Wind tunnel-based noise detection method and device, storage medium and electronic equipment

Abstract

This application discloses a wind tunnel-based noise detection method, apparatus, storage medium, and electronic device. The method includes: when testing a test piece in a wind tunnel, acquiring a measured noise spectrum in the testing environment and obtaining the operating frequency of the wind tunnel's inverter; matching the background noise spectrum corresponding to the operating frequency in a background noise spectrum library based on the operating frequency; aligning and dividing the measured noise spectrum and the background noise spectrum using a frequency band division rule adapted to the inverter's operating frequency to obtain multiple frequency band units; based on the acoustic energy density of the measured noise spectrum and the background noise spectrum within each frequency band unit, extracting the energy component corresponding to the background noise spectrum from the measured noise spectrum through directional orthogonal projection, and reconstructing the spectrum of the residual energy after extraction to obtain the effective aerodynamic noise spectrum of the test piece; and calculating the noise detection result of the test piece based on the effective aerodynamic noise spectrum.

Description

Technical Field

[0001] This application relates to the field of wind tunnel-based noise detection technology, and more particularly to a wind tunnel-based noise detection method, apparatus, storage medium, and electronic device. Background Technology

[0002] Current technologies for detecting aerodynamic noise in wind tunnels typically subtract the background noise from the total noise signal collected by the sensors under empty wind tunnel conditions, using this difference as the aerodynamic noise data of the test specimen. However, in the actual wind tunnel acoustic field, the aerodynamic noise of the test specimen and the background noise of the wind tunnel couple and interfere with each other. The superposition process of the wind tunnel background noise and the aerodynamic noise of the test specimen is nonlinear, not a simple linear superposition. Current technologies, using a linear subtraction method, cannot eliminate the energy changes caused by coupling interference, making it difficult to accurately separate the background noise component. This results in a significant deviation between the obtained test specimen noise data and the actual aerodynamic noise, leading to inaccurate test results. Summary of the Invention

[0003] In view of the above problems, this application provides a wind tunnel-based noise detection method, apparatus, storage medium, and electronic device.

[0004] To solve the above-mentioned technical problems, this application proposes the following solution: Firstly, this application provides a wind tunnel-based noise detection method, comprising: when testing a test piece in a wind tunnel, acquiring a measured noise spectrum in the testing environment and obtaining the operating frequency of the wind tunnel's inverter; matching the operating frequency with the corresponding background noise spectrum in a background noise spectrum library, the background noise spectrum library including the correspondence between the operating frequency of the wind tunnel's inverter and the background noise spectrum in an empty wind field without the test piece; using a frequency band division rule adapted to the inverter's operating frequency to align and divide the measured noise spectrum and the background noise spectrum to obtain multiple frequency band units; based on the acoustic energy density of the measured noise spectrum and the background noise spectrum within each frequency band unit, extracting the energy component corresponding to the background noise spectrum from the measured noise spectrum through directional orthogonal projection, and reconstructing the spectrum of the residual energy after extraction to obtain the effective aerodynamic noise spectrum of the test piece; and calculating the noise detection result of the test piece based on the effective aerodynamic noise spectrum.

[0005] Secondly, this application provides a wind tunnel-based noise detection device, which includes: The acquisition module is used to acquire the measured noise spectrum in the test environment and obtain the operating frequency of the wind tunnel's frequency converter when the test piece is being tested in the wind tunnel. The determination module is used to match the background noise spectrum corresponding to the operating frequency in the background noise spectrum library based on the operating frequency. The background noise spectrum library includes the correspondence between the operating frequency of the wind tunnel inverter and the background noise spectrum in an empty wind field without the test specimen. The partitioning module is used to align and partition the measured noise spectrum and the background noise spectrum using a frequency band partitioning rule adapted to the inverter's operating frequency, thereby obtaining multiple frequency band units; The stripping module is used to strip the energy component corresponding to the background noise spectrum from the measured noise spectrum by means of directional orthogonal projection, based on the acoustic energy density of the measured noise spectrum and the background noise spectrum within each frequency band unit. The reconstruction module is used to reconstruct the spectrum of the residual energy after stripping to obtain the effective aerodynamic noise spectrum of the test piece. The detection module is used to calculate the noise detection results of the test piece based on the effective aerodynamic noise spectrum.

[0006] To achieve the above objectives, according to a third aspect of this application, a storage medium is provided, the storage medium including a stored program, wherein, when the program is executed, the device where the storage medium is located is controlled to perform the wind tunnel-based noise detection method of the first aspect described above.

[0007] To achieve the above objectives, according to a fourth aspect of this application, an electronic device is provided, the device including at least one processor, and at least one memory and bus connected to the processor; wherein the processor and memory communicate with each other through the bus; the processor is used to call program instructions in the memory to execute the wind tunnel-based noise detection method of the first aspect described above.

[0008] By employing the above-described technical solution, the technical solution provided in this application has at least the following advantages: This application uses the acoustic energy density of measured noise and background noise within each frequency band unit to strip the background noise energy component using directional orthogonal projection. This method uses the energy correlation of two sets of signals within the same frequency band as the calculation basis, first determining the actual energy proportion and contribution component of the background noise in the measured signal, and then removing interference terms according to the energy distribution ratio of the real sound field, rather than subtracting fixed amplitudes. Therefore, it can adapt to the nonlinear energy changes in nonlinear superimposed sound fields and accurately eliminate the actual contribution of wind tunnel background noise. Spectral reconstruction of the residual energy after stripping can correct the signal shape according to the continuous change law of frequency domain energy, eliminating amplitude distortion and frequency distortion caused by nonlinear superposition and interference, and retaining the original aerodynamic noise characteristics generated only by the specimen, thus obtaining true and effective data unaffected by sound field coupling.

[0009] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0010] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the scope of this application. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings: Figure 1 A schematic flowchart of a wind tunnel-based noise detection method provided in an embodiment of this application is shown. Figure 2 This illustration shows a flowchart of a process for stripping the background noise spectrum according to an embodiment of this application; Figure 3 This illustration shows a schematic diagram of a residual energy spectrum reconstruction process provided in an embodiment of this application; Figure 4 A schematic diagram of a wind tunnel-based noise detection device provided in an embodiment of this application is shown. Figure 5 A schematic diagram of the structure of an electronic device provided in an embodiment of this application is shown. Detailed Implementation

[0011] Exemplary embodiments of the present application will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present application are shown in the drawings, it should be understood that the present application may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this application will be thorough and complete, and will fully convey the scope of the present application to those skilled in the art.

[0012] In the embodiments of this application, the terms "first," "second," etc., do not have a logical or temporal dependency, nor do they limit the quantity or execution order. It should also be understood that although the following description uses the terms "first," "second," etc., to describe various elements, these elements should not be limited by the terms. These terms are merely used to distinguish one element from another.

[0013] In this application, the term "at least one" means one or more, and the term "multiple" means two or more.

