Aircraft engine fan acoustic modal detection method and system based on semi-threshold iterative shrinkage

Abstract

A method and system for detecting acoustic modes of an aircraft engine fan based on semi-threshold iterative contraction are disclosed, wherein the order of the rotor-static interference mode is calculated by the aircraft engine fan model parameters, and the acoustic mode monitoring range, the number of microphone sensors, and the installation angle are determined; a sensing matrix is ​​constructed according to the sensor layout scheme and the monitoring range, and L 1 / 2 The norm regularization term constructs a compressed sensing model for acoustic mode detection; based on the semi-threshold representation algorithm, combined with the number of rotation-static interference modes, the threshold parameters in the semi-threshold representation algorithm are set, and the regularization parameters of the compressed sensing model are corrected to improve the model; based on the microphone sensor array, the sound field information in the fan duct is obtained, the rotation-static interference acoustic mode is solved by the model, and the acoustic mode amplitude is output after comparison with the calculation result. The disclosure can improve the recognition accuracy of the acoustic mode of the aircraft engine fan, and the result is also more sparse.

Description

Technical Field

[0001] The invention belongs to the technical field of aircraft engine noise testing, in particular to an aircraft engine fan sound mode detection method based on semi-threshold iterative contraction. Background Art

[0002] With the widespread application of high bypass ratio turbofan aircraft engines, fan noise has gradually become the dominant component of aircraft engine noise, which has attracted widespread social attention. According to its formation mode, fan noise is divided into single-tone noise at each blade passing frequency and broadband noise distributed in the entire frequency domain. Among them, under subsonic flight conditions, the intensity of single-tone noise is significantly greater than the other noise components, and the acoustic mode in single-tone noise is the dominant component of single-tone noise of aircraft engine fans. The refined noise reduction design of aircraft engine fans requires a deep understanding of the generation and propagation of acoustic modes in single-tone noise, and the decomposition of the modal structure of single-tone noise and the detection of acoustic modes play an important guiding role in this regard.

[0003] The use of annular microphone arrays to decompose the circumferential acoustic modes of aircraft engine fan single-tone noise is currently the most commonly used method in the field of aircraft engine fan noise acoustic mode decomposition. The classic uniformly arranged microphone array has the disadvantages of high testing cost, complex test system, and poor anti-failure ability, making it difficult to apply to the actual aircraft engine duct acoustic mode measurement. The compressed sensing model of acoustic mode detection based on non-uniform layout L1 norm sparse regularization often underestimates the amplitude of the dominant mode of rotor-static interference, and the recognition success rate is greatly reduced under strong noise background. In view of the problems existing in the classical acoustic mode decomposition method and the acoustic mode detection method with L1 norm sparse regularization, combined with the sparse prior of the single-tone noise acoustic mode in the wavenumber domain, the L1 norm sparse regularization method with few measurement points and non-uniform microphone layout is established. 1 / 2 The norm sparse regularized acoustic modal detection model improves the cleanliness and robustness of the circumferential acoustic modal detection of aircraft engine fan single-tone noise, and realizes the accurate identification of the acoustic modal order and amplitude of aircraft engine fan noise by a few microphones.

[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 in this country to a person of ordinary skill in the art. Summary of the invention

[0005] In view of the problems existing in the prior art, the present invention proposes a method for detecting the acoustic modes of an aircraft engine fan based on semi-threshold iterative contraction. 1 / 2 The acoustic modal detection model based on norm sparse regularization term solves the problems of complex testing of traditional detection methods and significant underestimation of the amplitude of L1 norm detection methods. It successfully and accurately identifies the acoustic modes of aircraft engines through a small number of sensors and fast calculations.

