High-precision and low-wear intelligent turning device and method for grinding machine

Abstract

The invention relates to the technical field of high-precision grinding machines, in particular to a high-precision and low-abrasion grinding machine intelligent rotation device and method, and the device and method are applied to a grinding machine. The vibration monitoring unit is installed on the rotating table mechanism and used for collecting vibration signals of the rotating table mechanism in the operation process; the servo driving unit is in transmission connection with the rotating table mechanism and is used for driving the rotating table mechanism to rotate according to preset motion parameters; the spindle control unit is connected with the grinding spindle and used for controlling grinding parameters of the grinding wheel; the central controller is in communication connection with the vibration monitoring unit, the servo driving unit and the main shaft control unit; according to the method, the wear feature fingerprints are constructed and motion-process collaborative optimization is realized, so that the intelligent level of the rotary table is improved from state monitoring to active control, and the problem of precision maintenance under the heavy-load eccentric wear working condition is solved.

Description

Technical Field

[0001] This invention relates to the field of high-precision grinding machine technology, and in particular to a high-precision, low-wear intelligent rotary device and method for grinding machines. Background Technology

[0002] The rotary table is one of the core functional components of modern high-precision grinding machines, undertaking key tasks such as workpiece bearing, indexing and positioning, and rotary feed. With the increasing demands for machining accuracy of complex curved surface parts in aerospace, mold manufacturing and other fields, the dynamic accuracy, wear resistance and long-term stability of the rotary table are facing severe challenges.

[0003] In existing technologies, high-precision rotary tables typically employ a servo motor + encoder closed-loop control scheme and are equipped with low-friction transmission components such as ceramic bearings and ball screws to reduce wear and extend service life. Meanwhile, some advanced devices have added a vibration monitoring module to provide real-time feedback on the equipment's operating status and issue an alarm when wear exceeds tolerance, so that maintenance can be performed by stopping the machine.

[0004] However, in actual machining, especially when the workpiece has a large overhang (such as large molds or blade machining), the overturning torque generated by the cutting force can cause localized uneven wear in the slewing bearing. Once uneven wear occurs, the rotational resistance torque will fluctuate periodically. This fluctuation will be transmitted back to the grinding zone, causing instantaneous changes in the grinding depth, which in turn can induce surface vibration marks or even grinding burns on the workpiece. Traditional vibration monitoring methods can only issue an "after-the-fact alarm" after the wear has accumulated to a certain extent. At this time, the workpiece often already has quality defects, and frequent downtime for maintenance seriously affects production efficiency. In addition, existing control methods cannot dynamically adjust motion parameters and grinding processes according to the wear state, making it difficult to proactively avoid its impact on machining quality in the early stages of wear.

[0005] Therefore, in view of the technical problem in the prior art that it is difficult to intervene in time after the rotary table wears unevenly, which leads to a decline in processing quality, a high-precision, low-wear intelligent rotary grinding device and method are proposed. Summary of the Invention

[0006] In order to overcome the problem that it is difficult to intervene in time and actively when the rotary table under the existing technology is worn out, which leads to a decline in processing quality.

[0007] The technical solution of this invention is: a high-precision, low-wear intelligent rotary device for grinding machines, applied to a grinding machine, the grinding machine including a grinding spindle and a grinding wheel mounted on the grinding spindle, comprising:

[0008] A rotary table mechanism is used to carry the workpiece and drive its rotation. A vibration monitoring unit is installed on the rotary table mechanism to collect vibration signals of the rotary table mechanism during operation. A servo drive unit is connected to the rotary table mechanism for driving the rotary table mechanism to perform rotary motion according to preset motion parameters. A spindle control unit, connected to the grinding spindle, is used to control the grinding parameters of the grinding wheel; The central controller is communicatively connected to the vibration monitoring unit, the servo drive unit, and the spindle control unit, respectively. The central controller is configured as follows: The vibration signal is received, and time-frequency domain analysis is performed on the vibration signal to extract feature values ​​related to the wear state of the mechanical transmission components in the rotary table mechanism, forming a wear feature fingerprint; Based on the wear feature fingerprint, a first control command is generated and sent to the servo drive unit to adjust the preset motion parameters; and based on the wear feature fingerprint, a second control command is generated and sent to the spindle control unit to adjust the grinding parameters.

[0009] Preferably, during the grinding process, the vibration monitoring unit continuously collects the vibration signals of the rotary table mechanism and transmits them to the central controller. The central controller, through time-frequency domain analysis, separates the frequency components related to fault characteristics such as bearing raceway damage and lead screw wear from the complex vibration signals, quantifying them as wear characteristic fingerprints. These fingerprints can quantitatively characterize the current wear state and degree of the mechanical transmission components. Subsequently, based on this fingerprint information, the central controller sends instructions to the servo drive unit to optimize the motion trajectory parameters of the table (such as jerk and acceleration curve) to avoid coupling between motion excitation and resonant frequencies caused by wear. On the other hand, it sends instructions to the spindle control unit to fine-tune the grinding depth or feed rate to reduce the cutting load in the wear aggravation zone. Through the above dual-channel coordinated adjustment, the active suppression of wear effects is achieved.

[0010] Preferably, the central controller is further configured as follows: Establish a benchmark fingerprint database, which contains feature value templates corresponding to different degrees of wear. The currently extracted wear feature fingerprint is compared with the feature value template in the benchmark fingerprint database to calculate the wear intensity index; Based on the threshold range of the wear intensity index, determine the generation strategy of the first control command and / or the second control command.

[0011] Preferably, the central controller is configured to perform time-frequency domain analysis on the vibration signal to extract wear feature fingerprints, specifically including: The vibration signal is subjected to order tracking processing and converted into a stationary signal in the angular domain; Spectral analysis is performed on the stationary signal in the angle domain to extract the frequency band energy corresponding to the fault characteristic frequency of the mechanical transmission component, which is then used as the feature value.

[0012] Preferably, the preset motion parameters include the jerk and / or acceleration curve of the rotary table mechanism; The central controller is configured to generate a first control command to adjust the preset motion parameters, specifically including: analyzing the dominant frequency component in the wear feature fingerprint, adjusting the parameters of the jerk and / or acceleration curve, so that the vibration energy excited by the rotational motion avoids the frequency band corresponding to the dominant frequency component.