[0014] It should also be understood that the term “if” can be interpreted as “when” or “upon”, or “in response to determination” or “in response to detection”. Similarly, depending on the context, the phrase “if determination…” or “if detection [the stated condition or event]” can be interpreted as “when determination…” or “in response to determination…” or “when detection [the stated condition or event]” or “in response to detection [the stated condition or event]”.

[0015] In wind tunnel testing, the traditional method for detecting aerodynamic noise in specimens typically involves direct subtraction: first, the overall noise signal after the specimen is placed is collected, then the background noise signal under no-load conditions in the wind tunnel is collected, and the difference between the two sets of data is calculated and directly used as the aerodynamic noise data of the specimen itself. However, in the real wind tunnel flow field and acoustic environment, the noise generated by the specimen under airflow is not independent of the background noise generated by the wind tunnel fan, duct, and airflow disturbance. The two sound sources interact during propagation, reflection, and scattering, forming a complex mixed sound field. The energy distribution, frequency characteristics, and amplitude changes of this sound field all follow a nonlinear superposition law and cannot be represented by a simple linear addition or subtraction relationship. The traditional direct subtraction method, based solely on the assumption of linear superposition, cannot distinguish the true proportion of the two sound sources, cannot correct the energy distortion and frequency shift caused by the interaction of the sound fields, and cannot completely eliminate the interference components of background noise. Therefore, the extracted noise data is distorted, failing to accurately reflect the true acoustic characteristics of the specimen under airflow and making it difficult to guarantee the accuracy and reliability of the test results.

[0016] Based on this, this application proposes a wind tunnel-based noise detection method. The wind tunnel-based noise detection method will be described in detail below with reference to the accompanying drawings. Figure 1 This application provides a flowchart of a wind tunnel-based noise detection method. Specifically, it includes the following steps: Step 110: When testing the test piece in the wind tunnel, collect the measured noise spectrum in the test environment and obtain the operating frequency of the wind tunnel's frequency converter.

[0017] The test specimens applicable to this application cover various products and components that are prone to generating aerodynamic noise under airflow, including building doors and windows, ventilation louvers, curtain wall components, industrial heat dissipation devices, ventilation equipment, cabinet air duct systems, as well as rail transit vehicle components, exposed automotive structural parts, drone fuselages and propellers, and other structural products involving airflow. These test specimens are subjected to airflow scouring and flow field excitation during actual use, thus generating aerodynamic noise. The aerodynamic noise referred to here is the noise generated by the interaction between airflow and the surface of the test specimen, inducing structural vibration or forming turbulent vortices and shedding, typically manifesting as broadband random noise or localized narrowband howling. This type of noise not only affects the comfort of product use and the surrounding acoustic environment but also directly reflects the aerodynamic design defects, structural vibration characteristics, and flow field stability of the test specimen; therefore, it is necessary to test the aerodynamic noise of the test specimens.

[0018] A wind tunnel is a testing device that can artificially generate controllable and stable airflow. Through a drive system, it directs airflow within a closed or semi-closed channel, simulating the airflow conditions experienced by a test specimen in a real-world application scenario. During operation, conventional wind tunnels inevitably generate background noise due to their internal fans, guide structures, airflow turbulence, and drive system. Background noise refers to the background noise generated by the wind tunnel itself during operation in an empty wind field without a test specimen installed; it is also an inherent interference noise component in the testing environment. In actual testing, the signals collected by sensors are a mixture of the wind tunnel's background noise and the specimen's aerodynamic noise, making it impossible to directly distinguish and obtain the true noise characteristics of the specimen itself. Therefore, this application proposes a wind tunnel-based noise detection method.

[0019] When conducting aerodynamic noise testing on a test specimen in a wind tunnel, the acoustic acquisition system is first set up and debugged. Acoustic sensors are deployed in a pre-defined acoustic measurement area outside the wind tunnel test section. The sensor installation positions avoid areas directly impacted by airflow and areas of strong structural vibration, and are equipped with wind shields and vibration-damping bases to reduce the impact of airflow disturbance and structural conducted noise on the acquired signals. After the test begins, the wind tunnel operates to a stable condition, and the test specimen is in the set installation posture and airflow conditions. The acoustic sensors continuously acquire the sound pressure time-domain signal of the test environment according to the preset sampling rate. The acquired signal is a mixed signal formed by the superposition of the aerodynamic noise of the test specimen and the background noise of the wind tunnel. Next, the acquired time-domain signal is preprocessed sequentially by anti-aliasing filtering, DC component removal, segmented windowing, and overlapping averaging to suppress random noise and signal drift. Then, the preprocessed time-domain signal is converted to the frequency domain by fast Fourier transform to obtain the measured noise spectrum based on frequency distribution and characterized by sound pressure level or sound energy density.

[0020] The airflow velocity in a wind tunnel is controlled by a frequency converter that adjusts the speed of the driven fan. The operating frequency of the frequency converter directly determines the fan speed, the wind tunnel load, and the stability of the flow field. The magnitude and spectral distribution of the wind tunnel's inherent noise are strongly correlated with the frequency converter's operating frequency; at the same wind speed, there are significant differences in inherent noise corresponding to different frequency converter operating frequencies. Therefore, this application acquires the operating frequency of the wind tunnel's frequency converter simultaneously with the acquisition of acoustic signals. In one embodiment, the currently stable output operating frequency of the frequency converter is acquired in real-time via the wind tunnel's built-in control system communication interface.

[0021] Step 120: Match the background noise spectrum in the background noise spectrum library corresponding to the operating frequency based on the operating frequency.

[0022] The background noise spectrum library in this application is used to record the background noise information generated by the wind tunnel itself during operation in an empty wind field. The background noise spectrum library stores multiple sets of correspondences, each set containing a frequency converter operating frequency and the background noise spectrum of the wind tunnel at that operating frequency without a test specimen. During the calibration process in the empty wind field, no test specimen is installed in the test section. The wind tunnel operates stably at multiple different frequency converter operating frequencies. The noise signal corresponding to each frequency is acquired using an acoustic acquisition system, and the signal is converted into a spectrum form and stored in association with the frequency, ultimately forming a background noise spectrum library that can be directly accessed.

[0023] Because the background noise generated by a wind tunnel at different operating frequencies varies significantly in both energy magnitude and spectral distribution, only background noise data that matches the current test conditions can accurately reflect the interference components in the test environment. Based on this characteristic, a matching operation is performed in the background noise spectrum library according to the inverter operating frequency obtained in step 110 to determine the wind tunnel background noise spectrum corresponding to this test.

[0024] During the specific matching process, the currently acquired operating frequency is used as an index to search the background noise spectrum library. When there is a record in the background noise spectrum library that matches the operating frequency, the corresponding background noise spectrum is directly retrieved as the wind tunnel background noise spectrum under the current operating conditions.

[0025] In actual testing, there may be situations where the current inverter operating frequency is not included in the background noise spectrum library, making it impossible to directly obtain a matching background noise spectrum. To address this, this application selects multiple calibration frequencies adjacent to the operating frequency from the background noise spectrum library; analyzes whether the energy spectrum of the noise at the operating frequency satisfies a positive correlation function with the multiple calibration frequencies; if so, using the multiple calibration frequencies as independent variables and the acoustic energy density corresponding to the noise spectrum of each frequency band as the dependent variable, and based on the relative position between the operating frequency and the multiple calibration frequencies, performs a linear weighted calculation of the acoustic energy density of each frequency band to obtain the background noise spectrum corresponding to the operating frequency.