[0006] The objective of the present invention is achieved through the following technical solution: a method for detecting the acoustic modal of an aircraft engine fan based on semi-threshold iterative shrinkage comprises the following steps:

[0007] In the first step, the acoustic modal order m of the fan single-tone noise is calculated according to the number of rotor blades B and the number of stator blades V of the aircraft engine fan, and the acoustic modal monitoring range and the number of microphone sensors n are determined based on the maximum modal order of the acoustic modal order m. CS and the mounting angle of the microphone sensor

[0008] In the second step, the dimension of the spatial circumferential wave number domain basis vector is determined by the acoustic mode monitoring range to construct an acoustic mode sparse representation dictionary; the number of microphone sensors n is determined by the number of microphone sensors n. CS And the microphone sensor installation angle Construct the observation matrix; merge the observation matrix and the sparse representation dictionary into the perception matrix, and establish the L 1 / 2 Compressed sensing optimization model with norm sparse regularization term;

[0009] In the third step, based on the semi-threshold representation algorithm, the number of acoustic modes m of the single-tone noise is set as the sparsity K of the trans-static interference mode, and the threshold parameter T in the semi-threshold representation algorithm is adaptively set by the sparsity K. (k) =[z (k) ] K+1 , correct the regularization parameter of the compressed sensing model Then improve the compressed sensing optimization model;

[0010] In the fourth step, the sound field information in the fan duct is obtained using a microphone sensor array, the acoustic mode is solved using the compressed sensing optimization model, and the acoustic mode amplitude is output after comparison with the calculated result.

[0011] In the method described, in the first step, the acoustic modal order m of the fan single-tone noise is m=nB±kV, wherein n represents the pressure pulsation order caused by the unsteady aerodynamic force caused by the fan rotation-static interference, and k represents an integer.

[0012] In the method described above, in the first step, the number n of microphone sensors in the microphone array is determined. CS and the microphone sensor installation angle,

[0013] S101. Calculate the number of sensors n required for modal detection based on the Nyquist sampling theorem NS , and its relationship with the acoustic mode order m is: NS ≥2m+1;n NS The installation angle of each sensor is

[0014] S102, select the number of microphone sensors n CS , which is related to the number of sensors n under Nyquist sampling theorem NS The relationship is: n CS <<n NS ;

[0015] S103, randomly select n CS The installation position of the actual sensor is used to determine the circumferential installation angle of the sensor in the microphone array at the same axial installation position.

[0016] In the method described, in the second step, a method based on L 1 / 2 The compressed sensing model of the norm sparse regularization term includes the following steps:

[0017] S201. Using the spatial circumferential wavenumber domain basis vectors as atoms of the sparse dictionary Φ, construct an acoustic mode sparse representation model: P = ΦS, where: Represents n NS The complex amplitude of the sound pressure signal measured by the uniformly distributed microphones at the frequency ω, Φ represents the sparse representation dictionary, and S represents the expression of the sound pressure signal measured by the microphone sensor in the acoustic modal domain;

[0018] S202, by the actual number of microphone sensors n CS And installation angle Construct the observation matrix Ψ:

[0019]

[0020] S203, based on the measurement matrix Ψ and the sparse representation dictionary Φ, a compressed sensing model of acoustic modal sparse decomposition of single-tone noise is established as follows:

[0021] y=ΨP=ΨΦS=ΘS,

[0022] Where y represents n CS The sound pressure signal measured by the microphone array, Θ represents the perception matrix of the compressed sensing model, and S represents the expression of the sound pressure signal measured by the microphone sensor in the acoustic modal domain;

[0023] S204, establish based on L 1 / 2 Compressed sensing unconstrained optimization model with norm sparse regularization term:

[0024]

[0025] Among them, λ represents the regularization parameter, ||·||2 represents the 2-norm, and ||·|| 1 / 2 represent Norm.

[0026] In the method described, in the third step, the dominant mode of the rotation-static interference usually has only one order m, and the sparsity K of the rotation-static interference mode is determined to be 1, and the threshold parameter T in the semi-threshold representation algorithm is adaptively set according to the sparsity K. (k) =[z (k) ] K+1 , where [z (0) ] K+1 =0,[z (k) ] K+1 The regularization parameters of the compressed sensing model are automatically updated with the number of iterations k. Improve compressed sensing optimization models.

[0027] In the method, in the fourth step, using a microphone sensor array to obtain sound field information in the fan duct, and using the compressed sensing optimization model to solve the acoustic mode includes the following steps:

[0028] S401, using a microphone sensor array to obtain the sound field information in the fan duct, and using Fourier transform to extract the sound pressure vector of the microphone sensor array at a frequency ω

[0029] S402, based on the semi-threshold representation theory, according to the regularization parameter, using the algorithm of iterative shrinkage threshold to solve the compressed sensing unconstrained optimization model to obtain the solution result of the acoustic mode;

[0030] S403, determining whether the acoustic mode solution result is equal to the modal order, and if so, outputting the acoustic mode solution result.

