Acoustic control method and device for controlling propagation of acoustic waves in the vicinity of a wall

Abstract

The present invention relates to an acoustic control method, comprising a step a) in which a number Nc of cells are placed on a wall, these comprising a control unit, a power supply, a set of Nm microphones, and a loudspeaker, each cell being such that the loudspeaker is driven by the control unit depending on signals measured by the microphones, a step b) in which each microphone of each cell measures the acoustic pressure of the acoustic waves, the voltage across the terminals of the loudspeaker of each cell also being measured, a step c) in which the control unit estimates the acoustic pressure and / or its tangential spatial derivatives at the loudspeaker, then applies a control law defining, from the voltage measurement and the estimation of the acoustic pressure and / or its tangential spatial derivatives, the amplitude of the electrical signal sent to the loudspeaker so as to determine a generalised acoustic impedance Zdet for the loudspeaker, and a step d) in which a master device controls all of the control units iteratively to adjust the generalised acoustic impedance Zdet for each cell.

Description

Acoustic control method and device for controlling the propagation of acoustic waves in the vicinity of a wall

[0001] The present invention relates to a method and device for controlling the propagation of acoustic waves in the vicinity of a wall.

[0002] Reducing noise pollution from transportation and human activity has become a major challenge. The use of passive coatings (foams, Helmholtz resonators, etc.) in buildings or vehicles does not always allow for adapting acoustic absorption to the sound waves in the vicinity of these coatings. Weight and size constraints often render these devices (active and passive) ineffective, particularly for low frequencies.

[0003] Acoustic treatment techniques are generally based on the use of absorbent materials such as foam or structured cellular materials. For example, in certain building or transportation applications, acoustic liners with a distribution of Helmholtz resonators at low and mid frequencies and foam at high frequencies are used.

[0004] The reduction achieved remains below a few decibels at low frequencies. The effectiveness of conventional sound-absorbing treatments is linked to the thickness of the materials, and therefore limited by size and added mass, not to mention the problems of water and pollutant absorption in these porous materials. All these techniques are passive and offer no adaptive or selective noise treatment capabilities. Nor do they offer the ability to direct emissions.

[0005] Active noise control techniques were developed as early as the 1980s to address these technological challenges, and their applications span fields as diverse as consumer audio and transportation. However, the size and energy consumption of acoustic treatment systems limit their effectiveness for many potential applications.

[0006] Application WO 2021 / 122394 describes a method and device for controlling the propagation of acoustic waves near a wall. The device comprises a master device for controlling a set of cells consisting primarily of a control unit, a loudspeaker connected to the control unit, and a set of microphones connected to the control unit. The control is achieved by means of control laws that determine the intensity of the electrical signal to be sent to each loudspeaker based on the signals measured by the microphones. This device makes it possible to obtain a target generalized acoustic impedance for each loudspeaker, such that a fraction of the acoustic waves is absorbed by the diaphragm of each loudspeaker.

[0007] However, the method and device described in application WO 2021 / 122394 do not sufficiently adapt to changes in the acoustic environment, particularly temperature variations. The electrical components of the device described in application WO 2021 / 122394 are susceptible to aging and malfunctioning, potentially disrupting the generalized impedance determination. The method described in application WO 2021 / 122394 is prone to error accumulation and may therefore generate noise instead of absorbing sound waves. Error accumulation is more likely to occur because the method is iterative and the generalized acoustic impedance is determined in an open loop. In particular, the method described in application WO 2021 / 122394 is prone to generating errors at both low and high frequencies.

[0008] There is therefore a need for an acoustic control process and device that is more adaptable to changes in the acoustic environment and / or to the aging of acoustic cell components, in particular by limiting errors and especially the accumulation of errors in determining an acoustic impedance.

[0009] The invention proposes a method and a device for locally, iteratively and adaptively controlling the generalized acoustic impedance of a loudspeaker to efficiently absorb acoustic waves with low energy consumption.

[0010] As a reminder, acoustic impedance is a common and well-known physical quantity that corresponds to the ratio between sound pressure and sound velocity. The generalized acoustic impedance of the loudspeaker is defined as the impedance of the loudspeaker, determined by the control method according to the invention, and designed to handle as efficiently as possible the acoustic waves causing the loudspeaker diaphragm to vibrate. The objective is to create an acoustic silencer, that is, to prevent sound from escaping. This can be achieved by absorbing / dissipating or reflecting the wave, so that, in both cases, it does not reach the user's ears.

[0011] The device comprises a first set of acoustic transducers, each consisting of microphones and a loudspeaker. A second set of the device consists of the electronic signal processing and real-time control components.

[0012] The device is cellular, with each cell integrating a speaker, microphones, and the electronics for processing and managing signals.

[0013] Regarding the process, each cell is autonomous and executes a control law whose parameters can be determined and updated via an integrated interface. This interface allows management of the cell network and access to the inputs and outputs of the entire system.

[0014] Similarly, the power supply to the device is distributed throughout all the components. The invention relates more specifically to the modular nature of the system.

