A method for enhancing random resonance of a variable frequency signal

Abstract

This invention belongs to the field of signal processing technology, specifically relating to a method for enhancing the stochastic resonance of variable frequency signals. The method includes setting the fault characteristic frequency of interest as the target frequency; calculating the target frequency curve based on the rotational frequency and the fault characteristic order; dividing the vibration signal into multiple sub-signals according to the target frequency curve, with each sub-signal's target frequency decomposed and grouped into the same node; performing envelope demodulation on each sub-signal to obtain multiple envelope sub-signals; performing multi-scale noise adjustment on each envelope sub-signal to obtain multiple reconstructed envelope sub-signals, which are then input into a normalized stochastic resonance system to obtain multiple output sub-signals, which are then combined; and performing time-frequency analysis or order analysis on the combined signal to determine whether a fault exists in the variable speed equipment. This invention can solve the problem of segmented signal processing when the frequency variation range is large, greatly enhancing the target frequency, eliminating noise influence, and thus achieving accurate detection of minor faults in rotating machinery.

Description

Technical Field

[0001] This invention belongs to the field of signal processing technology, and specifically relates to a method for enhancing random resonance of frequency conversion signals. Background Technology

[0002] Rotating machinery typically operates in harsh environments, making its critical components prone to failure. This can impact the stability, reliability, and safety of the entire mechanical system. Furthermore, signal acquisition is always affected by interference noise from vibrations of other mechanical parts, data acquisition and transmission systems, etc. This significantly reduces the signal-to-noise ratio of the measured signals, affecting the quality of fault characteristics and further decreasing the accuracy of fault detection. Moreover, useful signals reflecting fault information are very weak in the early stages of a fault and are often submerged in strong, random background noise, making them difficult to detect directly.

[0003] To address the aforementioned issues, representative methods include parameter-tuned stochastic resonance models and multi-scale noise-tuned stochastic resonance models. Traditional stochastic resonance models are based on the assumption that the signal to be detected is periodic. However, in real-world industrial scenarios, the signal to be detected may be aperiodic. For example, during the start-up or shutdown phase of rotating machinery, the characteristic frequencies of the mechanical vibration signal change with the rotational speed. In this case, the signal to be detected is a variable-frequency signal, and the target frequency changes over time, posing a challenge to the detection of weak signals based on stochastic resonance.

[0004] Theoretically, traditional stochastic resonance models can also be used for frequency conversion signal enhancement because frequency conversion signals can be considered as fixed-frequency signals for a short period of time. However, since parameter-tuned stochastic resonance models and multi-scale noise-tuned stochastic resonance models require different parameters to be optimized for different target frequencies, there is a problem of how to optimize the parameters in frequency conversion signal detection. Variable parameter stochastic resonance models adapt the scale factor of the parameter-tuned stochastic resonance model to the target frequency, requiring only optimization of the initial value of the scale factor, and can be used for frequency conversion signal enhancement detection.

[0005] Variable-parameter stochastic resonance models are effective in detecting time-varying signals at a single frequency. However, in practical engineering, mechanical vibration signals often contain multiple components, making it difficult for this method to concentrate noise energy at the target frequency, and the system output still suffers from interference from other frequency components. Therefore, the existing technology has at least the following drawbacks:

[0006] Noise energy is difficult to transfer to the target frequency in a targeted manner;

[0007] The signal-to-noise ratio of the output of a stochastic resonance system is low;

[0008] It cannot effectively detect multiple frequency conversion components in a signal. Summary of the Invention

[0009] The purpose of this invention is to provide a method for enhancing the random resonance of frequency conversion signals to solve the problems mentioned in the background art.

[0010] The present invention achieves the above objectives through the following technical solutions:

[0011] A method for enhancing stochastic resonance of frequency conversion signals includes:

[0012] S1. Set the fault characteristic frequency of interest for the variable speed equipment as the target frequency, and calculate the target frequency curve based on the rotational frequency and the fault characteristic order ratio;

[0013] S2. Based on the target frequency curve, the vibration signal of the variable speed equipment is divided into multiple sub-signals. The target frequency of each sub-signal is divided into the same node after wavelet packet decomposition.

