System and method for manipulating the spectral composition and / or for reducing an acoustic power emitted by a source

Abstract

The invention relates to a system and a method for manipulating and / or reducing the spectral composition of a first acoustic power emitted by an acoustic source. The aim of the invention is to eliminate the disadvantages of the prior art and to provide a system by means of which an acoustic power emitted by an acoustic source can be manipulated and / or reduced. This is achieved by means of the features specified in the claims.

Description

System and method for manipulating the spectral composition and / or reducing the acoustic power emitted by a source Description

[0001] The invention relates to a system and a method for manipulating and / or reducing the spectral composition of a first acoustic power emitted by an acoustic source.

[0002] A related method and system for actively reducing the noise emissions of jet engines is disclosed, for example, in German patent DE 101 35 566 A1, in which a sound field is measured using a plurality of sensors. This sound field consists of a primary sound field and a secondary sound field superimposed on it for noise reduction. Actuators are controlled to generate the secondary sound field by means of a control device, which minimizes the measured sound field based on a known transfer function between the actuators and the sensors. Several vibration modes are determined from the sound field, and their amplitudes are used as input for the control device. Description of the invention

[0003] The object of the invention is to eliminate the disadvantages of the prior art and to provide a system by means of which the emitted acoustic power of an acoustic source can be manipulated and / or reduced.

[0004] This problem is solved by the features listed in the claims.

[0005] A system for manipulating and / or reducing the spectral composition of an acoustic source S is being developed. emitted first acoustic power P s provided, whereby the first acoustic power P s a first acoustic intensity field l s i forms. The system according to the invention comprises N sound intensity actuators A1, A2, -, A n , A N on, where each of these sound intensity actuators A1, A2, -, A n , A N each with a distance R SII, RS,2> -> Rs,n> Rs, N and a radiation direction around the source S, but within a virtual surface C s , is arranged. Furthermore, each of the N sound intensity actuators A1, A2,..., A n , A N set up to create a second acoustic intensity field I defined in amplitude, frequency and phase A 1 , l A n until Aj W with a second acoustic power P A to generate. The system according to the invention further comprises M sound intensity probes I 1 , 1 2 , I m , I M on, which are arranged on the virtual surface Cs, wherein the M sound intensity probes I 1 , I 2 , I m , I MEach of the K axes measures the component of an incident intensity vector I™ projected onto the respective axis direction. An electronic signal processing and active control device is configured to determine the total emitted acoustic power P, formed from the first acoustic power P s the source S and the second power P A the N sound intensity actuators are manipulated or minimized.

[0006] According to various embodiments, the acoustic source S is composed of L individual acoustic sources.

[0007] The emitted first acoustic power P s is preferably formed from a first active acoustic power P s active or reactive performance P s reactive or a sum of both according to P s = P s active + j ■ Ps_reactive (with j 2 = -1 imaginary unit).

[0008] Furthermore, the emitted first acoustic power P s the integral of the intensity Is^Cg of the source S over the virtually imagined surface C s .

[0009] According to various design variants, the N sound intensity actuators are A1, A2,..., A n , A N within the area defined by C s limited area using a probabilistic approach Method, preferably using a Poisson disc sampling method, distributed.

[0010] A spatial orientation (rotational position) of the N sound intensity actuators A1, A2,..., A is preferred. n , A N can be determined by means of a predetermined probability distribution, preferably a Poisson distribution or a Gaussian distribution.

[0011] According to various design variants, the N sound intensity actuators are A1, A2,..., A n , A Ncontrollable systems, preferably with which an acoustic intensity field I defined in amplitude, frequency and phase A 1 , I A 2 , •••, I A n , I AN The sound intensity is generated individually by each of these sound intensity actuators. The adjustable systems can be, for example, loudspeakers, underwater loudspeakers, and / or structure-borne sound transducers. The use of other adjustable systems is also conceivable.

[0012] According to various embodiments, each sound intensity probe I 1 , I 2 , •••, I m , I M on each surface segment C S 1 , C S 2 , C SM arranged 1 , 1 2 , •••, I m , I M can be uniaxial (one-dimensional), biaxial (two-dimensional), triaxial (three-dimensional) or multiaxial (multi-dimensional).

[0013] Furthermore, according to various embodiments, the M sound intensity probes are sensors or sensor combinations for measuring acoustic active and reactive intensity. Preferably, the M sound intensity probes are pu probes and / or pp probes.

[0014] Furthermore, the system according to the invention can include reference sensors, wherein the number of reference sensors corresponds to the number L of individual sources. The reference sensors can also be configured to measure quantities that are directly related to the acoustic emission of the source S. Preferably, the L reference sensors are intensity probes, microphones, accelerometers, velocity sensors, displacement sensors, pressure sensors, hydrophones, rotational speed sensors, and / or laser vibrometers. The use of reference sensors other than those mentioned here is conceivable.

[0015] To solve the problem, a method for manipulating the spectral composition and / or reducing the acoustic power P emitted by a source S is also required. s by means of a system for manipulating and / or reducing the spectral composition of a first acoustic power P emitted by an acoustic source S s The procedure is provided. It comprises the following steps: a. Defining an envelope area C s , which is a virtual, physically non-existent, area for accounting for the first acoustic power P emitted by the source S s , where the enclosing surface C s the source S completely encloses, b. Arrangement of M intensity probes I 1 , I 2 ,..., I m , I M on surface C s , where each of the M intensity probes I 1 , I 2 ,..., I m , I Ma sub-area segment C S1 , C S2 , C SM the envelope area C s is assigned c. Measuring an incident first intensity vector Z" 1 (with / "? 1 ], I™ k up to 1™ K axis directions) using the M intensity probes and determination of its components of Z" 1 ( / ^, I™ Y and I™ z ) with respect to a reference coordinate system and measuring an incident second intensity vector I™ (with / / 1 , 1™ to I™ axis directions) using the M intensity probes and determining its components of T™ ( ™ X 'I™Y unc i with respect to a reference coordinate system. In the case that only the (noise) source or only the actuators are switched on, for example only Z" 1or only I™ can be determined. In the case where both are "activated", the superimposed field is measured (superposition principle). d. Calculating the first acoustic power P s , by summing individual first acoustic partial powers P Sm , which extend over the sub-area segments C Sm are emitted, and by means of Ps,m = f if - n^ - dA c S,m to be calculated, where n™ is a surface normal vector, and to calculate a second acoustic power P A of N sound intensity actuators A, A2,..., A n , A N "by summing individual second acoustic partial powers P" Am , which extend over the sub-area segments C Sm are emitted, and by means of P A , m = / T^ - n^ - dA calculated, where n™ is a surface normal vector. e. Active control of the system by means of a control system, such that a total emitted acoustic power P, formed from the first acoustic power P s the source S and the second acoustic power P A The N sound intensity actuators are manipulated or minimized according to: p = ' m=iPm> where Pm = L + -n^ -dA = f ■ n™ ■ dA = P Sm + P A m , where P m The determined total power of an intensity probe or the partial sound power on a surface segment corresponds to the control parameters q for the N sound intensity actuators A1, A2,..., A n , A N will be based on P, P m or weighted sizes of P or P m The control parameters q are determined and these control parameters are sent to the N sound intensity actuators A1, A2,..., A n , A Ntransmitted in order to determine the emitted acoustic first power P using these control parameters q s to manipulate and / or reduce. This will affect the superimposed total field of I™ and on area C Sm (or a partial area) is manipulated, thus manipulating the total radiated acoustic power.

[0016] Each surface segment C is given preference S 1 , C S 2 , C S M A surface normal vector nJ, nj, •••, n™, rio is assigned, where riQ, UQ, •••, n™, n is constant over the respective assigned surface segment C. S 1 , C S 2 , C SM is.

[0017] Furthermore, the envelope area C is preferred. s for the division into area sub-segments C S 1 , C S 2 , C S Mdiscretized with the assumption that acoustic state variables, in particular sound pressure, particle velocity, density and / or temperature, are to a first approximation constant over all surface sub-segments C Sm are.

[0018] Furthermore, the problem is solved with a control method for controlling the system according to the invention and / or in a method according to the invention, wherein control parameters q for the N sound intensity actuators A1, A2,..., A n , A N The following procedural steps will be used to determine this: a. Performing a system identification, wherein the transfer function G (in the frequency domain) between the N sound intensity actuators A1, A2,..., A n , A N and the partial powers P measured by the sound intensity probes A m , where a G-matrix is ​​determined to determine the transfer functions, and b. Calculation of the control parameters q of the N sound intensity actuators A1, A2,..., A n , A N by solving a system of equations according to e = p s,x + G ■ q - P x ,soii = p x + G ■ q = 0 using the transfer function G, such that an error signal e is minimized or becomes zero, where P sx P S m , P s , m , weight' P s , weight or P is and P X soll a target value for P Sm , P s , m , we ight' Ps, weight- The separation P A and P s This is not done in the regulation presented here, as only the superimposed field is measured. However, P corresponds to A the secondary power generated by G*q.

[0019] The transfer function Z between the N sound intensity actuators and the L acoustic sources or L reference sensors and / or the transfer function H between the first acoustic source S, composed of up to L sub-sources or detected via L reference sensors, and the measured acoustic partial powers P is also preferred in the control method according to the invention. S m or P Sm weight' where for the determination of the Transfer functions Z and H each result in a Z-matrix and an H-matrix. These are primarily determined and used for the purpose of utilizing a feedback path (further details are explained in Figure 7d i).

[0020] Depending on the implementation, the control procedure is carried out using a Single Input - Single Output (SISO) method, a Single Output - Multiple Expansion (SISO-ME) method, a Single Input - Multiple Output (SIMO) method, or a Multiple Input - Multiple Output (MIMO) method, which will be explained in more detail later.

[0021] The control procedure is operated according to various implementation variants. a. without reference sensors (feedback operating mode) or b. with reference sensors (feedforward operating mode) or c. without reference sensors but with a synthetic reference signal generated by the control system (hybrid operating mode).

[0022] Preferably, the control parameters q for the actuators are calculated based on the error signals by a. An FBFxLMS algorithm using either a Tikhonov regularization or a Moore-Penrose pseudoinverse, where the regularization depends on the problem being addressed: If fewer actuators than intensity probes are used, then the Moore-Penrose pseudoinverse is preferred. Conversely, if there are many actuators and fewer intensity probes, regularization or a singular value decomposition (SVD) variant is preferred. b. Or by a mapped FBFxSVD algorithm using singular value decomposition (SVD), truncated SVD (TSVD), and / or in combination with a Tikhonov regularization or a comparable method, such as Lasso.

[0023] Furthermore, a control method for the general application of the FBFxSVD or FxSVD method in control structures is disclosed, whereby these methods are transferable to any linear problems of the form Ax = b for which a real-time solution for x is sought. In particular, the transferability to classic ANC problems for minimizing sound pressure using the LMS method, as well as in SIMO and MIMO systems, is mentioned. Implementation of the invention

[0024] The invention will be explained in more detail using one or more exemplary embodiments. For this purpose, we will show... Figure 1 Basic configuration of the system for determining the acoustic power S, Figure 2 Basic configuration of the system with sound field actuators, Figure 3 Basic configuration of the system with sound field actuators with Main radiation direction towards the acoustic source, Figure 3b Basic configuration of the system with sound field actuators with Main radiation direction away from the acoustic source, Figure 4 Basic configuration of the system with intensity probes Figure 5 Setup of the intensity probes, Figure 6 Basics of SPCS control, Figure 7a Block diagram: FxLMS, Figure 7b Block diagram: FxSVD, Figure 7c Ratio of computation speeds between FxLMS and FxSVD for different numbers of sound intensity actuators, Figure 7d Block diagram: Feedback Cancellation FxLMS (FBFxLMS), Figure 8 Block diagram of system identification with determination of the Matrix G, Figure 9 Block diagram of system identification with determination of the Matrix Z, Figure 10 Block diagram of system identification with determination of the Matrix H, Figure 11 Block diagram of an SPCS system with FBFxLMS-based control Figure 12 Block diagram of an SPCS system with FBFxLMS-based Hybrid control Figure 13 Block diagram of an SPCS system with FxLMS-based feedback control Figure 14 Representation of 3D intensity vectors measured on the basis of the hydroacoustic radiation from a ship propeller a) Top view and b) Side view Figure 15 shows how to determine the location of the minimum distance between two skew lines. Figure 16 Single configuration of the SPCS system for use in a pipeline Figure 17 Arrangement of the sound intensity probes when using the SPCS system in pipelines Figure 18 Dual configuration of the SPCS system for use in pipelines,

[0025] The description refers to the accompanying drawings, which illustrate specific embodiments in which the arrangement according to the invention can be implemented. In this respect, directional terminology such as "top," "bottom," etc., is used with reference to the orientation of the described drawings. Directional terminology serves for illustration and is in no way restrictive.

