An adaptive sound guiding and noise reduction system and method based on multi-band sound wave cancellation

Abstract

The application discloses a kind of adaptive flow guiding noise reduction system and method based on multi-band sound wave cancellation, comprising: noise collection and spectrum analysis module, for collecting real-time noise signal and environmental parameter in engine test run process, and carrying out real-time spectrum analysis to noise signal, output each frequency band noise amplitude, peak frequency and energy proportion;Multi-band active sound wave cancellation module is connected with the noise collection and spectrum analysis module, for 20 ~ 1000Hz low frequency noise, using multi-band adaptive sound wave cancellation mode to generate cancellation sound wave, the present application uses frequency band collaborative noise reduction strategy, different noise reduction means is used for the characteristics of different frequency band noise, wherein 20 ~ 1000Hz low frequency strong noise uses multi-band adaptive active sound wave cancellation, 1 ~ 8kHz high frequency irritating noise uses adaptive flow guiding and sound absorption structure collaborative control, 8kHz above noise uses passive sound absorption supplement, solve the problem that traditional single noise reduction mode cannot realize effective noise reduction in full frequency band.

Description

Technical Field

[0001] This invention relates to the field of noise reduction system technology, and in particular to an adaptive flow-guiding noise reduction system and method based on multi-band sound wave cancellation. Background Technology

[0002] Ground testing of aircraft engines is a critical testing phase before engine assembly, maintenance, and delivery. The testing process generates strong, broadband noise in the 20Hz–20kHz range, with maximum sound pressure levels exceeding 135dB(A). Its noise spectrum characteristics are complex: the low-frequency range (20–1000Hz) is dominated by high-energy roaring noise with strong penetrating power; the mid-to-high frequency range (1–8kHz) includes sharp jet noise and turbulence noise; and the high-frequency range (>8kHz) is characterized by broadband hissing. Existing noise reduction technologies have significant drawbacks: Passive noise reduction has limitations: Traditional soundproof enclosures and anechoic towers are effective for mid-to-high frequency noise above 1kHz, but their attenuation capability for low-frequency noise (20-1000Hz) is extremely weak (typically <10dB). Active noise cancellation (ANC) has poor stability: Existing ANC systems mostly use a single algorithm to process the entire frequency band. In a large space with high sound intensity and strong airflow disturbance, such as a test bench, problems such as algorithm divergence, slow convergence speed (especially at low frequencies), and untimely compensation for secondary channel distortion easily occur, making it difficult to cope with the transient noise surge under engine afterburner conditions. Lack of a collaborative mechanism: Existing technologies lack a frequency-band collaborative strategy for "low-frequency active cancellation plus mid-to-high frequency passive / flow-guiding control plus high-frequency sound absorption," resulting in low overall noise reduction efficiency across the entire frequency band and an inability to dynamically optimize the sound field while ensuring engine exhaust back pressure safety.

[0003] Therefore, there is an urgent need to develop an adaptive collaborative system that can adapt to unsteady working conditions and accurately reduce noise across the entire frequency band. Summary of the Invention

[0004] The purpose of this invention is to provide an adaptive flow-guiding noise reduction system and method based on multi-band acoustic wave cancellation, so as to solve the problems existing in the prior art.

[0005] To achieve the above objectives, the present invention provides an adaptive flow-guiding noise reduction system based on multi-band acoustic wave cancellation, comprising:

[0006] The noise acquisition and spectrum analysis module is used to acquire real-time noise signals and environmental parameters during engine testing, and to perform real-time spectrum analysis on the noise signals, outputting the noise amplitude, peak frequency and energy percentage of each frequency band.

[0007] A multi-band active acoustic wave cancellation module is connected to the noise acquisition and spectrum analysis module and is used to generate cancellation sound waves for low-frequency noise of 20 to 1000 Hz using a multi-band adaptive acoustic wave cancellation method.

[0008] The adaptive flow guidance execution module is used to optimize the sound wave propagation path and absorption efficiency by dynamically adjusting the orientation of the flow guidance structure and the sound absorption parameters for mid-to-high frequency noise in the range of 1 to 8 kHz.

[0009] The collaborative control and feedback module is connected to the noise acquisition and spectrum analysis module, the multi-band active acoustic wave cancellation module, and the adaptive flow guidance execution module, respectively. The collaborative control and feedback module is used to allocate the working weight of each noise reduction module according to the spectrum analysis results, output control commands, and iteratively update the algorithm parameters and structural attitude according to the residual noise signal.

[0010] The modules communicate with each other via real-time industrial Ethernet to ensure data exchange with a control cycle of less than 10ms, forming a fully closed-loop noise reduction system that enables real-time sensing, calculation, cancellation, and adjustment.

[0011] Preferably, the noise acquisition and spectrum analysis module includes:

[0012] The multi-channel microphone array is distributed and the sampling rate is dynamically set according to the highest analysis frequency, not lower than 44.1kHz;

[0013] Environmental sensors are used to collect wind speed and pressure signals;

[0014] The spectrum analysis unit adopts a real-time spectrum analysis method combining frame-based windowed FFT with multi-band smoothing. The frame length is 512 points, the frame shift is 256 points, and the window function is Hanning window.

