A method for decoupling coupled vibration signals from multiple vibration sources in a hydroelectric generator unit.

Abstract

This invention discloses a decoupling method for multi-source coupled vibration signals in a hydroelectric generator, comprising the following steps: acquiring horizontal and vertical vibration signals of the hydroelectric generator at a preset frequency and preprocessing them to obtain measured vibration signals; weighting the measured vibration signals using a constructed time-domain convolution window and then performing a Fourier transform to obtain a constant frequency; performing a second-order Taylor expansion on the instantaneous phasors of the measured vibration signals in amplitude and phase representation to obtain a multi-source coupled vibration signal model; solving the error equation between the measured vibration signals and the model to obtain the instantaneous vector of the measured vibration signals, which is the modulation signal of the original vibration signals, thus achieving decoupling of the vibration signals. This invention, based on the accurate measurement of the original frequency of the vibration signal by constructing a novel window function, establishes a vibration signal model under multi-source coupling based on Taylor expansion, and then achieves decoupling of the vibration signals under fast time-varying and high-frequency conditions by solving the model parameters.

Description

Technical Field

[0001] This invention belongs to the field of vibration signal processing of hydropower units, and specifically relates to a decoupling method for multi-source coupled vibration signals in hydropower units. Background Technology

[0002] The operational reliability of hydropower units has a crucial impact on the safety and stability of hydropower stations. However, the motion state of hydropower units is affected by the coupling effects of hydraulic, mechanical, and electrical factors. For example, the mechanical motion of the rotating mechanical parts of the unit will cause changes in the magnetic field between the water flow and the air gap of the motor; changes in the water flow field in the unit will also cause changes in the magnetic field between the air gap of the motor, and at the same time, will cause mechanical vibration. Therefore, hydropower units are multi-source nonlinear systems with water-mechanical-electrical coupling, and the vibration signals will also exhibit strong multi-source coupling characteristics.

[0003] Existing vibration signal processing methods are mainly divided into time domain, frequency domain, and time-frequency domain methods. Time domain methods typically calculate the mathematical characteristics or probability distribution of the signal itself as the basis for detection, such as cross-correlation functions and probability density functions, but they are easily affected by noise. Frequency domain methods based on Fourier transform can obtain the characteristic spectrum of vibration signals, but they lack time-domain descriptive information for non-stationary signals. Time-frequency domain methods can characterize signal features from both time and frequency dimensions, and have advantages in processing time-varying signals.

[0004] Although there is a lot of research on the fault feature extraction of non-stationary vibration signals of hydropower units, there is not much research on the identification and separation of multi-source coupled vibration signals of hydropower units. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide a decoupling method for multi-source coupled vibration signals in hydropower units to address the shortcomings of the prior art, thereby identifying and separating the multi-source coupled vibration signals of hydropower units.

[0006] To solve the above-mentioned technical problems, the present invention includes:

[0007] A method for decoupling coupled vibration signals from multiple sources in a hydroelectric generator unit includes the following steps:

[0008] S1. Collect vibration signals of the hydropower unit in the horizontal and vertical directions according to the preset frequency;

[0009] S2. Perform preliminary noise reduction processing on the collected vibration signals to obtain the measured vibration signals;

[0010] S3. Construct a temporal convolutional window;

[0011] S4. The measured vibration signal is weighted using the constructed time-domain convolution window, and the weighted measured vibration signal is subjected to Fourier transform to obtain a constant frequency.

[0012] S5. Perform a second-order Taylor expansion on the instantaneous phasor of the measured vibration signal in terms of amplitude and phase, and use the resulting approximate signal as a multi-source coupled vibration signal model.

[0013] S6. Solve the error equation of the measured vibration signal and the multi-source coupled vibration signal model to obtain the instantaneous vector of the measured vibration signal, which is the modulation signal of the original vibration signal, thereby realizing the decoupling of the vibration signal.

[0014] Furthermore, in step S2, the collected vibration signal is initially denoised using a filter and a low-rank sparse decomposition method.

[0015] Furthermore, in step S2, the measured vibration signal is represented by x(n), where n = 0, 1, ..., N, and N is an odd number.