[0031] An aircraft engine fan acoustic modal detection system for implementing the method comprises:

[0032] A perception matrix construction module is used to construct a perception matrix between the sound pressure vector measured by the sound array sensor and the sound mode vector in the fan single-tone noise;

[0033] The sound field measurement module is used to measure the sound field information propagated in the duct to the installation position of the sound array when the fan is working;

[0034] Acoustic mode recognition module is used to construct L-based 1 / 2 A compressed sensing model of the norm is proposed, and a semi-threshold iterative shrinkage method is used to solve the compressed sensing model to detect the dominant sound mode propagating in the pipeline.

[0035] In the system, the sound field measurement module includes a microphone array for measuring sound field information.

[0036] Technical features and effects of the present invention:

[0037] 1. The present invention uses fewer sound measurement points to realize the detection of the rotation-static interference mode, which saves the number of sensors and installation costs of the microphone array compared to the traditional sound mode detection method;

[0038] 2. The present invention uses a semi-threshold iteration method to obtain the precise amplitude of the rotation-static interference mode. Compared with the solution of the L1 norm sparse regularization term model, the detection accuracy is greatly improved and the result is more sparse;

[0039] 3. The present invention utilizes a threshold adaptive strategy to reduce the setting of regularization parameters in the model, improves the robustness of the method, and can adapt to different speed conditions for acoustic mode detection, with a wide range of adaptability.

[0040] The above description is only an overview of the technical solution of the present invention. In order to make the technical means of the present invention clearer and to enable those skilled in the art to implement it according to the contents of the specification, and to make the above and other purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described below by way of example. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] By reading the detailed description of the preferred specific embodiments below, various other advantages and benefits of the present invention will become clear to those of ordinary skill in the art. The drawings in the specification are only for the purpose of illustrating the preferred embodiments and are not considered to be limitations of the present invention. Obviously, the drawings described below are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings without creative work. Moreover, the same reference numerals are used to represent the same components throughout the drawings.

[0042] In the attached picture:

[0043] Figure 1 is a flow chart of the present disclosure;

[0044] Figure 2 It is a schematic diagram of an aircraft engine fan acoustic mode detection device based on semi-threshold iterative contraction provided by an embodiment of the present disclosure;

[0045] Figure 3 It is a schematic diagram of the sound pressure and the observed vector amplitude measured by the microphone sound array sensor provided by one embodiment of the present disclosure;

[0046] Figure 4 is a flow chart of a semi-threshold iterative shrinkage algorithm of the present disclosure;

[0047] Figure 5(a) to Figure 5(b) The single-tone noise acoustic mode provided by an embodiment of the present disclosure is based on the L1 norm and the L 1 / 2Schematic diagram of acoustic modal amplitude comparison of norm detection; Figure 5(a) shows L 1 / 2 norm; Figure 5(b) shows the L1 norm.

[0048] The present invention is further explained below in conjunction with the accompanying drawings and embodiments. DETAILED DESCRIPTION

[0049] The following will refer to the attached Figure 1 to Figure 5(b) Specific embodiments of the present invention are described in more detail. 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 described herein. On the contrary, 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.

[0050] 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 present invention. The scope of protection of the present invention shall be determined by the attached claims.

[0051] To facilitate understanding of the embodiments of the present invention, further explanation will be given below by taking specific embodiments as examples in conjunction with the accompanying drawings, and each of the accompanying drawings does not constitute a limitation on the embodiments of the present invention.

