An acoustic identification method and system for compressor flutter faults

Abstract

This invention discloses an acoustic identification method and system for compressor flutter faults, relating to the field of engine condition monitoring and fault diagnosis technology. The method includes the following steps: optimizing the acoustic measurement point array layout and installing sound pressure sensors according to the array layout; acquiring signals through the sound pressure sensor array and preprocessing the signals; identifying non-integer-order abnormal frequencies and detecting frequency lock-in on the preprocessed signals; calculating the complex coherence between each channel signal and the reference channel and the phase of each channel signal at the locked frequency point; calculating the circumferential order k and wave direction of the flutter using a spatial matching algorithm; performing flutter identification and outputting the flutter identification result. This method can accurately identify the circumferential order and traveling wave direction online, effectively alleviating spatial aliasing problems. Furthermore, the algorithm is lightweight and easy to deploy in real-time on air, providing an efficient and reliable engineering solution for early warning and stability monitoring of compressor flutter.

Description

Technical Field

[0001] This invention relates to the field of engine condition monitoring and fault diagnosis technology, and in particular to an acoustic identification method and system for compressor flutter faults. Background Technology

[0002] As a key component of the engine, the health of engine blades is directly related to flight safety. Only by establishing an efficient flutter prediction and early warning mechanism can we ensure the long-term safe and reliable operation of the engine.

[0003] In existing technologies, compressor flutter identification often involves placing strain gauges or fast-response surface pressure sensors on the blades to obtain unsteady aerodynamic forces and modal responses, and directly determining whether the region has entered the negative damping region from the damping estimation; or using blade tip timing vibration measurement technology, which specifically uses sensors to sense the arrival time of the blade, and processes the time series through different blade tip timing processing algorithms to obtain blade vibration information.

[0004] These sensors are complex to arrange and costly, and are prone to interfering with the engine flow field. Furthermore, the strain gauge signals are susceptible to electromagnetic interference. The blade tip timing technology has insufficient resolution at high speeds. Both of them are unable to stably capture non-integer frequency characteristics in the early stages of flutter and cannot directly obtain circumferential vibration modal information. This limits the timeliness and accuracy of early warnings and is not conducive to engineering-based online monitoring and real-time alarms.

[0005] Therefore, an acoustic identification method and system for compressor flutter faults are provided to solve the above problems. Summary of the Invention

[0006] The purpose of this invention is to provide an acoustic identification method and system for compressor flutter faults. The method utilizes multi-point acoustic sensors installed through casing openings to collect signals, and combines non-integer frequency locking and high-coherence phase locking to achieve online automatic and accurate identification of flutter faults.

[0007] To achieve the above objectives, the present invention provides an acoustic identification method for compressor flutter faults, comprising the following steps:

[0008] S1: Optimize the layout of the acoustic measurement point array and install the sound pressure sensor according to the array layout;

[0009] S2: Acquire signals through a sound pressure sensor array and preprocess the signals;

[0010] S3: Perform non-integer-order abnormal frequency identification and frequency lock detection on the preprocessed signal;

[0011] S31: Map the spectrum obtained from the spectrum calculation from the frequency domain to the order domain. order domain Specifically set as follows:

[0012] ;

[0013] S32: During the runtime phase, calculate the current frame order spectrum to obtain abnormal orders. Abnormal order Specifically set as follows:

[0014] ;

[0015] in, Indicates absolute amplitude;

[0016] S33: Set the non-integer order determination threshold This yields candidate indices that are farthest from the integer order. The specific calculation method is set as follows:

[0017] ;

[0018] in, Represents the floor function;

[0019] S34: Introduce a continuous judgment strategy to perform non-integer frequency locking judgment on the maintenance counter for each candidate order, and obtain an alarm output. Alarm output Specifically set as follows:

[0020] ;

[0021] in, Indicates candidate frequency, Indicates the candidate amplitude. Indicates the candidate order;

[0022] S35: In the set of anomalous non-integer frequencies The nearest non-integer frequency combination is selected, and statistical calculations are performed on the non-integer frequency combination to obtain the variance of the abnormal frequencies at multiple measurement points. Variance of anomaly frequency at multiple measurement points Specifically set as follows:

[0023] ;

[0024] in, Indicates the number of acoustic measurement points. Indicates the peak frequency at the measuring point. This represents the average abnormal frequency of a combination of non-integer frequency orders.

