Fault diagnosis method of main bearing of manufacturing equipment based on digital twin

Abstract

Disclosed is a fault diagnosis method and system for the main bearing of manufacturing equipment based on digital twins, including a physical entity module and a digital twin module. In the physical entity module, the bearing vibration signal is acquired in real time based on a data acquisition device, and the acquired vibration signal is input into a physical diagnosis model to obtain the characteristic frequency of the early fault of the bearing, and the location of the occurrence of the early fault of the bearing is determined by comparing it with the theoretical fault characteristic frequency calculated in the digital twin module. In addition, the working parameters of the bearing under real-time working conditions are acquired based on the data acquisition device, and a simulation model of the working state of the bearing under synchronous working conditions is constructed in combination with the bearing material performance parameters. The location of the fault is verified by auxiliary verification of the simulation results, and the simulation model is optimized and adjusted based on the diagnosis results of the physical diagnosis model. The digital twin is introduced to realize auxiliary verification of the diagnosis results of the physical diagnosis model, thereby realizing rapid capture and accurate positioning of the early fault of the bearing.

Description

Technical Field

[0001] The present application belongs to the field of intelligent operation, maintenance and health management of intelligent manufacturing equipment, and specifically relates to a fault diagnosis method for main bearings of manufacturing equipment based on digital twins. Background Art

[0002] With the continuous development of industrial technology, mechanical equipment tends to be large-scale and intelligent. As an indispensable key component in industrial rotating equipment, rolling bearings play an important role in rotating machinery. Due to the harsh working environment and heavy workload, bearings are inevitably damaged after long-term work. If the bearing fault is not discovered in time, it may cause a series of mechanical damage and even lead to catastrophic production accidents. On the contrary, if the bearing fault can be diagnosed in the early stage, the accident can be avoided through timely maintenance, which is of great significance and value to the health and maintenance of the equipment. However, early failure means that the wear of the bearing is not obvious, and the cyclic pulse signal generated by the bearing defect is relatively weak. In addition, the vibration signal collected in the industrial scene contains interference from background noise and environmental noise, which increases the difficulty of bearing fault feature extraction. Therefore, strong noise interference makes the early fault diagnosis of bearings a huge challenge.

[0003] The cyclic pulse signal stimulated by the bearing fault is an important carrier of bearing health. In the actual working environment, when the noise intensity in the collected vibration signal increases, the fault information will be submerged by the noise. The traditional method will weaken the energy of the available information while filtering the noise component. Therefore, when extracting early fault features under strong noise interference, the filtering effect is not ideal. Digital twin is a technology that can realize the interactive fusion of physical entities and virtual models, with the characteristics of high synchronization and real-time mapping. In the field of intelligent operation and health management (PHM), digital twin is based on the synchronous mapping and real-time interaction of physical and virtual devices and precise PHM services to form a new model of equipment health management. The invention patent "Equipment Fault Diagnosis Method, Device and System Based on Digital Twin Model" (CN110442936B) establishes a deep association between the physical system and the digital twin system, thereby realizing quantitative analysis and precise positioning of the target equipment fault.

[0004] By analyzing the existing mainstream fault diagnosis methods, although various signal processing technologies have been used for mechanical fault detection, such as wavelet transform, singular spectrum decomposition, integrated empirical mode decomposition, spectral kurtosis, etc., certain theoretical results have been achieved in the field of fault diagnosis, but the basic characteristics of the algorithm constrain them, which makes them limited in effect when processing low signal-to-noise ratio vibration signals collected under actual working conditions. The fault diagnosis method of the main bearing of the manufacturing equipment based on the digital twin proposed in the present invention establishes a fault diagnosis model based on the frequency-weighted energy operator and the power spectrum fusion, which can effectively filter out noise interference in the frequency domain signal and retain the real signal to the greatest extent, so that the early fault of the bearing can be diagnosed in a timely and effective manner. In addition, by introducing the digital twin technology, the working state of the bearing is synchronously mapped to provide auxiliary verification for the accurate identification and rapid location of the fault. Summary of the invention

[0005] In view of this, the embodiment of the present application proposes a fault diagnosis method for the main bearing of manufacturing equipment based on digital twins. The entity diagnosis model established based on the power spectrum fusion of the frequency-weighted energy operator can not only accurately diagnose various types of weak bearing faults under strong noise interference, but also has higher effectiveness and robustness. In addition, the model only relies on the real-time acquisition signal itself for diagnosis without the need for additional auxiliary means, thus solving the problem of the existing fault diagnosis model's over-reliance on signal filtering and expert experience.