[0052] The present disclosure provides an aircraft engine fan acoustic modal detection method based on semi-threshold iterative shrinkage, comprising the following steps:

[0053] In the first step (S1), according to the fan model of the aircraft engine, the rotation-static interference modal order in the fan single-tone noise is calculated, and the acoustic mode monitoring range, the number of microphone sensors, the axial installation position and the circumferential installation angle are determined according to the maximum modal order of interest;

[0054] In the second step (S2), the dimension of the spatial circumferential wave number domain basis vector is determined by the acoustic mode monitoring range, and the acoustic mode sparse representation dictionary is constructed; the measurement matrix is ​​constructed by the number of microphone sensors and the installation position; the measurement matrix and the sparse dictionary are combined into a perception matrix, and an L-based 1 / 2 Compressed sensing optimization model with norm sparse regularization term;

[0055] In the third step (S3), based on the semi-threshold representation algorithm, the acoustic mode sparsity is determined in combination with the physical sparse prior information of the acoustic mode of the rotation-static interference mode, and the threshold adaptive strategy is used to improve the model and automatically set the regularization parameter;

[0056] In the fourth step (S4), a stable operation test of the aircraft engine compressor is carried out, and the sound field information in the fan duct is obtained using a microphone sensor array. The acoustic mode is solved using the above model and compared with the calculated results to output the acoustic mode amplitude.

[0057] In a preferred embodiment, in the first step (S1), the fan single-tone noise acoustic modal order m is calculated based on the number of rotor blades B and the number of stator blades V of the aircraft engine fan: m = nB ± kV, wherein n represents the pressure pulsation order caused by the unsteady aerodynamic force caused by the rotation-stator interference of the fan, and k represents an integer.

[0058] In a preferred embodiment, in the first step (S1), the number of sensors n in the microphone array is determined. CS and installation angle The following steps are involved:

[0059] S101. Calculate the number of sensors n required for modal detection based on the Nyquist sampling theorem NS , and its relationship with the acoustic mode order m is: NS ≥2m+1; According to Nyquist's equidistant sampling theorem, n NS The installation angle of each sensor is

[0060] S102, selecting the appropriate number of microphone sensors n CS , which is related to the number of sensors n under Nyquist sampling theorem NS The relationship is: n CS <<n NS ;

[0061] S103, randomly select n CS The installation position of the real sensor is used to determine the circumferential installation angle of the sensor in the microphone array at the same axial installation position.

[0062] In a preferred embodiment, in the second step (S2), a L-based 1 / 2 The compressed sensing model of the norm sparse regularization term includes the following steps:

[0063] S201. Using the spatial circumferential wavenumber domain basis vectors as atoms of the sparse dictionary Φ, a sparse representation model of acoustic modes is constructed:

[0064] P=ΦS

[0065] in, Represents n NS The complex amplitude of the sound pressure signal measured by the uniformly distributed microphones at the frequency ω, Φ represents the sparse dictionary, and S represents the expression of the sound pressure signal measured by the microphone sensor in the acoustic modal domain;

[0066] S202, by the actual number of microphone sensors n CS And installation location Construct the observation matrix Ψ:

[0067]

[0068] S203, based on the above-mentioned measurement matrix Ψ and sparse dictionary Φ, a compressed sensing model of sparse decomposition of single-tone noise acoustic modal is established as follows:

[0069] y=ΨP=ΨΦS=ΘS

[0070] Where y represents n CS The sound pressure signal measured by the acoustic array, Θ represents the perception matrix of the compressed sensing model, and S represents the expression of the sound pressure signal measured by the microphone sensor in the acoustic modal domain;

[0071] S204, establish based on L 1 / 2 Compressed sensing unconstrained optimization model with norm sparse regularization term:

[0072]

[0073] Where y represents n CS The sound pressure signal measured by the acoustic array, Θ represents the perception matrix in the compressed sensing model, S represents the expression of the sound pressure signal measured by the microphone sensor in the acoustic modal domain, λ represents the regularization parameter, ||·||2 represents the L2 norm, ||·|| 1 / 2 Represents L 1 / 2 Norm.

[0074] In a preferred embodiment, in the third step (S3), the dominant mode of the rotation-static interference usually has only one order m, and the sparsity K of the rotation-static interference mode is determined to be 1, and the threshold parameter T in the semi-threshold representation algorithm is adaptively set according to the sparsity K. (k) =[z (k) ] K+1 , where [z (0) ] K+1 =0,[z (k) ] K+1 The regularization parameters of the compressed sensing model are automatically updated with the number of iterations k. Improve compressed sensing optimization models;

[0075] In a preferred embodiment, in the fourth step (S4), a stable operation test of an aircraft engine compressor is carried out, and the sound field information in the fan duct is obtained by using a microphone sensor array. Solving the acoustic mode using the above model includes the following steps:

