JP2923476B2 - Adaptive active silencer - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an active silencer which silences noise by radiating sound waves having the same sound pressure and opposite phase in the propagation path of noise in the duct, and in particular, the control is deteriorated due to various causes. The present invention relates to an adaptive active silencer that appropriately takes measures against abnormalities that occur.

[0002]

2. Description of the Related Art Conventionally, as a method of silencing noise propagating in a duct, a passive silencing method such as a method of absorbing sound by a sound absorbing material adhered to the duct has been adopted. There are problems such as.

On the other hand, active silencers have been actively studied which simultaneously radiate sound waves having the same sound pressure and opposite phase to the sound waves of the propagation noise in the duct into the duct, and muffle the sound by interference of both sound waves. .

FIG. 6 is a structural explanatory view showing a conventional active silencer. That is, the noise propagating in the duct 1 is transmitted by the first microphone 2 to the first microphone input signal x (t).
Is detected, and the first microphone input signal x (t) is input to the adaptive digital filter 4 and the adaptive control algorithm 5 of the sound elimination sound generation circuit 3.

On the other hand, a second microphone input signal e (t) is detected by the second microphone 6, and the second microphone input signal e (t) is input to the adaptive control algorithm 5 of the sound-muffling sound generation circuit 3.

The sound-absorbing sound wave generating circuit 3 does not simply generate an electric signal for radiating sound waves having the same sound pressure and opposite phase with respect to the propagation noise in the duct 1 obtained from the first microphone 2. The following calculation is performed in order to generate a noise-canceling signal so that an optimal sound-canceling sound wave can be emitted at that time with respect to the sound wave of the noise in the duct 1 at the place where the sound-canceling speaker 7 is installed.

(A) In the system, when a sound wave propagates in the duct 1 or a signal passes through the circuit, the transfer function is affected by a unique transfer function such as a time delay or a change in frequency characteristics. Is performed.

(B) Based on the adaptive control algorithm 5 from the first microphone input signal x (t) of the first microphone 2 and the second microphone input signal e (t) of the second microphone 6, the second microphone 6 2 microphone input signal e
A coefficient is calculated so that (t) approaches zero (that is, noise is effectively suppressed). The coefficient is used as a filter coefficient of the adaptive digital filter 4 to perform an operation of convolving the first microphone input signal x (t) from the first microphone 2 to generate a noise-elimination electric signal.

The silencing electric signal generated by the silencing sound wave generating circuit 3 drives the silencing speaker 7, and the silencing speaker 7 converts the silencing electric signal into a silencing sound wave and radiates it into the duct 1. .

As described above, the noise propagating in the duct 1 is simultaneously radiated into the duct 1 with the noise canceling sound having the same sound pressure and the opposite phase to the sound propagating in the duct 1 so as to cause the two sound waves to interfere with each other. I do.

An active silencer that muffles the propagation noise in the duct 1 by radiating a muffling sound wave from the muffling speaker 7 generates an anti-phase sound based on the adaptive control algorithm 5. The adaptive control may diverge due to changes in noise or external noise.

In this case, an abnormal sound is generated from the sound deadening speaker 7, and a loud sound is propagated downstream of the sound deadening device.
The silencer may be a noise source. Since the muffling sound wave generation circuit 3 itself cannot self-diagnose the control abnormality, it is necessary to provide another function of avoiding the abnormality. However, in the conventional abnormality detection method, it is necessary to detect what abnormality is occurring. And could not be properly treated.

[0013]

SUMMARY OF THE INVENTION The present invention has been made in view of the above circumstances, and calculates input power of two microphones and output power of a speaker to detect what kind of abnormality has occurred. It is an object of the present invention to provide an adaptive active silencer that performs an appropriate process and restores normal control again.

