Rotary machine rotor rub-impact diagnosis method based on vibration signal deconstruction and frequency modulation feature Anti-noise enhancement

Abstract

Disclosed is a rotary machine rotor rub-impact diagnosis method based on vibration signal deconstruction and frequency modulation feature anti-noise enhancement. The method comprises: on the basis of second-order integral transformation and high-pass filtering, converting a vibration acceleration signal collected from a rotary mechanical apparatus into a vibration displacement signal; on the basis of an improved variational mode decomposition method, carrying out targeted extraction on a rotation frequency component in the vibration displacement signal; by means of normalized Hilbert transform based on an orthogonal derivative, calculating and estimating the instantaneous change characteristic of a basic frequency of the rotation frequency component to obtain an instantaneous frequency; inputting potential energy parameters and a step length of a Runge-Kutta calculation process, and carrying out intra-wave modulation feature anti-noise enhancement by means of an optimal random resonance system optimized by means of a particle swarm optimization algorithm; and carrying out FFT on the instantaneous frequency of an intra-wave frequency modulation feature subjected to anti-noise enhancement, and on the basis of distribution characteristics of a harmonic amplitude related to a rotation frequency of a rotor, diagnosing and identifying a rub-impact fault of the rotor of a rotary machine. The present invention can accurately diagnose and discriminate the rotor rub-impact fault under complex noise interference.

Description

Rotor rub diagnosis method for rotating machinery based on vibration signal deconstruction and frequency modulation feature noise reduction enhancement Technical Field

[0001] The present invention belongs to the field of rotating machinery state monitoring and fault diagnosis, and in particular relates to a rotating machinery rotor rub diagnosis method based on vibration signal deconstruction and frequency modulation feature anti-noise enhancement. Background Art

[0002] Rotating machinery such as wind turbines, rotary compressors, centrifugal pumps, and generator sets plays a vital role in numerous modern applications, including mining, metallurgy, and energy development. Structural rotors are among the most critical components of these machines. Due to prolonged operation under extremely harsh operating conditions, structural rotors may develop various potential faults, such as imbalance, cracks, misalignment, looseness, vortex, and warping. Rubbing, a common fault exhibiting nonlinear dynamic behavior, occurs when the rotor quasi-periodically impacts the stator of the rotor system. The mechanism is typically due to initial subtle imbalances and slight misalignment of the rotor. This is particularly true when the shaft is operating at extremely high speeds and the radial clearance between the stator and rotor is very small, making rubbing highly susceptible to rotor failure. In fact, with higher requirements for assembly precision and stricter regulations, the machine structures of most power equipment today have become relatively complex and compact. Their components often operate with extremely small clearances, increasing the potential for frictional impact within the rotor system. Once a rubbing fault occurs, it causes dynamic instability in the rotor system, significantly reducing the machine's operating accuracy and adversely affecting the structural properties and performance of related power equipment. It also causes additional defects and damage to related components such as bearings and impellers, ultimately leading to catastrophic system-level failure. Therefore, performing condition monitoring and fault diagnosis on the rotor system and promptly detecting the onset of early rubbing faults in rotating machinery rotors can effectively improve the prognosis and health management of the equipment, thereby ensuring safe and stable industrial production. This is of extremely important practical engineering significance.

[0003] Currently, the following methods are commonly used to monitor and diagnose rotor rubbing faults in rotating machinery:

[0004] (1) Based on conventional spectrum analysis technology, the Fourier spectrum of the collected vibration signal is mainly observed to detect the corresponding rotational frequency, half frequency and high-order harmonic frequency of the fault characteristics including synchronous vibration, subsynchronous vibration and supersynchronous vibration; however, these frequency characteristic signs may also appear when other types of machine faults occur, such as rotor shaft cracks and mechanical looseness, and these characteristic representations are not significant in the early stages of the rotor friction fault of the rotating machinery and are difficult to be detected in time.

[0005] (2) To observe the impact generated by the quasi-periodic contact and separation process between the rotor and stator in the time domain signal, it is usually necessary to use a method such as signal deconstruction to effectively separate and extract this component, mainly including empirical mode decomposition, blind source separation and empirical wavelet transform; however, when other components of the rotating machinery fail, similar fault signs may also occur, such as gear defects and local damage to rolling bearings, and different noise distribution forms and noise levels also tend to affect the effectiveness and usability of the signal deconstruction results.

[0006] (3) Data-driven intelligent methods mainly include an end-to-end process of "data preprocessing - feature extraction - feature selection / dimensionality reduction - feature learning - diagnosis output". The mainstream methods involved are support vector machines, random forests, deep belief networks, convolutional neural networks, etc. However, on the one hand, these methods usually require a large number of labeled, identically distributed, high-quality data samples for training and learning, which is difficult to meet under realistic conditions. On the other hand, the non-transparency and unexplainability of those deep intelligent diagnostic models make them "black boxes", and the diagnostic conclusions they output are difficult to be fully trusted in practical applications.

[0007] (4) Rotating machinery rotor rub fault diagnosis technology based on nonlinear random dynamic behavior analysis mainly includes Volterra series identification and nonlinear output frequency response equation estimation methods; however, the computational complexity of such methods is usually high, and a certain amount of computational analysis time is required to obtain reliable rotor rub fault diagnosis results, which is usually difficult to meet the timeliness of implementing online monitoring and diagnosis in actual engineering applications.

