Voiceprint continuous monitoring method, system, medium and equipment for near stall state of gas compressor

Abstract

The invention discloses a gas compressor near stall state voiceprint continuous monitoring method and system, a medium and equipment, and the method comprises the steps: randomly selecting a preset installation number of microphone measuring points at an inlet to form a microphone array, and carrying out the synchronous collection of multi-channel sound pressure time domain signals; the method comprises the following steps: adding a rectangular window to a multi-channel sound pressure time domain signal, constructing an acoustic time domain signal segment, carrying out fast Fourier transform on the multi-channel acoustic time domain signal segment in a current time domain window to form a plurality of sound pressure frequency spectrums, judging whether a plurality of asynchronous single-tone characteristic frequencies which are non-integer multiples of current rotation frequency exist in the sound pressure frequency spectrums, and if yes, judging whether the sound pressure frequency spectrums exist in the sound pressure frequency spectrums; the method comprises the following steps: constructing a multi-channel microphone continuous frequency domain observation matrix of an asynchronous single-tone characteristic frequency, constructing an orthogonal Fourier transfer matrix based on a microphone sensor installation azimuth angle and a modal highest order, and establishing a compressor near-stall characteristic modal reconstruction model under a non-uniform few-measurement-point condition; and the amplitude is solved based on a block orthogonal matching pursuit algorithm, and compressor near-stall working condition monitoring based on voiceprint features is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of compressor state monitoring, and in particular to a method, system, medium and equipment for continuous monitoring of the soundprint of a compressor in a near-stall state. Background Art

[0002] As aircraft requirements for flight altitude, speed, and maneuverability increase, the compressor's operating environment becomes increasingly harsh. When an engine enters an aerodynamically unstable operating state, it can degrade performance, increase vibration stress on the compressor's rotor blades, increase thermal loads and stresses on the turbine, reduce the stable operating range of the combustion chamber, and even damage the engine's structural integrity, posing a serious threat to flight safety.

[0003] By monitoring the unsteady pressure pulsations of a compressor near stall conditions, the precursor anomalies exhibited during near stall conditions can be detected and diagnosed, which is of great significance for the design of available stability margins and aerodynamic stability control of aircraft engine compression systems. Currently, compressor status monitoring typically involves installing strain gauges, pressure probes, and other sensors on rotating blades. This is costly, and the invasive measurement method can interfere with the compressor flow field. Acoustic sensors offer high sensitivity, short transmission paths, and non-invasive measurement. Compressors typically generate strong single-tone noise during near stall conditions, and acoustic signals are a key form of fault response. However, little research has been conducted on the acoustic signature of compressors near stall conditions, and a near-stall monitoring method for compressors based on acoustic signatures has yet to be established.

[0004] The above information disclosed in this Background section is only for enhancement of understanding of the background of the invention and therefore it may contain information that does not form the prior art that is already known to a person of ordinary skill in the art. Summary of the Invention

[0005] The present invention provides a method, system, medium, and equipment for continuous monitoring of the near-stall soundprint of a compressor. A small number of non-uniform annular microphone arrays are arranged at the compressor inlet. The frequency spectrum analysis module of the acoustic array test signal determines the asynchronous characteristic single-tone frequency of the compressor near-stall operating condition. A multi-channel microphone continuous frequency domain measurement matrix of the asynchronous single-tone characteristic frequency is constructed. A reconstruction model of the compressor near-stall characteristic mode with a small number of non-uniform measurement points is established. The amplitude of the compressor near-stall characteristic acoustic mode is solved based on the block orthogonal matching pursuit algorithm, thereby realizing monitoring of the compressor near-stall operating condition based on soundprint features. This method realizes the real-time monitoring of the near-stall characteristic mode amplitude based on a small number of acoustic measurement points for the first time, providing an effective tool for engine control in near-stall operating conditions. At the same time, this method uses an acoustic monitoring method with a more sensitive response, a shorter transmission path, and is non-invasive, solving the problem that traditional static pressure sensors have difficulty in achieving dynamic sensitive feedback when monitoring near-stall operating conditions.