[0015] The spectrum analysis unit incorporates the MCRA minimum control recursive averaging algorithm for accurate spectrum estimation of unsteady-state test noise, with a spectrum analysis delay of no more than 50ms.

[0016] The noise acquisition and spectrum analysis module also has a built-in noise threshold setting unit. When the noise amplitude exceeds the preset threshold, a high-intensity noise reduction mode is automatically triggered.

[0017] Preferably, the multi-band active acoustic cancellation module is based on the sub-band FXLMS algorithm and includes:

[0018] The orthogonal mirror filter bank is used to decompose the noise signal into 8 to 16 narrow sub-bands, and sets independent step size and weight for different sub-bands. The step size adjustment range is 0.001 to 0.01.

[0019] The secondary sound source array is arranged in a distributed manner, with a frequency response range of 20Hz to 1000Hz.

[0020] The secondary channel online identification unit is used to compensate for acoustic path delay and distortion from the loudspeaker to the error microphone in real time, with an identification accuracy of no less than 95%.

[0021] Specifically, a fast convergence algorithm is used for the 20–200 Hz ultra-low frequency noise sub-band, with a convergence time of no more than 0.5 s; a highly stable algorithm is used for the 200–1000 Hz low frequency noise sub-band.

[0022] Preferably, the adaptive flow redirection execution module includes:

[0023] The adjustable guide vanes are driven by a stepper motor and have an adjustable angle range of 0° to 90°. They can be adjusted while ensuring that the exhaust back pressure does not exceed the set threshold. The vane surface has a sound-absorbing coating and the adjustment response time does not exceed 0.3s.

[0024] The depth of the resonant cavity is dynamically adjusted by an electric adjustment mechanism, with an adjustment range of 50–200 mm.

[0025] The sound-absorbing structure has an open area ratio that is dynamically switched by a solenoid valve, with a switching range of 20% to 60%.

[0026] The control logic configuration of the adaptive flow guidance execution module is as follows:

[0027] When turbulent noise is dominant in the 1-3kHz range, increase the guide vane angle to 60°-90°.

[0028] When the high-frequency noise of 3-8kHz is dominant, the opening ratio of the sound-absorbing structure is adjusted to 40%-60%, and the depth of the resonant cavity is adjusted to 80-150mm.

[0029] When the wind speed is ≥20m / s, the angle of the guide vanes is limited to 30°~60°.

[0030] Preferably, the collaborative control and feedback module includes:

[0031] The collaborative controller, which uses a PLC controller, is used to automatically switch the system's operating mode and allocate weights based on spectrum information;

[0032] The system dynamically adjusts the weight ratio of each dominant mode, ensuring that in the low-frequency dominant mode, the output power of the multi-band active acoustic wave cancellation module is higher than that of the adaptive current guiding execution module. For example:

[0033] Low-frequency dominant mode: The weight of the multi-band active acoustic wave cancellation module is set to 80%, and the adaptive flow guidance execution module maintains a basic posture;

[0034] Mid-to-high frequency dominant mode: The weight of the adaptive flow guidance execution module is set to 70%, and the weight of the multi-band active acoustic wave cancellation module is adjusted to 30%;

[0035] Full-band wideband noise mode: the weight of the multi-band active acoustic wave cancellation module is 50% to 70%, and the weight of the adaptive flow guidance execution module is 30% to 50%;

[0036] The residual noise acquisition unit is used to acquire the residual noise signal after noise reduction in real time and feed it back to the collaborative controller.

[0037] The collaborative controller iteratively updates the step size and weight of the multi-band active acoustic wave cancellation module, as well as the guide vane angle, resonant cavity depth and sound absorption structure opening ratio of the adaptive guide execution module, based on the difference between the residual noise signal and the preset threshold, to ensure that the system response time is less than 1 second.

[0038] Preferably, the collaborative control and feedback module also includes a built-in fault diagnosis unit and a data storage unit;

[0039] The fault diagnosis unit is used to monitor the working status of each module in real time and automatically switch to the standby working mode when a fault is detected.

[0040] The data storage unit is used to store at least one year of noise collection data, system operating parameters, and noise reduction effect data.

[0041] An adaptive flow-guiding noise reduction method based on multi-band acoustic wave cancellation, applied to the aforementioned adaptive flow-guiding noise reduction system based on multi-band acoustic wave cancellation, includes the following steps:

[0042] Step S1: The noise signal and environmental parameters are collected in real time through the noise acquisition and spectrum analysis module, and spectrum analysis is performed to output the noise characteristics of each frequency band;

[0043] Step S2: The collaborative control and feedback module determines the current dominant noise frequency band based on the noise characteristics, selects the corresponding working mode, and assigns working weights to the multi-band active acoustic wave cancellation module and the adaptive flow guidance execution module.

[0044] Step S3: The multi-band active acoustic wave cancellation module generates cancellation acoustic waves using the sub-band FXLMS algorithm for low-frequency noise of 20-1000Hz; at the same time, the adaptive flow guiding execution module dynamically adjusts the angle of the flow guiding blades, the depth of the resonant cavity, and the opening ratio of the sound-absorbing structure for mid-to-high frequency noise of 1-8kHz.