[0008] (5) Rotor rub diagnosis technology based on the extraction of amplitude and frequency modulation features of vibration signals. This type of method often involves time-frequency domain analysis of signals, mainly including short-time Fourier transform, continuous wavelet transform, Hilbert-Huang transform and adaptive linear frequency modulation mode decomposition. However, this type of method is usually less robust in the presence of noise interference, and often fails when analyzing and processing vibration data with low signal-to-noise ratio, which greatly restricts the practical application of these methods.

[0009] As can be seen, the various existing methods for monitoring and diagnosing rotor rub faults in rotating machinery all have their own drawbacks and shortcomings. Therefore, there is a need to develop efficient rotor rub fault diagnosis methods that can promptly detect the onset of early rotor rub faults in complex environments. This can effectively improve the prognosis and health management of equipment, and has extremely important practical engineering significance. Furthermore, since displacement signals typically have a high signal-to-noise ratio and can accurately reflect the vibration state of the rotor, most of the above methods use vibration displacement signals to diagnose and analyze rotor systems. However, it should be noted that while displacement sensors offer advantages such as a wide linear range, zero-frequency response, strong anti-interference capabilities, and ease of calibration, their practical application is not as widespread as that of acceleration sensors. They are generally only suitable for monitoring rotor systems in large rotating machinery, primarily due to complex installation requirements, economic considerations, and other factors. Therefore, rotor rub fault diagnosis methods based on vibration displacement signal analysis may be difficult to implement in some cases. Therefore, it is necessary to develop more universal and versatile advanced diagnostic strategies, as rotor rub faults can occur in various rotating machinery systems.

[0010] Summary of the Invention

[0011] In order to solve the above-mentioned problems in the prior art, the present invention provides a rotating machinery rotor rub diagnosis method based on vibration signal deconstruction and frequency modulation feature anti-noise enhancement, which can diagnose and identify rotor rub faults of power equipment rotating machinery based on vibration acceleration signal analysis under complex noise interference.

[0012] The purpose of the present invention is achieved through the following technical solutions:

[0013] A rotating machinery rotor rub diagnosis method based on vibration signal deconstruction and frequency modulation feature noise reduction enhancement, comprising:

[0014] Step 1: Collect signal sequences from rotating mechanical equipment through vibration acceleration sensors;

[0015] Step 2: Based on second-order integral transformation and high-pass filtering, the collected vibration acceleration signal is converted into a vibration displacement signal;

[0016] Step 3: Targeted extraction of the frequency conversion component in the vibration displacement signal based on the improved variational mode decomposition method;

[0017] Step 4: For the extracted frequency conversion component, use the normalized Hilbert transform based on orthogonal derivatives to calculate and estimate the instantaneous change characteristics of its fundamental frequency to obtain the instantaneous frequency;

[0018] Step 5: Input the calculated instantaneous frequency sequence into the optimal stochastic resonance system to enhance the noise resistance of the intra-wave modulation feature. The potential energy parameters of the stochastic resonance system and the step size of the Runge-Kutta calculation process are optimized by the particle swarm optimization algorithm to obtain the optimal stochastic resonance system.

[0019] Step 6: Perform fast Fourier transform processing on the instantaneous frequency of the intra-wave frequency modulation feature after anti-noise enhancement, and diagnose and identify the rotor rubbing fault of the rotating machinery based on the distribution characteristics of the harmonic amplitude related to the rotor rotation frequency.

[0020] Compared with the prior art, the present invention has the following beneficial effects:

[0021] In response to the defects and shortcomings of existing power equipment rotating machinery rotor rubbing fault diagnosis and analysis technology, the present invention proposes to use vibration acceleration signal analysis and processing as the main method, convert the vibration acceleration signal into a vibration displacement signal through a second-order integral transform, and target the rotation frequency component based on the signal deconstruction method, then calculate and estimate the instantaneous frequency of the rotation frequency component, and then input the obtained instantaneous frequency sequence into the optimal stochastic resonance system for intra-wave modulation feature anti-noise enhancement, and then perform fast Fourier transform processing on the output instantaneous frequency with enhanced frequency modulation feature, and finally diagnose and identify the rotating machinery rotor rubbing fault based on the distribution characteristics of the harmonic amplitude related to the rotor rotation frequency. This method is based on the more common and widely used vibration acceleration signal processing technology, has excellent robustness to the inevitable complex noise interference in the signal acquisition and signal conversion processing process in industrial environments, and is of great significance for practical engineering applications. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] FIG1 is a flow chart of a method for diagnosing rotor rub of a rotating machinery based on vibration signal deconstruction and frequency modulation feature anti-noise enhancement according to the present invention.

[0023] FIG2 is a time domain waveform diagram of a vibration acceleration signal sequence collected from a centrifugal pump by a vibration acceleration sensor in an embodiment of the present invention.

[0024] FIG3 is a time domain waveform diagram of a vibration displacement signal converted from a vibration acceleration signal by second-order integral transformation and high-pass filtering in an embodiment of the present invention.

[0025] FIG4 is a spectrum diagram of a vibration displacement signal obtained by converting a vibration acceleration signal into a vibration displacement signal through second-order integral transformation and high-pass filtering in an embodiment of the present invention.

