A machine tool spindle bearing fault detection method based on pull-type weighted modal decomposition

Abstract

A machine tool spindle bearing fault detection method based on pull type weighted modal decomposition, comprising the following processes, S1: collecting vibration signal data; S2: periodically intercepting the collected vibration signal to improve the analysis accuracy of the Laplace transform; S3: calculating the periodic pulse integrated features in each frequency band based on the periodic harmonic product enhancement principle; S4: weighting and integrating the narrowband filtered signals after the Laplace transform according to the integrated features in step S3; S5: performing envelope spectrum analysis on the weighted integrated signals in step S4 to qualitatively judge the bearing fault; S6: using the spindle bearing damage evaluation mechanism established in step S5 to statistically analyze the influence of different spindle bearing damage degrees on high-speed machine tool machining under dynamic working conditions. The present application improves the detection accuracy of weak faults, is beneficial to quantitative evaluation of bearing damage size, solves the problem of workpiece vibration marks caused by spindle bearing disturbance interference, breaks through the difficulty of online quantitative diagnosis, and provides technical support for safe operation of the spindle bearing.

Description

Technical Field

[0001] This invention relates to the field of abnormal detection technology in machine tool processing, and in particular to a method for detecting machine tool spindle bearing faults based on Lagrange weighted mode decomposition. Background Technology

[0002] High-speed machine tools play a crucial role in modern manufacturing. Spindle bearings are a core component of these machines, and their reliability and stability directly impact machining quality and production efficiency. However, in actual production, under complex thermal coupling and rapid acceleration / deceleration conditions, spindle bearings inevitably experience wear, pitting, and cracking. If a spindle bearing malfunctions and the monitoring system fails to alarm, it can lead to anything from increased workpiece scrap rates to machine tool damage or even serious accidents. Research indicates that equipping high-speed machine tools with machining process monitoring systems can effectively improve machining safety and machine tool utilization, significantly enhancing enterprise efficiency and attracting widespread attention.

[0003] In recent years, bearing fault diagnosis application software has become a research hotspot in the industry and has achieved some good results. Among them, the method of monitoring bearings based on vibration signals has become a hotspot in this field. Specifically, signal processing technology and feature recognition algorithms can be used to evaluate the bearing operating status relatively effectively, mainly including the following technical solutions: (1) Empirical Mode Decomposition (EMD) method is used to monitor the condition of high-speed rail bearings; (2) Variational Mode Decomposition (VMD) and energy spectrum entropy are used to detect anomalies in wind turbine bearings; (3) A novel Singular Value Decomposition (SVD) method is constructed for fault diagnosis of rail transit bearings. Although the above-mentioned technical solutions have achieved bearing fault detection to a certain extent, due to the limitations of the technical solutions, in practical applications, due to the complex thermo-mechanical coupling and rapid acceleration and deceleration disturbances during high-speed machining, it is difficult to guarantee the accuracy and reliability of the results by directly analyzing vibration signals to evaluate the condition of the spindle bearing. Therefore, it is particularly necessary to extract useful information reflecting the operating condition of the spindle bearing from multiple disturbance states, so as to effectively realize the detection of spindle bearing faults. Summary of the Invention

[0004] To overcome the shortcomings of existing technologies, which lack an effective solution for detecting machine tool spindle bearing faults and thus cannot effectively guarantee the safe and reliable operation of spindle bearings, this invention provides a machine tool spindle bearing fault detection method based on Lagrange weighted mode decomposition. This method, through the combined action of related steps, takes a time-frequency domain dual-dimensional perspective, deeply mines the periodic transient pulse characteristics based on the spindle bearing fault response mechanism, enhances weak fault characteristics through weighted optimization, improves the detection accuracy of early bearing faults, and finally verifies its effectiveness through actual data collection. When practically applied to spindle bearing fault detection, it exhibits advantages such as high detection accuracy and strong generalization ability, providing favorable technical support for the safe and stable operation of spindle bearings.

