A vibration active control method, test method, device, vehicle, equipment and medium based on a feedback FXLMS algorithm

Abstract

The application provides a vibration active control method, a test method, a device, a vehicle, equipment and a medium based on a feedback FXLMS algorithm, can reproduce the required reference signal input of the FXLMS algorithm, realizes the calculation and real-time adjustment of the control signal, avoids the difficulty of obtaining the reference signal of the feedforward control system, and solves the problem of instability in the actual application of AVC. The method comprises the following steps: acquiring the initial response of the measured plate caused by external excitation by continuously collecting the initial response of the measured plate caused by external excitation in the current time step of the error sensor; according to the initial response and the pre-set primary path transfer characteristic function, the excitation signal applied to the measured plate by the external excitation in the current time step is reproduced; the primary path is the transmission path from the excitation point of the external excitation on the measured plate to the error sensor; all the excitation signals reproduced in the current time step are input into the FXLMS algorithm as the reference signal, and the control signal output to the actuator in the current time step iteration is obtained.

Description

Technical Field

[0001] This invention belongs to the field of vehicle NVH technology and is used to reduce low-frequency vibration and radiated sound of thin-walled panels in automobiles. Specifically, it involves an FXLMS adaptive vibration active control method, test method, device, vehicle, equipment and medium based on feedback control. Background Technology

[0002] With the continuous development of the automotive industry, vehicle comfort has become an important evaluation indicator. NVH performance significantly impacts the passenger experience. Controlling NVH performance generally involves reducing noise transmitted to the ears of passengers through various means, minimizing discomfort caused by vibration noise. In addition, vibrations of thin-walled components (such as the headliner, floor, or tailgate) can generate low-frequency noise. Low-frequency noise below 30Hz is usually inaudible but can cause discomfort such as ear pressure, while low-frequency noise between 30-50Hz can produce a drumming sound, resulting in a poor passenger experience. Traditional methods for controlling low-frequency noise include adding mass blocks or vibration absorbers, which lack flexibility and increase vehicle weight, hindering weight reduction. With the development of digital filtering and intelligent technologies, Active Noise Control (ANC) technology can be used to improve automotive NVH performance.

[0003] ANC (Active Noise Cancellation) technology includes two types: source control and noise reduction. Active Vibration Control (AVC) is a source control type that can reduce low-frequency vibrations and radiated noise caused by board vibration. The principle of AVC technology is to generate a control force through a controller, which acts on the controlled object (vibrating board) to cancel out the vibration response generated by the original excitation, thereby reducing the vibration velocity of the board surface and suppressing the radiated noise. The AVC system measures the vibration velocity of the controlled object's surface through sensors, transmits the vibration velocity signal to the controller, and the controller processes the input signal through a corresponding control algorithm, outputting a control force that acts on the controlled object to achieve the vibration control effect.

[0004] Classic control types include feedforward control and feedback control. Feedforward control is an open-loop system that requires the controller to be designed based on external disturbance signals, thus necessitating the acquisition of reference signals related to the excitation source beforehand. Feedback control, on the other hand, is a closed-loop control method that does not require prior acquisition of external excitation information and can achieve control based solely on the error signal. FXLMS (Filter-LeastMean Square) is a classic ANC control algorithm that processes the input signal through filtering and outputs a control signal. As an adaptive control algorithm, FXLMS can adjust control parameters in real time based on the system's response characteristics, making it more practical than classic control algorithms (such as PID control).

[0005] The FXLMS algorithm requires weighted processing of the reference signal, and therefore is usually combined with a feedforward system. For example, CN109932906 proposes a feedforward FXLMS algorithm based on extended secondary channels to control the actuators in the active mounts of an automotive engine. This allows the actuators to generate an excitation force that can counteract the vibration response, thereby suppressing engine vibration through the active mounts. While this method has achieved good results in simulations, in engineering applications, the excitation sources and paths causing vibration in automotive components are often complex, making it difficult to obtain the reference signal in advance, resulting in unstable feedforward control performance.

[0006] The reference signal in the FXLMS method needs to be related to the external excitation. To address the difficulty in measuring the external excitation in real time, CN110794681A proposes a feedforward FXLMS filtering method for narrowband active vibration systems. This method synthesizes a reference signal based on the vibration response signal, replacing the reference signal acquisition at the input end, and achieves good vibration suppression. However, the reference signal in this method is artificially synthesized and differs from the actual excitation signal, leading to insufficient control accuracy and effect. Summary of the Invention

[0007] To address the aforementioned problems, this invention proposes a vibration active control method, testing method, device, and vehicle based on the feedback FXLMS algorithm. By reproducing the reference signal input required by the FXLMS algorithm, the control signal can be calculated and adjusted in real time, avoiding the problem of instability in practical AVC applications caused by the difficulty in obtaining reference signals in feedforward control systems.

[0008] The technical solution of this invention is as follows:

[0009] This invention provides a vibration active control method based on the feedback FXLMS algorithm, comprising:

[0010] The error sensor acquires multiple consecutive initial responses of the tested board due to external excitation within the current time step.