[0016] Furthermore, in step S3, the third-order maximum sidelobe attenuation window of length M1 and the third-order minimum sidelobe Nuttall window of length M2 are convolved in the temporal domain and zero-padding is performed to obtain a temporal convolution window w(m) of length M, where M = M1 + M2, m = 0, 1, ..., M. <N。

[0017] Furthermore, in step S4, the weighted measured vibration signal is x. w (m)=w(m)x(m), for the weighted measured vibration signal x w After performing a Fourier transform on (m), a constant frequency f0 is obtained.

[0018] Furthermore, in step S4, the constant frequency f0 is obtained by using two-line interpolation discrete Fourier transform or three-line interpolation discrete Fourier transform.

[0019] Furthermore, in step S5, the amplitude and phase of the measured vibration signal x(n) are expressed in the following forms:

[0020]

[0021] Where A(n) and Let f represent the instantaneous amplitude and instantaneous phase of the measured vibration signal x(n), respectively. s The preset frequency is i, where i is the imaginary unit.

[0022] Instantaneous phasor of measured vibration signal in amplitude and phase representation A second-order Taylor expansion yields the multi-source coupled vibration signal model:

[0023]

[0024] in, p′(n0) and p″(n0) represent the first and second derivatives of the instantaneous phasor p(n) at n0, respectively.

[0025] Furthermore, in step S6, the measured vibration signal x(n) is coupled with the multi-source vibration signal model. The multiple sets of error equations are:

[0026]

[0027] Furthermore, using the least squares method, the error equations e of multiple sets are obtained. n The instantaneous vector p(n) is obtained by finding p(n0), p′(n0), and p″(n0) when they are at their minimum values.

[0028] The beneficial effects of this invention are:

[0029] 1) A novel window function is constructed to measure the original frequency of the vibration signal, effectively reducing spectral leakage in the discrete frequency correction method; 2) A vibration signal model under multi-source coupling is established, and the Taylor expansion (i.e., higher-order polynomial) is used to accurately estimate the instantaneous frequency and instantaneous amplitude of the signal; 3) By accurately solving the Taylor expansion coefficients, a highly readable dynamic signal feature characterization method is established to complete the mapping of the dynamic features of the vibration signal under multi-source coupling. Detailed Implementation

[0030] To facilitate understanding of the present invention, specific embodiments are described in further detail below. Those skilled in the art should understand that the embodiments described are merely illustrative and should not be construed as limiting the scope of the invention.

[0031] This invention provides a method for decoupling coupled vibration signals from multiple vibration sources in a hydroelectric generator, comprising the following steps:

[0032] S1, according to the preset frequency f s Vibration signals of the hydropower unit in the horizontal and vertical directions are collected.

[0033] S2. Perform preliminary noise reduction on the collected vibration signal to obtain the measured vibration signal x(n), where n = 0, 1, ..., N, and N is an odd number.

[0034] In this step, for the vibration signals with high noise levels that are collected, preliminary noise reduction processing is performed using methods such as filters and low-rank sparse decomposition.

[0035] S3. Constructing a temporal convolutional window: A third-order maximum sidelobe attenuation window of length M1 and a third-order minimum sidelobe Nuttall window of length M2 are convolved in the temporal domain and zero-padding is applied to obtain a temporal convolutional window w(m) of length M, where M = M1 + M2, m = 0, 1, ..., M. <N。

[0036] S4. The measured vibration signal x(n) is weighted using the constructed time-domain convolutional window w(m), and the weighted measured vibration signal x is then processed. w Performing a Fourier transform on (m) = w(m)x(m) yields a constant frequency f0.

[0037] In this step, the constant frequency f0 can be obtained using discrete spectrum correction methods such as two-line interpolation discrete Fourier transform and three-line interpolation discrete Fourier transform.

[0038] S5. Perform a second-order Taylor expansion on the instantaneous phasor of the measured vibration signal in amplitude and phase representation, and use the resulting approximate signal as a multi-source coupled vibration signal model.

[0039] In this step, the amplitude and phase of the measured vibration signal x(n) are expressed as follows:

[0040]

[0041] Where A(n) and Let f represent the instantaneous amplitude and instantaneous phase of the measured vibration signal x(n), respectively. s The preset frequency is denoted by i, where i is the imaginary unit.

