阅读说明：本技术 一种用于通风管道噪声的降噪系统及降噪方法 (Noise reduction system and method for noise of ventilating duct ) 是由 张晓晖 杨松楠 刘浩林 赵伟 于 2021-07-21 设计创作，主要内容包括：本发明公开了一种用于通风管道噪声的降噪系统,包括信号采集模块,信号采集模块通过导线连接有控制器,控制器通过导线连接有信号输出模块,解决了现有降噪装置对通风管道的噪声消除作用不大的问题。本发明还公开了一种用于通风管道噪声的降噪方法：在通风管道内壁上固接参考麦克风、误差麦克风和抗噪扬声器,并将误差麦克风靠近噪声源设置；误差麦克风和参考麦克风采集声音信号,并将采集到的声音信号通过LC滤波器进行预处理,得到模拟信号；A/D转换器将模拟信号转换为数字信号；数控制器采用降噪算法计算出噪声控制信号；抗噪扬声器发出抗噪声波,与噪声相互干涉,完成对通风管道噪声的降噪处理。(The invention discloses a noise reduction system for noise of a ventilating duct, which comprises a signal acquisition module, wherein the signal acquisition module is connected with a controller through a lead, and the controller is connected with a signal output module through a lead, so that the problem that the existing noise reduction device has small noise elimination effect on the ventilating duct is solved. The invention also discloses a noise reduction method for the noise of the ventilating duct, which comprises the following steps: a reference microphone, an error microphone and an anti-noise loudspeaker are fixedly connected to the inner wall of the ventilation pipeline, and the error microphone is arranged close to a noise source; collecting sound signals by an error microphone and a reference microphone, and preprocessing the collected sound signals by an LC filter to obtain analog signals; the A/D converter converts the analog signal into a digital signal; the digital controller calculates a noise control signal by adopting a noise reduction algorithm; the anti-noise loudspeaker sends anti-noise waves which interfere with noise mutually, and noise reduction processing of the ventilation pipeline noise is completed.)

1. A noise reduction system for noise of a ventilation pipeline is characterized by comprising a signal acquisition module, wherein the signal acquisition module is connected with a controller through a lead, an LC filter, an A/D converter, a filter, a coder-decoder and a power amplifier are connected on board in the controller, and the controller is connected with a signal output module through a lead;

the signal acquisition module comprises an error microphone and a reference microphone, and the error microphone and the reference microphone are both connected with the controller through leads;

the signal output module comprises an anti-noise loudspeaker, and the anti-noise loudspeaker is connected with the controller through a lead.

2. The system of claim 1, wherein the error microphone and the reference microphone are electret condenser microphones.

3. A noise reduction method for ventilation duct noise, which is implemented by applying the noise reduction system for ventilation duct noise according to any one of claims 1-2, and is characterized by comprising the following steps:

step 1, fixedly connecting a reference microphone, an error microphone and an anti-noise loudspeaker on the inner wall of a ventilation pipeline, and arranging the error microphone close to a noise source;

step 2, collecting sound signals by an error microphone and a reference microphone, and preprocessing the collected sound signals by an LC filter to obtain analog signals;

step 3, transmitting the analog signal in the step 2 to an A/D converter, and converting the analog signal into a digital signal;

step 4, the digital signals in the step 3 are transmitted to a control system through I2S communication, and the control system calculates noise control signals by adopting a noise reduction algorithm;

step 5, the noise control signal calculated in the step 4 is transmitted to a coder-decoder through I2S communication, the noise signal passing through the coder-decoder passes through a filter and a power amplifier in sequence, and finally, the signal is output to an anti-noise loudspeaker;

and 6, the anti-noise loudspeaker sends out anti-noise waves, noise reduction points are achieved through the secondary channel, the anti-noise waves interfere with noise, and noise reduction processing of the ventilation pipeline is completed.

4. The noise reduction method for the ventilation duct noise according to claim 3, wherein in the step 1, shock absorbing gaskets are respectively installed between the reference microphone, the error microphone and the inner wall of the duct.