[0025] S36: Calculate the variance of anomaly frequencies at multiple measurement points Average frequency of abnormalities normalized ratio According to the normalized ratio Determine the frequency locking condition and normalize the ratio. Specifically set as follows:

[0026] ;

[0027] S4: At the frequency lock point, calculate the complex coherence between each channel signal and the reference channel, and the phase of each channel signal;

[0028] S5: Calculate the circumferential order k and wave direction of flutter using a spatial matching algorithm;

[0029] S6: Perform flutter identification and output the flutter identification results.

[0030] Preferably, step S1 specifically includes the following steps:

[0031] S11: Determine the number K of sound pressure sensors based on the modal number;

[0032] S12: Based on the modal phase relationship, reconstruct the sound field structure of N uniformly distributed sound pressure microphones through the frequency domain signals of K non-uniformly distributed sound pressure sensors, and confirm the installation position of K sound pressure sensors.

[0033] S13: Based on the casing opening at the installation location opposite the rotor, install K sound pressure sensors flush on the rotor interface.

[0034] Preferably, step S2 specifically includes the following steps:

[0035] S21: Acoustic signals and speed pulse signals during compressor operation are collected using K acoustic pressure sensors to obtain a discrete time-domain sequence. ];

[0036] S22: Perform frame segmentation calculation on the acoustic signal to obtain the number of frames. Specifically, it is set as follows:

[0037] ;

[0038] in, Indicates the total number of signal sampling points. Indicates frame length, Indicates frame shift;

[0039] S23: Perform window function weighting on the acoustic signal and calculate the... Windowing signal of a frame , No. Windowing signal of a frame Specifically set as follows:

[0040] ;

[0041] ;

[0042] in, Indicates the first Frame signal segment, Indicates the index of the intra-frame sampling point. Indicates the window function;

[0043] S24: Perform spectrum calculation on the acoustic signal, and perform Fourier transform frame by frame to obtain the... Frequency domain components of frame signal Frequency domain components Specifically set as follows:

[0044] ;

[0045] in, Indicates frequency index, Represents a complex exponential function. Represents the imaginary unit. Indicates the vibration pitch diameter. Indicates the sampling frequency;

[0046] S25: Calculate the rotational frequency from the rotational speed pulse signal. Frequency conversion Specifically set as follows:

[0047] ;

[0048] in, This indicates the time interval between two rising edge pulses.

[0049] Preferably, step S4 specifically includes the following steps:

[0050] S41: Computer box Complex spectrum of each acoustic measurement point Complex spectrum Specifically set as follows:

[0051] ;

[0052] in, Indicates the frequency of the frequency lock. The common narrowband component representing flutter, Indicates the first The amplitude at each acoustic measurement point Indicates the first The phase of each acoustic measurement point Indicates local noise;

[0053] S42: At the frequency lock frequency At this point, calculate the first... Cross spectrum between channel and reference channel , No. Channel power spectrum and reference channel power spectrum mutual spectrum , No. Channel power spectrum and reference channel power spectrum Specifically set as follows:

[0054] ;

[0055] ;

[0056] ;

[0057] in, Indicates the first Channel at frequency lock The complex spectrum at that location, Represents the complex conjugate of the reference channel spectrum. Indicates the reference channel at the frequency lock. The complex spectrum at that location;

[0058] S43: Calculate the... Complex coherence between channel and reference channel Complex coherence Specifically set as follows:

[0059] ;

[0060] in, Represents the set of complex numbers;

[0061] S44: Calculate complex coherence amplitude and phase Amplitude and phase Specifically set as follows:

[0062] ;

[0063] ;

[0064] in, Indicates the first The phase angle of the channel, Indicates the phase angle of the reference channel;

[0065] If at least n channels satisfy the amplitude Execute step S5; otherwise, the vibration pitch diameter will not be output for this frame. .