[0026] Specifically, since the background noise spectrum library only contains pre-calibrated discrete frequency points, while the actual operating frequency used in the test may not have been pre-calibrated, it is necessary to first select the calibration frequency closest to the current operating frequency as a substitute. During the selection process, calibration frequencies that are numerically adjacent to the current operating frequency are screened from the background noise spectrum library.

[0027] Next, we analyze whether the noise energy spectrum at the current operating frequency satisfies a positive correlation function with multiple calibration frequencies. This positive correlation function means that as the value of the calibration frequency increases, the acoustic energy density of each frequency band in the corresponding noise energy spectrum exhibits a monotonically increasing trend, forming a functional relationship of unidirectional change. Analyzing and verifying this positive correlation function is to confirm that the variation law of the wind tunnel background noise energy spectrum with the inverter's operating frequency is continuous and stable, conforming to the inherent physical characteristics of wind tunnel background noise varying with the wind turbine speed. This ensures the legality and validity of the subsequent linear weighted calculation premise, avoids distortion of the fitted background noise spectrum due to abnormal variation law, and ensures that the obtained background noise spectrum matches the actual operating conditions at the current operating frequency.

[0028] If the two are determined to satisfy a positive correlation function, it indicates that the background noise energy spectrum changes continuously and stably with the frequency converter's operating frequency, and conforms to the inherent physical characteristics of wind tunnel background noise changing with fan speed. This provides a valid premise for obtaining the background noise spectrum at the current operating frequency through linear weighted calculation and fitting. Based on this, linear weighted calculation is performed independently for each frequency band unit after frequency band division. Specifically, multiple selected calibration frequencies are used as independent variables, and the acoustic energy density corresponding to the current frequency band at each calibration frequency is used as the dependent variable. Weights are assigned based on the closeness of the values ​​between the current operating frequency and each calibration frequency. The smaller the difference between the current operating frequency and a certain calibration frequency, the closer the fan speed and airflow conditions are, and the greater the weight corresponding to that calibration frequency. Based on the assigned weights, the acoustic energy density of the current frequency band at each calibration frequency is linearly weighted and summed to obtain the acoustic energy density corresponding to that frequency band at the current operating frequency. After calculating the acoustic energy density of all frequency band units in sequence according to the above method, the calculation results of each frequency band are integrated in order of noise frequency to finally form a complete background noise spectrum that matches the actual operating conditions of the current operating frequency.

[0029] Step 130: Using a frequency band division rule adapted to the inverter's operating frequency, align and divide the measured noise spectrum and the background noise spectrum to obtain multiple frequency band units.

[0030] Wind tunnels rely on frequency converters to drive internal fans, thereby generating the airflow required for testing. The operating frequency of the frequency converter directly determines the actual speed of the fan. When the fan speed changes, the vibration frequency generated by its own rotation, the intensity of the disturbance created by the airflow, and the resonance position of the duct structure all change accordingly, resulting in different frequency distribution characteristics of the wind tunnel's inherent noise floor. As the fan speed changes, the frequency points where the energy is most concentrated in the inherent noise floor and the most representative main noise frequency also shift significantly. When the frequency converter operates at a lower frequency, the fan speed is slower, mainly generating low-frequency vibrations and low-frequency airflow noise, with noise energy concentrated in a lower and relatively wide frequency range. When the frequency converter operates at a higher frequency, the fan speed increases, the frequency range covered by vibration and airflow disturbances is wider, the high-frequency noise component increases, and the noise energy distribution is more dispersed, requiring more detailed frequency band division for accurate characterization. If a fixed frequency band division method is used, when the inverter's operating frequency changes, the energy distribution of the background noise and the position of the main frequency will shift, but the frequency band interval will not be adjusted accordingly. This will cause the energy characteristics of the background noise and the measured noise to not fall into the corresponding frequency band range, resulting in a disordered frequency band correspondence. Consequently, subsequent noise separation processing will be biased and will not be able to truly reflect the noise distribution characteristics under the current operating conditions.

[0031] After determining the wind tunnel's background noise spectrum to match the current operating conditions, to avoid inaccurate frequency band correspondence due to changes in noise energy distribution with operating frequency, this application employs a frequency band division rule adapted to the inverter's operating frequency obtained in step 110 to align and divide the measured noise spectrum and the background noise spectrum. This frequency band division rule is based on the correspondence between wind turbine speed and the dominant noise frequency, dividing the operating frequency into three intervals: low frequency, mid frequency, and high frequency. Different intervals correspond to different frequency band division strategies, ensuring that the frequency band division results match the noise distribution characteristics under the current operating conditions. When determining the frequency band range, the theoretical dominant frequency of the wind tunnel at the current operating frequency is used as the center, expanding outwards to obtain the effective frequency range. This range is then used as the starting and ending frequencies for this division, ensuring that all effective noise energy is included within the division range. The bandwidth is adaptively determined based on the operating frequency, with a wider bandwidth used in the low-frequency interval to match the concentrated energy distribution characteristics. A narrower bandwidth is used in the high-frequency range to adapt to the dispersed energy distribution. This method dynamically determines the width of individual bandwidths, ensuring the bandwidth division matches the actual noise distribution under current operating conditions. After determining the appropriate bandwidth division rules, the measured noise spectrum and the background noise spectrum are synchronously segmented using the same method. During segmentation, the starting and ending frequencies and bandwidths of the two sets of spectra are kept consistent, ensuring alignment in the frequency dimension. This synchronous alignment ensures that each bandwidth of the measured noise matches its corresponding bandwidth in subsequent processing, avoiding characteristic comparison deviations caused by frequency peak shifts or bandwidth misalignments, thus guaranteeing the accuracy and reliability of subsequent noise separation calculations. After synchronous alignment, the two sets of spectra are divided into multiple bandwidth units of equal number and corresponding frequency ranges. Each bandwidth unit contains an independent frequency range and corresponding noise energy information.

[0032] After aligning and dividing the measured noise spectrum and the background noise spectrum, the energy change trends of the measured noise and the background noise are highly similar in some frequency bands, and the effective component of the aerodynamic noise of the specimen itself accounts for a very low proportion. If such frequency bands are directly used for subsequent noise separation and energy calculation, a large amount of interference information will be introduced, reducing the noise separation accuracy and weakening the reliability of the detection results. Therefore, this application identifies and filters the effectiveness of all frequency bands based on the similarity characteristics of the acoustic energy density gradient, retaining only the frequency bands containing the effective aerodynamic noise component of the specimen for subsequent processing. Specifically, the acoustic energy density gradient along the frequency dimension of the measured noise spectrum and the background noise spectrum in each frequency band is calculated; the similarity between the measured noise acoustic energy density gradient and the background noise acoustic energy density gradient in the same frequency band is calculated to obtain the gradient matching degree; and frequency bands with a gradient matching degree higher than a preset matching threshold are removed from multiple frequency bands.

