A pipe flow-induced noise testing system and method

Abstract

This invention proposes a method for testing flow-induced noise in pipelines, comprising the following steps: Step S1: Determine the measurement positions at the upstream and downstream flow channels of the component under test, normalize the measured signals, and calculate the pressure power spectral density of the pipe wall; Step S2: Calculate the time-domain cross-correlation coefficient of the wall pressure fluctuations and analyze the correlation characteristics between the wall pressure pulsations; Step S3: Perform spectral eigenorthogonal decomposition on the pressure pulsation field and achieve spatiotemporal decoupling through time-domain-frequency-domain conversion; Step S4: Extract the dominant flow-induced noise modes, construct an empirical frequency-wavenumber spectrum, and decompose the energy distribution of the fluctuations into contributions from different frequency and wavenumber combinations; Step S5: Analyze the decoupled flow-induced noise characteristic results and identify the amplitude-frequency characteristics of the flow-induced noise; This invention can realize real-time measurement of pressure pulsations in pipelines, achieve decoupling analysis of flow field pulsations and sound pressure pulsations in the measured signals, and identify the characteristics and generation mechanism of pipeline flow-induced noise.

Description

Technical Field

[0001] This invention relates to the field of testing technology, and in particular to a pipeline flow-induced noise testing system and method. Background Technology

[0002] Pipelines are ubiquitous in engineering applications and daily life. From water supply systems to energy transmission, pipelines perform a variety of important functions. However, the noise generated by fluids flowing through pipelines often has adverse effects on the working environment and human health. The sources of pipeline noise are very complex, including the pulsation of the fluid flow itself and the interaction of fluids as they flow through components within the pipe. With increasing awareness of the health effects of noise, how to effectively test and control pipeline noise has become an urgent problem to be solved. Existing pipeline noise testing methods are mostly limited to specific frequency ranges and have not yet provided effective testing methods for high-frequency noise in complex flow environments. Therefore, developing a new pipeline noise testing system and method is not only of great significance for improving the comfort of the working environment, but also provides necessary data support for a better understanding of the flow characteristics of fluids in pipelines, thereby providing a theoretical basis for noise control and pipeline design optimization.

[0003] Existing noise testing methods include "sound intensity method", "microphone array method", and "sound pressure measurement method". These methods mainly infer the noise source indirectly by measuring sound pressure or sound intensity. However, in pipeline systems with high-frequency noise and complex noise sources, these methods usually cannot accurately capture the full picture of the noise.

[0004] This invention proposes a novel testing system and method that uses 10 microphone dynamic pressure sensors to measure pressure pulsation signals and employs advanced decoupling analysis techniques to distinguish the pressure pulsation signals from different characteristics of noise sources. This method can more accurately capture the spatial distribution and frequency characteristics of pipeline noise sources, providing more precise data support for noise source identification and control. Summary of the Invention

[0005] This invention proposes a pipeline flow-induced noise testing system and method, namely a novel testing system and method for measuring pipeline flow-induced noise based on high-frequency dynamic pressure signals. It can realize real-time measurement of pressure pulsation in the pipeline, and can also perform decoupling analysis of flow field pulsation and sound pressure pulsation on the measured signal, thereby identifying the characteristics and generation mechanism of pipeline flow-induced noise.

[0006] The present invention adopts the following technical solution.

[0007] A method for testing flow-induced noise in a pipeline includes the following steps;

[0008] Step S1: Determine the measurement positions of several microphone high-frequency dynamic pressure signals at the upstream and downstream flow channels of the component under test. Normalize the pulsating pressure signals in the pipe measured at these positions using the flow velocity in the pipe collected by the Pitot tube, calculate the power spectral density of the pressure on the inner wall of the pipe, and explore the dominant characteristics of the flow structure, wall pressure pulsation and far-field noise.

[0009] Step S2: Calculate the time-domain cross-correlation coefficient of wall pressure fluctuations, analyze the correlation characteristics between wall pressure pulsations, and intuitively show the correlation of physical quantities at various locations;

[0010] Step S3: Perform spectral eigenorthogonal decomposition on the pressure pulsation field, and achieve spatiotemporal decoupling through time-domain to frequency-domain transformation, thereby optimally identifying the distribution characteristics of flow-induced noise;

[0011] Step S4: Perform principal component analysis on the signal to extract the dominant flow-induced noise mode, perform two-dimensional Fourier transform, construct an empirical frequency-wavenumber spectrum, and decompose the energy distribution of the fluctuation into the contribution of different frequency and wavenumber combinations.

[0012] Step S5: Analyze the decoupled flow-induced noise characteristics and identify the amplitude-frequency characteristics of the flow-induced noise.

[0013] The test system includes an air source system, a data acquisition system, a data processing system, and an acoustic system; the air source system includes a silent fan (2), a frequency converter, a flexible hose (3), a sound insulation pad (4), and a Pitot tube (5) located in the upstream flow channel; the data acquisition system includes the element under test (7) located between the upstream and downstream flow channels, and also includes an upstream microphone array (6) and a downstream microphone array (8) for measuring the high-frequency dynamic pressure signal of the microphone, and each microphone dynamic pressure sensor includes a microphone pointing towards the flow channel;

[0014] The component under test is a replaceable pipe;

[0015] The data processing system includes a workstation (1) for running data processing algorithms; the acoustic system includes an exponential diffuser (9) and an anechoic chamber (10) connected sequentially in the downstream channel.

[0016] The method of using the testing system includes the following steps;

[0017] Step 1: Calculate the shear number of the wave in the pipe based on the inner diameter of the measured component, the kinematic viscosity of the fluid medium, and the frequency of the measured component. Determine the installation spacing of each microphone based on the calculation results.

[0018] Step 2: Start the equipment and put the silent fan into suction mode;

[0019] Step 3: Calibrate the microphone of the dynamic pressure sensor during the test using a standard microphone.