[0026] FIG5 is a time-frequency diagram of the frequency conversion component in the vibration displacement signal extracted by the improved variational mode decomposition method in an embodiment of the present invention.

[0027] FIG6 is a diagram of the instantaneous frequency spectrum of the frequency conversion component estimated using the normalized Hilbert transform based on orthogonal derivatives in an embodiment of the present invention.

[0028] FIG7 is a diagram of an instantaneous frequency spectrum with enhanced frequency modulation characteristics output by the optimal stochastic resonance system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0029] The present invention will be described in detail below with reference to the accompanying drawings and preferred embodiments, and the purpose and effects of the present invention will become more apparent. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0030] The present invention is based on vibration acceleration signal analysis. First, the vibration acceleration signal is converted into a vibration displacement signal through a second-order integral transform. Then, the rotation frequency component is targetedly extracted through a signal deconstruction method. The instantaneous frequency of the rotation frequency component is calculated and estimated. Then, the obtained instantaneous frequency sequence is input into the optimal stochastic resonance system for intra-wave modulation feature anti-noise enhancement. Finally, the instantaneous frequency with enhanced frequency modulation feature of the optimal stochastic resonance output is subjected to fast Fourier transform processing. Based on the distribution characteristics of the harmonic amplitude related to the rotor rotation frequency, the rotor friction fault of the rotating machinery is diagnosed and identified, thereby providing important technical support for timely monitoring and control of the operating status of the rotating machinery rotor system of power equipment, and safe and efficient operation and maintenance.

[0031] As shown in Figure 1, a rotating machinery rotor rub diagnosis method based on vibration signal deconstruction and frequency modulation feature noise reduction enhancement includes the following steps:

[0032] Step 1: Collect signal sequences from rotating mechanical equipment through vibration acceleration sensors.

[0033] In step 1, the signal sequence is collected from the rotating mechanical equipment through the vibration acceleration sensor, which is recorded as N is the total number of sampling points, and the sampling frequency is recorded as f s ; and during the signal acquisition process, the frequency response parameters of the vibration acceleration sensor used should be no less than 2kHz, the sampling frequency of the vibration signal sequence should be no less than 5.12kHz and no more than 100kHz, and the sampling time should be no less than 1s.

[0034] Step 2: Based on the second-order integral transformation and high-pass filtering, the collected vibration acceleration signal is converted into a vibration displacement signal. Step 2 includes the following sub-steps:

[0035] (2-1) First, perform an integration to convert the vibration acceleration signal Converted into vibration velocity signal as follows:

[0036] Where ΔT = 1 / f s , S a By signal The result is obtained by filtering with a high-pass filter; in order to avoid the adverse effects of the extremely low-frequency components in the original vibration signal on the final integration result, the cutoff frequency W of the high-pass filter is BG1 Usually 10Hz to 25Hz is used to obtain better signal integration results.

[0037] (2-2) Next, for the vibration velocity signal obtained Similarly, it is first filtered through a high-pass filter to remove the extremely low-frequency components. The cutoff frequency of the filter is W BG2 According to experience, 10Hz~25Hz is selected to obtain the filtered signal S v Then for S v Perform integration operation to obtain the final vibration displacement signal S x : S x (1)=(S v (1)+S v (2)) / 2×ΔT S x (i) = S x (i-1)+(S v (i)+S v (i+1)) / 2×ΔT,i=2,3,...,N

[0038] In step 2, a high-pass filter with a stopband attenuation of 60 dB is used to perform zero-phase filtering on the vibration acceleration signal and the vibration velocity signal respectively, and the high-pass filter attenuates frequencies lower than the specified passband frequency, which can compensate for the delay introduced by the digital filter.

[0039] Step 3: Targeted extraction of the frequency conversion component in the vibration displacement signal based on the improved variational mode decomposition method. Step 3 includes the following sub-steps:

[0040] (3-1) First, regarding the signal S x A series of eigenmode equations u k (t), k∈(1,2,...,K), K is the total number of modes, which is defined as an AM-FM signal and can be expressed as:

[0041] Among them A k (t) and denote the instantaneous amplitude and instantaneous phase respectively, and is a non-decreasing function, that is, the instantaneous frequency always satisfies and In comparison, A k(t) and ω k The rate of change of (t) is almost extremely slow, so these eigenmode functions can be regarded as a series of bandwidth-limited signals, which can be expressed by the variational method as follows:

[0042] where δ(·) is the Dirac equation, * is the convolution operator; BW k is u k (t) is the norm of the bandwidth in the frequency domain. At this point, the constrained variational problem can be recast as:

[0043] Among them, ω k represents the center frequencies of the K intrinsic mode functions, k∈(1,2,...,K).

[0044] (3-2) Introducing the quadratic penalty term and the Lagrange multiplier term to solve the minimization problem is as follows:

[0045] Where α represents the bandwidth balance parameter, and λ(t) is the Lagrange multiplier factor. Solving this equation using the alternating direction Lagrange multiplier method yields the following equation about the vibration displacement signal S: x K eigenmode functions u of (t) k (t),k∈(1,2,...,K).