[0033] Specifically, the acoustic energy density gradient is used to characterize the rate and trend of change of noise energy amplitude with frequency within a frequency band. It can intuitively reflect the fluctuation characteristics of noise energy within the same frequency band and effectively distinguish the distribution differences between aerodynamic noise of the test specimen and background noise of the wind tunnel. Based on this characteristic, this application first traverses all aligned and divided frequency bands. For each independent frequency band, the energy amplitude data of the measured noise spectrum and the background noise spectrum within that frequency band are extracted. With frequency as the independent variable and acoustic energy density as the dependent variable, the acoustic energy density gradient of the measured noise spectrum along the frequency dimension within each frequency band and the acoustic energy density gradient of the background noise spectrum along the frequency dimension within the same frequency band are calculated.

[0034] When calculating the acoustic energy density gradient, for each frequency band, a set of discrete frequency points corresponding to the measured noise spectrum within that frequency band and the corresponding acoustic energy density values ​​are first extracted. Then, the gradient is calculated along the frequency dimension using a finite difference method. Using adjacent frequency points as calculation units, the acoustic energy density of the previous frequency point is subtracted from the acoustic energy density of the subsequent frequency point, and then divided by the frequency interval between the two frequency points to obtain the rate of change of acoustic energy density within that interval. After traversing all frequency points within the frequency band and completing the calculation, the rate of change of all intervals is averaged to finally obtain the acoustic energy density gradient along the frequency dimension of the measured noise spectrum within that frequency band. Using the same calculation method, the discrete frequency points and acoustic energy density values ​​of the background noise spectrum within the same frequency band are processed to obtain the acoustic energy density gradient along the frequency dimension of the background noise spectrum within the corresponding frequency band.

[0035] After obtaining two sets of acoustic energy density gradient data within the same frequency band, the similarity between the two sets of gradient data is calculated to obtain the gradient matching degree. This application uses a cosine similarity algorithm to measure the similarity of the changing trends of the two sets of gradients. The acoustic energy density gradient of the measured noise is used as the first vector, and the acoustic energy density gradient of the background noise is used as the second vector. The cosine value of the angle between the two vectors is calculated. The closer the cosine value is to 1, the more consistent the changing trends of the two sets of gradients are; the closer the cosine value is to 0, the more obvious the difference in their changing trends are. The result of this cosine similarity calculation is the gradient matching degree corresponding to the current frequency band. It should be noted that the above cosine similarity algorithm is only an exemplary illustration of this application and does not constitute a specific limitation of this application. In other embodiments, for the similarity calculation of two sets of acoustic energy density gradient data within the same frequency band, correlation coefficient, Euclidean distance, or other conventional similarity measurement methods in the art can also be used to obtain the corresponding gradient matching degree.

[0036] After calculating the gradient matching degree for each frequency band unit, it is compared with a preset matching threshold, which is the critical value for distinguishing between effective and ineffective frequency bands. If the gradient matching degree of a frequency band unit is higher than the preset matching threshold, it indicates that the measured noise in that frequency band is mainly composed of wind tunnel background noise, and the aerodynamic noise component of the specimen cannot be effectively separated and identified, thus lacking subsequent analysis value. In this case, frequency band units with gradient matching degrees higher than the preset matching threshold are removed from all frequency band units, and only effective frequency band units with gradient matching degrees lower than or equal to the preset matching threshold are retained, thereby eliminating interference information and improving data validity.

[0037] Step 140: Based on the acoustic energy density of the measured noise spectrum and the background noise spectrum in each frequency band unit, the energy component corresponding to the background noise spectrum is extracted from the measured noise spectrum by directional orthogonal projection, and the spectrum of the residual energy after extraction is reconstructed to obtain the effective aerodynamic noise spectrum of the test piece.

[0038] After screening the effective frequency band units, in order to separate the aerodynamic noise component generated solely by the test device from the mixed measured noise, this application performs energy fitting, spatial compensation, orthogonal projection stripping, and spectrum reconstruction on the retained effective frequency band units. These processes eliminate energy attenuation errors caused by spatial propagation and strip the energy component corresponding to the background noise from the measured noise, ultimately restoring the true and continuous effective aerodynamic noise spectrum of the test device. The specific processing steps are as follows.

[0039] Step 210: Perform frequency domain energy fitting on the measured noise spectrum and the background noise spectrum in each frequency band unit to obtain the spectral energy sequence corresponding to the measured noise and the spectral energy sequence corresponding to the background noise in each frequency band unit.

[0040] The original spectrum consists of energy values ​​at discrete frequency points. The data distribution is scattered and lacks continuity. Therefore, this application uses frequency domain energy fitting to smooth the energy distribution within the frequency band, forming a complete and unified energy distribution form.

[0041] In practice, for each selected and retained frequency band, the acoustic energy density amplitude data corresponding to each discrete frequency point of the measured noise spectrum within that frequency band, as well as the frequency dimension energy distribution data corresponding to each time sampling point within the sampling period, are first obtained. Based on the center frequency of the frequency band and the frequency range within that band, a kernel function method is used for time-frequency energy fitting. The fitting uses time and frequency as the two independent variables and acoustic energy density amplitude as the dependent variable. Iterative optimization is used to minimize the error between the fitting result and the acoustic energy density data of the original measured spectrum. After completing the model training, the fitting model is applied to the full time sampling interval and full frequency range of the current frequency band to obtain a continuous and smooth spectral energy sequence corresponding to the measured noise within each frequency band. This sequence completely covers the noise energy information corresponding to all time points and frequency intervals within the current frequency band.

[0042] Using the same fitting model construction method and parameter configuration, frequency domain energy fitting processing is performed on the background noise spectrum within the same frequency band unit to obtain the spectral energy sequence corresponding to the background noise in each frequency band unit. By using a unified fitting model and parameter configuration, the consistency of the measured noise and background noise energy sequences in the time and frequency dimensions and the consistency of their distribution characteristics are ensured, avoiding deviations in subsequent energy stripping processing due to differences in fitting rules.

[0043] Step 220: Apply spatial propagation weights to the spectral energy sequence within each frequency band unit based on the sound wave propagation path of the detection environment and the orientation information of the sensor array.

[0044] During the propagation of sound waves from the sound source to the acoustic sensor, energy attenuation occurs due to factors such as propagation distance, air absorption, and incident angle. This results in a difference between the sound energy density amplitude collected by the sensor and the actual sound energy density amplitude radiated by the sound source. Furthermore, the spatial propagation attenuation characteristics of the measured noise and the background noise within the same frequency band remain consistent. To correct the energy deviation caused by spatial propagation and to unify the energy benchmark between the measured noise and the background noise, this application applies a spatial propagation weight to the spectral energy sequence within each frequency band.

