A method for online identification of non-synchronous vibration of rotor blades based on limited measuring points

Abstract

The application provides a rotor blade non-synchronous vibration online identification method based on limited measuring points, and belongs to the technical field of aero-engines. The method comprises the following steps: obtaining a pulsating pressure signal and a blade tip timing signal, and constructing an autocorrelation function of the pulsating pressure signal and the blade tip timing signal; processing the autocorrelation function of the pulsating pressure signal and the blade tip timing signal to obtain a power spectral density of the pulsating pressure signal and the blade tip timing signal; performing frequency sweeping on the power spectral density in a predetermined frequency range based on an adaptive threshold, extracting a peak value signal above the adaptive threshold, and thus identifying a power spectral density peak frequency of the pulsating pressure signal and the blade tip timing signal; based on the peak frequency in the power spectral density of the pulsating pressure signal and the blade tip timing signal and a rotating speed signal, carrying out a blade non-synchronous vibration fluid-solid coupling characteristic frequency reconstruction, and obtaining a circumferential acoustic modal characteristic propagation value; and identifying the blade non-synchronous vibration according to the circumferential acoustic modal propagation characteristic value.

Description

Technical Field

[0001] This application belongs to the field of aero-engine technology, and specifically relates to an online identification method for asynchronous vibration of rotor blades based on a limited number of measuring points. Background Technology

[0002] Asynchronous blade vibration is a complex unsteady field flow phenomenon involving multiple disciplines and multi-field coupling. When asynchronous blade vibration occurs, the blade vibration stress increases sharply, which may even cause blade breakage. The frequency of asynchronous blade vibration is generally a non-integer multiple of the rotational frequency, and it has frequency-locking characteristics near the resonant speed, accompanied by flow field pulsation modes and structural vibration pitch-locking characteristics.

[0003] To achieve asynchronous blade vibration identification, it is necessary to synchronously identify the flow field inside the flow channel and the blade vibration frequency and modes online. Since asynchronous blade vibration is mostly high-nodal-diameter vibration, multiple dynamic measurement points need to be arranged inside the flow channel and on the blade for asynchronous blade vibration identification. However, in the ground test environment of the whole machine or core machine, the limited number of dynamic measurement points cannot meet the requirements of high-order mode identification. The blade asynchronous vibration monitoring method based on the principle of pulsation and blade tip timing cannot be implemented due to the huge amount of dynamic data. It can only perform feature analysis of asynchronous blade vibration in offline state, which cannot meet the urgent need for online monitoring of blade vibration through sparse measurement points in the whole machine and component environment.

[0004] One existing approach is a blade asynchronous vibration identification method based on pulsating pressure and blade dynamic stress measurement points. This method utilizes a unified dynamic testing platform combining pulsating pressure arrays and dynamic stress. It acquires the inlet pulsating and blade vibration frequencies and modal characteristics during compressor testing using pulsating pressure array sensors, enabling synchronous identification of high signal-to-noise ratio pulsating signals and structural responses within the compressor. However, this approach requires multiple microphones to be arranged circumferentially in the compressor casing and on the bladed disk to identify high-nodal-diameter vibrations. However, in ground testing scenarios for the entire engine or core unit, the number of compressor casing pulsating pressure measurement points is limited, making it impossible to meet the requirements for high-order modal identification. Furthermore, it suffers from low survival rate and poor reusability in high-speed environments, resulting in high testing costs. Therefore, it cannot meet the reliability testing and monitoring requirements of the core unit or the entire engine, and it cannot achieve asynchronous blade vibration monitoring and identification.