[0046] (3-3) Based on the improved variational mode decomposition method, the vibration displacement signal S x Extract the frequency conversion component in the signal S x Before performing variational mode decomposition and eigenmode function extraction, the mode number K is set to 1, the initial bandwidth balance parameter α is set to 5000, and the initial center frequency of the mode is set to the rotation frequency f roc The MTFI target frequency index is used to evaluate the spectral clarity of the signal deconstruction results, where the optimal parameters of the deconstruction algorithm are obtained by alternating forward and reverse iterations.

[0047] The calculation formula of the target frequency index MTFI is as follows: F * (f roc )=max[F(f roc -0.02f roc ,f roc +0.02f roc )]

[0048] Where F(f j ) indicates the output signal at frequency f j The amplitude of the Fourier spectrum corresponding to * (f roc) indicates that at frequency f roc The maximum frequency amplitude is within the range of ±2% of the value near.

[0049] Finally, the eigenmode function u corresponding to the maximum MTFI value is target This is the extracted frequency conversion component.

[0050] Step (3-3) includes the following sub-steps:

[0051] (3-3-1) Input vibration displacement signal S x , initialize the iteration count Iter = 0, set the maximum number of iterations MaxIter = 500, and set the search step size of the bandwidth balance parameter to Δα = 500;

[0052] (3-3-2) Using variational mode decomposition to analyze the vibration displacement signal S x Perform a deconstruction to obtain the intrinsic mode function u0 and its target frequency index MTFI0;

[0053] (3-3-3) Start the loop, let Iter = Iter + 1, and change the bandwidth balance parameters to α1 = α0 + Δα and α2 = α0 - Δα respectively; Next, substitute the obtained two bandwidth balance parameter values ​​α1 and α2 into the variational mode decomposition algorithm for the vibration displacement signal S x Deconstruct and get u 1,Iter and u 2,Iter and the corresponding MTFI 1,Iter and MTFI 2,Iter ;

[0054] (3-3-4) If MTFI0>MTFI 1,Iter And MTFI0>MTFI 2,Iter , then the loop is terminated and the final target frequency-related component u is output target =u0; if MTFI0 does not meet the above conditions, then update MTFI0=max(MTFI 1,Iter ,MTFI 2,Iter )as well as

[0055] (3-3-5) If Iter>MTFI, the loop is terminated and the vibration displacement signal S is output using the variational mode decomposition method under the action of the latest bandwidth balance parameter α0. x Deconstruct and get the result u target ; Otherwise, return to step (3-3-3);

[0056] (3-3-6) End the loop and output the final target frequency component u target .

[0057] In steps (3-3-2), (3-3-3) and (3-3-5), the vibration displacement signal S is decomposed using the variational mode decomposition method. x Deconstruction includes the following sub-steps:

[0058] (1) Let u k The Fourier transforms of (t), f(t) and λ(t) are and initialization and And the iteration count n=0;

[0059] (2) Iteration count n = n + 1;

[0060] (3) For k = 1, 2, ..., K, update all As follows:

[0061] Update ω at the same time k :

[0062] (4) Update

[0063] Wherein τ is the noise tolerance factor. In order to avoid the influence of strong noise on the signal deconstruction result and since completely reconstructing the signal is not the ultimate goal of the present invention, the noise tolerance factor is set to τ=0 to regulate the influence of noise on the deconstruction process.

[0064] (5) Repeat (2)-(4) until the termination condition is met:

[0065] Where ε represents the convergence factor, and ε = 1 × 10 -6 .

[0066] Step 4: For the extracted frequency conversion component, use the normalized Hilbert transform based on orthogonal derivatives to calculate and estimate the instantaneous change characteristics of its fundamental frequency to obtain the instantaneous frequency. Step 4 includes the following sub-steps:

[0067] (4-1) For a single-component AM-FM signal g(t), if its amplitude is not normalized, all the maximum extreme values ​​in its absolute value form are searched, and then a cubic spline function is fitted to these extreme values ​​to obtain the empirical envelope function B0(t);

[0068] (4-2) Normalize the amplitude of the signal g(t), that is, g1(t) = g(t) / B0(t)

[0069] If the amplitude of g1(t) is still not completely normalized, return to step (4-1) until the Lth empirical envelope function B L The deviation of the amplitude of (t) from 1 is less than 10 -5 When the output signal g L+1 (t) is considered a pure FM signal;

[0070] (4-3) Calculate the instantaneous amplitude of signal g(t): Signal B(t) = B0(t)·B1(t)·…·B L (t), recalculate the frequency modulation signal F(t) of g(t): F(t)=g(t) / B(t)

[0071] (4-4) Order Then the instantaneous phase φ(t) is estimated as: φ(t) = arctan[Q(t) / F(t)]

[0072] Based on this, the instantaneous frequency ω(t) is calculated and estimated as:

[0073] And it always satisfies ω(t)>0.

[0074] Step 5: Input the calculated instantaneous frequency sequence into the optimal stochastic resonance system to enhance the noise resistance of the intra-wave modulation feature. The potential energy parameters of the stochastic resonance system and the step size of the Runge-Kutta calculation process are optimized by the particle swarm optimization algorithm to obtain the optimal stochastic resonance system.