[0026] It is understood that other embodiments may be used and structural or logical modifications made without deviating from the scope of protection of the present invention. It is understood that the features of the various exemplary embodiments described herein may be combined with one another, unless specifically stated otherwise. The following detailed description is therefore not to be interpreted as restrictive, and the scope of protection of the present invention is defined by the appended claims.

[0027] In the figures, identical or similar elements are provided with identical reference symbols where appropriate. 1. Basic structure of the SPCS system

[0028] The present invention describes a system (SPCS - Sound Power Control System) with which the acoustic power P emitted by an acoustic source S can be controlled. sits spectral composition can be manipulated and / or reduced. The acoustic source S is to be understood as a compact source composed of L individual acoustic sources. The acoustic power P s The acoustic energy W, which results from the acoustic energy emitted per unit time (Eq. 1), is fundamentally responsible for the formation of all other essential acoustic field quantities, such as acoustic pressure p, velocity. or the intensity / zThis relationship is described in Eq. 1. The acoustic energy W is composed of a potential and a kinetic energy component. p0 is the fluid density without an acoustic field, or is referred to as the rest density, c is the speed of sound at which disturbances propagate in the fluid, and p is the local fluid density in the acoustic field. The subscript i represents the directions with respect to the chosen coordinate space or the global coordinate system. GCS / 7. Furthermore, the Einstein summation convention shall apply. ||*||2 represents the Euclidean norm. ™ = W pot + W kin 1 f / P 2 “ G / .l = - I - 7 + p ■ || Vj H2 )dV with i = 1,2,3 or x,y,z 2 Jy \Po ' c )

[0029] The emitted acoustic power P s / 4 of source S / 1 is now for the formation of the intensity field l s i / 2 responsible and can be expressed as the integral of the intensity Is^Cg) / 5 over the virtually imagined surface C s / 3 are represented; Eq. 2. The area C s The source S must be completely enclosed, or in the general case, in the mathematical sense, it must be closed. P s = I (l s i(Cs) ■ n0i(Cs)} d A with i = 1,2,3 or x,y,z Eq.2 J c s y ' 7

[0030] I s ,i C s ') are the components of the acoustic intensity vector on C s and the components n 0 ( (C s ) / 6 of the surface normal vector on C s (Eq 2).

[0031] If the acoustic field quantities are defined in the complex number space (C, then the components of the intensity vector l can be s i as a sum of active l s active, reactive component Is-re active, i are represented; Eq. 3. In Eq. 3, j represents the imaginary unit, with j 2 = -1. Is,i Is_active,i + j ' ^S_reactive,i Eq.3

[0032] According to Eq. 3, the acoustic power can therefore also be represented as the sum of active and reactive components. Fs Ps_active + j ' ^S reactive Eq.4

[0033] The acoustic power P s The emission emitted by S is an integral, distance-invariant quantity. This means that it is independent of the size and shape of the surface C. s , taking into account the above requirement for C sThe acoustic power is always the same. With regard to the active manipulation of the spectral composition and / or the reduction of the acoustic power emitted by source S by a control system, it can be considered a great advantage that this is based on a distance-invariant quantity. All other relevant acoustic field quantities are, with respect to the underlying acoustic source, distance-dependent. The second advantage of the invention lies in the fact that the control system is based on an acoustic energy quantity, in this case the acoustic power or intensity. If the acoustic energy quantities are manipulated or possibly reduced, this automatically also affects other acoustic field quantities, such as the acoustic pressure p. sThe crucial point and advantage of this invention is that the reverse is not true. Another crucial advantage of this invention is that acoustic sources outside of C s have no influence on the SPCS control system. This can be explained by the fact that for all acoustic sources outside of C s , Eq.2 becomes identically zero. The fourth crucial advantage lies in the fact that, if all components of Is^Cg) are known, directional information regarding the acoustic source S is always also given.

[0034] To control the acoustic power emitted by S, A2-A NSound intensity actuators 8 are arranged around S. Preferably, the actuators are oriented with their main radiation direction towards the acoustic source, as shown in Fig. 3. Alternatively, the actuators can also be oriented with their main radiation direction away from the source, as shown in Fig. 3b. In the general case, the actuators can be located within the area defined by C. s The limited area is preferably distributed using the Poisson disc sampling method or another probability-based method. The spatial orientation (rotational position) of the actuators, defined by three rotation angles (Euler angles) of the main radiation direction relative to the directions of the GCS / 7 (Figs. 1 to 5), is also determined by a predetermined probability distribution, preferably a Poisson distribution or alternatively another distribution such as a Gaussian distribution.

[0035] The positions of the individual actuators are defined with respect to the global coordinate system GCS / 7 (Fig.2) and have distances of 9 R to the center of the acoustic source S or any other arbitrary reference point. S 1 , R Sn to R SN ) (Fig.2).

[0036] All controllable systems capable of generating an acoustic intensity field defined in amplitude, frequency and phase can be used as sound intensity actuators. A 1 , I A>n to I A N ) (Fig. 3 and Fig. 3b) can be generated individually by each sound intensity actuator 8. Preferably, loudspeakers, underwater loudspeakers, structure-borne sound transducers or comparable systems are used for this purpose.

[0037] To build the SPCS control system, the virtual surface C is now used. s multi-axis intensity probes 15 (f 1 , I m to I M(Fig. 4). Within the SPCS system, preferably pu probes, pp probes, or any other type of sensor or sensor combination can be used to measure the acoustic active and reactive intensity. For example, if acoustic intensity probes based on the pp principle are used, the axes of the intensity probes can be composed of at least two microphones, two hydrophones, or two pressure sensors spaced apart Ar. Any other combination of two sensors for the simultaneous measurement of the acoustic pressure and determination of the spatial acoustic pressure gradient can also be used. Likewise, any sensor combination that simultaneously measures the acoustic pressure p and the components of the acoustic velocity can be used. or can measure the component projected in a specific direction. The positions of the individual intensity probes are defined in the coordinates of the GCS / 7 (Fig. 1 to Fig. 5) and are at distances of 14 (Z?) to the center of the acoustic source or any other arbitrary reference point. s / 1 , R Siim to R S I IM ) (Fig. 4). The individual probes are assigned the surface sub-segments 11 (C S1 , C Sm up to C SM ) (Fig.4) with the respective surface normal vectors 12 (nj, n™ to n" ) (Fig.4). Eq.5 is assumed to apply. M m

[0038] In the technical implementation of the invention, the continuous sub-areas C S 1 , C Sm up to C SM approximated in a discretized form, where riQ, n™ to n" are constant over the respective sub-areas. Within the SPCS control kernel, the surface sub-segments are CS1 , C s , m up to C SM as well as the surface normal vectors nJ, n™ to ", which are defined with respect to the GCS / 7 (Fig.1 to Fig.5), are stored electronically.

[0039] The individual m intensity probes measure using I™, If 1 up to I™ the portion of the incident intensity vector projected onto the respective axis direction " 1 / 15e. In the concrete technical implementation of the invention, ID intensity probes 15a, 2D intensity probes 15b, 3D intensity probes 15c or intensity probes with more than 3 axis directions 15d can be used.

[0040] When using ID intensity probes 15a, the probe axis is surface normal on the subsurface C. S m to orient. When using 2D intensity probes 15b, the plane spanned by the two probe axes is surface normal on the subsurface C. s>mto orient. The orientation of the axes in the measurement of the projected fraction of I s on the area segment This is achieved through the direction vector n™ 1; n™ k to n™ K 15e described. The superscript m denotes the probe position or number, the subscript k the probe axis under consideration. Within the SPCS control kernel, the direction vectors n™ k to n™ K The individual probe axes with respect to the defined GCS are electronically stored. -rS,m- r n I,l, X n I,l> Y n I,l, Z jS,m „m „m ri m y 2 rt I,2, X n I,2, Y n I,3, Z 's,x jS,m = n jm *3 I,3, X n I,3, Y n I,3, Z *S, Y GL 6 jS,m „m jm rL I,k, X rL I,k, Y rL I,k, Z Js,z. jS,m \- n I, K, X rLI, K, Y n I, K, Zl U K J 1 0 0' mT ■ n m ■ Z™ =- n mT ■ Js,m;. with T = Y ■ 0 1 0 GL 7 Lo 0 1. jm l S n mT ■ n m + r T ■ r ) ■ n GL8 mT ■ l s ' m Y > 0 for K = 2; y > 0 for K > 3; for K = 1, I™ follows directly. GL9 fs,m.

[0041] In the technically implemented SPCS system, the components I projected onto the probe axis directions are now measured at the probe position m. s f m , I k m to I s f m of the complete intensity vector I™ on C Sm , which The coordinates of GCS 7 are defined and measured. The goal now is to determine the components of f™, specifically I™. x , I™ Y and I™ z, with respect to the GCS. This can be achieved by solving the system of equations Eq. 6. Formally, Eq. 6 is first transformed into Eq. 7 by a Tikhonov regularization, from which the general solution according to Eq. 8 is then obtained. The choice of the regularization parameter y depends on the number of probe axes K with which the projected components of I are determined. s on sub-area C s>m The possible configurations for choosing y, depending on K, are described in Eq. 9. For the special case K = 1, the calculation proceeds directly with the measured projected intensity components. The calculation of the individual acoustic partial powers P Sm / 13 (Fig. 4) which are emitted over the sub-areas Cs,m can be calculated using Eq.10. P Sm = j I^ - n^ - dA GI.1O Cs,m

[0042] The total emitted acoustic power P sis calculated by summing the partial services according to GL.11. M PS = PS, my GLU m=l

[0043] The regulation within the SPCS system is based either on the acoustic partial powers according to Eq. 10, the total power according to Eq. 11, or on the weighted acoustic partial powers Ps,m, weight according to Eq. 12 or the weighted total power according to Eq. 13. The individual weighting factors w m are preferably specified by the user of the SPCS system and stored electronically within the SPCS control system. M ^S,m, weight Ps,m ' mH 1 GL12 m=l M 1 p S, weight - y / 1 PS,m, weight GL13 m=l

[0044] In the preceding description of the mathematical basis of the SPCS system, equations 1 to 13 have always been described from the perspective of the acoustic source S. These are referred to as primary quantities in the following. However, with active control by the SPCS system, acoustic power components are also generated or emitted by the sound intensity actuators. These are referred to as secondary quantities in the following. Equations 2 to 13 can also be formulated without loss of generality with respect to the acoustic emissions from the actuators of the SPCS system and are given the subscript A (equations 14 to 24). P A = I (j A i (C s ) ■ n o i (C s )^ dA with i = 1,2,3 or x,y,z Eq. 14 J C S ^A,i ^A_active,i + j ' ^A_reactive,i GI.15 PA PA_active + j ' PA_reactive GI.16 -rA,m- A r n I,l, X n I,l> Y nI,l, Z jA,m A rt ri I,2, X n I,2, Y n I,3, Z ‘A m , X jA,m = jm n „m. jm A I,3, X n I,3, Y n I,3, Z ‘A, Y H _ <4 GI.17 jA,m „m jm A rL I,k, X rL I,k, Y rL I,k, Z JA, Z. jA,m U K J \- n I, K, X rL I, K, Y n I, K, Zl ■ n m nf 1 . jA.m. p _ y. Gl.18 Lo n mT ■ n m + r T ■ r ) ■ n mT ■ l A ’ m Gl.19 Y > 0 für K = 2; y > 0 für K > 3; für K = 1 folgt direkt Iff _ fA,m. GI.2O P A ,m = f Tff ■ n™ ■ dA GI.21 c S,m M PA = PA,m G / .22 m=l M PA, m, weight PA,™ ' mH ' W m GI.23 m=l M PA, Weight y / p 1 A,m, weight GI.24 m=l

[0045] Within the SPCS system, active control is achieved by manipulating or minimizing the superimposed primary acoustic power parameters of the source S and the secondary power parameters of the N sound intensity actuators, i.e., the total emitted acoustic power. ■ n0■ dA — P s , m + P^m G / .25 MP GI.26 m=l M PM, weight ' P m ' w m> m Ü GI.27 m=l M p 1 weight = y / p 1 m, weight GI.28 m=l

[0046] 1.1 Control procedures of the SPCS system

[0047] The SPCS system can be controlled using four different methods. The specific method used depends on the characteristics of the acoustic source S. These characteristics include, for example, the spatial radiation pattern (symmetrical, asymmetrical) and the temporal characteristics of the acoustic radiation. The latter describes whether the radiation pattern of S is intermittent or constant over time. The following section describes the control methods achievable with the SPCS system, outlining their advantages and disadvantages. The calculation time or complexity is only considered a secondary criterion.