[0006] A method for continuously monitoring the soundprint of a compressor in a near-stall state includes:

[0007] In the first step, according to the number of rotor blades of the compressor Number of stator blades Calculate the highest circumferential modal order of the compressor Number of microphones ,A predetermined number of microphone measurement points are randomly selected at the entrance to form a microphone array, and multi-channel sound pressure time domain signals are collected synchronously;

[0008] In the second step, a rectangular window is added to the multi-channel acoustic pressure time domain signal to construct an acoustic time domain signal segment, and a fast Fourier transform is performed on the multi-channel acoustic time domain signal segment within the current time domain window to form multiple sound pressure spectra. It is determined whether there are multiple non-synchronous single-tone characteristic frequencies that are non-integer multiples of the current rotation frequency in the sound pressure spectrum. If not, it indicates that the compressor is in normal working condition and the second step is repeated. If so, further monitoring of the near-stall condition is carried out.

[0009] In the third step, a multi-channel microphone continuous frequency domain measurement matrix of asynchronous single-tone characteristic frequencies is constructed, an orthogonal Fourier transfer matrix is constructed based on the microphone sensor installation azimuth angle and the highest modal order, and a near-stall characteristic mode reconstruction model of the compressor is established with a non-uniform and small number of measurement points.

[0010] In the fourth step, the amplitude of the compressor near-stall characteristic mode reconstruction model is solved based on the block orthogonal matching pursuit algorithm to realize the compressor near-stall condition monitoring based on the soundprint feature.

[0011] In the method for continuous monitoring of compressor near-stall state soundprint, the first step includes:

[0012] Step S101: Calculate the highest circumferential modal order, where the modal order of the circumferential acoustic mode is , represents the pressure pulsation order caused by the unsteady aerodynamic force caused by the fan's rotation-stationary interference. Represents a non-negative integer, determining the highest circumferential modal order , determine the measurable modal range as ;

[0013] Step S102: Setting the number of microphones in the virtual ring and installation position, according to Nyquist-Shannon sampling theory, the number of microphones With the highest modal order The relationship is , the virtual microphone array is evenly arranged in the circumferential direction, and the spacing between each virtual grid point is ;

[0014] Step S103: randomly install K microphones on the casing wall in the circumferential direction, where K is approximately the number of microphones under Nyquist sampling. 50% of the installation angle .

[0015] In the method for continuous monitoring of compressor near-stall soundprint, the second step (S2) includes the following steps:

[0016] Step S201: The sampling frequency determines the window length of the acoustic time domain signal segment, and a rectangular window is added to the multi-channel sound pressure time domain signal. The window length is and , is any integer, is the sampling frequency of the microphone sensor; and every Add windows to the time domain signal of the sound pressure of the multi-channel signal, and the overlapping length of each window ;

[0017] Step S202: the current acoustic time domain signal segment Perform fast Fourier transform on the K sound pressure time domain signals: ,in, Represents the current acoustic time domain signal fragment matrix of the sound pressure time domain signal measured by the non-uniform microphone array. The dimension of the current time domain signal fragment matrix is , represents the discrete Fourier transform, For multi-channel arrays The frequency domain matrix of the acoustic time domain signal fragment has the dimension , , which is expressed as follows:

[0018]

[0019] in ,

[0020] Step S203: Based on the obtained frequency domain matrix , draw the sound pressure spectrum measured by sensors at different positions, and observe whether there are multiple integer multiples of the frequency The non-synchronous single-tone characteristic frequency outside : If not, repeat the second step (S2); if present, further carry out monitoring of the near-stall condition.

[0021] In the method for continuous monitoring of compressor near-stall state soundprint, the third step includes:

[0022] Step S301: Establishing asynchronous single-tone characteristic frequency Multi-channel microphone continuous frequency domain measurement matrix :

[0023]

[0024] in, Represents an acoustic time domain signal segment The non-synchronous single-tone characteristic frequency of the kth measurement point of the non-uniform array is Amplitude,

[0025] Step S302: Construct an orthogonal Fourier transfer matrix ,

[0026]

[0027] in ,

[0028] Step S303: Define the acoustic mode matrix , establish the compressor near-stall characteristic mode reconstruction model,

[0029]

[0030] in, The matrix row vector, is the number of dominant modes of interest, is the structural sparse norm and has:

[0031] ,

[0032] It is an indicator function, that is, it returns 1 when the condition is true.

