Method for decoupling multi-component vibration signals of underwater rotating machinery based on order analysis

Abstract

The application discloses a kind of underwater rotating machinery multi-component vibration signal decoupling methods based on order analysis.Method includes: first, the discrete time domain vibration signal of underwater rotating machinery is obtained by superimposing multiple harmonic components;Then the discrete time domain vibration signal is transformed into angular domain signal, performs discrete Fourier transform, and extracts the amplitude spectrum of each order component;On this basis, the order amplitude spectrum is identified by using multi-peak detection algorithm, and the main order component order set is screened out;Finally, the time domain waveform corresponding to the main order component is extracted by order decoupling method.The application can effectively decouple the multi-component vibration signal in underwater rotating machinery, and provide quantitative basis for locating excitation source, analyzing excitation source characteristics or fault diagnosis.

Description

TECHNICAL FIELD

[0001] The application relates to a vibration signal decoupling method and relates to the technical field of underwater rotating machinery vibration signal analysis. BACKGROUND

[0002] In the field of marine equipment, the vibration signal of underwater rotating machinery (such as a propeller, a driving motor and a pump) contains key information representing the running state of the equipment. Due to the variable speed and variable load characteristics of the underwater machinery, the vibration signal presents the significant characteristics of non-stationary and multi-component coupling. The existing vibration signal analysis methods (such as wavelet transform and empirical mode decomposition) mainly face the following technical bottlenecks in the application of underwater equipment: 1) when the machinery is in the time-varying speed working condition, the non-stationary characteristics of the signal will cause the frequency spectrum energy to spread, produce the frequency band widening effect, the vibration signal analysis technology will produce the frequency spectrum blur phenomenon, the frequency resolution of the harmonic component is significantly reduced, and the effective tracking of the speed-related characteristics is difficult to realize; 2) the underwater environment has stronger environmental noise, the unique wideband noise and the harmonic vibration of the machinery vibration signal are coupled, the conventional signal analysis means is difficult to effectively separate the signals, and the signal-to-noise ratio is reduced in the vibration signal feature extraction process; 3) the existing adaptive signal decomposition methods (such as empirical mode decomposition and variational mode decomposition) have poor adaptability when processing multi-component signals, it is difficult to accurately extract the order component directly related to the rotating excitation, and the prior knowledge of the rotating machinery vibration signal characteristics is lacked, so that the component separation effect is not ideal. Especially when the signal contains multiple similar order components or the instantaneous frequency suddenly changes, the traditional method is often difficult to realize the component separation with clear physical meaning.

[0003] Therefore, an analysis method capable of adaptively tracking the time-varying speed and accurately decoupling the multi-order components is required to solve the analysis limitations of the traditional method under the underwater variable speed and variable load working condition. SUMMARY

[0004] In order to solve the problems in the background art, the application provides a multi-component vibration signal decoupling method for underwater rotating machinery based on order analysis. The method can accurately extract the specific order component, and provide a quantitative basis for locating the excitation source, analyzing the excitation source characteristics or performing fault diagnosis.

[0005] The technical scheme adopted by the application is:

[0006] The multi-component vibration signal decoupling method for underwater rotating machinery based on order analysis comprises:

[0007] Step a) first performs multi-component vibration signal characterization, and obtains the discrete time domain vibration signal of the underwater rotating machinery superimposed by multiple harmonic components.

[0008] Step b) based on the real-time speed signal of the underwater rotating machinery, performing angle domain Fourier transform on the discrete time domain vibration signal to obtain the amplitude spectrum of each order harmonic component.

[0009] Step c) using a multi-peak detection algorithm to identify the main order component for the amplitude spectrum of each order harmonic component, and further screening out the main order harmonic component order set.

[0010] Step d) based on the real-time speed signal of the underwater rotating machinery and the main order harmonic component order set, using the order component decoupling method to adaptively track and decouple each main order harmonic component, and finally extracting the time domain waveform of each main order harmonic component to identify the excitation source of each main order harmonic component, and realizing the decoupling of the multi-component vibration signal of the underwater rotating machinery.