5. The method for reducing noise of ventilation ducts according to claim 3, wherein the noise reduction algorithm in step 4 is specifically:

step 4.1, the signal received by the reference microphone is x, and the signal received by the error microphone is e;

step 4.2, if the signal received by the reference microphone at the time n in step 4.1 is x (n), calculating an output signal u (n) at the time n of the filter according to the formula (1):

in formula (1), M is the number of taps of the filter, w_i(n) is the ith filter tap weight at the nth moment;

step 4.3, using the pipeline estimation modelCorrecting the calculation deviation caused by different lengths and thicknesses of the pipelines according to the following formula (2):

in the formula (2), the first and second groups,for the pipeline estimation model at time n, i ═ 0, 1, 2, …, M-1]

x_f(n) using a pipeline estimation modelCorrected signals arriving at the error microphone;

step 4.4, as shown in the following formula (3), by iteratively updating the weight coefficient of the adaptive filter W,

W(n+1)＝W(n)-2μ(n)e(n)x_f(n) (3)

in formula (3), W (n) is the weight coefficient at time n, W (n +1) is the weight coefficient at the next time, μ (n) is the convergence step at time n, e (n) is the signal received by the error microphone at time n,

in order to ensure the stability of the algorithm, formula 4 is a value range of the convergence step length at n moments, and formula 5 is a step length update strategy at n moments:

in the formula (5), a is an amplitude adjusting parameter, b is an amplitude adjusting parameter, and e (n-1) is a signal received by the error microphone at the moment of n-1;

step 4.5, step 4.1 to step 4.4, until the mean square error function j (n) at time n reaches the minimum, as shown in equation (6), the minimum of j (n) depends on the actual situation:

in the formula (6), d (n) is a signal received by the error microphone when the noise reduction is not turned on at time n.

6. The method for reducing noise of the ventilation duct according to claim 3, wherein since the acoustic characteristics of different ducts are different, in order to eliminate the error of the noise reduction algorithm in the step 4, before the noise control calculation, the duct model is sufficiently excited by using Gaussian white noise to obtain an estimated value of the duct acoustic model.

7. The method for reducing noise of the ventilation duct according to claim 6, wherein the modeling step of the duct is as follows:

step 1, using a Gaussian noise signal as an input signal of a filter and as a secondary path estimation model at n timeAnd a reference input signal of a minimum mean square error algorithm;

step 2, receiving the modeling signal at the n moment by the error microphoneThe reference microphone n receives a signal u (n) at the moment, and the following formula (7):

in the formula (7), S_i(n) is the actual acoustic model of the pipeline;

and 3, calculating a secondary path estimation model through a formula (8):

in the formula (8), the first and second groups,is an output signal generated by a pipeline estimation model;

step 4, calculating the output difference e (n) generated by the pipeline actual model and the pipeline estimation model at the time n according to the formula (9):

step 5, updating the pipeline estimation model by adopting a minimum mean square error algorithm through a formula (10)Coefficient (c):

in the formula (10), μ is a fixed step length for modeling;

step 6, repeating the steps 2 to 5 until the error signal e (n) meets the set requirementStoringAnd the numerical value is used by the noise reduction algorithm.

8. Method for reducing the noise of ventilation ducts according to claim 7, characterized in that the estimation model used for duct modeling is an estimation modelFor an FIR type filter with M-25, the step value used in the modeling is 0.001, and the reference signal is a sinusoidal signal with amplitude 1.

Technical Field

The invention belongs to the technical field of noise reduction equipment of a ventilating duct, and relates to a noise reduction system for the ventilating duct and a noise reduction method for the ventilating duct.

Background

With the continuous progress and development of industrial technology, the influence of living noise and industrial noise on people is gradually increased. Long term exposure to noise can have a significant negative impact on the physical and mental well-being of a person. How to effectively reduce noise pollution becomes a difficult problem which needs to be solved urgently. The traditional passive noise reduction device and the traditional noise reduction method have small effect of eliminating low-frequency noise in a ventilation pipeline, and the volume of a noise absorption system needs to be increased to obtain a better noise reduction effect, but the construction cost and the construction difficulty are increased. The active noise control method is used as a new noise reduction means, has a good noise reduction effect aiming at medium and low frequency noise, has the advantages of small volume, wide application range and the like, and can be used for solving various low frequency noise pollution. Noise generated by an axial flow fan in a fresh Air system (HVAC) is low-frequency noise, and due to the waveguide characteristic of a pipeline, the low-frequency noise is transmitted to a far place along the pipeline, so that an efficient method is needed for reducing the pipeline noise.

