A machining spindle intelligent pre-tightening force control method and control system

Abstract

The application discloses a machining spindle intelligent pre-tightening force control method and a control system. In the healthy state of the spindle, a plurality of known pre-tightening forces are applied through a pre-tightening force executing mechanism, the dynamic stiffness values of the spindle under each pre-tightening force are measured, a pre-tightening force-dynamic stiffness mapping relationship is established, a micro excitation signal is applied to the spindle system during the machining interval or the idle stage of the spindle, the spindle system signals are synchronously collected, adaptive filtering processing is performed on the collected signals, the current dynamic stiffness value, the thermal influence factor and the wear health index are calculated, the pre-tightening force compensation amount is calculated through a nonlinear adaptive model based on the deviation of the current dynamic stiffness value and the target stiffness value and the fusion of the thermal influence factor and the wear health index, the data is iteratively updated according to the pre-tightening force compensation amount, and the dynamic intelligent adjustment of the spindle pre-tightening force is realized through the combination of multi-modal perception, the nonlinear adaptive model, hierarchical driving control and long-term tracking re-calibration, thereby bringing multi-level significant beneficial effects.

Description

Technical Field

[0001] This invention belongs to the field of spindle preload control technology, specifically a method and control system for intelligent preload control of machining spindles. Background Technology

[0002] In the field of CNC machine tool machining, the spindle, as a core functional component, directly determines the final quality of the workpiece through its dynamic performance and machining accuracy. Currently, most machining spindles adopt a fixed preload method, which means that during the assembly stage, the bearing preload is set to a fixed value through mechanical structures such as spacers and locking nuts. Although this preload method is simple in structure and low in cost, it has significant inherent defects: to meet the high requirements for spindle rigidity during low-speed heavy-load cutting, a large preload is usually required. However, this will cause a sharp increase in bearing frictional heat during high-speed spindle operation, resulting in significant temperature rise and thermal deformation, which not only affects machining accuracy but also seriously shortens the bearing life. Conversely, if a small preload is used to control the temperature rise under high-speed conditions, the spindle will exhibit insufficient rigidity and increased vibration under low-speed, high-cutting-force conditions, thus affecting the surface quality and shape accuracy of the machined parts.

[0003] Therefore, a method is needed that can sense the preload drift of the spindle bearing online and perform intelligent dynamic compensation accordingly to overcome the preload failure caused by long-term wear and temperature changes, and ensure that the spindle maintains stable dynamic performance and machining accuracy under different working conditions and throughout its entire life cycle. Summary of the Invention

[0004] The purpose of this invention is to provide a method and control system for intelligent preload control of machining spindles to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] A method for intelligent preload control of a machining spindle includes the following steps:

[0007] S1: Under healthy spindle conditions, apply multiple known preloads via a preload actuator, measure the spindle dynamic stiffness under each preload, establish a preload-dynamic stiffness mapping relationship, and determine the target stiffness range. ;

[0008] S2: During machining breaks or when the spindle is idle, apply a small excitation signal to the spindle system and simultaneously collect multi-axis vibration acceleration signals, multi-point temperature signals, and acoustic emission signals of the spindle system.

[0009] S3: Perform adaptive filtering on the acquired signal and calculate the current dynamic stiffness value. Thermal Influence Factors and wear and tear health index ;

[0010] S4: Based on the deviation between the current dynamic stiffness value and the target stiffness value And integrate thermal influence factors and wear and tear health index The preload compensation amount is calculated using a nonlinear adaptive model. ;

[0011] S5: Based on the preload compensation amount The graded drive hydraulic preload actuator dynamically adjusts the preload of the spindle bearing, wherein the single adjustment range does not exceed 10% of the maximum preload.

[0012] S6: Repeat steps S2-S3 to obtain the compensated spindle dynamic stiffness value. If the deviation from the target stiffness value does not fall within the preset tolerance range, then repeat steps S4-S5 for iterative compensation until the convergence criterion is met or the maximum number of iterations is reached.

[0013] Further technical solutions include S7: During long-term operation, monitor the trend changes in the dynamic stiffness of the spindle and the wear health index. When the recalibration trigger condition is met, re-execute step S1 to update the preload-dynamic stiffness mapping relationship.