[0066] Preferably, step S5 specifically includes the following steps:

[0067] S51: If the frequency is locked Given a traveling wave of a single circumferential order, calculate the position of the circumferential angle. steady phase at steady-state phase Specifically set as follows:

[0068] ;

[0069] in, Indicates the input angle variable;

[0070] S52: Calculate the candidate order Spatial matching score Matching score Specifically set as follows:

[0071] ;

[0072] ;

[0073] in, The unit phase vector representing the phase of the measuring point. Indicates the angular position. Represents frequency components;

[0074] S53: Spatial matching score The maximum value is set to If the ratio of the highest score to the second-highest score is greater than the separation threshold... If it is true, it is considered valid; otherwise, it is considered invalid.

[0075] S54: Wave direction is determined by the slope of the phase-angle linear fitting. If the slope is greater than 0, the wave direction is forward; if the slope is less than 0, the wave direction is backward. This yields the directed order. .

[0076] An acoustic identification system for an acoustic identification method of compressor flutter fault includes a data acquisition unit, a preprocessing unit, a non-integer-order abnormal frequency identification unit, a cross-channel consensus frequency locking unit, a coherence and phase extraction unit, a circumferential mode identification unit, and a visualization and data management unit connected in sequence.

[0077] Preferably, the acquisition unit is used to acquire acoustic and speed pulse signals during compressor operation and output a time-domain signal sequence. The preprocessing unit is used to perform frame segmentation and windowing processing on the acoustic signal, calculate the spectrum of each frame, and calculate the rotational frequency in real time for the speed pulse signal in frames. The non-integer order abnormal frequency identification unit is used to generate frequency-locking candidates only for non-integer order narrowband targets on the order axis. The cross-channel consensus frequency-locking unit is used to perform frequency-locking pass determination frame by frame and record the average of the two frequencies as the frequency-locking frequency. The coherence and phase extraction unit is used in The complex coherence of each channel relative to the reference channel is calculated, and the amplitude and phase are extracted. The circumferential mode recognition unit is used to identify the circumferential order and wave direction based on the angular position and phase of the measurement point. The visualization and data management unit is used to display the time-frequency locking curve and time-frequency locking curve in real time. curve.

[0078] Therefore, the present invention employs the above-mentioned acoustic identification method and acoustic identification system for compressor flutter faults, which has the following beneficial effects:

[0079] (1) This scheme uses an acoustic sensor array installed outside the casing for monitoring, which avoids complex operations such as arranging strain gauges on the blade surface. The sensor arrangement is simple and has little interference with the flow field.

[0080] (2) This scheme uses non-integer frequency locking and high coherence phase locking technology to stably capture the weak asynchronous vibration characteristics in the early stage of flutter, which significantly improves the accuracy and reliability of early identification.

[0081] (3) This scheme effectively eliminates the interference of engine speed fluctuations and strong background noise by mapping the spectrum to the order domain and avoiding integer order frequencies. At the same time, it adopts a cross-channel consensus frequency locking mechanism and a continuous judgment strategy to ensure the stability of flutter frequency identification and greatly reduce false alarms caused by occasional interference.

[0082] (4) This scheme combines spatial matching algorithm and complex coherence analysis, which can identify key modal information such as the circumferential order and traveling wave direction of flutter online. It not only realizes qualitative alarm of flutter phenomenon, but also provides quantitative basis for in-depth fault diagnosis and stability analysis, and has high early warning value.

[0083] (5) The entire algorithm of this solution does not require complex interpolation calculations, has low computational load, and is easy to implement in real-time processing and online monitoring in airborne embedded systems.