[0015] The invention relates to a method for acoustically controlling the propagation of acoustic waves in the vicinity of a wall, the control method comprising: - a step a) in which a number Nc, at least equal to 1, of cells is affixed to the wall, each cell comprising a control unit, a power supply, a set of Nm microphones connected to the control unit, and a loudspeaker connected to the control unit; - a step b) in which each microphone of each cell measures the acoustic pressure of the acoustic waves, these acoustic pressure measurements being returned to the cell's control unit, the voltage across the loudspeaker of each cell also being measured and this voltage measurement also being returned to the cell's control unit; - a step c) in which the cell's control unit estimates the acoustic pressure and / or its tangential spatial derivatives at the level of the cell's loudspeaker.then applies a control law, determined from the voltage measurement and the estimation of the sound pressure and / or its tangential spatial derivatives, which sets the intensity of the electrical signal sent to the loudspeaker in such a way as to determine a generalized acoustic impedance Zdet for the loudspeaker, - a step d) in which the cell control unit sends the electrical signal to the loudspeaker so that a fraction of the acoustic waves is processed by absorption and / or reflection by the loudspeaker diaphragm, a master device iteratively controlling all the control units to adjust the generalized acoustic impedance Zdet for each cell.

[0016] According to the invention, the control unit, in step c) estimates either the sound pressure at the loudspeaker, or its spatial derivative, or both.

[0017] The acoustic control method according to the invention is better adapted to changes in the acoustic environment and / or to the aging of acoustic cell components, particularly due to the voltage measurement across each loudspeaker. Specifically, the acoustic control method is better adapted to temperature changes.

[0018] The acoustic control method according to the invention improves the absorption of sound waves near a wall, notably by taking into account a voltage measurement across each loudspeaker to determine a generalized acoustic impedance. The determination of the generalized acoustic impedance, for example, better accounts for changes in the acoustic environment and / or the aging of acoustic cell components.

[0019] In particular, by using both a voltage measurement across the speaker terminals of each cell and a sound pressure measurement, the determination of a generalized acoustic impedance is improved compared to a generalized acoustic impedance determined solely from a sound pressure measurement. For example, measuring the voltage across each speaker terminal allows for a more precise estimation of the speaker diaphragm velocity and, consequently, the speaker's acoustic impedance.

[0020] The acoustic control method limits errors, and especially error accumulation, in determining acoustic impedance, particularly due to the voltage measurement across each loudspeaker. For example, the acoustic control method is more accurate at both low and high frequencies. Errors in determining the generalized acoustic impedance are also detected more quickly. Error correction between a setpoint impedance and the actual impedance is facilitated. Error accumulation during the implementation of the acoustic control method to determine generalized acoustic impedance is limited, even with iterations of open-loop generalized acoustic impedance determination. In particular, the acoustic control method prevents the generation of noise from each cell due to error accumulation during generalized acoustic impedance determination.

[0021] By measuring the voltage across each loudspeaker and minimizing the risk of error accumulation, the overall acoustic impedance of each cell can be determined within each cell or within each group of cells. Each cell or group of cells is capable of self-calibrating and self-adjusting over time and despite variations in the acoustic environment.

[0022] The control method according to the invention is implemented with cells having a small footprint and which are particularly effective at attenuating low frequency noise compared to a passive coating.

[0023] Using an iterative process, the control law parameters are adjusted for each cell as long as the insertion loss remains below a predetermined threshold. Once the threshold is reached, step c) is applied, which optimizes the control law parameters to obtain the generalized acoustic impedance Zdet for the loudspeaker. This generalized acoustic impedance Zdet is then adapted to absorb acoustic waves near the loudspeaker.

[0024] Optional features of the invention, complementary or alternative, are set out below.

[0025] According to one embodiment, the control method is a method of acoustic wave absorption by the cells. The generalized acoustic impedance Zdet is determined in such a way as to optimize the absorption and / or reflection of acoustic waves by the cell speakers. Preferably, the speakers can be configured to absorb spectrally different acoustic waves, for example, to effectively dampen tonal and broadband noise. The optimization method measures the effect of the impedance setpoint (corrected for errors) on the insertion loss (IL), i.e., the acoustic attenuation caused by the system. The target generalized acoustic impedance Zdet comprises a so-called "local" component based on the pressure in front of the cell and a so-called "distributed" component based on the pressure gradient. This impedance is expressed as a transfer function.Typically, the initial target impedance is a fraction of the impedance corresponding to an air layer.

[0026] According to certain characteristics, the acoustic control method according to the invention is a method of processing, by absorption and / or reflection of acoustic waves by the cells (1), the generalized acoustic impedance Zdet being a predetermined target value to maximize the processing by absorption and / or reflection of acoustic waves by the loudspeakers (11) of the cells (1).

[0027] According to certain characteristics, the control process implements a control loop which includes the following steps: Start, A1: loading of a generic acoustic model, A2: assignment of a control law to at least one of the cells, A3: calculation of the parameters associated with the control law, A4: application of the control law to the cell, A5: generation of a reference signal, A6: acquisition of the signal by the microphones, A7: calculation of the insertion loss IL, A8: comparison of the calculated insertion loss IL with a predetermined insertion loss value IL0 corresponding to the determination of the generalized acoustic impedance Zdet, A9: return to A3 to adapt the parameters of the control law to minimize the error on the measured impedance and maximize the insertion loss if IL is strictly less than IL0, or end of the control process if IL is greater than IL0.