[0045] Step S4: Collect the noise-reduced residual noise signal through the residual noise acquisition unit and feed it back to the collaborative control and feedback module;

[0046] Step S5: The collaborative control and feedback module iteratively updates the algorithm parameters and structural attitude based on the residual noise signal to achieve real-time noise reduction in a fully closed loop.

[0047] Preferably, in step S2...

[0048] If the energy proportion in the 20-1000Hz frequency band exceeds the preset threshold, the system will enter the low-frequency dominant mode.

[0049] If the energy proportion in the 1-8kHz frequency band exceeds the preset threshold, the system will enter the mid-to-high frequency dominance mode.

[0050] If the noise amplitude across the entire frequency band exceeds the preset threshold or the engine is in afterburner mode, then the system will enter full-frequency wideband noise mode.

[0051] Compared with the prior art, the present invention has the following beneficial effects:

[0052] 1. This invention employs a frequency-band collaborative noise reduction strategy, using differentiated noise reduction methods tailored to the characteristics of noise in different frequency bands. Specifically, strong low-frequency noise (20–1000 Hz) is canceled using multi-band adaptive active sound wave cancellation; harsh mid-to-high frequency noise (1–8 kHz) is controlled by adaptive flow guidance and sound-absorbing structures; and noise above 8 kHz is supplemented by passive sound absorption. This solves the problem that traditional single noise reduction methods cannot achieve effective noise reduction across the entire frequency band, achieving precise suppression of wideband noise. Simultaneously, the frequency-band design allows each module to focus on its own advantageous frequency band, improving noise reduction efficiency and accuracy. Compared to existing composite noise reduction systems, the overall noise reduction across the entire frequency band is improved by 5–10 dB(A).

[0053] 2. This invention uses online identification and compensation of acoustic path delay and distortion through secondary channels, and adopts differentiated algorithms for different low-frequency subbands, taking into account both the convergence speed of active noise cancellation and system stability. It solves the problem of insufficient stability of single active noise cancellation in high-intensity and large-space scenarios of engine test benches. The system response time is less than 1 second, which can quickly adapt to the non-steady-state noise changes during engine testing. Especially under engine afterburner conditions, it can still maintain a stable noise reduction effect and avoid cancellation failure.

[0054] 3. The adaptive flow guiding execution module of the present invention can dynamically adjust the angle of the guide vanes, the parameters of the resonant cavity, and the opening ratio of the sound-absorbing structure according to the noise frequency band and wind speed conditions. While suppressing mid-to-high frequency noise, it also takes into account airflow resistance, avoids adverse effects on the airflow field of engine test, ensures the accuracy of engine test data, and improves the adaptability and engineering practicality of the system. Compared with the existing flow guiding noise reduction structure, the noise reduction efficiency of the adaptive flow guiding structure of the present invention is improved by more than 20%.

[0055] 4. This invention constructs a fully closed-loop noise reduction system. Through real-time feedback of residual noise, parameters and structural attitude are iteratively updated. The system has a fast response speed, can adapt to the non-steady-state noise changes of the engine under all operating conditions, and has a stable noise reduction effect. It can meet the requirements of relevant industry standards, improve the working environment of test personnel, and reduce the risk of noise pollution. At the same time, the system has a modular design, which makes installation, debugging and maintenance convenient. It can be extended to high-power broadband noise scenarios such as fans, compressors, and power test benches, with a wide range of applications and significant economic and social benefits.

[0056] 5. The present invention sets up a fault diagnosis unit and a data storage unit in the collaborative control and feedback module, which can monitor the system operating status in real time, detect and handle module faults in a timely manner, and ensure the continuous and stable operation of the system. At the same time, the stored operating data can be used for subsequent system performance optimization and fault diagnosis, reducing system maintenance costs and improving system reliability and service life. Attached Figure Description

[0057] Figure 1 This is a schematic diagram of the workflow of the noise acquisition and spectrum analysis module of the present invention;

[0058] Figure 2 This is a flowchart of the collaborative control logic of the system of the present invention. Detailed Implementation

[0059] The technical solution of the present invention will now be described in detail with reference to the accompanying drawings. It should be noted that the following embodiments are only used to further explain and illustrate the present invention and do not constitute a limitation on the scope of protection of the present invention. Equivalent substitutions or conventional modifications made by those skilled in the art based on the spirit and principles of the present invention should all be included within the scope of protection of the present invention.

[0060] like Figure 1 As shown, the present invention provides an adaptive flow-guiding noise reduction system based on multi-band acoustic wave cancellation, comprising:

[0061] The noise acquisition and spectrum analysis module is used to acquire real-time noise signals and environmental parameters during engine testing, and to perform real-time spectrum analysis on the noise signals, outputting the noise amplitude, peak frequency and energy percentage of each frequency band.

[0062] A multi-band active acoustic wave cancellation module is connected to the noise acquisition and spectrum analysis module and is used to generate cancellation sound waves for low-frequency noise of 20 to 1000 Hz using a multi-band adaptive acoustic wave cancellation method.

[0063] The adaptive flow guidance execution module is used to optimize the sound wave propagation path and absorption efficiency by dynamically adjusting the orientation of the flow guidance structure and the sound absorption parameters for mid-to-high frequency noise in the range of 1 to 8 kHz.

