KR920004775B1 - Active noise control method of duct - Google Patents

Abstract

The apparatus includes one or more of detecting sensors (40) for detecting noises in a noisy environment. One or more of offsetting sound generators (61,62) offsets the inputted noises to minimize the output noises, and one or more of sound feed-back compensating filters (71,72) compensate the sound fed back to the detecting sensors. One or more of main control filters (81,82) receive signals from a subtractor (92) to output offsetted sounds and signals, and one or more of error sensors (51,52) detect the non-offsetted portions of the output of the offsetted sound wave generators (61,62). A main control filter LMS adapted algorithm (90) controls the main control filters (81,82).

Description

Active noise attenuation device in duct and applied acoustic environment

1 is an overall block diagram for explaining the apparatus of the present invention.

2 is a connection state diagram of each component of the apparatus of the present invention.

3 is a diagram illustrating an acoustic feedback compensation filter estimation technique used in the present invention.

4 is a diagram showing an error path compensation filter estimating method used in the apparatus of the present invention.

5 is a detailed diagram of an error path compensation filter group and an LMS adaptation algorithm in the present invention.

6 and 7 are block diagrams illustrating a conventional active noise attenuation method.

* Explanation of symbols for main parts of the drawings

10: Applied acoustic environment 21: Input noise

22: output noise 30: error path compensation filter group

40: detection sensor 51,52: error sensor

61,62: first and second offset generator 71,72: first and second acoustic feedback compensation filter

81,82: first and second subject filter 90: LMS adaptation algorithm

100: control unit

140: Digital Signal Processing Microprocessor

The present invention relates to an active noise attenuation device using an adaptive filter in a duct, pipe, exhaust pipe, indoor space, or a similar acoustic environment (hereinafter referred to as an applied acoustic environment). By using a signal converter (microphone, pressure gauge, accelerometer, detection sensor, etc.) and one or more error detection signal converters (hereinafter, referred to as error sensors) to improve the observability of the active noise control system Using one or more noise canceling sound generators (such as loudspeakers and vibration suppressors) to improve the controllability of the active noise attenuation system and transmit it as a plane wave in the applied acoustic environment. Vibration in duct and applied acoustic environments that not only increases the reduction of noise, but also enables the control of noise transmitted in the higher order mode. It relates to a noise damping device.

Conventionally, sound absorbing materials have been mainly used to reduce noise in the acoustic environment mentioned above.

However, in the case of using such a sound absorbing material, the noise reduction measures are effective in the high frequency region of 500 Hz or more, but the effect is very small in the low frequency region of 500 Hz or less.

The low frequency noise below 500HZ causes vibrations of various structures, destroying the environment requiring precise work, and even damaging the structures, and disturbing the psychology of the person exposed to these noises. Of course, it can also cause hearing loss.

In addition, when the sound absorbing material is used to reduce noise in a place where a clean environment is required, there is a problem that it is difficult to preserve the clean environment due to the fine particles falling off from the sound absorbing material.

In order to solve these problems, instead of passive noise control using sound absorbing materials, sound waves having the same amplitude and phase difference 180 ° are artificially generated to cancel artificially generated sound waves and unwanted noises. Active Noise Control has been developed.

6 and 7 show typical conventional active noise attenuation methods, and in FIG. 6, two speakers 61 and 62 artificially generating predetermined sound waves; The offset sound waves radiated from the speaker are propagated only in the direction of the output noise 22 by using the time delay device 99 for giving a difference between the sound waves output from the speakers 61 and 62 so that the offset sound waves are input sound waves. It is a method to prevent the detection again by the detection sensor 40 installed on the (21) side.

In the case of using a directional speaker array, an acoustic feedback phenomenon in which offset sound waves are detected by a detection sensor can be detected to prevent divergence of an active noise attenuator due to a feedback loop, but a speaker array having directionality required precision is required. It is preferable to use a non-directional speaker which is very difficult to install and requires a lot of installation and manufacturing costs, and is easy to install and low in manufacturing and installation cost, and the acoustic feedback phenomenon is compensated by an active noise control program.