[0048] Control method 1: Single Input - Single Output (SISO) Input = P or P weight (Eq. 26 or Eq. 28) Output = A single control signal for all N actuators. Target criterion: Minimization of the total acoustic power P via C s is emitted. Advantages: Simple and stable control for stationary acoustic sources with symmetrical radiation characteristics. Disadvantage: Problematic with intermittent symmetrical / asymmetrical metric radiation characteristics.

[0049] Control Method 2: Single Output - Multiple Expansion (SISO-E) Input = P or P weight (Eq. 26 or Eq. 28) Output = A uniform output signal which is converted outside the SPCS control loop into the individual control signals of the N actuators according to fixed criteria. Target criterion: Minimization of the total acoustic power P via C s is emitted. Advantages: Simple and stable control, as the SPCS control works in the SISO method, for stationary acoustic sources with asymmetrical radiation characteristics. Disadvantage: Problematic with intermittent radiation characteristics.

[0050] Control Method 3: Single Input - Multiple Output (SIMO) Input = P or P weight (Eq. 26 or Eq. 28) Output = an individual control signal for each of the N actuators Target criterion: Minimization of the total acoustic power P via C s is emitted Advantages: Suitable for sources with intermittent, asymmetrical radiation characteristics. More stable control than variant 4, as it regulates the total integral power. Disadvantage: More complex control algorithms than variants 1 and 2, and the possibility of control instabilities occurring.

[0051] Control Method 4: Multiple Input - Multiple Output (MI MO) Input = P m or P m , weig ht (GI-25 or Gl.27) Output = an individual control signal for each of the N actuators Target criterion: Minimization of the acoustic partial power P m the individual C s>m be emitted. Advantages: Suitable for sources with intermittent, asymmetrical radiation characteristics. Individual control of all intensity actuators to minimize the individual acoustic power levels P. m on C s>m . Disadvantage: More complex control algorithms than variants 1, 2 and 3, and the possibility of control instabilities occurring.

[0052] To further explain the SPCS control, we refer to the general representations in Fig. 6. Here, we assume a compact acoustic source S, which can be understood as a combination of up to L individual sources. In the practical implementation of SPCS, the exact value of L is irrelevant.

[0053] The acoustic source S is responsible for the emission of the acoustic power P. s via C s or for the emission of P Sm via C s>m responsible. Furthermore, there are N sound intensity actuators that control the acoustic power P. A via C s or the partial services P A m via C A>m , measured by M acoustic intensity probes, emit.

[0054] G e C MxN is the transfer function between the acoustic power quantities P A m or P AimiWeig ht and the control signals of the sound intensity actuators. It should be noted that from here on, only transfer functions will be discussed. Depending on whether the control core of the SPCS system operates in the frequency domain or the time domain, G is either the transfer function in the frequency domain or the impulse response in the time domain.

[0055] The acoustic performance parameters P are used in the control loop of the SPCS system. A or P A , weig If work has been done, G is reduced to G e (C 1xW The complex vector qe (C Wx1 describes the source strengths of the sound intensity actuators.

[0056] Z e C LxN describes the transfer functions between the N sound intensity actuators and the L acoustic sources, se (C Lx1 is the complex vector of acoustic source strengths. Z e C LxNThe values ​​of the reference sensors (L and N) are usually not directly known in the practical implementation of SPCS, but can only be determined indirectly via one or more reference sensors. Z represents the transfer functions between L reference sensors and N sound intensity actuators, and s represents the source or signal strength of the reference sensors. The reference sensors are preferably positioned close to the acoustic source S to guarantee a high correlation or coherence between the acoustic signal of the source S and the reference sensor signals. Any type of sensor capable of measuring quantities directly related to the acoustic emission of the source S can be used as a reference sensor. These preferably include microphones, intensity probes, accelerometers, velocity sensors, displacement sensors, pressure sensors, hydrophones, speed sensors, laser vibrometers, or comparable sensors or systems.When actively controlling the SPCS system, preferably one of the L reference sensors is actively used in the control core. However, control using multiple reference sensors is also possible in principle.

[0057] H e C MxL is the transfer function between the acoustic source S, composed of up to L sub-sources or detected via L reference sensors, and the measured acoustic partial powers P. Sm or P Sm weight- VVird in the SPCS control core with the total acoustic performance P s or P SiWe By working closely together, the dimension of H is reduced to H e C. lxL In the practical implementation of SPCS, H cannot be directly measured, but only based on reference sensors. Thus, H represents the transfer function between L reference sensors and M intensity probes.

[0058] The actual control within the SPCS system is based on the superimposed primary acoustic power components of the source S and the secondary acoustic power of the sound intensity actuators according to Eq. 25, Eq. 26, Eq. 27 or Eq. 28. In the internal control of the SPCS system, in practical implementation, one of the error vectors e from Eqs. 29 - Eq. 32 is used, specifying a setpoint (subscript setpoint). e m = P m ~ Pm,should' mit e e G / .29 p c — p — p ■ with m r m, weight r m, weight, shall' c 1= GI.30 e = P — P so u; with e GC lxl GI.31 Pweight Pweight, shall' TTUt 6 GC

[0059] Regardless of which of equations 29 to 32 are used for control within the SPCS system, the following general relationships result. For calculating the optimal control parameters or source strengths q opt For the sound intensity actuators, the system of equations (Eq. 33) must be solved such that the error signal e is minimized or becomes zero. P x can represent P m , P miWeight , P weig ht or nur P stands. e = Px ~ Px,soii = 0 GI.33

[0060] Using the transfer functions according to Fig. 6 and the following explanations, Eq. 34 follows. P x Eq. 34 represents the difference between the emitted acoustic power of the source S and the setpoint to which the SPCS system should regulate, so that the error e is minimized or becomes identically zero. e ~ Ps,x + G ■ q - P x ,soii - Px + G ■ q - 0 G / .34

[0061] The solution of Eq. 34 for calculating the optimal source strengths q opt The sound intensity actuators are achieved purely formally by inverting the matrix G and are shown in Eq. 35. q op t = -G- 1 ■ Px G / .35

[0062] The method used to invert matrix G in Eq. 35 depends on the number of sound intensity actuators N and sound intensity probes M. The following configurations are possible and are covered by the functionality of the SPCS system.

[0063] Variant 1: N > M underdetermined system of equations that is solved after prior regularization using Thikhonov, singular value decomposition (SVD) or an equivalent method

[0064] Variant 2: N = M a specific system of equations that can be solved directly

[0065] Variant 3: N < M overdetermined system of equations that is solved using Moore-Penrose pseudoinverse, Thikhonov regularization, singular value decomposition (SVD) or an equivalent method

[0066] In solution variant 1, the number of sound intensity actuators is greater than the number of intensity probes, resulting in an underdetermined system of equations. This can be solved by using regularization methods, preferably Tikhonov, singular value decomposition, Lasso, or another equivalent method. After Tikhonov regularization (Eq. 36) and specifying the regularization parameter ip, q can be solved using Eq. 37. opt calculated according to Eq. 38. ( G H ■ G + T) ■ q opt = -G H ■ P^; with T = rp ■ I NxN Eq. 36 G^kh = ( G H - G + ^)- 1 GI.37 Q opt, Thikh G^tkh ' G ' Px G / .38

[0067] The calculation of q opt Alternatively, it can also be done according to Eq. 41, where first the singular value decomposition of the matrix G (Eq. 39) and based on this its inverse G y. D calculated according to Eq. 40. G = U - S - V H G / .39 GSVD = V ' S + U H ; with S + = Pseudoinverse of S GI.40 opt, sv D = ~G S y D ■ P x GI.41

[0068] Another possibility is to q optto determine iteratively using a gradient descent method. This is one possible solution variant in the real-time control kernel of the SPCS system and is preferably implemented using the Least Mean Square (LMS) or the Normalized Least Mean Square (NLMS) algorithm. In the following, only NLMS will be referred to, even though both variants are implemented in the SPCS system. Derived variants using reference sensors, such as the filtered LMS (FxLMS) or the Feedback Canceling FxLMS (FBFxLMS) algorithm, can also be implemented by the control kernel of the SPCS system. The iterative solution q opt Using Tikhonov regularization and the Normalized LMS algorithm (NLMS), it is formally determined according to the formation rule in Eq. 42. H Qopt(i 1) Qopt(. G) "t ■ QoptCO + Px, Tikh) ■ G T ikh GI.42

[0069] In GI.42, i denotes the current iteration step, G Tikh the matrix G (Eq. 43) modified for the solution by Tikhonov regularization and Px ikh e i nen modified input vector (Eq. 44). G Tikh = (G H ■ G + V); with V = ip ■ I NxN G / .43 Px*, Ti h = G H - P x * GI.44

[0070] The step size or convergence rate of the method is controlled by the parameter T] in Eq.42 which must lie in the interval 0 < 77 < 2.

[0071] In solution variant 2, the number of sound intensity actuators equals the number of intensity probes, resulting in a precisely determined system of equations, allowing the inverse of G to be calculated directly. For the direct solution, Eq. 36 simplifies to Eq. 45. q op t = -G- 1 ■ Px G / .45

[0072] Equations 39 to 41 remain fully valid. In the iterative solution, equation 42 simplifies to equation 46. / yj \ H \ I II r 112 ) ' (( G ' ^optCO + Px) ■ G) j GI.46 Vlrn lb / /

[0073] The use of a regularization method in the active operation of the SPCS system, as described in detail in solution variant 1, is also fundamentally possible here and may be advantageous. Likewise, active control by the SPCS system can be based on the singular value decomposition described in Eqs. 39 to 41.

[0074] In solution variant 3, the number of sound intensity actuators is smaller than the number of intensity probes, resulting in an overdetermined system of equations that can be solved using the classical method of squared errors. Formally, this only requires setting the Tikhonov regularization parameter rp = 0 in Eq. 37 to obtain the solution for q. opt Equation 38 can be calculated directly. Equations 39 to 41 can also be used without restriction for a direct solution. Within the SPCS system, the iterative approach is preferred. Solution of q opt Equation 42 is used with Q = 0. The use of a regularization method in the active operation of the SPCS system, as described in detail in solution variant 1, is also fundamentally possible and may be advantageous. The use of singular value decomposition, described in Equations 39 to 41, in active control by the SPCS system is also possible. 1.2 Further comments on the control procedure in the SPCS system:

[0075] In the previous presentation, the possible solutions to Eq. 33 were based solely on P. XsoThis is illustrated in Figure 7a. In technical implementation, this corresponds to a "feedback" control system, which, however, is only promising with certain acoustic characteristics of the source S, such as pronounced periodicity. Internal control in the SPCS system is therefore preferably achieved using the "feedforward" strategy (operating mode of the SPCS system: feedforward). The diagram (Fig. 7a) serves as an aid to explain this. Furthermore, it should be noted that the following explanations assume that the SPCS system operates in the frequency domain, which means that any convolution operations can be represented as simple complex-valued multiplications. Without loss of generality, the following statements also apply to the time domain or to the case where the SPCS system operates in the time domain.

[0076] In the "feedforward" strategy, the control signal q for the sound intensity actuators is determined based on a convolution between a reference sensor signal s and a filter w. Since "acoustic feedback" from the actuators to the reference sensors is not yet considered in this example, s = s', where s' is the direct signal component from the source S to the reference sensor. The acoustic intensities generated by the individual actuators are transferred to the individual intensity probes by the transfer function G. In this context, G is also referred to as the secondary path. At the individual intensity probes of the SPCS system, the acoustic signal is superimposed on the signal. Intensity field s* ■ H of the source S. The superimposed signal is called the error signal e, which is to be minimized or made zero by adjusting the filter weights in w (see also FCU (Fig. 11 - reference numeral 30)), e.g., using the LMS or NLMS algorithm. Equations 42 and 46 can then be reformulated as 42b and 46b, respectively. ( / n \ \ \ ilil7 W7 — ■ 1 - ii2 ' ( e * H ' ( G ' s )nkh) H GL42b S) Tikh ||2 / y ( / n \ 777 — ' ( gH ' ( G ' s )) Eq. 46b \ll ü ' S II2 / / e, = G H ■ e GL47a (G ■ s^Tikh = (.(. G ■ ■ (G ■ s) + V); with V = rp ■ I NxN G / .47b

[0077] In the case of regularization according to Tikhonov or when using the Moore-Penrose pseudoinverse with ip = 0 (Eq. 42b), the modified error signal e* (Eq. 47a) and the modified form of the transfer function G (Eq. 47b) are to be formed in the NLMS control loop (see also FCU (Fig. 11 - reference numeral 30)).