[0006] Furthermore, the digital twin is introduced to realize the synchronous mapping of the diagnostic results of the physical diagnostic model, so as to quickly capture and accurately locate the early faults of the bearing.

[0007] In order to achieve the above purpose, this application adopts the following technical solutions:

[0008] A fault diagnosis method for a main bearing of manufacturing equipment based on a digital twin comprises a physical entity module and a digital twin module.

[0009] Preferably, the physical entity module comprises a physical diagnostic model, comprising the following steps:

[0010] S1, obtaining a real-time vibration signal of the bearing based on a data acquisition device, segmenting the vibration signal based on a sliding window function to obtain a sample signal to be fused, and constructing a Hankel matrix of a time domain segment based on the segmented sample signal;

[0011] S2, for the sample signal represented by each row in the Hankel matrix, the power spectrum of the corresponding signal is calculated, and the Hankel matrix in the frequency domain is constructed. The power spectrum matrix of the sample signal is fused by the mean fusion method to obtain the reconstructed power spectrum signal;

[0012] S3, demodulate the fused power spectrum to eliminate the modulation phenomenon and further amplify the periodic impact characteristics caused by the bearing fault;

[0013] S4, the demodulated power spectrum is converted to the time domain using inverse Fourier transform, and the square envelope spectrum analysis of the time signal is performed to diagnose the bearing condition.

[0014] For example, the method may further include: S5, evaluating the obtained square envelope spectrum of the vibration signal to determine the validity of the diagnosis result under the current number of samples.

[0015] For example, if the end condition is not met, the process loops from S1 to S5.

[0016] Preferably, the digital twin module includes a theoretical calculation model and a simulation model.

[0017] Preferably, the theoretical calculation model comprises the following steps:

[0018] S11, confirm the model of the monitored bearing and find the size parameters of the corresponding model bearing;

[0019] S12, according to the theoretical formula of bearing fault characteristic frequency, the theoretical values ​​of fault characteristic frequencies of different parts of the bearing are calculated.

[0020] Preferably, the simulation model comprises the following steps:

[0021] S21, acquiring working condition information of the bearing under real-time working conditions based on a data acquisition device; for example, the working condition information includes information such as the speed and load of the bearing, and working environment parameters of the bearing;

[0022] S22, finding the material performance parameters of the monitored bearing, such as density, hardness, Poisson's ratio, etc., and establishing a three-dimensional simulation model consistent with the working state of the bearing; for example, the working state of the bearing includes the working states of the bearing assembly, meshing, force drive, etc.

[0023] S23, based on the parameters obtained in S21 and S22 and the three-dimensional simulation model, a simulation model synchronously mapped with the actual working state of the bearing is constructed.

[0024] S24, obtaining the working condition information of the physical entity bearing in real time through the data acquisition device, and inputting the working condition information of the physical entity bearing obtained in real time into the simulation model to simulate the bearing damage.

[0025] Preferably, in S1, the segmentation expression of the sample signal to be fused obtained by segmenting the vibration signal f(x) using a sliding window function is:

[0026]

[0027] In the formula, S(l, iτ) is the sliding window function, l is the length of the window, τ is the sliding length of the window function, and i is the number of sliding times. is the sample signal obtained.

[0028] The Hankel matrix constructed based on the sample signal is expressed as:

[0029]

[0030] Preferably, in S2, the frequency domain Hankel matrix constructed based on the sample signal is expressed as:

[0031]

[0032] In the formula, Represents the power spectrum of the sample signal. The specific expression of the power spectrum signal reconstructed by the fusion method is:

[0033]

[0034] In the formula, Represents the reconstructed power spectrum signal.