Disclosure of Invention

The invention aims to provide a noise reduction system for ventilating duct noise, which solves the problem that the existing noise reduction device has little effect on eliminating low-frequency noise in a ventilating duct.

The invention also aims to provide a noise reduction method for ventilating duct noise, which realizes self-adaptive active noise reduction by constructing the mapping relation between vibration and noise in real time.

The technical scheme adopted by the invention is that the noise reduction system for the noise of the ventilating duct comprises a signal acquisition module, wherein the signal acquisition module is connected with a controller through a lead, an LC filter, an A/D converter, a filter, a coder-decoder and a power amplifier are connected on board in the controller, and the controller is connected with a signal output module through a lead;

the signal acquisition module comprises an error microphone and a reference microphone, and the error microphone and the reference microphone are both connected with the controller through leads;

the signal output module comprises an anti-noise loudspeaker which is connected with the controller through a lead

The technical solution of the present invention is also characterized in that,

the error microphone and the reference microphone both adopt electret condenser microphones.

The invention adopts another technical scheme that the noise reduction method for the noise of the ventilating duct is implemented according to the following steps:

step 1, fixedly connecting a reference microphone, an error microphone and an anti-noise loudspeaker on the inner wall of a ventilation pipeline, and arranging the error microphone close to a noise source;

step 2, collecting sound signals by an error microphone and a reference microphone, and preprocessing the collected sound signals by an LC filter to obtain analog signals;

step 3, transmitting the analog signal in the step 2 to an A/D converter, and converting the analog signal into a digital signal;

step 4, the digital signals in the step 3 are transmitted to a controller through I2S communication, and the controller calculates noise control signals by adopting a noise reduction algorithm;

step 5, the noise control signal calculated in the step 4 is transmitted to a coder-decoder through I2S communication, the noise signal passing through the coder-decoder passes through a filter and a power amplifier in sequence, and finally, the signal is output to an anti-noise loudspeaker;

and 6, the anti-noise loudspeaker sends out anti-noise waves, noise reduction points are achieved through the secondary channel, the anti-noise waves interfere with noise, and noise reduction processing of the ventilation pipeline is completed.

Another solution according to the invention is also characterized in that,

in the step 1, shock-absorbing gaskets are arranged between the reference microphone, the error microphone, the anti-noise loudspeaker and the inner wall of the pipeline.

The noise reduction algorithm in the step 4 specifically comprises the following steps:

step 4.1, the signal received by the reference microphone is x, and the signal received by the error microphone is e;

step 4.2, if the signal received by the reference microphone at the time n in step 4.1 is x (n), calculating an output signal u (n) at the time n of the filter according to the formula (1):

in formula (1), M is the number of taps of the filter, w_i(n) is the ith filter tap weight at the nth moment;

step 4.3, using the pipeline estimation modelCorrecting the calculation deviation caused by different lengths and thicknesses of the pipelines according to the following formula (2):

in the formula (2), the first and second groups,for the pipeline estimation model at time n, i ═ 0, 1, 2]x_f(n) using a pipeline estimation modelCorrected signals arriving at the error microphone;

step 4.4, as shown in the following formula (3), by iteratively updating the weight coefficient of the adaptive filter W,

W(n+1)＝W(n)-2μ(n)e(n)x_f(n) (3)

in formula (3), W (n) is the weight coefficient at time n, W (n +1) is the weight coefficient at the next time, μ (n) is the convergence step at time n, e (n) is the signal received by the error microphone at time n,

in order to ensure the stability of the algorithm, formula 4 is a value range of the convergence step length at n moments, and formula 5 is a step length update strategy at n moments:

in the formula (5), a is an amplitude adjusting parameter, b is an amplitude adjusting parameter, and e (n-1) is a signal received by the error microphone at the moment of n-1;

step 4.5, step 4.1 to step 4.4, until the mean square error function j (n) at time n reaches the minimum, as shown in equation (6), the minimum of j (n) depends on the actual situation:

in the formula (6), d (n) is a signal received by the error microphone when the noise reduction is not turned on at time n.