[0005] The technical solution adopted by this invention to solve its technical problem is:

[0006] A method for fault detection of machine tool spindle bearings based on Lagrange weighted mode decomposition, characterized by the following steps: S1: Acquiring vibration signal data; specifically, using a vibration sensor installed near the high-speed machine tool spindle to collect vibration signals during the machining process in real time, and the collected vibration signals are denoted as... Where N represents the number of sample points for data acquisition; S2: The acquired vibration signal x is truncated over an integer period to improve the accuracy of Laplace transform analysis. The signal is truncated over an integer period according to the signal sampling frequency and the spindle rotation speed, with data truncated once every 10,000 revolutions. The truncated signal is denoted as... S3: Perform Laplace Fourier transform analysis on the intercepted signal, evaluate the periodic pulse characteristics in each frequency band based on the time-frequency dual-dimensional characterization after the Laplace transform, and calculate the integrated characteristics of periodic pulses in each frequency band based on the principle of periodic harmonic product enhancement; S4: Perform weighted integration on the narrowband filtered signal after the Laplace transform based on the integrated characteristics in step S3 to enhance the weak damage characteristics caused by early bearing failure; S5: Perform envelope spectrum analysis on the weighted integrated signal in step S4, and then use it in the spindle bearing application software system to qualitatively identify bearing failures, and evaluate the spindle bearing damage state based on the bearing failure characteristic frequency, and determine the spindle operating state based on the evaluation and detection results; S6: Utilize the spindle bearing damage scale evaluation mechanism established in step S5 to statistically analyze the impact of different spindle bearing damage degrees on high-speed machine tool processing under dynamic working conditions, specifically including the impact on machining accuracy and tool life.

[0007] Furthermore, in step S1, the sampling frequency is set to 50kHz, the machine tool spindle speed is 25314RPM, and the duration of each data acquisition is 2s.

[0008] Furthermore, in step S3, the formula is used... The intercepted signal was analyzed using a Laplace Fourier transform, where... , Represents the Laplace transform coefficients. This indicates that k and q are coprime numbers. These are the basis functions of the Laplace transform;

[0009] Furthermore, in step S4, the formula is used... The narrowband filtered signal after Laplace transform is weighted and integrated, where... This represents the weighting coefficient, q=50.

[0010] Furthermore, in step S5, the formula is used... Envelope spectrum analysis was performed on the weighted integrated signal, where Indicates Fourier transform, This represents Hilbert envelope analysis.

[0011] Furthermore, in step S6, it is necessary to follow the formula. In step S5, the sparse impulse characteristic data of the weighted integrated signal are calculated, where, This indicates the averaging operator operation. This represents the data obtained in the nth experiment.

[0012] Compared with the prior art, the beneficial effects of the present invention are: (1) The present invention performs time-frequency dual-dimensional characterization of data based on Laplace transform, which solves the problem that the single dimension of the prior art is difficult to fully characterize the data features, gets rid of the limitations of traditional technology in single-dimensional data processing, and improves the detection accuracy of early weak faults; (2) The integrated weighted integration algorithm constructed by the present invention has the advantages of high computational efficiency, clear characterization mechanism and strong anti-interference ability. The designed weak feature sharpening method is conducive to quantitatively assessing the bearing damage scale, and provides an effective tool for accurate diagnosis of spindle bearing faults; (3) The spindle bearing damage scale assessment mechanism designed by the present invention is conducive to enhancing the stability of high-speed machining and ensuring the surface quality of the workpiece, solving the problem of workpiece vibration marks caused by spindle bearing disturbance, breaking through the problem of complex dynamic interference causing the inability to make online quantitative diagnosis, and providing favorable technical support for the safe and stable operation of spindle bearings. Attached Figure Description

[0013] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0014] Figure 1 This is a schematic diagram of the installation position of the vibration sensor of the present invention.

[0015] Figure 2 This is a schematic diagram of a bearing outer ring failure according to the present invention.