[0005] Another approach in the prior art is an offline identification method for asynchronous blade vibration based on the principles of pulsation and blade tip timing, such as... Figure 1As shown, this scheme installs multiple blade tip timing probes 3 on the casing 2, and a rotor probe 42 is set on the turntable part of the blade 1. A stator probe 41 is set on the casing 2. The rotor probe 42 and the stator probe 41 together constitute a speed probe assembly. The signals collected by the blade tip timing probes 3 and the speed probe assembly are sent to the signal acquisition instrument 5 for signal analysis. In this scheme, the relative displacement of the blade tip is intermittently measured by the blade tip timing probes 3. The original vibration process of the blade 1 is reconstructed based on the measured values, and the parameters of the entire vibration process of the blade 1 are analyzed. For a blade rotor system with a given number of blades Nr, blade tip diameter D, and blade speed n, when the blade speed n and blade tip diameter D are fixed, there is a certain correspondence between the blade tip arrival time difference Δt and the blade tip amplitude ΔA, such as... Figure 2 The image shows the blade pulse and rotational speed pulse sequences acquired by the blade tip timing probe and rotational speed probe components. Here, signal x(t) is the pulse signal for one rotation, x1(t) is the ideal blade tip pulse signal, x2(t) is the actual blade tip pulse signal, t1 is the arrival time of the ideal blade tip, and t2 is the arrival time of the actual blade tip. When blade 1 is not vibrating, the blade pulse sequence is as follows: Figure 2 As shown in the signal x1(t), when the blade vibrates, the resulting blade pulses will exhibit varying degrees of lead or lag, such as... Figure 2 The signal x2(t) is shown in the diagram. A blade tip timing probe 3 is installed at the corresponding casing of the rotor blade under test. As the blade rotates, the probe senses the passage of the blade tip in real time. A series of blade pulse signals are obtained via a preamplifier and signal preprocessor. The time difference Δt between each blade tip arrival at the sensor is calculated using the test system counter board. This allows for the accumulation of the arrival time difference sequence for each blade. Combined with the acquired rotational speed positioning signal, the blade tip amplitude ΔA can be calculated. Then, the actual vibration displacement of the blade tip at at least one blade tip timing probe 3 is obtained using the blade tip timing system. The relationship between the actual vibration displacement of the blade tip and time is obtained, and the actual vibration frequency, actual vibration amplitude, and actual vibration pitch diameter of the blade corresponding to the actual vibration displacement of the blade tip are calculated to determine the actual mode shape of the blade. However, the existing technical solutions have the following shortcomings: 1) This method is mainly based on the blade tip timing principle for asynchronous blade vibration monitoring, which cannot achieve synchronous acquisition with dynamic pressure measurement points. It can only analyze the blade vibration frequency and pitch diameter online, but cannot analyze the asynchronous blade vibration online; 2) This technical solution requires multiple timing measurement points (usually no less than 5) to be arranged on the casing, which results in high testing costs. It is generally used in compressor tests and is rarely used in core engine or whole machine scenarios. Summary of the Invention

[0006] The purpose of this application is to provide an online identification method for asynchronous vibration of rotor blades based on a limited number of measuring points, so as to solve or alleviate at least one of the problems in the background art.

[0007] The technical solution of this application is: an online identification method for asynchronous vibration of rotor blades based on a limited number of measuring points, including:

[0008] Acquire the pulsating pressure signal and the blade tip timing signal, and construct the autocorrelation function of the pulsating pressure signal and the blade tip timing signal;

[0009] The autocorrelation function of the pulsating pressure signal and the tip timing signal is processed to obtain the power spectral density of the pulsating pressure signal and the tip timing signal.

[0010] Based on the adaptive threshold, the power spectral density of the pulsating pressure signal and the blade tip timing signal is swept within a predetermined frequency range, and the peak signal above the adaptive threshold is extracted, thereby identifying the peak frequency of the power spectral density of the pulsating pressure signal and the blade tip timing signal.

[0011] Based on the peak frequency and rotational speed signal in the power spectral density of the pulsating pressure signal and the blade tip timing signal, fluid-structure interaction characteristic frequency reconstruction for asynchronous blade vibration is carried out to obtain the circumferential acoustic mode characteristic propagation value.

[0012] Identification of asynchronous vibration of blades based on circumferential acoustic mode propagation characteristic values.

[0013] In at least one embodiment of this application, a single pulsating pressure sensor is arranged on the axial front side of the rotor blade to obtain a pulsating pressure signal, and two blade tip timing sensors are arranged circumferentially at the blade tip of the rotor blade to obtain a blade tip timing signal.

[0014] In at least one embodiment of this application, the autocorrelation function R of the pulsating pressure signal is obtained by integrating the product of the pulsating pressure signal at different times. xx (τ):

[0015]

[0016] In the formula, x(t) is the pulsating pressure signal obtained by a single pulsating pressure sensor;

[0017] x(t+τ) is the delayed pulsating pressure signal;

[0018] τ is the delay time of the pulsating pressure signal;

[0019] t represents time;

[0020] T is the length of the pulsating pressure signal.