Because the acoustic characteristics of different pipelines are different, in order to eliminate the error of the noise reduction algorithm in the step 4, the pipeline model is fully excited by using Gaussian white noise before noise control calculation, and the estimation value of the pipeline acoustic model is obtained.

The modeling steps of the pipeline are as follows:

step 1, using a Gaussian noise signal as an input signal of a filter and as a secondary path estimation model at n timeAnd a reference input signal of a minimum mean square error algorithm;

step 2, receiving the modeling signal at the n moment by the error microphoneThe reference microphone n receives a signal u (n) at the moment, and the following formula (7):

in the formula (7), S_i(n) is an acoustic model of the pipeline reality；

And 3, calculating a secondary path estimation model through a formula (8):

in the formula (8), the first and second groups,is an output signal generated by a pipeline estimation model;

step 4, calculating the output difference e (n) generated by the pipeline actual model and the pipeline estimation model at the time n according to the formula (9):

step 5, updating the pipeline estimation model by adopting a minimum mean square error algorithm through a formula (10)Coefficient (c):

in the formula (10), μ is a fixed step length for modeling;

step 6, repeating the steps 2 to 5 until the error signals e (n) meet the set requirements, and storing the error signals e (n)And the numerical value is used by the noise reduction algorithm.

Estimation model for pipeline modelingFor an FIR type filter with M-25, the step value used in the modeling is 0.001, and the reference signal is a sinusoidal signal with amplitude 1.

The invention has the beneficial effects that:

compared with the traditional passive noise reduction method, the noise reduction effect of the noise reduction method for the ventilating duct noise is greatly improved, the low-frequency noise of the ventilating duct can be eliminated, and meanwhile, the system structure is small in size. The test input signal is composed of white Gaussian noise and a sinusoidal signal, the reference signal is a 200Hz sinusoidal signal, and the test noise is 20 dB. The sub-path modeling uses a 16-tap FIR filter, where Pz is [0.01,0.25,0.5,1,0.5,0.25,0.01], Sz is [ Pz × 0.25 ], and μ is 0.1.

Drawings

FIG. 1 is a schematic view of the structural installation of a noise reduction system for ventilation duct noise of the present invention;

FIG. 2 is a flowchart illustrating the operation of the software portion of the noise reduction system for ventilation duct noise according to the present invention;

FIG. 3 is a flow chart of a noise reduction algorithm for a noise reduction method for ventilation duct noise according to the present invention;

FIG. 4 is a duct modeling flow diagram of a method of noise reduction for ventilation duct noise of the present invention;

fig. 5 is a diagram of an actual noise reduction effect of a noise reduction method for ventilation duct noise of the present invention;

fig. 6 is a noise reduction effect analysis chart of a noise reduction method for ventilation duct noise of the present invention.

Detailed Description

The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

The invention discloses a noise reduction system for noise of a ventilation pipeline, which comprises a signal acquisition module, wherein the signal acquisition module is connected with a controller through a lead, an LC filter, an A/D converter, a filter, a coder-decoder and a power amplifier are connected on an onboard inside the controller, the controller is connected with a signal output module through a lead, the model of a chip of the controller is STM32F746, and the model of the coder-decoder is WM 8994.

The signal acquisition module comprises an error microphone and a reference microphone, and the error microphone and the reference microphone are both connected with the controller through leads;

the signal output module comprises an anti-noise loudspeaker, and the anti-noise loudspeaker is connected with the controller through a lead.

The error microphone and the reference microphone both adopt electret condenser microphones.

As shown in fig. 2, the software portion mainly includes four modules: the system comprises a USB AUDIO STACK module, a board-level support packet module, a HAL library driving module and a USB initialization module. Firstly, the HAL library driving module is used for carrying out initialization operation on I2C, I2S and a clock which need to be used, mapping output pins of signals, and connecting a main chip and a peripheral chip.

Then, the main function uses the USB device descriptor to initialize the USB interface, and by short-circuiting Vbus of the USB with D-and setting D + to a pull-up state, at this time, Windows will think that the USB interface has a full-speed device, and complete the enumeration process of the USB. The data transmission process comprises the following steps: when the Codec end receives signals from the error microphone and the reference microphone, the signals are transmitted to the main controller through I2S. And when the controller sends data to the Codec, the same audio data of one frame is sent to the PC end through the USB, and the recording and analysis are carried out at the PC end.