The active noise attenuation method of FIG. 6 can only control the noise propagating in the plane wave, and the attenuation performance of the noise propagating in the higher order mode is weak.

7 is an active noise attenuation method using non-directional speakers, and the acoustic feedback phenomenon is compensated by a feedback compensation filter 71, and two main control filters 81 and 82 having an Infinite Impulse Response (IIR) filter form are shown. use.

In this method, only the control of the noise propagating through the plane wave is possible, and since the control filter has an IIR filter type, the stability of the active noise attenuator is not guaranteed due to the characteristics of the IIR filter.

As shown in FIG. 6 and FIG. 7, the conventional active noise attenuation method is limited to the control of noise propagated by plane waves.

A method of synthesizing the active noise attenuation system of FIG. 7 in parallel has been published for the control of the noise in the higher order mode. However, this method uses the same number of error sensors and offset generators, and each offset generator. Is controlled only by the error sensor and detection sensor that are paired with it.

In this case, it is difficult to solve cross-coupling between the error sensors and the offset acoustic wave generators that are not paired with each other, thereby degrading active noise attenuation performance.

In the active noise control method of the present invention, since the signals of all the error sensors collectively control the active noise attenuation filter, such performance degradation does not occur.

An object of the present invention is to improve the problems of passive noise attenuation countermeasures and conventional offset active noise attenuation countermeasures as mentioned above, using one or more detection sensors, error sensors and offset generators. The present invention provides an active noise attenuation device in a duct and an applied acoustic environment that can remove noise transmitted in a plane wave and a higher order mode in an applied acoustic environment by improving the engine measurement and controllability of the present invention. The present invention is described in detail as follows.

As shown in FIG. 1, the apparatus includes: a detection sensor 40 for detecting input noise 21 from a noise source flowing into an applied acoustic environment 10 of an active noise attenuation system; First and second offset sound generators 61 and 62 for outputting a predetermined sound wave to cancel the input noise 21 to minimize the output noise 22; First and second acoustic feedback compensation filters (71, 72) for compensating for the phenomenon that the offset sound waves generated by the first and second offset sound generators (61, 62) are fed back to the detection sensor (40); A first subtractor (91) for subtracting the acoustic feedback compensation signal output from the first acoustic feedback compensation filter (71) from the input noise detection signal (101) output from the detection sensor (40); A second subtractor (92) for subtracting the acoustic feedback compensation signal output from the second acoustic feedback compensation filter (72) from the output signal of the first subtractor (91); First and second main control filters 81 and 82 for receiving predetermined signals output from the second subtractor 92 and outputting offset sound signals; An error path compensation filter group 37 for receiving an output signal of the second subtractor 92 and modeling an acoustic environment applied to an active noise attenuation system; An error sensor 51.52 for detecting a portion of the input noise 21 not canceled by offset sound waves, a predetermined signal output from the error path compensation filter group 30, and an output from the error sensors 51 and 52; The LMS adaptation algorithm 90 which controls the first and second main control filters 81 and 82 by receiving the error signal to be inputted is interconnected.

In addition, the acoustic feedback compensation filter (71, 72), the main control filter (81, 82), the error path compensation filter group 30, the subtractor (91, 92) and the LMS adaptation algorithm (90) are the apparatus of the present invention. The program is embodied and stored in a digital signal processing professional microprocessor to be used for, which is collectively referred to as an active noise attenuator controller 100.

In addition to the components used to implement the device of the present invention, as shown in Figure 2 pre-amplifier 110 for amplifying the small signal output from the input noise detection sensor 40 and the error sensor (51, 52), respectively Wow ; An anti-aliasing filter 120 for removing unnecessary high frequency components from the output signal of the preamplifier 110; An A / D converter 130 for converting an analog signal output from the antialiasing filter 120 into a predetermined digital signal; An offset sound wave is generated by analyzing a predetermined signal output from the A / D converter 130 using a program stored in the program memory 150 and an input noise detection signal and various filter coefficients stored in the data memory 160. A digital signal processing specialized microprocessor 140 for outputting a predetermined digital signal for making a digital signal; A D / A converter 170 for converting the digital offset sound signal output from the digital signal processing microprocessor 140 into an analog signal; A reconstruction filter 180 for removing high frequency components included in the analog offset sound waves output from the D / A converter 170; And a power amplifier 190 for driving the offset sound generators 61 and 62 by amplifying the signal output from the regenerative generation filter 180.