[0011] Based on multiple consecutive initial responses and a pre-defined primary path transfer characteristic function, the excitation signal applied to the test board by the external excitation in the current time step is reproduced; the primary path is the transmission path from the excitation point on the test board to the error sensor.

[0012] All excitation signals reproduced within the current time step are used as reference signals and input into the FXLMS algorithm to obtain the control signal output to the actuator during the current time step iteration.

[0013] Preferably, the step of reproducing the excitation signal applied to the test board by external excitation within the current time step based on multiple consecutive initial responses and a pre-set primary path transfer characteristic function includes:

[0014] The frequency domain response at each moment is obtained by performing Fourier transform on the initial response collected by the error sensor at each moment within the current time step.

[0015] Based on the frequency domain response at each time step and the pre-defined primary path transfer characteristic function, the frequency domain excitation signal at each time step is obtained by using the principles of time domain convolution and frequency domain multiplication.

[0016] Perform an inverse Fourier transform on the frequency domain excitation signal at each time step to obtain the time domain excitation signal at each time step.

[0017] Preferably, the step of using the excitation signal reproduced in the current time step as a reference signal input into the FXLMS algorithm to obtain the control signal output to the actuator at the current time step iteration includes:

[0018] Based on the reference signal obtained in the previous time step iteration and the error response after the previous time step iteration, the filter weight coefficients for the current time step iteration are obtained.

[0019] Generate a reference signal sequence vector based on all excitation signals reproduced within the current time step;

[0020] Based on the reference signal sequence vector and the filter weight coefficients obtained by the FXLMS filter at the current time step iteration, the control signal output to the actuator at the current time step iteration is obtained.

[0021] Preferably, the step of obtaining the filter weight coefficients for the current time step iteration based on the reference signal obtained in the previous time step iteration and the error response after the previous time step iteration includes:

[0022] The filtered reference signal is obtained after the reference signal obtained in the previous time step iteration is estimated by the secondary path; the secondary path is the transmission path from the point of action of the actuator on the test board to the error sensor.

[0023] Obtain the error signal based on the initial response at the previous time step iteration and the control response obtained after the previous time step iteration;

[0024] Input the filter weight coefficients, filter reference signal, and error signal from the previous time step iteration into the FXLMS filter to obtain the filter weight coefficients for the current time step iteration.

[0025] In the first time step iteration, the filter weight coefficient is 0.

[0026] Preferably, the pre-defined primary pathway transport characteristic function is obtained through a primary pathway identification experiment.

[0027] Preferably, the method further includes:

[0028] The control signal is convolved with a pre-defined secondary path transfer characteristic function to obtain the control response at the error sensor during the current time step iteration.

[0029] Based on the initial response at the current time step iteration and the control response at the error sensor, the error signal after the current time step iteration is obtained.

[0030] Preferably, the pre-defined secondary path transmission characteristic function is obtained through a secondary path identification experiment.

[0031] This invention provides a vibration active control device based on feedback FXLMS, comprising:

[0032] The acquisition module is used to acquire multiple consecutive initial responses of the tested board due to external excitation, collected by the error sensor at the current time step.

[0033] The reference signal reproduction module is used to reproduce the excitation signal applied to the board under test by external excitation in the current time step based on multiple consecutive initial responses and a pre-set primary path transfer characteristic function; the primary path is the transmission path from the excitation point on the board under test to the error sensor.

[0034] The control signal output module is used to input all excitation signals reproduced in the current time step as reference signals into the FXLMS algorithm to obtain the control signal output to the actuator during the current time step iteration.

[0035] Preferably, the reference signal reproduction module includes:

[0036] The frequency domain response output unit is used to perform Fourier transform on the initial response collected by the error sensor at each moment within the current time step to obtain the frequency domain response at each moment.

[0037] The frequency domain excitation signal output unit is used to obtain the frequency domain excitation signal at each time based on the frequency domain response at each time and the pre-set primary path transfer characteristic function, using the principles of time domain convolution and frequency domain multiplication.

[0038] The time-domain excitation signal output unit is used to perform inverse Fourier transform on the frequency-domain excitation signal at each time step to obtain the time-domain excitation signal at each time step.

[0039] Preferably, the control signal output module includes:

[0040] The FXLMS filter weight coefficient determination unit is used to obtain the filter weight coefficients for the current time step iteration based on the reference signal obtained in the previous time step iteration and the error response after the previous time step iteration.

[0041] The reference signal sequence vector determination unit is used to generate a reference signal sequence vector based on all excitation signals reproduced in the current time step;

[0042] The control signal determination unit is used to obtain the control signal output to the actuator at the current time step iteration based on the reference signal sequence vector and the filter weight coefficients obtained by the FXLMS filter at the current time step iteration.

[0043] Preferably, the FXLMS filter weight coefficient determination unit includes:

[0044] The filtered reference signal acquisition subunit is used to acquire the filtered reference signal obtained from the reference signal obtained in the previous time step iteration after estimation by the secondary path; the secondary path is the transmission path from the actuator to the error sensor;

[0045] The error signal acquisition subunit is used to acquire the error signal based on the initial response at the previous time step iteration and the control response obtained after the previous time step iteration.

[0046] The filter weight coefficient determination subunit is used to obtain the filter weight coefficients for the current time step iteration based on the filter weight coefficients, filter reference signal, and error signal from the previous time step iteration.