[0076] S401, using a microphone sensor array to obtain the sound field information in the fan duct, and using Fourier transform to extract the sound pressure vector of the microphone sensor array at a frequency ω

[0077] S402, according to the regularization parameter λ set by the threshold adaptive strategy, using the algorithm of iterative shrinkage threshold to solve the compressed sensing unconstrained optimization model for the acoustic modal detection of the single-tone noise of the fan of the aircraft engine, including the following steps:

[0078] (1) Set the error threshold ε0, the number of iterations k = 0, and the initial solution S 0 =0, initial residual Threshold T (k) =0, sparsity K;

[0079] (2) Calculate the intermediate variables of the acoustic mode amplitude

[0080] (3) Set the threshold T according to the sparsity K (k) =[z (k) ] K+1 , where [z (k) ] K+1 Represents z (k) Update the regularization parameter Calculate intermediate variables

[0081] (4) Traverse each dimension of the modal amplitude The component is compared with the threshold T (k) Compare the size, when renew in For S (k+1) The i-th dimension component of is θ (k) The i-th dimension component of ; otherwise, update

[0082] (5) The number of iterations is increased by k = k + 1; and the residual is updated. Determine whether the stop condition r is reached k ≤ε 0 If yes, output S (k) ; Otherwise, continue (2) to iterate.

[0083] S403, determining whether the solution result of the acoustic mode is equal to the order of the rotation-static interference mode obtained in the first step, and outputting the solution result of the acoustic mode.

[0084] The present disclosure also provides an aircraft engine fan acoustic modal detection system based on semi-threshold iterative shrinkage, comprising the following modules:

[0085] A perception matrix construction module is used to construct a perception matrix between the sound pressure vector measured by the sound array sensor and the sound mode vector in the fan single-tone noise;

[0086] The sound field measurement module is used to measure the sound field information propagated in the duct to the installation position of the sound array when the fan is working;

[0087] Acoustic mode recognition module is used to construct L-based 1 / 2 A compressed sensing model of the norm is proposed, and a semi-threshold iterative shrinkage method is used to solve the compressed sensing model to detect the dominant sound mode propagating in the pipeline.

[0088] In a preferred embodiment, the sound field information is measured by a microphone array.

[0089] In a preferred embodiment, the compressed sensing model is:

[0090]

[0091] Where y represents n CS The sound pressure signal measured by the acoustic array, Θ represents the perception matrix in the compressed sensing model, S represents the expression of the sound pressure signal measured by the microphone sensor in the acoustic modal domain, λ represents the regularization parameter, ||·||2 represents the L2 norm, ||·|| 1 / 2 Represents L 1 / 2 Norm.

[0092] The present invention is further described below in conjunction with Figures 1 to 5(b) and using a specific simulation example instead of the experimental example. It should be emphasized that the following description is merely exemplary, and the application of the present invention is not limited to the following examples.

[0093] Figure 1 The present invention is a flowchart of an aircraft engine fan acoustic modal detection method based on semi-threshold iterative shrinkage. The method uses an aircraft engine fan model to calculate the rotation-static interference modal order in the fan single-tone noise, and determines the acoustic modal monitoring range, the number of microphone sensors, the axial installation position and the circumferential installation angle according to the maximum modal order of interest; determines the spatial circumferential wavenumber domain basis vector dimension by the acoustic modal monitoring range, constructs an acoustic modal sparse representation dictionary, constructs an observation matrix by the number of microphone sensors and the installation position, merges the observation matrix and the sparse dictionary into a perception matrix, and establishes an acoustic modal detection method based on L 1 / 2The compressed sensing optimization model of the norm sparse regularization term is further carried out to conduct stable operation tests of aircraft engine compressors, use microphone sensor arrays to obtain the sound field information in the fan duct, use the above model to solve the acoustic mode, and compare it with the calculated results to output the acoustic mode amplitude. The specific steps are as follows:

[0094] 1) Assuming that the number of rotor blades of the aircraft engine fan is B=16 and the number of stator blades is V=6, according to the calculation formula of the acoustic mode order of the fan single-tone noise m=nB±kV, n=1 is usually taken to represent the pressure pulsation order caused by the unsteady aerodynamic force caused by the fan rotor-stator interference as 1, and k=1 is taken to represent the first-order acoustic mode, and m=10 is calculated;