And finally, the transmitted error microphone sound signal and a reference sound signal received by a reference microphone enter an HAL Driver layer, an improved active noise control algorithm calculates the A/D conversion data of the Codec register in the HAL Driver layer, and after the calculation of one frame of audio data is finished, the USB interrupt is triggered to send one frame of processed data to the PC end, and the frame of processed data is called back to the USB Conf module. Then repeating the process, continuously receiving new data for processing, and completing the whole process of iteration and record analysis. The USB Audio Class (UAC) module is a standard data I/O module and is a general protocol for Audio transmission among different devices. The UAC module receives audio data from WM8994 and transfers the audio data to PC by frame, and transmits the data calculated by PC to WM8994 by frame, the UAC transmission format set by the invention is 44.1Khz sampling rate and 16bit sampling digit.

The invention discloses a noise reduction method for noise of a ventilating duct, which is implemented by applying a noise reduction system for the noise of the ventilating duct according to the following steps:

step 1, fixedly connecting a reference microphone, an error microphone and an anti-noise loudspeaker on the inner wall of a ventilation pipeline, and arranging the error microphone close to a noise source;

step 2, collecting sound signals by an error microphone and a reference microphone, and preprocessing the collected sound signals by an LC filter to obtain analog signals;

step 3, transmitting the analog signal in the step 2 to an A/D converter, and converting the analog signal into a digital signal;

step 4, the digital signals in the step 3 are transmitted to a control system through I2S communication, and the control system calculates noise control signals by adopting a noise reduction algorithm;

step 5, the noise control signal calculated in the step 4 is transmitted to a coder-decoder through I2S communication, the noise signal passing through the coder-decoder passes through a filter and a power amplifier in sequence, and finally the signal is output to an anti-noise loudspeaker;

and 6, the anti-noise loudspeaker sends out anti-noise waves, noise reduction points are achieved through the secondary channel, the anti-noise waves interfere with noise, and noise reduction processing of the ventilation pipeline is completed.

Shock-absorbing gaskets are arranged between the reference microphone, the error microphone and the inner wall of the pipeline.

As shown in fig. 3, the noise reduction algorithm in step 4 specifically includes:

step 4.1, the signal received by the reference microphone is x, and the signal received by the error microphone is e;

step 4.2, if the signal received by the reference microphone at the time n in step 4.1 is x (n), calculating an output signal u (n) at the time n of the filter according to the formula (1):

in formula (1), M is the number of taps of the filter, w_i(n) is the ith filter tap weight at the nth moment;

step 4.3, using the pipeline estimation modelCorrecting the calculation deviation caused by different lengths and thicknesses of the pipelines according to the following formula (2):

in the formula (2), the first and second groups,for the pipeline estimation model at time n, i ═ 0, 1, 2, …, M-1]x_f(n) using a pipeline estimation modelCorrected signals arriving at the error microphone;

step 4.4, as shown in the following formula (3), by iteratively updating the weight coefficient of the adaptive filter W,

W(n+1)＝W(n)-2μ(n)e(n)x_f(n) (3)

in formula (3), W (n) is the weight coefficient at time n, W (n +1) is the weight coefficient at the next time, μ (n) is the convergence step at time n, e (n) is the signal received by the error microphone at time n,

in order to ensure the stability of the algorithm, formula 4 is a value range of the convergence step length at n moments, and formula 5 is a step length update strategy at n moments:

in the formula (5), a is an amplitude adjusting parameter, b is an amplitude adjusting parameter, and e (n-1) is a signal received by the error microphone at the moment of n-1;

step 4.5, step 4.1 to step 4.4, until the mean square error function j (n) at time n reaches the minimum, as shown in equation (6), the minimum of j (n) depends on the actual situation:

in the formula (6), d (n) is a signal received by the error microphone when the noise reduction is not turned on at time n.