[0014] S3. Demodulate the envelope of each sub-signal to obtain multiple envelope sub-signals;

[0015] S4. Perform multi-scale noise adjustment on each envelope sub-signal to obtain multiple reconstructed envelope sub-signals;

[0016] S5. Input each reconstructed envelope sub-signal into the normalized stochastic resonance system to obtain multiple output sub-signals;

[0017] S6. Combine the output sub-signals and perform time-frequency analysis or order ratio analysis on the combined signals to determine whether there is a fault in the variable speed equipment.

[0018] As a further optimization of the present invention, in step S1, the target fault of the variable speed equipment, the fault characteristic frequency is a multiple of the rotational frequency, and the multiple is the fault characteristic order ratio.

[0019] As a further optimization of the present invention, step S2 includes:

[0020] S201, Let the vibration signal of the equipment be... Where N is the number of signal sampling points, and the target frequency is... The minimum wavelet packet decomposition cutoff level J corresponding to the initial target frequency is determined according to the following inequality. min ,Right now:

[0021] ;

[0022] S202, Take the cutoff decomposition level as J = J min +1, re-determine the wavelet packet node position j of the target frequency according to the following inequality. f :

[0023] ;

[0024] The frequency range of the target frequency band can then be determined as follows: ;

[0025] S203. As the target frequency changes, the boundary point x(i1) of the first signal segmentation is determined according to the following inequality:

[0026] ;

[0027] The first segmented sub-signal is:

[0028] ;

[0029] The target frequencies of this sub-signal are all within the frequency band. Within the range, the second segmented sub-signal x2 starts from x(i1) with a target frequency of f. d (i1), determine the boundary point x(i2) for the second signal segmentation using the above method, until the signal x = {x(i), i = 1, 2, ..., N} is completely segmented, then x = [x1, x2, ..., x L ], where L is the number of segmented sub-signals.

[0030] As a further optimization of the present invention, step S4 includes:

[0031] Select sub-signal x1 and perform J-level decomposition on it using WPT to obtain the wavelet packet coefficient set. :

[0032] ;

[0033] Among them, C i,1 These are the wavelet packet coefficients corresponding to the i-th wavelet packet node of the J-th layer of signal x1, arranged in ascending order of frequency content. Multi-scale noise adjustment is performed on these wavelet packet coefficients using the following method:

[0034] ;

[0035] Where α, β, and γ are all adjustment parameters and are all positive real numbers, the adjusted wavelet packet coefficient set is:

[0036] ;

[0037] The adjusted wavelet packet coefficients are reconstructed into a new sub-signal using a wavelet packet decomposition and reconstruction algorithm. .

[0038] As a further optimization of the present invention, step S5 includes:

[0039] S501, adjust the sub-signal Continuous time signal Input into the normalized bistable stochastic resonance model:

[0040] ;

[0041] Obtain the system output signal y out1 ;

[0042] S502. Optimize the adjustment parameters α, β, and γ using an optimization algorithm to obtain the optimal output signal y. out1， Among them, the optimization algorithms include at least genetic algorithms and particle swarm optimization algorithms;

[0043] S503. Define a time-frequency domain signal-to-noise ratio as the optimization objective function, and its mathematical expression is as follows:

[0044] ;

[0045] in, The target frequency f d (i) In signal y out1 Amplitude in the time spectrum, A sum It is the sum of the amplitudes of all frequencies in the time spectrum at all time points. Each adjusted sub-signal is sequentially input into the stochastic resonance model to obtain multiple output sub-signals.

[0046] As a further optimization of the present invention, step S6 includes: converting the system output sub-signal y out1 The amplitude is normalized to the range [-1, 1] to obtain the normalized signal y1. The normalized output signals corresponding to all sub-signals are combined in the following way to obtain the output combined signal y:

[0047] ;

[0048] Perform time-frequency analysis or order ratio analysis on the output combined signal, observe whether there is a fault characteristic frequency curve or fault characteristic order ratio in the time-frequency diagram or order ratio spectrum, and determine whether the equipment has a corresponding fault.