[0078] Another way to calculate the filter weights w within a feedforward is to calculate the necessary inverse of G using the singular value decomposition (SVD) according to Eqs. 39 to 41. Following the terminology FxLMS, ​​we call the SVD-based method FxSVD. For this purpose, the NLMS block in Fig. 7a is replaced by the SVD block (Fig. 7b); see also FCU (Fig. 11 - reference numeral 30). Possible uses of SVD in the SPCS control core - Option 1:

[0079] The first application option is more theoretical in nature and consists of calculating the SVD based on the convolution between G and s. However, this has the crucial disadvantage that the SVD must be recalculated in "real time" in the control kernel for each new data block. Due to the high numerical requirements of the SVD, this is not yet feasible with the currently available systems used in the control core of the SPCS system, such as FPGAs (Field Programmable Gate Arrays) or DSPs, or comparable systems, or such systems are the subject of current research and development. Possible uses of SVD in the SPCS control core - Option 2:

[0080] The second option is to calculate the inverse of G, which is needed for control, "offline" after system identification using the SVD. In the technical implementation of the SPCS system, the calculation of G yD According to Eq. 40, the signal is transferred to the OCU (Fig. 11 - reference 34) and, before the start of active control operation, to the SVD block (Fig. 7b) or the FCU (Fig. 11 - reference 30). The convolution of the reference signal s = s* with G is omitted (G = 1) and is instead performed indirectly in the SVD block (see Fig. 7b or also FCU (Fig. 11 - reference 30)). Using Eq. 40 and Eq. 48, Ggy is calculated. D Mod calculated. G svD. Mod = GSVD 0 s i' m it 0 = Hadamard Division GI.48

[0081] The optimal filter weights w can then be directly calculated from GI.48 using GI.49. w opt = {^svD, Mod ' e) 0 s GI.49

[0082] Optionally, it is also possible to omit the full SVD variant Gy from the control core of the SPCS system. D ModInstead of using a method that uses all singular values, it is better to work with a variant that uses only a limited number of singular values ​​to approximate the inverse. This variant is called "Truncated SVD" or TSVD and yields the inverse G^ VD Mod It is also possible to calculate the SVD based on the matrix G modified by Tikhonov regularization. Tikh (Eq. 43) to carry out and its inverse G^ ksVD to determine.

[0083] A combination of both optional variants is also possible in the SPCS system and yields the inverse G^ kTSVD If one of the two variants with Tikhonov regularization is used within the framework of SVD, the modified error signal e* (Eq. 47a) must also be used.

[0084] The use of SVD has the decisive advantage that the iterative solution of w optUsing NLMS and the associated iterative inversion of G is no longer necessary, and the solution can instead be calculated very quickly and efficiently directly in the real-time kernel of the SPCS system, based on simple matrix operations. The computationally intensive inversion of G has already been performed "offline".

[0085] In numerical experiments conducted with an Ml MO setup of up to N = 500 sound intensity actuators and M = 10 acoustic intensity probes M, the ratio of computation times between the NLMS variant and the SVD variant was F speedup = t NLMS / t SVD , in the range between 9000 - 15000. For massive MIMO systems, with e.g. N = 500 sound intensity actuators and M = 10 acoustic intensity probes, the ratio increases significantly again and is in the range of ^speedup ~ 50000.

[0086] With the FxSVD concept shown here, the SPCS system can be implemented based on a MIMO control method, in which significantly more sound intensity actuators and intensity probes are used than would be possible with the FxLMS concept due to the computation times.

[0087] Independently of the present invention, the SPCS system, the FxSVD concept can be used for all applications where active control, similar to that shown in Figs. 7a and 7b, is based on an adaptive filter and the calculation of the filter weights w is performed by an LMS-based algorithm. In particular, the FxSVD concept can be applied throughout the entire Active Noise Control (ANC) field wherever control based on the LMS algorithm or algorithms derived from it is currently used.

[0088] In general, the use of the FxSVD concept in ANC systems promises significant advantages, such that it makes massive MIMO control concepts technically feasible in the first place, using a large number of actuators and sensors.

[0089] Within the SPCS system, the control core has the option of operating in both NLMS and SVD modes.

[0090] The previous descriptions and explanations have not yet addressed the issue of "acoustic feedback" from the actuators to the reference sensors. Taking into account the actuators' effect on the reference sensor signal results in the FBFxLMS algorithm (Fig. 7d).

[0091] The feedback effect of the actuators s A = s ■ w ■ Z on the reference sensor signal s = s* The + is calculated from this via a feedback path before the convolution with the filter w takes place. The neutralization of the “acoustic feedback” can also be implemented in the same way for the FxSVD algorithm (Fig.7b), whereby the previously made remarks apply and this variant is referred to as FBFxSVD. 1.2 Configuration of the SPCS system for use in pipelines and ducts

[0092] The preceding explanations of the SPCS system were based on the example of acoustic power emission into a free field or the outside environment; see also Figures 1 to 4. However, the SPCS system can also be used within pipelines or ducts to modify and / or reduce the spectral composition of the acoustic power emitted by a source S. Within pipelines and ducts, pumps, compressors, fans, or similar systems would be the primary emitters of acoustic power. Figure 16 shows the schematic setup of the SPCS system as a single configuration, on one side of the source S, within a pipeline. This configuration is preferably used when the source S is a The main direction of acoustic radiation is determined by the source S. If the source S is a pump, fan, or compressor, for example, the main direction of acoustic radiation usually coincides with the direction in which the fluid is conveyed. This is referred to as the pressure side of the source S.

[0093] The N sound intensity actuators (Fig. 16 - reference numeral 8) and the M sound intensity probes (Fig. 16 - reference numeral 15) are arranged on the pressure side of the source S. The orientation of the main radiation direction of the individual sound intensity actuators, which are integrated into the pipe or duct wall, relative to the main radiation direction of the source S can be chosen arbitrarily and can be different for each sound intensity actuator. The distance of the N sound intensity actuators to the source S can also vary. Single-axis (Fig. 17 - reference numeral 15a), dual-axis (Fig. 17 - reference numeral 15b), or multi-axis (Fig. 17 - reference numeral 15c and Fig. 17 - reference numeral 15d) probes are used to measure the acoustic intensity. Preferably, these are pu probes, pp probes, or any other type of sensor or sensor combination for measuring the acoustic active and reactive intensity. See notes “1.Basic structure of the SPCS system. There are two variants for the arrangement of the acoustic intensity probes, designated as (Fig. 17 - Optl) and (Fig. 17 - Opt2). In the Optl variant, the intensity probes are either integrated flush with the wall directly into the pipe or duct wall or installed in close proximity to the wall. The technically simplest solution is the flush integration of 1-axis or 2-axis intensity probes according to the pp principle. If, for example, acoustic intensity probes according to the pp principle are used, the axes of the intensity probes can consist of at least two microphones, two hydrophones, or two pressure sensors spaced a distance Ar apart. Any other combination of two sensors for the simultaneous measurement of acoustic pressure and determination of the spatial pressure gradient can also be used.

[0094] The positions of the individual intensity probes are defined in the coordinates of the GCS (Fig. 16 - reference numeral 7). The distance of the individual intensity probes to the source S can vary. If acoustic intensity fields with pronounced three-dimensionality predominate within the pipeline or channel, additional intensity probes can be integrated directly into the pipeline (cross-section) or the channel itself (Fig. 17 - Opt2). If a background flow field is present, the intensity probes must be adequately shielded from the influence of the flow using suitable measures.

[0095] The allocation of surface elements C S 1 , C Sm up to C SM (see also Fig.) 17 - Reference numeral 11) and the respective surface normal vectors n0, n™ to n" to the individual intensity probes are preferably determined by meshing the cross-sectional area of ​​the pipeline or channel, with the positions of the intensity probes as support points. Any numerical meshing method can be used for this purpose.

[0096] If required by the radiation characteristics of the source S or the acoustic properties of the pipeline or duct, sound intensity probes and sound intensity actuators can be integrated on both sides of S; see Fig. 18. This would be the case, for example, with pumps, fans, or compressors if significant acoustic power emission also occurs via the intake side. Sound intensity probes and actuators can be part of a single dual SPCS system (variant 1), as shown in Fig. 18, or operated via two independently operating SPCS systems (variant 2). If there is strong acoustic feedback between the sound intensity actuators on one side of the source S and the sound intensity probes on the other side of the source S, variant 1 is preferably used. If this is not the case, variant 2 is preferably used. 2. Operating modes of the SPCS system

[0097] Explanations of the abbreviations used in the block diagrams can be found in section 2.2.7. 2.1. System identification procedure of the SPCS system

[0098] The APCU of the SPCS system (Fig. 8 - reference numeral 25) first performs system identification before manipulating the spectral composition and / or reducing the acoustic power emitted by source S. During system identification, the transfer functions contained in the matrices H, Z, and G are determined. H is determined under one or more steady-state operating conditions of the sound power source S. 2.1.1 Determining the matrix G

[0099] To determine the matrix G, the sound intensity actuators (Fig. 8 - reference numeral 8) are driven sequentially using white noise or purely harmonic signals. The drive signals (Fig. 8 - reference numeral 23) are generated separately and sequentially for each sound intensity actuator by a signal generator SGU (Fig. 8 - reference numeral 22) and fed into the respective amplifier units APU (Fig. 8 - reference numeral 24). The emission of an acoustic intensity field by the driven actuator n results in a change on the sub-areas C. Sm (Fig.4 - reference numeral 11) generates an acoustic intensity field distribution.

[0100] The intensity probes (Fig. 8 - reference numeral 15) measure the projected intensity components onto the individual probe axes; see also Fig. 5. The signals of the individual projected intensity components (Fig. 8 - 16), which were generated by the intensity actuator n, are converted within the IAPCU (Fig. 8 - reference numeral 17) by the GICU (Fig. 8 - reference numeral 18) according to Eqs. 17 to 20 onto T™. with respect to the GCS (Figs. 1 to 5 - reference numeral 7), converted. The direction vectors of the probe axes n^, n™ required for this. k to n™ K intensity probes 1,2, are stored electronically in the OCU (Fig.8 - reference 34) and are transferred as a signal (Fig.8 - reference 41) to the GICU (Fig.8 - reference 18) of the IAPCU (Fig.8 - reference 17).

[0101] The environmental conditions at the time of system identification are measured by the EDU (Fig. 8 - reference 38) and transferred as signals (Fig. 8 - reference 40) to the GICU (Fig. 8 - reference 18), where they can be used further as needed. The environmental conditions always measured by the EDU (Fig. 8 - reference 38) include static pressure, temperature, and flow velocity in both magnitude and direction. The flow direction is determined by the EDU (Fig. 8 - reference 38) by measuring the three velocity components with respect to the GCS (Fig. 1 to Fig. 5 - reference 7) or by measuring two angles relative to two basis vectors of the GCS (Fig. 1 to Fig. 5 - reference 7). In the event that the fluid in which the SPCS system is used is normal ambient air, the relative humidity will continue to be measured by the EDU (Fig.8 - reference numeral 38).If the fluid in which the SPCS system is used is water, the salinity is measured by the EDU (Fig. 8 - reference numeral 38). The EDU (Fig. 8 - reference numeral 38) is equipped with the necessary sensors / measuring systems for measuring the defined parameters, or such sensors / systems can be connected to it.

[0102] From the signals of the individual intensity probes (Fig. 8 - reference numeral 19), the acoustic partial powers on the partial areas C are then calculated in the APSU (Fig. 8 - reference numeral 20) according to equations 21 to 24. Sm determined. The required areas of the sub-areas C Sm , as well as the surface normal vectors n™ are stored electronically in the OCU (Fig. 8 - reference 34) and are transmitted as a signal (Fig. 8 - reference 41) into the APSU (Fig. 8 - reference 20) is transferred from IAPCU (Fig. 8 - reference 17). The weighting factors w mare automatically set to one during the system identification of matrix G (Eq. 23) by the IAPCU (Fig. 8 - reference numeral 17) in the APSU (Fig. 8 - reference numeral 20).

[0103] The specific acoustic partial power outputs are transmitted as separate signals (Fig. 8 - reference 21) to an Adaptive Filter Unit (ADFU) (Fig. 8 - reference 26), which constitutes the actual control core of the APCU (Fig. 8 - reference 25) of the SPCS system. The control signal (Fig. 8 - reference 23) from the SGU (Fig. 8 - reference 22) is transmitted to the ADFU (Fig. 8 - reference 26) and convolved with the filter weights of the FU (Fig. 8 - reference 27). The output signal of the FU (Fig. 8 - reference 32) is then blended with the signal (Fig. 8 - reference 21) of the acoustic partial power output on area segment C. SmThe signal is subtracted and fed to the FCU (Fig. 8 - reference 30) as a signal (Fig. 8 - reference 33). The NLMS algorithm is implemented within the FCU (Fig. 8 - reference 30). The control signal (Fig. 8 - reference 23) is further convolved with the filter weights of the SPFU (Fig. 8 - reference 28). In "System Identification" operating mode, all filter weights of the SPFU are identical (1).