[0047] In the first time step iteration, the filter weight coefficient is 0.

[0048] Preferably, the device further includes:

[0049] The control response determination module is used to convolve the control signal with a pre-set secondary path transfer characteristic function to obtain the control response at the error sensor during the current time step iteration.

[0050] The error signal determination module is used to obtain the error signal after the current time step iteration based on the initial response at the current time step iteration and the control response at the error sensor.

[0051] This invention also provides a vibration active control testing method based on feedback FXLMS, comprising:

[0052] Select the test board and establish the connection between the test board and the error sensor, the piezoelectric ceramic as the actuator, and the vibrator as the external excitation source. The error sensor, piezoelectric ceramic, and vibrator are electrically connected to the controller.

[0053] The controller acquires multiple consecutive initial responses of the test board generated by the exciter within the current time step collected by the error sensor at the current time step.

[0054] Based on multiple consecutive initial responses and a pre-defined primary path transfer characteristic function, the excitation signal applied to the test board by the external excitation in the current time step is reproduced; the primary path is the transfer path from the excitation point of the vibrator on the test board to the error sensor.

[0055] All excitation signals reproduced within the current time step are used as reference signals and input into the FXLMS algorithm to obtain the control signal output to the actuator during the current time step iteration.

[0056] Preferably, the primary path transfer characteristic function is obtained by performing a primary path identification experiment on the primary path formed by the excitation point of the vibrator on the test plate and the error sensor.

[0057] Preferably, the method further includes:

[0058] The control signal is convolved with a pre-defined secondary path transfer characteristic function to obtain the control response at the error sensor during the current time step iteration.

[0059] Based on the initial response at the current time step iteration and the control response at the error sensor, the error signal after the current time step iteration is obtained.

[0060] Preferably, the secondary path transfer characteristic function is obtained by performing a secondary path identification experiment on the secondary path formed by the point of action of the piezoelectric ceramic on the test plate and the error sensor.

[0061] The present invention also provides a vehicle including the above-described active vibration control device based on feedback FXLMS.

[0062] The present invention also provides a control device, including a processor, a memory, and a program or instructions stored in the memory and executable on the processor, wherein when the program or instructions are executed by the processor, the steps of the vibration active control method based on the feedback FXLMS algorithm described above are implemented.

[0063] The present invention also provides a readable storage medium on which a program or instruction is stored, which, when executed by a processor, implements the steps of the vibration active control method based on the feedback FXLMS algorithm described above.

[0064] The beneficial effects of this invention are as follows:

[0065] The excitation signal of the excitation source within a time step is reproduced by back-reasoning multiple consecutive initial responses and primary path transfer characteristic functions. The reproduced excitation signal is used as the reference signal required in the feedback FXLMS algorithm, instead of directly acquiring the excitation signal in the feedforward FXLMS control of the prior art. The method of acquiring the response signal on the board under test is simpler and more accurate, and can overcome the problems of external interference when directly acquiring the excitation signal and the problem of the excitation source being undeterminable due to the complexity of the excitation source. Specifically: (1) Using the feedback FXLMS algorithm can overcome external interference and is suitable for situations where the external excitation is complex and the excitation source is difficult to determine. Therefore, it has more applicable scenarios than the feedforward control system; (2) Using Fourier transform pairs to reproduce the reference signal of the feedback FXLMS algorithm avoids the problem of difficulty in obtaining the reference signal in feedforward control and instability in the actual application of AVC. Moreover, the reproduced reference signal reproduces the real external excitation signal, avoiding the problem of poor correlation between the reference signal and the real external excitation signal caused by artificially synthesized reference signal. This allows the FXLMS algorithm to better adapt to the real external excitation, taking into account the advantages of accuracy and practicality. Attached Figure Description

[0066] Figure 1 A schematic diagram of a system for active control of plate vibration;

[0067] Figure 2 This is a schematic diagram of the feedback-based FXLMS control algorithm.

[0068] Figure 3 This is a path identification model based on LMS;

[0069] Figure 4 The response and error signals are based on feedback FXLMS;

[0070] Figure 5 The response and error signals are based on feedforward FXLMS.

[0071] Figure 6 This is a schematic diagram of the active vibration control method in this embodiment;

[0072] Figure 7 This is a schematic diagram of the active vibration control method in this embodiment;

[0073] Figure 8 This is a schematic diagram of the vibration active control device in this embodiment. Detailed Implementation

[0074] The embodiments of the present invention will be described below with reference to the accompanying drawings and preferred embodiments. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be understood that the preferred embodiments are only for illustrating the present invention and not for limiting the scope of protection of the present invention.

[0075] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Therefore, the drawings only show the components related to the present invention and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0076] Existing active noise control for sheet metal is based on feedforward control, which directly acquires the excitation signal. However, due to external interference and the potential presence of multiple external excitation sources, the traditional method of directly acquiring the excitation signal is complex and difficult to implement. Therefore, this application provides an active vibration control method based on the feedback FXLMS algorithm, such as... Figure 6 As shown, the method includes:

[0077] Step S101: Obtain multiple consecutive initial responses of the tested board due to external excitation collected by the error sensor in the current time step.