[0020] Step 4: Within the range of interest, obtain the amplitude-frequency and phase-frequency characteristic curves and transfer function, verify whether the self-power spectrum of the signal to be calibrated matches the standard signal, and then reconstruct the experimental signal data;

[0021] Step 5: If the power spectrum of the signal to be calibrated does not match that of the standard signal, repeat step 4;

[0022] Step 6: Install the component under test and use a Pitot tube to collect the flow velocity information inside the tube;

[0023] Step 7: Obtain the pulsating pressure signal inside the pipe. Acquire the high-frequency dynamic pressure signals provided by ten high-frequency dynamic pressure sensors under conditions of no device under test and with device under test installed, and calculate their decibel values.

[0024] Step 8: Calculate the power spectral density of the pressure on the inner wall of the pipe to explore the dominant characteristics of the flow structure, wall pressure pulsation, and far-field noise;

[0025] Step 9: Calculate the time-domain cross-correlation coefficient of wall pressure fluctuations and analyze the correlation characteristics between wall pressure pulsations;

[0026] Step 10: Perform spectral eigenorthogonal decomposition on the pressure pulsation field, achieve spatiotemporal decoupling through time-domain to frequency-domain transformation, and obtain the optimal flow field mode that has both time orthogonality and spatial orthogonality.

[0027] Step 11: Perform principal component analysis on the pressure pulsation field to extract the dominant flow-induced noise mode; perform two-dimensional Fourier transform on the pressure field reconstructed by the opposite mode decomposition to construct the empirical frequency-wavenumber spectrum;

[0028] Step 12: Analyze the characteristics of the decoupled flow-induced noise, identify the frequency and amplitude characteristics of the flow-induced noise, and analyze its generation mechanism.

[0029] In step one, the installation spacing of each microphone With pipe diameter noise frequency and speed of sound The relation should satisfy:

[0030] Formula 1.

[0031] In step nine, the mutual coherence coefficient of wall pressure pulsations is defined as follows:

[0032] Formula 2;

[0033] In the formula, For reference point location, For reference point spacing, The cross-power spectrum between two points is defined as follows:

[0034] Formula 3;

[0035] In the formula, It is the cross-correlation function of pressure pulsations between two points.

[0036] Step ten, the method for implementing spectral eigenorthogonal decomposition includes the following steps:

[0037] Step S10: Given a set of transient field sequences, first subtract the time mean from each transient field. , obtain the pulsating field ( = 1, 2, …, ), reconstructed as:

[0038] Formula 4;

[0039] Step S20: The energy of the transient field collected at each time step can be represented by the spatial inner product:

[0040] Formula 5;

[0041] in, Let be a positive definite Hermitian matrix, representing the weights of each component, and Ω be the spatial domain of the decomposition. Indicates complex conjugate;

[0042] Step S30: Decompose the instantaneous energy and seek the mode with the optimal spacetime inner product:

[0043] Formula 6;

[0044] Step S40: Divide the transient field set into The segments contain some overlapping measurement sequences, and each series contains An instantaneous field, namely

[0045] Formula 7;

[0046] Step S50: Perform a time-domain weighted discrete Fourier transform on each sequence. The Fourier transform data matrix is ​​obtained as follows:

[0047] Formula 8;

[0048] in, Representing the The first frequency at the frequency The Fourier realization, in the _ ... The matrix of all Fourier realizations at each frequency is:

[0049] Formula 9;

[0050] Step S60: Modality and corresponding energy value Through CSD matrix The eigenvectors and eigenvalues ​​are obtained from the eigenvalues;

[0051] Formula 10;

[0052] Step S70: Obtain the coefficients of the SPOD mode in the Fourier transform. It is a column matrix. , corresponding to the The SPOD modes at each frequency are:

[0053] Formula 11;

[0054] in, The SPOD mode energies are arranged from high to low, i.e. ,and This refers to the modal characteristics of noise.

[0055] In step eleven, a two-dimensional Fourier transform is performed on the pressure field reconstructed by modal decomposition to construct an empirical frequency-wavenumber spectrum, expressed by the formula:

[0056] Formula 12;

[0057] in, Represents angular frequency. and These are the window functions in the spatial and time domains, respectively; the formula for calculating the wavenumber-frequency spectrum is:

[0058] Formula 13.

[0059] This spectrum (empirical frequency-wavenumber spectrum) decomposes the energy distribution of fluctuations into contributions from different combinations of frequencies and wavenumbers. After temporal and spatial Fourier transforms, the reconstructed pressure field displays clear energy transport ridges. The slope of these ridges represents the transport velocity of the energy components. For modes identified as fluid kinetic energy, this portion of the reconstruction dissipates rapidly with the flow in the pipe. However, acoustic energy propagates rapidly along the pipe, penetrating the entire pipe system. This enables the identification and characterization of flow-induced noise components.

[0060] The upstream microphone array includes five microphone dynamic pressure sensors PT1, PT2, PT3, PT4, and PT5 equidistantly arranged upstream of the flow channel, and the downstream microphone array includes five microphone dynamic pressure sensors PT6, PT7, PT8, PT9, and PT10 equidistantly arranged downstream of the flow channel.

[0061] In step one, the microphone dynamic pressure sensor adopts a pin-hole installation method, that is, a small hole with a diameter of 1 mm is drilled in the pipe wall for pipe wall pressure measurement, so as to reduce the influence of the small hole on the flow of the pipe wall; a wedge-shaped cavity is set between the microphone and the small hole to fully receive the pulsating pressure information transmitted by the small hole.

[0062] According to this connection method, the resonant frequency of the signal transmission cavity can be calculated using Formula 1, which is approximately 5.5 kHz, meeting the experimental measurement requirements.

[0063] Formula A1;

[0064] In step three, the experimental microphones are calibrated sequentially using a standard microphone. The calibration principle is based on the interchangeability technology of the standard microphone and the microphone to be calibrated. An acoustic coupling cavity microphone calibration device is used, the main body of which is composed of a long cylinder with an inner diameter of 85 mm. The inside of the wall is filled with a sponge with high sound absorption performance. A speaker is installed at one end of the device, and the standard microphone and the microphone to be tested are installed at the other end.