[0075] Step 5 includes the following sub-steps:

[0076] (5-1) Construct a stochastic resonance system, which is expressed as follows:

[0077] Where x represents the trajectory of the particle, S(t) represents a weak periodic signal, N(t) represents Gaussian white noise, and U(x) represents the potential function;

[0078] (5-2) Using the reflection symmetric quartic potential function as U(x),

[0079] Where a and b represent the positive parameters of the potential barrier and potential well of the control potential function U(x), respectively;

[0080] The stochastic resonance system is derived as:

[0081] (5-3) Solve the stochastic resonance system based on the fourth-order Runge-Kutta method and obtain k1=ax(i)-bx(i) 3 +S(i)+N(i) k2=a(x(i)+hk1 / 2)-b(x(i)+hk1 / 2)3 +S(i)+N(i) k3=a(x(i)+hk2 / 2)-b(x(i)+hk2 / 2) 3 +S(i+1)+N(i+1) k4=a(x(i)+hk3)-b(x(i)+hk3) 3 +S(i+1)+N(i+1) x(i+1)=x(i)+(k1+2k2+2k3+k4)h / 6

[0082] Where x(i), S(i) and N(i) represent the discrete forms of signals x(t), S(t) and N(t), respectively, and h represents the step size of the calculation;

[0083] (5-4) The output x(i) of the stochastic resonance system is evaluated using the signal-to-noise ratio (SNR).

[0084] Where P(·) represents the power spectrum of signal x(i), Δf=f s / N is the spectrum resolution, f m is the target frequency component, that is, the frequency conversion, so f m =f roc ; round(·) represents the rounding operator;

[0085] (5-5) For a certain input instantaneous frequency signal instantaneous frequency ω(t), which covers the components S(t) and N(t), it is necessary to properly adjust the values ​​of a, b and h to achieve the optimal stochastic resonance effect, as follows:

[0086] The present invention uses particle swarm optimization algorithm to search for the optimal (a * ,b * ,h * ) combination, and the optimal stochastic resonance system under this parameter combination performs intra-wave modulation feature noise enhancement on the input instantaneous frequency ω(t), and the output is obtained

[0087] Step (5-5) includes the following sub-steps:

[0088] (5-5-1) Input the estimated instantaneous frequency ω(t) into the stochastic resonance system, let S(t)+N(t)=ω(t), and initialize the number of particle swarms N p=50, set the dimension of the parameter search space to 3, set the search upper and lower bounds of the parameters (a, b, h) to LB = [0, 0, 0] and HB = [0.05, 5000, 0.05] respectively, set the particle inertia weight to w = 0.6, set the acceleration constant to γ1 = γ2 = 1.5, and the absolute value of the maximum velocity of the particle motion is less than 0.8; initialize the iteration count IterPSO = 0, set the maximum number of iterations MaxIterPSO = 500, and set the precision coefficient of the iterative process to ε P =0.001;

[0089] (5-5-2) Randomly initialize each particle and substitute the parameters of these particles into the stochastic resonance system to calculate the corresponding output; find the current global optimal output and its SNR0;

[0090] (5-5-3) Start the loop, set IterPSO = IterPSO + 1, update and adjust the position and speed of each particle, and re-substitute the parameters of these particles into the stochastic resonance system to calculate the corresponding output; find the current global optimal output and its SNR IterPSO ;

[0091] (5-5-4) Calculate the signal-to-noise ratio change ΔSNR = |SNR0-SNR IterPSO |, and update SNR0=max(SNR0,SNR IterPSO )and

[0092] (5-5-5) If IterPSO>MaxIterPSO or ΔSNR<ε P , the loop is terminated and output Otherwise, return to (5-5-3);

[0093] (5-5-6) End the loop and output the final optimal parameter combination (a * ,b * ,h * ) is the output result of the intra-wave modulation feature noise enhancement of the input instantaneous frequency ω(t) under the optimal stochastic resonance system

[0094] Step 6: Perform fast Fourier transform processing on the instantaneous frequency of the intra-wave frequency modulation feature after anti-noise enhancement, and diagnose and identify the rotor rubbing fault of the rotating machinery based on the distribution characteristics of the harmonic amplitude related to the rotor rotation frequency.

[0095] In step 6, the instantaneous frequency The Fourier spectrum of the rotor frequency f rocThe presence of a rubbing fault in the rotor of a rotating machine can be determined only when the distribution characteristics of the relevant harmonic amplitudes meet the following two conditions:

[0096] (1) There is a clear rotor rotation frequency f that can be clearly distinguished from noise and other mechanical vibration components roc and its harmonics;

[0097] (2) Double the frequency f roc The amplitude is the maximum.

[0098] The effect of the method of the present invention is demonstrated below through a specific embodiment.

[0099] S01, a vibration acceleration sensor is used to collect a signal sequence from a centrifugal pump that may have a potential rotor rubbing fault. The time domain waveform of the vibration acceleration signal is shown in Figure 2. The centrifugal pump runs at a speed of 1500 rpm, so its rotor rotation frequency f roc It can be calculated as 25Hz, the sampling frequency of the vibration acceleration sensor is 5120Hz, and the sampling time is 2s. It is difficult to observe the characteristic signs associated with the centrifugal pump rotor rubbing fault from the time domain waveform of the vibration acceleration signal sequence in Figure 2.

[0100] S02, the collected vibration acceleration signal is converted into a vibration displacement signal based on second-order integral transformation and high-pass filtering, wherein the cutoff frequency W of the high-pass filter in the two high-pass filtering processes is BG1 and W BG2 The frequency domain waveform of the converted vibration displacement signal is shown in Figure 3, and its spectrum is shown in Figure 4. As can be seen from Figure 4, except for the more significant rotation frequency f roc In addition to the above (marked with ○), there are other frequency components.