[0016] Figure 3This is a schematic flowchart of the workflow of a machine tool spindle bearing fault detection method based on Lagrange weighted mode decomposition according to the present invention.

[0017] Figure 4 This is a schematic diagram of the original signal and the original signal envelope spectrum of the present invention.

[0018] Figure 5 (a) is a schematic diagram of the spectral kurtosis of the present invention.

[0019] Figure 5 (b) is a schematic diagram of the spectral coherence of the present invention.

[0020] Figure 5 (c) is a schematic diagram of the narrowband filtered signal of the present invention.

[0021] Figure 5 (d) is a schematic diagram of the envelope spectrum of the filtered signal of the present invention.

[0022] Figure 6 (a) is a two-dimensional schematic diagram of the time-frequency dual-dimensional representation of the Laplace transform of the present invention.

[0023] Figure 6 (b) is a three-dimensional schematic diagram of the time-frequency dual-dimensional Laplace transform representation of the present invention.

[0024] Figure 7 This invention integrates a schematic diagram of weight distribution and a schematic diagram of the Laplace weighted mode decomposition envelope spectrum.

[0025] Figure 8 This is a schematic diagram illustrating the assessment of the damage level of the spindle bearing in this invention. Detailed Implementation

[0026] This invention specifically uses high-speed milling as an example to verify the effectiveness of the invention. The high-speed machining tool structure involved is as follows: Figure 1 As shown; to fully evaluate the performance of the present invention, a specific diagnosis of the outer ring fault of the spindle bearing was performed. The bearing model is SKF6306, as shown. Figure 2 As shown, the characteristic frequency of the outer ring failure of the bearing is 1288.5 Hz.

[0027] Figure 3 As shown, a machine tool spindle bearing fault detection method based on Lagrange weighted mode decomposition includes the following steps: Step 1: Figure 1 As shown, a vibration sensor installed near the spindle of a high-speed machine tool was used to collect vibration signals in real time. The sampling frequency was set to 50kHz, the spindle speed was 25314RPM, and each data collection session lasted 2 seconds. The collected data was recorded as follows: Where N represents the number of data acquisition points, this step is used to obtain monitoring information that reflects the health status of the high-speed machine tool. Step 2: Perform integer-cycle truncation on the acquired vibration signal x to improve the accuracy of Laplace transform analysis; specifically, the acquired vibration signal is as follows: Figure 4 As shown in (a), it is difficult to identify any valid fault information from the figure, such as Figure 4 As shown in (b), envelope spectrum analysis was performed on the original signal. The red dashed line in the figure represents the characteristic frequency of the bearing outer ring fault and its harmonics. It can be seen from the figure that there is no obvious spectral line corresponding to the bearing outer ring fault. In contrast, this invention uses spectral kurtosis and spectral coherence to process the acquired vibration signal, such as... Figure 5 As shown, from Figure 5 As can be seen from (b)-(c), under strong interference and multi-modulation conditions, traditional methods struggle to identify subtle fault characteristics. Furthermore, to facilitate rapid analysis, the collected data is truncated in whole cycles according to the signal sampling frequency and spindle rotation speed. Data is truncated once every 10,000 spindle revolutions (calculated to extract 20,000 data points each time, with a time interval of 0.4 seconds). The truncated signal is denoted as... M=20000 represents the number of data sample points (length) to be intercepted. This step is used to intercept the collected data by an integer rotation period, which improves data acquisition efficiency on the one hand and subsequent analysis accuracy on the other.

[0028] Figure 3 As shown, step three: Figure 6 As shown, the data intercepted during the entire rotation period are analyzed using Ramanujan Fourier Transform (RFT) according to formula (1). Then, the periodic pulse characteristics in each frequency band are evaluated based on the time-frequency dual-dimensional characterization after the Ramanujan transform (specifically from...). Figure 6 As can be seen from the transformed time-frequency dual-dimensional characterization, the periodic pulse characteristics in each frequency band are evaluated, and the periodic pulse integration characteristics in each frequency band are calculated based on the periodic harmonic product enhancement principle. In this invention, the frequency band is divided into 50 periods.