[0049] The beneficial effects of this invention are as follows:

[0050] Compared with existing technologies, this invention discloses a method for enhancing the random resonance of frequency conversion signals. It designs a signal partitioning criterion, improves the multi-scale noise adjustment method, and defines a time-frequency domain signal-to-noise ratio (SNR) calculation method as the objective function for parameter optimization. This makes the detection of target frequencies more targeted, enhances the diversity and refinement of multi-scale noise adjustment modes, and realizes the calculation and parameter optimization of the SNR of frequency conversion signals. This method has at least the following advantages: it can target frequencies in a targeted manner; it has a high output signal-to-noise ratio; it can sequentially enhance and detect multi-component frequency conversion signals; and it has good fault diagnosis effects. Attached Figure Description

[0051] Figure 1 This is a flowchart of a method for enhancing random resonance of frequency conversion signals disclosed in an embodiment of the present invention;

[0052] Figure 2 (a) Time-domain waveform and (b) Rotational frequency curve of the mixed fault vibration signal of the inner and outer rings of the variable speed bearing provided in the embodiments of the present invention;

[0053] Figure 3 for Figure 2 (a) Time-frequency diagram and (b) Spectrum of the envelope signal of the vibration signal;

[0054] Figure 4 To analyze using the variable parameter stochastic resonance method Figure 2 Results of the vibration signal: (a) Time-frequency diagram of the output signal, (b) Order spectrum of the output signal;

[0055] Figure 5 To analyze using the method of this invention Figure 2 Vibration signal results (detection of outer ring fault): (a) Time-frequency diagram of output combined signal, (b) Order spectrum of output combined signal;

[0056] Figure 6 To analyze using the method of this invention Figure 2 Results of vibration signals (detection of inner ring faults): (a) Time-frequency diagram of output combined signal, (b) Order spectrum of output combined signal. Detailed Implementation

[0057] The present application will now be described in further detail with reference to the accompanying drawings. It should be noted that the following specific embodiments are only used to further illustrate the present application and should not be construed as limiting the scope of protection of the present application. Those skilled in the art can make some non-essential improvements and adjustments to the present application based on the above application content.

[0058] Example 1

[0059] like Figure 1 As shown, this embodiment proposes a method for enhancing the stochastic resonance of a frequency conversion signal, including:

[0060] S1. Setting the target fault characteristic frequency: Set the fault characteristic frequency of interest for the variable speed equipment as the target frequency, and calculate the target frequency curve based on the rotational frequency and the fault characteristic order ratio.

[0061] In this embodiment, the target fault of the variable speed equipment in step S1 is a common fault of rotating machinery. The fault characteristic frequency is a multiple of the rotation frequency, and the multiple is the fault characteristic order ratio. The multiple can be determined in advance by analyzing the fault generation mechanism.

[0062] S2. Equipment vibration signal segmentation: The vibration signal of the variable speed equipment is segmented into multiple sub-signals according to the target frequency curve. The target frequency of each sub-signal is divided into the same node after wavelet packet decomposition.

[0063] As a further preferred option, step S2 includes:

[0064] S201, Assume the equipment vibration signal is... Where N is the number of signal sampling points, and the target frequency is... The minimum wavelet packet decomposition cutoff level J corresponding to the initial target frequency is determined according to the following inequality. min ,Right now:

[0065] ;

[0066] S202. To achieve finer noise conditioning and reduce noise interference within the decomposition band (i.e., the target frequency band) where the target frequency is located, the cutoff decomposition level is set to J = J min +1, re-determine the wavelet packet node position j of the target frequency according to the following inequality. f (Starting point is 1):

[0067] ;

[0068] The frequency range of the target frequency band can then be determined as follows: ;

[0069] S203. Therefore, as the target frequency changes, the boundary point x(i1) of the first signal segmentation is determined according to the following inequality:

[0070] ;

[0071] The first segmented sub-signal is:

[0072] ;

[0073] The target frequencies of this sub-signal are all within the frequency band. Within the range, the second segmented sub-signal x2 starts from x(i1) with a target frequency of f. d (i1), determine the boundary point x(i2) for the second signal segmentation using the above method, until the signal x = {x(i), i = 1, 2, ..., N} is completely segmented, then x = [x1, x2, ..., x L ], where L is the number of segmented sub-signals.