[0014]

In order to achieve the above object, an adaptive active silencer according to the present invention comprises a first microphone for detecting a propagation noise in a duct, and a radiating sound wave in the duct. Active muffling comprising a speaker, a second microphone for detecting a muffling situation in the duct, and a controller to which input signals of the first microphone and the second microphone are applied and which calculates muffling sound waves radiated from the speaker. calculated in device, a first calculation means for calculating an input power P ₁ of the first microphone, a second calculation means for calculating an input power P ₂ of the second microphone, the output power P _out of the speaker a third calculating means for, the input power P ₂ and the ratio P of the P ₁
Second comparison means for detecting an abnormality in comparison with the first comparison means the reference value or the output power P _out, which detects an abnormality of ₂ / P ₁ is compared with a reference value, the first comparison And an abnormality processing unit for processing the abnormality when the abnormality is detected by the comparison unit or the second comparison unit.

In the adaptive active silencer according to the present invention, the first calculating means, the second calculating means, and the third
Is characterized by performing digital operation using the following equation.

Σ (t) = αz ² (t) + (1−α) σ
(T-1) Here, z (t) is a signal whose power is to be obtained, σ (t) is power P ₁ , P ₂ , P _out , and α is a forgetting coefficient.

[0017]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Embodiments of the present invention will be described below in detail with reference to the drawings. FIG. 1 is a structural explanatory view showing an adaptive active silencer according to an embodiment of the present invention, and FIG.
Is a characteristic diagram showing an example of a short-term abnormality of the control according to the present invention, and FIG. 3 is a characteristic diagram showing an example of a long-term abnormality of the control according to the present invention. In the figure, reference numeral 1 denotes a duct, 2 denotes a first microphone for detecting a propagation noise in the duct 1, and 3 denotes a muffling sound wave generating circuit of a controller for calculating muffling sound waves. As shown in FIG. It comprises an adaptive control algorithm 5. 6 is a second microphone for detecting a muffling situation, 7 is a muffling speaker that emits a muffling sound wave, 10 is an abnormality determination processing circuit, 11 is a power calculation circuit,
12 is a power calculation circuit, 13 is a comparison circuit, 14 is an abnormality counter, 15 is an abnormality processing circuit, 16 is a power calculation circuit,
17 is a comparison circuit, 18 is an abnormality counter, 19 is a power calculation circuit, 20 is a comparison circuit, 21 is an abnormality counter, 22 is a timer, 23 is a periodic inspection circuit, 24 is an A / D (analog / digital) converter, 25 Is D / A (digital /
An analog (A / D) converter 26 is an A / D (analog / digital) converter.

In such an adaptive active silencer, the first microphone input signal x (t) of the first microphone 2 and the second microphone input signal e (t) of the second microphone 6 are calculated as follows.
The coefficient w (i) of the adaptive digital filter 4 of the noise elimination sound generation circuit 3 is updated at the rate of the convergence coefficient μ based on the equation (1). w _n (i) = w ₀ (i) + μe (t) × (t−i) (1) Therefore, the amount of the coefficient w (i) of the adaptive digital filter 4 to be updated is It depends on the magnitude of the first microphone input signal x (t), the second microphone input signal e (t), and the convergence coefficient μ. The first microphone input signal x (t) and the second microphone input signal e (t) vary depending on the magnitude of the noise radiated from the noise source, and the convergence coefficient μ is used as a control parameter at the time of starting the control. Set according to one microphone input signal x (t). When the set value of the convergence coefficient μ is large with respect to the magnitudes of the first microphone input signal x (t) and the second microphone input signal e (t), the update amount of the coefficient w (i) of the adaptive digital filter 4 becomes large. And the adaptation speed increases, but diverges if the convergence coefficient μ is too large. On the other hand, if the convergence coefficient μ is too small, on the contrary, the adaptation speed becomes slow, and it becomes impossible to follow the acoustic change in the duct 1.

The present invention is characterized in that the following three abnormalities are distinguished and detected, and appropriate measures are taken. [Short-term abnormalities]: A sudden deterioration of the sound reduction effect caused by sudden divergence of control due to the incorporation of external disturbances and the like, and abnormal sound emission from the speaker 7.

For example, consider that the operating condition of the fan of the duct 1 has changed and the noise has increased. As shown in FIG. 2, the first microphone input signal x (t) of the first microphone 2
Rises from 60 [dB] to 80 [dB], the second microphone input signal e (t) of the second microphone 6 also becomes 45 [dB].
B] to 65 [dB].