[0021] In at least one embodiment of this application, the autocorrelation function R of the blade tip timing signal is obtained by summing the time delays of the two blade tip timing signals. xx (ε):

[0022]

[0023] In the formula, x(n) is the timing signal of the first blade tip obtained by the timing sensor;

[0024] x(n+ε) is the timing signal of the blade tip obtained by the second blade tip timing sensor;

[0025] ε is the delay time between the two leaf tip timing signals;

[0026] n represents the measurement point timing of the leaf tip timing signal;

[0027] N is the length of the leaf tip timing signal.

[0028] In at least one embodiment of this application, the frequency range of the power spectral density sweep of the blade tip timing signal is the frequency of the Mth vibration mode before blade vibration. The scanning frequency range of the power spectral density of the pulsating pressure signal is determined according to the frequency range of the power spectral density sweep of the blade tip timing signal segment. The scanning frequency range of the power spectral density of the pulsating pressure signal is three times the scanning frequency range of the dynamic stress signal.

[0029] In at least one embodiment of this application, the adaptive threshold is the power spectral density mean + N times the power spectral density standard deviation, wherein N ranges from 2 to 3.

[0030] In at least one embodiment of this application, the circumferential acoustic modal propagation characteristic value m satisfies:

[0031] m=(f R ±f S ) / f N

[0032] In the formula, f N f is the rotor rotation frequency. R For aerodynamic characteristic frequency structure, f S These are the modal characteristic frequencies.

[0033] In at least one embodiment of this application, the process of identifying asynchronous vibrations of the blade based on circumferential acoustic mode propagation characteristic values ​​is as follows:

[0034] During the experiment, the circumferential acoustic mode propagation characteristic value calculated in real time will change with the rotational speed of the blade. When the circumferential acoustic mode propagation characteristic value m calculated in real time is not an approximate integer value, it is judged that the blade vibration is good at this time.

[0035] When the real-time calculated circumferential acoustic mode propagation characteristic value m is an approximately integer value and greater than a predetermined value, it is determined that the blade is at risk of asynchronous vibration.

[0036] When there is a risk of asynchronous vibration in the blade, and the real-time calculated circumferential acoustic mode propagation characteristic value does not change with the blade rotation speed, it is determined that the blade has already experienced asynchronous vibration.

[0037] In at least one embodiment of this application, the predetermined value is not less than 5. Attached Figure Description

[0038] To more clearly illustrate the technical solutions provided in this application, the accompanying drawings will be briefly described below. Obviously, the drawings described below are merely some embodiments of this application.

[0039] Figure 1 This is a schematic diagram of a blade asynchronous vibration detection scheme based on the principles of pulsation and blade tip timing.

[0040] Figure 2 The sequence of blade pulses and rotational speed pulses acquired by the blade tip timing probe and rotational speed probe components.

[0041] Figure 3 This is a schematic diagram of the online identification method for asynchronous vibration of rotor blades based on a limited number of measuring points proposed in this application.

[0042] Figure 4 This is a schematic diagram showing the arrangement of the pulsating pressure sensor and the blade tip timing sensor in this application. Detailed Implementation

[0043] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions in the embodiments of this application will be described in more detail below with reference to the accompanying drawings.

[0044] To address the limitation of measurement points for monitoring asynchronous vibration of blades in the context of a complete compressor or core engine, this application provides an online identification method for asynchronous vibration of rotor blades based on a limited number of measurement points. The method can achieve rapid online identification of asynchronous vibration of blades based on a single pulsating pressure sensor and two blade tip timing sensors. This method can be directly applied to the monitoring of asynchronous vibration of blades in real compressors or complete engines.

[0045] like Figure 3 As shown, the online identification method for asynchronous vibration of rotor blades based on finite measurement points provided in this application includes:

[0046] Step S10: Obtain the pulsating pressure signal and the leaf tip timing signal, and construct the autocorrelation function of the pulsating pressure signal and the leaf tip timing signal based on the pulsating pressure signal and the leaf tip timing signal, respectively.

[0047] like Figure 4As shown, in this application, a single pulsating pressure sensor 101 is arranged on the axially forward casing 104 of the rotor blade 103 to obtain pulsating pressure signals, and two blade tip timing sensors 102 are arranged circumferentially on the blade tip casing 104 of the rotor blade 103 to obtain blade tip timing signals. The blade tip timing sensors 102 can be evenly distributed or non-uniformly distributed when arranged circumferentially.