On the other hand, the error path compensation filter group 30 is made by connecting a plurality of error path compensation filters 31 to 34 in parallel as shown in FIG. 5, and the LMS adaptation algorithm 90 performs the error path compensation. A plurality of multipliers 131 to 134 comparing the output signals 161 to 164 output from the filter group 30 and the error signals 105 and 106 detected by the error sensors 51 and 52; Summers 141 and 142 for summing output signals of the multipliers 131 to 134; The amplifiers 151 and 152 which amplify the output signals of the summers 141 and 142 to provide coefficient change rates of the first and second main control filters 81 and 82 are interconnected.

Referring to Figure 1 the effect of the device of the present invention having such a configuration as follows.

First, since the active noise attenuation principle of the present invention cancels the input noise by artificially generating sound waves having the same amplitude as the input noise 21 and having a phase difference of 180 degrees, the input noise must be accurately measured.

Therefore, the present invention uses the detection sensor 40 installed near the noise source for the measurement of the input noise (21). However, in the signal measured by the detection sensor 40 during the operation of the active noise control device, not only the components of the day noise 21 radiated from the noise source but also the offset sound components radiated from the omnidirectional offset generators 61 and 62 are included. .

As described above, a phenomenon in which the offset sound wave propagates in the direction of the output noise 22 and performs the required noise canceling effect is also propagated toward the detection sensor and measured again by the detection sensor.

In addition, when the detection sensor measurement signal including the sound wave component emitted from the offset sound generator is directly transmitted to the active noise control main control filters 81 and 82, the output signal of the main control filter returns to the input of the main control filter. A feedback loop is formed that not only attenuates the performance of the active noise attenuator but also becomes unstable, leading to divergence of the control system.

Therefore, in the present invention, the acoustic feedback component is removed from the detection sensor measurement signal (101 in FIG. 1) using the acoustic feedback compensation filters 71 and 72 for modeling the acoustic feedback phenomenon to remove the acoustic feedback phenomenon. The feedback loop was prevented from forming.

That is, since the acoustic feedback compensation filters 71 and 72 are filters for modeling the transfer functions of the offset acoustic generators 61 and 62 and the detection sensor 40, the acoustic feedback components are removed from the subtractors 91 and 92.

The main control filters 81 and 82 then accept the detection sensor output signal 101 from which the acoustic feedback component has been removed as an input and synthesize the predetermined signals 103 and 104 of the waveform capable of canceling the input noise 21. To the offset acoustic generators 61 and 62.

At this time, the optimum state of the sound wave to be output from the offset sound generator (61, 62) in order to cancel the input noise 21 is a sound wave having the same amplitude as the input noise and the phase difference is 180 °, in this case the input noise is It will be completely offset by the offset noise and there will be no output noise.

However, in order to model a given acoustic environment, a complete cancellation may occur because various filters 30, 71, 72, 81, and 82 included in the active attenuation controller 100 cannot accurately represent the actual acoustic environment. In addition, in many practical applications, the acoustic environment and the characteristics of the input noise change over time, so that the active noise attenuation device of the present invention reconstructs the portion of the input noise that is not canceled by the offset sound wave to cope with such a change. The performance improvement of the active noise attenuation system was made by using adaptive filtering which detects (51, 52) and feeds back the error signal into the active noise attenuator.

As mentioned above, the main control filters 81 and 82 are configured as adaptive filters to cope with a change in a given acoustic environment and noise source characteristics over time, and have a form of finite impulse response (FIR) that ensures stability. .

This main control adaptive filter 81, 82 uses the input signal 102 of the main control filter which has passed through the error path compensation filter group 30 and the error signals 105, 106 from the error sensors 51, 52, The main control filter coefficients are designed to be adapted at every sample time so that the main control filter outputs the offset sound generating signals 103 and 104 which allow the sum of squares of the error sensor signals 105 and 106 to approach zero.