[0064] The collaborative control and feedback module is connected to the noise acquisition and spectrum analysis module, the multi-band active acoustic wave cancellation module, and the adaptive flow guidance execution module, respectively. The collaborative control and feedback module is used to allocate the working weight of each noise reduction module according to the spectrum analysis results, output control commands, and iteratively update the algorithm parameters and structural attitude according to the residual noise signal.

[0065] The modules communicate with each other via real-time industrial Ethernet to ensure data exchange with a control cycle of less than 10ms, forming a fully closed-loop noise reduction system that enables real-time sensing, calculation, cancellation, and adjustment.

[0066] The noise acquisition and spectrum analysis module includes:

[0067] The multi-channel microphone array is distributed and the sampling rate is dynamically set according to the highest analysis frequency, not lower than 44.1kHz;

[0068] Environmental sensors are used to collect wind speed and pressure signals;

[0069] The spectrum analysis unit adopts a real-time spectrum analysis method combining frame-based windowed FFT with multi-band smoothing. The frame length is 512 points, the frame shift is 256 points, and the window function is Hanning window.

[0070] The frame length of 512 points and the frame shift of 256 points are set based on the following: At a sampling rate of 44.1kHz, 512 points correspond to a time resolution of approximately 11.6ms, which is sufficient to capture the periodic characteristics of noise above 20Hz (the period of 20Hz is 50ms), while ensuring that the frequency domain resolution of the FFT calculation (Δf≈86Hz) meets the requirements of frequency band control. The 50% overlap (frame shift of 256) is to reduce information loss caused by windowing and improve the smoothness of the spectrum estimation.

[0071] The spectrum analysis unit incorporates the MCRA minimum control recursive averaging algorithm for accurate spectrum estimation of unsteady-state test noise, with a spectrum analysis delay of no more than 50ms.

[0072] The principle of the MCRA minimum value controlled recursive average algorithm is as follows:

[0073] Perform a short-time Fourier transform on the input signal to calculate the local energy of each frame and each frequency point;

[0074] Speech / noise presence probability estimation: The noise presence probability p is calculated using the ratio of local energy to global minimum energy through the sigmoid function;

[0075] Smooth update: The recursive average coefficient α is dynamically adjusted according to the probability p. When it is determined to be pure noise, α approaches 1 (quickly update the noise spectrum); when it is determined to contain a signal, α approaches 0 (preserve the original noise spectrum estimate).

[0076] Output: Real-time updated ambient noise power spectral density estimate, used for subsequent Wiener filtering or as a reference weight for ANC;

[0077] Parameter settings: The forgetting factor is usually set to 0.9~0.98, and the smoothing window length corresponds to a frame shift of 256 points.

[0078] The noise acquisition and spectrum analysis module also has a built-in noise threshold setting unit. When the noise amplitude exceeds the preset threshold, a high-intensity noise reduction mode is automatically triggered.

[0079] The multi-band active acoustic cancellation module, based on the sub-band FXLMS algorithm, includes:

[0080] The orthogonal mirror filter bank is used to decompose the noise signal into 8 to 16 narrow sub-bands, and sets independent step size and weight for different sub-bands. The step size adjustment range is 0.001 to 0.01.

[0081] The step size of 0.001 to 0.01 is based on experimental values ​​that represent a trade-off between stability and convergence speed.

[0082] 0.01: Applicable to ultra-low frequencies of 20-200Hz. This frequency band has high energy and relatively slow changes. Large step sizes can achieve fast convergence of <0.5s.

[0083] 0.001: Applicable to 200-1000Hz, this frequency band has complex modes. Small step size can prevent algorithm divergence caused by secondary channel modeling errors and ensure stability under high sound intensity.

[0084] The secondary sound source array is arranged in a distributed manner, with a frequency response range of 20Hz to 1000Hz.

[0085] The secondary channel online identification unit is used to compensate for acoustic path delay and distortion from the loudspeaker to the error microphone in real time, with an identification accuracy of no less than 95%.

[0086] Specifically, a fast convergence algorithm is used for the 20–200 Hz ultra-low frequency noise sub-band, with a convergence time of no more than 0.5 s; a highly stable algorithm is used for the 200–1000 Hz low frequency noise sub-band.

[0087] The principle of the subband FXLMS algorithm is as follows:

[0088] Subband division: The full-band (20-1000Hz) is decomposed into M=8~16 subbands using an orthogonal mirror filter bank, with each subband having a bandwidth of approximately 1000 / M Hz.

[0089] Filter order: The adaptive filter order N for each subband can be significantly reduced. For the 20-200Hz ultra-low frequency subband, due to the long wavelength, the required order is less (e.g., N=32~64); for the 200-1000Hz subband, the order can be appropriately increased (e.g., N=64~128), and the total computational load is much smaller than that for full-band processing.

[0090] Weight update formula:

[0091] In the k-th sub-band, the weight update formula at time n is:

[0092] W k(n+1)=Wk(n)+μk.ek(n)Xk′(n)

[0093] Where: W_k: filter weight vector of the k-th sub-band; u_k: independent step size of the k-th sub-band (range 0.001~0.01); larger values ​​are used for low-frequency sub-bands to speed up convergence, and smaller values ​​are used for high-frequency sub-bands to ensure stability; e_k(n): error signal (residual noise) of the k-th sub-band; X'_k(n): the reference signal after being filtered by the **secondary channel estimation model $\hat{S}(z)$.