[0028] Acoustic impedance is based on a direct measurement of the pressure and pressure gradient in front of the cell. However, it relies on an estimation of the diaphragm velocity derived from a theoretical loudspeaker model. This velocity estimation by a model is sensitive to environmental variations, particularly temperature, and consequently generates errors between the target impedance in the setpoint and the impedance actually achieved by the system's diaphragms. Using voltage provides a direct measurement of the velocity, eliminating these errors and thus allowing the system to accurately achieve the target acoustic impedance. The optimization process measures the effect of the impedance setpoint (corrected for errors) on insertion losses (IL), i.e., the acoustic attenuation caused by the system.

[0029] According to certain characteristics, the control process implements a control loop which includes the following steps: Start, A1: loading a generic acoustic model, A2: assigning a control law to at least one of the cells, A3: calculating the parameters associated with the control law, A4: applying the control law to the cell, A5: generating a reference signal, A6: acquiring the signal by the microphones, A7: calculating the insertion loss IL, A8: comparing the calculated insertion loss IL with a predetermined insertion loss value IL0 corresponding to the determination of the generalized acoustic impedance Zdet, A9: returning to A3 to adapt the parameters of the control law to minimize the difference between an estimated acoustic impedance of the loudspeaker and the generalized acoustic impedance Zdet, and to maximize the insertion loss if IL is strictly less than IL0.or end of the control process if IL is greater than IL0.

[0030] Insertion loss is a common and known physical quantity that corresponds to the reduction in acoustic power level caused by inserting an acoustic control device into a duct in place of a rigid-walled duct section.

[0031] It is also possible to adjust the parameters of the control law as long as the value of a reference physical quantity other than the insertion loss is sufficiently close to a predetermined value. This other reference physical quantity is, for example, a transmission loss, an absorption coefficient, and / or a target impedance.

[0032] According to a particular design feature, the fraction of acoustic waves absorbed by the speaker diaphragm is converted into electrical energy used to power the cells. The cells are then at least partially self-powered.

[0033] According to a particular implementation, the generalized acoustic impedance Zdet is determined from the Thiele and Small model according to the equations [Math 4] With x the position of the speaker diaphragm, Mms, Rms, Cms the equivalent mass, damping and compliance coefficients of the speaker, p, pb the acoustic pressures in front and behind the diaphragm, Sd the equivalent area of ​​the piston, Bl the electromagnetic force factor, i the current in the speaker coil, u the voltage across the speaker coil, Re, Le, the electrical resistance and inductance of the speaker coil.

[0034] According to a particular implementation, the acoustic impedance is determined by the control law in the following way: the intensity of the current (i) which must be sent to the loudspeaker (11) is determined from the voltage (u), the acoustic pressure (p), in the form of a sum of filters with infinite impulse response denoted IIR, for "Infinite Impulse Response" in English, whose dynamics are materialized by at least two transfer functions Hu and Hp: [Math 1] With Hu and Hp, which can be written in discrete time as a fraction of polynomials in z: [Math 2] With (ai, bi) the real coefficients of the equation and (m, n) the integers corresponding to the order of the filter. Knowing that the property according to which z -1 is a pure delay of a sampling period giving the recurrent control equation between an output at time k denoted (yk) and an input at time k denoted (Xk): [Math 3] Given that the current control signal in the speaker coil depends on the acoustic pressure, its gradient, and the voltage, the complete control equation can be written as the sum of two recursive equations of the previous form: ytot = yu + yp, with yu depending on the measured voltage and yp depending on the estimated acoustic pressure and possibly the estimated acoustic pressure gradient.

[0035] According to another particular implementation, the acoustic impedance is determined by the control law in the following way: the current intensity (i) is determined from the voltage (u), the acoustic pressure (p), the acoustic pressure gradient grad(p), in the form of a sum of filters with infinite impulse response denoted IIR, for "Infinite Impulse Response" in English, whose dynamics are materialized by transfer functions Hu, Hgp and Hp: [Math 5] With Hu, Hgp and Hp which can be written in discrete time as a fraction of polynomials in z: [Math 2] With (ai, bi) the real coefficients of the equation and (m, n) the integers corresponding to the order of the filter. Knowing that the property according to which z -1 is a pure delay of a sampling period giving the recurrent control equation between an output at time k denoted (yk) and an input at time k denoted (Xk): [Math 3] Given that the current control signal in the speaker coil depends on the acoustic pressure, its gradient, and the voltage, the complete control equation can be written as the sum of two recursive equations of the previous form: ytot = yu + yp, with yu depending on the measured voltage and yp depending on the estimated acoustic pressure and possibly the estimated acoustic pressure gradient.

[0036] In general, the intensity value is preferably calculated by an infinite impulse response (IIR) filter, as a function of the voltage across the speaker terminals and the acoustic pressure of the speaker diaphragm.

[0037] According to a particular implementation feature, the control law is defined for an acoustic wave having a frequency between 20 Hz and 20 kHz.