[0078] Step S102: Based on multiple consecutive initial responses and a pre-set primary path transfer characteristic function, reproduce the excitation signal applied to the test board by the external excitation in the current time step; the primary path is the transmission path from the excitation point on the test board to the error sensor.

[0079] Step S103: All excitation signals reproduced within the current time step are used as reference signals and input into the FXLMS algorithm to obtain the control signal output to the actuator during the current time step iteration.

[0080] In this embodiment, error sensors are arranged at thin-walled interior panels (such as the headliner, floor, or tailgate) that are prone to vibration. These error sensors are used to collect responses from these panels. Specifically, the error sensors periodically collect responses from the thin-walled interior panels. Within one time step of an FXLMS filter, the error sensors continuously collect response data multiple times, resulting in multiple d(t) values. In this embodiment, the number of responses collected by the error sensors within the current time step is determined based on the time step length of the filter in the feedback FXLMS algorithm. For example, if the time step length of the filter is 128, then the error sensors will collect responses from the thin-walled interior panels 128 times within the current time step.

[0081] The data collected by the error sensor is converted from digital data to digital data and then transmitted to the controller.

[0082] For the controller, it needs to combine the 128 response data points collected within this time step to reproduce the excitation signal. In this embodiment, step S102 includes:

[0083] Step S1021: Perform Fourier transform on the initial response collected by the error sensor at each moment within the current time step to obtain the frequency domain response at each moment.

[0084] Step S1022: Based on the frequency domain response at each time point and the pre-set primary path transfer characteristic function, the frequency domain excitation signal at each time point is obtained using the principles of time domain convolution and frequency domain multiplication.

[0085] Step S1023: Perform inverse Fourier transform on the frequency domain excitation signal at each time point to obtain the time domain excitation signal at each time point.

[0086] Specifically, within the current time step, assuming the initial response acquired by the error sensor at time t (where t represents any of the 128 signal acquisition times within the current time step) is d(t), the frequency domain signal D(jω) at time t is obtained by performing a Fourier transform on the initial response d(t) at time t. Furthermore, using the principle of time-domain convolution and frequency-domain multiplication, the frequency-domain expression of the excitation signal can be obtained: X(jω) = D(jω) / H(jω), where H(jω) is the primary path transfer characteristic function, which is obtained through prior primary path identification experiments. In practical vehicle applications, the primary path transfer characteristic function is a pre-defined function. Further, performing an inverse Fourier transform on X(jω) can reproduce the time-domain excitation signal x(t) at time t. By repeating the above steps, the time-domain excitation signals for all moments within that time step can be obtained.

[0087] Then, the sequence vector consisting of the temporal excitation signals x(1) to x(128) at 128 acquisition times within this time step is used as the reference signal x(n) for the iteration input of this time step. Here, n refers to time step n. For example, when n refers to the current time step, n-1 refers to the previous current time step and n+1 refers to the next current time step.

[0088] Specifically, step S103 includes:

[0089] Step S1031: Based on the reference signal x(n-1) obtained in the previous time step iteration and the error response y(n-1) after the previous time step iteration, obtain the filter weight coefficient w(n) for the current time step iteration.

[0090] Step S1032: Based on all excitation signals reproduced in the current time step (i.e., x(1) to x(128)), generate a reference signal sequence vector x(n);

[0091] Step S1033: The reference signal sequence vector x(n) and the filter weight coefficient w(n) obtained by the FXLMS filter at the current time step iteration are used to obtain the control signal y(n) output to the actuator at the current time step iteration.

[0092] In step S1031, to obtain the filter weight coefficients w(n) at the current time step iteration, it is necessary to:

[0093] The filtered reference signal x'(n-1) is obtained by estimating the reference signal x(n-1) obtained in the previous time step iteration through the secondary path; the secondary path is the transmission path from the point of action of the actuator on the measured board to the error sensor; here, Where g'2 is the result of estimating the secondary pathway transmission characteristics g2, and L is the weighting order.

[0094] Obtain the error signal e(n-1) based on the initial response d(n-1) at the previous time step iteration and the control response y'(n-1) obtained after the previous time step iteration; the error signal e(n-1) is the difference between the initial response d(n-1) and the control response y'(n-1).

[0095] Then, the filter weight coefficients w(n-1), the filter reference signal x'(n-1), and the error signal e(n-1) from the previous time step iteration are input into the FXLMS filter. The filter weight coefficients w(n) for the current time step iteration are obtained through the FXLMS adaptive filtering method; that is, w(n) = w(n-1) + 2μe(n-1)x'(n-1), where μ represents the step size factor of the FXLMS filter. By pre-setting, the value of μ is between 0 and 1, such as 0.04. Among them, the filter weight coefficients w(1) in the first time step iteration are 0.

[0096] like Figure 7 In this embodiment of the invention, the method further includes:

[0097] Step S104: Convolve the control signal y(n) with the pre-set secondary path transmission characteristic g2 to obtain the control response y'(n) at the error sensor during the current time step iteration; where y'(n) = g2*y(n), and the secondary path transmission characteristic g2 is obtained through secondary path identification experiments;

[0098] Step S105: Based on the initial response d(n) at the current time step iteration and the control response y'(n) at the error sensor, obtain the error signal e(n) after the current time step iteration, i.e., e(n) = d(n) - y'(n).