[0013] Preferably, the grinding parameters include the grinding depth and / or feed rate of the grinding wheel; The central controller is configured to generate second control commands to adjust the grinding parameters, specifically including: Identify the angle range in the wear fingerprint where the feature value exceeds a preset threshold; When the rotary table mechanism drives the workpiece into the angle range, it sends a command to the spindle control unit to reduce the grinding depth and / or feed rate within that range.

[0014] Preferably, the rotary table mechanism includes: Base; A rotary table is used to support workpieces; A ceramic bearing is disposed between the base and the rotary table; A ball screw assembly, connected to the servo drive unit, is used to drive the rotary table to rotate. The vibration monitoring unit is installed on the base or near the ceramic bearing.

[0015] A high-precision, low-wear intelligent rotary method for grinding machines, applied to the aforementioned high-precision, low-wear intelligent rotary device for grinding machines, includes the following steps: S1: The vibration signal of the rotary table mechanism is collected in real time through the vibration monitoring unit; S2: The vibration signal is analyzed in the time and frequency domain by the central controller to extract feature values ​​related to the wear state of mechanical transmission components and form a wear feature fingerprint. S3: Based on the wear feature fingerprint, generate a first control command and send it to the servo drive unit to adjust the preset motion parameters of the rotary table mechanism; S4: Based on the wear feature fingerprint, generate a second control command and send it to the spindle control unit to adjust the grinding parameters of the grinding wheel.

[0016] Preferably, step S2, which involves extracting wear feature fingerprints, specifically includes: S21: Perform wavelet packet decomposition or short-time Fourier transform on the vibration signal to obtain the time spectrum. S22: In the time-frequency spectrum, extract the frequency band energy corresponding to the raceway fault characteristic frequency of the ceramic bearing and / or the pitch error characteristic frequency of the ball screw pair to form a multi-dimensional feature vector as the wear feature fingerprint.

[0017] Preferably, step S3 specifically includes: S31: Compare the wear feature fingerprint with the pre-stored reference fingerprint and calculate the wear intensity index; S32: If the wear intensity index exceeds the first preset threshold, then analyze the dominant frequency components in the wear feature fingerprint; S33: Adjust the acceleration and / or acceleration curve in the preset motion parameters so that the vibration energy generated by the rotary table mechanism during acceleration and deceleration avoids the resonance band corresponding to the dominant frequency component.

[0018] Preferably, step S4 specifically includes: S41: Combining the angular position information of the rotary table mechanism, analyze the law of wear feature fingerprint change with angle, and identify the wear aggravation angle range where the feature value exceeds the second preset threshold. S42: Store the wear aggravation angle range information to the central controller; S43: During the current and subsequent machining processes, when the rotary table mechanism drives the workpiece into the wear aggravation angle range, it sends an instruction to the spindle control unit to reduce the grinding depth and / or feed rate within the angle range, and restores the original grinding parameters after exiting the angle range.

[0019] The beneficial effects of this invention are: 1. This invention enables early quantitative perception of the microscopic wear state of core components such as bearings and lead screws by constructing wear feature fingerprints; by combining time-frequency domain analysis with order tracking technology, the system can identify anomalies in the early stage of wear and before they cause processing quality defects, thus gaining a time window for proactive intervention. 2. This invention uses wear fingerprints to drive two control loops simultaneously: first, it adjusts the motion parameters (jerk / acceleration curve) of the rotary table itself to avoid amplifying the wear effect from the excitation source; second, it adjusts the process parameters (grinding depth / feed speed) of the grinding spindle in conjunction with the rotation to reduce the impact on the wear zone from the load end. The dual-channel synergistic optimization forms a three-dimensional suppression of the wear effect, which is better than single-dimensional adjustment. For example, when the system detects an increase in the bearing fault characteristic frequency energy in the 90°~120° range, it optimizes the S-curve through this range to avoid the resonance frequency and temporarily reduces the grinding depth in this range. The combination of the two can effectively eliminate the risk of vibration marks and burns. 3. By establishing a benchmark fingerprint database and calculating the wear intensity index, this invention achieves refined management of graded response; for light wear, only motion parameters are optimized to avoid affecting processing efficiency; for moderate wear, both motion and grinding parameters are optimized to maintain processing capacity as much as possible while protecting the equipment; for heavy wear, a maintenance warning is issued; this intervention strategy takes into account both equipment protection and production efficiency, avoiding the production capacity loss caused by traditional downtime maintenance methods. 4. This invention slows down the rate of uneven wear by implementing temporary load reduction protection in the wear-accelerated area; under typical heavy-load overhang conditions, this technology can extend the effective service life of the slewing bearing and improve the accuracy retention time of the ball screw; correspondingly, unplanned equipment downtime is reduced and maintenance costs are significantly lowered. Attached Figure Description

[0020] Figure 1 The diagram shown is a three-dimensional structural schematic of the intelligent rotary device for grinding machines with high precision and low wear according to the present invention. Figure 2 The diagram shows the steps of the intelligent rotary grinding method for high precision and low wear according to the present invention. Figure 3 The flowchart shown is a sub-flowchart of the intelligent rotary grinding method for high precision and low wear of the present invention for extracting wear feature fingerprints. Figure 4 The flowchart shown is a sub-flowchart of the intelligent rotary grinding method for high-precision, low-wear grinding machines according to the present invention, which links grinding parameters for control. Figure 5 The diagram shown is a cross-sectional view of the intelligent rotary grinding device for high-precision, low-wear grinding machines according to the present invention. Explanation of reference numerals in the attached drawings: 1. Rotary table mechanism; 2. Vibration monitoring unit; 3. Servo drive unit; 11. Base; 12. Ceramic bearing; 13. Ball screw pair. Detailed Implementation