In equation (1), the first microphone input signal x
(T), since both the second microphone input signal e (t) become large, the adaptive digital filter 4 of the muffling sound generation circuit 3
The update amount of the coefficient w (i) becomes too large, and the control diverges rapidly as shown in FIG. At this time, a very loud sound is radiated from the speaker 7, and a loud sound than the original additional noise source may propagate to the downstream side.

[Long-term abnormality] A gradual deterioration of the sound reduction effect caused by the inability of adaptive control to follow subtle changes in the duct 1 system. For example, consider that the temperature in duct 1 has changed and the speed of sound in duct 1 has changed.

As shown in FIG. 3, the second microphone input signal e (t) of the second microphone 6 is controlled even though the first microphone input signal x (t) of the first microphone 2 hardly changes. Without adaptation, the sound volume decreases slowly.

When the speed of sound changes, the time required for the noise to reach the speaker 7 from the first microphone 2 changes.
It is necessary to update the coefficient w (i) of the adaptive digital filter 4 of the muffling sound generation circuit 3. However, if the convergence coefficient μ is too small with respect to the magnitudes of the first microphone input signal x (t) and the second microphone input signal e (t), the update amount is small and follows the sound speed in the duct 1 quite easily. Will not do. In this case, the noise reduction effect gradually deteriorates.

[Abnormal output] An abnormal sound emission from the speaker 7 that occurs when the output power P _out of the speaker 7 exceeds its allowable amount even if the control is performed normally. Hereinafter, an example of the control abnormality processing will be described with reference to FIG. 1 in terms of power measurement, abnormality determination, abnormality processing, and periodic inspection.

First, the measurement of power will be described. That is, the power calculation circuit 11 is provided immediately after the A / D converter 24, and the power calculation circuit 1 is provided immediately after the A / D converter 26.
2 and 16 are provided, and a power calculation circuit 19 is provided immediately before the D / A converter 25. Each of the power calculation circuits 11, 1
2, 16 and 19 each perform a moving average by digital operation based on the equation (2) using a primary IIR filter. σ (t) = αz ² (t) + (1−α) σ (t−1) (2) In equation (2), z (t) is a signal for which power is to be obtained. ,
σ (t) is the calculated signal power (P ₁ , P _2sht , P _2lng , P
_out ), α is a forgetting factor (Forgetting F
By setting the value of the forgetting coefficient α appropriately, a sudden power change or a slow power change can be detected. For example, as the input power of the second microphone 6, P _2sht (α is almost 1) and P _2lng (α is almost 0) are separately calculated to detect a sharp control divergence and a gradual control divergence separately. Can be performed.

The first microphone input signal x (t) of the first microphone 2 is applied to the power calculation circuit 11 via the A / D converter 24, and the input power P ₁ of the first microphone 2 is input.
Is calculated.

The power calculation circuit 12 is connected to the second microphone 6
Of the second microphone 6 is input via the A / D converter 26 and the input short-term power P
_{Calculate 2sht} .

The power calculation circuit 16 includes the second microphone 6
Is input via the A / D converter 26, and the input long-term power P of the second microphone 6 is
_{Calculate 2lng} .

The power calculating circuit 19 calculates the output power P _out of the speaker 7 to which the sound-absorbing sound signal from the sound-absorbing sound wave generating circuit 3 is added. Next, the determination of the abnormality will be described.

That is, control abnormalities can be detected by comparing the values of P _2sht / P ₁ and P _2lng / P ₁ , which are the power ratios of the sound reduction by the active silencer, with the reference values of the sound reduction. . Further, by comparing the output power P _out of the speaker 7 with a reference value such as the allowable output of the speaker 7, an abnormal sound output of the speaker 7 can be detected.

The comparison circuit 13 receives the input power P ₁ from the power calculation circuit 11 and the input short-term power P _2sht from the power calculation circuit 12, and compares P _2sht / P ₁ with the short-term abnormal reference value P _{0sht. P} 2sht / P _1> to determine short-term abnormal reference value is a rapid deterioration of the control by correctness of P _0Sht short abnormal Te.