[0045] In one implementation, a spatial propagation gain model is constructed for each frequency band unit based on the propagation distance of the sound wave in the detection environment, the air attenuation coefficient, and the azimuth and elevation angles of the sensor array relative to the sound source. Based on the spatial propagation gain model, corresponding spatial propagation weights are applied to the measured noise spectrum energy sequence and the background noise spectrum energy sequence in each frequency band unit.

[0046] Specifically, in the detection environment, sound waves will experience energy attenuation as they propagate from the sound source to the acoustic sensor. In order to restore the true radiated energy of the sound source, this application compensates for the attenuation effect.

[0047] In actual calculations, the actual propagation distance of the sound wave from the sound source to the sensor is first determined, and the airborne sound attenuation coefficient corresponding to the center frequency of each frequency band unit is obtained. Simultaneously, the azimuth and elevation angles of the sensor array relative to the sound source are measured. The propagation distance is used to calculate the spherical spread attenuation, i.e., the sound energy density decreases inversely with the square of the distance traveled. The airborne sound attenuation coefficient is used to calculate the absorption loss of sound waves by the air at different frequencies; the higher the frequency, the more significant the attenuation. The azimuth and elevation angles are used to correct for the influence of different incident directions on the sensor response.

[0048] After obtaining the three types of parameters, the attenuation compensation for each type is calculated. For spherical extension compensation, the square of the ratio of the fixed reference distance to the actual propagation distance is used; the larger the distance, the larger the compensation. For air absorption compensation, it is obtained by multiplying the center frequency of the frequency band, the airborne sound attenuation coefficient, and the propagation distance; the higher the frequency, the larger the attenuation coefficient, and the farther the propagation distance, the larger the corresponding compensation. For angle correction compensation, based on the sensor's directional response characteristics, the azimuth and elevation angles are substituted into the directional response curve to obtain the acquisition correction coefficient at the corresponding angle. The three types of attenuation compensation are linearly superimposed to obtain a comprehensive attenuation compensation factor used to restore the true energy of the sound source. This comprehensive attenuation compensation factor is the spatial propagation gain corresponding to the current frequency band unit. Following the same calculation steps, all the selected and retained frequency band units are processed sequentially to obtain the spatial propagation gain corresponding to each frequency band unit, ultimately forming a spatial propagation weight sequence corresponding to each frequency band unit.

[0049] After obtaining the spatial propagation weight sequence, the spatial propagation gain corresponding to each frequency band unit is multiplied by the energy amplitude of each frame of the measured noise spectrum energy sequence in that frequency band. Then, the same calculation method is used to weight the background noise spectrum energy sequence so that the measured noise and the background noise maintain a unified energy benchmark under the influence of propagation attenuation, thereby restoring the energy distribution that conforms to the true radiation state.

[0050] Step 230: Based on the background noise spectrum energy sequence after applying spatial propagation weights, calculate the energy projection component of the background noise spectrum energy sequence on the measured noise spectrum energy sequence, and subtract the energy projection component from the measured noise spectrum energy sequence to remove the energy component corresponding to the background noise spectrum from the measured noise spectrum.

[0051] After applying the same spatial propagation weight to both the measured noise and the background noise, the spectral energy sequences of the two types of noise were normalized under the same spatial propagation reference, and their energy amplitude distribution accurately corresponded to the radiated energy of the actual sound source. At this point, the measured noise spectral energy sequence simultaneously contained the aerodynamic noise component generated by the test specimen and the background noise component generated by the wind tunnel environment. The background noise, as a non-target interference component, needed to be completely removed from the mixed signal, retaining only the effective noise energy corresponding to the test specimen.

[0052] In practice, the background noise spectrum energy sequence is used as the reference vector, and the measured noise spectrum energy sequence is used as the vector to be decomposed. The two vectors have one-to-one corresponding energy amplitude points at the same frequency dimension. During the projection calculation, the inner product of the background noise spectrum energy sequence is first calculated, which involves squaring the energy amplitudes corresponding to all frequency points in the sequence and summing the squares to obtain the squared magnitude of the background noise vector. This value characterizes the overall energy intensity of the background noise energy sequence. Then, the dot product between the measured noise spectrum energy sequence and the background noise spectrum energy sequence is calculated. The energy amplitudes of the two sequences at the same frequency points are multiplied one by one and summed to obtain the inner product value characterizing the correlation between their energy distributions.

[0053] Dividing the above dot product value by the square of the magnitude of the background noise vector yields the projection coefficient corresponding to the current frequency band cell. The projection coefficient objectively reflects the proportion of energy of the background noise in the measured noise. The larger the projection coefficient, the higher the contribution of the background noise to the measured noise; the smaller the projection coefficient, the stronger the dominance of the specimen's aerodynamic noise in the measured noise. Multiplying the obtained projection coefficient by each energy amplitude in the background noise spectrum energy sequence yields the projected energy component. This projected energy component conforms to the energy variation law of the background noise and represents the energy contributed by the wind tunnel background noise in the measured noise; it is also the target component that needs to be separated from the measured noise.

[0054] Finally, the measured noise spectrum energy sequence is subtracted from the projected energy components within the corresponding frequency band, and the amplitude subtraction is performed point by point to obtain the stripped residual energy sequence. This residual energy sequence eliminates all energy components related to the background noise, removes the interference effects of wind tunnel operation, and retains only the aerodynamic noise energy generated by the test specimen under airflow.

[0055] After stripping away the background noise components, each effective frequency band unit retains only the residual energy sequence corresponding to the test specimen. These sequences are independent and segmented discretely distributed in the frequency dimension, failing to directly reflect the complete spectral shape of the specimen's aerodynamic noise. To obtain continuous, smooth noise data that closely matches the actual radiation characteristics, the discrete residual energy sequences need to be spliced, smoothed, and converted in format. The specific process is as follows.

[0056] Step 310: Based on the center frequency and amplitude distribution of each frequency band unit, the residual spectral energy sequence after removing the background noise is connected to form a continuous spectral energy sequence.

[0057] The residual spectral energy sequence after stripping the background noise exists independently in each of the selected frequency band units. Different frequency band units are separated from each other on the frequency axis. Direct combination will result in frequency discontinuities and amplitude abrupt changes, and cannot form a continuous and complete noise spectrum.

[0058] During the transition process, all effective frequency bands are first arranged in ascending order of center frequency based on their own center frequencies to determine the overall frequency distribution order of the spectrum. For each pair of adjacent frequency bands, the highest frequency of the preceding band and the lowest frequency of the following band are determined. The common boundary between these two frequency points is used as the transition point. Simultaneously, the energy amplitude of the preceding band at its highest frequency and the energy amplitude of the following band at its lowest frequency are extracted. Using the transition point as the center, a weighted average of the two amplitudes is calculated. The calculated average amplitude is used as the unified amplitude at the transition point, replacing the original independent amplitudes of the two bands at the boundary, thus ensuring a smooth transition of amplitude between adjacent bands at the transition point.

[0059] Following the above method, all adjacent frequency band units are traversed sequentially and the connection process is completed. The originally segmented and independent residual spectral energy sequences are connected into a whole, and finally a complete spectral energy sequence covering the entire effective frequency range and with continuous frequency is formed.