[0084] The method of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0085] Figure 1 This is a flowchart of an acoustic identification method for compressor flutter faults according to the present invention;

[0086] Figure 2 This is a schematic diagram of frequency locking and flutter identification curves of the present invention, wherein (a) is a schematic diagram of frequency locking, and (b) is the actual flutter alarm point;

[0087] Figure 3 This is a structural diagram of an acoustic identification system for compressor flutter faults according to the present invention.

[0088] The system comprises: 1. Acquisition unit; 2. Preprocessing unit; 3. Non-integer-order abnormal frequency identification unit; 4. Cross-channel consensus frequency locking unit; 5. Coherence and phase extraction unit; 6. Circumferential mode identification unit; and 7. Visualization and data management unit. Detailed Implementation

[0089] The method of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0090] Unless otherwise defined, the methodological or scientific terms used in this invention shall have the ordinary meaning as understood by one of ordinary skill in the art to which this invention pertains.

[0091] The terms "comprising" or "including" as used in this invention mean that the element preceding the term encompasses the element listed after the term, and do not exclude the possibility of encompassing other elements as well. Terms such as "inner," "outer," "upper," and "lower" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. When the absolute position of the described object changes, the relative positional relationship may also change accordingly. In this invention, unless otherwise explicitly specified and limited, the term "attached" and similar terms should be interpreted broadly. For example, it can refer to a fixed connection, a detachable connection, or an integral part; it can refer to a direct connection or an indirect connection through an intermediate medium; it can refer to the internal communication of two elements or the interaction relationship between two elements. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0092] Example 1

[0093] This embodiment verifies the effectiveness of the proposed solution using test data of compressor flutter from engine model A.

[0094] like Figures 1 to 2 As shown, the present invention provides an acoustic identification method for compressor flutter faults, comprising the following steps:

[0095] S1: Optimize the layout of the acoustic measurement point array and install the sound pressure sensor according to the array layout;

[0096] Step S1 specifically includes the following steps:

[0097] S11: Determine the number K of sound pressure sensors based on the modal number;

[0098] S12: Based on the modal phase relationship, reconstruct the sound field structure of N uniformly distributed sound pressure microphones through the frequency domain signals of K non-uniformly distributed sound pressure sensors, and confirm the installation position of K sound pressure sensors.

[0099] S13: Based on the casing opening at the installation location opposite the rotor, install K sound pressure sensors flush on the rotor interface.

[0100] S2: Signal acquisition is performed using a sound pressure sensor array. The acquisition frequency response is set to 3.15Hz-20kHz, and the acquisition sensitivity is set to 50mV / Pa. To ensure sufficient sample data for subsequent short-time Fourier transform, the sampling rate is 25.6kHz. The signal is then preprocessed.

[0101] Step S2 specifically includes the following steps:

[0102] S21: Acoustic signals and speed pulse signals during compressor operation are collected using K acoustic pressure sensors to obtain a discrete time-domain sequence. ];

[0103] S22: Perform frame segmentation calculation on the acoustic signal to obtain the number of frames. Specifically, it is set as follows:

[0104] ;

[0105] in, Indicates the total number of signal sampling points. Indicates frame length, Indicates frame shift;

[0106] S23: Perform window function weighting on the acoustic signal. The window function can be any one of the Hamming window, Hanning window, rectangular window, Blackman window, or an equivalent window. Calculate the... Windowing signal of a frame , No. Windowing signal of a frame Specifically set as follows:

[0107] ;

[0108] ;

[0109] in, Indicates the first Frame signal segment, Indicates the index of the intra-frame sampling point. Indicates the window function;

[0110] S24: Perform spectrum calculation on the acoustic signal, and perform Fourier transform frame by frame. The number of Fourier transform points should not be less than 8192 points to obtain the [number of points]. Frequency domain components of frame signal Frequency domain components Specifically set as follows:

[0111] ;

[0112] in, Indicates frequency index, Represents a complex exponential function. Represents the imaginary unit. Indicates the vibration pitch diameter. Indicates the sampling frequency;

[0113] S25: Calculate the rotational frequency from the rotational speed pulse signal. Frequency conversion Specifically set as follows:

[0114] ;

[0115] in, This indicates the time interval between two rising edge pulses.