[0038] The invention also relates to an acoustic control device for controlling the propagation of acoustic waves near a wall. The control device comprises an array of cells, at least one (Nc), each containing a control unit, a loudspeaker connected to the control unit, a set of microphones (Nm) connected to the control unit, and a power supply. The control unit is configured to drive the loudspeaker based on signals measured by the microphones. Each microphone in each cell is capable of measuring the acoustic pressure of the acoustic waves, and these acoustic pressure measurements are fed back to the cell's control unit. The voltage across the loudspeaker in each cell is also measured, and this voltage measurement is also fed back to the cell's control unit.The control unit is configured to estimate the sound pressure level and / or its tangential spatial derivatives at the loudspeaker. The control unit is configured to apply the control law, determined from the voltage measurement and the estimate of the sound pressure level and / or its tangential spatial derivatives, which sets the intensity of the electrical signal sent to the loudspeaker in order to determine a generalized acoustic impedance Zdet for the loudspeaker. The control device is configured to implement a control method as defined above.The master device is configured to drive all the control units iteratively with the following steps: Start, A1: load a generic acoustic model, A2: assign a control law to at least one of the cells, A3: calculate the parameters associated with the control law, A4: apply the control law to the cell, A5: generate a reference signal, A6: acquire the signal by the microphones, A7: calculate the insertion loss IL, A8: compare the calculated insertion loss IL with a predetermined insertion loss value IL0 corresponding to the determination of the generalized acoustic impedance Zdet, A9: return to A3 to adapt the parameters of the control law to minimize the error on the measured impedance and maximize the insertion loss if IL is strictly less than IL0, or end of the control process if IL is greater than IL0.

[0039] According to other features, the invention relates to an acoustic control device for controlling the propagation of acoustic waves in the vicinity of a wall (2), characterized in that it comprises an assembly of a number Nc, at least equal to 1, of cells (1) each comprising a control unit (12), a loudspeaker (11) connected to the control unit (12), an array of Nm microphones (10) connected to the control unit (12) and a power supply, the control unit (12) being configured to drive the loudspeaker (11) according to the signals measured by the microphones (10), each microphone (10) of each cell (1) being capable of measuring the acoustic pressure of the acoustic waves, these acoustic pressure measurements being returned to the control unit (12) of the cell (1), the voltage across the loudspeaker (11) of each cell (1) also being measured,this voltage measurement being also fed back to the control unit (12) of the cell (1), the control unit (12) being configured to estimate the acoustic pressure and / or its tangential spatial derivatives at the loudspeaker (11), and configured to apply the control law, determining from the voltage measurement and the estimation of the acoustic pressure and / or its tangential spatial derivatives, the intensity of the electrical signal which is sent to the loudspeaker (11) so as to determine a generalized acoustic impedance Zdet for the loudspeaker (11), the control device being configured to implement a control method according to any one of the preceding claims, the master device (C) being configured to drive all the control units (12) iteratively with the following steps: Start, A1: loading a generic acoustic model, A2: assigning a control law to at least one of the cells,A3: Calculation of parameters associated with the control law, A4: Application of the control law to the cell, A5: Generation of a reference signal, A6: Acquisition of the signal by the microphones, A7: Calculation of the insertion loss IL, A8: Comparison of the calculated insertion loss IL with a predetermined insertion loss value IL0 corresponding to the determination of the generalized acoustic impedance Zdet, A9: Return to A3 to adapt the parameters of the control law to minimize the difference between an estimated acoustic impedance of the loudspeaker and the generalized acoustic impedance Zdet, and maximize the insertion loss if IL is strictly less than IL0, or end of the control process if IL is greater than IL0.

[0040] The multiple cells and the multiple microphones per cell allow for a modular and distributed control system. In particular, the control law can vary from one cell to another. The control system can operate with the other cells if one of the cells malfunctions. Due to the iterative control law, the control system adapts to the ambient sound conditions near the loudspeakers.

[0041] Depending on a particular design feature, each cell contains between 3 and 5 microphones, preferably 4 microphones.

[0042] In one embodiment, the control unit includes a microcontroller, preferably of the ARM or FPGA type. In another embodiment, the master device includes a microcontroller, preferably of the ARM or FPGA type.

[0043] The invention also relates to an air and / or ventilation duct incorporating a control device as defined above, the generalized acoustic impedance Zdet of each loudspeaker of each cell of the control device according to a control law according to a control method as defined above to optimize the acoustic absorption inside the duct.

[0044] Alternatively, the invention relates to an air and / or ventilation duct incorporating a control device as defined above, the generalized acoustic impedance Zdet of each loudspeaker of each cell of the control device following a control law according to a control method as defined above to maximize the acoustic absorption inside the duct.

[0045] Alternatively, the control device can be integrated into an acoustic panel.

[0046] The invention will be better understood upon reading the following description of preferred embodiments, given by way of simple figurative and non-limiting example, and accompanied by figures among which: - represents a schematic view of an acoustic control device according to a first embodiment of the invention; and - represents a detail of an acoustic cell of a control device according to the first embodiment of the invention.

[0047] The present description is given as a non-limiting example of implementation.

[0048] The device according to the invention aims to transform an electroacoustic transducer into a versatile electroacoustic resonator capable of absorbing sound energy in a space or containing this energy between two adjacent spaces by adapting a generalized acoustic impedance, in order to obtain the desired noise reduction.