[0065] During calibration, white noise is first generated by a signal generator, and after being amplified by a power amplifier, a plane wave signal is emitted by a speaker. The two microphones simultaneously measure the signal. Based on the transfer function of a linear invariant system, the amplitude-frequency and phase-frequency characteristics of the microphone under test can be obtained.

[0066] by and y These are used as timing signals for the standard microphone and the microphone to be calibrated, respectively. , and These are represented as the self-power spectra of the two microphones and the cross-power spectrum between them, respectively. The cross-power spectral density function is a complex number, expressed as:

[0067] Formula A2;

[0068] The transfer function between the standard microphone and the microphone to be calibrated is

[0069] Formula A3;

[0070] The transfer function contains information about the amplitude-frequency and phase-frequency relationships. Based on this, the pressure signal data obtained from the actual measurement of each microphone to be calibrated is reconstructed. According to the discrete Fourier transform, the output signal of the linear array microphone is:

[0071] Formula A4;

[0072] in Formula A5;

[0073] The reconstructed signal is obtained through inverse discrete Fourier transform.

[0074] Formula A6;

[0075] This allowed for the calibration of all 10 test microphones, improving measurement accuracy.

[0076] In step eight, the data collected by the microphone is the transient pressure pulsation value. It is the instantaneous value minus the mean. To obtain, that is:

[0077] Formula A7;

[0078] Its Fourier transform is:

[0079] Formula A8;

[0080] Its one-sided self-power spectral density is defined as:

[0081] Formula A9;

[0082] In the formula, for The complex conjugate;

[0083] In step nine, the time-domain cross-correlation coefficient of the wall pressure fluctuation is calculated to analyze the correlation characteristics between wall pressure pulsations. This visually demonstrates the correlation of physical quantities at various locations. The cross-correlation function is essentially the dimensionless result of the cross-power spectrum of pressure pulsations measured by the dynamic pressure sensors PT1, PT2, PT3, PT4, PT5, PT6, PT7, PT8, PT9, and PT10 at various points on the wall. When the cross-correlation function presents a vertical line perpendicular to the frequency axis, it indicates that the signals are highly coherent at each location, and the measurement is determined to show significant flow-induced noise characteristics.

[0084] The component under test is a rigid, straight pipe with no bends and a rigid wall, used for noise testing in the frequency range of 0-10kHz.

[0085] This invention proposes a novel testing system and method for measuring pipeline flow-induced noise based on high-frequency dynamic pressure signals from ten microphones. It can not only achieve real-time measurement of pressure pulsation within the pipeline, but also perform decoupled analysis of flow field pulsation and sound pressure pulsation on the measured signals, thereby identifying the characteristics and generation mechanism of pipeline flow-induced noise, providing a reference for the measurement and analysis of pipeline flow-induced noise. Attached Figure Description

[0086] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments:

[0087] Appendix Figure 1 This is a schematic diagram of the test method for measuring pipeline flow-induced noise based on high-frequency dynamic pressure signals from ten microphones, provided in an embodiment of the present invention.

[0088] Appendix Figure 2 This is a schematic diagram of a pipeline flow-induced noise measurement system based on high-frequency dynamic pressure signals from ten microphones, provided in an embodiment of the present invention.

[0089] In the diagram: 1. Workstation (i.e., the control unit containing the signal acquisition module and the signal processing module); 2. Silent fan; 3. Flexible hose; 4. Sound insulation pad; 5. Pitot tube; 6. Upstream microphone array (i.e., microphone dynamic pressure sensors PT1, PT2, PT3, PT4, PT5); 7. Component under test; 8. Downstream microphone array (i.e., microphone dynamic pressure sensors PT6, PT7, PT8, PT9, PT10); 9. Exponential diffuser tube; 10. Anechoic chamber. Detailed Implementation

[0090] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0091] It should be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0092] The terms “comprising” and “including” indicate the presence of the described feature, whole, step, operation, element, and / or component, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components, and / or collections thereof. The terms “and / or” refer to any combination of one or more of the associated listed items and all possible combinations, and include such combinations.

[0093] like Figure 1 As shown, a method for testing flow-induced noise in a pipeline includes the following steps;

[0094] Step S1: Determine the measurement positions of several microphone high-frequency dynamic pressure signals at the upstream and downstream flow channels of the component under test. Normalize the pulsating pressure signals in the pipe measured at these positions using the flow velocity in the pipe collected by the Pitot tube, calculate the power spectral density of the pressure on the inner wall of the pipe, and explore the dominant characteristics of the flow structure, wall pressure pulsation and far-field noise.

[0095] Step S2: Calculate the time-domain cross-correlation coefficient of wall pressure fluctuations, analyze the correlation characteristics between wall pressure pulsations, and intuitively show the correlation of physical quantities at various locations;

[0096] Step S3: Perform spectral eigenorthogonal decomposition on the pressure pulsation field, and achieve spatiotemporal decoupling through time-domain to frequency-domain transformation, thereby optimally identifying the distribution characteristics of flow-induced noise;

[0097] Step S4: Perform principal component analysis on the signal to extract the dominant flow-induced noise mode, perform two-dimensional Fourier transform, construct an empirical frequency-wavenumber spectrum, and decompose the energy distribution of the fluctuation into the contribution of different frequency and wavenumber combinations.

[0098] Step S5: Analyze the decoupled flow-induced noise characteristics and identify the amplitude-frequency characteristics of the flow-induced noise.

[0099] The testing system includes an air supply system, a data acquisition system, a data processing system, and an acoustic system. The air supply system includes a silent fan 2, a frequency converter, a flexible hose 3, a sound insulation pad 4, and a Pitot tube 5 located in the upstream flow channel. The data acquisition system includes the element under test 7 located between the upstream and downstream flow channels, and also includes an upstream microphone array 6 and a downstream microphone array 8 for measuring the high-frequency dynamic pressure signal of the microphone. Each microphone dynamic pressure sensor includes a microphone pointing towards the flow channel.

