RU2798744C1 - Active vibration damping method - Google Patents

Abstract

FIELD: vibroacoustics.

SUBSTANCE: invention relates to methods for actively suppressing vibration effects. A method for active vibration damping includes: placing residual vibration sensors in places requiring vibration damping, transmitting a signal from reference sensors and signals from residual vibration sensors to the unit for calculating the parameters of a compensating action device, determining the parameters of the compensating signal shaping filters, generating compensating signals when the help of the conversion by the forming filters of the signal from the reference sensors. At the same time, the parameters of the shaping filters of the compensating signal are determined by a time-limited sample of signals from the reference sensors and residual vibration sensors, which does not require calculation and constant correction of the parameters of the shaping filters in real time, and the least squares method is used as a computational method.

EFFECT: reduction in the tuning time of the algorithm for determining the parameters of the compensating signal while simultaneously reducing the requirements for computing power of the signal processing device used by eliminating high-performance real-time computing.

1 cl, 1 dwg

Description

The invention relates to mechanical engineering and can be used to reduce the level of vibration fields in the elements of devices, machines and mechanisms, and is also applicable to create an active vibration isolation platform designed to place equipment sensitive to vibration effects on it, in particular, relates to methods for active vibration control.

Traditionally, to reduce the level of vibration fields, passive systems (shock absorbers and inertial dampers) and active systems (acting on the object of vibration protection with an acu- tor) are used.

Passive methods and systems for damping vibrational fields, despite the simplicity and relatively low cost of implementation, have the disadvantage of low efficiency of passive methods at low frequencies.

Another way to solve such problems is active damping, that is, the use of controlled vibration emitters that create an impact that reduces the total level of vibrations in a given area of space. In this case, the purpose of the problem of active damping is the formation of the necessary signals applied to controlled emitters.

The difficulty lies in the fact that in order to implement the known methods of active damping, it is necessary to use a real-time digital signal processing device.

According to patent RU2572664 , publ. 01/20/2016, IPC G10K 11/178, a device for active vibration damping is known, containing sensors for the period of rotation of the shafts of mechanisms, equal to the number of compensating sources and control receivers installed on the foundation structure, the electronic path of the damping system, consisting of an analyzer, a discrete integrator, a synthesizer, a shaper exponential functions, a period meter, moreover, the outputs of the sensors of the period of rotation of the shafts of the mechanisms are connected to the inputs of the generator of exponential functions and the inputs of the period meter, the outputs of the control receivers are connected to the first inputs of the analyzer, the output of the period meter is connected to the second input of the analyzer, the output of the generator of exponential functions is connected to the third input analyzer and the first input of the synthesizer, characterized in that it is equipped with a modal analyzer and a modal synthesizer, while the output of the analyzer is connected in series with the input of the modal analyzer, discrete integrator, modal synthesizer and the second input of the synthesizer, the outputs of which are connected to the inputs of compensating sources.

However, the analog device has the disadvantage that it is necessary to continuously adjust the control signals of the compensating sources during the operation of the device, which requires the use of a high-performance digital signal processing device in real time.

According to US5365594, publ. 11/15/1994, IPC G10K11/178; H03H21/00; G10K11/16, a signal processing system for controlling vibrations is known, in which noise-free signals of a primary source and at least one secondary source of periodic vibrations are used as input, in which the signals are processed to obtain an output signal representing noise-free interference signal from the primary source. Active sound or vibration suppression device, contains:

• a measuring device with an output for creating a compensating signal;

• a set of sound or vibration sensors;

• a set of controlled emitters of sound or vibrations;

• ADC for signal processing of the specified sound or vibration sensors.

• control controller for generation of suppression signal

The disadvantages of the analogue method include the impossibility of tracking by the proposed device of rapid changes in the frequency of vibration exposure, for example, when changing the speed of rotation of the mechanism drive or changing the load.

The closest analogue (prototype) in technical essence of the proposed method is a method and device for active suppression of the main noise source to create the desired noise level in at least one location, known from patent WO9424970, publ. 11/10/1994, IPC G10K11/178; H04B1/12; A61F11/06; H03B29/00; H04B15/00, comprising the following steps:

• control of at least one actuator by means of a processor;

• selection of locations for residual vibration sensors;

• providing the processor with the primary noise reference signal in block format;

• direction of the output signal of the error sensor in block format to the processor;

• modeling of the relationship between the operation of actuators and the output signal of the error sensor using adaptive filters;

• formation of a suppression signal from the actuators controlled by the processor, based on the model and the reference signal of the primary noise;

• adaptation of the filter coefficients in response to the comparison of the error signals with the model in the processor.

However, the prototype method has the disadvantage that it is necessary to continuously adapt the coefficients of the shaping filters during the operation of the system using gradient methods, which requires the use of a high-performance digital signal processing device that performs calculations in real time. This increases the time for setting up algorithms for determining the parameters of the compensating signal, increases the cost and complicates the practical implementation of the compensation system.

The technical problem solved by this invention is to reduce the tuning time of the algorithm for determining the parameters of the compensating signal while reducing the requirements for computing power of the signal processing device used by eliminating high-performance real-time computing.

The technical result in the proposed method is achieved by the fact that, like the prototype, it includes placing residual vibration sensors in places requiring vibration damping, transmitting a signal from reference sensors and signals from residual vibration sensors to a block for calculating the parameters of a compensating action device, determining the parameters of the forming compensating signal filters, generation of compensating signals by converting the signal from reference sensors by forming filters.