[0099] The above method in this embodiment reproduces the excitation signal x(t) of the excitation source by back-reasoning the initial response and the primary path transfer characteristic function. The reproduced excitation signal x(t) is used as the reference signal required in the feedback FXLMS algorithm, instead of directly acquiring the excitation signal in the feedforward FXLMS control of the prior art. The method of acquiring the response signal on the board under test is simpler and more accurate, and can overcome the problems of external interference when directly acquiring the excitation signal and the problem of the excitation source being difficult to determine due to the complexity of the excitation source. Specifically: (1) Using the feedback FXLMS algorithm can overcome external interference and is suitable for situations where the external excitation is complex and the excitation source is difficult to determine. Therefore, it has more applicable scenarios than the feedforward control system; (2) Using Fourier transform pairs to reproduce the reference signal of the feedback FXLMS algorithm avoids the problem of difficulty in obtaining the reference signal in feedforward control and instability in the actual application of AVC. Moreover, the constructed signal reproduces the external excitation signal, taking into account the advantages of accuracy and practicality.

[0100] This embodiment also provides a method for verifying the effectiveness of the above-mentioned active vibration control method, such as... Figure 1In this embodiment, the main components required for active vibration control testing of the plate include a test plate, a vibrator, an error sensor, a controller with an integrated control algorithm, a piezoelectric ceramic, a piezoelectric ceramic drive power supply, and a power amplifier. In the active control system, the input signal is the excitation signal generated by the vibrator, simulating external excitation. The integrated controller processes the initial response of the test plate to the external excitation and outputs a voltage signal that can control the actuator (piezoelectric ceramic). The piezoelectric ceramic converts the electrical signal into a force signal, thereby generating a control force on the surface of the plate, gradually weakening and suppressing the vibration of the plate.

[0101] Therefore, closed-loop feedback control is more suitable for AVC systems. This invention proposes combining the FXLMS algorithm with feedback control, using the primary path to transmit the characteristic function and initial response signal to reproduce the excitation signal applied to the board by external stimuli, and constructing the reference signal input required by the FXLMS algorithm, thus achieving both algorithm stability and practicality. Furthermore, the effectiveness of the improved method is verified through simulation testing of the board's active control system.

[0102] Combination Figure 1 and Figure 2 Taking a thin steel plate fixed at both ends as an example, this paper simulates a thin-walled panel in an automobile and establishes an active vibration control system based on feedback FXLMS. First, the response functions of the primary and secondary paths required for the simulation need to be tested using an experimental bench (i.e., the transmission characteristics of the primary and secondary paths are tested). The response of the primary path G1 is the response from the exciter to the error sensor, and the response of the secondary path G2 is the response at the error sensor, driven by the piezoelectric ceramic driving power supply. The measurement uses methods such as... Figure 3 The LMS offline identification method shown can identify the filter coefficients of the corresponding paths through experimentally measured responses. After obtaining the transmission characteristics of each path, the active control of plate vibration can be simulated based on the feedback FXLMS method. In this simulation, four actuators (piezoelectric ceramics), one error sensor, and one excitation source are selected. Therefore, there are four secondary paths, and the transmission characteristics from the piezoelectric element to the error sensor are formed by the superposition of the transmission characteristics of the four secondary paths.

[0103] The specific implementation process of the simulation for active control of plate vibration based on feedback FXLMS is as follows:

[0104] 1. The placement area of ​​piezoelectric ceramics on the plate surface is determined by modal simulation analysis. For plates that are fixed on both sides, the location with the largest modal strain energy is concentrated at the four corners of the plate. Therefore, piezoelectric ceramics are placed at the four corners of the plate.

[0105] 2. The frequency response function H(jω) of the primary path was tested using the constructed experimental setup.

[0106] 3. Two sinusoidal signals are superimposed as the excitation signal x(n). The amplitude of both sinusoidal signals is 1, and their frequencies are ω1 = 80 rad / s and ω2 = 110 rad / s, respectively.

[0107] x1(n) = sin80t

[0108] x2(n)=sin110t

[0109] x(n) = x1(n) + x2(n)

[0110] 4. Set the simulation time and step size. The simulation time is 10s, and the time step size is consistent with the sampling time of the input excitation signal, which is 0.0001s.