[0104] The output signal (Fig. 8 - reference 29) is then also fed to the FCU (Fig. 8 - reference 30). The FCU (Fig. 8 - reference 30) iteratively calculates, preferably using the NLMS algorithm, new filter coefficients from the two input signals (Fig. 8 - reference 29) and (Fig. 8 - reference 33), which are then fed to the FU (Fig. 8 - reference 27) as a signal (Fig. 8 - reference 31). This process continues until the convergence criterion defined in the ADFU (Fig. 8 - reference 26) is met.

[0105] The result is the transfer function G. mn in the form of the filter weights of the FU (Fig. 8 - reference numeral 27). These are sent to the OCU (Fig. 8 - reference numeral 34) as a signal (Fig. 8 - reference numeral 35). The data is transferred and stored electronically there together with the environmental conditions measured by the EDU (Fig. 8 - reference 38) and transferred as a signal (Fig. 8 - reference 39). The procedure is repeated sequentially for all ne {1,2,...,fV} sound intensity actuators and the me {1,2,..., M} acoustic partial powers on the sub-areas until all transfer functions contained in G have been determined. Using this methodology, a database structure is implemented in the OCU (Fig. 8 - reference 34) in which the transfer functions in G are linked to the environmental conditions.

[0106] In addition to determining the matrix G with the transfer functions G mn a second matrix G^ n ec3m x n Determined during the identification process. These are the transfer functions between the sound intensity actuators n and the 3 components of the acoustic intensity vector measured by the intensity probes m in the coordinates of the GCS (Fig. 2 - reference numeral 7). To determine G^ n The signal from the GICU (Fig. 8 - 19) with the three intensity vector components is not converted into an acoustic partial power on the partial area C. Sm The signals are not converted by the APSU (Fig. 8 - reference 19), but are transferred directly as 3 separate signals (Fig. 8 - reference 21) to the FCU (Fig. 8 - reference 30) of the AFU (Fig. 8 - reference 26). The determination of the individual transfer functions in G' nn This is then carried out in the same way as with Matrix G. mn . 2.1.2 Determining the matrix Z

[0107] The determination of the transfer functions in matrix Z proceeds almost identically to that for G. The sound intensity actuators (Fig. 9 - reference numeral 8) are driven sequentially using white noise or purely harmonic signals.

[0108] The control signals (Fig. 9 - reference numeral 23) are generated separately and sequentially for each sound intensity actuator by a signal generator SGU (Fig. 9 - reference numeral 22) and fed into the The respective amplifier units APU (Fig.9 - reference 24) are fed in.

[0109] The emission of an acoustic intensity field by the driven actuator n generates an intensity field distribution, which is measured by each of the L reference sensors in the RSU (Fig. 9 - reference 36) and transmitted as a separate signal (Fig. 9 - reference 37) to the Adaptive Filter Unit ADFU (Fig. 9 - reference 26) of the APCU (Fig. 9 - reference 25) of the SPCS system. The drive signal (Fig. 9 - reference 23) from the SGU (Fig. 9 - reference 22) is transmitted to the ADFU (Fig. 9 - reference 26) and convolved with the filter weights of the FU (Fig. 9 - reference 27).

[0110] The output signal of the frequency converter (Fig. 9 - reference 32) is then subtracted from the signal (Fig. 9 - reference 37) and fed to the frequency control unit (FCU) (Fig. 9 - reference 30) as signal (Fig. 9 - reference 33). The control signal (Fig. 9 - reference 23) is further convolved with the filter weights of the frequency converter (SPFU) (Fig. 9 - reference 28). In the "System Identification" operating mode, all filter weights of the SPFU (Fig. 9 - reference 28) are identically set to one. The output signal (Fig. 9 - reference 29) is then also fed to the FCU (Fig. 9 - reference 30). The FCU (Fig. 9 - reference 30) iteratively calculates, preferably using the NLMS algorithm, new filter coefficients from the two input signals (Fig. 9 - reference 29) and (Fig. 9 - reference 33), which are then fed to the FU (Fig. 9 - reference 27) as a signal (Fig. 9 - reference 31). This process continues until the desired value is reached in the ADFU (Fig.9 - Reference 26) the defined convergence criterion is met.

[0111] The result is the transfer function Z. ln in the form of the filter weights of the FU (Fig. 9 - reference 27). These are transferred to the OCU (Fig. 9 - reference 34) as a signal (Fig. 9 - 35) and stored electronically there together with the environmental conditions measured by the EDU (Fig. 9 - reference 38) and transferred as a signal (Fig. 9 - reference 39). Details on the measurement of the environmental conditions are in section "2.1.1 The procedures described in detail in sections “Determination of Matrix G” and “2.1.4 Further Remarks on the System Identification Process” apply analogously here. The procedure is repeated sequentially by controlling the {1,2,...,fV} sound intensity actuators until all transfer functions between the actuators and the {1,2,...,L] reference sensors contained in Z have been determined. Using this method, a database structure is implemented in the OCU (Fig. 9 - reference numeral 34) in which the transfer functions in Z are linked to the environmental conditions. The transfer functions contained in Z are required during operation of the SPCS system when using the FBFxLMS or FBFxSVD algorithm to account for or neutralize the acoustic feedback between actuators and reference sensors.

[0112] 2.1.3 Determination of the matrix H

[0113] The determination of the transfer functions in the matrix H is carried out at one or more operating points of the source S.

[0114] The acoustic intensity field generated during operation of the source is detected by the intensity probes (Fig. 10 - reference numeral 15), whereby the projected intensity components onto the individual probe axes are measured; see also Fig. 5.

[0115] The signals of the individual projected intensity components (Fig. 10 - reference numeral 16) are converted within the IAPCU (Fig. 10 - reference numeral 17) by the GICU (Fig. 10 - reference numeral 18) according to Eqs. 6 to 9 to f™, with respect to the components in the GCS (e.g., Fig. 4 - reference numeral 7). The direction vectors of the probe axes required for this are until intensity probes 1,2, are stored electronically in the OCU (Fig.10 - reference 34) and are transferred as a signal (Fig.10 - reference 41) to the GICU (Fig.10 - reference 18) of the IAPCU (Fig.10 - reference 17).

[0116] The environmental conditions at the time of system identification are measured by the EDU (Fig. 10 – 38) and transferred as signals (Fig. 10 – reference 40) to the GICU (Fig. 8 – reference 18), where they can be used further if needed. Details on the measurement of the environmental conditions are described in detail in section “1.1 Determination of Matrix G” and apply analogously here.

[0117] From the signals of the individual intensity probes (Fig. 10 - reference numeral 19), the acoustic partial powers on the partial areas C are then calculated in the APSU (Fig. 10 - reference numeral 20) according to Eqs. 10 to 13. Sm determined. The required areas of the sub-areas C SmThe surface normal vectors n™ are stored electronically in the OCU (Fig. 10 - reference 34) and are transferred as a signal (Fig. 10 - reference 41) to the APSU (Fig. 10 - reference 20) of the IAPCU (Fig. 10 - reference 17). The weighting factors w m are automatically set to 1 during the system identification of matrix G (Eq. 23) by the IAPCU (Fig. 10 - reference numeral 17) in the APSU (Fig. 10 - reference numeral 20).

[0118] The specific acoustic partial powers (Fig. 10 - reference 21) are transferred to the Adaptive Filter Unit ADFU (Fig. 10 - reference 26) of the APCU (Fig. 10 - reference 25) of the SPCS system. Furthermore, the values ​​on the sub-areas C Sm certain partial services (Fig.10 - reference 21) are transferred to the OCU (Fig.10 - reference 34) and stored there electronically.

[0119] The emission of the acoustic intensity field by the source S generates an intensity field distribution, which is directly or indirectly detected by each of the L reference sensors in the RSU (Fig. 10 - reference 36) and transmitted as a separate signal (Fig. 10 - reference 37) to the Adaptive Filter Unit ADFU (Fig. 10 - reference 26) of the APCU (Fig. 10 - reference 25) of the SPCS system. The signal (Fig. 10 - reference 37) is then convolved with the filter weights of the FU (Fig. 10 - reference 27). The output signal The signal from the frequency converter (FCU) (Fig. 10 - reference 32) is subtracted from the signal (Fig. 10 - reference 21) and fed to the frequency converter unit (FCU) (Fig. 10 - 30) as the signal (Fig. 10 - reference 33). The reference sensor signal (Fig. 10 - reference 37) is further convolved with the filter weights of the frequency converter unit (FCU) (Fig. 10 - reference 28). In the "System Identification" operating mode, all filter weights of the FCU are identical.

[0120] The output signal (Fig. 10 - reference 29) is then also fed to the FCU (Fig. 10 - reference 30). The FCU (Fig. 10 - reference 30) iteratively calculates, preferably using the NLMS algorithm, new filter coefficients from the two input signals (Fig. 10 - reference 29) and (Fig. 10 - reference 33), which are then fed to the FU (Fig. 10 - reference 27) as a signal (Fig. 10 - reference 31). This process continues until the convergence criterion defined in the ADFU (Fig. 10 - reference 26) is met.

[0121] The result is the transfer function H. mlin the form of the filter weights of the FU (Fig. 10 - reference 27). These are transferred to the OCU (Fig. 10 - reference 34) as a separate signal (Fig. 10 - reference 35) and there, together with the ambient conditions measured by the EDU (Fig. 10 - reference 38), transferred as a signal (Fig. 10 - reference 39), are stored electronically. The procedure is performed for a selected operating point of the source S for all me {1,2,..., M} acoustic partial powers P. Sm and the l E {1,2,..., L] reference sensors repeatedly until all transfer functions contained in H are determined.

[0122] The procedure is then repeated for all other selected operating points of source S, and the transfer functions are also stored electronically in the OCU (Fig. 10 - reference 34). This method creates a database structure in the OCU (Fig. 10 - reference 34) in which the transfer functions in H are linked to specific operating point data of source S and the environmental conditions. 2.1.4 Further remarks on the system identification process

[0123] As part of the system identification process, immediately before the start of regular operation of the SPCS system, the environmental conditions are also recorded by the EDU (Figs. 8 to 12 - reference numeral 38). The transfer functions contained in G, H, and Z are valid for these environmental conditions. If there are significant changes in the environmental conditions during the operation of the SPCS system (e.g., an increase in temperature, a drop in pressure, etc.), this simultaneously leads to changes in G, H, and Z. This can negatively affect the system's performance and necessitate re-identification. Depending on the medium in which the acoustic power is emitted by the source S, re-identification due to changed environmental conditions is typically required when there is a deviation of > 2.5% of the current speed of sound c. aktuell compared to the speed of sound c SysIdentduring system identification by the (9CZ / (Fig.8 to Fig.12 - reference numeral 34) in active control operation of the SPCS system.

[0124] Over a certain operating period of the SPCS system, system identifications are repeatedly carried out and the determined transfer functions in G, H and Z together with the environmental conditions are electronically stored in the OCU (Fig. 8 to Fig. 12 - reference numeral 34).

[0125] This creates a database over an extended period in which the environmental parameters are linked to the transfer functions G, H, and Z. During active control operation of the SPCS system, the EDU (Figs. 8 to 12 - reference 38) also continuously monitors the environmental conditions. If significant changes in the environmental conditions, resulting in a change in the speed of sound c in the medium and a deterioration in the SPCS system performance, are detected by the OCU (Figs. 8 to 12 - reference 34), updated transfer functions G, H, and Z can be transferred to the control core of the SPCS system by the OCU (Figs. 8 to 12 - reference 34). The updated transfer functions more closely reflect the new, current environmental conditions at the time of control by the SPCS system. The updated transfer functions contained in G, H, and Z are preferably calculated using a multidimensional interpolation method (multivariate interpolation) based on a multitude of available transfer functions and environmental conditions from previous system identifications by the SPCS system. This multidimensional interpolation method can be based, for example, on n-linear, n-cubic, spline, kriging, nearest-neighbor, radial basis functions, or inverse distance weighting interpolation, or another equivalent method.The entire described procedure achieves a learning process of the SPCS system, which after a certain period of time leads to the fact that no system identifications need to be carried out, or only to a very limited extent, because the available data space from transfer functions and associated environmental conditions is sufficiently large.