[0060] Step 320: Based on the air attenuation coefficient and sound wave propagation distance of the test environment, perform amplitude smoothing on the continuous spectrum energy sequence to form the effective aerodynamic noise spectrum of the test piece.

[0061] After completing the frequency band connection, a continuous spectral energy sequence is obtained. Due to the influence of the previous frequency band division, amplitude compensation, and propagation attenuation correction, the sequence still exhibits local amplitude fluctuations and small jumps between adjacent frequency points. These fluctuations are not characteristic of the actual aerodynamic noise of the specimen, but rather non-stationary interference introduced during data processing, which directly leads to distortion of the subsequently generated noise signal. Simultaneously, the attenuation effect of sound waves propagating through the air exhibits continuous variation characteristics across the entire frequency band. Single-band independent compensation alone cannot guarantee the uniformity of energy variation across the entire frequency band. To eliminate local amplitude fluctuations, restore the smooth and natural energy variation law of the specimen's aerodynamic noise, and make the spectral curve more closely resemble the actual physical radiation characteristics, this application performs amplitude smoothing processing on the continuous spectral energy sequence.

[0062] In one implementation, the sound wave attenuation amount corresponding to each frequency band unit is calculated based on the air attenuation coefficient and the sound wave propagation distance; the spectral energy of each frequency band unit is reverse-compensated based on the sound wave attenuation amount to obtain the amplitude compensation coefficient corresponding to each frequency band unit; based on the difference between the amplitude compensation coefficients of adjacent frequency band units, the continuous spectral energy sequence is segmented and constrained to smooth it, so that the amplitude change between adjacent frequency band units satisfies the preset gradient constraint.

[0063] Specifically, firstly, based on the center frequency of each frequency band, the corresponding air attenuation coefficient, and a fixed sound wave propagation distance, the sound wave attenuation calculation formula is substituted to calculate the spherical spread attenuation and air absorption attenuation respectively. These are then summed to obtain the total sound wave attenuation generated by each frequency band unit in the propagation path. This attenuation is used to perform reverse compensation on the spectral energy of the corresponding frequency band, replenishing the energy lost during propagation to the original amplitude level, thus restoring the energy amplitude to the true radiation state of the sound source, thereby obtaining the amplitude compensation coefficient corresponding to each frequency band unit. Subsequently, the difference between the amplitude compensation coefficients of two adjacent frequency band units is calculated to determine the amplitude variation amplitude between frequency bands. When the difference exceeds the stable range, the spectral amplitude in the connecting area is constrained and smoothed. The amplitude point is adjusted through a linear transition method to control the amplitude variation between adjacent frequency bands within a preset gradient range, avoiding abrupt changes and abnormal jitter. Finally, the effective aerodynamic noise spectrum of the test device is obtained.

[0064] Step 150: Calculate the noise detection results of the test piece based on the effective aerodynamic noise spectrum.

[0065] After obtaining a smooth and continuous effective aerodynamic noise spectrum that has been freed from wind tunnel background noise interference, this data fully reflects the noise radiation characteristics of the specimen itself. Standardized noise detection results can be generated using conventional statistical and computational methods in the acoustic field. Specifically, the sound pressure level time-series signal across the entire frequency band of the effective aerodynamic noise spectrum is extracted first. Signal data within a stable sampling period is selected as the calculation object, eliminating interference from non-stationary segments at the beginning and end of the signal. Time-domain statistics are performed on the sound pressure level time-series signal within the effective period to calculate the total equivalent continuous sound pressure level, which characterizes the average noise intensity of the specimen throughout the entire testing period. Simultaneously, the sound pressure level data for each frequency band are integrated in the frequency domain to obtain the total sound level across the entire frequency band, directly reflecting the overall radiation level of the specimen's aerodynamic noise.

[0066] The peak frequencies and corresponding peak sound pressure levels in the spectrum are further extracted to determine the main contributing frequencies and maximum noise intensity of the aerodynamic noise of the specimen. Combined with the preset detection and evaluation standards, the calculated key parameters such as total sound level, peak sound pressure level, and peak frequency are organized to form a complete detection result that includes noise intensity, frequency characteristics, and time-domain stability.

[0067] In summary, this application determines the background noise spectrum consistent with the operating conditions by matching the inverter's operating frequency, ensuring a precise correspondence between the background noise benchmark and the actual test conditions, and avoiding fundamental errors caused by background noise mismatch. By employing rules adapted to the operating frequency to align the measured noise and background noise into frequency bands, the two sets of spectra can be strictly correlated in the frequency domain. Based on the acoustic energy density of each frequency band, directional orthogonal projection is used to strip away the background noise component, enabling the calculation and removal of interference components according to the true energy distribution ratio. This overcomes the shortcomings of traditional direct subtraction, which cannot adapt to nonlinear superimposed sound fields, and completely separates the true noise components of the specimen from the mixed signal. Spectrum reconstruction processing of the residual energy corrects amplitude distortion and frequency discontinuities in discrete frequency bands, restoring a continuous, smooth, and effective aerodynamic noise spectrum unaffected by sound field coupling interference. Finally, the detection results are calculated based on clean and effective data, improving the accuracy of noise extraction and the reliability of detection. This accurately reflects the actual aerodynamic noise level of the test specimen under wind tunnel conditions, solving the core problems of detection distortion and inaccurate results in traditional methods.

[0068] It is understood that, in order to achieve the functions in the above embodiments, the computer device includes hardware structures and / or software modules corresponding to the execution of each function. Those skilled in the art should readily recognize that, based on the units and method steps described in conjunction with the embodiments disclosed in this application, this application can be implemented in hardware or a combination of hardware and computer software. Whether a function is executed by hardware or by computer software driving hardware depends on the specific application scenario and design constraints of the technical solution.

[0069] Furthermore, as a response to the above Figure 1 The implementation of the method embodiment shown in this application provides a wind tunnel-based noise detection device. The embodiment of this device corresponds to the foregoing method embodiments. For ease of reading, this embodiment will not repeat the details of the foregoing method embodiments one by one, but it should be clear that the device in this embodiment can correspondingly implement all the contents of the foregoing method embodiments. Specifically, as shown... Figure 4 As shown, the wind tunnel-based noise detection device 400 includes: The acquisition module 410 is used to acquire the measured noise spectrum in the test environment and obtain the operating frequency of the wind tunnel's frequency converter when the test piece is being tested in the wind tunnel. The determination module 420 is used to match the background noise spectrum corresponding to the operating frequency in the background noise spectrum library based on the operating frequency. The background noise spectrum library includes the correspondence between the operating frequency of the wind tunnel inverter and the background noise spectrum in an empty wind field without the test piece. The partitioning module 430 is used to align and partition the measured noise spectrum and the background noise spectrum using a frequency band partitioning rule adapted to the inverter's operating frequency, thereby obtaining multiple frequency band units; The stripping module 440 is used to strip the energy component corresponding to the background noise spectrum from the measured noise spectrum by directional orthogonal projection based on the acoustic energy density of the measured noise spectrum and the background noise spectrum in each frequency band unit. The reconstruction module 450 is used to reconstruct the spectrum of the residual energy after stripping to obtain the effective aerodynamic noise spectrum of the test piece. The detection module 460 is used to calculate the noise detection results of the test piece based on the effective aerodynamic noise spectrum.