[0100] The component under test is a replaceable pipe;

[0101] The data processing system includes a workstation 1 for running data processing algorithms; the acoustic system includes an exponential diffuser 9 and an anechoic chamber 10 connected sequentially in the downstream flow channel.

[0102] The method of using the testing system includes the following steps;

[0103] Step 1: Calculate the shear number of the wave in the pipe based on the inner diameter of the measured component, the kinematic viscosity of the fluid medium, and the frequency of the measured component. Determine the installation spacing of each microphone based on the calculation results.

[0104] Step 2: Start the equipment and put the silent fan into suction mode;

[0105] Step 3: Use a standard microphone to calibrate the microphone of the dynamic pressure sensor during the test (calibrate within the range of interest).

[0106] Step 4: Within the range of interest, obtain the amplitude-frequency and phase-frequency characteristic curves and transfer function, verify whether the self-power spectrum of the signal to be calibrated matches the standard signal, and then reconstruct the experimental signal data;

[0107] Step 5: If the power spectrum of the signal to be calibrated does not match that of the standard signal, repeat step 4;

[0108] Step 6: Install the component under test and use a Pitot tube to collect the flow velocity information inside the tube;

[0109] Step 7: Obtain the pulsating pressure signal inside the pipe. Acquire the high-frequency dynamic pressure signals provided by ten high-frequency dynamic pressure sensors under conditions of no device under test and with device under test installed, and calculate their decibel values.

[0110] Step 8: Calculate the power spectral density of the pressure on the inner wall of the pipe to explore the dominant characteristics of the flow structure, wall pressure pulsation, and far-field noise;

[0111] Step 9: Calculate the time-domain cross-correlation coefficient of wall pressure fluctuations and analyze the correlation characteristics between wall pressure pulsations;

[0112] Step 10: Perform spectral eigenorthogonal decomposition on the pressure pulsation field, achieve spatiotemporal decoupling through time-domain to frequency-domain transformation, and obtain the optimal flow field mode that has both time orthogonality and spatial orthogonality.

[0113] Step 11: Perform principal component analysis on the pressure pulsation field to extract the dominant flow-induced noise mode; perform two-dimensional Fourier transform on the pressure field reconstructed by the opposite mode decomposition to construct the empirical frequency-wavenumber spectrum;

[0114] Step 12: Analyze the characteristics of the decoupled flow-induced noise, identify the frequency and amplitude characteristics of the flow-induced noise, and analyze its generation mechanism.

[0115] In step one, the installation spacing of each microphone With pipe diameter noise frequency and speed of sound The relation should satisfy:

[0116] Formula 1.

[0117] In step nine, the mutual coherence coefficient of wall pressure pulsations is defined as follows:

[0118] Formula 2;

[0119] In the formula, For reference point location, For reference point spacing, The cross-power spectrum between two points is defined as follows:

[0120] Formula 3;

[0121] In the formula, It is the cross-correlation function of pressure pulsations between two points.

[0122] Step ten, the method for implementing spectral eigenorthogonal decomposition includes the following steps:

[0123] Step S10: Given a set of transient field sequences, first subtract the time mean from each transient field. , obtain the pulsating field ( = 1, 2, …, ), reconstructed as:

[0124] Formula 4;

[0125] Step S20: The energy of the transient field collected at each time step can be represented by the spatial inner product:

[0126] Formula 5;

[0127] in, Let be a positive definite Hermitian matrix, representing the weights of each component, and Ω be the spatial domain of the decomposition. Indicates complex conjugate;

[0128] Step S30: Decompose the instantaneous energy and seek the mode with the optimal spacetime inner product:

[0129] Formula 6;

[0130] Step S40: Divide the transient field set into The segments contain some overlapping measurement sequences, and each series contains An instantaneous field, namely

[0131] Formula 7;

[0132] Step S50: Perform a time-domain weighted discrete Fourier transform on each sequence. The Fourier transform data matrix is ​​obtained as follows:

[0133] Formula 8;

[0134] in, Representing the The first frequency at the frequency The Fourier realization, in the _ ... The matrix of all Fourier realizations at each frequency is:

[0135] Formula 9;

[0136] Step S60: Modality and corresponding energy value Through CSD matrix The eigenvectors and eigenvalues ​​are obtained from the eigenvalues;

[0137] Formula 10;

[0138] Step S70: Obtain the coefficients of the SPOD mode in the Fourier transform. It is a column matrix. , corresponding to the The SPOD modes at each frequency are:

[0139] Formula 11;

[0140] in, The SPOD mode energies are arranged from high to low, i.e. ,and This refers to the modal characteristics of noise.

[0141] In step eleven, a two-dimensional Fourier transform is performed on the pressure field reconstructed by modal decomposition to construct an empirical frequency-wavenumber spectrum, expressed by the formula:

[0142] Formula 12;

[0143] in, Represents angular frequency. and These are the window functions in the spatial and time domains, respectively; the formula for calculating the wavenumber-frequency spectrum is:

[0144] Formula 13.

[0145] This spectrum (empirical frequency-wavenumber spectrum) decomposes the energy distribution of fluctuations into contributions from different combinations of frequencies and wavenumbers. After temporal and spatial Fourier transforms, the reconstructed pressure field displays clear energy transport ridges. The slope of these ridges represents the transport velocity of the energy components. For modes identified as fluid kinetic energy, this portion of the reconstruction dissipates rapidly with the flow in the pipe. However, acoustic energy propagates rapidly along the pipe, penetrating the entire pipe system. This enables the identification and characterization of flow-induced noise components.

[0146] The upstream microphone array includes five microphone dynamic pressure sensors PT1, PT2, PT3, PT4, and PT5 equidistantly arranged upstream of the flow channel, and the downstream microphone array includes five microphone dynamic pressure sensors PT6, PT7, PT8, PT9, and PT10 equidistantly arranged downstream of the flow channel.