[0033] In the method for continuous monitoring of compressor near-stall state soundprint, the fourth step includes:

[0034] Step S401: Set the residual matrix , index number set , structural sparsity is the number of dominant modes of interest, the number of iterations ;

[0035] Step S402: Update the number of iterations ;

[0036] Step S403: Calculate support set index ;

[0037] Step S404: Update the index sequence number set ;

[0038] Step S405: Calculate the acoustic modal amplitude vector corresponding to the support set

[0039] ;

[0040] Step S406: Update the residual matrix ;

[0041] Step S407: Check the number of iterations Whether the structural sparsity is achieved , the amplitude monitoring result of the compressor near-stall characteristic mode reconstruction model is output, otherwise, the process returns to step S402.

[0042] In the method for continuous monitoring of soundprints of a compressor in a near-stall state, the compressor is a compressor of an aircraft engine.

[0043] In the method for continuous monitoring of soundprints of a compressor in a near-stall state, the compressor includes a fan structure.

[0044] A fan blade synchronous vibration identification system for implementing the method includes:

[0045] The acoustic field measurement module is used to measure the number of rotor blades of the compressor Number of stator blades Calculate the highest circumferential modal order of the compressor Number of microphones ,A predetermined number of microphone measurement points are randomly selected at the entrance to form a microphone array, and multi-channel sound pressure time domain signals are collected synchronously;

[0046] A spectrum analysis module is used to add a rectangular window to the multi-channel sound pressure time domain signal to construct an acoustic time domain signal segment, and perform a fast Fourier transform on the multi-channel acoustic time domain signal segment within the current time domain window to form multiple sound pressure spectra, and determine whether there are multiple asynchronous single-tone characteristic frequencies in the sound pressure spectrum that are non-integer multiples of the current frequency conversion;

[0047] A construction module is used to construct a multi-channel microphone continuous frequency domain measurement matrix of asynchronous single-tone characteristic frequencies, construct an orthogonal Fourier transfer matrix based on the microphone sensor installation azimuth angle and the highest modal order, and establish a near-stall characteristic mode reconstruction model for a compressor with a non-uniform and small number of measurement points;

[0048] The monitoring module solves the amplitude of the compressor near-stall characteristic mode reconstruction model based on the block orthogonal matching pursuit algorithm, realizing the near-stall condition monitoring of the compressor based on the soundprint characteristics.

[0049] A computer storage medium includes computer instructions, which, when executed on a computer, cause the computer to execute the method described above.

[0050] An electronic device, comprising:

[0051] A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein:

[0052] When the processor executes the program, the method described is implemented.

[0053] Compared with existing technologies, the present invention has the following advantages: It utilizes acoustic information to monitor the near-stall point of the compression system, which is more sensitive and adaptable than traditional interstage static pressure monitoring. It reveals the group sparsity characteristics of the acoustic modal monitoring problem, and solves the acoustic modal reconstruction problem more accurately and robustly using the block orthogonal matching pursuit algorithm. It enables continuous-time monitoring of the near-stall characteristic modal amplitudes with a small number of measurement points, significantly streamlining the test system. The spatial modal characteristics of the acoustic signal are defined as the acoustic signature of the compressor near-stall point. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Various other advantages and benefits of the present invention will become apparent to those skilled in the art upon reading the detailed description of the preferred embodiments below. The accompanying drawings are intended only to illustrate preferred embodiments and are not to be construed as limiting the present invention. It should be understood that the drawings described below are merely examples of the present invention, and that those skilled in the art will be able to derive other drawings from these drawings without inventive effort. Throughout the drawings, identical reference numerals are used to denote identical components.