[0074] S3, Sub-signal envelope demodulation: Perform envelope demodulation on each sub-signal to obtain multiple envelope sub-signals;

[0075] Understandably, the purpose of envelope demodulation is to demodulate the low-frequency modulation frequency in the high-frequency resonant frequency, which is the target fault characteristic frequency.

[0076] S4. Multi-scale noise conditioning of envelope sub-signals: Multi-scale noise conditioning is performed on each envelope sub-signal to obtain multiple reconstructed envelope sub-signals;

[0077] As a further preferred option, step S4 includes:

[0078] Select sub-signal x1 and perform J-level decomposition on it using WPT to obtain the wavelet packet coefficient set. :

[0079] ;

[0080] Among them, C i,1 These are the wavelet packet coefficients corresponding to the i-th wavelet packet node of the J-th layer of signal x1, arranged in ascending order of frequency content. Multi-scale noise adjustment is performed on these wavelet packet coefficients using the following method:

[0081] ;

[0082] Where α, β, and γ are all adjustment parameters and are all positive real numbers, the adjusted wavelet packet coefficient set is:

[0083] ;

[0084] The adjusted wavelet packet coefficients are reconstructed into a new sub-signal using a wavelet packet decomposition and reconstruction algorithm. .

[0085] S5. Reconstructed envelope sub-signal stochastic resonance enhancement: Each reconstructed envelope sub-signal is input into the normalized stochastic resonance system to obtain multiple output sub-signals;

[0086] As a further preferred option, step S5 includes:

[0087] S501, adjust the sub-signal Continuous time signal Input into the normalized bistable stochastic resonance model:

[0088] ;

[0089] Obtain the system output signal y out1 ;

[0090] S502. Optimize the adjustment parameters α, β, and γ using an optimization algorithm to obtain the optimal output signal y. out1， Among them, the optimization algorithms include at least genetic algorithms and particle swarm optimization algorithms;

[0091] S503. Define a time-frequency domain signal-to-noise ratio as the optimization objective function, and its mathematical expression is as follows:

[0092] ;

[0093] in, The target frequency f d (i) In signal y out1 Amplitude in the time spectrum, A sum It is the sum of the amplitudes of all frequencies in the time spectrum at all time points. Each adjusted sub-signal is sequentially input into the stochastic resonance model to obtain multiple output sub-signals.

[0094] S6. Fault diagnosis of variable speed equipment: Combine the output sub-signals and perform time-frequency analysis or order ratio analysis on the combined signals to determine whether there is a fault in the variable speed equipment.

[0095] As a further preferred embodiment, step S6 includes: outputting the system sub-signal y out1 The amplitude is normalized to the range [-1, 1] to obtain the normalized signal y1. The normalized output signals corresponding to all sub-signals are combined in the following way to obtain the output combined signal y:

[0096] ;

[0097] Perform time-frequency analysis or order ratio analysis on the output combined signal, observe whether there is a fault characteristic frequency curve or fault characteristic order ratio in the time-frequency diagram or order ratio spectrum, and determine whether the equipment has a corresponding fault.

[0098] In this embodiment, addressing the problem of noise energy being difficult to transfer to the target frequency in variable-parameter stochastic resonance models for frequency conversion signal detection, a segmented multi-scale noise-tuned stochastic resonance method is proposed, leveraging the significant advantages of multi-scale noise-tuned stochastic resonance models in weak signal enhancement. This method divides the signal according to the range of target frequency variation, performs multi-scale noise tuning segmentally, and utilizes stochastic resonance to enhance the target frequency. This invention solves the problem of segmented signal processing when the frequency variation range is large, significantly enhances the target frequency, eliminates noise influence, and thus achieves accurate detection of weak faults in rotating machinery.

[0099] To better understand the technical solution and effects of the present invention, the above embodiments will be described in detail below with reference to specific examples.

[0100] Taking the vibration signal of a mixed fault between the inner and outer rings of a variable speed bearing as an example. The bearing model is ER16K. An ICP accelerometer (model 623C01) is mounted on the bearing housing, collecting vibration signals at a sampling frequency of 200 kHz. An incremental encoder (EPC, model 775) is mounted on the shaft to measure its rotational speed. Based on the bearing structural parameters, the characteristic order ratio of the inner ring fault can be calculated to be O. n = 5.43, the characteristic order ratio of the bearing outer ring fault is 0 w = 3.57.