[0147] In step one, the microphone dynamic pressure sensor adopts a pin-hole installation method, that is, a small hole with a diameter of 1 mm is drilled in the pipe wall for pipe wall pressure measurement, so as to reduce the influence of the small hole on the flow of the pipe wall; a wedge-shaped cavity is set between the microphone and the small hole to fully receive the pulsating pressure information transmitted by the small hole.

[0148] According to this connection method, the resonant frequency of the signal transmission cavity can be calculated using Formula 1, which is approximately 5.5 kHz, meeting the experimental measurement requirements.

[0149] Formula A1;

[0150] In step three, the experimental microphones are calibrated sequentially using a standard microphone. The calibration principle is based on the interchangeability technology of the standard microphone and the microphone to be calibrated. An acoustic coupling cavity microphone calibration device is used, the main body of which is composed of a long cylinder with an inner diameter of 85 mm. The inside of the wall is filled with a sponge with high sound absorption performance. A speaker is installed at one end of the device, and the standard microphone and the microphone to be tested are installed at the other end.

[0151] During calibration, white noise is first generated by a signal generator, and after being amplified by a power amplifier, a plane wave signal is emitted by a speaker. The two microphones simultaneously measure the signal. Based on the transfer function of a linear invariant system, the amplitude-frequency and phase-frequency characteristics of the microphone under test can be obtained.

[0152] by and y These are used as timing signals for the standard microphone and the microphone to be calibrated, respectively. , and These are represented as the self-power spectra of the two microphones and the cross-power spectrum between them, respectively. The cross-power spectral density function is a complex number, expressed as:

[0153] Formula A2;

[0154] The transfer function between the standard microphone and the microphone to be calibrated is

[0155] Formula A3;

[0156] The transfer function contains information about the amplitude-frequency and phase-frequency relationships. Based on this, the pressure signal data obtained from the actual measurement of each microphone to be calibrated is reconstructed. According to the discrete Fourier transform, the output signal of the linear array microphone is:

[0157] Formula A4;

[0158] in Formula A5;

[0159] The reconstructed signal is obtained through inverse discrete Fourier transform.

[0160] Formula A6;

[0161] This allowed for the calibration of all 10 test microphones, improving measurement accuracy.

[0162] In step eight, the data collected by the microphone is the transient pressure pulsation value. It is the instantaneous value minus the mean. To obtain, that is:

[0163] Formula A7;

[0164] Its Fourier transform is:

[0165] Formula A8;

[0166] Its one-sided self-power spectral density is defined as:

[0167] Formula A9;

[0168] In the formula, for The complex conjugate;

[0169] In step nine, the time-domain cross-correlation coefficient of the wall pressure fluctuation is calculated to analyze the correlation characteristics between wall pressure pulsations. This visually demonstrates the correlation of physical quantities at various locations. The cross-correlation function is essentially the dimensionless result of the cross-power spectrum of pressure pulsations measured by the dynamic pressure sensors PT1, PT2, PT3, PT4, PT5, PT6, PT7, PT8, PT9, and PT10 at various points on the wall. When the cross-correlation function presents a vertical line perpendicular to the frequency axis, it indicates that the signals are highly coherent at each location, and the measurement is determined to show significant flow-induced noise characteristics.

[0170] The component under test is a rigid, straight pipe with no bends and a rigid wall, used for noise testing in the frequency range of 0-10kHz.

[0171] Example 1:

[0172] As shown in the figure, a novel pipeline flow-induced noise testing system and method are disclosed. The testing system is driven by a silent fan and analyzes the flow field pulsation and flow-induced noise generated by the tested element through decoupling calculations of high-frequency dynamic pressure signals collected by ten dynamic pressure sensors at the pipeline. The system is characterized by: the ten high-frequency dynamic pressure sensors including five PT1, PT2, PT3, PT4, and PT5 equidistantly arranged upstream (from the tested element to the silent fan), and five PT6, PT7, PT8, PT9, and PT10 equidistantly arranged downstream (from the exponential diffuser to the tested element); the silent fan operates in a suction mode to reduce the influence of background noise.

[0173] The calculation process for measuring the current-induced noise of the measured component includes the following steps:

[0174] Step 1: Calculate the shear number of the wave in the pipe based on the pipe's inner diameter, the kinematic viscosity of the fluid medium, and the frequency of the measured element. Determine the installation position and spacing of each microphone based on the calculation results.

[0175] Step 2: Start the equipment; the silent fan will operate in suction mode to reduce the impact of background noise.

[0176] Step 3: Calibrate the experimental microphone. Use a standard microphone to calibrate the experimental microphone within the range of interest in the experiment.

[0177] Step 4: Within the range of interest, obtain the corresponding amplitude-frequency and phase-frequency characteristic curves, and derive their amplitude-frequency and phase-frequency transfer functions to verify whether the self-power spectrum of the signal to be calibrated matches the standard signal. Subsequently, the experimentally acquired signal is reconstructed, and the reconstructed data is essentially identical to the characteristics of the real pressure pulsation signal.

[0178] Step 5: If the power spectrum of the signal to be calibrated does not match that of the standard signal, repeat step 4 to recalibrate the microphone.

[0179] Step Six: Install the component under test and use a Pitot tube to collect the flow velocity information inside the tube. The Pitot tube is placed in the air source section for flow velocity measurement. The distance between the Pitot tube and the test section is long enough, and the flow velocity is recorded in real time by an air flow meter to ensure that the Reynolds number inside the tube is maintained at the target operating condition for different components under test.

[0180] Step 7: Obtain the pulsating pressure signal inside the pipe. Acquire high-frequency dynamic pressure signals from ten high-frequency dynamic pressure sensors (PT1, PT2, PT3, PT4, PT5, PT6, PT7, PT8, PT9, PT10) under both conditions with and without the device under test (DUT), and calculate their decibel values. Ensure that the dynamic pressure signal acquired by the microphone after installing the DUT is at least 10 dB higher than the signal without the DUT.

[0181] Step 8: Based on the high-frequency dynamic pressure signal provided by the dynamic pressure sensor in Step 7, calculate the pressure power spectral density of the inner wall of the pipe, and explore the dominant characteristics of the flow structure, wall pressure pulsation and far-field noise.