[0011] In step b), first, an interpolation algorithm is used to convert the discrete time domain vibration signal into an angle domain multi-component vibration signal, which contains the cumulative angle obtained based on the real-time speed signal of the underwater rotating machinery; then, after discrete Fourier transform, the complex amplitude of each order harmonic component is extracted, and then the amplitude spectrum is obtained.

[0012] The step c) is specifically as follows:

[0013] Step c1) for each order harmonic component, performing first-order difference processing on the amplitude spectrum of the current order harmonic component to obtain the amplitude variation rate of the adjacent frequency points of the current order harmonic component.

[0014] Step c2) performs local maximum value discrimination to construct a candidate peak set C.

[0015] Step c3) based on a preset vibration amplitude threshold A, screening out the main order harmonic component order set P from the candidate peak set C to eliminate the pseudo-peak value caused by the harmonic sidelobes.

[0016] In step c2), the candidate peak set C is specifically as follows:

[0017] C={k|Δ|Y[k]|<-ε∩Δ|Y[k-1]|>ε}

[0018] Δ|Y[k]|=|Y[k]|-|Y[k-1]|,k≥2

[0019] Wherein, Delta |Y[k]| and Delta |Y[k-1]| are the amplitude variation rates of the adjacent frequency points of the k order and k-1 order harmonic components of the vibration signal respectively, |Y[k]| and |Y[k-1]| are the amplitudes of the k order and k-1 order harmonic components respectively, Y[k] and Y[k-1] are the complex amplitudes of the k order and k-1 order harmonic components respectively; epsilon is a preset noise suppression threshold.

[0020] In the step c3), the main order harmonic component order set P is specifically as follows:

[0021] P={k∈C||Y[k]|≥A}

[0022] Wherein, |Y[k]| is the amplitude of the k order harmonic component, Y[k] is the complex amplitude of the k order harmonic component.

[0023] The step d) is specifically as follows:

[0024] The step d1) obtains the order number of each main order harmonic component according to the main order harmonic component order set P, and further obtains the frequency trajectory of each main order harmonic component.

[0025] The step d2) uses an order tracking method to track each main order harmonic component based on the frequency trajectory of each main order harmonic component, thereby constructing a minimum target model of each main order harmonic component, extracting the time domain waveform of each main order harmonic component after decoupling each main order harmonic component, and realizing adaptive tracking and decoupling of each main order harmonic component.

[0026] In the step d1), for each main order harmonic component, the frequency trajectory F r [n] of the main order harmonic component is specifically as follows:

[0027]

[0028] Wherein, pr is the order number of the current main order harmonic component; R[n] is the real-time rotating speed of the underwater rotating machinery at the nth sampling.

[0029] In the step d2), the minimum target model of each main order harmonic component is specifically as follows:

[0030]

[0031] Using the order tracking method to track the main order component can be expressed as:

[0032] Delta 2 x pr [n] = epsilon[n]

[0033]

[0034] Wherein, y[n] is the time domain waveform of each main order harmonic component; Δ 2 is a second-order difference operator; x pr [n] is the pr main order harmonic component under the n th sampling; y[n] is the observation signal under the n th sampling; ε[n] is the process noise under the n th sampling, which controls the smoothness of the decoupled vibration signal; η[n] is the measurement noise under the n th sampling, which is used to control the fitting error between the decoupled signal and the original observation signal.

[0035] By optimizing the minimum target model, the smoothness and fitting accuracy of the main order component x p [n] are balanced, and the adaptive tracking and decoupling of the main order component are realized.

[0036] The electronic device of the present application comprises a memory and a processor coupled to each other, wherein the memory stores program data, and the processor calls the program data to execute the method as described above.

[0037] The computer readable storage medium of the present application has program data stored thereon, and the program data is executed by the processor to realize the method as described above.

[0038] The method of the present application has the following innovations: first, the order-based signal analysis method can enhance the physical interpretability of the signal, and convert the non-stationary signal with time-varying frequency into a stationary signal with order stability; second, by focusing on specific order components, background noise unrelated to the rotating speed is filtered out, thereby improving the feature extraction accuracy and robustness; third, under time-varying conditions, the homologous vibration components can be accurately decoupled, and the vibration signal components of the underwater rotating machinery are separated.