[0111] 5. Identify the primary and secondary paths, select a filter length of 128, and use methods such as... Figure 3 The LMS identification model shown uses the measured response as the simulation expectation input. The weight coefficients obtained from the simulation are the filter coefficients representing the corresponding path transmission characteristics. The primary path filter coefficients (i.e., primary path transmission characteristics) and the four secondary path filter coefficients are g1, g2, g3, g4, g5, g6, g7, g8, g9, g1, g1, g1, g2, g3, g4, g5, g6, g7, g8, g9, g1, g1, g1, g2, g3, g1, g2, g3, g1, g2, g3, g4, g5, g1, g1, g2, g3, g1, g2, g3, g1, g2, g3, g4 ...2, g3, g1, g2, g3, g4, g1, g2, g 21 g 22 g 23 g 24 Therefore, the secondary path transfer characteristic g2 from the piezoelectric element to the error sensor can be expressed as:

[0112] g2=g 21 +g 22 +g 23 +g 24

[0113] 6. Convolve the identified primary path filter coefficients g1 with the excitation signal x(t) at time t to obtain the primary response d(t), where:

[0114] d(t) = g1*x(t)

[0115] 7. Perform a Fourier transform on the time-domain response signal d(t) to obtain the frequency-domain signal D(jω), where:

[0116]

[0117] 8. Using the principle of time-domain convolution and frequency-domain multiplication, the frequency-domain representation of the excitation signal can be obtained, that is:

[0118] X(jω)=D(jω) / H(jω)

[0119] 9. The inverse Fourier transform of X(jω) can reproduce the time-domain excitation signal x(t) at time t, that is:

[0120]

[0121] 10. The reproduced time-domain excitation signal x(t) is used as a reference signal input to the filter update module. A reference signal sequence (column vector) x(n) is generated using the time-domain excitation signals x(t) at multiple times within the current time step. The reference signal is then filtered using the secondary path filter coefficients g2 to obtain the filtered reference signal x'(n), i.e.:

[0122]

[0123] Simultaneously, the excitation signal x(n) reconstructed in the previous step is input into the FXLMS algorithm as a reference signal, and substituted into the control signal calculation formula, i.e.:

[0124]

[0125] Where L represents the weighting order of the FXLMS algorithm, which is 128 in the simulation. Therefore, x(n) is the reference signal sequence composed of the reference signal values ​​of 128 time steps in the nth iteration.

[0126] 11. Input the filtered reference signal x'(n-1) and error signal e(n-1) obtained in the previous time step iteration into the FXLMS filter. After FXLMS adaptive filtering, the iterative formula for the filter weight coefficients w(n) can be obtained, that is:

[0127] w(n)=w(n-1)+2μe(n-1)x'(n-1)

[0128] Among them, the step size factor μ is adaptively updated as the iteration proceeds, and x'(n-1) is obtained by estimating the reference signal x(n-1) through the secondary path.

[0129] 12. Convolve the control signal y(n) with the secondary channel transfer characteristic function g2 to obtain the control response y'(n) at the error sensor, i.e.:

[0130] y'(n) = g2*y(n)

[0131] The vibration active control method based on feedback FXLMS was tested and verified. This method was compared with the feedforward-based FXLMS vibration active control method. The comparison of the initial response signal and error signal under the two methods was also discussed. Figure 4 and Figure 5As shown, the error signal of the vibration active control method based on feedback FXLMS can converge well to a small value, and the iterative process is stable, proving the effectiveness and accuracy of the proposed method. Combined with... Figure 4 and Figure 5 It can be seen that although the error signal of the feedback-based FXLMS active vibration control method at the final convergence state is greater than that of the feedforward-based FXLMS active vibration control method, this is because the accuracy of the feedback control system itself is lower than that of the feedforward control system. Therefore, in engineering applications, the feedback-based FXLMS active vibration control method can overcome all external disturbances and form a closed-loop control system without acquiring external excitation signals. Compared with the feedforward control system, it has better flexibility and applicability.

[0132] like Figure 8 This embodiment also provides a vibration active control device based on feedback FXLMS, including:

[0133] The acquisition module 201 is used to acquire the initial response d(n) of the tested board caused by external excitation, which is continuously collected by the error sensor in the current time step.

[0134] The reference signal reproduction module 202 is used to reproduce the excitation signal applied to the test board by the external excitation in the current time step according to the initial response d(n) and the preset primary path transmission characteristic function H(jω); the primary path G1 is the transmission path from the excitation point on the test board to the error sensor.

[0135] The control signal output module 203 is used to input all the excitation signals reproduced in the current time step as reference signals into the FXLMS algorithm to obtain the control signal output to the actuator during the current time step iteration.

[0136] Reference signal reproduction module 202 corresponds to Figure 2 The reference signal reproduction module in the middle.

[0137] like Figure 8 The reference signal reproduction module 202 includes:

[0138] The frequency domain response output unit 2021 is used to perform Fourier transform on the initial response collected by the error sensor at each moment in the current time step to obtain the frequency domain response at each moment.

[0139] The frequency domain excitation signal output unit 2022 is used to obtain the frequency domain excitation signal at each time based on the frequency domain response at each time and the pre-set primary path transfer characteristic function, using the principles of time domain convolution and frequency domain multiplication.

[0140] The time-domain excitation signal output unit 2023 is used to perform inverse Fourier transform on the frequency-domain excitation signal at each time to obtain the time-domain excitation signal at each time.

[0141] like Figure 8 The control signal output module 203 includes:

[0142] The FXLMS filter weight coefficient determination unit 2031 is used to obtain the filter weight coefficient w(n) at the current time step iteration based on the reference signal x(n-1) obtained in the previous time step iteration and the error response y(n-1) after the previous time step iteration.