[0029] (1)

[0030] In formula (1), , Represents the Laplace transform coefficients. This indicates that k and q are coprime numbers. The Laplace transform basis function is used in this step to enhance the fault pulse characteristics through RFT. Step 4: According to formula (2), the integrated features in step 3 are used to weight and integrate the narrowband filtered signal after the Laplace transform to enhance the weak damage characteristics caused by early bearing faults; specifically, the weight distribution for each cycle is as follows: Figure 7 As shown in (a), it can be seen from the figure that the Laplace transform can adaptively evaluate the periodic pulse characteristics of different frequency bands, and sharpen the weak fault characteristics based on the principle of periodic pulse integration enhancement.

[0031] (2)

[0032] In formula (2), The weighting coefficient is represented by q=50 in this invention. This step involves integrating weighted sharpening of weak fault features.

[0033] Figure 3 As shown, step five: qualitatively determine the bearing fault, and evaluate the spindle bearing damage state based on the bearing fault characteristic frequency, and determine the spindle operating state based on the evaluation and detection results. Specifically, according to formula (3), the weighted integrated signal in step four is subjected to envelope spectrum analysis, such as... Figure 7 As shown in (b), the characteristic frequency of the bearing outer ring failure and its harmonics can be clearly seen from the figure, which indicates that the bearing outer ring is damaged. This is exactly consistent with the actual bearing failure, which shows the effectiveness of the present invention.

[0034] (3)

[0035] in, Indicates Fourier transform, This refers to Hilbert envelope analysis, a step that qualitatively detects bearing faults through envelope demodulation analysis.

[0036] Step Six: Utilizing the spindle bearing damage assessment mechanism established in Step Five, the impact of different bearing damage levels on high-speed machine tool machining under dynamic operating conditions is statistically analyzed, including machining accuracy and tool life. The analysis results show that, compared to existing modal decomposition methods, this invention can not only effectively sharpen early weak characteristics of spindle bearings and improve bearing fault detection accuracy, but also quantitatively perceive the spindle health status, providing an effective solution for ensuring the safety and stability of high-speed machine tools and improving machine tool utilization. Specifically, such as... Figure 8 As shown, the sparse pulse characteristics of the weighted integrated signal in step five are calculated according to formula (4), and the bearing damage degree under 533 experimental tests is statistically analyzed. It can be seen from the figure that the constructed damage scale assessment index can effectively sense the health status of the spindle bearing. When the bearing has slight damage, it can detect and alarm in time to improve machining accuracy and tool life. The actual analysis results show that, compared with the existing methods, the method of the present invention can not only effectively detect the early weak characteristics of the spindle bearing, but also the constructed sparse pulse characteristics can quantitatively sense the health status of the bearing. This provides an effective solution for ensuring the healthy and stable operation of the high-speed machining process.

[0037] (4)

[0038] in, This indicates the averaging operator operation. This represents the data obtained in the nth experiment.

[0039] Through all the above technical solutions, this invention, based on the periodicity principle of fault characteristics and using Laplace transform to perform time-frequency dual-dimensional characterization of data, solves the problem that existing technologies cannot comprehensively characterize data features using a single dimension. It overcomes the limitations of traditional single-dimensional data processing and improves the detection accuracy of early, subtle faults. The integrated weighted algorithm constructed in this invention has the advantages of high computational efficiency, clear characterization mechanism, and strong anti-interference ability. The designed subtle feature sharpening method is beneficial for quantitatively assessing bearing damage scale, providing an effective tool for accurate diagnosis of spindle bearing faults. The spindle bearing damage scale assessment mechanism designed in this invention helps enhance high-speed machining stability and ensure workpiece surface quality, solving the problem of workpiece vibration marks caused by spindle bearing disturbances. It overcomes the difficulty of online quantitative diagnosis caused by complex dynamic interference, providing favorable technical support for the safe and stable operation of spindle bearings.