In addition, the adaptive filter technique used in the present invention is called a Filtered-x LMS adaptive filter technique, and FIG. 5 shows the adaptive filter technique of the present invention in detail.

That is, the comparison of the error signals 105, 106 and the error path compensation filter group output signals 161, 162, 163, and 164 is performed by a plurality of multipliers 131, 132, 133, and 134, respectively. The coefficients of the main control filters 81 and 82 are adapted every sample time since they are added to the main control filters 81 and 82 through the amplifiers 151 and 152 which adjust the rate of change of the main control filter coefficients after being summed by (141, 142). Will fluctuate.

As described above, according to the present invention, the noise generated in the applied acoustic environment is detected by at least one or more detection sensors and at least one error sensor, thereby making it possible to measure plane waves and high-order mode noise present in the applied acoustic environment, and offset sound waves. Is fired using at least one offset sonar generator, which enables the emission of plane waves and high-order mode noises, and improves the observability and controllability of the active noise attenuation system. Can attenuate

In other words, by providing a plurality of detection sensors, error sensors, and offset acoustic wave generators, it is possible to attenuate high-order modes including plane waves in an applied acoustic environment.

In addition, the filter coefficients of the acoustic feedback compensation filters 71 and 72 and the error path compensation filter 30 included in the active noise control unit 100 other than the main control filters 81 and 82 are respectively determined before the active noise canceller. As shown in FIG. 3 and FIG. 4, it is obtained by using an LMS adaptive system identification technique, and the obtained filter coefficient is stored and used in the data memory (160 in FIG. 2).

The acoustic feedback compensation filter is a transfer function for modeling the acoustic environment between the offset acoustic wave generators 61 and 62 and the detection sensor 40, and thus a white noise signal is generated using a noise signal generator (303 in FIG. 3). Generate and drive an offset acoustic wave generator with this white noise signal, measure the sound pressure at the detection sensor 40, and estimate the adaptive least-squares system given in equation (1) using the measured white noise signal and the detection sensor signal. Obtain the acoustic feedback compensation filter coefficients according to the algorithm.

f (n + 1) = f (n) + μ _F e (n) x (n)... … … … … … … … … … (One)

Where f (n) = acoustic feedback compensation filter coefficient at the nth sampling time

μ _F = adaptive filter coefficient adaptation rate

e (n) = modeling error signal (301 in FIG. 3)

x (n) = white noise signal (302 in FIG. 3)

The acoustic feedback compensation filter has a form of FIR filter which guarantees stability. The acoustic feedback compensation filter uses the filter coefficient stored in the data memory (160 in FIG. 2) and the offset acoustic signal (103 or 104 in FIG. 1). ) And an algorithm for outputting the acoustic feedback compensation signal.

Where y (k) = acoustic feedback compensation signal (output signal of the acoustic feedback compensation filter)

x (k) = offset sound generation signal (input signal of acoustic feedback compensation filter)

f (i) = coefficient of acoustic feedback compensation filter

L = length of the acoustic feedback compensation filter

In addition, the error path compensation filter 30 is a second filter for modeling an acoustic environment to which the active noise attenuation system is applied. The error path compensation filter 30 is used to filter the acoustic environment from each offset generator 61 and 62 to each error sensor 51 and 52. It has a digital FIR filter shape to model.

4 shows the coefficient estimation method of the error path compensation filter. The error path compensation filter is a transfer function for modeling an acoustic environment between the offset sound generators 61 and 62 and the error sensors 51 and 52, thereby generating a white noise signal using a noise signal generator (403 in FIG. 4). Drive the offset sonic generator with this white noise signal, measure the sound pressure at the error sensor 51 or 52, and estimate the adaptive least-squares system given in equation (3) using the measured white noise signal and the error sensor signal. Obtain the acoustic feedback compensation filter coefficients according to the algorithm.

c (n + 1) = c (n) + μ _c (e (n) x (n) …………………………… (3)

Where c (n) = error path compensation filter coefficient at the nth sampling time

μ _c = adaptive filter coefficient adaptation rate

e (n) = modeling error signal (401 in FIG. 4)

x (n) = white noise signal (402 in FIG. 4)

The error path compensation filter has a form of a FIR filter which guarantees stability, and is formed by using a filter coefficient stored in a data memory (160 in FIG. 2) and a detection sensor signal (102 in FIG. 1) from which acoustic feedback is removed. It consists of an algorithm that outputs the acoustic feedback compensation signal by performing the calculation given in (4).