[0049] The acoustic impedance presented by the membrane of the electroacoustic transducer to the surrounding sound field can thus be made transparent, absorbent or insulating to incident sound waves, depending on the transfer function used.

[0050] A moving part of the loudspeaker, such as the diaphragm, dust cover, and voice coil, is set in motion when subjected to an exogenous acoustic pressure field. It oscillates back and forth along the transducer's axis of symmetry and is returned to an equilibrium position by a spider and peripheral suspensions. The movement of the voice coil, itself immersed in a magnetic field generated by a permanent magnet, creates an electromotive force, which is translated into an induced voltage across the transducer's electrical terminals.

[0051] This induced voltage mirrors the acoustic disturbance that causes the movement of the moving part, but it also depends on the internal dynamics of the loudspeaker system and the acoustic radiation conditions. It constitutes the input to the control device, which establishes a calculated compensating electrical current to apply a mechanical force to the diaphragm, tailored to achieve the desired acoustic effect of attenuating noise near the diaphragm.

[0052] By controlling the current flowing through the voice coil of each loudspeaker, it is possible to control the overall acoustic impedance of the cartridge. Controlling the overall acoustic impedance—that is, the ratio of the estimated acoustic pressure exerted on the loudspeaker diaphragm to the diaphragm's velocity as determined by a control law—results in a significant reduction in the electrical energy consumed for noise reduction.

[0053] The developed control system allows for the simultaneous control of Nc active loudspeaker cells, with Nc being at least equal to 1. Furthermore, the control system's architecture enables real-time modification of the dynamics of each implemented filter. Programming the generalized acoustic impedance, imposed locally on the active surface applied to the wall, thus facilitates the implementation of various control strategies.

[0054] The bandwidth of interest extends from 20 to 20000 Hertz, and in particular in the context of civil engineering applications from 20 to 1500 Hertz.

[0055] To ensure a reduced footprint and allow effective control in the targeted frequency band, the wall can be subdivided into local control zones of five centimetres per side.

[0056] As shown in figures 1 and 2, the device consists of Nc = 4 identical and autonomous cells 1 composed of a loudspeaker 11, Nm microphones 10, an electronic signal conditioning board, a digital calculation board and a power supply, the whole representing the control unit 12.

[0057] Each cell contains between 3 and 5 microphones 10, preferably 4. Each loudspeaker 11 is controlled by a power supply managed by a specifically developed digital processing board. The four microphones 10 of each cell 1 allow for the estimation of the average sound pressure at the center of the diaphragm of each loudspeaker. The difference in sound pressure between the right and left boundaries of the cell allows for the evaluation of the spatial sound pressure gradient along the wave propagation axis in the duct.

[0058] In terms of operation and in relation to the control system, the control device measures the sound pressure at microphones 10 to deduce the sound pressure at the loudspeaker diaphragm 11. After conditioning in a processing unit 13, the signals are digitized by an analog-to-digital converter, also known as an ADC. The average sound pressure at the center of the diaphragm and / or the spatial derivative of the sound pressure at the diaphragm is estimated from the microphone measurements. The control law is then determined by the processing unit 12, also taking into account the voltage measured across the loudspeaker. The current setpoint resulting from the previous calculation is generated by a digital-to-analog converter, also known as a DAC. Finally, a current source drives the current flowing through the loudspeaker 11.

[0059] The loudspeakers 11, for example, are driven by a current source based on 150mA operational amplifiers. The chosen form is an improved Howland source, stable in the case of inductive loads such as loudspeakers.

[0060] Advantageously, the fraction of acoustic waves absorbed by the speaker diaphragm 11 is converted into electrical energy to power each of the cells.

[0061] In more detail, the control method according to the invention includes a step a) in which a number Nc, at least equal to 1, of cells 1 are affixed to the wall, each cell consisting of a control unit 12, a loudspeaker 11 connected to the control unit 12, a set of microphones 10 connected to the control unit 12 and a power supply.

[0062] The control method includes a step b) in which each microphone 10 of each cell 1 measures the acoustic pressure of the acoustic waves, the voltage across the loudspeaker of each cell also being measured, these acoustic pressure measurements and this voltage measurement being returned to the control unit 12 of the cell.

[0063] The control method includes a step (c) in which the control unit 12 estimates the sound pressure at the loudspeaker and / or its spatial derivative, and then determines, from the voltage measurement and the estimate of the sound pressure and / or its spatial derivative, by means of the control law, the intensity of the electrical signal sent to the loudspeaker 11 so as to determine an acoustic impedance Zdet for the loudspeaker 11. According to the invention, the control unit 12, in this step, estimates either the sound pressure at the loudspeaker 11, or its spatial derivative, or both. Using the spatial derivatives of the sound pressure advantageously allows for consideration of the rates of change of the sound pressure field on the wall of the acoustic treatment and for taking into account the effective velocities of noise propagation along the wall.

[0064] The control method includes a step d) in which the control unit 12 sends the electrical signal to the loudspeaker 11, so that a first fraction of the acoustic waves is absorbed by the loudspeaker diaphragm, and a second remaining fraction is reflected.