[0143] The reference signal sequence vector determination unit 2032 is used to generate a reference signal sequence vector x(n) based on all excitation signals reproduced in the current time step;

[0144] The control signal determination unit 2033 is used to input the reference signal sequence vector x(n) and the filter weight coefficient w(n) at the current time step iteration into the FXLMS filter to obtain the control signal y(n) output to the actuator at the current time step iteration.

[0145] Among them, the FXLMS filter weight coefficient determination unit 2031 corresponds to Figure 2 The FXLMS filter in the image.

[0146] like Figure 8 The FXLMS filter weight coefficient determination unit 2031 includes:

[0147] The filter reference signal acquisition subunit 20311 is used to acquire the filter reference signal x'(n-1) obtained from the reference signal x(n-1) obtained in the previous time step iteration after estimation by the secondary path; the secondary path is the transmission path from the actuator to the error sensor;

[0148] Error signal acquisition subunit 20312 is used to acquire error signal e(n-1) based on the initial response d(n-1) at the previous time step iteration and the control response y(n-1) obtained after the previous time step iteration;

[0149] The filter weight coefficient determination subunit 20313 is used to obtain the filter weight coefficient w(n) at the current time step iteration based on the filter weight coefficient w(n-1), the filter reference signal x'(n-1), and the error signal e(n-1) at the previous time step iteration.

[0150] In the first time step iteration, the filter weight coefficient w(1) is 0.

[0151] like Figure 8 The device further includes:

[0152] The control response determination module 204 is used to convolve the control signal with a pre-set secondary path transfer characteristic function to obtain the control response at the error sensor during the current time step iteration.

[0153] The error signal determination module 205 is used to obtain the error signal after the current time step iteration based on the initial response at the current time step iteration and the control response at the error sensor.

[0154] The present invention also provides a vehicle including the above-described active vibration control device based on feedback FXLMS.

[0155] The present invention also provides a control device, including a processor, a memory, and a program or instructions stored in the memory and executable on the processor, wherein when the program or instructions are executed by the processor, the steps of the vibration active control method based on the feedback FXLMS algorithm described above are implemented.

[0156] The present invention also provides a readable storage medium on which a program or instruction is stored, which, when executed by a processor, implements the steps of the vibration active control method based on the feedback FXLMS algorithm described above.

[0157] The above embodiments are preferred embodiments of the present invention, but the embodiments of the present invention are not limited to the above embodiments. Any changes, modifications, substitutions, combinations, or simplifications made without departing from the spirit and principle of the present invention shall be considered equivalent substitutions and shall be included within the protection scope of the present invention.

Claims

1. A vibration active control method based on the feedback FXLMS algorithm, characterized in that, include: Acquire multiple consecutive initial responses of the tested board due to external excitation, collected by the error sensor within the current time step; Based on multiple consecutive initial responses and a pre-defined primary path transfer characteristic function, the excitation signal applied to the test board by the external excitation in the current time step is reproduced; the primary path is the transmission path from the excitation point on the test board to the error sensor. The steps of reproducing all excitation signals within the current time step as reference signals and inputting them into the FXLMS algorithm to obtain the control signal output to the actuator during the current time step iteration include: reproducing the excitation signals applied to the tested board by external excitation within the current time step based on multiple consecutive initial responses and a pre-defined primary path transfer characteristic function. The frequency domain response at each moment is obtained by performing Fourier transform on the initial response collected by the error sensor at each moment within the current time step. Based on the frequency domain response at each time step and the pre-defined primary path transfer characteristic function, the frequency domain excitation signal at each time step is obtained by using the principles of time domain convolution and frequency domain multiplication. Perform an inverse Fourier transform on the frequency domain excitation signal at each time step to obtain the time domain excitation signal at each time step.

2. The vibration active control method based on the feedback FXLMS algorithm according to claim 1, characterized in that, The steps involved in using the excitation signal reproduced in the current time step as a reference signal input into the FXLMS algorithm to obtain the control signal output to the actuator during the current time step iteration include: Based on the reference signal obtained in the previous time step iteration and the error response after the previous time step iteration, the filter weight coefficients for the current time step iteration are obtained. Generate a reference signal sequence vector based on all excitation signals reproduced within the current time step; Based on the reference signal sequence vector and the filter weight coefficients obtained by the FXLMS filter at the current time step iteration, the control signal output to the actuator at the current time step iteration is obtained.

3. The vibration active control method based on the feedback FXLMS algorithm according to claim 2, characterized in that, The steps for obtaining the filter weight coefficients for the current time step iteration based on the reference signal obtained in the previous time step iteration and the error response after the previous time step iteration include: The filtered reference signal is obtained after the reference signal obtained in the previous time step iteration is estimated by the secondary path; the secondary path is the transmission path from the point of action of the actuator on the test board to the error sensor. Obtain the error signal based on the initial response at the previous time step iteration and the control response obtained after the previous time step iteration; Input the filter weight coefficients, filter reference signal, and error signal from the previous time step iteration into the FXLMS filter to obtain the filter weight coefficients for the current time step iteration. In the first time step iteration, the filter weight coefficient is 0.

4. The vibration active control method based on the feedback FXLMS algorithm according to claim 1, characterized in that, The pre-defined primary pathway transport characteristic function was obtained through primary pathway identification experiments.