Where g (k) = error path compensation filter output at the kth sampling time

h (k) = Error path compensation filter input at the kth sampling time

(Detection sensor signal through sound feedback compensation process)

C (i) = coefficient of error path compensation filter

M = depth of error path compensation filter (number of coefficients)

In addition, the preamplifier 110 in FIG. 2 amplifies a minute signal from the detection sensor 40 and the error sensors 51 and 52 into a state suitable for various signal processing in the future. The passed analog signal is converted into a digital signal through the A / D converter 130 and input to the digital signal processing microprocessor 140, wherein the analog signal output from the preamplifier is input to the A / D converter 130. Before passing through the anti-aliasing (aliasing) filter 120, the anti-aliasing filter to prevent the aliasing phenomenon occurs during the A / D conversion.

In addition, the digital signal input to the digital signal processing professional microprocessor 140 is analyzed by an active noise control program stored in the predetermined program memory 150 and the data memory 160, and then digitally generated for offset sound wave generation. Output the signal.

Thereafter, the offset sound digital signal output from the digital signal processing unit 140 is converted into an analog signal by the D / A converter 170, and regeneration is generated to remove unnecessary high frequency components generated when the digital signal is converted into an analog signal. Pass through the filter (Reconstruction Filter: 180).

The output of this regeneration filter is amplified by the power amplifier 190 so that the offset sonic generator 60 can be operated and sent to the offset sonic generator.

As described above, according to the present invention, noise generated in an applied acoustic environment such as a duct is sensed by at least one detection sensor, and generates sound waves having the same amplitude and different phases by 180 ° through at least one offset sound generator. At the same time, the error signal detected by at least one error sensor is fed back and applied to at least one offset acoustic wave generator, thereby improving the observability and controllability of the active noise attenuation system. It is a useful invention that can attenuate even the noise transmitted in the mode.

Claims (2)

At least one detection sensor 40 for detecting input noise 21 from a noise source flowing into the applied acoustic environment 10; At least one offset sound generator 61, 62 which receives a predetermined signal from the main control filters 81 and 82, respectively, and outputs a predetermined sound wave to cancel the input noise 21 to minimize the output noise 22. )Wow ; At least one acoustic feedback compensation filter (71,72) for compensating for the phenomenon that the offset sound waves generated by the offset sound generators (61, 62) are fed back to the detection sensor; A subtractor 91 for subtracting an acoustic feedback compensation signal output from the acoustic feedback compensation filter 71 from a predetermined input noise detection signal 101 output from the detection sensor; A subtractor 92 for subtracting the acoustic feedback compensation signal output from the acoustic feedback compensation filter 72 from the output signal of the subtractor 91; At least one main control filter (81,82) for receiving a predetermined signal output from the subtractor (92) and outputting an offset sound wave signal; An error path compensation filter group 30 which receives a predetermined signal output from the subtractor 92 and increases the stability of the main control filter adaptive algorithm; At least one error sensor (51, 52) installed at a rear side of the offset sound generator (61, 62) for detecting a portion of the input noise not canceled by the offset sound wave; A main control filter LMS adaptation algorithm for controlling the main control filters 81 and 82 by receiving a predetermined signal output from the error path compensation filter group 30 and an error signal detected by the error sensors 51 and 52, respectively. 90) The active noise attenuation device in the duct and applied acoustic environment, characterized in that to cancel the input noise by interconnecting. 2. The active noise attenuation device according to claim 1, wherein the error path compensation filter group (30) is formed by connecting a plurality of error path compensation filters (31 to 34) in parallel with each other.