[0116] S3: Perform non-integer-order abnormal frequency identification and frequency lock detection on the preprocessed signal;

[0117] Step S3 specifically includes the following steps:

[0118] S31: To eliminate the influence of engine speed fluctuations and clean frequency on the recognition results, the spectrum obtained from spectrum calculation is mapped from the frequency domain to the order domain. order domain Specifically set as follows:

[0119] ;

[0120] S32: During the runtime phase, calculate the current frame order spectrum to obtain abnormal orders. Abnormal order Specifically set as follows:

[0121] ;

[0122] in, Indicates absolute amplitude;

[0123] S33: Since the actual frequency of asynchronous resonance in blades is usually located at a non-integer order position, a non-integer order judgment threshold is set. This yields candidate indices that are farthest from the integer order. The specific calculation method is set as follows:

[0124] ;

[0125] in, Represents the floor function;

[0126] S34: To further reduce false alarms caused by sporadic interference, a continuous judgment strategy is introduced. For each candidate order, a non-integer order frequency locking judgment is performed on the maintenance counter. When two consecutive frames of a certain order meet the above conditions, it is determined to be a stable non-integer order frequency locking, triggering an alarm output. Alarm output Specifically set as follows:

[0127] ;

[0128] in, Indicates candidate frequency, Indicates the candidate amplitude. Indicates the candidate order;

[0129] S35: Set of Abnormal Non-Integer Frequency The order range is [3,15], the integer order avoidance bandwidth is ±0.05-±0.10EO, the continuous frame threshold is 2 frames, and the amplitude threshold is not less than 80-110dB.

[0130] In the set of anomalous non-integer frequencies The nearest non-integer frequency combination is selected, and statistical calculations are performed on the non-integer frequency combination to quantify frequency stability in real-time monitoring and obtain the variance of abnormal frequencies at multiple measurement points. Variance of anomaly frequency at multiple measurement points Specifically set as follows:

[0131] ;

[0132] in, Indicates the number of acoustic measurement points. Indicates the peak frequency at the measuring point. This represents the average abnormal frequency of a combination of non-integer frequency orders.

[0133] S36: Calculate the variance of anomaly frequencies at multiple measurement points Average frequency of abnormalities normalized ratio Normalized ratio Less than 0.0001, based on the normalized ratio Determine the frequency locking condition and normalize the ratio. Specifically set as follows:

[0134] .

[0135] S4: At the frequency lock point, calculate the complex coherence between each channel signal and the reference channel, and the phase of each channel signal;

[0136] Step S4 specifically includes the following steps:

[0137] S41: Computer box Complex spectrum of each acoustic measurement point Complex spectrum Specifically set as follows:

[0138] ;

[0139] in, Indicates the frequency of the frequency lock. The common narrowband component representing flutter, Indicates the first The amplitude at each acoustic measurement point Indicates the first The phase of each acoustic measurement point Indicates local noise;

[0140] S42: At the frequency lock frequency At this point, calculate the first... Cross spectrum between channel and reference channel , No. Channel power spectrum and reference channel power spectrum mutual spectrum , No. Channel power spectrum and reference channel power spectrum Specifically set as follows:

[0141] ;

[0142] ;

[0143] ;

[0144] in, Indicates the first Channel at frequency lock The complex spectrum at that location, Represents the complex conjugate of the reference channel spectrum. Indicates the reference channel at the frequency lock. The complex spectrum at that location;

[0145] S43: Calculate the... Complex coherence between channel and reference channel Complex coherence Specifically set as follows:

[0146] ;

[0147] in, Represents the set of complex numbers;

[0148] S44: Calculate complex coherence amplitude and phase Amplitude The closer to 1, the more coherent; that is, the first... Channel and reference channel at frequency lock The higher the degree of linear correlation.