[0021] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0022] Example 1: Please see Figure 1 , Figure 2 , Figure 3 , Figure 4 and Figure 5 This invention provides an embodiment: a high-precision, low-wear intelligent rotary device for a grinding machine, applied to a grinding machine including a grinding spindle and a grinding wheel mounted on the grinding spindle, comprising: Rotary worktable mechanism 1 is used to carry the workpiece and drive it to rotate; Vibration monitoring unit 2 is installed on rotary table mechanism 1 and is used to collect vibration signals of rotary table mechanism 1 during operation; The servo drive unit 3 is connected to the rotary table mechanism 1 for driving the rotary table mechanism 1 to perform rotary motion according to preset motion parameters. The spindle control unit, connected to the grinding spindle, is used to control the grinding parameters of the grinding wheel; The central controller is communicatively connected to the vibration monitoring unit 2, the servo drive unit 3, and the spindle control unit, respectively. The central controller is configured as follows: The vibration signal is received and analyzed in the time and frequency domain to extract the feature values ​​related to the wear state of the mechanical transmission components in the rotary table mechanism 1, forming a wear feature fingerprint. Based on the wear feature fingerprint, a first control command is generated and sent to the servo drive unit 3 to adjust the preset motion parameters; and based on the wear feature fingerprint, a second control command is generated and sent to the spindle control unit to adjust the grinding parameters.

[0023] During the grinding process, the vibration monitoring unit 2 continuously collects the vibration signals of the rotary table mechanism 1 and transmits them to the central controller. The central controller uses time-frequency domain analysis to separate the frequency components related to fault characteristics such as bearing raceway damage and lead screw wear from the complex vibration signals, quantifying them as wear characteristic fingerprints. These fingerprints can quantitatively characterize the current wear state and degree of the mechanical transmission components. Subsequently, based on this fingerprint information, the central controller sends instructions to the servo drive unit 3 to optimize the motion trajectory parameters of the table (such as jerk and acceleration curve) to avoid coupling between motion excitation and resonant frequencies caused by wear. On the other hand, it sends instructions to the spindle control unit to fine-tune the grinding depth or feed rate to reduce the cutting load in the wear aggravation zone. Through the above dual-channel coordinated adjustment, the active suppression of wear effects is achieved.

[0024] Furthermore, the central controller is also configured as follows: Establish a benchmark fingerprint database, which contains feature value templates corresponding to different wear levels; Compare the currently extracted wear feature fingerprint with the eigenvalue template in the reference fingerprint library to calculate the wear intensity index; Determine the generation strategy of the first control instruction and / or the second control instruction according to the threshold interval where the wear intensity index is located.

[0025] Among them, the central controller has a built-in reference fingerprint library, which can be established through factory calibration or self-learning in the early healthy state of the equipment, and stores the typical feature vectors at each stage from normal wear to severe wear; during operation, by calculating the similarity between the current fingerprint and the template (such as Mahalanobis distance, Euclidean distance), the wear intensity index W is obtained; the system presets multiple thresholds (such as warning threshold W1, intervention threshold W2, shutdown threshold W3), and responds in grades according to the interval where W is located: when W≤W1, keep normal processing; when W1<W≤W2, only optimize the motion parameters; when W2<W≤W3, optimize both the motion parameters and the grinding parameters; when W>W3, issue a maintenance warning; this grading strategy realizes refined control and avoids over-intervention affecting efficiency.

[0026] Furthermore, the central controller is configured to perform time-frequency domain analysis on the vibration signal to extract the wear feature fingerprint, specifically including: Perform order tracking processing on the vibration signal to convert it into an angle-domain stationary signal; Perform spectrum analysis on the angle-domain stationary signal, and extract the band energy corresponding to the fault characteristic frequency of the mechanical transmission component as the eigenvalue.

[0027] Among them, since the rotational speed of the workbench may change during processing, directly performing spectrum analysis on the time-domain signal will cause frequency ambiguity; the present invention adopts the order tracking technology to perform equal-angle resampling on the vibration signal based on the workbench rotation angle, and converts the time-varying signal into an angle-domain stationary signal; then perform order spectrum analysis, and according to the fault characteristic orders (such as inner ring passing order, outer ring passing order, ball circulation order, etc.) pre-calculated based on the bearing geometric parameters and the lead of the lead screw, extract the amplitude or energy at the corresponding order as the wear feature value; this method eliminates the interference of rotational speed fluctuations and makes the extracted features more stable.

[0028] Furthermore, the preset motion parameters include the jerk and / or acceleration curve of the rotary workbench mechanism 1; The central controller is configured to generate a first control instruction to adjust the preset motion parameters, specifically including: analyzing the dominant frequency component in the wear feature fingerprint, and adjusting the parameters of the jerk and / or acceleration curve to make the vibration energy excited by the rotary motion avoid the frequency band corresponding to the dominant frequency component.

[0029] Among them, when the wear feature fingerprint shows a significant increase in the energy of a certain characteristic frequency component (such as the frequency corresponding to the bearing failure order), it means that the vicinity of that frequency may become a weak resonance point of the system; the servo drive unit 3 usually uses an S-shaped acceleration and deceleration curve to control the movement of the worktable, and the selection of its acceleration and deceleration directly affects the distribution of the vibration spectrum excited during the movement; the central controller changes the energy distribution of the motion excitation by adjusting the S-curve parameters (such as acceleration time, constant speed time, etc.), so that it avoids the frequency band where the wear feature frequency is located in the spectrum, avoids resonance amplification of the wear effect, and thus suppresses the vibration transmission to the processing area.

[0030] Furthermore, grinding parameters include the grinding depth and / or feed rate of the grinding wheel; The central controller is configured to generate a second control command to adjust grinding parameters, specifically including: Identify angle ranges in wear fingerprints where feature values ​​exceed a preset threshold; When the rotary table mechanism 1 drives the workpiece into the angle range, it sends a command to the spindle control unit to reduce the grinding depth and / or feed rate in that range.

[0031] The wear fingerprint is recorded synchronously with the corner position of the worktable, forming a "fingerprint-angle" map. The central controller analyzes this map to locate the angle range where the feature value is abnormally high (i.e., the area with the most severe wear). In subsequent processing, when the worktable carries the workpiece into this angle range, the central controller sends a real-time command to the spindle control unit to temporarily reduce the grinding depth (e.g., from 0.02mm to 0.015mm) or feed rate (e.g., from 1000mm / min to 700mm / min) in this range, thereby reducing the additional impact of the cutting load on the wear area. After exiting the range, the original parameters are automatically restored, which protects the worn parts and maintains the overall processing efficiency to the maximum extent.