The comparison circuit 17 receives the input power P ₁ from the power calculation circuit 11 and the input long-term power P _2lng from the power calculation circuit 16 and compares P _2lng / P ₁ with the long-term abnormality reference value P _{0lng. P} 2lng / P _1> to determine long-term abnormal reference value is a gradual worsening of the control by correctness of P _0Lng long abnormal Te.

The comparison circuit 20 receives the output power P _out of the speaker 7 from the power calculation circuit 19 and
The output power P _out and compared with the speaker abnormal reference value P _0Out is P _out> determines speaker abnormal output in the form of the output abnormal by correctness of the speaker abnormal reference value _P 0out.

Next, the processing of the abnormality will be described. That is, abnormality processing is performed by combining the following three processings according to the type of abnormality.

(A) Re-identification of Transfer Function Required for Control If the transfer function required for control has changed, an abnormality occurs again even if control is restarted. Therefore, the transfer function required for the control is identified again.

(B) Resetting of the convergence coefficient μ The short-term abnormality occurs when the convergence coefficient μ is too large and the adaptive control diverges. Therefore, the convergence coefficient μ is reduced.

The long-term abnormality occurs when the convergence coefficient μ is too small to perform the adaptive control in time. Therefore, the convergence coefficient μ is increased. (C) The scale-down of the adaptive digital filter The abnormal output may be caused by any cause.
Occurs when the adaptive digital filter 4 becomes too large. Therefore, the adaptive digital filter 4 is scaled down.

When the result of the comparison by the comparison circuit 13 is positive and a short-term abnormality is determined, the abnormality counter 14 is incremented, and the abnormality processing circuit 15 re-identifies and controls the transfer function necessary for the control according to the short-term abnormality. The convergence coefficient μ is increased.

When the result of comparison by the comparison circuit 17 is positive and a long-term abnormality is determined, the abnormality counter 18 is incremented, and the abnormality processing circuit 15 re-identifies and controls a transfer function required for control according to the long-term abnormality. The convergence coefficient μ is reduced.

When the result of the comparison by the comparison circuit 20 is positive and the output is determined to be abnormal, the abnormality counter 21 is incremented. Further, the abnormality processing circuit 15 outputs the adaptive digital filter 4 of the muffling sound wave generating circuit 3 according to the output abnormality. The coefficient w (i) is scaled down, and the transfer function required for control is re-identified.

The abnormality processing circuit 15 is provided with an abnormality counter 1
When the number of abnormalities exceeds the allowable number of times, the operations of the sound-muffling sound wave generating circuit 3 and the fan of the duct 1 are stopped. Next, the periodic inspection will be described.

That is, the periodic check circuit 23 is periodically operated from the timer 24, and re-identification of a transfer function required for control, control check such as resetting of the control convergence coefficient μ, and the like are performed.
FIGS. 4A, 4B, and 4C are characteristic diagrams showing an example of abnormal processing of a short-term abnormality according to the present invention, which is an abnormal processing when a sudden divergence occurs due to an increase in additional sound sources.

FIG. 4A shows an example of the input levels of the first microphone input signal x (t) of the first microphone 2 and the second microphone input signal e (t) of the second microphone 6, and FIG.
4B shows an example of the calculated signal powers P ₁ , P _2sht , and P _2lng of the power calculation circuits 11, ₁₂ , and _16. FIG. 4C shows that the comparison circuit 13 uses P _2sht / P ₁ and a short-term abnormal reference value. P
_0sht and is compared _P 2sht / P _1> short abnormal reference value P
_{When 0sht} is positive and a short-term abnormality is determined, short-term abnormality detection is performed, and the abnormality processing circuit 15 performs re-identification of a transfer function necessary for control according to the short-term abnormality and abnormal processing of _increasing the control convergence coefficient μ. Resume control. In this case, it is determined that P _2lng / P ₁ > long-term abnormality reference value P _0lng in the comparison circuit 17 is erroneous and is not a long-term abnormality.