[0094] Level Channel Online Identification Method:

[0095] The auxiliary noise method is adopted, and low-level white noise or maximum length sequence (M sequence) is injected into the secondary sound source as a probe signal. By comparing the injected signal with the probe signal response received by the error microphone, the coefficients of the secondary channel model \hat{S}(z) are updated in real time using the normalized LMS algorithm to compensate for acoustic path distortion caused by temperature and airflow changes.

[0096] The adaptive flow redirection execution module includes:

[0097] The adjustable guide vanes, driven by a stepper motor, have an adjustable angle range of 0° to 90°. Adjustment is made while ensuring the exhaust back pressure does not exceed a set threshold. The vane surface has a sound-absorbing coating, and the adjustment response time is no more than 0.3 seconds. In the 1-3kHz frequency band, as the vane angle increases from 0° to 75°, the noise reduction increases non-linearly, with a maximum gain of 15dB. However, beyond 75°, the net noise reduction decreases due to increased regenerated noise from airflow separation. Therefore, setting the dominant mode to 60°–90° is the optimal balance point.

[0098] The depth of the resonant cavity is dynamically adjusted by an electric adjustment mechanism, with an adjustment range of 50–200 mm.

[0099] The sound-absorbing structure has an open area ratio that is dynamically switched by a solenoid valve, with a switching range of 20% to 60%.

[0100] In the 3-8kHz frequency range, the porosity of the sound-absorbing structure directly affects the Helmholtz resonant frequency.

[0101] When the aperture ratio is 20%, the resonant frequency is relatively low, making it suitable for absorbing low-frequency components;

[0102] When the aperture ratio is increased to 40%–60%, the resonant frequency shifts to higher frequencies, and the flow resistance characteristics are optimized, increasing the absorption coefficient for sharp noise in the 3–8 kHz range from 0.4 to over 0.85.

[0103] The correlation formula is approximately:

[0104] ;

[0105] Where P is the aperture ratio, V is the cavity volume, and L is the neck depth. By dynamically adjusting P and L, the peak frequency of noise can be tracked in real time.

[0106] The guide vane angle is set between 0° and 90° based on the following: 0° is fully open (minimum resistance), and 90° is fully closed (strongest sound insulation but maximum back pressure). Experiments show that the 60°–90° range significantly improves the scattering and absorption of mid-to-high frequency turbulent noise, while exceeding 90° will cause a sharp increase in back pressure, affecting engine safety.

[0107] The control logic configuration of the adaptive flow guidance execution module is as follows:

[0108] When turbulent noise is dominant in the 1-3kHz range, increase the guide vane angle to 60°-90°.

[0109] When the high-frequency noise of 3-8kHz is dominant, the opening ratio of the sound-absorbing structure is adjusted to 40%-60%, and the depth of the resonant cavity is adjusted to 80-150mm.

[0110] When the wind speed is ≥20m / s, the angle of the guide vanes is limited to 30°~60°.

[0111] It is precisely because of the dynamic matching of the above parameters that the system can achieve 10-15dB attenuation in the low frequency band by relying on active cancellation, and 15-20dB attenuation in the mid-to-high frequency band by relying on current-guided tuning, ultimately achieving a comprehensive improvement of 5-10dB(A) across the entire frequency band.

[0112] The collaborative control and feedback module includes:

[0113] The collaborative controller, which uses a PLC controller, is used to automatically switch the system's operating mode and allocate weights based on spectrum information;

[0114] The system dynamically adjusts the weight ratio of each dominant mode, ensuring that in the low-frequency dominant mode, the output power of the multi-band active acoustic wave cancellation module is higher than that of the adaptive current guiding execution module. For example:

[0115] Low-frequency dominant mode: The weight of the multi-band active acoustic wave cancellation module is set to 80%, and the adaptive flow guidance execution module maintains a basic posture;

[0116] Mid-to-high frequency dominant mode: The weight of the adaptive flow guidance execution module is set to 70%, and the weight of the multi-band active acoustic wave cancellation module is adjusted to 30%;

[0117] Full-band wideband noise mode: the weight of the multi-band active acoustic wave cancellation module is 50% to 70%, and the weight of the adaptive flow guidance execution module is 30% to 50%;

[0118] The PLC (Programmable Logic Controller) internally runs a fuzzy PID weighted algorithm, with the following specific logic:

[0119] Weight calculation logic:

[0120] Let Elow be the normalized energy of the 20-1000Hz frequency band, and Emid be the normalized energy of the 1-8kHz frequency band.

[0121] Define low-frequency coefficients ;

[0122] High frequency coefficients in definition ;

[0123] Active noise reduction weight WANC = 0.3 + 0.5 × α

[0124] That is: the weight is 0.8 for pure low frequency and 0.3 for pure high frequency to ensure basic stability;

[0125] The residual noise acquisition unit is used to acquire the residual noise signal after noise reduction in real time and feed it back to the collaborative controller.