[0035] Preferably, in S3, the specific method of demodulating the reconstructed power spectrum is a frequency-weighted energy operator, and its expression is:

[0036]

[0037] In the formula, S y (f) represents the processed power spectrum, Γ represents the frequency-weighted energy operator processing process, F(f) represents the power spectrum signal to be processed, and H(f) is the Hilbert transform of F(f).

[0038] Preferably, in S4, the specific process of converting the demodulated power spectrum signal into a time domain signal and performing power spectrum analysis is:

[0039] C p (n)=|ζ -1 {S y (f)}| (6)

[0040]

[0041] In the formula, ζ and ζ -1 Respectively represent Fourier transform and its inverse transform, C p (n) represents the time domain signal obtained by conversion, SES(f) represents the obtained square envelope spectrum, and N is the length of the data.

[0042] Preferably, in S5, the square envelope spectrum needs to be normalized before being evaluated.

[0043] Preferably, in S5, the evaluation method of the square envelope spectrum is to quantify the diagnosis result by calculating the Gini index of the square envelope spectrum.

[0044] Preferably, in S5, when the Gini index of the square envelope spectrum is greater than 0.65, it indicates that the diagnosis result meets the requirements; when the Gini index of the square envelope spectrum is less than 0.65, the diagnosis result does not meet the requirements; when the diagnosis result does not meet the requirements, the process loops from S1 to S5.

[0045] Preferably, the fault type and location of the bearing fault are determined by comparing the fault characteristic frequency obtained by the physical diagnosis model with the fault characteristic frequency obtained by the theoretical calculation model in the digital twin. In addition, the fault location and type generated by the simulation model are used to assist in verifying the diagnosis results of the physical diagnosis model.

[0046] It can be seen from the above technical solutions that, compared with the prior art, the present application provides a fault diagnosis method for the main bearing of manufacturing equipment based on digital twins, which has the following beneficial effects:

[0047] (1) The vibration signal diagnosed by the fault diagnosis method of the main bearing of manufacturing equipment based on digital twin proposed in this application is a real-time collected signal of the main bearing of the CNC machine tool. It does not rely on the historical fault data and empirical knowledge of the bearing under the same working conditions. It can identify the early fault of the bearing only by processing and analyzing the real-time collected data. It has high practical significance for monitoring the health status of the bearing and ensuring the stable operation of the bearing.

[0048] (2) The fault diagnosis method for the main bearing of manufacturing equipment based on digital twin proposed in this application has extremely strong anti-noise performance. Even if the bearing fault intensity is weak and the collected signal is contaminated by strong noise, the method can still extract the fault characteristics of the bearing and diagnose weak bearing faults, which is of great significance for timely and accurate understanding of the health status of the equipment.

[0049] (3) The fault diagnosis method for the main bearing of manufacturing equipment based on digital twin proposed in this application does not require additional parameter input and filtering processing when diagnosing early bearing faults under strong noise interference, thereby avoiding the influence of human factors and filtering effects on the diagnosis results and having a high degree of autonomy.

[0050] (4) The fault diagnosis method for the main bearing of manufacturing equipment based on digital twin proposed in this application makes full use of the random characteristics of the noise signal. By means of power spectrum fusion, it reduces noise interference while retaining weak fault information, thereby greatly improving the accuracy and effectiveness of early fault diagnosis of bearings.

[0051] (5) The fault diagnosis method for the main bearing of manufacturing equipment based on digital twin proposed in this application can not only diagnose the early faults of the inner ring and outer ring of the bearing, but also has good effect on the early diagnosis of bearing rolling element faults under strong noise interference, while most fault diagnosis methods are difficult to apply to bearing rolling element fault diagnosis.