[0039] The present application has the following advantages:

[0040] The method of the present application models the time domain vibration signal as the superposition of multiple harmonic components, and combines the direct correlation between the time-varying frequency and the rotating speed of the rotating machinery to accurately characterize the multi-component vibration characteristics under the non-stationary rotating speed. The method overcomes the limitations of the traditional Fourier transform under variable speed conditions, and improves the frequency analysis accuracy. At the same time, the stable tracking of the target frequency is maintained under complex working conditions, and the decoupling failure problem caused by the time-varying frequency in the traditional vibration signal analysis method is solved. The present application can effectively decouple the multi-component vibration signal in the underwater rotating machinery, and provides a quantitative basis for positioning the excitation source, analyzing the excitation source characteristics or performing fault diagnosis. BRIEF DESCRIPTION OF DRAWINGS

[0041] Figure 1 The flowchart of the method of the present application;

[0042] Figure 2 The angle domain Fourier transform amplitude waveform of the embodiment of the present application;

[0043] Figure 3 Figure 2 is a frequency trajectory diagram corresponding to the main harmonic component of an embodiment of the present application;

[0044] Figure 4 Figure 3 is a time-domain waveform diagram corresponding to the main harmonic component of an embodiment of the present application. DETAILED DESCRIPTION

[0045] In order to make the person skilled in the art better understand the present application, the technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the drawings, and the embodiments and various drawings cannot limit the protection scope of the present application. Obviously, the described embodiments are only some embodiments of the present application, and all other embodiments obtained by the person skilled in the art without creative labor on the basis of the embodiments in the present application should belong to the protection scope of the present application.

[0046] As shown in Figure 1 the present application, the order analysis-based underwater rotating machinery multi-component vibration signal decoupling method is as follows:

[0047] Step a) first performs multi-component vibration signal characterization to obtain the discrete time-domain vibration signal of the underwater rotating machinery superimposed by multiple harmonic components. In specific implementation, the sensor is arranged as follows: a piezoelectric acceleration sensor (sensitivity: 10 mV / m·s 2 ; range: ±50g; frequency response: 0.5-7000Hz; resolution: 0.0005g) is installed at the radial position of the deep-sea motor shell of the underwater electro-hydraulic actuator to collect the vibration signal of the motor surface. The signal collection system is as follows: a 24-bit high-precision data acquisition card is used to continuously collect the motor vibration signal at a sampling frequency of 10 kHz. During the experiment, the motor speed is linearly increased from 1000 RPM to 2000 RPM at a speed-up rate of 50 RPM / s. The deep-sea motor noise is mainly caused by the electromagnetic-mechanical coupling between the stator and the rotor. When the stator winding magnetic field interacts with the rotor permanent magnet magnetic field, a specific time-space distributed electromagnetic field is generated at the air gap, which causes the deformation of the stator tooth under the action of alternating magnetic force, and further causes periodic vibration. When the motor structure is fixed, the vibration characteristics are dominated by the speed-related harmonic components; if the motor is in an accelerating state, the rotor angle θ changes with time, and the vibration signal presents a non-stationary characteristic, and its discrete form has the same structure as the form of the multi-component vibration signal. Therefore, the multi-component signal analysis method can effectively extract each frequency component in the motor vibration signal and further analyze the change trend of each harmonic component.

[0048] Each harmonic component contains time-varying amplitude, time-varying frequency related to real-time rotating speed of the underwater rotating machinery, and initial phase, and the time-varying frequency is associated with the order fundamental frequency and the rotating frequency. The vibration waveform of the underwater rotating machinery is closely related to the harmonic components generated by the periodic rotation thereof, and the discrete time-domain vibration signal x[n] can be expressed as the superposition of a plurality of harmonic components, i.e.

[0049]

[0050] x k [n]=A k [n]cos(2πf k [n]·n+φ k )

[0051] wherein x k [n] is the k-th order harmonic component at the n-th sampling; A k [n] is the time-varying amplitude of the k-th order harmonic component at the n-th sampling; f k [n] is the time-varying frequency of the k-th order harmonic component at the n-th sampling related to the rotating speed; and φ k is the initial phase of the k-th order harmonic component.

[0052] The time-varying frequency f k [n] at the n-th sampling is further defined according to the relationship between the harmonic and the rotating speed as follows:

[0053]

[0054] wherein R[n] is the real-time rotating speed of the underwater rotating machinery at the n-th sampling, and O R is the order fundamental frequency based on the rotating frequency.