[0126] Sufficiently large means the following: The speed of sound c is calculated from the current ambient conditions using the OCU (Figs. 8 to 12 - reference numeral 34). aktuM The calculation revealed a deviation of more than 2.5% from the speed of sound c. SysIdent Detected during system identification.

[0127] From the data set of environmental conditions and transfer functions available in the OCU (Fig. 8 to Fig. 12 - reference 34), a sufficient number of data sets with environmental conditions must now be identified by the OCU (Fig. 8 to Fig. 12 - reference 34), according to the selected interpolation method and the number of data points required, for which a speed of sound c SysIdent This includes a deviation of < 2.5% from the current speed of sound c. aktuell Based on the identified datasets, approach functions for one of the listed interpolation methods are then parameterized. Subsequently, the updated transfer functions for the matrices G and H are calculated. and Z is determined by the OCU (Fig.8 to Fig.12 - reference numeral 34) by means of interpolation. 2.1.5 Processing of the data from the system identification for the active manipulation and / or reduction of the acoustic power emitted by S

[0128] The transfer functions H, Z, and G were determined in the system identification mode of the SPCS system and stored electronically in the OCU (Figs. 8 to 12 - reference numeral 34). Before actively manipulating the spectral composition and / or reducing the acoustic power emitted by S, it must be determined which of the four control methods available in the SPCS system should be used.

[0129] The stored transfer functions, essentially G or a variant thereof, are used within the SPCS system, can be directly used for Multiple Input - Multiple Output (MIMO) control, after specifying the setpoint parameters P. mso u or P mweight, soil, are used by the real-time control kernel ADFU (Fig. 8 to Fig. 10 - reference numeral 26) of the SPCS system.

[0130] A Single Input-Multiple Output (SIMO) control system is used according to the specified setpoint parameter P. soll or P weightiSOll To target, the individual lines of G are summed by the OCU (Fig. 8 to Fig. 12 - reference numeral 34), so that G SIMO as per equation 50. M G S! MO = G , 50 m=l

[0131] Regarding the SIMO regulation, the statements in section “1.2 Further remarks on the control procedure in the SPCS system” apply analogously, with the restriction that G is replaced by G SIMO It needs to be replaced.

[0132] In a single input - single output (SISO) control system, the setpoint parameters P are specified. soll or P weightiSOll, the individual rows and columns of G are summed by the OCU (Fig. 11 to Fig. 12 - reference numeral 34), so that G SISO accordingly Eq. 51 is generated. The simplest type of SISO control using the SPCS system is based on phase-coherent and source-intensity control of all sound intensity actuators. NMG SISO = 2 Z G ”’ GL51 n=lm=l

[0133] In the Single Input - Single Output - Multiple Expansion (SISO-ME) control system, the ADFU control core (Figs. 11 to 12 - reference 26) of the SPCS system provides individual control signals for each sound intensity actuator, despite internal SISO control. The division of the FU (Figs. 11 to 12 - reference 27) output signal (Figs. 11 to 12 - reference 32) into the individual control signals for the individual actuators takes place in the PSU (Figs. 11 to 12 - reference 37).

[0134] The system identification yields the transfer functions G and G. SISO known. Furthermore, in the context of identifying the transfer function H for specific operating points of the source S, the acoustic partial powers P were determined. SimiX as well as P sx stored in the OCU (Fig.11 to Fig.12 - reference 34).

[0135] The OCU (Figs. 11 to 12 - reference numeral 34) of the SPCS system allows the optimal, uniform, and phase-coherent control signal for the actuators to be determined directly "offline" according to Eq. 52a. For "feedforward" control of the SPCS system, Eq. 52b applies, where the optimal filter weights w opt°x are determined, with which the reference sensor signal s (Fig.11 to Fig.12 - reference numeral 37) is folded in the FU (Fig.11 to Fig.12 - reference numeral 27).

[0136] The resulting acoustic power P A The sound intensity actuators are given by Eq. 53. Subscript X can also represent the weighted acoustic power of source S. "SISO _ Ps,x opt,x QSISO Gl.52a vv siso = - Eq. 52b opt, X QSISO. s P — 'SISO, „SIS0 C q O pt,x GI.53

[0137] Based on a singular value decomposition of G, its inverse can be given according to Eq. 54. G -1 = V ■ S + U H ; with S + = Pseudoinverse of S GI.54

[0138] If the magnitude of the vector with the acoustic partial powers Ps,m, X' which were determined for a specific operating point of the source S, within the framework of the identification of H and are stored in the OCU (Fig.11 to Fig.12 - reference numeral 34), is normalized to 1 (Eq.55), a normalized but individual control qop™ for each sound intensity actuator can be calculated (Eq.56). tri X Ps, Norm, X 11 “ |7 Eq.55 II P S,m, X || qopt' 'x = 1 ' Ps, Norm, X GI.56 p SISO — ME _ 1 „ r XYC ■ n n °rm <31 J / ZJ U Vopt, X

[0139] The scaling factor P^ISO-ME is then calculated. nac|1Q| 57 is calculated for the specific operating point of source S. Equations 50 to 57 are calculated "offline" in the OCU (Figs. 11 to 12 - reference 34) and qop™ and psiso-ME are stored electronically in it. In active SISO-ME control mode of the SPCS system, the output signal (Figs. 11 to 12 - reference 27) is calculated by the frequency converter (Figs. 11 to 12 - reference 32) and transferred to the PSU (Figs. 11 to 12 - reference 37). In the PSU (Figs. 11 to 12 - reference 37), the signal is split into the individual control signals for the sound intensity actuators. c lopt°x~ ME nac h Eq. 58. In this context, division refers to the This refers to the specification of an individual amplitude and phase position for each actuator. “SISO—ME _ n norm riSISO—ME n _ n norm riSISO—ME r'SISO “SISO / - / co R opt, X — Hopt, X P X — q Opt, X P X ü Hopt, XU / .5Ö

[0140] qop S t°xi n Eq. 58 is the output signal (Fig. 11 to Fig. 12 - reference numeral 32) from the frequency converter (Fig. 11 to Fig. 12 - reference numeral 27). qopt™' Fy IS0 ~ ME as well as G SISO are transferred by the OCU (Fig.11 to Fig.12 - reference 34) to the PSU (Fig.11 to Fig.12 - reference 37) as signals (Fig.11 to Fig.12 - reference 43) before the start of the active control operation of the SPCS system. 2.2 Manipulation and / or reduction of the acoustic power emitted by source S by the SPCS system

[0141] After the SPCS system has completed system identification, the spectral composition of the acoustic power emitted by source S can be manipulated and / or reduced. In the following explanations, we assume that the SPCS system operates in "feedforward" mode with "feedback cancellation" (Fig. 11). However, the SPCS system can also operate in "feedback" mode or in a hybrid mode combining "feedforward" with synthetic reference sensor signals. "Feedback" means that the SPCS system controls the acoustic power emitted by S without any reference sensor signal (Fig. 11 - reference numeral 37).

[0142] In hybrid mode (Fig. 12), the system also operates without a direct reference sensor signal (Fig. 11 - reference numeral 37), but synthetic reference signals (Fig. 12 - reference numeral 23) are provided via the SGU (Fig. 12 - reference numeral 22). These are preferably purely harmonic signals with fixed amplitudes, frequencies, and phase angles.

[0143] Before commissioning the SPCS system, the user determines which of the possible control methods SISO, SISO-ME, SIMO or MIMO should be used. Furthermore, The method by which the FCU (Fig. 11 - reference numeral 30) of the control core ADFU (Fig. 11 - reference numeral 26) of the SPCS system is to operate, based on FBFxLMS or FBFxSVD, is specified; see also “1.2 Further remarks on the control method in the SPCS system”. Furthermore, the user must specify which of the L reference sensors is to be used for active control by the SPCS system. 2.2.1 Explanations regarding the use of the FBFxLMS control kernel

[0144] If the control method 3 (SIMO) is used in combination with the FBFxLMS algorithm, the Tikhonov regularization parameter I|J must be set according to Eq. 43. This can either be done by the user of the SPCS system or is set automatically by the OCU (Fig. 11 - reference numeral 34).

[0145] If control method 4 (MIMO) is to be used, the Tikhonov regularization parameter I|J must also be specified if the number of sound intensity actuators N is greater than the number of intensity probes M (N > M). This can again be done by the user of the SPCS system or is automatically determined by the OCU (Fig. 11 - reference 34).

[0146] In the case N < M, the SPCS system can be used with Tikhonov regularization or only with the Moore-Penrose pseudoinverse ( I|J = 0). If Tikhonov regularization is used, I|J must be specified by the user or is automatically set by the OCU (Fig. 11 - reference 34). When using Moore-Penrose pseudoinverses, I|J = 0 is automatically set by the OCU (Fig. 11 - reference 34).

[0147] For the case N = M, the SPCS system can operate using Tikhonov regularization, Moore-Penrose pseudoinverses, or a direct solution. If Tikhonov regularization is used, I|J must be specified by the user or is automatically determined by the OCU (Fig. 11 - reference 34). When using Moore-Penrose pseudoinverses, I|J = 0 is also automatically set by the OCU (Fig. 11 - reference 34).

[0148] The mathematical relationships are explained in "Details on solution variant 1-3". 2.2.2 Explanations regarding the use of the FBFxSVD control core

[0149] When using an SVD-based control kernel (FBFxSVD) within the FCU (Fig. 11 - reference numeral 30), no user-defined settings are required by default. Optionally, the SVD control kernel can operate in TSVD mode, as described in section "1.2 Further Remarks on the Control Method in the SPCS System". Both SVD variants can operate either directly based on the transfer functions of matrix G or based on matrices modified by Tikhonov regularization; see also "Applications of SVD in the SPCS Control Kernel - Option 2".

[0150] When working in TSVD mode, the user must specify how many singular values ​​should be used to form G^SVD, or this is automatically determined by the OCU (Fig.11 - reference 34).

[0151] If Tikhonov regularization is used within the SVD, I|J must be specified by the user or is automatically determined by the OCU (Fig. 11 - reference numeral 34) and G^ is used. ksVD , or in the TSVD variant G^ kTSVD Definitely. The optional features can only be used with the SIMO and MIMO control methods.

[0152] The matrices SVD and MIMO are therefore available for the MIMO control procedure. T SVD, MIMO' GT ISVD, MIMO or G T ^kTSVD MIM0 for selection. For simplicity, only G^ will be used in the following. S VD, MIMO, where all four variants can be meant. The matrices G are correspondingly represented. S y D SIM0 , G T VD SIM0 , G T ^ ksVD SIM0 or G T ^kTSVD SIM0 when using the SIMO control procedure for selection.

[0153] For simplicity, the following will only refer to GX VD and SIMO, although this can also refer to all four variants. The SISO control procedure only uses Gy. D SIS0 to select. 2.2.3 Further explanations regarding the configuration of the SPCS system before the start of active control

[0154] The target values ​​P XiSO ii according to Eq. 33 on which the SPCS system is to rule, as well as the weighting factors w m for the weighting of the acoustic partial performances P m on sub-areas C Sm are stored electronically in the OCU (Fig. 11 - reference 34) and are transferred as a signal (Fig. 11 - reference 41) to the APSU (Fig. 11 - reference 20) of the IAPCU (Fig. 11 - reference 17). The required areas of the sub-areas C SmThe surface normal vectors n™ are also stored electronically in the OCU (Fig. 10 - reference 34) and are transferred as a signal (Fig. 11 - reference 41) to the APSU (Fig. 11 - reference 20) of the IAPCU (Fig. 11 - reference 17). If no weighting factors are specified, w m = 1 is automatically specified by the APSU (Fig. 11 - reference numeral 20). If the SPCS system is specified with the weight factors w m If the acoustic power emitted by S is to be manipulated and / or reduced in its spectral composition, this is preferably done only when using the SISO, SIMO or MIMO control method.

[0155] The environmental conditions are continuously measured by the EDU (Fig. 11 - reference 38) during active control of the SPCS system and transferred as signals (Fig. 11 - reference 40) to the GICU (Fig. 11 - reference 18), where they can be used further as needed. Details on the measurement of the environmental conditions are described in detail in section "2.1.4 Further Remarks on the System Identification Process" and apply analogously here.

[0156] If real-time control is achieved using the FBFxLMS algorithm, the matrix G is used by the OCU (Fig. 11 - reference numeral 34); in the case of MIMO control, the matrix G is used. SIMO corresponding to Eq. 50 in SIMO control or matrix G SISO In SISO control according to Eq. 51, the signal (Fig. 11 - reference 41) is transferred to the SPFU (Fig. 11 - reference 28) and FCU (Fig. 11 - reference 30). The matrices are processed in both the SPFU (Fig. 11 - reference 28) and the FCU (Fig. 11). - Reference numeral 30) is stored electronically. If regularization according to Tikhonov is used, the defined parameter is indicated via the signal (Fig. 11 - reference numeral 41). For the solution variant using the Moore-Penrose pseudoinverse I|J = 0, the data is transferred to the FCU (Fig. 11 - 30) and stored electronically there. Furthermore, the transfer functions of matrix Z belonging to the selected reference sensor are transmitted via the signal (Fig. 11 - reference 41) to the FBFU (Fig. 11 - reference 44), which is required for feedback cancellation.