[0065] A master device C, equipped with an interface card, allows for efficient communication with the control unit 12 of each individual cell 1 via a graphical user interface. The master device C is configured to control all the control units 12 using a learning loop. The control unit 12 and / or the master device C may include, for example, a microcontroller, preferably ARM or FPGA type, to determine the control law of cell 1. An ARM processor is a processor built on a RISC architecture, short for "Advanced RISC Machine." An FPGA, short for "Field Programmable Gate Array," is an electrically configurable logic component used to build a prototype of a future ASIC.

[0066] As an example, the loop may include a first "Start" step to initiate the process. This is followed by a step A1 in which a generic acoustic model is initiated and defined by equation [Math3] below. Then, in step A2, a control law is assigned to at least one of the cells. In A3, the parameters associated with the control laws are calculated. In A4, the control law is applied to the cell. To verify the suitability of the device consisting of all the cells in terms of generalized impedance, in A5, a reference signal is generated. This reference signal is actually "noise" initiated by the loudspeaker or by an external component, which is collected in step A6 by the microphones to initiate the control loop. Step A6 allows the microphones to collect the signal. The insertion loss IL must then be calculated in A7.At A8, the insertion loss IL is compared with a predetermined insertion loss value IL0 to check if the insertion loss is greater than the minimum value IL0 corresponding to the desired generalized impedance. At A9, the master device C loops back to A3 to adjust the control law parameters to minimize the error in the measured impedance if IL < IL0. Otherwise, the loop terminates with the End command.

[0067] Thus, if the insertion loss IL is less than a minimum value, the master device C restarts the loop to refine the control laws. This process is repeated until the desired generalized impedance Zdet is obtained. When the predetermined insertion loss threshold IL0 is reached, step c) is applied, which adjusts the appropriate control law parameters to obtain the determined, i.e., targeted, generalized acoustic impedance Zdet for loudspeaker 11 for acoustic waves near loudspeaker 11.

[0068] It is possible to calibrate each of the cells 1 simultaneously, just as it is possible to calibrate the cells 1 one after the other. The implemented control laws are infinite impulse response (IIR) filters. The filter output depends both on the state of the inputs, i.e., the voltage and sound pressure and / or the sound pressure gradient, and on the output, namely the current flowing through the speaker coil, at time t and at previous times.

[0069] The control law is determined in discrete time using a recursive equation. This recursive equation is derived from the expression of the transfer function representing the targeted dynamics. In particular, the following equivalence relation, d / dt = jω = s, is used, which allows conversion from the time-domain representation to the harmonic and Laplace frequency domain representations.

[0070] Acoustic impedance is determined by the control law as follows: the current intensity (i) is determined from the voltage (u), and the acoustic pressure (p) is in the form of a sum of infinite impulse response (IIR) filters. The sum of IIR filters is represented by two transfer functions, Hu and Hp: [Math 1] With Hu applying to the tension (u) and Hp applying to the acoustic pressure (p). With Hu and Hp which can be written in discrete time as a fraction of polynomials in z: [Math 2] With (ai, bi) the real coefficients of the equation and (m, n) the integers corresponding to the order of the filter, knowing that the property that z-1 is a pure delay of a sampling period giving the recurrent control equation between an output at time k denoted (yk) and an input at time k denoted (Xk): [Math 3] knowing that the signal driving the current in the speaker coil depends on the acoustic pressure, its gradient, and the voltage, the complete control equation is written as the sum of two recursive equations of the previous form: ytot = yu + yp, with yu depending on the measured voltage and yp depending on the estimated acoustic pressure and possibly the estimated acoustic pressure gradient.

[0071] Equations [Math 2] and [Math 3] are classic generic definitions of filtering techniques that allow us to express, with equation [Math 1], the desired dynamics of the current intensity (i) with respect to the acoustic pressure (p) and the voltage, in the form of a sum of filters with infinite impulse response.

[0072] Equations [Math 1], [Math 2] and [Math 3] are for example implemented by modeling the behavior of the loudspeaker 11 and in particular of its diaphragm by the Thiele and Small model according to equations [Math 4] With x the position of the speaker diaphragm, Mms, Rms, Cms the equivalent mass, damping and compliance coefficients of the speaker, p, pb the acoustic pressures in front and behind the diaphragm, Sd the equivalent area of ​​the piston, Bl the electromagnetic force factor, i the current in the speaker coil, u the voltage across the speaker coil, Re, Le, the electrical resistance and inductance of the speaker coil.

[0073] Alternatively, the acoustic impedance is determined by the control law in the following way: the current intensity (i) is determined from the voltage (u), the acoustic pressure (p), the acoustic pressure gradient grad(p), in the form of a sum of filters with infinite impulse response denoted IIR, for "Infinite Impulse Response", whose dynamics are materialized by transfer functions Hu, Hgp and Hp:[Math 5] With Hu applying to the tension, Hp applying to the acoustic pressure p and Hgp applying to its gradient grad(p). With Hu, Hgp and Hp which can be written in discrete time as a fraction of polynomials in z: [Math 2] With (ai, bi) the real coefficients of the equation and (m, n) the integers corresponding to the order of the filter, knowing that the property that z-1 is a pure delay of a sampling period giving the recurrent control equation between an output at time k (yk) and an input at time k (Xk): [Math 3] knowing that the signal driving the current in the speaker coil depends on the acoustic pressure, its gradient, and the voltage, the complete control equation is written as the sum of two recursive equations of the previous form: ytot = yu + yp, with yu depending on the measured voltage and yp depending on the estimated acoustic pressure and possibly the estimated acoustic pressure gradient.