[0182] Step Nine: Calculate the time-domain cross-correlation coefficient of the wall pressure fluctuations, analyze the correlation characteristics between wall pressure pulsations, and intuitively display the correlation of physical quantities at various locations. The cross-correlation function is essentially a dimensionless result of the cross-power spectrum of pressure pulsations at various points on the wall (PT1, PT2, PT3, PT4, PT5, PT6, PT7, PT8, PT9, PT10). When the coherence function presents a vertical line perpendicular to the frequency axis, it indicates that the signal is highly coherent at all locations, which can be considered as exhibiting significant flow-induced noise characteristics.

[0183] Step 10: Perform spectral eigenorthogonal decomposition on the pressure pulsation field. Spatiotemporal decoupling is achieved through time-domain to frequency-domain transformation, yielding optimal flow field modes that possess both temporal and spatial orthogonality (i.e., spatial-frequency orientation). At each frequency, the spectral eigenorthogonal decomposition yields a set of energy-ranked orthogonal modes, thus optimally identifying the flow-induced noise distribution characteristics.

[0184] Step 11: Perform principal component analysis on the pressure pulsation field to extract the dominant flow-induced noise modes. A two-dimensional Fourier transform is performed on the reconstructed pressure field using the opposing mode decomposition to construct an empirical frequency-wavenumber spectrum. This spectrum decomposes the energy distribution of the fluctuations into contributions from different frequency and wavenumber combinations. After time and space Fourier transforms, the reconstructed pressure field displays clear energy transport ridges. The slopes of these ridges represent the transport velocity of the energy components. For modes identified as fluid kinetic energy, this portion of the reconstruction result dissipates rapidly with the flow in the pipe. Meanwhile, the acoustic energy of flow-induced noise propagates rapidly along the pipe, penetrating the entire pipe system. This enables the characteristic identification of the noise.

[0185] Step 12: Analyze the characteristics of the decoupled flow-induced noise, identify the frequency and amplitude characteristics of the flow-induced noise, and analyze its generation mechanism.

[0186] In step one, a microphone is mounted on the pipe wall for measuring the pulsating pressure on the pipe surface. The traditional method is flush-mounted, where the microphone's pressure-sensing diaphragm is parallel to the inner surface of the pipe wall. Due to the relatively large size of the microphone, this method undoubtedly affects the spatial resolution of the signal and the effective frequency of the measurement. Furthermore, this mounting method is difficult to align, and drilling too large a hole in the pipe wall poses a risk of affecting flow and noise characteristics. To avoid these problems, a pin-hole mounting method is adopted, where a small hole with a diameter of 1 mm is drilled in the pipe wall for pipe wall pressure measurement, minimizing the impact of the hole on the pipe wall flow. A wedge-shaped cavity is formed between the microphone and the hole to fully receive the pulsating pressure information transmitted through the hole. With this connection method, the resonant frequency of the signal transmission cavity can be calculated using Formula 1, which is approximately 5.5 kHz, meeting the experimental measurement requirements.

[0187]

[0188] In step three, standard microphones are used to calibrate the experimental microphones sequentially. The calibration principle is based on the interchangeability technology of the standard microphone and the microphone under test. An acoustic coupling cavity microphone calibration device is used, the main body of which consists of a long cylinder with an inner diameter of 85 mm. The interior of the device is filled with high-absorption sound-insulating foam material. A speaker is mounted at one end of the device, and the standard microphone and the microphone under test are simultaneously mounted at the other end. During calibration, a signal generator emits white noise, which is amplified by a power amplifier and then emitted as a plane wave signal by the speaker. Both microphones simultaneously measure this signal. Based on the transfer function of a linearly invariant system, the amplitude-frequency and phase-frequency characteristics of the microphone under test can be obtained. and y These are used as timing signals for the standard microphone and the microphone to be calibrated, respectively. , and These are represented as the self-power spectra of the two microphones and the cross-power spectrum between them, respectively. The cross-power spectral density function is a complex number and can be expressed as:

[0189]

[0190] The transfer function between the standard microphone and the microphone to be calibrated is

[0191]

[0192] This transfer function contains information about both amplitude and phase frequency relationships, which can be used to reconstruct the pressure signal data obtained from actual measurements of each microphone to be calibrated. According to the Discrete Fourier Transform, the output signal of the linear array microphone is...

[0193]

[0194] in

[0195]

[0196] The reconstructed signal can be obtained through inverse discrete Fourier transform.

[0197]

[0198] This allowed for the calibration of all 10 test microphones, improving measurement accuracy.

[0199] In step eight, the data collected by the microphone is the transient pressure pulsation value. It is the instantaneous value minus the mean. To obtain, that is:

[0200]

[0201] Its Fourier transform is:

[0202]

[0203] Its one-sided self-power spectral density is defined as:

[0204]

[0205] In the formula, for .

[0206] In step nine, the mutual coherence coefficient of wall pressure pulsations is defined as:

[0207]

[0208] In the formula, For reference point location, For reference point spacing, The cross-power spectrum between two points is defined as follows:

[0209]

[0210] In the formula, It is the cross-correlation function of pressure pulsations between two points.

[0211] In step ten, the spectral eigenorthogonal decomposition is implemented as follows:

[0212] S10: Given a set of transient field sequences, first subtract the time mean from each transient field. , obtain the pulsating field ( = 1, 2, …, ), reconstructed as:

[0213]

[0214] S20: The energy of the transient field collected at each time step can be represented by the spatial inner product:

[0215]

[0216] in, Let be a positive definite Hermitian matrix, representing the weights of each component, and Ω be the spatial domain of the decomposition. It indicates complex conjugate.

[0217] S30: Decompose instantaneous energy and seek the mode with optimal spacetime inner product:

[0218]

[0219] Solving the eigenvalue problem in the frequency domain can be achieved using the two-point spatiotemporal correlation matrix of the Fourier transform.