[0055] Step b) is based on the real-time rotating speed signal of the underwater rotating machinery, and the amplitude spectrum of each order harmonic component is obtained after the angle-domain Fourier transform of the discrete time-domain vibration signal, specifically, first, the interpolation algorithm is used to convert the discrete time-domain vibration signal into an angle-domain multi-component vibration signal, and the cumulative angle based on the real-time rotating speed signal of the underwater rotating machinery is contained in the angle-domain multi-component vibration signal; then the complex amplitude of each order harmonic component is extracted after the discrete Fourier transform, and the amplitude spectrum is further obtained.

[0056] The angle-domain Fourier transform is specifically as follows:

[0057] First, the angle-domain signal acquisition is performed: the time-domain vibration signal is converted into the angle-domain multi-component vibration signal y[n] at the n-th sampling through the interpolation algorithm, specifically as follows:

[0058]

[0059] Wherein, N is the total sampling point number of the vibration signal; h(·) is an interpolation kernel function, such as linear interpolation, spline interpolation, etc., in the embodiment, linear interpolation method is used for interpolation processing; θ[n] is the cumulative angle of the underwater rotating machinery under the n-th sampling; Δθ is the angle resolution, in the embodiment, the angle resolution Δθ=0.002496.

[0060] Then, order domain spectrum analysis is carried out: the discrete Fourier transform is performed on the angle domain multi-component vibration signal y[n], and the order component is extracted, specifically as follows:

[0061]

[0062] Wherein, Y[k] is the complex amplitude of the k-th order harmonic component, and the absolute value thereof is the amplitude |Y[k]|, representing the intensity of the order component; j is an imaginary unit.

[0063] As shown in Figure 2 , it is the angle domain Fourier transform amplitude waveform of the embodiment, and the distribution characteristics of each order component are clearly presented. Especially at the 23rd, 25th, 31st and 38th orders, the amplitude appears obvious peak value, indicating that the main components of the vibration signal are concentrated in these specific orders, and by analyzing the order amplitude of the multi-component vibration signal, the main frequency components of the signal can be effectively extracted.

[0064] Step c) the amplitude spectrum of each order harmonic component is subjected to a multi-peak detection algorithm to identify the main order component, and then a main order harmonic component order set is screened out, specifically as follows:

[0065] Step c1) for each order harmonic component, the amplitude spectrum of the current order harmonic component is subjected to first-order difference processing to obtain the amplitude variation rate of adjacent frequency points of the current order harmonic component.

[0066] Step c2) local maximum value discrimination is carried out to construct a candidate peak set C, specifically as follows:

[0067] C={k|Δ|Y[k]|<-ε∩Δ|Y[k-1]|>ε}

[0068] Δ|Y[k]|=|Y[k]|-|Y[k-1]|,k≥2

[0069] Wherein, Δ|Y[k]| and Δ|Y[k-1]| are the amplitude variation rates of adjacent frequency points of the k-th order and the k-1-th order vibration signal harmonic components, respectively; |Y[k]| and |Y[k-1]| are the amplitudes of the k-th order and the k-1-th order harmonic components, respectively; Y[k] and Y[k-1] are the complex amplitudes of the k-th order and the k-1-th order harmonic components, respectively; ε is a preset noise suppression threshold, in the embodiment, the noise suppression threshold ε=0.1.

[0070] Step c3) based on a preset vibration amplitude threshold A, each main harmonic component order set P is screened out from the candidate peak set C to eliminate the pseudo-peak caused by the harmonic sidelobe, specifically as follows:

[0071] P = {k e C ||Y[k]| >= A}

[0072] Wherein, |Y[k]| is the amplitude of k order harmonic component, Y[k] is the complex amplitude of k order harmonic component; in this embodiment, the preset threshold A = 3m / s 2 .

[0073] After multi-peak detection, 23.2, 24.8, 31, 38.2 are selected as main order components.