[0101] In step S03, we used an improved variational mode decomposition method to target the rotational frequency component in the vibration displacement signal and calculated and estimated the instantaneous variation characteristics (i.e., frequency modulation characteristics) of its fundamental frequency using a normalized Hilbert transform based on orthogonal derivatives. The resulting time-frequency plot of the rotational frequency component is shown in Figure 5. It can be observed that the rotational frequency of the vibration signal exhibits intra-wave modulation, indicating a potential fault in the centrifugal pump rotor. However, the specific fault type still needs to be further clarified.

[0102] S04, draw the instantaneous frequency spectrum of the rotation frequency component estimated by the normalized Hilbert transform based on the orthogonal derivative, as shown in Figure 6. It can be seen that due to the interference of noise and other complex factors, the Fourier spectrum of the instantaneous frequency is not consistent with the rotor rotation frequency f roc The distribution characteristics of the relevant harmonic amplitudes cannot meet the judgment criteria for rotor rubbing failure, so it is still difficult to determine whether the centrifugal pump has a rotor rubbing failure.

[0103] S05, the calculated instantaneous frequency sequence is input into the optimal stochastic resonance system to enhance the intra-wave modulation feature noise reduction, wherein the potential energy parameters of the stochastic resonance system and the step size of the Runge-Kutta calculation process are optimized by the particle swarm optimization algorithm, and the optimal (a * ,b * ,h * ) parameter combination is (2.9×10 -4 ,4.4×10 3 ,0.0015). The instantaneous frequency of the frequency modulation characteristic enhancement of the optimal stochastic resonance system output The spectrum is shown in Figure 7, which shows a clear rotor rotation frequency f that can be clearly distinguished from noise and other mechanical vibration components. roc and its harmonics, and double the frequency f roc The amplitude is the largest. Therefore, it can be determined that the centrifugal pump has a rotor rubbing fault during operation.

[0104] Those skilled in the art will understand that the foregoing descriptions are merely preferred embodiments of the invention and are not intended to limit the invention. Although the invention has been described in detail with reference to the foregoing examples, those skilled in the art will still be able to modify the technical solutions described in the foregoing examples or substitute equivalents for some of the technical features therein. Any modifications, equivalent substitutions, etc. made within the spirit and principles of the invention shall be included within the scope of protection of the invention.

Claims

1. A rotating machinery rotor rubbing diagnosis method based on vibration signal deconstruction and frequency modulation feature anti-noise enhancement, characterized in that: include: Step 1: Collect signal sequence from rotating mechanical equipment through vibration acceleration sensor; Step 2: Based on second-order integral transformation and high-pass filtering, the collected vibration acceleration signal is converted into a vibration displacement signal; Step 3: Targeted extraction of the frequency conversion component in the vibration displacement signal based on the improved variational mode decomposition method; Step 4: For the extracted frequency conversion component, use the normalized Hilbert transform based on orthogonal derivatives to calculate and estimate the instantaneous change characteristics of its basic frequency to obtain the instantaneous frequency; Step 5: Input the calculated instantaneous frequency sequence into the optimal stochastic resonance system to enhance the noise resistance of the intra-wave modulation feature, where the potential energy parameters of the stochastic resonance system and the step size of the Runge-Kutta calculation process are optimized by the particle swarm optimization algorithm to obtain the optimal stochastic resonance system; Step 6: Perform fast Fourier transform processing on the instantaneous frequency of the intra-wave frequency modulation feature after anti-noise enhancement, and diagnose and identify the rotor friction fault of the rotating machinery based on the distribution characteristics of the harmonic amplitude related to the rotor rotation frequency.

2. The rotating machinery rotor rubbing diagnosis method based on vibration signal deconstruction and frequency modulation feature anti-noise enhancement according to claim 1 is characterized in that: In the step 1, a signal sequence is collected from the rotating mechanical equipment through a vibration acceleration sensor, which is recorded as N is the total number of sampling points, and the sampling frequency is recorded as f s ; and during the signal acquisition process, the frequency response parameters of the vibration acceleration sensor used should be no less than 2kHz, the sampling frequency of the vibration signal sequence should be no less than 5.12kHz and no more than 100kHz, and the sampling time should be no less than 1s.

3. The rotating machinery rotor rubbing diagnosis method based on vibration signal deconstruction and frequency modulation feature anti-noise enhancement according to claim 1 is characterized in that: The step 2 includes the following sub-steps: (2-1) First, perform an integration to convert the vibration acceleration signal Converted into vibration velocity signal as follows: Where, ΔT = 1 / f s , S a By signal After filtering through a high-pass filter, the cutoff frequency of the high-pass filter is W BG1 Take 10Hz~25Hz; (2-2) For the obtained vibration velocity signal Similarly, it is filtered through a high-pass filter to remove the extremely low-frequency components. The cutoff frequency of the filter is W BG2 Take 10Hz~25Hz to get the filtered signal S v ; Then for S v Perform integration operation to obtain the final vibration displacement signal S x : S x (1)=(S v (1)+S v (2)) / 2×ΔT S x (i)=S x (i-1)+(S v (i)+S v (i+1)) / 2×ΔT,i=2,3,...,N In step 2, a high-pass filter with a stopband attenuation of 60 dB is used to perform zero-phase filtering on the vibration acceleration signal and the vibration velocity signal, and the high-pass filter attenuates frequencies lower than the specified passband frequency, which can compensate for the delay introduced by the digital filter.