[0055] In the attached figure:

[0056] Figure 1 is a flow chart of the present disclosure;

[0057] Figure 2 Schematic diagram of a method and device for continuous monitoring of compressor near-stall state soundprint based on structural sparse features provided by one embodiment of the present disclosure;

[0058] Figure 3 This is a spectrum diagram of a sound pressure signal at a near-stall operating condition provided by an embodiment of the present disclosure;

[0059] FIG4(a) is a diagram of an embodiment of the present disclosure provided in Figure 3 The result of acoustic mode decomposition at the characteristic frequency 14.7EO; FIG4(b) is provided by an embodiment of the present disclosure. Figure 3 Schematic diagram of the acoustic mode decomposition results at the characteristic frequency 15.7EO;

[0060] Figure 5(a) to Figure 5(b) FIG4(a) is a schematic diagram of the acoustic modal amplitude monitoring results of two modes provided by an embodiment of the present disclosure;

[0061] Figure 6(a) to Figure 6(b) FIG4( b ) is a schematic diagram of the acoustic modal amplitude monitoring results of two modes provided by an embodiment of the present disclosure.

[0062] The present invention will be further explained below with reference to the accompanying drawings and embodiments. DETAILED DESCRIPTION

[0063] Specific embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. Although specific embodiments of the present invention are shown in the accompanying drawings, it should be understood that the present invention can be implemented in various forms and should not be limited by the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of the present invention and to fully convey the scope of the present invention to those skilled in the art.

[0064] It should be noted that certain words are used in the specification and claims to refer to specific components. Those skilled in the art should understand that technicians may use different nouns to refer to the same component. This specification and claims do not use the difference in nouns as a way to distinguish components, but use the difference in the functions of the components as the criterion for distinction. As mentioned throughout the specification and claims, "including" or "comprising" is an open term, so it should be interpreted as "including but not limited to". The subsequent description of the specification is a preferred embodiment of the present invention, but the description is based on the general principles of the specification and is not intended to limit the scope of the invention. The scope of protection of the present invention shall be as defined in the attached claims.

[0065] To facilitate understanding of the embodiments of the present invention, further explanation will be given below using specific embodiments as examples in conjunction with the accompanying drawings, and the accompanying drawings do not constitute a limitation on the embodiments of the present invention.

[0066] like Figures 1 to 6(b) As shown, the method for continuous monitoring of compressor near-stall state soundprint includes the following steps:

[0067] In the first step S1, according to the number of rotor blades of the compressor Number of stator blades Calculate the highest circumferential modal order of the compressor Number of microphones ,A predetermined number of microphone measurement points are randomly selected at the entrance to form a microphone array, and multi-channel sound pressure time domain signals are collected synchronously;

[0068] In the second step S2, a rectangular window is added to the multi-channel acoustic pressure time domain signal to construct an acoustic time domain signal segment, and a fast Fourier transform is performed on the multi-channel acoustic time domain signal segment within the current time domain window to form multiple sound pressure spectra. It is determined whether there are multiple non-synchronous single-tone characteristic frequencies that are non-integer multiples of the current rotation frequency in the sound pressure spectrum: if not, it indicates that the compressor is in normal working condition, and the second step S2 is repeated; if so, further monitoring of the near-stall condition is carried out;

[0069] In the third step S3, a multi-channel microphone continuous frequency domain measurement matrix of non-synchronous single-tone characteristic frequencies is constructed, an orthogonal Fourier transfer matrix is constructed based on the microphone sensor installation azimuth angle and the highest modal order, and a near-stall characteristic mode reconstruction model of the compressor is established with a non-uniform small number of measurement points;

[0070] In the fourth step S4, the amplitude of the compressor near-stall characteristic mode reconstruction model is solved based on the block orthogonal matching pursuit algorithm to realize the compressor near-stall condition monitoring based on the soundprint feature.

[0071] In a preferred embodiment of the method for continuous monitoring of compressor near-stall state soundprint, the first step S1 includes:

[0072] Step S101: Calculate the highest circumferential modal order, where the modal order of the circumferential acoustic mode is , represents the pressure pulsation order caused by the unsteady aerodynamic force caused by the fan's rotation-stationary interference. Represents a non-negative integer, determining the highest circumferential modal order , determine the measurable state range as ;

[0073] Step S102: Setting the number of microphones in the virtual ring and installation location, Nyquist-Shannon sampling, number of microphones With the highest modal order The relationship is , the virtual microphone array is evenly arranged in the circumferential direction, and the spacing between each virtual grid point is ;

[0074] Step S103: randomly install K microphones on the casing wall in the circumferential direction, where K is approximately the number of microphones under Nyquist sampling. 50% of the installation angle .