New in the proposed method is that the parameters of the shaping filters of the compensating signal are determined by a time-limited sample of signals from the reference sensors and residual vibration sensors, which does not require calculation and constant correction of the parameters of the shaping filters in real time, and as a computational method is used least square method.

The method is carried out as follows.

Signals are registered for a certain period of time from a reference sensor 2 located near the source of compensated vibrations 1 and sensors of residual vibrations 3. The length of the time period is determined experimentally depending on the specific implementation of the device.

The impulse characteristics of the shaping filters are calculated in the block for calculating the parameters of the compensating action device 5.

The calculated coefficients are transferred to the digital signal processing device 4, which provides the calculation of the convolution of the signal from the reference sensor 2 with the impulse responses of the shaping filters and the formation of compensating signals supplied to the controlled emitters 6.

The impulse responses of the shaping filters are calculated in the frequency domain. In this case, the level of residual vibrations on each of the frequency components is minimized.

The calculation is made as follows:

The field recorded by the residual vibration sensors 3 consists of the field created by the compensated source and the fields created by controlled emitters 6:

Y _p - the field created by the vibration source 1, registered by the reference sensors 2 at the points of their location:

Y _p = Pv , (1)

where v is the signal recorded by the reference sensors 2, P is the transfer matrix from the vibration source to the residual vibration sensors 3.

Y _s - field created by controlled emitters 6, recorded by residual vibration sensors 3 at the points of their location:

Y _s = Sy, (2)

where S is the transfer matrix from controlled emitters 6 to residual vibration sensors 3 at a fixed frequency, y are the voltages applied to controlled emitters 6.

y = Wv , (3)

where – W is a matrix composed of the AFC of shaping filters.

Then the residual field recorded by the sensors of residual vibrations e :

e = Y _p + Y _s = Pv + Sy (4)

Solving the problem of minimizing the power of residual vibrations by the least squares method, taking into account regularization, we obtain the equation:

By solving equation (5) and introducing regularization, one can find a solution for the matrix of the AFC of the shaping filters W and, using the obtained AFC, determine the impulse responses through the inverse Fourier transform.

Where:

A = 2 S ^H S , B = vv ^H , (7)

C = -2 S ^H Y _p v ^H (8)

операция векторизации матрицы,

произведение Кронекера (блочное произведение матриц), α - коэффициент регуляризации, выбираемый под конкретную систему активного гашения (определяется экспериментально), I - единичная матрица. ^Н – знак эрмитового сопряжения, ^T - знак транспонирования.Where

matrix vectorization operation,

Kronecker product (block product of matrices), α - regularization coefficient chosen for a specific active damping system (determined experimentally), I - identity matrix. ^H is the Hermitian conjugation sign, ^T - transpose sign.

Impulse characteristics of shaping filters:

H \u003d F ^-1 (W) (9)

обратное преобразование Фурье.Where

inverse Fourier transform.

и

Таким образом, импульсные характеристики формирующих фильтров определяются механическими параметрами системы и не зависят от времени. Это позволяет осуществлять вычисление H не в режиме реального времени, а по определенной ограниченной по времени выборке сигналов. Таким образом, не требуется высокопроизводительного устройства цифровой обработки сигналов, производящего вычисления в режиме реального времени. Taking into account equation (1), one can see that the solution does not depend on

And

Thus, the impulse responses of shaping filters are determined by the mechanical parameters of the system and do not depend on time. This makes it possible to calculate H not in real time, but according to a certain time-limited sample of signals. Thus, a high-performance digital signal processing device that performs real-time calculations is not required.

According to the recorded signals and the calculated impulse responses of the shaping filters H , signals are generated that are fed to the controlled emitters.

In addition, the use of the least squares method instead of gradient methods can dramatically increase the rate of convergence of the algorithm used.

This method is quite simply implemented by using a PC and a signal processing device connected to multichannel ADCs and DACs. The purpose of the signal processing device (4) in this case will be only the calculation of the convolution of signals from the reference sensors (2) with the calculated impulse responses, without the need to calculate the characteristics of the shaping filters in real time.

The operation of the damping algorithm is impossible without knowledge of the transfer matrices from the controlled elements to the residual vibration sensors. ( S - transfer matrix from controlled emitters 6 to sensors of residual vibrations 3). To determine the transfer matrix data S, a separate calibration procedure must be used. This procedure consists in applying a calibration signal (for example, short pulses, bandpass noise, broadband modulated signal) to each of the controlled radiators 6 and calculating the required transfer characteristics from the signals recorded from the residual vibration sensors 3.

In FIG. 1 shows the layout of sensors and emitters, the block for calculating the parameters of the compensating action device and the digital signal processing device according to the proposed method.

Designations:

1 - source of compensated noise

2 - reference sensor

3 - sensors of residual vibrations

4 - digital signal processing device

5 - block for calculating the parameters of the compensating action device

6 - controlled emitters

Claims (1)

A method for active vibration damping, which includes: placing residual vibration sensors in places requiring vibration damping, transmitting a signal from reference sensors and signals from residual vibration sensors to a block for calculating the parameters of a compensating action device, determining the parameters of the compensating signal shaping filters, generating compensating signals when by means of converting the signal from the reference sensors by the shaping filters, characterized in that the parameters of the shaping filters of the compensating signal are determined by a time-limited sample of signals from the reference sensors and residual vibration sensors, which does not require calculation and constant correction of the parameters of the shaping filters in real time, and as a computational method, the least squares method is used.

Images