[0149] Amplitude and phase Specifically set as follows:

[0150] ;

[0151] ;

[0152] in, Indicates the first The phase angle of the channel, Indicates the phase angle of the reference channel;

[0153] If at least n channels satisfy the amplitude In this embodiment If the value is set to the range of 0.7-0.9, proceed to step S5; otherwise, the vibration pitch diameter will not be output for this frame. In this embodiment, the value of 𝑀 is not less than 3.

[0154] S5: Calculate the circumferential order k and wave direction of flutter using a spatial matching algorithm;

[0155] Step S5 specifically includes the following steps:

[0156] S51: If the frequency is locked The location is a single circumferential traveling wave. In this embodiment, the circumferential angle position is calculated using the latter traveling wave as an example. steady phase at steady-state phase Specifically set as follows:

[0157] ;

[0158] in, Indicates the input angle variable;

[0159] S52: Calculate the candidate order Spatial matching score Matching score Specifically set as follows:

[0160] ;

[0161] ;

[0162] in, The unit phase vector representing the phase of the measuring point. Indicates the angular position. Represents frequency components;

[0163] S53: Spatial matching score The maximum value is set to If the ratio of the highest score to the second-highest score is greater than the separation threshold... ,Right now If the resolution is within the acceptable range, it is considered valid; otherwise, it is considered invalid. In this embodiment, the resolution threshold is... Take 1.08;

[0164] S54: Wave direction is determined by the slope of the phase-angle linear fitting. If the slope is greater than 0, the wave direction is forward; if the slope is less than 0, the wave direction is backward. This yields the directed order. .

[0165] S6: Perform flutter identification and output the flutter identification results.

[0166] like Figure 3 As shown, an acoustic identification system for an acoustic identification method of compressor flutter fault includes a data acquisition unit 1, a preprocessing unit 2, a non-integer-order abnormal frequency identification unit 3, a cross-channel consensus frequency locking unit 4, a coherence and phase extraction unit 5, a circumferential mode identification unit 6, and a visualization and data management unit 7 connected in sequence. Each functional unit can be implemented by an independent hardware circuit, or by a processor running a software module, or by a combination of hardware and software.

[0167] Acquisition unit 1 is used to acquire acoustic and speed pulse signals during compressor operation and output time-domain signal sequences.

[0168] The preprocessing unit 2 is used to perform framing and windowing processing on the acoustic signal, calculate the spectrum of each frame, and calculate the rotational frequency in real time for the rotational speed pulse signal in frames.

[0169] The non-integer-order abnormal frequency identification unit 3 is used to generate frequency-locking candidates only for non-integer-order narrowband targets on the order axis, within a preset order interval, for example... Within the range, avoid integer-order neighborhoods and select peak values ​​that exceed the amplitude threshold and appear across consecutive frames.

[0170] Cross-channel consensus frequency locking unit 4 is used to perform frequency locking pass determination frame by frame. For example, within the same frame, when the four abnormal frequencies simultaneously meet the condition... When the frame passes the frequency locking test, the average of the two frequencies is recorded as the frequency locking frequency. .

[0171] The coherence and phase extraction unit 5 is used in The complex coherence of each channel relative to the reference channel is calculated, and the amplitude and phase are extracted, only if at least 𝑀 channels satisfy . If the phase vector and coherence vector of the frame are output, the next unit is entered; otherwise, only the frequency lock is recorded and the mode is not output.

[0172] The circumferential modal recognition unit 6 is used to identify the circumferential order and wave direction based on the angular position and phase of the measurement point. If the ratio of the maximum score to the second largest score is greater than φ, it is considered valid. Then, the wave direction is determined by the sign of the slope of the phase-angle linear fitting: a value greater than 0 indicates a forward wave, and a value less than 0 indicates a backward wave, thus obtaining the directed order. .

[0173] Visualization and Data Management Unit 7 is used to display time-locked frequency curves and time- in real time. curve.

[0174] Example 2

[0175] This embodiment provides a computer-readable storage medium on which a program is stored and runs on a processor.