[0075] In a preferred embodiment of the method for continuous monitoring of compressor near-stall soundprint, the second step S2 includes the following steps:

[0076] Step S201: The sampling frequency determines the window length of the acoustic time domain signal segment, and a rectangular window is added to the multi-channel sound pressure time domain signal. The window length is and , is any integer, is the sampling frequency of the microphone sensor; and every Add windows to the time domain signal of the sound pressure of the multi-channel signal, and the overlapping length of each window ;

[0077] Step S202: the current acoustic time domain signal segment Perform fast Fourier transform on the K sound pressure time domain signals: ,in, Represents the current acoustic time domain signal fragment matrix of the sound pressure time domain signal measured by the non-uniform microphone array. The dimension of the current time domain signal fragment matrix is , represents the discrete Fourier transform, For multi-channel arrays The frequency domain matrix of the acoustic time domain signal fragment has the dimension , , which is expressed as follows:

[0078]

[0079] in ,

[0080] Step S203: Based on the obtained frequency domain matrix , draw the sound pressure spectrum measured by sensors at different positions, and observe whether there are multiple integer multiples of the frequency The non-synchronous single-tone characteristic frequency outside : If not present, repeat the second step S2; if present, further carry out monitoring of the near-stall condition.

[0081] In a preferred embodiment of the method for continuous monitoring of compressor near-stall state soundprint, the third step S3 includes:

[0082] Step S301: Establishing asynchronous single-tone characteristic frequency Multi-channel microphone continuous frequency domain measurement matrix :

[0083]

[0084] in, Represents an acoustic time domain signal segment The non-synchronous single-tone characteristic frequency of the kth measurement point of the non-uniform array is Amplitude,

[0085] Step S302: Construct an orthogonal Fourier transfer matrix ,

[0086]

[0087] in ,

[0088] Step S303: Define the acoustic mode matrix , establish the compressor near-stall characteristic mode reconstruction model,

[0089]

[0090] in, The matrix row vector, is the number of dominant modes of interest, is the structural sparse norm and has:

[0091] ,

[0092] It is an indicator function, that is, it returns 1 when the condition is true.

[0093] In a preferred embodiment of the method for continuous monitoring of compressor near-stall state soundprint, the fourth step S4 includes:

[0094] Step S401: Set the residual matrix , index number set , structural sparsity is the number of dominant modes of interest, the number of iterations ;

[0095] Step S402: Update the number of iterations ;

[0096] Step S403: Calculate support set index ;

[0097] Step S404: Update the index sequence number set ;

[0098] Step S405: Calculate the acoustic modal amplitude vector corresponding to the support set

[0099] ;

[0100] Step S406: Update the residual matrix ;

[0101] Step S407: Check the number of iterations Whether the structural sparsity is achieved , the amplitude monitoring result of the compressor near-stall characteristic mode reconstruction model is output, otherwise, the process returns to step S402.

[0102] In a preferred embodiment of the method for continuous monitoring of compressor soundprints in a near-stall state, the compressor is a compressor of an aircraft engine.

[0103] In a preferred embodiment of the method for continuous monitoring of soundprints in a near-stall state of a compressor, the compressor includes a fan structure.

[0104] In a preferred embodiment of the method for continuous monitoring of compressor near-stall state soundprint, the method includes:

[0105] The acoustic field measurement module is used to measure the number of rotor blades of the compressor Number of stator blades Calculate the highest circumferential modal order of the compressor Number of microphones ,A predetermined number of microphone measurement points are randomly selected at the entrance to form a microphone array, and multi-channel sound pressure time domain signals are collected synchronously;

[0106] A spectrum analysis module is used to add a rectangular window to the multi-channel sound pressure time domain signal to construct an acoustic time domain signal segment, and perform a fast Fourier transform on the multi-channel acoustic time domain signal segment within the current time domain window to form multiple sound pressure spectra, and determine whether there are multiple asynchronous single-tone characteristic frequencies in the sound pressure spectrum that are non-integer multiples of the current frequency conversion;

[0107] A construction module is used to construct a multi-channel microphone continuous frequency domain measurement matrix of asynchronous single-tone characteristic frequencies, construct an orthogonal Fourier transfer matrix based on the microphone sensor installation azimuth angle and the highest modal order, and establish a near-stall characteristic mode reconstruction model for a compressor with a non-uniform and small number of measurement points;

[0108] The monitoring module solves the amplitude of the compressor near-stall characteristic mode reconstruction model based on the block orthogonal matching pursuit algorithm, realizing the near-stall condition monitoring of the compressor based on the soundprint characteristics.