[0032] Furthermore, the rotary table mechanism 1 includes: Base 11; A rotary table is used to support workpieces; A ceramic bearing 12 is disposed between the base 11 and the rotary table; The ball screw assembly 13 is connected to the servo drive unit 3 and is used to drive the rotary table to rotate. The vibration monitoring unit 2 is installed on the base 11 or near the ceramic bearing 12.

[0033] Among them, the ceramic bearing 12 has the advantages of low density, high stiffness and good thermal stability, which can reduce frictional heat generation; the ball screw pair 13 realizes precision rotary transmission; the vibration monitoring unit 2 is preferably installed at the base 11 near the bearing, where the vibration signal contains the richest bearing and screw fault information, and the signal transmission path is short and the attenuation is reduced, which is conducive to the extraction of early weak wear characteristics.

[0034] A high-precision, low-wear intelligent rotary method for grinding machines, applied to the aforementioned high-precision, low-wear intelligent rotary device for grinding machines, includes the following steps: S1: The vibration signal of the rotary table mechanism 1 is collected in real time through the vibration monitoring unit 2; S2: The vibration signal is analyzed in the time and frequency domain by the central controller to extract feature values ​​related to the wear state of mechanical transmission components and form a wear feature fingerprint. S3: Based on the wear feature fingerprint, generate the first control command and send it to the servo drive unit 3 to adjust the preset motion parameters of the rotary table mechanism 1; S4: Based on the wear feature fingerprint, generate a second control command and send it to the spindle control unit to adjust the grinding parameters of the grinding wheel.

[0035] Furthermore, step S2 involves extracting wear feature fingerprints, specifically including: S21: Perform wavelet packet decomposition or short-time Fourier transform on the vibration signal to obtain the time spectrum. S22: In the time-frequency spectrum, extract the frequency band energy corresponding to the raceway fault characteristic frequency of the ceramic bearing 12 and / or the pitch error characteristic frequency of the ball screw pair 13 to form a multi-dimensional feature vector as a wear feature fingerprint.

[0036] Among them, wavelet packet decomposition can finely divide vibration signals into different frequency bands, which is especially suitable for extracting transient impact features; short-time Fourier transform is suitable for analyzing time-varying signals; using these two time-frequency analysis tools, the energy bands corresponding to bearing fault characteristic frequencies (such as inner ring fault frequency f_i, outer ring fault frequency f_o) and screw fault characteristic frequencies (such as ball passing frequency f_b) can be accurately located on the time-frequency spectrum; the energy values, energy proportions, kurtosis and other indicators of these frequency bands can be extracted and combined into multi-dimensional feature vectors to form a wear fingerprint with physical interpretability.

[0037] Furthermore, step S3 specifically includes: S31: Compare the wear feature fingerprint with the pre-stored reference fingerprint and calculate the wear intensity index; S32: If the wear intensity index exceeds the first preset threshold, analyze the dominant frequency components in the wear feature fingerprint; S33: Adjust the acceleration and / or acceleration curve in the preset motion parameters so that the vibration energy generated by the rotary table mechanism 1 during acceleration and deceleration avoids the resonance band corresponding to the dominant frequency component.

[0038] Specifically, quantitative evaluation is achieved by calculating the wear intensity index W. When W exceeds the first threshold (such as W1), it indicates that wear has begun to affect the dynamic characteristics of the system. At this point, the energy proportion of each frequency band in the wear feature fingerprint is further analyzed to find the dominant frequency f_d with the largest energy. Subsequently, in the servo drive parameters, the jerk value of the S-shaped acceleration and deceleration curve is optimized to change the spectral envelope of the motion excitation, so that it forms a "concave" at f_d, thereby reducing the excitation intensity at this frequency point and avoiding resonance amplification.

[0039] Furthermore, step S4 specifically includes: S41: Combining the rotation position information of the rotary table mechanism 1, analyze the law of wear feature fingerprint change with angle, and identify the wear aggravation angle range where the feature value exceeds the second preset threshold. S42: Store the information on the wear aggravation angle range to the central controller; S43: During the current and subsequent machining processes, when the rotary table mechanism 1 drives the workpiece into the wear intensification angle range, a command is sent to the spindle control unit to reduce the grinding depth and / or feed rate in that angle range, and to restore the original grinding parameters after exiting the angle range.

[0040] Since uneven wear typically occurs within a specific angular range (such as the 90°~120° range where the maximum overturning moment is applied), the wear characteristic fingerprint will exhibit a periodic variation with the angle. By using the rotational angle information retained through order tracking, a "wear characteristic value-angle" curve can be plotted. The system automatically searches for continuous angular intervals on the curve where the characteristic value exceeds the second threshold (such as the characteristic value corresponding to W2) and marks them as areas of intensified wear. In subsequent machining, when the table rotation angle enters this marked interval, the central controller immediately triggers linkage protection: it sends a temporary deload command to the spindle control unit to reduce the grinding depth or feed rate in this interval, and automatically recovers after rotating out of the interval. This process is completed within the interpolation cycle of the CNC system, and the operator is unaware of it, but it effectively protects the worn parts and maintains the machining quality.