[0101] Reference Appendix Figure 2 , Figure 2 This is a time-domain waveform diagram and corresponding frequency curve of the mixed fault vibration signal of the inner and outer rings of the variable speed bearing provided in this embodiment of the invention. The waveform diagram shows the fault impact pulse, but also contains a lot of background noise. The bearing frequency first decreases and then increases.

[0102] Because the sampling frequency was too high, it increased the computational load and memory consumption for signal analysis. Since fault information is mainly concentrated in the low-frequency region of the signal envelope, the vibration signal was downsampled to a sampling frequency of 5 kHz. The envelope of the downsampled vibration signal was then extracted, followed by time-frequency analysis and order ratio analysis. The resulting time-frequency plot and order ratio spectrum are shown below. Figure 3 As shown. Since most of the energy of the envelope signal is concentrated in the low frequency range, Figure 3 The time-frequency plot in (a) is the result of filtering out frequency components below 10 Hz. The plot contains many frequency conversion signal components that vary with the rotation frequency, as well as some background noise, which affects the detection of the inner and outer ring fault characteristic frequencies of the target. Figure 3 (b) can be used to find the outer ring fault characteristic order O. w However, there is no inner ring fault characteristic order ratio O n Therefore, it can only detect faults in the outer ring of the bearing, but not in the inner ring.

[0103] The vibration signal was analyzed using the variable-parameter stochastic resonance method, and the results are as follows: Figure 4 As shown. These results also filter out low-frequency information below 10 Hz, allowing observation of frequency information above 10 Hz. Some frequency curves varying with rotational frequency can be faintly observed in the time-frequency plot; only the outer ring fault characteristic order O with a low amplitude exists in the order spectrum. w Therefore, this method can only detect faults on the outer ring, and the detection effect is not good.

[0104] The signal was processed using the enhancement method disclosed in this invention. The outer ring fault characteristic frequency and the inner ring fault characteristic frequency were used as target frequencies for detection, and the results are as follows: Figure 5 and Figure 6 As shown, a continuous curve of the target frequency can be observed in both time-frequency plots, and the characteristic order of the target fault is also very prominent in both order ratio spectra. Therefore, the proposed method can detect individual faults corresponding to mixed faults separately, and the detection results do not affect each other. This experimental analysis result proves the effectiveness and superiority of the method of the present invention.

[0105] In summary, the enhancement method proposed in this invention first divides the signal into multiple sub-signals based on the target frequency variation curve. The target frequency of each sub-signal is grouped into the same wavelet packet node after wavelet packet decomposition. Then, multi-scale noise adjustment is performed on the envelopes of the sub-signals. Next, the adjusted envelope sub-signals are input into a normalized stochastic resonance model, and the amplitude of the output signal is normalized and concatenated to obtain the output combined signal. Finally, time-frequency analysis or order ratio analysis is performed on the output combined signal to detect whether a target fault exists in the signal. The analysis results of the mixed fault signal of the variable speed bearing show that the method of this invention can detect individual faults separately. A continuous curve of the target fault characteristic frequency can be observed in the time-frequency graph of the output combined signal, and a clear target fault characteristic order ratio can be found in the signal order ratio spectrum. Therefore, the method of this invention can effectively detect whether a fault exists in a variable speed bearing and accurately determine what type of fault it contains. The method of this invention overcomes the problems of traditional random resonance enhancement methods for frequency conversion signals, such as difficulty in selectively transferring noise energy to the target frequency, low signal-to-noise ratio of the output random resonance system, and inability to effectively detect multiple frequency conversion components in the signal. It has the advantages of being able to selectively enhance the target frequency, having a high signal-to-noise ratio of the output signal, being able to sequentially enhance and detect multi-component frequency conversion signals, and having good fault diagnosis effect. It is of great significance for the effective detection of weak faults in rotating machinery.

[0106] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of the present invention are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. The semiconductor medium can be a solid-state drive.