[0042] Instantaneous phasor of measured vibration signal in amplitude and phase representation A second-order Taylor expansion yields a multi-source coupled vibration signal model.

[0043]

[0044] in, p′(n0) and p″(n0) represent the first and second derivatives of the instantaneous phasor p(n) at n0, respectively.

[0045] S6. Obtain the vibration signal model that couples the measured vibration signal x(n) with the multi-source vibration signal using the least squares method. Multiple sets of error equations By finding the smallest values ​​of p(n0), p′(n0), and p″(n0), we obtain the instantaneous vector p(n) of the measured vibration signal x(n), which is the modulation signal of the original vibration signal, thus achieving decoupling of the vibration signal.

[0046] This invention achieves accurate measurement of the original frequency of vibration signals by constructing a novel time-domain convolution window function. It then establishes a vibration signal model under multi-source coupling based on Taylor expansion and solves the model parameters to decouple the vibration signal under fast time-varying and high-frequency conditions.

[0047] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for decoupling multi-source coupled vibration signals in a hydroelectric generator unit, characterized in that, Includes the following steps: S1. Collect vibration signals of the hydropower unit in the horizontal and vertical directions according to the preset frequency; S2. Perform preliminary noise reduction processing on the collected vibration signals to obtain the measured vibration signals; S3. Construct a temporal convolutional window; in step S3, the window with a length of... M The third-order maximum sidelobe attenuation window and its length are 1 M After performing temporal convolution on a 2-order third-order minimum sidelobe Nuttall window and padding with zeros, we obtain a length of... M Temporal convolution window ,in, M = M 1+ M 2, m =0,1,…, M , M < N , N It is an odd number; S4. The measured vibration signal is weighted using the constructed time-domain convolution window, and a Fourier transform is performed on the weighted measured vibration signal to obtain a constant frequency. ; S5. Perform a second-order Taylor expansion on the instantaneous phasor of the measured vibration signal in amplitude and phase representation form, and use the resulting approximate signal as a multi-source coupled vibration signal model; in step S5, the measured vibration signal The amplitude and phase are expressed in the following forms: in, and These represent the measured vibration signals respectively. The instantaneous amplitude and instantaneous phase, f s For preset frequency, i The imaginary unit; Instantaneous phasor of measured vibration signal in amplitude and phase representation A second-order Taylor expansion yields the multi-source coupled vibration signal model: ； in, , , They represent instantaneous phasors respectively. exist n The first and second derivatives of 0; S6. Solve the error equation of the measured vibration signal and the multi-source coupled vibration signal model to obtain the instantaneous vector of the measured vibration signal, which is the modulation signal of the original vibration signal, thereby realizing the decoupling of the vibration signal.

2. The decoupling method for multi-source coupled vibration signals in a hydropower unit according to claim 1, characterized in that: In step S2, the collected vibration signal is initially denoised using a filter and a low-rank sparse decomposition method.

3. The decoupling method for multi-source coupled vibration signals in a hydropower unit according to claim 1, characterized in that: In step S2, the measured vibration signal is used It means that among them n =0, 1, …, N .

4. The decoupling method for multi-source coupled vibration signals in a hydropower unit according to claim 1, characterized in that: In step S4, the weighted measured vibration signal is: The weighted measured vibration signal A constant frequency is obtained after performing a Fourier transform. f 0.

5. The decoupling method for multi-source coupled vibration signals in a hydropower unit according to claim 4, characterized in that: In step S4, the constant frequency is obtained by solving the problem using either two-spectral-line interpolation discrete Fourier transform or three-spectral-line interpolation discrete Fourier transform. f 0.

6. The decoupling method for multi-source coupled vibration signals in a hydropower unit according to claim 5, characterized in that: In step S6, the measured vibration signal Vibration signal model coupled with multiple vibration sources The multiple sets of error equations are: 。 7. The decoupling method for multi-source coupled vibration signals in a hydropower unit according to claim 6, characterized in that: The least squares method is used to obtain multiple sets of error equations. Minimum , , , to obtain the instantaneous vector .

Images