[0041] Through the above steps, this invention achieves early quantitative perception of the microscopic wear state of core components such as bearings and lead screws by constructing wear feature fingerprints. By combining time-frequency domain analysis with order tracking technology, the system can identify anomalies in the early stages of wear, before they cause machining quality defects, thus gaining a time window for proactive intervention. This invention uses wear fingerprints to drive two control loops simultaneously: first, adjusting the motion parameters (jerk / acceleration curve) of the rotary table itself to avoid resonance amplification of the wear effect from the excitation source; second, adjusting the process parameters (grinding depth / feed speed) of the grinding spindle in a coordinated manner to reduce the impact on the wear zone from the load end. Dual-channel collaborative optimization forms a three-dimensional suppression of wear effects, which is superior to single-dimensional adjustments. For example, when the system identifies an increase in the bearing fault characteristic frequency energy in the 90°~120° range, it optimizes the process... The S-curve in this range avoids the resonant frequency, and the grinding depth in this range is temporarily reduced. The combination of these two measures can effectively eliminate the risk of vibration marks and burns. By establishing a benchmark fingerprint database and calculating the wear intensity index, this invention achieves refined management of graded responses. In the case of light wear, only motion parameters are optimized to avoid affecting processing efficiency. In the case of moderate wear, both motion and grinding parameters are optimized to maintain processing capacity as much as possible while protecting the equipment. In the case of heavy wear, a maintenance warning is issued. This intervention strategy takes into account both equipment protection and production efficiency, avoiding the production capacity loss caused by traditional downtime maintenance methods. This invention slows down the development rate of uneven wear by implementing temporary load reduction protection in the wear aggravation zone. Under typical heavy-load overhang conditions, this technology can extend the effective service life of the slewing bearing and improve the accuracy retention time of the ball screw. Correspondingly, unplanned equipment downtime is reduced, and maintenance costs are significantly reduced.

[0042] Example 2: Optionally, this embodiment provides a high-precision, low-wear intelligent rotary device for grinding machines, applied to a five-axis linkage CNC grinding machine, which includes a grinding spindle and a grinding wheel mounted on the grinding spindle.

[0043] The device includes: The rotary table mechanism 1 is used to carry the workpiece and drive it to rotate. Specifically, the mechanism includes a base 11, a rotary table, a ceramic bearing 12, and a ball screw pair 13. The base 11 is fixed to the machine tool bed. The ceramic bearing 12 is located between the base 11 and the rotary table. It is a high-precision angular contact ceramic ball bearing with low density, high rigidity, and good thermal stability. The ball screw pair 13 is connected to the servo drive unit 3 and drives the rotary table to achieve precise rotary motion through a nut.

[0044] Vibration monitoring unit 2 is installed on rotary table mechanism 1 to collect vibration signals during operation. In this embodiment, a triaxial ICP type accelerometer (sensitivity 100mV / g, frequency response range 0.5Hz~10kHz) is used and installed on base 11 near ceramic bearing 12. The signal transmission path is short and the attenuation is small at this position, which is conducive to extracting early weak wear characteristics. The sensor is connected to the data acquisition card of the central controller through a shielded cable.

[0045] The servo drive unit 3 is connected to the ball screw pair 13 of the rotary table mechanism 1 and is used to drive the rotary table to rotate according to preset motion parameters. The unit includes a servo driver and a servo motor. The motor has a built-in high-precision absolute encoder (23-bit resolution) to realize closed-loop position control.

[0046] The spindle control unit, connected to the grinding spindle, is used to control the grinding parameters of the grinding wheel. This unit receives instructions from the central controller and can adjust the spindle speed, grinding depth, and feed rate in real time.

[0047] The central controller adopts an embedded control system based on FPGA+ARM architecture, with a built-in high-speed data acquisition card (sampling rate up to 50kHz) and model calculation module; the central controller is connected to the vibration monitoring unit 2, servo drive unit 3 and spindle control unit via industrial Ethernet.

[0048] The central controller is configured to execute the following core control logic: First, the vibration signal collected in real time by the vibration monitoring unit 2 is received, and time-frequency domain analysis is performed on the signal to extract feature values ​​related to the wear state of the mechanical transmission components, forming a wear feature fingerprint; the specific extraction process includes: (1) Preprocess the original vibration signal, including removing the DC component and anti-aliasing filtering; (2) Perform order tracking processing: Since the speed of the workbench may change, direct spectrum analysis will produce frequency ambiguity; In this embodiment, the order tracking technology is used to synchronously collect the rotation pulse signal of the servo motor encoder, and use this as a reference to resample the vibration signal at equal angles, converting the non-stationary signal in the time domain into a stationary signal in the angle domain; the resampling angle interval Δθ is set to 0.1° to ensure that the high-frequency components are not aliased.

[0049] (3) Perform a fast Fourier transform on the stationary signal in the angle domain to obtain the order spectrum; calculate the fault characteristic order in advance based on the geometric parameters of the ceramic bearing 12 and the lead of the ball screw: Bearing outer ring fault characteristic order:

[0050] Bearing inner ring fault characteristic order:

[0051] Ball passing through the characteristic order of the lead screw:

[0052] Among them, Z is the number of rolling elements, d is the diameter of the rolling element, D is the pitch diameter of the bearing, α is the contact angle, L is the lead of the lead screw, and P is the lead screw stroke corresponding to the ball circulation period; in this embodiment, Z = 20, d = 12.7 mm, D = 180 mm, α = 15°, L = 20 mm, P = 40 mm, and it is calculated that O_o≈8.5, O_i≈11.5, O_b = 0.5.

[0053] (4) Extract the amplitudes or energies at the above characteristic orders O_o, O_i, O_b and their harmonic frequencies in the order spectrum to form a multi-dimensional feature vector F = [E_o, E_i, E_b, K_o, K_i, K_b], where E represents the frequency band energy and K represents the kurtosis index; this feature vector is the wear feature fingerprint at the current moment.

[0054] Secondly, the central controller has a built-in reference fingerprint library, which is established through no-load running and simulated load tests of the equipment in a healthy state before leaving the factory, and stores the eigenvalue templates corresponding to different wear degrees (normal, slight, medium, severe); during operation, compare the currently extracted wear feature fingerprint F with each template F_ref in the reference fingerprint library to calculate the wear intensity index W; in this embodiment, the Mahalanobis distance is used to calculate the similarity:

[0055] Among them, Σ is the covariance matrix of the reference fingerprint library; the larger the value of W, the higher the deviation degree of the current wear state from the healthy state.

[0056] The system presets three thresholds: warning threshold W1 = 0.6, intervention threshold W2 = 0.8, shutdown threshold W3 = 1.2; according to the interval where W is located, determine the hierarchical response strategy: If W≤W1, it is determined to be in a normal state, and the current processing parameters remain unchanged; If W1 < W≤W2, it is determined to be slightly worn, and the motion parameter optimization is started; If W2 < W≤W3, it is determined to be moderately worn, and at the same time, the motion parameter optimization and grinding parameter adjustment are started; If W > W3, it is determined to be severely worn, and a maintenance warning signal is issued.