[0014] In a further technical solution, the micro-amplitude excitation signal in step S2 is generated by a piezoelectric ceramic vibrator, an electromagnetic vibrator, or a controlled spindle motor torque ripple integrated into the spindle system, with an excitation frequency range of 0.1-2kHz and an amplitude of less than 5μm.

[0015] In a further technical solution, the adaptive filtering process described in step S3 adopts the least mean square algorithm, the main sensor signal is the acceleration signal close to the excitation point, the reference sensor signal is the acceleration signal far from the excitation point, and the adaptive weight update step size μ is set to 0.01-0.05.

[0016] A further technical solution is the thermal influence factor mentioned in step S3. The calculation formula is:

[0017]

[0018] in, The average temperature rise of the bearing, For the axial temperature gradient of the bearing, Integrating over time, , , These are the coefficients calibrated experimentally.

[0019] A further technical solution is the wear health index mentioned in step S3. The calculation formula is:

[0020]

[0021] in, This represents the ratio of the bearing wear characteristic frequency band energy to the baseline energy. This is the ratio of the current acoustic emission intensity to the baseline intensity. For equivalent runtime, , , , These are the weighting coefficient and the attenuation coefficient.

[0022] A further technical solution is that the nonlinear adaptive model mentioned in step S4 is:

[0023]

[0024] in, This represents the normalized deviation between the current dynamic stiffness value and the target stiffness value. The preload from the previous moment, , , , , As adaptive parameters, the initial values ​​are obtained from calibration, and the online update rate is limited by: .

[0025] A further technical solution, the hierarchical driving in step S5 includes:

[0026] Fast compensation loop with a response time of less than 100 milliseconds, based on the current dynamic stiffness deviation. K is adjusted using proportional-integral-derivative methods;

[0027] Slow optimization loop with a response time of 1-10 seconds, based on thermal influence factor. and wear and tear health index Perform feedforward compensation and feedback optimization;

[0028] The protective constraint module ensures that the preload adjustment rate does not exceed 50 N / s and the cumulative adjustment amount does not exceed 80% of the bearing's ultimate load.

[0029] In a further technical solution, the convergence criterion in step S6 includes any of the following conditions:

[0030] <1% of the target stiffness value;

[0031] Two consecutive preload adjustments <5N;

[0032] The number of iterations reached the preset maximum of 5.

[0033] In a further technical solution, the recalibration triggering condition in step S7 includes any of the following conditions:

[0034] The 30-day sliding average change in the dynamic stiffness of the main shaft exceeds 8% of the initial value;

[0035] Wear and tear health index It dropped below 0.7;

[0036] The correlation coefficient of the temperature-stiffness relationship curve is less than 0.9;

[0037] The cumulative running time exceeds the preset scheduled maintenance cycle.

[0038] A machining spindle control system includes a multimodal sensing unit, comprising at least three triaxial accelerometers, at least three temperature sensors, and at least one broadband acoustic emission sensor arranged at a bearing housing.

[0039] The signal processing and feature extraction module is connected to the multimodal sensing unit and is used to adaptively filter the acquired signal and calculate the current dynamic stiffness value, thermal influence factor and wear health index.

[0040] A nonlinear adaptive controller, connected to the signal processing and feature extraction module, calculates the preload compensation amount through a nonlinear adaptive model based on the nonlinear mapping relationship between preload and dynamic stiffness and the target stiffness range, integrating the current dynamic stiffness deviation, thermal influence factor, and wear health index.

[0041] A preload actuator includes a housing, which has at least one hydraulic channel penetrating the space inside and outside the housing.

[0042] The preload ring is connected to the inner side of the housing through a two-way sealing structure. It has a flange on its outer side, which abuts against the outer ring of the main shaft bearing. Its inner side corresponds to the output end of the oil pressure channel, forming a closed oil pressure chamber.

[0043] Several elastic elements are disposed between the inner side of the preload ring and the housing to provide initial preload force under zero oil pressure conditions;

[0044] A hydraulic servo system is used to provide controllable hydraulic pressure to the hydraulic channels;

[0045] The oil temperature sensor, integrated into the oil circuit of the housing, is used to monitor the oil temperature;

[0046] The oil pressure compensation module corrects the oil pressure-preload mapping relationship based on the oil temperature sensor signal.