[0052] (6) In addition to constructing a physical diagnosis model based on the actual collected vibration signal, the fault diagnosis method for the main bearing of manufacturing equipment based on digital twin proposed in this application also establishes a digital twin model of the main bearing of intelligent manufacturing equipment. By simulating the real working conditions, the simulation results of bearing damage are obtained, which serves as an auxiliary verification of the physical diagnosis model, thereby realizing the rapid capture and accurate positioning of early bearing faults. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0054] Figure 1 A flowchart of a fault diagnosis method for a main bearing of a manufacturing equipment based on a digital twin provided in an embodiment of the present application;

[0055] Figure 2(a) is a time domain diagram of the bearing outer ring fault signal;

[0056] Figure 2(b) is the Hilbert spectrum of the bearing outer race fault signal;

[0057] Figure 2(c) is the time domain diagram of the bearing outer ring fault signal after adding Gaussian noise;

[0058] Figure 2(d) is the Hilbert spectrum of the bearing outer race fault signal after adding Gaussian noise;

[0059] Figure 3 The diagnosis result of the bearing outer ring fault signal after adding noise by the method proposed in the embodiment of the present application;

[0060] Figure 4(a) is a time domain diagram of the bearing inner ring fault signal;

[0061] Figure 4(b) is the Hilbert spectrum of the bearing inner race fault signal;

[0062] Figure 4(c) is the time domain diagram of the bearing inner race fault signal after adding Gaussian noise;

[0063] Figure 4(d) is the Hilbert spectrum of the bearing inner race fault signal after adding Gaussian noise;

[0064] Figure 5The diagnosis result of the bearing inner ring fault signal after adding noise by the method proposed in the embodiment of the present application;

[0065] Figure 6(a) is a time domain diagram of the bearing rolling element fault signal;

[0066] FIG6( b ) is a Hilbert spectrum diagram of a bearing rolling element fault signal;

[0067] Figure 6(c) is the time domain diagram of the bearing rolling element fault signal after adding Gaussian noise;

[0068] Figure 6(d) is the Hilbert spectrum of the bearing rolling element fault signal after adding Gaussian noise;

[0069] Figure 7 The diagnosis result of the bearing rolling element fault signal after adding noise by the method proposed in the embodiment of the present application;

[0070] Figure 8 A block diagram of a fault diagnosis system for a main bearing of manufacturing equipment based on a digital twin provided in an embodiment of the present application. DETAILED DESCRIPTION

[0071] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0072] Example 1: Bearing outer ring failure.

[0073] See attached Figure 1 As shown, the embodiment of the present application discloses a fault diagnosis method for a main bearing of a manufacturing equipment based on a digital twin, comprising:

[0074] S1, based on the data acquisition device, obtain the real-time vibration signal of the bearing, as shown in Figure 2(a), which is the fault signal of the outer ring of the bearing. Since the fault impact is more obvious, the characteristic frequency of the outer ring fault of the bearing can be detected from the envelope spectrum shown in Figure 2(b). In order to verify the outstanding advantages of this application, Gaussian white noise interference is added to the collected signal. The time domain and frequency domain of the signal after adding noise are shown in Figure 2(c) and Figure 2(d) respectively. It can be seen from the figure that the original signal has been completely covered by the noise, and the characteristic frequency of the outer ring fault of the bearing cannot be observed in the envelope spectrum, which shows that the difficulty of fault diagnosis is greatly increased.

[0075] In this example, the vibration signal is segmented based on the sliding window function to obtain the sample signal to be fused, and the Hankel matrix of the time domain segment is constructed based on the segmented sample signal, so as to obtain different manifestations of noise in the signal. The segmentation process and the constructed Hankel matrix are shown in equations (1) and (2):

[0076]

[0077]

[0078] In the formula, S(l, iτ) is the sliding window function, l is the length of the window, τ is the sliding length of the window function, and i is the number of sliding times. is the sample signal obtained.

[0079] S2, for each row of the sample signal in the Hankel matrix, the power spectrum of the corresponding signal is calculated, and the frequency domain Hankel matrix is ​​constructed. The power spectrum matrix of the sample signal is fused by the mean fusion method to obtain the reconstructed power spectrum signal. The frequency domain Hankel matrix and the reconstructed power spectrum signal are shown in equations (3) and (4):

[0080]

[0081]

[0082] In the formula, represents the power spectrum of the sample signal, Represents the reconstructed power spectrum signal.