[0057] The central controller generates the first control instruction and sends it to the servo drive unit 3 to adjust the motion parameters, and generates the second control instruction and sends it to the spindle control unit to adjust the grinding parameters according to the classification results, so as to achieve the active perception and adaptive avoidance of the wear state.

[0058] Example 3: Optionally, based on Example 2, this example further details the specific implementation of motion parameter optimization.

[0059] When the wear intensity index W satisfies W1 < W ≤ W2, the central controller starts the motion parameter optimization subroutine. The specific steps are as follows: (1) Analyze the dominant frequency component in the wear feature fingerprint; sort the energy of each frequency band in the current feature vector F to find the feature frequency band with the largest energy, denoted as f_d; for example, if the energy E_o corresponding to the outer ring fault feature order O_o is the largest, the actual frequency value at the working table rotation frequency corresponding to the dominant frequency f_d is f_d = O_o × (n / 60), where n is the current working table speed (r / min); in this example, if n = 30 r / min, then f_d = 8.5 × 0.5 = 4.25 Hz.

[0060] (2) In the servo drive unit 3, the motion trajectory of the working table is controlled by an S-shaped acceleration and deceleration curve, and its acceleration expression is:

[0061] Among them, J is the jerk (mm / s³), A_max is the maximum acceleration (mm / s²), and T_1~T_7 are the time points of each stage; the vibration spectrum excited during the motion process is closely related to the J value. The larger the J value, the richer the high-frequency excitation components.

[0062] (3) The central controller adjusts the jerk J to change the spectral distribution of the motion excitation so as to form a "spectral depression" at the dominant frequency f_d; the specific algorithm is: according to the preset optimization objective function, solve the optimal J value:

[0063] Among them, S(f,J) is the theoretical spectral density function of the motion excitation at a given jerk J, which can be pre-calibrated through the transfer function model of the system; Δf is the frequency band width, which is taken as 0.5 Hz in this example; J_min and J_max are the adjustable ranges of the jerk, which are 1000 mm / s³ and 10000 mm / s respectively in this example; (4) Write the obtained optimal jerk J_opt into the motion control parameters of the servo drive to update the subsequent motion trajectory; after optimization, the integral value of the vibration energy excited by the working table during the acceleration and deceleration process near f_d is the smallest, thus avoiding the resonance amplification effect on the dominant frequency component caused by wear and suppressing the transmission of vibration to the processing area.

[0064] Example 4: Optionally, based on Embodiment 2, this embodiment further details the specific implementation method of the grinding parameter linkage control.

[0065] When the wear intensity index W satisfies W2 < W ≤ W3, while starting the optimization of the motion parameters, the central controller starts the grinding parameter linkage control subroutine; the specific steps are as follows: (1) Identify the angle range with increased wear; since eccentric wear usually occurs in a specific angle range (such as the area承受最大倾覆力矩的区域, the wear characteristic fingerprint will show a periodic change law with the angle; the central controller combines the position information of the workbench rotation angle encoder, synchronously records the continuously collected wear characteristic vector F and the rotation angle θ, and forms an "eigenvalue - angle" sequence; the sliding window method is used to calculate the mean value of the eigenvalues in each angle range:

[0066] where θ is the starting angle, Δθ is the sampling interval (in this embodiment, it is taken as 1°), and N is the window length (corresponding to 10°); the calculated mean value sequence reflects the change trend of the wear characteristics with the angle.

[0067] (2) Set the characteristic threshold E_th, which is taken as 3 times the standard deviation of the normal state eigenvalue mean in the reference fingerprint library in this embodiment; search for all continuous angle ranges that satisfy E(θ) > E_th; Record its starting angle θ_s and ending angle θ_e, and form a set Φ of angle ranges with increased wear = {[θ_s1, θ_e1], [θ_s2, θ_e2],...}; in this embodiment, it is identified that the eigenvalue in the 90° - 120° interval increases significantly and is marked as the wear - increased area; (3) Store the information of the angle range with increased wear in the non - volatile memory of the central controller; (4) During the current and subsequent machining processes, execute the linkage protection logic; the numerical control system reads the actual rotation angle θ_cur of the current workbench in each interpolation cycle (in this embodiment, it is 1ms), and judges whether θ_cur is within any of the angle ranges with increased wear: If θ_cur ∈ [θ_s, θ_e], then enter the linkage protection mode; the central controller sends an instruction to the spindle control unit to temporarily reduce the grinding parameters in this interval; in this embodiment, the grinding depth is reduced from the reference value a_p = 0.02mm to a_p' = 0.015mm (a reduction of 25%), and the feed rate is reduced from v_f = 1000mm / min to v_f' = 700mm / min (a reduction of 30%); to prevent machining surface steps caused by parameter mutation, a transition zone length Δθ_trans = 5° is set when entering and exiting the interval, and the parameters are smoothly adjusted by linear interpolation in the transition zone: It should be noted that the Chinese phrase "承受最大倾覆力矩的区域" in the original text seems to be incomplete or incorrect. It might need to be further corrected or clarified in the original content for a more accurate translation.

[0068] After exiting the wear acceleration zone, the grinding parameters automatically return to the reference value.

[0069] (5) After each processing is completed, the system records the effect of the linkage protection, including the change trend of wear characteristic values, the workpiece surface quality detection results, etc., for subsequent self-learning optimization.

[0070] Example 5: Optionally, this embodiment provides a specific structural configuration scheme.

[0071] The rotary table mechanism 1 includes: Base 11 is integrally cast from high-strength cast iron HT300 and undergoes aging treatment to eliminate internal stress; The rotary worktable has a diameter of Φ800mm and T-slots on the table surface for mounting workpieces or fixtures. Ceramic bearing 12 is an imported high-precision hybrid ceramic ball bearing, model HS7014-CT-P4S, with an inner diameter of 70mm, an outer diameter of 110mm, a width of 20mm, rolling elements of Si3N4 ceramic balls, a cage of phenolic resin, and a precision grade of P4. Ball screw assembly 13 uses a ground ball screw with a nominal diameter of 40mm, a lead of 20mm, and a precision grade of C3. It has a built-in preload nut to eliminate backlash. Servo drive unit 3 includes a Siemens 1FK7 series servo motor (rated torque 27Nm, maximum speed 3000r / min) and an S120 driver.