[0126] The collaborative controller iteratively updates the step size and weight of the multi-band active acoustic wave cancellation module, as well as the guide vane angle, resonant cavity depth and sound absorption structure opening ratio of the adaptive guide execution module, based on the difference between the residual noise signal and the preset threshold, to ensure that the system response time is less than 1 second.

[0127] The collaborative control and feedback module also has a built-in fault diagnosis unit and a data storage unit;

[0128] The fault diagnosis unit is used to monitor the working status of each module in real time and automatically switch to the standby working mode when a fault is detected.

[0129] Fault diagnosis logic:

[0130] Sensor malfunction: Monitor whether the microphone signal shows DC offset, clipping distortion, or no signal for a long time (<1mV for 5s).

[0131] Actuator malfunction: Monitor whether the motor current is overloaded, whether the position feedback deviates from the command by more than 5°, and whether the valve action timeout.

[0132] Algorithm failure: Monitor whether the weight vector diverges (norm exceeds the threshold) or whether the residual noise energy increases instead of decreasing for 3 control cycles.

[0133] Conditions and implementation for switching to standby mode:

[0134] Condition: Once any of the above critical faults is detected.

[0135] Implementation method:

[0136] Immediately cut off the high-power output of the active noise cancellation module to prevent howling.

[0137] Mechanical reset: Force the guide vanes to the "safe default position" (usually 30°, taking into account both noise reduction and minimum back pressure).

[0138] Mode downgrade: The system switches to "pure passive noise reduction mode", retaining only the sound absorption function of the fixed structure and sending an alarm code to the host computer.

[0139] Data logging: The waveform data before and after the fault occurs is locked and stored for subsequent analysis;

[0140] The data storage unit is used to store at least one year of noise collection data, system operating parameters, and noise reduction effect data.

[0141] Analog signals are transmitted between the microphone array, environmental sensors, and spectrum analysis unit via shielded twisted-pair cables.

[0142] The co-controller is connected to the stepper motor, solenoid valve, and speaker array via a power drive line to transmit PWM pulses or analog voltage control signals.

[0143] Signal flow direction:

[0144] Input: The analog sound pressure signal acquired by the multi-channel microphone array is converted into a digital signal by the ADC and then transmitted to the noise acquisition and spectrum analysis module.

[0145] Processing end: The analysis results of the noise acquisition and spectrum analysis module are transmitted to the co-controller via the EtherCAT bus.

[0146] Control end: The co-controller calculates the control commands and distributes them to the multi-band active acoustic cancellation module via the EtherCAT bus.

[0147] The collaborative controller outputs an inverted acoustic wave digital signal, which is then converted by a DAC and fed into a power amplifier. After amplification, the signal is fed into a secondary sound source array and then into an adaptive flow control module. The adaptive flow control module outputs motor step count / valve opening commands to the driver to adjust the blade angle / resonant cavity depth / aperture ratio.

[0148] Feedback end: The residual noise microphone collects signals and returns them to the collaborative controller to form a closed loop;

[0149] An adaptive flow-guiding noise reduction method based on multi-band acoustic wave cancellation, applied to an adaptive flow-guiding noise reduction system based on multi-band acoustic wave cancellation, includes the following steps:

[0150] Step S1: The noise signal and environmental parameters are collected in real time through the noise acquisition and spectrum analysis module, and spectrum analysis is performed to output the noise characteristics of each frequency band;

[0151] Step S2: The collaborative control and feedback module determines the current dominant noise frequency band based on the noise characteristics, selects the corresponding working mode, and assigns working weights to the multi-band active acoustic wave cancellation module and the adaptive flow guidance execution module.

[0152] Step S3: The multi-band active acoustic wave cancellation module generates cancellation acoustic waves using the sub-band FXLMS algorithm for low-frequency noise of 20-1000Hz; at the same time, the adaptive flow guiding execution module dynamically adjusts the angle of the flow guiding blades, the depth of the resonant cavity, and the opening ratio of the sound-absorbing structure for mid-to-high frequency noise of 1-8kHz.

[0153] Step S4: Collect the noise-reduced residual noise signal through the residual noise acquisition unit and feed it back to the collaborative control and feedback module;

[0154] Step S5: The collaborative control and feedback module iteratively updates the algorithm parameters and structural attitude based on the residual noise signal to achieve real-time noise reduction in a fully closed loop.

[0155] In step S2,

[0156] If the energy proportion in the 20-1000Hz frequency band exceeds the preset threshold, the system will enter the low-frequency dominant mode.

[0157] If the energy proportion in the 1-8kHz frequency band exceeds the preset threshold, the system will enter the mid-to-high frequency dominance mode.

[0158] If the noise amplitude across the entire frequency band exceeds the preset threshold or the engine is in afterburner mode, then the system will enter full-frequency wideband noise mode.

[0159] The main application scenarios of this invention are as follows:

[0160] Scenario 1: Engine start-up phase

[0161] Noise characteristics: The low-frequency roar gradually increases, with the main energy concentrated in the 50-200Hz range, and the sound pressure level rises slowly.

[0162] System response:

[0163] Mode determination: Spectrum analysis detected that the energy ratio of 20-1000Hz exceeded the threshold, and the co-controller switched to "low frequency dominant mode".

[0164] Weighting: Active noise reduction weight is set to 80%, and the flow guiding module weight is 20% (to maintain the basic posture and reduce startup resistance).