[0048] Since the asynchronous vibration of the blade is a complex fluid-structure interaction dynamic evolution process, this application uses the autocorrelation coefficient to perform autocorrelation analysis on the single-branch pulsating pressure signal and the two-point blade tip timing signal to obtain the autocorrelation function between the pulsating pressure signal and the blade tip timing signal characterizing the blade vibration.

[0049] For the pulsating pressure signal x(t), the autocorrelation function R is obtained by integrating the product of the pulsating pressure signal at different times. xx (τ), as shown in formula (1):

[0050]

[0051] In the formula, x(t) is the pulsating pressure signal obtained by a single pulsating pressure sensor;

[0052] x(t+τ) is the delayed pulsating pressure signal;

[0053] τ is the delay time of the pulsating pressure signal, which is determined based on the delay time of the blade tip timing signal.

[0054] t represents time;

[0055] T is the length of the pulsating pressure signal.

[0056] For the tip timing signal x(n), the autocorrelation function R is obtained by summing the time delays of the two tip timing signals. xx (ε), as shown in formula (2):

[0057]

[0058] In the formula, x(n) is the timing signal of the first blade tip obtained by the timing sensor;

[0059] x(n+ε) is the timing signal of the blade tip obtained by the second blade tip timing sensor;

[0060] ε is the delay time between the two blade tip timing signals, which is related to the casing angle difference, engine speed and number of rotor blades for a given blade tip timing measurement point;

[0061] n represents the measurement point timing of the leaf tip timing signal;

[0062] N is the length of the leaf tip timing signal.

[0063] Step S20: Perform a fast Fourier transform on the autocorrelation function of the pulsating pressure signal and the tip timing signal to obtain the power spectral density of the pulsating pressure signal and the tip timing signal.

[0064] Performing a Fast Fourier Transform (FFT) on the autocorrelation functions of the pulsating pressure signal and the blade tip timing signal obtained in step S10 allows for the online identification of the relationship curves between the main frequencies and their powers of the pulsating pressure signal and the blade tip timing signal. These curves represent the power spectral density of the pulsating pressure signal and the blade tip timing signal. Compared to signal spectral characteristics, this approach helps to accurately identify the pulsating frequency and blade vibration frequency within the channel, improving the signal-to-noise ratio of the characteristic signals.

[0065] In this application, the autocorrelation function R of the pulsating pressure signal is... xx The method for obtaining the power spectral density of the pulsating pressure signal by performing a fast Fourier transform (τ) is as follows:

[0066]

[0067] In the formula, F represents the Fourier transform;

[0068] -j represents the imaginary part;

[0069] f is the frequency;

[0070] e is a natural constant.

[0071] Similarly, the autocorrelation function R of the leaf tip time series signal xx The method for obtaining the power spectral density of the leaf tip time-series signal by performing a fast Fourier transform (ε) is as follows:

[0072]

[0073] Step S30: Based on the adaptive threshold, the power spectral density of the pulsating pressure signal and the blade tip timing signal is swept within a predetermined frequency range, and the peak signal above the adaptive threshold is extracted, thereby identifying the peak frequency of the power spectral density of the pulsating pressure signal and the blade tip timing signal. The peak frequency includes the aerodynamic characteristic frequency and the structural modal characteristic frequency.

[0074] Based on the power spectral density of the pulsating pressure signal and the blade tip timing signal obtained in step S20, peak spectrum scanning extraction is performed within a specific frequency range. In this application, the Mth vibration mode frequency before blade vibration is selected as the sweep frequency range of the blade tip timing signal power spectral density. In this application, the order of the vibration mode frequency is usually not less than 5; for example, the Mth vibration mode frequency can be 10. The pulsating pressure signal spectral sweep frequency range is determined according to the sweep frequency range of the blade tip timing signal. For asynchronous blade vibration, the spectral sweep frequency range of the pulsating pressure signal is 3 times the sweep frequency range of the dynamic stress signal.

[0075] The power spectral density of the pulsating pressure signal and the blade tip timing signal is swept according to the set start frequency and end frequency. The spectral interval during the sweep can be set to 1Hz, 2Hz or 5Hz, etc.