[0220] S40: Divide the transient field set into... The segments contain some overlapping measurement sequences, and each series contains An instantaneous field, namely

[0221]

[0222] Under the ergodic assumption, each sequence can be regarded as an independent representation of the noise field. The purpose of this data segmentation is to increase the size of the dataset.

[0223] S50: Perform a time-domain weighted discrete Fourier transform on each sequence. The Fourier transform data matrix is ​​obtained as follows:

[0224]

[0225] in, Representing the The first frequency at the frequency The Fourier realization, in the _ ... The matrix of all Fourier realizations at each frequency is:

[0226]

[0227] S60: Modal and corresponding energy value Through CSD matrix It is obtained from the eigenvectors and eigenvalues.

[0228]

[0229] S70: Obtaining the coefficients of the SPOD mode in the Fourier transform. It is a column matrix. , corresponding to the The SPOD modes at each frequency are:

[0230]

[0231] in, The SPOD mode energies are arranged from high to low, i.e. ,and This refers to the modal characteristics of noise.

[0232] In step eleven, an empirical frequency-wavenumber spectrum is constructed by performing a two-dimensional Fourier transform on the pressure field reconstructed from mode decomposition.

[0233]

[0234] in, Represents angular frequency. and These are the window functions for the spatial and time domains, respectively. Subsequently, the formula for calculating the wavenumber-frequency spectrum is:

[0235]

[0236] This spectrum decomposes the energy distribution of fluctuations into contributions from combinations of different frequencies and wavenumbers. After temporal and spatial Fourier transforms, the reconstructed pressure field reveals clear energy transport ridges. The slopes of these ridges represent the transport velocity of the energy components. For modes identified as fluid kinetic energy, this portion of the reconstruction dissipates rapidly with the flow in the pipe. However, acoustic energy propagates rapidly along the pipe, penetrating the entire pipe system. This enables the identification and characterization of flow-induced noise components.

[0237] Example 2:

[0238] like Figure 2 As shown, in this example:

[0239] A silent fan 2 provides stable airflow to the pipeline. The fan's noise and vibration are extremely low, and its impact on the experiment is negligible. Simultaneously, the fan speed can be controlled via a frequency converter, thereby controlling the airflow conditions in the test section to obtain a specified flow velocity. A flexible hose 3 isolates any remaining minor vibrations from the fan; a sound-insulating pad 4 further eliminates background noise caused by the fan and the environment; a Pitot tube 5 measures the flow velocity within the pipeline and is installed in the air source section. The distance between the Pitot tube and the test section is sufficiently long to ensure that the flow fully develops within the test section; the measured element 7 is replaceable and used to measure the pipeline flow-induced noise caused by different elements; an exponential diffuser 9 eliminates standing waves in the working section while minimizing sound wave reflections at both ends of the pipeline; an anechoic chamber 10 reduces background noise while preventing noise reflections from generating unnecessary standing waves and noise.

[0240] Microphone dynamic pressure sensors PT1, PT2, PT3, PT4, and PT5 at the upstream end of the duct and microphone dynamic pressure sensors PT6, PT7, PT8, PT9, and PT10 at the downstream end of the duct measure the flow-induced noise generated by the measured element within the duct.

[0241] The pipe-side pulsating pressure signal obtained in step seven is normalized using the pipe-side flow velocity acquired by the Pitot tube in step six. Then, based on the high-frequency dynamic pressure signal provided by the dynamic pressure sensor in step seven, the power spectral density of the pipe wall pressure is calculated to explore the dominant characteristics of the flow structure, wall pressure pulsation, and far-field noise. Next, according to step nine, the time-domain cross-correlation coefficient of the wall pressure fluctuation is calculated to analyze the correlation characteristics between wall pressure pulsations, visually demonstrating the correlation of physical quantities at various locations. According to step ten, the pressure pulsation field is subjected to spectral eigenorthogonal decomposition, achieving spatiotemporal decoupling through time-domain to frequency-domain transformation, resulting in the optimal flow field modes that possess both temporal and spatial orthogonality. At each frequency, the spectral eigenorthogonal decomposition yields a set of energy-ranked orthogonal modes, thus optimally identifying the distribution characteristics of flow-induced noise. According to step eleven, principal component analysis is performed on the pressure pulsation field to extract the dominant flow-induced noise modes. A two-dimensional Fourier transform was performed on the pressure field reconstructed by modal decomposition to construct an empirical frequency-wavenumber spectrum. This spectrum decomposes the energy distribution of the fluctuation into contributions from different combinations of frequencies and wavenumbers. Finally, the characteristics of the decoupled flow-induced noise were analyzed based on the results from step twelve to identify the frequency and amplitude characteristics of the flow-induced noise and to analyze its generation mechanism.