[0070] Furthermore, such as Figure 4 As shown, the stripping module 440 is specifically used to perform frequency domain energy fitting on the measured noise spectrum and the background noise spectrum in each frequency band unit to obtain the spectral energy sequence corresponding to the measured noise and the spectral energy sequence corresponding to the background noise in each frequency band unit; according to the sound wave propagation path and sensor array orientation information of the detection environment, a spatial propagation weight is applied to the spectral energy sequence in each frequency band unit; based on the background noise spectral energy sequence after applying the spatial propagation weight, the energy projection component of the background noise spectral energy sequence on the measured noise spectral energy sequence is calculated, and the energy projection component is subtracted from the measured noise spectral energy sequence to strip the energy component corresponding to the background noise spectrum from the measured noise spectrum.

[0071] Furthermore, such as Figure 4 As shown, the stripping module 440 is also used to determine the preset trend of amplitude change of each frequency band unit with the propagation distance and the corresponding allowable deviation range based on the propagation distance and propagation attenuation law of the sound wave in the detection environment; compare the amplitude change of the residual noise spectrum energy sequence in each frequency band unit with the preset trend; when the amplitude change of a certain frequency band unit exceeds the allowable deviation range, the frequency band unit is removed to obtain the verified residual noise spectrum energy sequence.

[0072] Furthermore, such as Figure 4 As shown, the reconstruction module 450 is specifically used to connect the residual spectral energy sequence after stripping the background noise based on the center frequency and amplitude distribution of each frequency band unit to form a continuous spectral energy sequence; and to perform amplitude smoothing on the continuous spectral energy sequence according to the air attenuation coefficient and sound wave propagation distance of the detection environment to form the effective aerodynamic noise spectrum of the test piece.

[0073] Furthermore, such as Figure 4As shown, the reconstruction module 450 is specifically used to calculate the sound wave attenuation corresponding to each frequency band unit according to the air attenuation coefficient and the sound wave propagation distance; to perform reverse compensation on the spectral energy of each frequency band unit based on the sound wave attenuation to obtain the amplitude compensation coefficient corresponding to each frequency band unit; and to perform segmented constraint smoothing on the continuous spectral energy sequence based on the difference between the amplitude compensation coefficients of adjacent frequency band units so that the amplitude change between adjacent frequency band units meets the preset gradient constraint.

[0074] Furthermore, such as Figure 4 As shown, the partitioning module 430 is also used to calculate the acoustic energy density gradient along the frequency dimension of the measured noise spectrum and the background noise spectrum in each frequency band unit after obtaining multiple frequency band units; to calculate the similarity between the measured noise acoustic energy density gradient and the background noise acoustic energy density gradient in the same frequency band unit to obtain the gradient matching degree; and to remove the frequency band units with a gradient matching degree higher than the preset matching threshold from the multiple frequency band units.

[0075] Furthermore, such as Figure 4 As shown, the determination module 420 is also used to select multiple calibration frequencies adjacent to the operating frequency from the background noise spectrum library; analyze whether the energy spectrum of the noise at the operating frequency and the multiple calibration frequencies satisfy a positive correlation function relationship; if so, with the multiple calibration frequencies as independent variables and the sound energy density corresponding to the noise spectrum of each frequency band as dependent variables, and according to the relative position between the operating frequency and the multiple calibration frequencies, the sound energy density of each frequency band is linearly weighted and calculated to obtain the background noise spectrum corresponding to the operating frequency.

[0076] Optionally, the wind tunnel-based noise detection device may be an electronic device with data processing capabilities, or a functional module within such electronic device, without limitation.

[0077] For example, the electronic device can be a server, which can be a single server or a server cluster consisting of multiple servers. As another example, the electronic device can be a mobile phone, tablet computer, desktop computer, laptop computer, handheld computer, notebook computer, ultra-mobile personal computer (UMPC), netbook, as well as cellular phones, personal digital assistants (PDAs), augmented reality (AR) devices, virtual reality (VR) devices, and other terminal devices. As yet another example, the electronic device can also be a recording device, video surveillance equipment, etc. This application does not impose any special limitations on the specific form of the electronic device.

[0078] The following example uses a wind tunnel-based noise detection device that is an electronic device. Figure 5 As shown, Figure 5 The hardware structure of an electronic device 500 provided in this application.

[0079] like Figure 5 As shown, the electronic device 500 includes a processor 510, a communication line 520, and a communication interface 530.

[0080] Optionally, the electronic device 500 may also include a memory 540. The processor 510, memory 540, and communication interface 530 can be connected via a communication line 520.

[0081] The processor 510 can be a central processing unit (CPU), a general-purpose processor, a network processor (NP), a digital signal processor (DSP), a microprocessor, a microcontroller, a programmable logic device (PLD), or any combination thereof. The processor 510 can also be any other device with processing capabilities, such as a circuit, device, or software module, without limitation.

[0082] In one example, processor 510 may include one or more CPUs, for example Figure 5 CPU0 and CPU1 in the CPU.

[0083] As an optional implementation, the electronic device 500 may include multiple processors; for example, in addition to processor 510, it may also include processor 570. A communication line 520 is used to transmit information between the components included in the electronic device 500.

[0084] Communication interface 530 is used for communication with other devices or other communication networks. These other communication networks can be Ethernet, Radio Access Network (RAN), Wireless Local Area Networks (WLAN), etc. Communication interface 530 can be a module, circuit, transceiver, or any device capable of enabling communication.

[0085] Memory 540 is used to store instructions. These instructions can be computer programs.

[0086] The memory 540 may be a read-only memory (ROM) or other type of static storage device capable of storing static information and / or instructions; it may also be a random access memory (RAM) or other type of dynamic storage device capable of storing information and / or instructions; it may also be an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage media, or other magnetic storage devices, etc., without limitation.

[0087] It should be noted that the memory 540 can exist independently of the processor 510, or it can be integrated with the processor 510. The memory 540 can be used to store instructions, program code, or some data, etc. The memory 540 can be located inside or outside the electronic device 500, without restriction.

[0088] The processor 510 is configured to execute instructions stored in the memory 540 to implement the communication method provided in the following embodiments of this application. For example, when the electronic device 500 is a terminal or a chip in a terminal, the processor 510 can execute instructions stored in the memory 540 to implement the steps performed by the sending end in the following embodiments of this application.

[0089] As an optional implementation, the electronic device 500 also includes an output device 550 and an input device 560. The output device 550 can be a display screen, speaker, or other device capable of outputting data from the electronic device 500 to the user. The input device 560 can be a keyboard, mouse, microphone, joystick, or other device capable of inputting data into the electronic device 500.

[0090] It should be pointed out that, Figure 5 The structure shown does not constitute a limitation on the electronic device, except... Figure 5 In addition to the components shown, the electronic device may include more or fewer components than illustrated, or combine certain components, or have different component arrangements.