[0076] The mean and standard deviation of the power spectral density are calculated in real time online. An adaptive threshold is determined based on the mean and standard deviation. In this application, the adaptive threshold is set as mean + N × standard deviation, where N = 2 to 3, for example, N = 2.5. Frequency sweeping identification is performed on the data above the adaptive threshold in the power spectral density. Local peak signals are filtered and extracted using the peak comparison method, thereby identifying the peak frequencies in the power spectral density of the pulsating pressure signal and the blade tip timing signal.

[0077] The identified pulsating pressure, tip timing power spectral density peak frequency, and rotational speed signal are uploaded online to the host computer and saved.

[0078] Step S40: Based on the peak frequency and rotational speed signal in the power spectral density of the pulsating pressure signal and the blade tip timing signal, perform fluid-structure interaction characteristic frequency reconstruction for asynchronous vibration of the blade to obtain circumferential acoustic mode propagation characteristic values.

[0079] When the blades vibrate asynchronously, there is a strong fluid-structure interaction relationship, as shown in formula (3), where the rotor rotation frequency is f. N The aerodynamic characteristic frequency is f R The structural modal characteristic frequency is f S The circumferential propagation modal eigenvalue is m:

[0080] f R ±f S =mf N (3)

[0081] Based on the above formula, peak combination feature reconstruction is performed on the peak frequencies of the power spectral density of the pulsating pressure signal and the blade tip timing signal acquired by the host computer, and the circumferential acoustic mode propagation characteristic values ​​are calculated:

[0082] m=(f R ±f S ) / fN .

[0083] Step 5: Identify asynchronous vibration of the blade based on the circumferential acoustic mode propagation characteristic value. During the test, if the circumferential acoustic mode propagation characteristic value m calculated in real time is an approximate integer value and is greater than a predetermined value, it is determined that there is a risk of asynchronous vibration of the blade.

[0084] It should be noted that, in this application, an approximate integer value generally refers to a value whose error from the nearest integer is no greater than 0.15. For example, if the calculated circumferential acoustic modal propagation characteristic value m is 2.02, and its error from the nearest integer value 2 is 0.02, then it meets the requirement of an approximate integer value. However, if the calculated circumferential acoustic modal propagation characteristic value m is 2.82, and its error from the nearest integer value 3 is 0.18, then it does not meet the requirement of an approximate integer value.

[0085] In some embodiments of this application, the predetermined value is not less than 5. For example, the predetermined value is 5.

[0086] During the experiment, the circumferential acoustic mode propagation characteristic value calculated online in real time will change with the rotational speed of the blade. When the real-time change of the calculated circumferential acoustic mode propagation characteristic value m is not an approximate integer value, it is judged that the blade vibration is good at this time.

[0087] During the experiment, the circumferential acoustic mode propagation characteristic value calculated in real time will change with the rotational speed of the blade. When the calculated circumferential acoustic mode propagation characteristic value m is an approximate integer value and greater than 5, it is judged that the blade is at risk of asynchronous vibration.

[0088] Given the risk of asynchronous vibration in the blade, if the circumferential acoustic mode propagation characteristic value calculated in real time does not change with the blade rotation speed, it is determined that the blade has already experienced asynchronous vibration.

[0089] This application provides a method for online identification of asynchronous vibration of rotor blades based on limited measuring points. It solves the problem of online identification of asynchronous blade vibration using finite and sparse measuring points when monitoring measuring points are limited in engineering environments such as the entire machine or core components. This application uses a non-contact measuring point system, consisting of a pulsating pressure measuring point formed by a single pulsating pressure sensor and an optical blade tip timing measuring point formed by two blade tip timing sensors, to identify the characteristics of asynchronous blade vibration online. This significantly reduces testing costs and enables rapid identification of asynchronous blade vibration under limited dynamic testing conditions, meeting the needs of core or entire machine blade vibration monitoring. Furthermore, this application uses a dynamic peak parameter scanning based on an autocorrelation function. By setting an adaptive threshold, it achieves the extraction and transmission of peak frequency features of pulsating pressure and blade vibration, enabling online analysis and coupling feature reconstruction of the peak frequency features of pulsating pressure and blade vibration. This results in online identification of asynchronous blade vibration characteristics, and the identification process is simple and reliable.