[0074] Equations [Math 2] and [Math 3] are equations which are classic generic definitions of filtering techniques which allow us to express with equation [Math 5] the desired dynamics of the current intensity (i) with respect to the acoustic pressure (p), the acoustic pressure gradient grad(p) and the voltage (u), in the form of a sum of filters with infinite impulse response.

[0075] Equations [Math 5], [Math 2] and [Math 3] are for example implemented by modeling the behavior of the loudspeaker 11 and in particular of its diaphragm by the Thiele and Small model according to equations [Math 4] With x the position of the speaker diaphragm, Mms, Rms, Cms the equivalent mass, damping and compliance coefficients of the speaker, p, pb the acoustic pressures in front and behind the diaphragm, Sd the equivalent area of ​​the piston, Bl the electromagnetic force factor, i the current in the speaker coil, u the voltage across the speaker coil, Re, Le, the electrical resistance and inductance of the speaker coil.

[0076] The electroacoustic control method and device according to the invention allows the implementation of a distributed control law based on an advection equation aimed at attenuating grazing acoustic waves in a tube.

[0077] The advantages of the invention are as follows: - the device can be programmed and the preferred direction of processing can be modified, - the device can be programmed in "self-learning" mode so as to define locally and in real time the optimal acoustic behavior, - the device is modular and can adopt several geometries, - the device allows the synthesis of an acoustic diode, i.e. the non-reciprocal propagation of waves, - the device allows the measurement of wall acoustic pressure fields in real time and therefore offers a source analysis capability, - the device is more robust than classical control approaches due to the distributed nature of the control units and due to the consideration of a voltage measurement and an acoustic pressure measurement, - the device is more efficient than other active systems, in terms of energy consumption.

[0078] Taking into account sound pressure and voltage measurements across each loudspeaker improves sound absorption, notably by limiting the accumulation of errors when determining each generalized impedance using the acoustic control process. The use of voltage allows for better results in both low and high frequencies. Furthermore, using voltage enables the device to automatically adapt to temperature changes without requiring an additional sensor.

[0079] Of course, various modifications can be made by a person skilled in the art to the invention just described without going out of scope of the disclosure of the invention.