[0095] 2) Calculate the number of sensors n required for modal detection based on the Nyquist sampling theorem NS ≥2m+1, take n NS =32 satisfies the condition, evenly distribute the sensors around the casing, and take n CS =6 is the number of sensors required for the compressed sensing method; according to the installation positions of 6 real sensors randomly selected, the circumferential installation angle θ of the sensors in the microphone array at the same axial installation position is determined to be [0°, 56.25°, 123.75°, 180°, 337.5°, 348.75°], such as Figure 2 As shown;

[0096] 3) Use the spatial circumferential wave number domain basis vectors to construct a sparse dictionary Φ and a sparse representation model of acoustic modes, P = Φs, and the actual number of microphone sensors n CS =32 and installation position θ = [0°, 56.25°, 123.75°, 180°, 337.5°, 348.75°] to construct the observation matrix Ψ, the size of Ψ is 6×32, and φ 1，1 =1,φ 2，6 =1,φ 3，12 =1,φ 4，17 =1,φ 5，31 =1,φ 6，32 =1,φ i，j is the element of the measurement matrix Ψ, and the rest of the positions are all 0; based on the above measurement matrix Ψ and sparse dictionary Φ, establish L 1 / 2 Compressed sensing unconstrained optimization model for sparse decomposition of single-tone noise acoustic modes with norm sparse regularization term:

[0097] 4) The dominant mode of the static-to-smooth interference has only one order m=10, and the acoustic mode sparsity K=1 is determined. The threshold parameter T in the semi-threshold representation algorithm is adaptively set by the sparsity K. (k) =[z (k) ] K+1 , where [z(0) ] K+1 =0,[z (k) ] K+1 The regularization parameter of the compressed sensing model will be automatically updated with the number of iterations k in the subsequent solution process. Improve compressed sensing optimization models;

[0098] 5) Construct a sound time domain signal with the dominant mode of the static-to-static interference m=10, add Gaussian white noise to simulate the background noise, perform Fourier transform on the time domain signal measured by the microphone sensor array, extract the sound pressure at the frequency ω=3000Hz, and add a certain degree of Gaussian white noise to simulate the background noise in the actual measurement, and obtain y=[y1(3000), y6(3000), y 12 (3000), y 17 (3000), y 31 (3000), y 32 (3000)] as the observation vector, such as Figure 3 As shown;

[0099] 6) By inputting the parameters of the compressed sensing model established in step 3) and the sound pressure amplitude y, according to Figure 4 The result is consistent with the m=10 calculated in step 1), and the output result is shown in Figure 5(a). The acoustic mode recognition result based on the L1 norm is similar, as shown in Figure 5(b). 1 / 2 The acoustic modal amplitude solved by the L1 model is 1.51Pa, the acoustic modal amplitude solved by the L1 model is 1.39Pa, and the acoustic modal amplitude measured by the traditional method is 1.511Pa. 1 / 2 Compared with the acoustic mode amplitude solved by the L1 model, the acoustic mode amplitude solved by the semi-threshold iterative shrinkage method is more stable and has higher solution accuracy.

Claims

1. A method for detecting the acoustic modal of an aircraft engine fan based on semi-threshold iterative shrinkage, the method comprising the following steps: In the first step (S1), according to the number of rotor blades of the aircraft engine fan Number of stator blades Calculation of the Rotor-Station Interference Modal Order of the Fan Single-tone Noise , based on the acoustic mode order The maximum modal order determines the acoustic mode monitoring range and the number of microphone sensors and the mounting angle of the microphone sensor ; In the second step (S2), the dimension of the spatial circumferential wavenumber domain basis vector is determined by the acoustic mode monitoring range to construct an acoustic mode sparse representation dictionary; the number of microphone sensors is determined by the number of spatial circumferential wavenumber domain basis vectors to construct an acoustic mode sparse representation dictionary; And the microphone sensor installation angle Construct the observation matrix; merge the observation matrix and the sparse representation dictionary into the perception matrix, and establish the L 1 / 2 Compressed sensing optimization model with norm sparse regularization term; In the third step (S3), based on the semi-threshold representation algorithm, the acoustic mode of the single-tone noise is combined The number of is set as the sparsity of the rotation-static interference mode , through the sparsity Adaptively set the threshold parameter in the semi-threshold representation algorithm , correct the regularization parameter of the compressed sensing model , thereby improving the compressed sensing optimization model; In the fourth step (S4), the sound field information in the fan duct is obtained by using a microphone sensor array, the acoustic mode is solved by using the compressed sensing optimization model, and the acoustic mode amplitude is output after comparison with the calculated result.