[0090] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A limited measurement point based online identification method for non-synchronous vibrations of a rotor blade, characterized in that, include: Acquire the pulsating pressure signal and the blade tip timing signal, and construct the autocorrelation function of the pulsating pressure signal and the blade tip timing signal; The autocorrelation function of the pulsating pressure signal and the tip timing signal is processed to obtain the power spectral density of the pulsating pressure signal and the tip timing signal. Based on the adaptive threshold, the power spectral density of the pulsating pressure signal and the blade tip timing signal is swept within a predetermined frequency range, and the peak signal above the adaptive threshold is extracted, thereby identifying the peak frequency of the power spectral density of the pulsating pressure signal and the blade tip timing signal. Based on the peak frequency and rotational speed signal in the power spectral density of the pulsating pressure signal and the blade tip timing signal, fluid-structure interaction characteristic frequency reconstruction for asynchronous blade vibration is carried out to obtain the circumferential acoustic mode characteristic propagation value. Identification of asynchronous vibration of blades based on circumferential acoustic mode propagation characteristic values.

2. The limited measurement point based online identification method of non-synchronous vibrations of a rotor blade as claimed in claim 1, characterized in that, A single pulsating pressure sensor is arranged on the axial front side of the rotor blade to obtain the pulsating pressure signal, and two blade tip timing sensors are arranged circumferentially at the blade tip to obtain the blade tip timing signal.

3. The limited measurement point based online identification method of non-synchronous vibrations of a rotor blade as claimed in claim 2, characterized in that, The autocorrelation function R of the pulsatile pressure signal is obtained by integrating the product of the pulsatile pressure signal at different times xx (τ): In the formula, x(t) is the pulsating pressure signal obtained by a single pulsating pressure sensor; x(t+τ) is the delayed pulsating pressure signal; τ is the delay time of the pulsating pressure signal; t represents time; T is the length of the pulsating pressure signal.

4. The limited measurement point based online identification method of non-synchronous vibrations of a rotor blade as claimed in claim 2, wherein, The autocorrelation function R of the blade tip timing signal is obtained by summing the time delays of two blade tip timing signals xx (epsilon): In the formula, x(n) is the timing signal of the first blade tip obtained by the timing sensor; x(n+ε) is the timing signal of the blade tip obtained by the second blade tip timing sensor; ε is the delay time between the two leaf tip timing signals; n represents the measurement point timing of the leaf tip timing signal; N is the length of the leaf tip timing signal.

5. The limited-sensor-based online identification method of non-synchronous vibrations of a rotor blade according to claim 1, characterized in that The frequency range of the power spectral density sweep of the blade tip timing signal is the frequency of the Mth vibration mode before the blade vibration. The scanning frequency range of the power spectral density of the pulsating pressure signal is determined according to the frequency range of the power spectral density sweep of the blade tip timing signal segment. The scanning frequency range of the power spectral density of the pulsating pressure signal is three times the scanning frequency range of the dynamic stress signal.

6. The limited measurement point based online identification method of non-synchronous vibrations of a rotor blade as claimed in claim 5, characterized in that, The adaptive threshold is the power spectral density mean plus N times the power spectral density standard deviation, where N ranges from 2 to 3.

7. The limited measurement point based online identification method of non-synchronous vibrations of a rotor blade as claimed in claim 6, characterized in that, The circumferential acoustic modal propagation characteristic value m satisfies: m = (f R ±f S ) / f N where f N is the rotor rotational frequency, f R is the aerodynamic feature frequency structure, f S is the modal feature frequency.

8. The online identification method for asynchronous vibration of rotor blades based on finite measurement points as described in claim 7, characterized in that, The process of identifying asynchronous vibrations of blades based on circumferential acoustic mode propagation characteristic values ​​is as follows: During the experiment, the circumferential acoustic mode propagation characteristic value calculated in real time will change with the rotational speed of the blade. When the circumferential acoustic mode propagation characteristic value m calculated in real time is not an approximate integer value, it is judged that the blade vibration is good at this time. When the real-time calculated circumferential acoustic mode propagation characteristic value m is an approximately integer value and greater than a predetermined value, it is determined that the blade is at risk of asynchronous vibration. When there is a risk of asynchronous vibration in the blade, and the real-time calculated circumferential acoustic mode propagation characteristic value does not change with the blade rotation speed, it is determined that the blade has already experienced asynchronous vibration.

9. The limited measurement point based online identification method of non-synchronous vibrations of a rotor blade as claimed in claim 8, characterized in that, The predetermined value is not less than 5.

Images