[0091] The wind tunnel-based noise detection device and application scenarios described in this application are intended to more clearly illustrate the technical solutions of this application and do not constitute a limitation on the technical solutions provided in this application. As those skilled in the art will know, with the evolution of wind tunnel-based noise detection devices and the emergence of new business scenarios, the technical solutions provided in this application are also applicable to similar technical problems.

[0092] This application provides a storage medium storing a program that, when executed by a processor, implements the wind tunnel-based noise detection method.

[0093] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0094] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0095] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A wind tunnel-based noise detection method, characterized in that, The method includes: When testing the test piece in a wind tunnel, the measured noise spectrum in the testing environment is collected, and the operating frequency of the frequency converter in the wind tunnel is obtained. Based on the operating frequency, the background noise spectrum corresponding to the operating frequency in the background noise spectrum library is matched. The background noise spectrum library includes the correspondence between the operating frequency of the wind tunnel inverter and the background noise spectrum in an empty wind field without test specimens. By adopting a frequency band division rule adapted to the operating frequency of the frequency converter, the measured noise spectrum and the background noise spectrum are aligned and divided to obtain multiple frequency band units; Based on the acoustic energy density of the measured noise spectrum and the background noise spectrum in each frequency band unit, the energy component corresponding to the background noise spectrum is stripped from the measured noise spectrum by directional orthogonal projection, and the spectrum of the residual energy after stripping is reconstructed to obtain the effective aerodynamic noise spectrum of the test piece. The noise detection result of the test piece is obtained by calculating based on the effective aerodynamic noise spectrum.

2. The method according to claim 1, characterized in that, Based on the acoustic energy density of the measured noise spectrum and the background noise spectrum within each frequency band unit, the energy component corresponding to the background noise spectrum is extracted from the measured noise spectrum through directional orthogonal projection, including: Frequency domain energy fitting is performed on the measured noise spectrum and the background noise spectrum in each frequency band unit to obtain the spectral energy sequence corresponding to the measured noise and the spectral energy sequence corresponding to the background noise in each frequency band unit. Based on the sound wave propagation path and sensor array orientation information of the detection environment, spatial propagation weights are applied to the spectral energy sequence within each frequency band unit; Based on the background noise spectrum energy sequence after applying spatial propagation weights, the energy projection component of the background noise spectrum energy sequence on the measured noise spectrum energy sequence is calculated, and the energy projection component is subtracted from the measured noise spectrum energy sequence to remove the energy component corresponding to the background noise spectrum from the measured noise spectrum.

3. The method according to claim 2, characterized in that, Before performing spectral reconstruction on the residual energy after stripping, the method further includes: Based on the propagation distance and propagation attenuation law of sound waves in the detection environment, the preset variation trend of amplitude of each frequency band unit with propagation distance and the corresponding allowable deviation range are determined. The amplitude variation of the residual noise spectrum energy sequence in each frequency band unit is compared with the preset variation trend; When the amplitude change of a certain frequency band unit exceeds the allowable deviation range, the frequency band unit is removed to obtain the verified residual noise spectrum energy sequence.

4. The method according to claim 2, characterized in that, The residual energy after stripping is reconstructed to obtain the effective aerodynamic noise spectrum of the test piece, including: Based on the center frequency and amplitude distribution of each frequency band unit, the residual spectral energy sequence after removing the background noise is connected to form a continuous spectral energy sequence. Based on the air attenuation coefficient and sound wave propagation distance of the detection environment, the amplitude smoothing process is applied to the continuous spectrum energy sequence to form the effective aerodynamic noise spectrum of the test piece.

5. The method according to claim 4, characterized in that, Based on the air attenuation coefficient and sound wave propagation distance of the detection environment, amplitude smoothing processing is performed on the continuous spectrum energy sequence, including: Calculate the sound wave attenuation for each frequency band unit based on the air attenuation coefficient and the sound wave propagation distance. Based on the sound wave attenuation, the spectral energy of each frequency band unit is reversed to obtain the amplitude compensation coefficient corresponding to each frequency band unit. Based on the difference between the amplitude compensation coefficients of adjacent frequency band units, the continuous spectrum energy sequence is segmented and constrained to smooth it, so that the amplitude variation between adjacent frequency band units meets the preset gradient constraint.

6. The method according to claim 1, characterized in that, After obtaining multiple frequency band units, the method further includes: Calculate the acoustic energy density gradient along the frequency dimension of the measured noise spectrum and the background noise spectrum in each frequency band unit; The similarity between the measured noise acoustic energy density gradient and the background noise acoustic energy density gradient within the same frequency band unit is calculated to obtain the gradient matching degree. Frequency band units with a gradient matching degree higher than a preset matching threshold are removed from the plurality of frequency band units.

7. The method according to any one of claims 1-6, characterized in that, When the background noise spectrum library does not contain a background noise spectrum corresponding to the operating frequency, the method further includes: Select multiple calibration frequencies adjacent to the operating frequency from the background noise spectrum library; Analyze whether the energy spectrum of the noise at the operating frequency satisfies a positive correlation function with the multiple calibration frequencies; If present, the acoustic energy density of each frequency band is linearly weighted and calculated based on the relative position between the operating frequency and the multiple calibration frequencies, using the multiple calibration frequencies as independent variables and the acoustic energy density corresponding to the noise spectrum of each frequency band as dependent variables, to obtain the background noise spectrum corresponding to the operating frequency.

8. A wind tunnel-based noise detection device, characterized in that, The device includes: The acquisition module is used to acquire the measured noise spectrum in the test environment and obtain the operating frequency of the frequency converter of the wind tunnel when the test piece is tested through the wind tunnel. The determination module is used to match the background noise spectrum corresponding to the operating frequency in the background noise spectrum library based on the operating frequency. The background noise spectrum library includes the correspondence between the operating frequency of the frequency converter of the wind tunnel and the background noise spectrum in an empty wind field without test specimens. The partitioning module is used to align and partition the measured noise spectrum and the background noise spectrum using a frequency band partitioning rule adapted to the operating frequency of the frequency converter, thereby obtaining multiple frequency band units; The stripping module is used to strip the energy component corresponding to the background noise spectrum from the measured noise spectrum by means of directional orthogonal projection, based on the acoustic energy density of the measured noise spectrum and the background noise spectrum in each frequency band unit. The reconstruction module is used to reconstruct the spectrum of the residual energy after stripping to obtain the effective aerodynamic noise spectrum of the test piece. The detection module is used to calculate the noise detection result of the test piece based on the effective aerodynamic noise spectrum.

9. A storage medium, characterized in that, The storage medium includes a stored program, wherein, when the program is executed, it controls the device where the storage medium is located to perform the wind tunnel-based noise detection method as described in any one of claims 1-7.

10. An electronic device, characterized in that, The device includes at least one processor, at least one memory connected to the processor, and a bus; wherein the processor and the memory communicate with each other through the bus; the processor is used to call program instructions in the memory to execute the wind tunnel-based noise detection method as described in any one of claims 1-7.

Images