5. The vibration active control method based on the feedback FXLMS algorithm according to claim 1, characterized in that, The method further includes: The control signal is convolved with a pre-defined secondary path transfer characteristic function to obtain the control response at the error sensor during the current time step iteration. Based on the initial response at the current time step iteration and the control response at the error sensor, the error signal after the current time step iteration is obtained.

6. The vibration active control method based on the feedback FXLMS algorithm according to claim 5, characterized in that, The pre-defined secondary path transmission characteristic function was obtained through secondary path identification experiments.

7. A vibration active control device based on feedback FXLMS, characterized in that, include: The acquisition module is used to acquire multiple consecutive initial responses of the tested board generated by external excitation within the current time step by the error sensor. The reference signal reproduction module is used to reproduce the excitation signal applied to the test board by external excitation within the current time step based on multiple consecutive initial responses and a pre-set primary path transfer characteristic function. The primary path G1 is the transmission path from the excitation point on the board under test to the error sensor; The control signal output module is used to input all excitation signals reproduced within the current time step as reference signals into the FXLMS algorithm to obtain the control signal output to the actuator during the current time step iteration; the reference signal reproduction module includes: The frequency domain response output unit is used to perform Fourier transform on the initial response collected by the error sensor at each moment within the current time step to obtain the frequency domain response at each moment. The frequency domain excitation signal output unit is used to obtain the frequency domain excitation signal at each time based on the frequency domain response at each time and the pre-set primary path transfer characteristic function, using the principles of time domain convolution and frequency domain multiplication. The time-domain excitation signal output unit is used to perform inverse Fourier transform on the frequency-domain excitation signal at each time step to obtain the time-domain excitation signal at each time step.

8. The vibration active control device based on feedback FXLMS according to claim 7, characterized in that, The control signal output module includes: The FXLMS filter weight coefficient determination unit is used to obtain the filter weight coefficients for the current time step iteration based on the reference signal obtained in the previous time step iteration and the error response after the previous time step iteration. The reference signal sequence vector determination unit is used to generate a reference signal sequence vector based on all excitation signals reproduced in the current time step; The control signal determination unit is used to obtain the control signal output to the actuator at the current time step iteration based on the reference signal sequence vector and the filter weight coefficients obtained by the FXLMS filter at the current time step iteration.

9. The vibration active control device based on feedback FXLMS according to claim 8, characterized in that, The FXLMS filter weight coefficient determination unit includes: The filtered reference signal acquisition subunit is used to acquire the filtered reference signal obtained from the reference signal obtained in the previous time step iteration after estimation by the secondary path; the secondary path is the transmission path from the actuator to the error sensor; The error signal acquisition subunit is used to acquire the error signal based on the initial response at the previous time step iteration and the control response obtained after the previous time step iteration. The filter weight coefficient determination subunit is used to obtain the filter weight coefficients for the current time step iteration based on the filter weight coefficients, filter reference signal, and error signal from the previous time step iteration. In the first time step iteration, the filter weight coefficient is 0.

10. The vibration active control device based on feedback FXLMS according to claim 7, characterized in that, The device further includes: The control response determination module is used to convolve the control signal with a pre-set secondary path transfer characteristic function to obtain the control response at the error sensor during the current time step iteration. The error signal determination module is used to obtain the error signal after the current time step iteration based on the initial response at the current time step iteration and the control response at the error sensor.

11. A vibration active control test method based on feedback FXLMS, characterized in that, include: Select the test board and establish the connection between the test board and the error sensor, the piezoelectric ceramic as the actuator, and the vibrator as the external excitation source. The error sensor, piezoelectric ceramic, and vibrator are electrically connected to the controller. The controller acquires the initial response of the test board caused by the exciter, which is continuously collected by the error sensor within the current time step; Based on the initial response and the pre-set primary path transmission characteristic function, the excitation signal applied to the test board by the external excitation in the current time step is reproduced; the primary path is the transmission path from the excitation point of the vibrator on the test board to the error sensor. All excitation signals reproduced within the current time step are used as reference signals and input into the FXLMS algorithm to obtain the control signal output to the actuator during the current time step iteration. The steps for reproducing the excitation signal applied to the test board by external excitation within the current time step based on the initial response and the pre-set primary path transmission characteristic function include: The frequency domain response at each moment is obtained by performing Fourier transform on the initial response collected by the error sensor at each moment within the current time step. Based on the frequency domain response at each time step and the pre-defined primary path transfer characteristic function, the frequency domain excitation signal at each time step is obtained by using the principles of time domain convolution and frequency domain multiplication. Perform an inverse Fourier transform on the frequency domain excitation signal at each time step to obtain the time domain excitation signal at each time step.

12. A vehicle, characterized in that, The active vibration control device based on feedback FXLMS as described in any one of claims 7-10.

13. A control device, comprising a processor, a memory, and a program or instructions stored in the memory and executable on the processor, wherein the program or instructions, when executed by the processor, implement the steps of the vibration active control method based on the feedback FXLMS algorithm as claimed in any one of claims 1 to 6.

14. A readable storage medium storing a program or instructions that, when executed by a processor, implement the steps of the vibration active control method as claimed in any one of claims 1 to 6.

Images