[0083] S3, the frequency-weighted energy operator is used to demodulate the fused power spectrum, eliminating the modulation phenomenon while further amplifying the periodic impact characteristics caused by the bearing fault. The frequency-weighted energy operator expression is:

[0084]

[0085] In the formula, S y (f) represents the processed power spectrum, Γ represents the frequency-weighted energy operator processing process, F(f) represents the power spectrum signal to be processed, and H(f) is the Hilbert transform of F(f).

[0086] S4, the demodulated power spectrum is converted to the time domain using inverse Fourier transform, and the square envelope spectrum analysis of the time signal is performed to diagnose the bearing condition. The specific process is shown in equations (6) and (7) respectively:

[0087] C p (n)=|ζ -1 {S y (f)}| (6)

[0088]

[0089] In the formula, ζ and ζ -1 They represent Fourier transform and its inverse transform, C p (n) represents the time domain signal obtained by conversion, SES(f) represents the obtained square envelope spectrum, and N is the length of the data.

[0090] S5, normalize the square envelope spectrum of the obtained vibration signal, and then calculate its Gini index to determine the effectiveness of the diagnosis result under the current sample size.

[0091] The Gini index can be expressed as:

[0092]

[0093] In the formula, SE[x] represents the data to be evaluated, ||SE[x]|| L1 represents the L1 norm of SE[x], SER[x] represents the arrangement of SE[x] from the minimum value to the maximum value, and N is the length of the data.

[0094] For example, when the Gini index of the square envelope spectrum is greater than 0.65, it indicates that the diagnosis result meets the requirements.

[0095] When the Gini index of the square envelope spectrum is less than 0.65, it indicates that the diagnosis result does not meet the requirement. When the diagnosis result does not meet the requirement, the process loops from S1 to S5.

[0096] Figure 8 A block diagram of a fault diagnosis system for a main bearing of a manufacturing equipment based on a digital twin provided in an embodiment of the present application. Figure 1 and Figure 8 The embodiment of the present application provides a fault diagnosis system for a main bearing of a manufacturing equipment based on a digital twin. The system includes a physical entity module and a digital twin module. The physical entity module includes a physical diagnosis model. The digital twin module includes a theoretical calculation model and a simulation model.

[0097] For example, the physical entity module may be used to perform the above steps S1 to S5.

[0098] In an embodiment of the present application, the digital twin module may perform the following steps.

[0099] For example, in the digital twin module, the theoretical calculation model can be used to perform steps S11 and S12.

[0100] In S11, the model of the monitored bearing is confirmed and the size parameters of the corresponding model bearing are found.

[0101] In S12, theoretical values ​​of characteristic frequencies of faults at different parts of the bearing are calculated according to a theoretical formula of characteristic frequencies of bearing faults.

[0102] For example, in the digital twin module, the simulation model can be used to perform steps S21 and S24.

[0103] In S21, the operating condition information such as the bearing speed and load and the working environment parameters under real-time operating conditions are obtained based on the data acquisition device.

[0104] In S22, the material performance parameters such as density, hardness, Poisson's ratio, etc. of the monitored bearing are found, and a three-dimensional simulation model consistent with the working state of the bearing is established.

[0105] In S23, a simulation model synchronously mapped with the actual working state of the bearing is constructed based on the parameters obtained in S21 and S22 and the three-dimensional simulation model.

[0106] In S24, relevant data is acquired in real time by a data acquisition device to drive the simulation model in the digital twin to run synchronously with the physical entity bearing, thereby simulating bearing damage.

[0107] In the embodiment of the present application, the fault characteristic frequency obtained by the physical diagnosis model is compared with the fault characteristic frequency obtained by the theoretical calculation model in the digital twin to determine the type and location of the bearing fault. In addition, the fault location and type generated by the simulation model are used to assist in verifying the diagnosis results of the physical diagnosis model.

[0108] In this example, after the algorithm is cycled 7 times, a satisfactory diagnostic result is obtained. Figure 3 As shown, the characteristic frequency of the bearing outer ring fault and its harmonics can be clearly identified from the diagnosis results and are consistent with the theoretically calculated characteristic frequency of the outer ring fault, indicating that the fault diagnosis method for the main bearing of the manufacturing equipment based on digital twin proposed in the embodiment of the present application is significantly effective.