Claims

Acoustic control method for controlling the propagation of acoustic waves in the vicinity of a wall (2), the control method comprising: - a step a) in which a number Nc, at least equal to 1, of cells (1) are affixed to the wall (2), each cell (1) comprising a control unit (12), a power supply, a set of Nm microphones (10) connected to the control unit (12), a loudspeaker (11) connected to the control unit (12), each cell (1) being designed so that the loudspeaker (11) is driven by the control unit (12) according to the signals measured by the microphones (10), - a step b) in which each microphone (10) of each cell (1) measures the acoustic pressure of the acoustic waves, these acoustic pressure measurements being returned to the control unit (12) of the cell (1),the voltage across the loudspeaker (11) of each cell (1) being also measured and this voltage measurement being also fed back to the control unit (12) of cell (1), - a step c) in which the control unit (12) of cell (1) estimates the acoustic pressure and / or its tangential spatial derivatives at the loudspeaker (11) of cell (1), then applies a control law, determining from the voltage measurement and the estimation of the acoustic pressure and / or its tangential spatial derivatives, the intensity of the electrical signal that is sent to the loudspeaker (11) so as to determine a generalized acoustic impedance Zdet for the loudspeaker (11), - a step d) in which the control unit (12) of cell (1) sends the electrical signal to the loudspeaker (11) such that a first fraction of the acoustic waves is absorbed and a second fraction of the waves is reflected by the diaphragm of the speaker (11),a master device (C) iteratively controlling all the control units (12) to adjust the generalized acoustic impedance Zdet for each cell (1). Acoustic control method according to claim 1, wherein the control method is a process for processing, by absorption and / or reflection of acoustic waves by the cells (1), the generalized acoustic impedance Zdet being determined so as to optimize the processing by absorption and / or reflection of acoustic waves by the loudspeakers (11) of the cells (1). Acoustic control method according to claim 1 or 2, wherein the control method implements a control loop which comprises the following steps: Start, A1: loading a generic acoustic model, A2: assigning a control law to at least one of the cells, A3: calculating the parameters associated with the control law, A4: applying the control law to the cell, A5: generating a reference signal, A6: acquiring the signal by the microphones, A7: calculating the insertion loss IL, A8: comparing the calculated insertion loss IL with a predetermined insertion loss value IL0 corresponding to the determination of the generalized acoustic impedance Zdet, A9: returning to A3 to adapt the parameters of the control law to minimize the error on the measured impedance and maximize the insertion loss if IL is strictly less than IL0, or ending the control method if IL is greater than IL0. Acoustic control method according to any one of the preceding claims, wherein the fraction of the acoustic waves absorbed by the loudspeaker diaphragm (11) is converted into electrical energy used for powering the cells (1). Acoustic control method according to any one of the preceding claims, wherein the generalized acoustic impedance Zdet is determined from the Thiele and Small model according to the equations With x the position of the speaker diaphragm, Mms, Rms, Cms the equivalent mass, damping and compliance coefficients of the speaker, p, pb the acoustic pressures in front and behind the diaphragm, Sd the equivalent area of ​​the piston, Bl the electromagnetic force factor, i the current in the speaker coil, u the voltage across the speaker coil, Re, Le, the electrical resistance and inductance of the speaker coil. Acoustic control method according to any one of the preceding claims, wherein the acoustic impedance is determined by the control law as follows: the current intensity (i) which must be sent to the loudspeaker (11) is determined from the voltage (u), the acoustic pressure (p), in the form of a sum of infinite impulse response filters denoted IIR whose dynamics are materialized by at least two transfer functions Hu and Hp, with Hu applying to the voltage (u) and Hp applying to the acoustic pressure (p), with Hu and Hp which can be written in discrete time as a fraction of polynomials in z: With (ai, bi) the real coefficients of the equation and (m, n) the integers corresponding to the order of the filter, knowing that the property according to which z -1is a pure delay of a sampling period giving the recurrent control equation between an output at time k denoted (yk) and an input at time k denoted (Xk): Given that the current control signal in the speaker coil depends on the acoustic pressure, its gradient, and the voltage, the complete control equation can be written as the sum of two recursive equations of the previous form: ytot = yu + yp, with yu depending on the measured voltage and yp depending on the estimated acoustic pressure and possibly the estimated acoustic pressure gradient. An acoustic control method according to the preceding claim, wherein the acoustic impedance is determined by the control law as follows: the current intensity (i) is determined from the voltage (u), the acoustic pressure (p), and the acoustic pressure gradient grad(p), in the form of a sum of infinite impulse response filters denoted IIR, whose dynamics are represented by transfer functions Hu, Hp, and Hgp , with Hu applying to the tension, Hp applying to the acoustic pressure (p) and Hgp applying to its gradient (grad(p)), with Hu, Hp and Hgp which can be written in discrete time as a fraction of polynomials in z: With (ai, bi) the real coefficients of the equation and (m, n) the integers corresponding to the order of the filter. Knowing that the property according to which z -1is a pure delay of a sampling period giving the recurrent control equation between an output at time k denoted (yk) and an input at time k denoted (Xk): Given that the current control signal in the speaker coil depends on the acoustic pressure, its gradient, and the voltage, the complete control equation can be written as the sum of two recursive equations of the previous form: ytot = yu + yp, with yu depending on the measured voltage and yp depending on the estimated acoustic pressure and possibly the estimated acoustic pressure gradient. Acoustic control method according to any one of the preceding claims, wherein the control law is defined for an acoustic wave having a frequency between 20 Hz and 20 kHz. Acoustic control device for controlling the propagation of acoustic waves in the vicinity of a wall (2), characterized in that it comprises an assembly of a number Nc, at least equal to 1, of cells (1) each comprising a control unit (12), a loudspeaker (11) connected to the control unit (12), an array of Nm microphones (10) connected to the control unit (12) and a power supply, the control unit (12) being configured to drive the loudspeaker (11) according to the signals measured by the microphones (10), each microphone (10) of each cell (1) being capable of measuring the acoustic pressure of the acoustic waves, these acoustic pressure measurements being returned to the control unit (12) of the cell (1), the voltage across the loudspeaker (11) of each cell (1) also being measured, this voltage measurement also being returned to the control unit (12) of the cell (1),the control unit (12) being configured to estimate the acoustic pressure and / or its tangential spatial derivatives at the loudspeaker (11), and configured to apply the control law, determining from the voltage measurement and the estimation of the acoustic pressure and / or its tangential spatial derivatives, the intensity of the electrical signal that is sent to the loudspeaker (11) so as to determine a generalized acoustic impedance Zdet for the loudspeaker (11), the control device being configured to implement a control method according to any one of the preceding claims, the master device (C) being configured to drive all the control units (12) iteratively with the following steps: Start, A1: loading a generic acoustic model, A2: assigning a control law to at least one of the cells, A3: calculating the parameters associated with the control law,A4: Application of the control law to the cell, A5: Generation of a reference signal, A6: Acquisition of the signal by the microphones, A7: Calculation of the insertion loss IL, A8: Comparison of the calculated insertion loss IL with a predetermined insertion loss value IL0 corresponding to the determination of the generalized acoustic impedance Zdet, A9: Return to A3 to adapt the parameters of the control law to minimize the error on the measured impedance and maximize the insertion loss if IL is strictly less than IL0, or end of the control process if IL is greater than IL0. Control device for the propagation of acoustic waves in the vicinity of a wall (2), according to claim 9, characterized in that each cell (1) comprises between 3 and 5 microphones (10), preferably 4 microphones (10). Control device for the propagation of acoustic waves in the vicinity of a wall (2), according to any one of the preceding claims 9 and 10, wherein the control unit (12) and / or the master device (C) comprises a microcontroller, preferably of the ARM or FPGA type. A ventilation and / or air duct comprising a control device according to any one of claims 9 to 11, the generalized acoustic impedance Zdet of each loudspeaker (11) of each cell (1) of the control device according to a control law according to a method according to claims 1 to 8, to optimize the absorption of acoustic waves inside the duct.

Images