[0072] The vibration monitoring unit 2 uses a PCB company model 356A16 triaxial ICP accelerometer with a sensitivity of 100mV / g, a range of ±50g, and a frequency range of 0.5~5kHz. It is installed on the side of the base 11 via a magnetic mount, 50mm away from the center line of the ceramic bearing 12. The sensor signal is connected to the data acquisition card of the central controller via a shielded cable.

[0073] The central controller uses Beckhoff CX5130 embedded controller, equipped with an Intel Atom processor, and has a built-in 2-channel analog input module EL3632 for acquiring vibration signals (sampling rate is configurable, up to 20kHz), as well as an EtherCAT bus communication module for real-time communication with the servo drive and spindle control unit.

[0074] Example 6: Optionally, this embodiment provides a complete implementation process for a high-precision, low-wear intelligent rotation method for grinding machines.

[0075] The method, applied to the apparatus of the foregoing embodiments, includes the following steps: S1: Real-time acquisition of vibration signals During the grinding process, the vibration monitoring unit 2 continuously collects triaxial vibration acceleration signals x(t), y(t), and z(t) at a sampling rate of 10kHz, and synchronously collects the rotation angle pulse signal of the servo motor encoder, which is then transmitted to the central controller.

[0076] S2: Wear Feature Fingerprint Extraction The central controller processes the vibration signals as follows: S21: Vector synthesis of the three-dimensional signals x(t), y(t), and z(t) is performed to obtain the combined vibration signal a(t) = √(x² + y² + z²), which comprehensively reflects the overall vibration energy; S22: Perform wavelet packet decomposition on a(t), select the db10 wavelet basis, and decompose the signal into 4 layers to obtain 16 wavelet packet coefficients with equal bandwidth; calculate the reconstructed signal energy E_j=∑|c_j(k)|² for each band, where c_j(k) is the wavelet packet coefficient of the j-th band. S23: Select the corresponding sensitive frequency band based on the fault characteristic frequency range of the bearing and the lead screw; in this embodiment, the theoretical value of the fault characteristic frequency of the outer ring of the bearing is f_o=O_o×n / 60. When n=30~300r / min, the range of f_o is 4.25~42.5Hz, corresponding to the first and second frequency bands of the wavelet packet; the fault characteristic frequency f_i of the inner ring is 5.75~57.5Hz, corresponding to the second and third frequency bands; the characteristic frequency f_b of the lead screw is 0.25~2.5Hz, corresponding to the first and second frequency bands. S24: Extract the energy values ​​of the above sensitive frequency bands and combine them with the kurtosis index K_j=(1 / N)∑((c_j(k)-μ_j) / σ_j)^4 of each frequency band, where μ_j is the mean, σ_j is the standard deviation, and N is the number of coefficients, to form a 12-dimensional wear feature vector F=[E_1,E_2,E_3,K_1,K_2,K_3].

[0077] S3: Wear Intensity Assessment and Motion Parameter Optimization S31: Calculate the Mahalanobis distance between the feature vector F and the normal state template F0 in the benchmark fingerprint database to obtain the wear intensity index W; The benchmark fingerprint database is established by collecting data from the device running continuously for 24 hours in a healthy state, and includes the mean value and covariance matrix of the feature vector under normal working conditions. S32: If W>0.6 (first preset threshold), then start motion parameter optimization; analyze the proportion of each energy component in F, find the dominant frequency band j_d with the largest energy; based on the frequency range corresponding to j_d, determine the center frequency f_c of the resonance frequency band to be avoided (take the center frequency of this frequency band). S33: Adjust the S-shaped acceleration / deceleration curve parameters in the servo drive unit 3 by using a lookup table or an online optimization algorithm; This embodiment has a pre-stored excitation spectrum characteristic table under different acceleration J values, and updates the motion control parameters by interpolating to find the J value that minimizes the spectrum amplitude at f_c.

[0078] S4: Grinding parameter linkage adjustment If W > 0.8 (second preset threshold), add the following operation based on step S3: S41: Combining the corner position information, plot the curve of energy change with angle for each sensitive frequency band, and use the peak detection algorithm to identify continuous angle intervals where the energy exceeds the threshold E_th (take 3 times the average value under normal conditions); S42: Store the identified wear aggravation angle range into the process parameter table of the central controller; S43: In subsequent processing, the current corner position is determined in real time. When entering the wear intensification zone, a command is sent to the spindle control unit to reduce the grinding depth and / or feed rate in that zone. After exiting, the feed rate is restored. The parameter transition method adopts the linear interpolation smoothing process of Example 4.

[0079] Example 7: Optionally, this embodiment adds self-learning and parameter adaptive optimization functions based on the above embodiments.

[0080] The central controller has a built-in reinforcement learning module. It uses the change in wear intensity index W ΔW and the workpiece surface roughness detection value Ra as inputs to the reward function, and uses graded thresholds (W1, W2), linkage reduction ratio (δ_a, δ_v), and transition zone length Δθ_trans as adjustable parameters. By continuously accumulating data in actual processing, these parameters are optimized to make the control strategy adapt to different workpieces, different tools and different wear stages.

[0081] For example, if the initial grinding depth reduction δ_a is set to 25%, after several processing cycles, if the upward trend of W value slows down and the surface quality is acceptable within the wear aggravation range, then this parameter is maintained; if the W value continues to rise rapidly, then δ_a is gradually increased to 30%, 35%, until the optimal balance point is found; at the same time, if it is found that W continues to rise even after load reduction in a certain range, then the threshold E_th of that range is automatically reduced, and a larger load reduction protection is entered in advance.

[0082] Through the aforementioned self-learning mechanism, the control method of this invention possesses adaptive evolution capabilities, enabling it to continuously optimize protection strategies as equipment ages, thereby achieving intelligent management throughout its entire lifecycle.