[0074] Step d) based on the real-time speed signal of the underwater rotating machinery and the main order harmonic component order set, the order component decoupling method is used to adaptively track and decouple each main order harmonic component, and finally the time domain waveform of each main order harmonic component is extracted to identify the excitation source of each main order harmonic component, and the multi-component vibration signal decoupling of the underwater rotating machinery is realized. Specifically as follows:

[0075] Step d1) according to the main order harmonic component order set P, the order number of each main order harmonic component is obtained, and then the frequency trajectory of each main order harmonic component is obtained. For each main order harmonic component, the frequency trajectory F r [n] of the main order harmonic component is specifically as follows:

[0076]

[0077] Wherein, pr is the order number of the current main order harmonic component; R[n] is the real-time speed of the underwater rotating machinery under the nth sampling; in this embodiment, P = {23.2, 24.8, 31, 38.2}.

[0078] Step d2) based on the frequency trajectory of each main order harmonic component, the order tracking method is used to track each main order harmonic component, so as to construct the minimum target model of each main order harmonic component, and the time domain waveform of each main order harmonic component is extracted after decoupling, adaptive tracking and decoupling of each main order harmonic component are realized, and Vold-Kalman filter is used to track and decompose the main order component in the order decoupling. The minimum target model of each main order harmonic component is specifically as follows:

[0079]

[0080] The order tracking method is used to track the main order component, which can be expressed as:

[0081] Delta 2 x pr[n] = ε [n]

[0082]

[0083] where J[n] is the time-domain waveform of each principal harmonic component; Δ 2 is the second-order difference operator; x pr [n] is the pr principal harmonic component at the nth sampling; y[n] is the observed signal at the nth sampling; ε[n] is the process noise at the nth sampling, which controls the smoothness of the decoupled vibration signal; η[n] is the measurement noise at the nth sampling, which is used to control the fitting error between the decoupled signal and the original observed signal.

[0084] By optimizing the minimum target model, the smoothness and fitting accuracy of the principal harmonic component x p [n] are balanced, and adaptive tracking and decoupling of the principal harmonic component are realized.

[0085] As shown in Figure 3 , the frequency trajectory corresponding to the principal harmonic component of the embodiment of the present application is shown, which shows the evolution law of the four characteristic orders in the process of changing the speed.

[0086] As shown in Figure 4 , the time-domain waveform corresponding to the principal harmonic component of the embodiment of the present application is shown, which is the time-domain waveform of the 23.2, 31, 24.8 and 38.2 order components, respectively. By adaptively decoupling signals of different orders, the time-domain waveform of each order can be clearly seen, which conforms to the actual physical characteristics.

[0087] The embodiment verifies the effectiveness of the present application in the decoupling of underwater variable working condition multi-component vibration signals, which is specifically shown as follows:

[0088] 1) The order amplitude spectrum extracted by the angle domain Fourier transform effectively identifies the order components related to the speed.

[0089] 2) The adaptive tracking ability of the method of the present application to the target order component under non-stationary signal conditions is verified, which solves the decoupling failure problem caused by the frequency mutation of the traditional method.

[0090] 3) By decoupling the principal harmonic component, the separated time-domain waveform presents clear independence, and the precise separation of the multi-component vibration signal is realized.

[0091] In summary, the method of the present application significantly improves the precision and reliability of the decoupling of underwater rotating machinery vibration signals, and provides strong technical support for the vibration analysis of equipment under variable speed and variable load working conditions.

[0092] The above-described specific embodiments further illustrate the purpose, technical solutions and beneficial effects of the present application, and it should be understood that the above-described is only a specific embodiment of the present application and is not intended to limit the present application, and any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the protection scope of the present application.