[0165] Action execution: The sub-band FXLMS algorithm is used for fast activation of the 50-200Hz sub-band, with a step size of 0.008 to accelerate convergence; the guide vanes remain stationary at 30°.

[0166] Effect: Low-frequency noise is suppressed by more than 10dB before the engine reaches idle speed, avoiding starting shock waves.

[0167] Scenario 2: Engine afterburner phase

[0168] Noise characteristics: A surge in noise across the entire frequency band, with sharp jet noise (3-8kHz), a dramatic increase in airflow velocity (>50m / s), accompanied by transient impacts.

[0169] System response:

[0170] Mode determination: If the amplitude of the full frequency band exceeds the standard and the wind speed sensor reading is ≥20m / s, immediately switch to "full frequency band wideband noise mode".

[0171] Weighting: Active noise reduction weight is adjusted to 50% (to prevent algorithm divergence under high sound intensity), and adaptive flow guidance weight is increased to 50%.

[0172] Action execution:

[0173] ANC module: The step size is automatically reduced to 0.002 to ensure stability, and the online identification frequency of the secondary channel is doubled (the model is updated every 10ms).

[0174] Airflow guiding module: Based on the wind speed limiting logic, the blade angle is locked in the range of 30°-60° (to avoid noise leakage caused by full opening and to avoid shutdown due to excessive back pressure caused by full closing); at the same time, the resonant cavity depth is adjusted to 100mm and the aperture ratio is adjusted to 50% to absorb 3-8kHz high frequency noise.

[0175] Effect: During the moment of acceleration, the system can complete the attitude adjustment within 0.8 seconds, effectively suppressing high-frequency harsh noise while ensuring smooth engine exhaust.

[0176] Scenario 3: Engine shutdown phase

[0177] Noise characteristics: High-frequency components disappear rapidly, while low-frequency components drift in frequency as the rotational speed decreases.

[0178] System response:

[0179] Mode determination: A continuous decrease in the total sound pressure level was detected, and the output gain of each module was gradually reduced.

[0180] Action execution: The guide vanes slowly return to 0° (fully open) to reduce residual airflow resistance; the ANC filter weights gradually decay to zero to prevent low-frequency oscillations during shutdown.

[0181] Data archiving: Automatically saves all data from the entire test run to the storage unit after shutdown.

[0182] The above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. An adaptive flow-guiding noise reduction system based on multi-band acoustic wave cancellation, characterized in that, include: The noise acquisition and spectrum analysis module is used to acquire real-time noise signals and environmental parameters during engine testing, and to perform real-time spectrum analysis on the noise signals, outputting the noise amplitude, peak frequency and energy percentage of each frequency band. A multi-band active acoustic wave cancellation module is connected to the noise acquisition and spectrum analysis module and is used to generate cancellation sound waves for low-frequency noise of 20 to 1000 Hz using a multi-band adaptive acoustic wave cancellation method. The adaptive flow guidance execution module is used to optimize the sound wave propagation path and absorption efficiency by dynamically adjusting the orientation of the flow guidance structure and the sound absorption parameters for mid-to-high frequency noise in the range of 1 to 8 kHz. The collaborative control and feedback module is connected to the noise acquisition and spectrum analysis module, the multi-band active acoustic wave cancellation module, and the adaptive flow guidance execution module, respectively. The collaborative control and feedback module is used to allocate the working weight of each noise reduction module according to the spectrum analysis results, output control commands, and iteratively update the algorithm parameters and structural attitude according to the residual noise signal. The modules communicate with each other via real-time industrial Ethernet to ensure data exchange with a control cycle of less than 10ms, forming a fully closed-loop noise reduction system that enables real-time sensing, calculation, cancellation, and adjustment.

2. The adaptive flow-guiding noise reduction system based on multi-band acoustic wave cancellation according to claim 1, characterized in that, The noise acquisition and spectrum analysis module includes: The multi-channel microphone array is distributed and the sampling rate is dynamically set according to the highest analysis frequency, not lower than 44.1kHz; Environmental sensors are used to collect wind speed and pressure signals; The spectrum analysis unit adopts a real-time spectrum analysis method combining frame-based windowed FFT with multi-band smoothing. The frame length is 512 points, the frame shift is 256 points, and the window function is Hanning window. The spectrum analysis unit incorporates the MCRA minimum control recursive averaging algorithm for accurate spectrum estimation of unsteady-state test noise, with a spectrum analysis delay of no more than 50ms. The noise acquisition and spectrum analysis module also has a built-in noise threshold setting unit. When the noise amplitude exceeds the preset threshold, a high-intensity noise reduction mode is automatically triggered.

3. The adaptive flow-guiding noise reduction system based on multi-band acoustic wave cancellation according to claim 1, characterized in that, The multi-band active acoustic cancellation module, based on the sub-band FXLMS algorithm, includes: The orthogonal mirror filter bank is used to decompose the noise signal into 8 to 16 narrow sub-bands, and set independent step size and weight for different sub-bands. The step size adjustment range is 0.001 to 0.