[0176] The remaining specific implementation methods are the same as in Example 1.

[0177] Therefore, the present invention adopts the above-mentioned acoustic identification method and acoustic identification system for compressor flutter faults. Through non-integer order frequency locking and cross-channel consensus mechanism, it effectively avoids integer order interference and achieves stable online identification of circumferential order and wave direction, providing key modal information for flutter early warning.

[0178] Finally, it should be noted that the above embodiments are only used to illustrate the method of the present invention and not to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the method of the present invention, and these modifications or equivalent substitutions should not cause the modified method to deviate from the spirit and scope of the method of the present invention.

Claims

1. A method of acoustic identification of a compressor surge fault, characterized in that, Includes the following steps: S1: Optimize the layout of the acoustic measurement point array and install the sound pressure sensor according to the array layout; S2: Acquire signals through a sound pressure sensor array and preprocess the signals; S3: Perform non-integer-order abnormal frequency identification and frequency lock detection on the preprocessed signal; S31: mapping the spectrum calculated from the frequency domain to the order domain , order domain Specifically, it is configured as: ； S32: In the running phase, the current frame order spectrum is calculated to obtain the abnormal order , the abnormal order Specifically set as: ； wherein denotes the absolute value; S33: setting a non-integer order determination threshold obtaining a candidate index far from an integer order The specific calculation method is set as: ； in, Represents the floor function; S34: Introduce a continuous judgment strategy to perform non-integer frequency locking judgment on the maintenance counter for each candidate order, and obtain an alarm output. Alarm output Specifically set as follows: ； in, Indicates candidate frequency, Indicates the candidate amplitude. Indicates the candidate order; S35: In the set of anomalous non-integer frequencies The nearest non-integer frequency combination is selected, and statistical calculations are performed on the non-integer frequency combination to obtain the variance of abnormal frequencies at multiple measurement points. Variance of anomaly frequency at multiple measurement points Specifically set as follows: ； in, Indicates the number of acoustic measurement points. Indicates the peak frequency at the measuring point. This represents the average abnormal frequency of a combination of non-integer frequency orders. S36: Calculate the variance of anomaly frequencies at multiple measurement points Average frequency of abnormalities normalized ratio According to the normalized ratio Determine the frequency locking condition and normalize the ratio. Specifically set as follows: ； S4: At the frequency lock point, calculate the complex coherence between each channel signal and the reference channel, and the phase of each channel signal; S5: Calculate the circumferential order k and wave direction of flutter using a spatial matching algorithm; S6: Perform flutter identification and output the flutter identification results.

2. The acoustic identification method for compressor flutter fault according to claim 1, characterized in that, Step S1 specifically includes the following steps: S11: Determine the number K of sound pressure sensors based on the modal number; S12: Based on the modal phase relationship, reconstruct the sound field structure of N uniformly distributed sound pressure microphones through the frequency domain signals of K non-uniformly distributed sound pressure sensors, and confirm the installation position of K sound pressure sensors. S13: Based on the casing opening at the installation location opposite the rotor, install K sound pressure sensors flush on the rotor interface.

3. The acoustic identification method for compressor flutter faults according to claim 2, characterized in that, Step S2 specifically includes the following steps: S21: Acoustic signals and speed pulse signals during compressor operation are collected using K acoustic pressure sensors to obtain a discrete time-domain sequence. ]; S22: Perform frame segmentation calculation on the acoustic signal to obtain the number of frames. Specifically, it is set as follows: ； in, Indicates the total number of signal sampling points. Indicates frame length, Indicates frame shift; S23: Perform window function weighting on the acoustic signal and calculate the... Windowing signal of a frame , No. Windowing signal of a frame Specifically set as follows: ； ； in, Indicates the first Frame signal segment, Indicates the index of the intra-frame sampling point. Indicates the window function; S24: Perform spectrum calculation on the acoustic signal, and perform Fourier transform frame by frame to obtain the... Frequency domain components of frame signal Frequency domain components Specifically set as follows: ； in, Indicates frequency index, Represents a complex exponential function. Represents the imaginary unit. Indicates the vibration pitch diameter. Indicates the sampling frequency; S25: Calculate the rotational frequency from the rotational speed pulse signal. Frequency conversion Specifically set as follows: ； in, This indicates the time interval between two rising edge pulses.