[0047] Further technical solutions also include a safety monitoring and fault tolerance module, which monitors the status of each component of the system in real time and automatically switches to a safe mode when a fault is detected.

[0048] A further technical solution is provided, wherein the security monitoring and fault tolerance module includes:

[0049] The sensor health monitoring unit detects the effectiveness and consistency of signals from each sensor in real time.

[0050] Hydraulic system fault diagnosis unit monitors hydraulic pressure build-up time, pressure stability, and signs of leakage;

[0051] When a serious fault is detected, the degradation strategy execution unit automatically switches to the fixed preload mode and maintains the most recent effective compensation value.

[0052] A further technical solution is that the nonlinear adaptive controller adopts a model reference adaptive control architecture, and the reference model is:

[0053]

[0054] in, For nominal dynamic stiffness, For transmission delay, , It is a time constant. For the Laplace operator.

[0055] Further technical solutions also include a predictive scheduling module, which predicts the processing interval length based on the CNC machining program. If the current interval time is insufficient to complete the full compensation process, the compensation is postponed to the next processing interval that meets the time requirements.

[0056] The beneficial effects of this invention are:

[0057] This invention achieves dynamic intelligent adjustment of spindle preload by combining multimodal sensing, nonlinear adaptive model, hierarchical drive control and long-term tracking recalibration, bringing significant benefits at multiple levels.

[0058] Other features and advantages of the present invention will be described in detail in the following detailed description section. Attached Figure Description

[0059] Figure 1 : A schematic diagram of the process of this invention.

[0060] Figure 2 The spindle cross-sectional structure of the present invention Figure 1 .

[0061] Figure 3 The spindle cross-sectional structure of the present invention Figure 2 .

[0062] Reference numerals: 1-Housing, 11-First channel, 12-First step, 2-Rotating shaft, 3-Bearing, 31-Outer ring, 32-Inner ring, 33-Ball body, 4-Preload ring, 41-Flange, 42-First contact part, 43-Second contact part, 44-Second step, 45-First seal, 46-Second seal, 5-Spring, 61-First limiting member, 62-Sleeve, 63-Second limiting member. Detailed Implementation

[0063] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0064] Please refer to Figure 1-3 ;

[0065] This invention discloses an intelligent preload control method for machining spindles, comprising the following steps:

[0066] S1: Initial calibration phase. Under healthy spindle conditions, multiple known preload forces are applied via the preload actuator. Healthy condition refers to the state before any operation is performed and the bearings are unworn. In existing technology, springs or other components are typically built into the spindle to apply pressure to the bearings; this pressure serves as the base preload force, but it is used as the zero-point standard during measurement. Multiple known preload forces are applied via the preload actuator. Measure the dynamic stiffness value of the spindle under various preloads. Establish the preload-dynamic stiffness mapping relationship and determine the target stiffness range. .

[0067] S2: During the online monitoring phase, during machining breaks or when the spindle is idle, a micro-amplitude excitation signal is applied to the spindle system. Simultaneously, multi-axis vibration acceleration signals, multi-point temperature signals, and acoustic emission signals of the spindle system are acquired. The micro-amplitude excitation signal is generated by a piezoelectric ceramic vibrator, an electromagnetic vibrator, or a controlled spindle motor torque ripple integrated into the spindle system. The excitation frequency range is 0.1-2kHz, and the amplitude is less than 5μm. Preferably, a sweep frequency excitation (0.1-1kHz linear sweep frequency, sweep rate 10Hz / s) is used to obtain wideband frequency response characteristics.

[0068] S3: Signal processing and feature extraction, performing adaptive filtering on the acquired signal, and calculating the current dynamic stiffness value. Thermal Influence Factors and wear and tear health index ;

[0069] Preferably, the current dynamic stiffness value Calculated using the frequency response function:

[0070]

[0071] in, For the Fourier transform of the vibration acceleration response, The Fourier transform of the excitation force signal (known or estimated by the exciter current). and For calculating the frequency band, 100-500Hz is preferred (covering the first bending mode of the principal axis). This represents the number of discrete frequency points within the frequency band.