2. The method according to claim 1, wherein: In the first step (S1), the acoustic modal order of the fan single-tone noise is for, ,in, represents the pressure pulsation order caused by the unsteady aerodynamic force caused by the fan's rotation and static interference. Represents an integer.

3. The method according to claim 1, wherein: In the first step (S1), the number of microphone sensors in the microphone array is determined and the microphone sensor installation angle, S101. Calculate the number of sensors required for modal detection based on Nyquist sampling theorem , which is related to the acoustic mode order The relationship is: ; The installation angle of each sensor is ; S102, select the number of microphone sensors , which is consistent with the number of sensors under Nyquist sampling theorem The relationship is: ; S103, Random Selection The installation position of the actual sensor is used to determine the circumferential installation angle of the sensor in the microphone array at the same axial installation position. .

4. The method according to claim 1, wherein: In the second step (S2), a L 1 / 2 The compressed sensing model of the norm sparse regularization term includes the following steps: S201, using spatial circumferential wavenumber domain basis vectors as sparse dictionary Atoms of the acoustic mode sparse representation model are constructed: ,in, represent The sound pressure signal measured by the uniformly distributed microphones is The complex amplitude under represents a sparse representation dictionary, Represents the expression of the sound pressure signal measured by the microphone sensor in the acoustic modal domain; S202, by actual number of microphone sensors And installation angle Construct observation matrix : ， S203, based on the observation matrix With sparse representation dictionary , the compressed sensing model of acoustic modal sparse decomposition of single-tone noise is established as: ， in, express The sound pressure signal measured by the microphone array is represents the perception matrix of the compressed sensing model, Represents the expression of the sound pressure signal measured by the microphone sensor in the acoustic modal domain; S204, establish based on L 1 / 2 Compressed sensing unconstrained optimization model with norm sparse regularization term: ， in, represents the regularization parameter, represents the 2-norm, represent Norm.

5. The method according to claim 4, wherein: In the third step (S3), the dominant mode of the rotation-static interference usually has only one order , determine the sparsity of the rotation-static interference mode , through the sparsity Adaptively set the threshold parameter in the semi-threshold representation algorithm ,in , As the number of iterations Automatically update and correct the regularization parameters of the compressed sensing model , improving the compressed sensing optimization model.

6. The method according to claim 1, wherein: In the fourth step (S4), the sound field information in the fan duct is obtained by using a microphone sensor array, and the acoustic mode is solved by using the compressed sensing optimization model, which includes the following steps: S401, using a microphone sensor array to obtain sound field information in the fan duct, and using Fourier transform to extract the frequency The sound pressure vector ; S402, based on the semi-threshold representation theory, according to the regularization parameter, using the algorithm of iterative shrinkage threshold to solve the compressed sensing unconstrained optimization model to obtain the solution result of the acoustic mode; S403, determining whether the acoustic mode solution result is equal to the modal order, and if so, outputting the acoustic mode solution result.

7. An aircraft engine fan acoustic modal detection system implementing the method according to any one of claims 1 to 6, comprising: A perception matrix construction module is used to construct a perception matrix between the sound pressure vector measured by the sound array sensor and the sound mode vector in the fan single-tone noise; The sound field measurement module is used to measure the sound field information propagated in the duct to the installation position of the sound array when the fan is working; Acoustic mode recognition module, used to construct L-based 1 / 2 A compressed sensing model of the norm is proposed, and a semi-threshold iterative shrinkage method is used to solve the compressed sensing model to detect the dominant sound mode propagating in the pipeline.

8. The system according to claim 7, wherein: The sound field measurement module includes a microphone array for measuring sound field information.

Images