If the secondary path is not modeled, the estimated value and the actual value of the input signal generate larger deviation, so that the convergence range of the LMS algorithm is reduced, the robustness of the ANC system is deteriorated, and the deviation caused by the secondary path also increases the integral operation amount of the control algorithm, so that the difficulty of real-time operation is increased. In the noise control process, the real-time performance of the system is ensured, and the secondary channel can reduce the stable convergence speed of the filter, so that the system divergence is caused in serious cases.

Because the controlled object of the active noise reduction system is a noise sound wave, rather than an electric signal generated by an electronic element, the output signal u (n) obtained by the adaptive filter needs to be converted into a sound wave signal, and a DAC is further used for converting the output signal into an analog signal, so that the cancellation of the noise wave and the noise signal can be realized. A transfer function exists between the anti-noise speaker and the error microphone, and the transfer function includes inherent delays between the anti-noise speaker and the error microphone and between the estimation value and the actual value, and a large deviation is generated between the estimation value and the actual value of the model due to the acoustic characteristics of the pipeline, so that the pipeline model needs to be sufficiently excited by white gaussian noise before noise control calculation is performed, and the estimation value of the pipeline model is obtained as shown in fig. 4.

The modeling steps of the pipeline are as follows:

step 1, using a Gaussian noise signal as an input signal of a filter and as a secondary path estimation model at n timeAnd a reference input signal of a minimum mean square error algorithm;

step 2, receiving the modeling signal at the n moment by the error microphoneThe reference microphone n receives a signal u (n) at the moment, and the following formula (7):

in the formula (7), S_i(n) is the actual acoustic model of the pipeline;

and 3, calculating a secondary path estimation model through a formula (8):

in the formula (8), the first and second groups,is an output signal generated by a pipeline estimation model;

step 4, calculating the output difference e (n) generated by the pipeline actual model and the pipeline estimation model at the time n according to the formula (9):

step 5, updating the pipeline estimation model by adopting a minimum mean square error algorithm through a formula (10)Coefficient (c):

in the formula (10), μ is a fixed step length for modeling;

step 6, repeating the steps 2 to 5 until the error signals e (n) meet the set requirements, and storing the error signals e (n)And the numerical value is used by the noise reduction algorithm.

Estimation model for pipeline modelingFor an FIR type filter with M-25, the step value used in the modeling is 0.001, and the reference signal is a sinusoidal signal with amplitude 1.

The effect of the present invention will be further explained in conjunction with simulation experiments.

1. Simulation conditions are as follows:

the experimental practical data processing software is Matlab2019b, the interface is Audio Tool Box, and the installation environment isThe computer of Core i5-4690, the size of the pipeline is 820mm 110mm, the size of the tee joint connector D is 120mm, the pipeline accords with GB/T5836.2 standard, the material is PVC, the type of the error microphone is ECM999, and 48V phantom power supply is used for excitation; the SPA2201 is used as a noise source, and Gaussian noise with 20dB of test noise is used in the experiment; the anti-noise speaker uses SPA 2201; the data acquisition card uses Yamaha UR 44; algorithm parameters set to a 0.5 and b 0.01 sample rate set to 44100, 100000 data points were tested and iterated 200 times in a loop using the 16 tap FIR filter initial parameters [0.01,0.01,0.02, 0.07,0.1,0.15, 0.2,0.3,0.4,0.3,0.2,0.1,0.009, 0.07,0.02,0.01, 0.009]。

2. Simulation content and results:

fig. 5 is a diagram of an actual noise reduction effect of the noise reduction platform, where a blue line represents an original noise signal received by the error microphone before noise reduction is performed, and a red line represents a residual noise signal received by the error microphone after the noise reduction system is turned on. As is apparent from fig. 5, after the noise reduction system performs a short iterative control process, the noise near the error microphone can be effectively reduced, and the noise reduction system has good stability after convergence, so that the noise reduction effect has continuity. FIG. 6 is a diagram showing that short-time Fourier analysis is performed on the original noise and the residual noise after noise reduction, and it can be seen that after the noise reduction is started, the method of the present invention can effectively reduce the low-frequency noise of 20-8kHz of 14-16dB, and the frequency band noise belongs to the main component of the noise in the pipeline, and the human ear has strong feeling on the noise. In addition, the noise of the frequency band is passively reduced and is difficult to reduce, and the method of the invention effectively solves the problem of poor noise reduction effect of low-frequency noise in the pipeline.