01. The secondary sound source array is arranged in a distributed manner, with a frequency response range of 20Hz to 1000Hz. The secondary channel online identification unit is used to compensate for acoustic path delay and distortion from the loudspeaker to the error microphone in real time, with an identification accuracy of no less than 95%. Specifically, a fast convergence algorithm is used for the 20~200Hz ultra-low frequency noise sub-band, with a convergence time of no more than 0.5s; a highly stable algorithm is used for the 200~1000Hz low frequency noise sub-band.

4. The adaptive flow-guiding noise reduction system based on multi-band acoustic wave cancellation according to claim 1, characterized in that, The adaptive flow redirection execution module includes: The adjustable guide vanes are driven by a stepper motor and have an adjustable angle range of 0° to 90°. They can be adjusted while ensuring that the exhaust back pressure does not exceed the set threshold. The vane surface has a sound-absorbing coating and the adjustment response time does not exceed 0.3s. The depth of the resonant cavity is dynamically adjusted by an electric adjustment mechanism, with an adjustment range of 50–200 mm. The sound-absorbing structure has an open area ratio that is dynamically switched by a solenoid valve, with a switching range of 20% to 60%. The control logic configuration of the adaptive flow guidance execution module is as follows: When turbulent noise is dominant in the 1-3kHz range, increase the guide vane angle to 60°-90°. When the high-frequency noise of 3-8kHz is dominant, the opening ratio of the sound-absorbing structure is adjusted to 40%-60%, and the depth of the resonant cavity is adjusted to 80-150mm. When the wind speed is ≥20m / s, the angle of the guide vanes is limited to 30°~60°.

5. The adaptive flow-guiding noise reduction system based on multi-band acoustic wave cancellation according to claim 1, characterized in that, The collaborative control and feedback module includes: The collaborative controller, which uses a PLC controller, is used to automatically switch the system's operating mode and allocate weights based on spectrum information; The system dynamically adjusts the weight ratio of each dominant mode, ensuring that in the low-frequency dominant mode, the output power of the multi-band active acoustic wave cancellation module is higher than that of the adaptive current guiding execution module. For example: Low-frequency dominant mode: The weight of the multi-band active acoustic wave cancellation module is set to 80%, and the adaptive flow guidance execution module maintains a basic posture; Mid-to-high frequency dominant mode: The weight of the adaptive flow guidance execution module is set to 70%, and the weight of the multi-band active acoustic wave cancellation module is adjusted to 30%; Full-band wideband noise mode: the weight of the multi-band active acoustic wave cancellation module is 50% to 70%, and the weight of the adaptive flow guidance execution module is 30% to 50%; The residual noise acquisition unit is used to acquire the residual noise signal after noise reduction in real time and feed it back to the collaborative controller. The collaborative controller iteratively updates the step size and weight of the multi-band active acoustic wave cancellation module, as well as the guide vane angle, resonant cavity depth and sound absorption structure opening ratio of the adaptive guide execution module, based on the difference between the residual noise signal and the preset threshold, to ensure that the system response time is less than 1 second.

6. The adaptive flow-guiding noise reduction system based on multi-band acoustic wave cancellation according to claim 5, characterized in that, The collaborative control and feedback module also has a built-in fault diagnosis unit and a data storage unit; The fault diagnosis unit is used to monitor the working status of each module in real time and automatically switch to the standby working mode when a fault is detected. The data storage unit is used to store at least one year of noise collection data, system operating parameters, and noise reduction effect data.

7. An adaptive flow-guiding noise reduction method based on multi-band acoustic wave cancellation, applied to the adaptive flow-guiding noise reduction system based on multi-band acoustic wave cancellation as described in any one of claims 1 to 6, characterized in that, Includes the following steps: Step S1: The noise signal and environmental parameters are collected in real time through the noise acquisition and spectrum analysis module, and spectrum analysis is performed to output the noise characteristics of each frequency band; Step S2: The collaborative control and feedback module determines the current dominant noise frequency band based on the noise characteristics, selects the corresponding working mode, and assigns working weights to the multi-band active acoustic wave cancellation module and the adaptive flow guidance execution module. Step S3: The multi-band active acoustic wave cancellation module generates cancellation acoustic waves using the sub-band FXLMS algorithm for low-frequency noise of 20-1000Hz; at the same time, the adaptive flow guiding execution module dynamically adjusts the angle of the flow guiding blades, the depth of the resonant cavity, and the opening ratio of the sound-absorbing structure for mid-to-high frequency noise of 1-8kHz. Step S4: Collect the noise-reduced residual noise signal through the residual noise acquisition unit and feed it back to the collaborative control and feedback module; Step S5: The collaborative control and feedback module iteratively updates the algorithm parameters and structural attitude based on the residual noise signal to achieve real-time noise reduction in a fully closed loop.

8. The adaptive flow-guiding noise reduction method based on multi-band acoustic wave cancellation according to claim 7, characterized in that, In step S2, If the energy proportion in the 20-1000Hz frequency band exceeds the preset threshold, the system will enter the low-frequency dominant mode. If the energy proportion in the 1-8kHz frequency band exceeds the preset threshold, the system will enter the mid-to-high frequency dominance mode. If the noise amplitude across the entire frequency band exceeds the preset threshold or the engine is in afterburner mode, then the system will enter full-frequency wideband noise mode.

Images