[0157] If real-time control is achieved using the FBFxSVD algorithm, the matrix G^ is calculated by the OCU (Fig. 11 - reference numeral 34). S VD_MIMO' in MIMO control, the matrix GX * S VD_SIMO in SIMO control or the Gsy matrix D SIS0In SISO control, the signal (Fig. 11 - 41) is transferred to the FCU (Fig. 11 - reference numeral 30) and stored electronically there; see also “1.2 Further remarks on the control procedure in the SPCS system”.

[0158] If Tikhonov regularization is also used in connection with SVD, the matrix G or G is additionally specified by OCU (Fig. 11 - reference numeral 34). SIMO as a signal (Fig.11 - reference numeral 41) to the FCU (Fig.11 - reference numeral 30) and stored electronically there in order to calculate the modified error signal according to Eq.47a when control is active.

[0159] Furthermore, when using the FBFxSVD algorithm, the matrix Z is transferred via the signal (Fig. 11 - reference 41) to the FBFU (Fig. 11 - 44), which is required for feedback cancellation. The SPFU (Fig. 11 - reference 28) is initialized to 1 when using the FBFxSVD algorithm, since the convolution operation between the signal (Fig. 11 - reference 46) and the SPFU (Fig. 11 - reference 28) is performed indirectly in the FCU (Fig. 11 - 30) compared to the FBFxLMS solution variant. The SPFU (Fig. 11 - reference 28), which is initialized to 1, therefore has no influence on the signal (Fig. 11 - reference 46), or rather, the signal (Fig. 11 - reference 46) is identical to the signal (Fig. 11 - reference 29).

[0160] If the SPCS system is operated in the SIMO-ME control method, the OCU (Fig. 11 - reference numeral 34) processes the “offline” calculated quantities qop™^ p according to Eq. 58 and the explanations in “2.1.5 Processing of the data from the system identification for the active manipulation and / or reduction of the acoustic power emitted by S”. EIS0 ~ ME as well as G SIS0 The signal is transferred to the PSU (Fig. 11 - reference 37) and stored electronically there. In the PSU (Fig. 11 - reference 37), during active control, the signal (Fig. 11 - reference 32) is split into the signal (Fig. 11 - 42) for individual control of the individual sound intensity actuators.

[0161] If, during active control operation of the SPCS system, the EDU (Fig. 11 - reference 38) measures a significant change in the environmental conditions and detects it based on the signal (Fig. 11 - reference 39) by the OCU (Fig. 11 - reference 34), then either the system identification procedure must be repeated, or the data set available in the OCU (Fig. 11 - reference 34) must be large enough to calculate the transfer functions contained in matrices G, H, and Z for the new environmental conditions using interpolation methods; see also "2.1.4 Further Remarks on the System Identification Process". The newly identified transfer functions, or those calculated by interpolation by the OCU (Fig. 11 - reference 34) in matrices G, H, and Z, are transferred to the individual components of the SPCS system as described above and stored there electronically.The update of the transfer functions in the matrices G, H and Z or derived quantities in the individual components of the control kernel takes place without interrupting the active control operation of the SPCS system. 2.2.4 Explanations regarding the active control of the SPCS system

[0162] In active control operation, the acoustic intensity fields emitted by source S and sound intensity actuators M are superimposed on the sub-areas C. Sm (Fig. 4 - reference numeral 11). The intensity probes (Fig. 11 - reference numeral 15) measure the projected, superimposed intensity components onto the individual probe axes; see also Fig. 5.

[0163] The signals of the individual projected intensity components (Fig. 11 - reference numeral 16) are converted within the IAPCU (Fig. 11 - reference numeral 17) by the GICU (Fig. 11 - reference numeral 18) according to Eqs. 6 to 9 or Eqs. 17 to 20 to I™, with respect to the GCS (Fig. 1 to Fig. 5 - reference numeral 7). The direction vectors of the probe axes n^, n™ required for this are k to n™ K intensity probes 1,2, The values ​​are stored electronically in the OCU (Fig. 11 - reference 34) and are transferred as a signal (Fig. 11 - reference 41) to the GICU (Fig. 11 - reference 18) of the IAPCU (Fig. 11 - reference 17) before the start of active control. The subscript T stands for Total and refers to the superposition of the intensity fields of the source S and the M sound intensity actuators.

[0164] Equations 6 to 9 (subscript and superscript S) and 17 to 20 (subscript and superscript A) are formally identical and were only introduced to distinguish between the intensity fields of source S and the sound intensity actuators, but can be used interchangeably to describe the conversion of the superimposed intensity fields to I™.

[0165] The environmental conditions are measured by the EDU (Fig.11 - reference 38) in the active control operation of the SPCS system and transferred as signals (Fig.11 - reference 40) to the GICU (Fig.11 - reference 18), where they can be used further if required.

[0166] From the signals / ” l The acoustic partial powers of the individual intensity probes (Fig. 11 - reference numeral 19) are then calculated in the APSU (Fig. 11 - reference numeral 20) according to Eqs. 25 to 28 on the partial areas C. Smor determines the overall performance. The required area of ​​the sub-areas C Sm , the surface normal vectors n™ and the weighting factors w m are stored electronically in the APSU (Fig. 11 - reference 20) of the IAPCU (Fig. 11 - reference 17). For the SISO, SISO-ME and SIMO control procedures, depending on the user's specifications, either the unweighted, summed total acoustic power P or the weighted variant P is used. weight as a signal (Fig.11 - reference 21) into the FCU (Fig.11 - reference 30) of the ADFU (Fig.11 - reference 26), which represents the actual control core of the APCU (Fig.11 - reference 25).

[0167] If the MIMO control method is used, either the unweighted acoustic partial powers P are used, depending on the user's specifications. m or the weighted partial performances P miWeightas separate signals (Fig.11 - reference 21) into the FCU (Fig.11 - reference 30) of the ADFU (Fig.11 - reference 26).

[0168] The reference sensor signal (Fig. 11 - reference 37) is transferred from the RSU (Fig. 11 - reference 36) to the node (Fig. 11 - reference 46). There, it is subtracted from the output signal (Fig. 11 - reference 45) from the FBFU (Fig. 11 - reference 26) to neutralize the feedback from the sound intensity actuators. The resulting signal (Fig. 11 - reference 47) is then forwarded to the SPFU (Fig. 11 - reference 28).

[0169] In the event that the FCU (Fig. 11 - reference numeral 30) operates on the basis of the FBFxLMS algorithm, the convolution with the transfer functions of matrix G in MIMO control takes place here. SIMOIn SIMO control or in SISO or SISO-ME control, the convolution with the transfer function G SIS0 . In the case that the FCU (Fig.11 - reference 30) operates on the basis of the FBFxSVD algorithm, the convolution weights in the SPFU (Fig.11 - reference 28) are initialized with 1, so that the signals (Fig.11 - reference 47) and (Fig.11 - reference 29) are identical.

[0170] The result of the convolution operation is transferred as a signal (Fig. 11 - reference 29) to the FCU (Fig. 11 - reference 30). Within the FCU (Fig. 11 - reference 30) the Calculation of the filter weights, which are transferred as a signal (Fig. 11 - reference 31) to the frequency converter (FCU) (Fig. 11 - reference 27). If the calculation of the filter weights is based on the FBFxLMS algorithm, then equations 42b are used in the FCU (Fig. 11 - reference 30), taking into account equations 47a and 47b, when using Tikhonov regularization or the Moore-Penrose pseudoinverse. Equation 46b is used for the direct solution using the FBFxLMS algorithm.

[0171] If the FBFxSVD algorithm is used within the FCU (Fig.11 - reference numeral 30), the filter weights for the FU (Fig.11 - reference numeral 27) are calculated on the basis of Eq.48 and Eq.49 and taking into account the explanations in “2.2.2 Explanations on the use of the FBFxSVD control core”.

[0172] After the feedback neutralization, the signals (Fig. 11 - reference 47) are also transferred to the frequency converter (Fig. 11 - reference 27), where convolution with the filter weights calculated by the frequency control unit (FCU) (Fig. 11 - reference 30) takes place. In the case of SISO or SISO-ME control, the output signal (Fig. 11 - reference 32) from the frequency converter (Fig. 11 - reference 27) is a single signal. In the case of SIMO or MIMO control, a separate signal is provided for each sound intensity actuator, bundled together as a single signal (Fig. 11 - reference 32). The signal (Fig. 11 - reference numeral 32) is transferred from the FU (Fig. 11 - reference numeral 27) to the PSU (Fig. 11 - reference numeral 37) and the FBFU (Fig. 11 - reference numeral 44). In the case of SISO-ME control, the calculation of the individual signals for each of the sound intensity actuators is carried out according to Eq. 58 and the explanations in "2.1".5. Processing of the data from the system identification for the active manipulation and / or reduction of the acoustic power emitted by S. In the case of SISO control, the input signal (Fig. 11 - reference 32) is duplicated unchanged for each of the sound intensity actuators in the PSU (Fig. 11 - reference 37). In SIMO or MIMO control, the individual and separate actuators are already used. A control signal is provided for each sound intensity actuator, so that the individual signals (Fig. 11 - reference 32) are only forwarded by the PSU (Fig. 11 - reference 37). The separate output signals from the PSU (Fig. 11 - reference 37) are transferred as a signal or signal bundle (Fig. 11 - reference 32) to the amplifier unit APU (Fig. 11 - reference 24) of the associated sound intensity actuator. There, the signal is converted into a usable signal for the sound intensity actuator, and finally, the desired acoustic intensity field is emitted.

[0173] The signal (Fig. 11 - reference 32), also transferred from the frequency converter (FTC) (Fig. 11 - reference 27) to the frequency converter (FBFC) (Fig. 11 - reference 44), is convolved and superimposed in the FBFC (Fig. 11 - reference 44) with the transfer functions for the selected reference sensor of matrix Z to calculate the feedback of the sound intensity actuators to the reference sensor signal. It should be noted again that in the case of SIMO or MIMO control, the signal (Fig. 11 - reference 32) contains an individual and separate signal for each sound intensity actuator. The output signal (Fig.11 - reference 45) of the FBFU (Fig.11 - reference 44) is then transferred to the node (Fig.11 - reference 46) and subtracted there from the reference sensor signal (Fig.11 - reference 37) provided by the RSU (Fig.11 - reference 36). 2.2.5 Explanation of the active control of the SPCS system in hybrid mode

[0174] In hybrid mode (Fig.12), control is carried out by the SPCS system without a direct reference sensor signal (Fig.11 - reference numeral 37), however, synthetic reference signals (Fig.12 - reference numeral 23) are provided via the SGU (Fig.12 - reference numeral 22).

[0175] These are preferably purely harmonic signals with precisely defined amplitudes, frequencies, and phases. Since, in this case, no feedback occurs from the sound intensity actuators to the (in this case, non-existent) reference sensors. The filter weights in the FBFU (Fig. 12 - reference 44) are initialized to 0. Therefore, the signal (Fig. 12 - reference 23) before the node (Fig. 12 - reference 46) is identical to the signal (Fig. 12 - reference 47) after it. No further changes occur in hybrid mode compared to true feedforward control according to Fig. 12 and the accompanying explanations. 2.2.6 Explanation of the active control of the SPCS system in feedback mode

[0176] The SPCS system can also be configured for feedback control (Fig. 13). In this case, control is based solely on the error signals (Fig. 13 - reference numeral 21) and the transfer functions from the matrices G, G. SIMO or G s / so which are stored electronically in the FCU (Fig. 13 - reference 30). The output signal (Fig. 13 - 31) from the FCU (Fig. 13 - reference 30) already corresponds to the input signal (Fig. 13 - reference 32) into the PSU (Fig. 13 - reference 37) and is merely forwarded by the FU (Fig. 13 - reference 27). The FCU (Fig. 13 - reference 30) thus directly calculates the control signals for the sound intensity actuators. The basic signal processing sequence continues to correspond to Fig. 12 and the accompanying explanations. 2.2.6.1 Explanation of the signals (Fig. 11 to Fig. 13 - reference numeral 48) and (Fig. 11 to Fig. 13 - reference numeral 49)

[0177] As previously described, within the IAPCU (Figs. 11 to 13 - reference 17), the signals of the individual projected intensity components (Figs. 11 to 13 - reference 18) are converted to I™, with respect to the GCS (Figs. 1 to 5 - reference 7), by the GICU (Figs. 11 to 13 - reference 18) according to Eqs. 6 to 9 or Eqs. 17 to 20. These are transferred as separate signals (Figs. 11 to 13 - reference 48) to the OCU (Figs. 11 to 13 - reference 34). There, the OCU (Figs. 11 to 13 - reference 34) processes the signals based on the transfer function G and the known control signals of the sound intensity actuators, which are transferred as a signal (Fig. 11 to Fig. 13 - reference numeral 49) into the OCU, the vectors 7” 1 calculated.