[0109] Example 2: Bearing inner ring failure.

[0110] See attached Figure 1 As shown, the embodiment of the present application discloses a fault diagnosis method for a main bearing of a manufacturing equipment based on a digital twin, comprising:

[0111] S1, based on the data acquisition device, obtain the real-time vibration signal of the bearing, as shown in Figure 4(a), which is the fault signal of the inner ring of the bearing. Since the fault impact is more obvious, the characteristic frequency of the inner ring fault of the bearing can be detected from the envelope spectrum shown in Figure 4(b). In order to verify the outstanding advantages of this application, Gaussian white noise interference is added to the collected signal. The time domain and frequency domain of the signal after adding noise are shown in Figure 4(c) and Figure 4(d), respectively. It can be seen from the figure that the original signal has been completely covered by the noise, and the characteristic frequency of the inner ring fault of the bearing cannot be observed in the envelope spectrum, which shows that the difficulty of fault diagnosis is greatly increased.

[0112] In this example, the vibration signal is segmented based on the sliding window function to obtain the sample signal to be fused, and the Hankel matrix of the time domain segment is constructed based on the segmented sample signal, so as to obtain different manifestations of noise in the signal. The segmentation process and the constructed Hankel matrix are shown in equations (1) and (2):

[0113]

[0114]

[0115] In the formula, S(l, iτ) is the sliding window function, l is the length of the window, τ is the sliding length of the window function, and i is the number of sliding times. is the sample signal obtained.

[0116] S2, for each row of the sample signal in the Hankel matrix, the power spectrum of the corresponding signal is calculated, and the frequency domain Hankel matrix is ​​constructed. The power spectrum matrix of the sample signal is fused by the mean fusion method to obtain the reconstructed power spectrum signal. The frequency domain Hankel matrix and the reconstructed power spectrum signal are shown in equations (3) and (4):

[0117]

[0118]

[0119] In the formula, represents the power spectrum of the sample signal, Represents the reconstructed power spectrum signal.

[0120] S3, the frequency-weighted energy operator is used to demodulate the fused power spectrum, eliminating the modulation phenomenon while further amplifying the periodic impact characteristics caused by the bearing fault. The frequency-weighted energy operator expression is:

[0121]

[0122] In the formula, S y(f) represents the processed power spectrum, Γ represents the frequency-weighted energy operator processing process, F(f) represents the power spectrum signal to be processed, and H(f) is the Hilbert transform of F(f).

[0123] S4, the demodulated power spectrum is converted to the time domain using inverse Fourier transform, and the square envelope spectrum analysis of the time signal is performed to diagnose the bearing condition. The specific process is shown in equations (6) and (7) respectively:

[0124] C p (n)=|ζ -1 {S y (f)}| (6)

[0125]

[0126] In the formula, ζ and ζ -1 They represent Fourier transform and its inverse transform, C p (n) represents the time domain signal obtained by conversion, SES(f) represents the obtained square envelope spectrum, and N is the length of the data.

[0127] S5, normalize the square envelope spectrum of the obtained vibration signal, and then calculate its Gini index to determine the validity of the diagnosis result under the current number of samples. If the end condition is not met, loop S1 to S5.

[0128] In this example, after the algorithm is cycled 7 times, a satisfactory diagnostic result is obtained. Figure 5 As shown, the characteristic frequency of the bearing inner ring fault and its harmonics can be clearly identified from the diagnosis results and are consistent with the theoretically calculated characteristic frequency of the inner ring fault, indicating that the fault diagnosis method for the main bearing of the manufacturing equipment based on digital twin proposed in the embodiment of the present application is significantly effective.

[0129] Example 3: Bearing rolling element failure.