4. The rotating machinery rotor rubbing diagnosis method based on vibration signal deconstruction and frequency modulation feature anti-noise enhancement according to claim 1 is characterized in that: The step three includes the following sub-steps: (3-1) Regarding the signal S x A series of eigenmode equations u k (t), k∈(1,2,...,K), K is the total number of modes, defined as a series of bandwidth-limited AM-FM signals, expressed as: Among them, A k (t) and denote the instantaneous amplitude and instantaneous phase respectively, and is a non-decreasing function, that is, the instantaneous frequency always satisfies It can be expressed by variational method as: where δ(·) is the Dirac equation, * is the convolution operator; BW k is u k (t) norm of bandwidth in the frequency domain; At this point, the constrained variational problem is reformulated as: Among them, ω k represents the center frequency of K intrinsic mode functions, k∈(1,2,...,K); (3-2) The quadratic penalty term and the Lagrange multiplier term are introduced to solve the constrained variational problem as follows: Where α represents the bandwidth balance parameter, and λ(t) is the Lagrange multiplier factor. Solving this equation by the alternating direction Lagrange multiplier method yields the following equation about the vibration displacement signal S: x (t) K eigenmode functions u k (t), k∈(1,2,...,K); (3-3) Based on the improved variational mode decomposition method, the vibration displacement signal S x The frequency conversion component in the signal S is extracted. x Before performing variational mode decomposition and eigenmode function extraction, the mode number K is set to 1, the initial bandwidth balance parameter α is set to 5000, and the initial center frequency of the mode is set to the rotation frequency f roc The target frequency index MTFI is used to evaluate the spectral clarity of the signal deconstruction result, where the optimal parameters of the deconstruction algorithm are obtained by searching for the optimal parameters through the forward and reverse alternating iteration method; The calculation formula of the target frequency index MTFI is as follows: F * (f roc )=max[F(f roc -0.02f roc ,f roc +0.02f roc )] Where F(f j ) indicates that the output signal is at frequency f j The amplitude of the Fourier spectrum corresponding to * (f roc ) indicates that at frequency f roc The maximum frequency amplitude within the range of ±2% of the value near; Finally, the intrinsic mode function u corresponding to the maximum MTFI value is target This is the extracted frequency conversion component.

5. The rotating machinery rotor rubbing diagnosis method based on vibration signal deconstruction and frequency modulation feature anti-noise enhancement according to claim 4 is characterized in that: The step (3-3) includes the following sub-steps: (3-3-1) Input vibration displacement signal S x , initialize the iteration count Iter = 0, set the maximum number of iterations MaxIter = 500, and set the search step size of the bandwidth balancing parameter to Δα = 500; (3-3-2) Using variational mode decomposition to analyze the vibration displacement signal S x Perform a deconstruction to obtain the intrinsic mode function u0 and its target frequency index MTFI0; (3-3-3) Start the loop, let Iter = Iter + 1, and change the bandwidth balance parameters to α1 = α0 + Δα and α2 = α0 - Δα respectively; Next, substitute the obtained two bandwidth balance parameter values ​​α1 and α2 into the variational mode decomposition algorithm for the vibration displacement signal S x Deconstruct and get u 1,Iter and u 2,Iter and the corresponding MTFI 1,Iter and MTFI 2,Iter ; (3-3-4) If MTFI0>MTFI 1,Iter And MTFI0>MTFI 2,Iter , the loop is terminated and the final target frequency-related component u is output target =u0; if MTFI0 does not meet the above conditions, then update MTFI0 = max(MTFI 1,Iter ,MTFI 2,Iter )as well as (3-3-5) If Iter>MaxIter, the loop is terminated and the vibration displacement signal S is output using the variational mode decomposition method under the action of the latest bandwidth balance parameter α0. x Deconstruct and get the result u target ; Otherwise, return to step (3-3-3); (3-3-6) End the loop and output the final target frequency component u target .

6. The rotating machinery rotor rubbing diagnosis method based on vibration signal deconstruction and frequency modulation feature anti-noise enhancement according to claim 5 is characterized in that: In the steps (3-3-2), (3-3-3) and (3-3-5), the vibration displacement signal S is decomposed using the variational mode decomposition method. x Deconstruction includes the following sub-steps: (1) Let u k The Fourier transforms of (t), f(t) and λ(t) are and initialization and And the iteration count n=0; (2) Iteration count n = n + 1; (3) For k = 1, 2, ..., K, update all As follows: Update at the same time k : (4) Update Where τ is the noise tolerance factor, τ = 0, which is used to regulate the impact of noise on the deconstruction process; (5) Repeat (2)-(4) until the termination condition is met: Where ε represents the convergence factor, ε=1×10 -6 .