[0107] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0108] In addition, the functional modules in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0109] If the aforementioned functions are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in the embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0110] The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for enhancing random resonance of frequency conversion signals, characterized in that, include: S1. Set the fault characteristic frequency of interest for the variable speed equipment as the target frequency, and calculate the target frequency curve based on the rotational frequency and the fault characteristic order ratio; S2. Based on the target frequency curve, the vibration signal of the variable speed equipment is divided into multiple sub-signals. The target frequency of each sub-signal is divided into the same node after wavelet packet decomposition. S3. Demodulate the envelope of each sub-signal to obtain multiple envelope sub-signals; S4. Perform multi-scale noise adjustment on each envelope sub-signal to obtain multiple reconstructed envelope sub-signals; S5. Input each reconstructed envelope sub-signal into the normalized stochastic resonance system to obtain multiple output sub-signals; S6. Combine the output sub-signals and perform time-frequency analysis or order ratio analysis on the combined signals to determine whether there is a fault in the variable speed equipment. Step S4 includes: The first segmented sub-signal x1 is selected and decomposed into J-levels using WPT to obtain the wavelet packet coefficient set. : ； Among them, C i,1 These are the wavelet packet coefficients corresponding to the i-th wavelet packet node of the J-th layer of signal x1, arranged in ascending order of frequency content. Multi-scale noise adjustment is performed on these wavelet packet coefficients using the following method: ； in, Let α, β, and γ be the wavelet packet node positions, and α, β, and γ be adjustment parameters, all of which are positive real numbers. The adjusted set of wavelet packet coefficients is: ； The adjusted wavelet packet coefficients are reconstructed into a new sub-signal using a wavelet packet decomposition and reconstruction algorithm. ; Step S5 includes: S501, adjust the sub-signal Continuous time signal Input into the normalized bistable stochastic resonance model: ； Obtain system output signal ; S502. Optimize the adjustment parameters α, β, and γ using an optimization algorithm to obtain the optimal output signal. Among them, the optimization algorithms include at least genetic algorithms and particle swarm optimization algorithms; S503. Define a time-frequency domain signal-to-noise ratio as the optimization objective function, and its mathematical expression is as follows: ； Where N is the number of signal sampling points, The target frequency f d (i) In signal y out1 Amplitude in the time spectrum, A sum It is the sum of the amplitudes of all frequencies in the time spectrum at all time points. Each adjusted sub-signal is sequentially input into the stochastic resonance model to obtain multiple output sub-signals.

2. The method for enhancing random resonance of frequency conversion signals according to claim 1, characterized in that: In step S1, the target fault of the variable speed equipment is a multiple of the rotational frequency, and the multiple is the fault characteristic order ratio.

3. The method for enhancing random resonance of frequency conversion signals according to claim 1, characterized in that: Step S2 includes: S201, Let the vibration signal of the equipment be... Where N is the number of signal sampling points, and the target frequency is... The minimum wavelet packet decomposition cutoff level J corresponding to the initial target frequency is determined according to the following inequality. min ,Right now: ； S202, Take the cutoff decomposition level as... Re-determine the wavelet packet node position of the target frequency based on the following inequality. : ； The frequency range of the target frequency band can then be determined as follows: ; S203. As the target frequency changes, determine the boundary point of the first signal segmentation according to the following inequality. : ； The first segmented sub-signal is: ； The target frequencies of the segmented sub-signals are all within the frequency band. Within the range, the second segmented sub-signal by Starting from the target frequency, the target frequency is... The boundary points for the second signal segmentation are determined using the above method. until the signal Until it is completely divided, then , where L is the number of segmented sub-signals.

4. The method for enhancing random resonance of frequency conversion signals according to claim 3, characterized in that: Step S6 includes: outputting the system sub-signal y out1 The amplitude is normalized to the range [-1, 1] to obtain the normalized signal y1. The normalized output signals corresponding to all sub-signals are combined in the following way to obtain the output combined signal y: ； Perform time-frequency analysis or order ratio analysis on the output combined signal, observe whether there is a fault characteristic frequency curve or fault characteristic order ratio in the time-frequency diagram or order ratio spectrum, and determine whether the equipment has a corresponding fault.

Images