[0072] Preferably, the adaptive filtering process employs the least mean square algorithm. The main sensor signal is the acceleration signal close to the excitation point, and the reference sensor signal is the acceleration signal far from the excitation point. The adaptive filtering algorithm is as follows:

[0073]

[0074]

[0075] in, The main sensor signal (acceleration signal near the excitation point). The reference sensor signal vector (acceleration signal far from the excitation point, 2-3 reference points can be selected) is used. For adaptive weight vectors, For adaptive step size, set to 0.01-0.05. This is an error signal;

[0076] The thermal influence factor The calculation formula is:

[0077]

[0078] in, The average temperature rise of the bearing, For the axial temperature gradient of the bearing, Integrating over time, , , These are the coefficients calibrated experimentally.

[0079] Preferably, the thermal influence coefficient , , Calibration method:

[0080] For example, set the spindle to run at a constant speed (e.g., 3000 rpm), record the temperature rise curves under different preloads, and fit the data using multiple linear regression. and , , The relationship.

[0081] 1. Average temperature rise of bearings :

[0082]

[0083] in, , , The temperature is represented by three points evenly distributed at 120° circumference around the outer ring of the bearing. The ambient temperature.

[0084] 2. Bearing axial temperature gradient :

[0085]

[0086] in , These are the front and rear end temperatures of the bearing, respectively. This refers to the axial length of the bearing.

[0087] 3. Temperature rise time integral Numerical integration is employed:

[0088]

[0089] in For the first The instantaneous temperature rise at each sampling moment, The sampling interval is denoted as .

[0090] Wear and tear health index The calculation formula is:

[0091]

[0092] in, This represents the ratio of the bearing wear characteristic frequency band energy to the baseline energy. This is the ratio of the current acoustic emission intensity to the baseline intensity. For equivalent runtime, , , , These are the weighting coefficient and the attenuation coefficient;

[0093] The weighting coefficients satisfy: ;

[0094] Attenuation coefficient

[0095] S4: Intelligent compensation calculation, based on the deviation between the current dynamic stiffness value and the target stiffness value. And integrate thermal influence factors and wear and tear health index The preload compensation amount is calculated using a nonlinear adaptive model. ;

[0096] The nonlinear adaptive model is as follows:

[0097]

[0098] in, This represents the normalized deviation between the current dynamic stiffness value and the target stiffness value. The preload from the previous moment, , , , , As adaptive parameters, the initial values ​​are obtained from calibration, and the online update rate is limited by: ;

[0099] Normalized dynamic stiffness deviation :

[0100] ,

[0101] Rate of change of preload at the previous moment: :

[0102]

[0103] in: To control the cycle.

[0104] In addition, through offline experiments, The classic range of values ​​for adaptive parameters is: Minimizes the error with the actual required compensation force. , , , , ,

[0105] S5: Graded drive adjustment, based on the preload compensation amount. The graded drive hydraulic preload actuator dynamically adjusts the preload of the spindle bearing, wherein the single adjustment range does not exceed 10% of the maximum preload.

[0106] The tiered drive includes: a fast compensation loop with a response time of less than 100 milliseconds, which effectively suppresses the instantaneous stiffness decrease caused by processing shocks and sudden load changes, improving dynamic stability; based on the current dynamic stiffness deviation... Proportional-integral-derivative control is performed; preferably, the control algorithm uses discrete PID.

[0107]

[0108] in , The sampling period is , .

[0109] A slow optimization loop with a response time of 1-10 seconds is used, employing feedforward compensation and feedback optimization based on the thermal influence factor (FT) and wear health index (H_w). Preferably, the control algorithm is a feedforward-feedback composite control.

[0110]

[0111] in For disturbance estimation (based on load current and speed prediction).

[0112] The protective constraint module ensures that the preload adjustment rate does not exceed 50 N / s, and the cumulative adjustment amount does not exceed 80% of the bearing's ultimate load. More specifically:

[0113] Preload adjustment rate: ;

[0114] Single adjustment range: ;

[0115] Cumulative adjustment amount: , ( (This refers to the bearing's ultimate load).

[0116] Triple protective constraints prevent overload damage and improve bearing mean time between failures (MTBF) by 50%–100%.