[0109] A computer storage medium includes computer instructions, which, when executed on a computer, cause the computer to execute the method described above.

[0110] An electronic device, comprising:

[0111] A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein:

[0112] When the processor executes the program, the method described is implemented.

[0113] In one embodiment, Figure 1 This is a flow chart of the method for continuous monitoring of the compressor near-stall state soundprint based on structural sparse features completed by the present invention. By arranging a small number of non-uniform annular microphone measuring points at the compressor, multi-channel acoustic signals of the compressor are synchronously collected, and acoustic time domain signal segments are constructed. The operating state of the compressor is judged according to the fast Fourier transform results in the current time domain window, and a reconstruction model of the compressor near-stall characteristic mode with a small number of non-uniform measuring points is established. The amplitude of the compressor near-stall characteristic sound mode is solved based on the block orthogonal matching pursuit algorithm, thereby realizing the monitoring of the compressor near-stall working condition based on soundprint features.

[0114] The schematic diagram of the aircraft engine fan structure used in the test is as follows: Figure 2 As shown, the number of first-stage rotor blades of the aircraft engine fan , number of guide blades , according to the calculation formula of the fan single-tone noise mode order , usually take The pressure pulsation order caused by the unsteady aerodynamic force caused by the fan's static interference is 1. Focus on the highest modal order ; Determine the measurable number as the modal range is , the number of measurable modes ;

[0115] A few microphones are randomly installed at the circumferential positions of the pipe inlet. The Nyquist-Shannon sampling theorem must be satisfied, that is, In this embodiment, , the number of measurement points of non-uniform annular array , the installation angle is ;

[0116] The sound pressure signal of the aircraft engine compressor is measured using a circular acoustic array. The sampling frequency of each sensor is The window length of the acoustic time domain signal segment is determined based on the sampling frequency. Data points, and in the continuous sampling process, the multi-channel signal is windowed every 1024 sampling points, that is, the overlapping length of each window ;

[0117] Perform fast Fourier transform on the sound pressure signal measured by 32 evenly distributed microphones in the current time domain signal segment to obtain the frequency domain matrix , the frequency domain matrix The size is . Draw the spectrum of the signal measured by each sensor, taking the installation position as Take the microphone signal as an example. At this time, the compressor speed is 9200r / min and the rotation frequency is 153.33Hz. The spectrum diagram near the stall point is as follows Figure 3 As shown, the horizontal axis EO is the multiple of the rotation frequency, the vertical axis is the sound pressure level, the black mark is the single tone frequency that is an integer multiple of the rotation frequency, and the red mark is the single tone frequency that is a non-integer multiple of the rotation frequency, that is, the abnormal single tone characteristic frequency.

[0118] Constructing a compressed sensing matrix ,in is the kth sensor installation angle, is the modal order of the jth circumferential modal wave, and its magnitude is .by Figure 3 Taking the two asynchronous frequencies of 14.7EO and 15.7EO as examples, we focus on the first two dominant modes at each asynchronous frequency and define , establish a sparse estimation model of acoustic modes with a few measurement points at each frequency:

[0119]

[0120] right Figure 3 The acoustic modal decomposition is performed at the abnormal peak frequency, and the acoustic modal spectra at each characteristic frequency are plotted. The acoustic modal spectrum at 14.7EO is shown in Figure 4(a), where the acoustic modal orders are observed to be +9 and -8; the acoustic modal spectrum at 15.7EO is shown in Figure 4(b), where the acoustic modal orders are observed to be +10 and -7.