[0178] These are the intensity vectors measured by the M intensity probes solely based on the emission of acoustic power by the sources S. These vectors are then used by the OCU (Figs. 11 to 13 - reference numeral 34) to locate the position and any changes in position of the source S in space.

[0179] To illustrate the procedure within the OCU (Figs. 11 to 13 - reference 34), Figs. 14 and 15 are used. In Fig. 14, the hydroacoustic intensity vectors from a measurement in front of a ship's propeller are shown in two views. It can be observed that the intensity vectors point away from the acoustic source. If the intensity vectors are traced back to their "origin" (backtracking), it can be seen that they converge at the center of the propeller or the acoustic source S. Mathematically, the problem is solved by the OCU (Figs. 11 to 13 - reference 34) such that, based on two measured intensity vectors, e.g., Z" l=1 and Z" l=2 two straight lines g t and g2 (Fig. 15) are defined in parametric form. In general, the lines g t and g2 is skew due to errors in intensity measurement. For the straight lines g tand g2 are then iteratively determined by the OCU (Fig. 11 to Fig. 13 - reference numeral 34) to find the point coordinates P 12 min with regard to the GCS (Fig. 1 to Fig. 5 - reference numeral 7) determines the distance between g t and g2minimal; see Fig. 15. This procedure is repeated for all possible two combinations of measured intensity vectors / Q. The graphical representation of all found points of minimum distance between two selected intensity vectors and the lines derived from them results in the light gray point distribution shown in Fig. 14. The position of the source S in space with respect to the GCS (Figs. 1 to 5 - reference numeral 7) is calculated by averaging the coordinates of all found points of minimum distance. In Fig. 14, this is the red dot, which is located in the center of the propeller.

[0180] This procedure is performed continuously within the OCU (Figs. 11 to 13 - reference 34) during active control, but with a different clock frequency than the active "real-time" control kernel ADFU (Figs. 11 to 13 - reference 26). If changes in the determined position of the source S are detected, the OCU (Figs. 11 to 13 - reference 34) can calculate updated transfer functions for the matrices G, H, and Z. The calculation of the updated transfer functions can be based on analytical acoustic models, which, for example, use combinations of acoustic monopole, dipole, quadrupole, or multipole sources to describe the acoustic emission of the source S with sufficient accuracy. Alternatively, numerical models for solving the inhomogeneous acoustic wave equation can also be used.The boundary element method (BEM), the finite element method (FEM), the finite difference method (FDM), or any other suitable method can be used to solve the inhomogeneous acoustic wave equation. The update of the transfer functions in the matrices G, H, and Z, or derived quantities, in the individual components of the control kernel is performed without interrupting the active control operation of the SPCS system. Reference sign SPCS - Sound Power Control System APCU = Acoustic Power Control Unit APCU = Acoustic Power Control Unit IAPCU = Intensity and Power Calculation Unit GICU = Global Intensity Calculation Unit APSU = Acoustic Power Summation Unit SGU = Signal Generation Unit OCU = Offline Calculation Unit RSU = Reference Sensor Unit SPFU = Secondary Path Filter Unit FBFU = Feedback Filter Unit FCU = Filter Calculation Unit FU = Filter Unit ADFU = Adaptive Filter Unit APU = Actuator Power Unit PSU = Phase Shift Unit EDU = Environment Data Unit GCS = Globales Koordinatensystem

Claims

Claims 1. Methods for manipulating the spectral composition and / or reducing the acoustic power P emitted by a source S s by means of a system for manipulating and / or reducing the spectral composition of a first acoustic power P emitted by an acoustic source S s , showing the following procedural steps: a. Defining an envelope area C s , which is a virtual, physically non-existent, area for accounting for the first acoustic power P emitted by the source S s , where the enclosing surface C s the source S completely encloses, b. Arrangement of M intensity probes I 1 , I 2 ,..., I m , I M on surface C s , where each of the M intensity probes I 1 , I 2 ,..., I m , l M a sub-area segment C SA , C S2 > C SiM the envelope area C s is assigned c. Measuring an incident first intensity vector I l / s" fc until I™ K Axis directions) using the M intensity probes and determination of its components of (P s n x , I™ Y and I™ z ) with respect to a reference coordinate system and measuring an incident second intensity vector I™ (mitz, up to I™ axis directions) using the M intensity probes and determining its components of 1 (l^ x , and I ™ z ) with respect to a reference coordinate system, d. Calculating the first acoustic power P s, by summing individual first acoustic partial powers P S m , which extend over the sub-area segments C S / m are emitted, and by means of — I n O c S,m be calculated, and calculate a second acoustic power P A of N sound intensity actuators A1, A2,..., A n , A N "by summing individual second acoustic partial powers P" Am , which extend over the sub-area segments C Siin are emitted, and by means of r PA.,m — — II 'A ■ n rl 0 m ■ d u / A i '-'S m are calculated, where n™ are surface normal vectors. e. Active control of the system by means of a control system, such that a total emitted acoustic power P, formed from the first acoustic power P sthe source S and the second acoustic power P A The N sound intensity actuators are manipulated or minimized according to: p = Im=i p m, whereby P m = f c (j™ + f™) " = P T l " = P S,m + P A,m> ^S,m L, S f m where control parameters or source strengths for the N sound intensity actuators A lt A 2t ..., A n , A N based on P, P m or weighted sizes of P or P m These control parameters or source strengths are determined and sent to the N sound intensity actuators A1, A2,..., A n , A N These parameters are transmitted in order to manipulate the emitted acoustic power.

2. Method according to claim 2, characterized in that each surface segment C S1 , C Si2 , ''', CSrm , C S M a surface normal vector n, n^, ™o 1 ' is assigned, where nj, n^, •••, n™, n" is constant over the respective assigned surface segment C Sil , C S 2 , C SM is.

3. Method according to claim 1 or 2, characterized in that the enveloping surface C s for the division into area sub-segments C S 1 , C S 2 , CS, M m 't is discretized under the assumption that acoustic state variables, in particular sound pressure, particle velocity, density and / or temperature, are to a first approximation constant over all surface sub-segments C Sm are.

4. System for manipulating and / or reducing the spectral composition of a first acoustic power P emitted by an acoustic source S sfor carrying out the method according to claim 1, wherein the first acoustic power P s a first acoustic intensity field I Sri trains, exhibiting system N sound intensity actuators A lt A 2t ..., A n , A N , where each sound intensity actuator J 41, J 42, - > A nr A N each with a distance ß Si i,Ä Sj2 , - > R s,n> R s, N and a direction of radiation around the source S, but within a surface C s , are arranged, where each of the N sound intensity actuators A1, A2,..., A nt A N is set up to create a second acoustic intensity field defined in amplitude, frequency and phase l A 1 , l A n until jjV with a second acoustic power P A to create M sound intensity probes 7 1 , 1 2 , I m , I M , which are arranged on the surface of Cs, where the M sound intensity probes 7 1 , I 2 , I m , I M each in K axis directions, measure the component of an incident intensity vector Z™ projected onto the respective axis direction. and at least one device for electronic signal processing and active control, wherein the at least one device for electronic signal processing and active control is configured to ensure that a total emitted acoustic power P, formed from the first acoustic power P s the source S and the second power P A the N sound intensity actuators are manipulated or minimized.

5. System according to claim 4, characterized in that the acoustic source S is composed of L individual acoustic sources.

6. System according to claim 4, characterized in that the emitted first acoustic power P s is formed from a first active acoustic power P s active or reactive performance P s reactive or a sum of both according to Ps = Ps-active + j ' Ps reactive (with j 2 = -1 imaginary unit).

7. System according to claim 4, characterized in that the emitted first acoustic power P s the integral of the intensity of the source S / s .i(C s ) via the virtually conceived surface C s is.

8. System according to claim 4, characterized in that the N sound intensity actuators A2,..., A n , A N within the area defined by C s are distributed over a limited area using a probabilistic method, preferably using a Poisson-Disc Sampling method.

9. System according to claim 4, characterized in that a spatial orientation of the N actuators A ll A2,..., A n , A N can be determined by means of a predetermined probability distribution, preferably a Poisson distribution or a Gaussian distribution.

10. System according to claim 4, characterized in that the N sound intensity actuators A1, A2,..., A n , A N Controllable systems are those with which an acoustic intensity field defined in amplitude, frequency and phase can be generated. Ail , l Ai2 , l A , n , I AiN is generated individually by each sound intensity actuator.

11. System according to claim 4, characterized in that the controllable systems are loudspeakers, underwater loudspeakers and / or structure-borne sound transducers.

12. System according to claim 4, characterized in that each sound intensity probe I 1 , 1 2 , I m, I M on each surface segment C S 1 , C S2 , “', C Sm , C SM is arranged.

13. System according to claim 4, characterized in that each sound intensity probe I 1 , 1 2 , I m , I M is designed as uniaxial (one-dimensional), biaxial (two-dimensional), triaxial (three-dimensional) or multiaxial (multi-dimensional).

14. System according to claim 4, characterized in that the M sound intensity probes are sensors or sensor combinations for measuring an acoustic active and reactive intensity, The preferred M sound intensity probes are pu probes and / or pp probes.

15. System according to claim 4 further comprising reference sensors, wherein the reference sensors correspond to the number of L individual sources.

16. System according to claim 15, characterized in that the reference sensors are configured to measure quantities which are directly related to the acoustic emission of the source S.

17. System according to claim 16, characterized in that the L reference sensors are intensity probes, microphones, accelerometers, velocity sensors, displacement sensors, pressure sensors, hydrophones, speed sensors and / or laser vibrometers.

18. Control method of a system according to claim 4 and / or of a system in a method according to claim 1, wherein control parameters for the N sound intensity actuators A-ZQ,..., A n , A N The following procedural steps will be used to determine this: a. A system identification is performed, whereby the following transfer functions (in the frequency domain) are determined: the transfer function G between the N sound intensity actuators A lr A2,..., An , A N and the partial powers P measured by the sound intensity probes A m , where a G-matrix is ​​determined to determine the transfer functions, b. Calculation of the control parameters q of the N sound intensity actuators A1, A2,..., A n , A N by solving a system of equations according to e = Ps,x + G ■ q - P x ,soii = Px + G ■ q = 0 using the transfer function G, such that an error signal e is minimized or becomes zero, where P SiX Ps,m> Ps, m, weight' Ps, weight or P is and P x , S(M a target value for P s.m , P St “ igh t, p s , weight or P.

19. Control method according to claim 18, characterized in that furthermore the transfer function 2 between the N sound intensity actuators and the L acoustic sources or L reference sensors and / or the transfer function H between the first acoustic source S, composed of up to L partial sources or detected via reference sensors, and the measured acoustic partial powers P Sm or P s , m , weight' where a Z- and / or H-matrix is ​​determined for the determination of the transfer functions z and H.

20. Control method according to claim 19, characterized in that the control is carried out by means of a Single Input - Single Output method (SISO) or by means of a Single Output - Multiple Expansion method (SISO-ME) or by means of a Single Input - Multiple Output method (SIMO) or by means of a Multiple Input - Multiple Output method (MIMO).

21. Control method according to one of claims 18-20, characterized in that the control method is operated a. without reference sensors (feedback operating mode) or b. with reference sensors (feedforward operating mode) or c. without reference sensors but with a synthetic reference signal generated by the control system (hybrid operating mode).

22. Control method according to one of claims 18 - 21, characterized in that the calculation of the control parameters q for the actuators is carried out on the basis of the error signals e by a. an FBFxLMS algorithm using a Tikhonov regularization or Moore Penrose pseudoinverses or b. a mapped FBFxSVD algorithm using a singular value decomposition (SVD), truncated SVD (TSVD) and / or in combination with a regularization according to Tikhonov.

23. Control method according to claim 20, characterized in that the methods are transferable to any linear problems of the form A ■ x = b for which a real-time capable solution is sought.

24. Control method according to claim 23, characterized in that it is transferable to classic ANC problems for minimizing sound pressure using the LMS method as well as in SIMO and MIMO systems.

Images