updated online in real time using the recursive least squares algorithm with a forgetting factor (FF-RLS). The specific update process is as follows:

[0053] S5.1 Transformation of Linear Regression Equation:

[0054] The above second-order ARX model is rewritten in standard linear regression form. ; where data observation vector Defined as This includes the historical pressure measurements of the current pressure measurement point at the two most recent moments; and the parameter vector to be fitted. This reflects the dynamic damping and inertial characteristics of the pipeline.

[0055] S5.2 Online Parameter Fitting and Recursive Update:

[0056] With sampling time With advancements, the FF-RLS algorithm tracks and updates the model parameter vector in real time by minimizing the weighted sum of squared errors. The algorithm employs techniques known in the art, and performs the following four recursive calculations sequentially for each sampling period:

[0057] 1) Calculate the Kalman gain vector;

[0058] 2) Solve for the equivalent prediction error term of the pneumatic measurement pipeline at the current pressure measuring point;

[0059] 3) Update the model parameter vector ;

[0060] 4) Update the error covariance matrix And utilize the forgetting factor The covariance matrix is ​​adaptively shrunk to suppress parameter estimation divergence.

[0061] Step S6: Move the probe to the next pressure measurement point. Construct a discrete characteristic equation using the ARX model parameters of the previous pressure measurement point after each shift and solve for the two poles of the corresponding pressure measurement point. Based on the ratio between the initial pressure measurement value at the first sampling moment of the current pressure measurement point and the pressure prediction value at the last sampling moment before the shift of the updated steady-state pressure prediction model of the previous pressure measurement point, correct the two poles of the previous pressure measurement point. Repeat steps S4 to S5 until the pressure measurement of all pressure measurement points is completed.

[0062] The ARX model is expressed in the form of a second-order difference equation. ,in For the first The pressure measuring point is at the first Pressure measurement at the sampling time For the first The pressure measuring point is at the first Pressure measurement at the sampling time For the first The pressure measuring point is at the first Pressure measurement at the sampling time , The first The coefficients of the ARX model parameters at the sampling time. For the first The constant term of the ARX model parameters at the sampling time. For the first White noise at the sampling time.

[0063] The pressure prediction model built based on ARX model parameters is as follows: ,in For the first The first pressure measurement point The pressure prediction value of the pressure prediction model corresponding to the ARX model parameters at the sampling time.

[0064] The discrete characteristic equation constructed based on the ARX model parameters of the pressure measurement points is as follows: And solve for the two poles of the corresponding pressure measuring point. , ,in For the first The variables in the discrete characteristic equation corresponding to each pressure measurement point , For the first The coefficients in the steady-state pressure prediction model for each pressure measurement point. The two poles after correction: , ,in , These are the corrected poles. , Two poles , The length of the mold, This is the pole scaling adjustment factor. , For the first The initial pressure measurement value at the first sampling time of each pressure measuring point. For the first The pressure prediction value of the steady-state pressure prediction model after updating the pressure measurement points at the last sampling time before the shift.

[0065] Updated initial ARX model parameter values: , , ,in , The first The coefficients of the initial ARX model parameters after updating each pressure measurement point. For the first The constant term of the initial ARX model parameters after updating each pressure measurement point.

[0066] In some other embodiments, if the predicted steady-state value in step S5 continues to fluctuate and the statistical variance is less than a preset threshold, a Fourier transform is performed on the pressure sequence and the synchronously acquired airflow angle sequence. If the characteristic frequency of the main frequency peak in the spectrum after the Fourier transform is within the physical frequency band corresponding to the characteristic Strouhal number of the blade shedding vortex, and the signal-to-noise ratio of the amplitude to the mean of the broadband background noise exceeds a preset judgment threshold, it is judged as physically unsteady, and the statistical mean of the data window is output. If the spectrum is broadband noise, it is judged as the aerodynamic measurement pipeline is unstable, and the process returns to step S3 or step S6 to continue sampling.

[0067] Those skilled in the art should understand that when this invention is applied to multi-hole pneumatic measurement scenarios such as five-hole or seven-hole probes, the control system will independently and synchronously execute the online parameter fitting and steady-state prediction process of steps S1 to S6 for each pressure measurement hole on the probe. Regarding the scheduling and control of the measurement point movement, a comprehensive command to move the probe to the next spatial measurement point can only be issued if and only if all parallel pressure measurement channels on the multi-hole probe individually meet the steady-state prediction confidence test condition of step S5 (or if an individual channel triggers the output cutoff in step S6).