7. The rotating machinery rotor rubbing diagnosis method based on vibration signal deconstruction and frequency modulation feature anti-noise enhancement according to claim 1 is characterized in that: The step 4 includes the following sub-steps: (4-1) For a single-component AM-FM signal g(t), if its amplitude is not normalized, search for all maximum extreme values ​​in its absolute value form, and then fit these extreme values ​​with a cubic spline function to obtain the empirical envelope function B0(t); (4-2) Normalize the amplitude of the signal g(t), that is: g1(t)=g(t) / B0(t) If the amplitude of g1(t) is still not completely normalized, return to step (4-1) until the Lth empirical envelope function B L The deviation of the amplitude of (t) from 1 is less than 10 -5 When the output signal g L+1 (t) is considered as a pure FM signal; (4-3) Calculate the instantaneous amplitude signal of signal g(t): B(t) = B0(t)·B1(t)·…·B L (t), recalculate the frequency modulated signal F(t) of g(t): F(t)=g(t) / B(t) (4-4) Order Then the instantaneous phase φ(t) is estimated as: φ(t)=arctan[Q(t) / F(t)] Based on this, the instantaneous frequency ω(t) is calculated and estimated as: And ω(t)>0 is always satisfied.

8. The rotating machinery rotor rubbing diagnosis method based on vibration signal deconstruction and frequency modulation feature anti-noise enhancement according to claim 1 is characterized in that: The step five includes the following sub-steps: (5-1) Construct a stochastic resonance system, which is expressed as follows: Among them, x represents the trajectory of the particle, S(t) represents a weak periodic signal, N(t) represents Gaussian white noise, and U(x) represents the potential function; (5-2) Using the reflection symmetric quartic potential function as U(x), Where a and b represent the positive parameters of the potential barrier and potential well of the control potential function U(x), respectively; The stochastic resonance system is derived as: (5-3) Based on the fourth-order Runge-Kutta method, we can solve the stochastic resonance system and obtain k1=ax(i)-bx(i) 3 +S(i)+N(i) k2=a(x(i)+hk1 / 2)-b(x(i)+hk1 / 2) 3 +S(i)+N(i) k3=a(x(i)+hk2 / 2)-b(x(i)+hk2 / 2) 3 +S(i+1)+N(i+1) k4=a(x(i)+hk3)-b(x(i)+hk3) 3 +S(i+1)+N(i+1) x(i+1)=x(i)+(k1+2k2+2k3+k4)h / 6 Where x(i), S(i) and N(i) represent the discrete forms of signals x(t), S(t) and N(t) respectively, and h represents the step size of the calculation; (5-4) The signal-to-noise ratio (SNR) is used to evaluate the output x(i) of the stochastic resonance system. Where P(·) represents the power spectrum of signal x(i), Δf = f s / N is the spectrum resolution, f m is the target frequency component, that is, the conversion frequency, so f m =f roc ; round(·) represents the rounding operator; (5-5) Use the particle swarm optimization algorithm to search for the optimal (a * ,b * ,h * ) combination, and the optimal stochastic resonance system under this parameter combination performs intra-wave modulation feature noise enhancement on the input instantaneous frequency ω(t), and the output is obtained 9. The rotating machinery rotor rubbing diagnosis method based on vibration signal deconstruction and frequency modulation feature anti-noise enhancement according to claim 8, characterized in that: The step (5-5) includes the following sub-steps: (5-5-1) Input the estimated instantaneous frequency ω(t) into the stochastic resonance system, let S(t)+N(t)=ω(t), and initialize the number of particle swarms N p = 50, the dimension of the parameter search space is set to 3, and the search upper and lower bounds of the parameters (a, b, h) are The bounds are set to LB = [0, 0, 0] and HB = [0.05, 5000, 0.05], the particle's inertial weight is set to w = 0.6, the acceleration constant is set to γ1 = γ2 = 1.5, and the absolute value of the maximum velocity of the particle is less than 0.8; the iteration count IterPSO is initialized to 0, the maximum number of iterations MaxIterPSO is set to 500, and the precision coefficient of the iterative process is set to ε P =0.001; (5-5-2) Randomly initialize each particle and substitute the parameters of these particles into the stochastic resonance system to calculate the corresponding output; find the current global optimal output and its SNR0; (5-5-3) Start the loop, set IterPSO = IterPSO + 1, update and adjust the position and speed of each particle, and substitute the parameters of these particles into the stochastic resonance system to calculate the corresponding output; find the current global optimal output and its SNR IterPSO ; (5-5-4) Calculate the signal-to-noise ratio change ΔSNR = |SNR0-SNR IterPSO |, and update SNR0=max(SNR0,SNR IterPSO )and (5-5-5) If IterPSO>MaxIterPSO or ΔSNR<ε P , the loop is terminated and output Otherwise, return to (5-5-3); (5-5-6) End the loop and output the final optimal parameter combination (a * ,b * ,h * The output result of the optimal stochastic resonance system under ) for the input instantaneous frequency ω(t) is enhanced by intra-wave modulation characteristics and noise reduction.

10. The rotating machinery rotor rubbing diagnosis method based on vibration signal deconstruction and frequency modulation feature anti-noise enhancement according to claim 1, characterized in that: In step 6, the instantaneous frequency The Fourier spectrum of the rotor frequency f roc The rotor of the rotating machinery can be judged to have a friction fault only when the distribution characteristics of the relevant harmonic amplitudes meet the following two conditions: (1) There is a clear rotor frequency f that can be clearly distinguished from noise and other mechanical vibration components roc and its harmonics; (2) One-fold rotation frequency f roc The amplitude is the maximum.

Images