Claims

1. A method for decoupling multi-component vibration signals of underwater rotating machinery based on order analysis, characterized in that, The application relates to a method for decoupling multi-component vibration signals of an underwater rotating machine. The method comprises the following steps: a) obtaining discrete time-domain vibration signals of the underwater rotating machine which are superimposed by multi-harmonic components; b) obtaining amplitude spectra of each order harmonic component by performing angle-domain Fourier transform on the discrete time-domain vibration signals based on a real-time rotating speed signal of the underwater rotating machine; c) identifying main order components by adopting a multi-peak detection algorithm on the amplitude spectra of each order harmonic component, and further screening a main order harmonic component order set; d) adaptively tracking and decoupling each main order harmonic component by adopting an order component decoupling method based on the real-time rotating speed signal of the underwater rotating machine and the main order harmonic component order set, and finally extracting time-domain waveforms of each main order harmonic component to realize decoupling of the multi-component vibration signals of the underwater rotating machine. In the step b), the discrete time-domain vibration signals are first converted into angle-domain multi-component vibration signals by adopting an interpolation algorithm, and the angle-domain multi-component vibration signals contain cumulative angles obtained based on the real-time rotating speed signal of the underwater rotating machine; then the complex amplitudes of each order harmonic component are extracted by performing discrete Fourier transform, and the amplitude spectra are obtained. The step c) is specifically as follows: In the step c1), the amplitude spectrum of the current order harmonic component is first subjected to first-order differential processing to obtain the amplitude variation rate of adjacent frequency points of the current order harmonic component; 2. The order analysis based decoupling method of multi-component vibration signals of underwater rotating machinery according to claim 1, characterized in that: In the step c2), local maximum value discrimination is performed to construct a candidate peak set C; 3. The order analysis based decoupling method of multi-component vibration signals of underwater rotating machinery according to claim 1, characterized in that: In the step c3), each main order harmonic component order set P is screened from the candidate peak set C based on a preset vibration amplitude threshold A. In the step c2), the candidate peak set C is specifically as follows: C={k|Delta|Y[k]|<-epsilon intersect DeltaY[k-1]|>epsilon} Delta|Y[k]|=|Y[k]|-|Y[k-1]|,k>=2 4. The order analysis based decoupling method of multi-component vibration signals of underwater rotating machinery according to claim 3, characterized in that: Wherein, Delta|Y[k]| and Delta|Y[k-1]| are the amplitude variation rates of adjacent frequency points of the kth order and the (k-1)th order harmonic components of the vibration signal, |Y[k]| and |Y[k-1]| are the amplitudes of the kth order and the (k-1)th order harmonic components, Y[k] and Y[k-1] are the complex amplitudes of the kth order and the (k-1)th order harmonic components; epsilon is a preset noise suppression threshold. In the step c3), the main order harmonic component order set P is specifically as follows: P={k element-of C||Y[k]|>=A} Wherein, |Y[k]| is the amplitude of the kth order harmonic component, and Y[k] is the complex amplitude of the kth order harmonic component.

5. The order analysis based decoupling method of multi-component vibration signals of underwater rotating machinery according to claim 3, characterized in that: The step d) is specifically as follows: In the step d1), the order number of each main order harmonic component is obtained according to the main order harmonic component order set P, and the frequency trajectory of each main order harmonic component is obtained; In the step d2), each main order harmonic component is tracked by using an order tracking method based on the frequency trajectory of each main order harmonic component, a minimum target model of each main order harmonic component is constructed, the time-domain waveforms of each main order harmonic component are extracted after decoupling of each main order harmonic component, and adaptive tracking and decoupling of each main order harmonic component are realized.

6. The order analysis based decoupling method of multi-component vibration signals of underwater rotating machinery according to claim 1, characterized in that: Wherein, pr is the order number of the current main order harmonic component; R[n] is the real-time rotating speed of the underwater rotating machine at the nth sampling. ​ ​ 7. The order analysis based decoupling method of multi-component vibration signals of underwater rotating machinery according to claim 6, characterized in that: In step d1), for each primary harmonic component, the frequency trajectory F r [n] In detail as follows: ​ 8. The order analysis based decoupling method of multi-component vibration signals of underwater rotating machinery according to claim 7, characterized in that: In the step d2), the minimization target model of each primary harmonic component is specifically as follows: Δ 2 x pr [n] = ε[n] where J[n] is the time-domain waveform of each primary harmonic component; Δ 2 is the second-order difference operator; x pr [n] is the prth primary harmonic component at the nth sampling; y[n] is the observed signal at the nth sampling; ε[n] is the process noise at the nth sampling; and η[n] is the measurement noise at the nth sampling.

9. An electronic device, comprising: comprising: a memory and a processor coupled to each other, wherein the memory stores program data, and the processor invokes the program data to execute the method according to any one of claims 1-8.

10. A computer readable storage medium having stored thereon program data, wherein, The program data, when executed by the processor, implements the method according to any one of claims 1-8. The program data, when executed by the processor, implements the method according to any one of claims 1-8.

Images