[0083] The embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the present invention is not limited to the above embodiments. Within the scope of knowledge possessed by those skilled in the art, various changes can be made without departing from the spirit of the present invention.

Claims

1. A high-precision, low-wear intelligent rotary device for a grinding machine, applied to a grinding machine comprising a grinding spindle and a grinding wheel mounted on the grinding spindle, characterized in that: include: A rotary table mechanism (1) is used to carry the workpiece and drive it to rotate; Vibration monitoring unit (2) is installed on the rotary table mechanism (1) and is used to collect vibration signals of the rotary table mechanism (1) during operation; The servo drive unit (3) is connected to the rotary table mechanism (1) for driving the rotary table mechanism (1) to rotate according to preset motion parameters; A spindle control unit, connected to the grinding spindle, is used to control the grinding parameters of the grinding wheel; The central controller is communicatively connected to the vibration monitoring unit (2), the servo drive unit (3), and the spindle control unit, respectively. The central controller is configured as follows: The vibration signal is received, and the vibration signal is analyzed in the time and frequency domain to extract the feature values ​​related to the wear state of the mechanical transmission components in the rotary table mechanism (1) to form a wear feature fingerprint; Based on the wear feature fingerprint, a first control command is generated and sent to the servo drive unit (3) to adjust the preset motion parameters; and based on the wear feature fingerprint, a second control command is generated and sent to the spindle control unit to adjust the grinding parameters.

2. The high-precision, low-wear intelligent rotary device for grinding machines according to claim 1, characterized in that: The central controller is also configured to: Establish a benchmark fingerprint database, which contains feature value templates corresponding to different degrees of wear. The currently extracted wear feature fingerprint is compared with the feature value template in the benchmark fingerprint database to calculate the wear intensity index; Based on the threshold range of the wear intensity index, determine the generation strategy of the first control command and / or the second control command.

3. The high-precision, low-wear intelligent rotary device for grinding machines according to claim 1, characterized in that: The central controller is configured to perform time-frequency domain analysis on the vibration signal to extract wear feature fingerprints, specifically including: The vibration signal is subjected to order tracking processing and converted into a stationary signal in the angular domain; Spectral analysis is performed on the stationary signal in the angle domain to extract the frequency band energy corresponding to the fault characteristic frequency of the mechanical transmission component, which is then used as the feature value.

4. The high-precision, low-wear intelligent rotary device for grinding machines according to claim 1, characterized in that: The preset motion parameters include the jerk and / or acceleration curve of the rotary table mechanism (1); The central controller is configured to generate a first control command to adjust the preset motion parameters, specifically including: analyzing the dominant frequency component in the wear feature fingerprint, adjusting the parameters of the jerk and / or acceleration curve, so that the vibration energy excited by the rotational motion avoids the frequency band corresponding to the dominant frequency component.

5. The high-precision, low-wear intelligent rotary device for grinding machines according to claim 1, characterized in that: The grinding parameters include the grinding depth and / or feed rate of the grinding wheel; The central controller is configured to generate second control commands to adjust the grinding parameters, specifically including: Identify the angle range in the wear fingerprint where the feature value exceeds a preset threshold; When the rotary table mechanism (1) drives the workpiece into the angle range, it sends a command to the spindle control unit to reduce the grinding depth and / or feed rate in the range.

6. The high-precision, low-wear intelligent rotary device for grinding machines according to claim 1, characterized in that: The rotary table mechanism (1) includes: Base (11); A rotary table is used to support workpieces; A ceramic bearing (12) is disposed between the base (11) and the rotary table; A ball screw assembly (13) is connected to the servo drive unit (3) and is used to drive the rotary table to rotate. The vibration monitoring unit (2) is installed on the base (11) or near the ceramic bearing (12).

7. A high-precision, low-wear intelligent rotary method for grinding machines, applied to the high-precision, low-wear intelligent rotary device for grinding machines as described in any one of claims 1-6, characterized in that: Includes the following steps: S1: The vibration signal of the rotary table mechanism (1) is collected in real time by the vibration monitoring unit (2); S2: The vibration signal is analyzed in the time and frequency domain by the central controller to extract feature values ​​related to the wear state of mechanical transmission components and form a wear feature fingerprint. S3: Based on the wear feature fingerprint, generate a first control command and send it to the servo drive unit (3) to adjust the preset motion parameters of the rotary table mechanism (1); S4: Based on the wear feature fingerprint, generate a second control command and send it to the spindle control unit to adjust the grinding parameters of the grinding wheel.

8. The intelligent rotation method for a grinding machine with high precision and low wear according to claim 7, characterized in that: The step S2 involves extracting wear feature fingerprints, specifically including: S21: Perform wavelet packet decomposition or short-time Fourier transform on the vibration signal to obtain the time-frequency spectrum. S22: In the time spectrum diagram, extract the frequency band energy corresponding to the raceway fault characteristic frequency of the ceramic bearing (12) and / or the pitch error characteristic frequency of the ball screw pair (13) to form a multi-dimensional feature vector as the wear feature fingerprint.

9. The intelligent rotation method for a grinding machine with high precision and low wear according to claim 7, characterized in that: Step S3 specifically includes: S31: Compare the wear feature fingerprint with the pre-stored reference fingerprint and calculate the wear intensity index; S32: If the wear intensity index exceeds the first preset threshold, analyze the dominant frequency components in the wear feature fingerprint; S33: Adjust the acceleration and / or acceleration curve in the preset motion parameters so that the vibration energy generated by the rotary table mechanism (1) during acceleration and deceleration avoids the resonance band corresponding to the dominant frequency component.

10. The intelligent rotation method for a grinding machine with high precision and low wear according to claim 7, characterized in that: Step S4 specifically includes: S41: Combining the angular position information of the rotary table mechanism (1), analyze the law of the wear feature fingerprint changing with the angle, and identify the wear aggravation angle range where the feature value exceeds the second preset threshold. S42: Store the wear aggravation angle range information to the central controller; S43: During the current and subsequent machining process, when the rotary table mechanism (1) drives the workpiece into the wear aggravation angle range, it sends an instruction to the spindle control unit to reduce the grinding depth and / or feed rate in the angle range, and restores the original grinding parameters after exiting the angle range.

Images