4. The acoustic identification method for compressor flutter fault according to claim 3, characterized in that, Step S4 specifically includes the following steps: S41: Computer box Complex spectrum of each acoustic measurement point Complex spectrum Specifically set as follows: ； in, Indicates the frequency of the frequency lock. The common narrowband component representing flutter, Indicates the first The amplitude at each acoustic measurement point Indicates the first The phase of each acoustic measurement point Indicates local noise; S42: At the frequency lock frequency At this point, calculate the first... Cross spectrum between channel and reference channel , No. Channel power spectrum and reference channel power spectrum mutual spectrum , No. Channel power spectrum and reference channel power spectrum Specifically set as follows: ； ； ； in, Indicates the first Channel at frequency lock The complex spectrum at that location, Represents the complex conjugate of the reference channel spectrum. Indicates the reference channel at the frequency lock. The complex spectrum at that location; S43: Calculate the... Complex coherence between channel and reference channel Complex coherence Specifically set as follows: ； in, Represents the set of complex numbers; S44: Calculate complex coherence amplitude and phase Amplitude and phase Specifically set as follows: ； ； in, Indicates the first The phase angle of the channel, Indicates the phase angle of the reference channel; If at least n channels satisfy the amplitude Execute step S5; otherwise, the vibration pitch diameter will not be output for this frame. .

5. The acoustic identification method for compressor flutter fault according to claim 4, characterized in that, Step S5 specifically includes the following steps: S51: If the frequency is locked Given a traveling wave of a single circumferential order, calculate the position of the circumferential angle. steady phase at steady-state phase Specifically set as follows: ； in, Indicates the input angle variable; S52: Calculate the candidate order Spatial matching score Matching score Specifically set as follows: ； ； in, The unit phase vector representing the phase of the measuring point. Indicates the angular position. Represents frequency components; S53: Spatial matching score The maximum value is set to If the ratio of the highest score to the second-highest score is greater than the separation threshold... If it is true, it is considered valid; otherwise, it is considered invalid. S54: Wave direction is determined by the slope of the phase-angle linear fitting. If the slope is greater than 0, the wave direction is forward; if the slope is less than 0, the wave direction is backward. This yields the directed order. .

6. An acoustic identification system for the acoustic identification method of compressor flutter fault according to any one of claims 1-5, characterized in that, It includes a data acquisition unit, a preprocessing unit, a non-integer-order abnormal frequency identification unit, a cross-channel consensus frequency locking unit, a coherence and phase extraction unit, a circumferential mode identification unit, and a visualization and data management unit connected in sequence.

7. The acoustic identification system of the acoustic identification method for compressor flutter fault according to claim 6, characterized in that, The acquisition unit is used to acquire acoustic and speed pulse signals during compressor operation and output a time-domain signal sequence. The preprocessing unit is used to perform frame segmentation and windowing processing on the acoustic signals, calculate the spectrum of each frame, and calculate the rotational frequency in real time for the speed pulse signals. The non-integer order abnormal frequency identification unit is used to generate frequency-locking candidates only for non-integer order narrowband targets on the order axis. The cross-channel consensus frequency-locking unit is used to perform frequency-locking pass determination frame by frame and record the average of the two frequencies as the frequency-locking frequency. The coherence and phase extraction unit is used in The complex coherence of each channel relative to the reference channel is calculated, and the amplitude and phase are extracted. The circumferential mode recognition unit is used to identify the circumferential order and wave direction based on the angular position and phase of the measurement point. The visualization and data management unit is used to display the time-frequency locking curve and time-frequency locking curve in real time. curve.

Images