[0130] See attached Figure 1 As shown, the embodiment of the present application discloses a fault diagnosis method for a main bearing of a manufacturing equipment based on a digital twin, comprising:

[0131] S1, based on the data acquisition device, obtain the real-time vibration signal of the bearing, as shown in Figure 6(a), which is the bearing rolling element fault signal. Since the fault impact is more obvious, the characteristic frequency of the bearing rolling element fault can be detected from the envelope spectrum shown in Figure 6(b). In order to verify the outstanding advantages of this application, Gaussian white noise interference is added to the collected signal. The time domain and frequency domain of the signal after adding noise are shown in Figure 6(c) and Figure 6(d), respectively. It can be seen from the figure that the original signal has been completely covered by the noise, and the characteristic frequency of the bearing rolling element fault cannot be observed in the envelope spectrum, which shows that the difficulty of fault diagnosis is greatly increased.

[0132] In this example, the vibration signal is segmented based on the sliding window function to obtain the sample signal to be fused, and the Hankel matrix of the time domain segment is constructed based on the segmented sample signal, so as to obtain different manifestations of noise in the signal. The segmentation process and the constructed Hankel matrix are shown in equations (1) and (2):

[0133]

[0134]

[0135] In the formula, S(l, iτ) is the sliding window function, l is the length of the window, τ is the sliding length of the window function, and i is the number of sliding times. is the sample signal obtained.

[0136] S2, for each row of the sample signal in the Hankel matrix, the power spectrum of the corresponding signal is calculated, and the frequency domain Hankel matrix is ​​constructed. The power spectrum matrix of the sample signal is fused by the mean fusion method to obtain the reconstructed power spectrum signal. The frequency domain Hankel matrix and the reconstructed power spectrum signal are shown in equations (3) and (4):

[0137]

[0138]

[0139] In the formula, represents the power spectrum of the sample signal, Represents the reconstructed power spectrum signal.

[0140] S3, the frequency-weighted energy operator is used to demodulate the fused power spectrum, eliminating the modulation phenomenon while further amplifying the periodic impact characteristics caused by the bearing fault. The frequency-weighted energy operator expression is:

[0141]

[0142] In the formula, S y (f) represents the processed power spectrum, Γ represents the frequency-weighted energy operator processing process, F(f) represents the power spectrum signal to be processed, and H(f) is the Hilbert transform of F(f).

[0143] S4, the demodulated power spectrum is converted to the time domain using inverse Fourier transform, and the square envelope spectrum analysis of the time signal is performed to diagnose the bearing condition. The specific process is shown in equations (6) and (7) respectively:

[0144] C p (n)=|ζ -1 {S y (f)}| (6)

[0145]

[0146] In the formula, ζ and ζ -1 They represent Fourier transform and its inverse transform, C p (n) represents the time domain signal obtained by conversion, SES(f) represents the obtained square envelope spectrum, and N is the length of the data.

[0147] S5, normalize the square envelope spectrum of the obtained vibration signal, and then calculate its Gini index to determine the validity of the diagnosis result under the current number of samples. If the end condition is not met, loop S1 to S5.

[0148] In this example, after the algorithm is cycled 4 times, a satisfactory diagnostic result is obtained. Figure 7 As shown, the characteristic frequency of the bearing rolling element fault and its harmonics can be clearly identified from the diagnosis results and are consistent with the theoretically calculated characteristic frequency of the bearing rolling element fault, indicating that the fault diagnosis method for the main bearing of the manufacturing equipment based on the digital twin proposed in the embodiment of the present application is significantly effective.

[0149] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0150] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A fault diagnosis method for main bearing of manufacturing equipment based on digital twin, It is characterized in that The following steps are involved: S1, obtaining a real-time vibration signal of the bearing based on a data acquisition device, segmenting the vibration signal based on a sliding window function to obtain a sample signal to be fused, and constructing a Hankel matrix of a time domain segment based on the segmented sample signal; S2, for the sample signal represented by each row in the Hankel matrix, the power spectrum of the corresponding signal is calculated, and the Hankel matrix in the frequency domain is constructed. The Hankel matrix in the frequency domain is fused by the average fusion method to obtain the reconstructed power spectrum signal; S3, demodulating the reconstructed power spectrum signal to eliminate the modulation phenomenon and amplify the periodic impact characteristics caused by the bearing fault; S4, using inverse Fourier transform to convert the demodulated power spectrum signal into the time domain to form a time domain signal, and performing square envelope spectrum analysis on the time domain signal to diagnose the bearing condition.