[0040] It will be apparent to those skilled in the art that the present invention is limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered in all respects as exemplary and not restrictive, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, it is intended that all variations falling within the meaning and scope of equivalents of the claims be included within the present invention.

[0041] Furthermore, it should be understood that although this specification describes the embodiments, the embodiments do not necessarily contain only one independent technical solution. This way of describing the specification is only for clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in the embodiments can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

Claims

1. A method for fault detection of machine tool spindle bearings based on Lagrange weighted mode decomposition, characterized in that, The process includes the following steps: S1: Acquire vibration signal data. Specifically, a vibration sensor installed near the spindle of a high-speed machine tool is used to collect vibration signals during the machining process in real time. The collected vibration signals are denoted as x = {x1, x2, ..., x...} N }, where N represents the number of sample points for data acquisition; S2: Perform integer-cycle truncation on the acquired vibration signal x to improve the accuracy of Laplace transform analysis. Truncate the signal according to the signal sampling frequency and the spindle rotation speed, truncate the data once every 10,000 revolutions, and the truncated signal is denoted as y = [y1, y2, ..., y]. M S3: Perform Laplace Fourier transform analysis on the intercepted signal, evaluate the periodic pulse characteristics in each frequency band based on the time-frequency dual-dimensional characterization after the Laplace transform, and calculate the integrated characteristics of periodic pulses in each frequency band based on the principle of periodic harmonic product enhancement; S4: Perform weighted integration on the narrowband filtered signal after the Laplace transform based on the integrated characteristics in step S3 to enhance the weak damage characteristics caused by early bearing failure; S5: Perform envelope spectrum analysis on the weighted integrated signal in step S4, and then use it in the spindle bearing application software system to qualitatively identify bearing failures, and evaluate the spindle bearing damage state based on the bearing failure characteristic frequency, and determine the spindle operating state based on the evaluation and detection results; S6: Utilize the spindle bearing damage scale evaluation mechanism established in step S5 to statistically analyze the impact of different spindle bearing damage degrees on high-speed machine tool processing under dynamic working conditions, specifically including the impact on machining accuracy and tool life.

2. The method for fault detection of machine tool spindle bearings based on Lagrange weighted mode decomposition according to claim 1, characterized in that, In step S1, the sampling frequency is set to 50kHz, the machine tool spindle speed is 25314RPM, and the duration of each data acquisition is 2s.

3. The method for fault detection of machine tool spindle bearings based on Lagrange weighted mode decomposition according to claim 1, characterized in that, In step S3, using the formula Perform Laplace Fourier transform analysis on the truncated signal, where k = 1, 2, ..., M, c q (m) represents the Laplace transform coefficients, (k,q)=1 indicates that k and q are coprime numbers, and exp(·) is the Laplace transform basis function.

4. The method for fault detection of machine tool spindle bearings based on Lagrange weighted mode decomposition according to claim 1, characterized in that, In step S4, using the formula The narrowband filtered signal after the Laplace transform is weighted and integrated, where R(q) represents the weighting coefficient and q = 50.

5. The method for fault detection of machine tool spindle bearings based on Lagrange weighted mode decomposition according to claim 1, characterized in that, In step S5, the weighted integrated signal is subjected to envelope spectrum analysis using the formula F(f)=FFT{Hilbert[x(m)]}, where FFT{·} represents Fourier transform and Hilbert[·] represents Hilbert envelope analysis.

6. The method for fault detection of machine tool spindle bearings based on Lagrange weighted mode decomposition according to claim 1, characterized in that, In step S6, it is necessary to follow the formula. Calculate the sparse pulse characteristic data of the weighted integrated signal in step S5, where <·> represents the averaging operator operation and x(n) represents the data obtained in the nth experiment.

Images