[0068] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A rapid measurement method for blade cascade aerodynamic probes based on online system identification, characterized in that, include: Step S1: Start collecting time-series pressure signals from the first pressure measurement point of the blade cascade test piece at the set sampling frequency, establish an ARX model describing the dynamic characteristics of the aerodynamic measurement pipeline at the first pressure measurement point, and construct a pressure prediction model for the first pressure measurement point of the blade cascade test piece based on the ARX model parameters. Step S2: Construct the discrete characteristic equation based on the ARX model parameters of the first pressure measuring point, and solve for the two poles of the first pressure measuring point; Step S3: After the probe is moved to the next pressure measurement point, the two poles of the first pressure measurement point are corrected based on the ratio between the initial pressure measurement value at the first sampling moment of the current pressure measurement point and the pressure prediction value at the last sampling moment of the pressure prediction model of the first pressure measurement point before the shift, so as to obtain the corrected two poles. Step S4: Update the initial values ​​of the ARX model parameters with the corrected two poles to obtain the initial ARX model for the current pressure measurement point; Step S5: Collect the pressure sequence of the current pressure measurement point, and use the recursive least squares method to update the initial ARX model parameters and the corresponding steady-state pressure prediction model of the current pressure measurement point in real time until the statistical variance of the predicted steady-state value of the steady-state pressure prediction model of the current pressure measurement point is less than the preset threshold. Then, terminate the pressure collection of the current pressure measurement point and output the predicted steady-state value. Step S6: Move the probe to the next pressure measurement point, construct the discrete characteristic equation using the ARX model parameters of the previous pressure measurement point, and solve for the two poles of the corresponding pressure measurement point. Based on the ratio between the initial pressure measurement value at the first sampling moment of the current pressure measurement point and the pressure prediction value at the last sampling moment before the shift of the updated steady-state pressure prediction model of the previous pressure measurement point, correct the two poles of the previous pressure measurement point. Repeat steps S4 to S5 until the pressure measurement of all pressure measurement points is completed.

2. The rapid measurement method for blade cascade aerodynamic probe according to claim 1, characterized in that, The ARX model is expressed in the form of a second-order difference equation. ,in For the first The pressure measuring point is at the first Pressure measurement at the sampling time For the first The pressure measuring point is at the first Pressure measurement at the sampling time For the first The pressure measuring point is at the first Pressure measurement at the sampling time , The first The coefficients of the ARX model parameters at the sampling time. For the first The constant term of the ARX model parameters at the sampling time. For the first White noise at the sampling time.

3. The rapid measurement method for blade cascade aerodynamic probe according to claim 2, characterized in that, The pressure prediction model built based on ARX model parameters is as follows: ,in For the first The first pressure measurement point The pressure prediction value of the pressure prediction model corresponding to the ARX model parameters at the sampling time.

4. The rapid measurement method for blade cascade aerodynamic probe according to claim 3, characterized in that, In step S1, the pressure prediction model corresponding to the pressure measurement value fluctuation amplitude of the first pressure measuring point is less than the preset fluctuation threshold, which is taken as the steady-state pressure prediction model of the first pressure measuring point. In step S3, the two poles of the first pressure measuring point are corrected according to the ratio between the initial pressure measurement value at the first sampling time of the current pressure measuring point and the pressure prediction value at the last sampling time of the steady-state pressure prediction model of the first pressure measuring point before the shift.

5. The rapid measurement method for blade cascade aerodynamic probe according to claim 4, characterized in that, The discrete characteristic equation constructed based on the ARX model parameters of the pressure measurement points is as follows: And solve for the two poles of the corresponding pressure measuring point. , ,in For the first The variables in the discrete characteristic equation corresponding to each pressure measurement point , For the first The coefficients in the steady-state pressure prediction model for each pressure measurement point.

6. The rapid measurement method for blade cascade aerodynamic probe according to claim 1, characterized in that, The two poles after correction: , ,in , These are the corrected poles. , Two poles , The length of the mold, This is the pole scaling adjustment factor. , For the first The initial pressure measurement value at the first sampling time of each pressure measuring point. For the first The pressure prediction value of the steady-state pressure prediction model after updating the pressure measurement points at the last sampling time before the shift.

7. The rapid measurement method for blade cascade aerodynamic probe according to claim 6, characterized in that, Updated initial ARX model parameter values: , , ,in , The first The coefficients of the initial ARX model parameters after updating for each pressure measurement point. For the first The constant term of the initial ARX model parameters after updating each pressure measurement point.

8. The rapid measurement method for blade cascade aerodynamic probe according to claim 1, characterized in that, If the predicted steady-state value in step S5 continues to fluctuate and the statistical variance is less than a preset threshold, perform Fourier transform on the pressure sequence and the synchronously acquired airflow angle sequence; if the characteristic frequency of the main frequency peak in the spectrum after Fourier transform is within the physical frequency band corresponding to the characteristic Strouhal number of the blade shedding vortex, and the signal-to-noise ratio of the amplitude to the mean of the broadband background noise exceeds the preset judgment threshold, it is judged as physically unsteady, and the statistical mean of the data window is output; if the spectrum is broadband noise, it is judged as the aerodynamic measurement pipeline is unstable, and return to step S3 or step S6 to continue sampling.

Images