[0121] The sparse model is solved by the block orthogonal matching pursuit algorithm. For comparison, the more advanced L1-norm array sparsity constraint is introduced for comparison, and the amplitudes of the two-order characteristic modes at 14.7EO and 15.7EO are monitored respectively. The monitoring results of the +9 and -8 two-order modes at 14.7EO are shown in Figures 5(a) and 5(b); the monitoring results of the +10 and -7 two-order modes at 15.7EO are shown in Figures 6(a) and 6(b). Among them, the black line represents the modal reference amplitude FSA obtained by uniform array observation of a large number of measurement points, the red line represents the modal amplitude BOMP measured based on the block orthogonal matching pursuit model described in this patent, and the blue line represents the modal amplitude SGL measured by the same advanced L1-norm array sparsity constraint method. It can be seen that the modal amplitude monitoring results based on the block orthogonal matching pursuit model described in this patent are closer to the monitoring results of a uniform array composed of a large number of measurement points. Obviously, the proposed compressor near-stall modal monitoring method based on block orthogonal matching pursuit is significantly better than the L1 norm array sparse constraint method. At the same time, the proposed method for continuous monitoring of compressor near-stall state soundprints based on structural sparse features can accurately reconstruct the time history of the characteristic mode amplitude at each frequency through a small number of measurement points. A significant increase in the amplitude of each characteristic mode at about the 6th second was observed, which can provide important reference significance for the aerodynamic stability monitoring and early warning of the compression system.

[0122] Although the embodiments of the present invention have been described above with reference to the accompanying drawings, the present invention is not limited to the above-mentioned specific embodiments and application fields. The above-mentioned specific embodiments are merely illustrative and instructive, and are not restrictive. A person skilled in the art, guided by this specification and without departing from the scope of protection of the claims of the present invention, may also devise various forms, all of which fall within the scope of protection of the present invention.

Claims

1. A method for continuous monitoring of compressor near-stall soundprint, characterized in that: The steps include: In the first step (S1), according to the number of rotor blades of the compressor Number of stator blades Calculate the highest circumferential modal order of the compressor Number of microphones ,A predetermined number of microphone measurement points are randomly selected at the entrance to form a microphone array, and multi-channel sound pressure time domain signals are collected synchronously; In the second step (S2), a rectangular window is added to the multi-channel acoustic pressure time domain signal to construct an acoustic time domain signal segment, and a fast Fourier transform is performed on the multi-channel acoustic time domain signal segment within the current time domain window to form multiple sound pressure spectra. It is determined whether there are multiple non-synchronous single-tone characteristic frequencies that are non-integer multiples of the current rotation frequency in the sound pressure spectrum: if not, it indicates that the compressor is in normal working condition, and the second step (S2) is repeated; if so, further monitoring of the near-stall condition is carried out; In the third step (S3), a multi-channel microphone continuous frequency domain measurement matrix of asynchronous single-tone characteristic frequency is constructed, an orthogonal Fourier transfer matrix is constructed based on the microphone sensor installation azimuth angle and the highest modal order, and a near-stall characteristic mode reconstruction model of the compressor with non-uniform and few measurement points is established; In the fourth step (S4), the amplitude of the compressor near-stall characteristic mode reconstruction model is solved based on the block orthogonal matching pursuit algorithm to realize the compressor near-stall condition monitoring based on the soundprint feature.

2. The method for continuous monitoring of compressor near-stall state soundprint according to claim 1, characterized in that: Preferably, the first step (S1) comprises, Step S101: Calculate the highest circumferential modal order, where the modal order of the circumferential acoustic mode is , represents the pressure pulsation order caused by the unsteady aerodynamic force caused by the fan's rotation-stationary interference. Represents a non-negative integer, determining the highest circumferential modal order , determine the measurable modal range as ; Step S102: Setting the number of microphones in the virtual ring and installation position, according to Nyquist-Shannon sampling theory, the number of microphones With the highest modal order The relationship is , the virtual microphone array is evenly arranged in the circumferential direction, and the spacing between each virtual grid point is ; Step S103: randomly install K microphones on the casing wall in the circumferential direction, where K is approximately the number of microphones under Nyquist sampling. 50% of the installation angle .