[0242] The above description is merely a preferred embodiment of the present invention and is used only to illustrate the method of the present invention and not to limit it. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method of testing for flow-induced noise in a pipe, the method comprising: The test system comprises a gas source system, a data acquisition system, a data processing system and an acoustic system; the gas source system comprises a mute fan (2), a frequency converter, a flexible hose (3), a sound insulation lining (4) and a Pitot tube (5) located at an upstream flow channel; the data acquisition system comprises a measured element (7) located between the upstream flow channel and a downstream flow channel, and further comprises an upstream microphone array (6) for measuring high-frequency dynamic pressure signals of microphones, a downstream microphone array (8), each microphone dynamic pressure sensor comprising a microphone pointing to the flow channel; The measured element is a replaceable pipeline; The data processing system comprises a workstation (1) for running a data processing algorithm; the acoustic system comprises an exponential diffuser (9) and a sound attenuation chamber (10) connected in sequence at the downstream flow channel; The test method comprises the following steps: Step one: according to the inner diameter of the measured element, the kinematic viscosity of the fluid medium and the frequency of the measured element, the shear number of the wave in the pipeline is solved and calculated, and the installation interval of each microphone is determined according to the calculation result; Step two: start the equipment, and make the mute fan work in the suction mode; Step three: calibrate the dynamic pressure sensor of the test microphone by using a standard microphone; The acoustic coupling cavity microphone calibration device is mainly composed of a long cylinder, the wall surface is filled with high sound absorption performance sponge sound insulation material, one end of the device is provided with a loudspeaker, and the other end is provided with a standard microphone and a to-be-measured microphone; During calibration, white noise is generated by a signal generator, and a plane wave signal is emitted by the loudspeaker after gain by a power amplifier; x(t) and y(t) are used as the time sequence signals of the standard microphone and the to-be-calibrated microphone respectively, and the transfer function is obtained; The signals collected by each to-be-calibrated microphone are reconstructed by inverse discrete Fourier transform; Step four: in the range of interest, the amplitude-frequency and phase-frequency characteristic curves and the transfer function are obtained, whether the to-be-calibrated signal self-power spectrum is consistent with the standard signal is verified, and then the experimental signal is reconstructed; Step five: if the to-be-calibrated signal self-power spectrum is not consistent with the standard signal, repeat step four; Step six: install the measured element, and collect the flow velocity information in the pipeline by using a Pitot tube; Step seven: obtain the fluctuating pressure signal in the pipeline, collect the high-frequency dynamic pressure signals provided by all high-frequency microphone dynamic pressure sensors in the upstream microphone array (6) and the downstream microphone array (8) under the conditions that the to-be-measured element is not installed and is installed respectively, and calculate the decibel values; the upstream microphone array comprises five equally spaced microphone dynamic pressure sensors PT1, PT2, PT3, PT4 and PT5 arranged upstream of the flow channel, and the downstream microphone array comprises five equally spaced microphone dynamic pressure sensors PT6, PT7, PT8, PT9 and PT10 arranged downstream of the flow channel; Step eight: calculate the wall surface pressure power spectrum density in the pipeline; Step nine: calculate the wall surface pressure fluctuation time domain cross-correlation coefficient, analyze the correlation characteristics between the wall surface pressure fluctuations, and intuitively display the correlation of physical quantities at each position; the mutual coherence function is essentially the dimensionless mutual power spectrum of the pressure fluctuations measured by each microphone dynamic pressure sensor, and is defined as: where x0is the reference point position, ξ is the reference point separation, Φ pp (ξ, f; x0) denotes the cross-power spectrum between two points, which is defined as: wherein R pp (ξ,τ;x0) is the cross-correlation function of the pressure fluctuations between two points; When the coherence function of formula 1 presents a vertical line perpendicular to the frequency axis, indicating that the signals at various positions are highly coherent, it is determined that the significant flow noise characteristics are exhibited; Step ten: performing spectral proper orthogonal decomposition on the pressure fluctuation field, realizing time-space decoupling through time-frequency domain conversion, and obtaining optimal flow field modes with time orthogonality and space orthogonality; Step eleven: performing principal component analysis on the pressure fluctuation field to extract dominant flow noise modes; through two-dimensional Fourier transform on the pressure field after mode decomposition reconstruction, an empirical frequency-wavenumber spectrum is constructed; Step twelve: analyzing the decoupled flow noise characteristic results to identify the frequency and amplitude-frequency characteristics of the flow noise; The measured element is a straight pipe with no bending and rigid pipe wall, which is used for noise testing in the frequency range of 0-10 kHz; In the test system, the silent fan provides stable airflow for the measured element pipe, and the noise and vibration of the fan are low enough to be ignored; the speed of the silent fan is controlled by a frequency converter, so as to control the airflow conditions of the test section to obtain the specified flow rate; The dynamic pressure sensor of the microphone adopts pinhole type installation: a small hole with a diameter of 1 mm is drilled on the pipe wall, and a wedge-shaped cavity is arranged between the microphone and the small hole.

2. The pipeline flow-induced noise testing method according to claim 1, wherein in step one, the relationship between the installation interval d of each microphone, the pipe diameter D, the noise frequency f and the sound speed c should satisfy:

3. The pipeline flow-induced noise testing method according to claim 1, wherein in step eleven, the calculation formula of the empirical frequency-wavenumber spectrum is: wherein ω represents the angular frequency, W(x) and W(t) are window functions in the spatial domain and the time domain respectively; the calculation formula of the wavenumber-frequency spectrum is:

4. The pipeline flow-induced noise testing method according to claim 1, wherein the Fourier transform thereof is: The wall surface pressure power spectral density is defined as:

5. The pipeline flow-induced noise testing method according to claim 1, wherein in step ten, the implementation method of the spectral proper orthogonal decomposition includes the following steps: In step eight, the data collected by the microphone is the value of the instantaneous pressure fluctuation p ′ (t), which is the instantaneous value minus the mean value resulting in, i.e.: Step S20: the energy of the transient field collected at each time can be expressed by spatial inner product: Step S30: decompose the instantaneous energy and seek the mode with optimal time-space inner product: wherein is the complex conjugate of P. ​ ​ Step S10: dividing the set of transient fields into measurement sequences with inter-segment overlap, for each set of transient field sequences, first subtract the time mean from each transient field resulting in the fluctuating field q i = q(t i )(i = 1, 2,..., n t ), reconstructed as: ​ where W is a positive definite Hermitian matrix representing the weights of the components, Ω is the spatial domain of decomposition, and (·) * denotes complex conjugation. ​ Step S40: dividing the transient field set into n blk sequences with some overlap, each series containing n fft transient fields, i.e. Step S50: Discrete Fourier transform with time domain weighting for each sequence Obtained Fourier transformed data matrix: wherein, represents the kth Fourier realization at the ith frequency, and the matrix of all Fourier realizations at the lth frequency is: Step S60: The modal Φ and the corresponding energy value λ can be obtained by the eigenvectors and eigenvalues of the CSD matrix ​ Step S70: Obtain the coefficient of SPOD mode Ψ in the Fourier transform l which is a column matrix Then the SPOD mode corresponding to the lth frequency is: wherein, is the SPOD modal energy ranked from high to low, i.e. and is the modal characteristic of the noise.

Images