2. The fault diagnosis method for the main bearing of manufacturing equipment based on digital twin according to claim 1, It is characterized in that The method further comprises: S5, evaluating the obtained square envelope spectrum of the time domain signal to determine the effectiveness of the diagnosis result of the fused sample signal under the current sample quantity.

3. The fault diagnosis method for the main bearing of manufacturing equipment based on digital twin according to claim 1 or 2, It is characterized in that The digital twin includes a theoretical calculation model and a simulation model, and the method further includes the following steps: S11, using the theoretical calculation model, confirming the model of the monitored bearing and finding the size parameters of the corresponding model bearing; S12, using the theoretical calculation model, according to the theoretical formula of bearing fault characteristic frequency, calculate the theoretical values ​​of fault characteristic frequencies of different parts of the bearing.

4. The fault diagnosis method for main bearing of manufacturing equipment based on digital twin according to claim 3, It is characterized in that The following steps are also included: S21, acquiring working condition information of the bearing under real-time working conditions based on a data acquisition device; S22, finding the material performance parameters of the monitored bearing and establishing a three-dimensional simulation model consistent with the working state of the bearing; S23, building a simulation model that is synchronously mapped with the actual working state of the bearing based on the parameters obtained in S21 and S22 and the three-dimensional simulation model; S24, obtaining the working condition information of the physical entity bearing in real time through the data acquisition device, and inputting the working condition information of the physical entity bearing obtained in real time into the simulation model to simulate the bearing damage.

5. The fault diagnosis method for main bearing of manufacturing equipment based on digital twin according to claim 1 or 2, It is characterized in that In S1, the specific expression of obtaining the sample signal to be fused by segmenting the vibration signal based on the sliding window function is: Where f(x) is the vibration signal, S(l, iτ) is the sliding window function, l is the length of the window, τ is the sliding length of the window function, and i is the number of sliding times. is the sample signal obtained.

6. The fault diagnosis method for main bearing of manufacturing equipment based on digital twin according to claim 5, It is characterized in that In S1, the Hankel matrix constructed based on the sample signal is expressed as:

7. The fault diagnosis method for main bearing of manufacturing equipment based on digital twin according to claim 1 or 2, It is characterized in that In S2, the constructed Hankel matrix in the frequency domain is expressed as: In the formula, Represents the power spectrum of the sample signal.

8. The fault diagnosis method for main bearing of manufacturing equipment based on digital twin according to claim 7, It is characterized in that The specific expression of the reconstructed power spectrum signal is: In the formula, Represents the reconstructed power spectrum signal.

9. The fault diagnosis method for main bearing of manufacturing equipment based on digital twin according to claim 1 or 2, It is characterized in that In S3, the demodulating the reconstructed power spectrum signal includes using a frequency weighted energy operator, which is expressed as: In the formula, S y (f) represents the processed power spectrum, Γ represents the frequency-weighted energy operator processing process, F(f) represents the power spectrum signal to be processed, and H(f) is the Hilbert transform of F(f).

10. The fault diagnosis method for main bearing of manufacturing equipment based on digital twin according to claim 9, It is characterized in that In S4, the specific process of performing square envelope spectrum analysis on the time domain signal is as follows: C p (n)=|ζ -1 {S y (f)}| (6) In the formula, ζ and ζ -1 Respectively represent Fourier transform and its inverse transform, C p (n) represents the time domain signal obtained by conversion, SES(f) represents the obtained square envelope spectrum, and N is the length of the data.

11. The fault diagnosis method for main bearing of manufacturing equipment based on digital twin according to claim 2, It is characterized in that In S5, evaluating the obtained square envelope spectrum of the vibration signal includes: quantifying the diagnosis result by calculating the Gini index of the square envelope spectrum.

12. The fault diagnosis method for main bearing of manufacturing equipment based on digital twin according to claim 11, It is characterized in that When the Gini index of the square envelope spectrum is greater than 0.65, it means that the diagnosis result meets the requirements; When the Gini index of the square envelope spectrum is less than 0.65, it indicates that the diagnosis result does not meet the requirement. When the diagnosis result does not meet the requirement, the process loops from S1 to S5.

Images