3. The method for continuous monitoring of compressor near-stall state soundprint according to claim 1, characterized in that: The second step (S2) includes the following steps: Step S201: The sampling frequency determines the window length of the acoustic time domain signal segment, and a rectangular window is added to the multi-channel sound pressure time domain signal. The window length is and , is any integer, is the sampling frequency of the microphone sensor; and every Add windows to the time domain signal of the sound pressure of the multi-channel signal, and the overlapping length of each window ; Step S202: the current acoustic time domain signal segment Perform fast Fourier transform on the K sound pressure time domain signals: ,in, Represents the current acoustic time domain signal fragment matrix of the sound pressure time domain signal measured by the non-uniform microphone array. The dimension of the current time domain signal fragment matrix is , represents the discrete Fourier transform, For multi-channel arrays The frequency domain matrix of the acoustic time domain signal fragment has the dimension , , which is expressed as follows: ， in , Step S203: Based on the obtained frequency domain matrix , draw the sound pressure spectrum measured by sensors at different positions, and observe whether there are multiple integer multiples of the frequency The non-synchronous single-tone characteristic frequency outside : If not, repeat the second step (S2); if present, further carry out monitoring of the near-stall condition.

4. The method for continuous monitoring of compressor near-stall state soundprint according to claim 3, characterized in that: The third step (S3) includes, Step S301: Establishing asynchronous single-tone characteristic frequency Multi-channel microphone continuous frequency domain measurement matrix : ， in, Represents an acoustic time domain signal segment The non-synchronous single-tone characteristic frequency of the kth measurement point of the non-uniform array is Amplitude, Step S302: Construct an orthogonal Fourier transfer matrix , ， in , Step S303: Define the acoustic mode matrix , establish the compressor near-stall characteristic mode reconstruction model, ， in, The matrix row vector, is the number of dominant modes of interest, is the structural sparse norm and has: ， It is an indicator function, that is, it returns 1 when the condition is true.

5. The method for continuous monitoring of compressor near-stall soundprint according to claim 4, characterized in that: The fourth step (S4) comprises, Step S401: Set the residual matrix , index number set , structural sparsity is the number of dominant modes of interest, the number of iterations ; Step S402: Update the number of iterations ; Step S403: Calculate support set index ; Step S404: Update the index sequence number set ; Step S405: Calculate the acoustic modal amplitude vector corresponding to the support set ； Step S406: Update the residual matrix ; Step S407: Check the number of iterations Whether the structural sparsity is achieved , the amplitude monitoring result of the compressor near-stall characteristic mode reconstruction model is output, otherwise, the process returns to step S402.

6. The method for continuous monitoring of compressor near-stall state soundprint according to claim 1, characterized in that: The compressor is the compressor of an aircraft engine.

7. The method for continuous monitoring of compressor near-stall state soundprint according to claim 1, characterized in that: The compressor includes a fan structure.

8. A fan blade synchronous vibration identification system implementing the method according to any one of claims 1 to 7, characterized in that: It includes: The acoustic field measurement module is used to measure the number of rotor blades of the compressor Number of stator blades Calculate the highest circumferential modal order of the compressor Number of microphones ,A predetermined number of microphone measurement points are randomly selected at the entrance to form a microphone array, and multi-channel sound pressure time domain signals are collected synchronously; A spectrum analysis module is used to add a rectangular window to the multi-channel sound pressure time domain signal to construct an acoustic time domain signal segment, and perform a fast Fourier transform on the multi-channel acoustic time domain signal segment within the current time domain window to form multiple sound pressure spectra, and determine whether there are multiple asynchronous single-tone characteristic frequencies in the sound pressure spectrum that are non-integer multiples of the current frequency conversion; A construction module is used to construct a multi-channel microphone continuous frequency domain measurement matrix of asynchronous single-tone characteristic frequencies, construct an orthogonal Fourier transfer matrix based on the microphone sensor installation azimuth angle and the highest modal order, and establish a near-stall characteristic mode reconstruction model for a compressor with a non-uniform and small number of measurement points; The monitoring module solves the amplitude of the compressor near-stall characteristic mode reconstruction model based on the block orthogonal matching pursuit algorithm, realizing the near-stall condition monitoring of the compressor based on the soundprint characteristics.

9. A computer storage medium, characterized in that The storage medium includes computer instructions, which, when executed on a computer, enable the computer to perform the method according to any one of claims 1 to 7.

10. An electronic device, characterized in that: The electronic device comprises: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the method according to any one of claims 1 to 7 is implemented.

Images