FIGS. 5 (a), 5 (b) and 5 (c) are characteristic diagrams showing an example of abnormal processing for a long-term abnormality according to the present invention, which is performed when the sound speed in the duct 1 changes due to a temperature change or the like. It is.

FIG. 5A shows an example of the input levels of the first microphone input signal x (t) of the first microphone 2 and the second microphone input signal e (t) of the second microphone 6, and FIG.
5B shows an example of the calculated signal powers P ₁ , P _2sht , and P _2lng of the power calculation circuits 11, ₁₂ , and _16. FIG. 5C shows that the comparison circuit 17 _calculates P _2lng / P ₁ as a long-term abnormal reference value. P
_0lng and is compared _P 2lng / P _1> Long abnormal reference value P
_{When 0lng} is positive and determined to be a long-term abnormality, the long-term abnormality is detected, and the abnormality processing circuit 15 performs re-identification of a transfer function necessary for control in accordance with the long-term abnormality and abnormal processing of _decreasing the control convergence coefficient μ. Resume control. In this case, it is determined that P _2sht / P ₁ > short-term abnormality reference value P _0sht in the comparison circuit 13 is incorrect and not a short-term abnormality.

[0047]

As described above, according to the present invention, the input power of the two microphones and the output power of the speaker are calculated, and what kind of abnormality has occurred is detected and appropriate processing is performed. It is possible to provide an adaptive active silencer that returns to normal control again.

[Brief description of the drawings]

FIG. 1 is a configuration explanatory view showing an adaptive active silencer according to an embodiment of the present invention.

FIG. 2 is a characteristic diagram showing an example of a short-term abnormality of control according to the present invention.

FIG. 3 is a characteristic diagram showing an example of a long-term abnormality of control according to the present invention.

FIG. 4 is a characteristic diagram illustrating an example of a short-term abnormality abnormality process according to the present invention.

FIG. 5 is a characteristic diagram showing an example of a long-term abnormality abnormality process according to the present invention.

FIG. 6 is a configuration explanatory view showing a conventional active silencer.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Duct 2 1st microphone 3 Sound canceling sound generation circuit 4 Adaptive digital filter 5 Adaptive control algorithm 6 2nd microphone 7 Sound canceling speaker 10 Abnormality judgment processing circuit 11 Power calculation circuit 12 Power calculation circuit 13 Comparison circuit 14 Abnormal counter 15 Abnormal processing Circuit 16 Power calculation circuit 17 Comparison circuit 18 Abnormal counter 19 Power calculation circuit 20 Comparison circuit 21 Abnormal counter 22 Timer 23 Periodic inspection circuit 24 A / D (analog / digital) converter 25 D / A (digital / analog) converter 26 A / D (analog / digital) converter

──────────────────────────────────────────────────続 き Continued on the front page (58) Field surveyed (Int.Cl. ⁶ , DB name) G05B 9/02 G10K 11/16 G10K 11/178

Claims (2)

(57) [Claims] A first microphone that detects noise propagated in the duct; a speaker that emits a sound-absorbing sound wave in the duct; a second microphone that detects a sound-muffling state in the duct; in a more becomes active silencer and a controller for calculating the sound-absorbing sound waves radiated from the input signal is applied the speaker of the second microphone, the calculated input power P ₁ of the first microphone 1
Calculating means for calculating the input power P ₂ of the second microphone.
And calculating means, first to detect a third calculating means for calculating an output power P _out of the speaker, the abnormality is compared with a reference value the ratio P ₂ / P ₁ of the input power P ₂ and P ₁ Or a second means for detecting an abnormality by comparing the output power P _out with a reference value.
An adaptive active silencer comprising: a comparison unit; and an abnormality processing unit that processes the abnormality when an abnormality is detected by comparing the first comparison unit or the second comparison unit. 2. The adaptive active silencer according to claim 1, wherein the first calculating means, the second calculating means, and the third calculating means perform digital operation using the following equation. σ (t) = αz ² (t) + (1−α) σ (t−1) where z (t) is a signal for which power is to be obtained, and σ (t) is power P ₁ , P ₂ , P _out , Α is the forgetting factor.