[0117] S6: Iterative verification and convergence, repeat steps S2-S3 to obtain the compensated spindle dynamic stiffness value. If the deviation from the target stiffness value does not fall within the preset tolerance range, then repeat steps S4-S5 for iterative compensation until the convergence criterion is met or the maximum number of iterations is reached.

[0118] Preferably, the convergence criterion includes any of the following conditions:

[0119] <1% of the target stiffness value;

[0120] Two consecutive preload adjustments <5N;

[0121] The number of iterations reached the preset maximum of 5.

[0122] S7: Long-term performance tracking and recalibration. During long-term operation, monitor the trend changes in spindle dynamic stiffness and wear health index. When the recalibration trigger condition is met, re-execute step S1 to update the preload-dynamic stiffness mapping relationship. Preferably, the recalibration trigger condition includes any of the following conditions:

[0123] The 30-day sliding average change in the dynamic stiffness of the spindle exceeded 8% of the initial value; wear health index Decreased to below 0.7; the correlation coefficient of the temperature-stiffness relationship curve <0.9; Cumulative running time exceeds the preset scheduled maintenance cycle;

[0124] More specifically, after each online measurement is completed in steps S2-S3, the system automatically calculates the average bearing temperature obtained in this measurement. and the corresponding spindle dynamic stiffness value As a data point The data is stored in a historical database. For every N valid data points accumulated (e.g., N=100), or after a fixed running time (e.g., every 24 hours), the system calculates the Pearson correlation coefficient based on the most recent M sets of data (e.g., M=200).

[0125]

[0126] in and These are the average values ​​of temperature and dynamic stiffness in this dataset, respectively. When the spindle performance degradation is mainly caused by linear thermal deformation, It usually remains above 0.9; when When the value remains below 0.9, it indicates that factors such as bearing wear and nonlinear thermal effects have become dominant, and the original preload-dynamic stiffness mapping relationship has deviated significantly, necessitating recalibration. Predictive maintenance mechanisms provide early warnings through the wear health index and temperature-stiffness correlation coefficient, reducing unplanned downtime by 80% and extending overhaul cycles by 30-50%.

[0127] The present invention provides a machining spindle intelligent preload control method and system, which achieves dynamic intelligent adjustment of spindle preload through the combination of multimodal perception, nonlinear adaptive model, hierarchical drive control and long-term tracking recalibration, bringing significant benefits at multiple levels;

[0128] 1. By monitoring the dynamic stiffness of the spindle in real time and intelligently compensating for it, the spindle is always kept in the optimal stiffness working range, the radial runout is reduced by 20-40%, the axial runout is reduced by 30-50%, and the online measurement and compensation of dynamic stiffness increases the dynamic stiffness value of the spindle by 15-25%, thus achieving better vibration resistance under the same preload conditions.

[0129] 2. Optimized preload reduces bearing frictional heat and fretting wear, lowering bearing temperature rise by 15-25% and wear rate by 30-40%. Optimized preload also reduces frictional torque, reducing spindle no-load power by 10-15%, resulting in significant energy savings.

[0130] 3. The automatic recalibration mechanism extends the recalibration interval from a fixed period of 3-6 months to 6-12 months, reducing the frequency of manual calibration by 80-90%. Digital health reports and maintenance recommendations provide data support for equipment management and reduce reliance on highly skilled commissioning personnel.

[0131] Based on the above control method, this invention also discloses a machining spindle control system, including a multimodal sensing unit comprising at least three triaxial accelerometers (not shown), at least three temperature sensors (not shown), and at least one broadband acoustic emission sensor (not shown) arranged at the bearing housing. The multiple triaxial accelerometers, temperature sensors, and broadband acoustic emission sensors arranged at the bearing housing achieve synchronous, high-precision acquisition of spindle vibration, temperature, and acoustic emission signals, improving signal integrity and reliability by 40-50% compared to traditional single-point sensors.

[0132] The signal processing and feature extraction module is connected to the multimodal sensing unit and is used to adaptively filter the acquired signal and calculate the current dynamic stiffness value, thermal influence factor and wear health index.

[0133] Adaptive filtering (LMS algorithm) can effectively suppress background noise, improve the signal-to-noise ratio and dynamic stiffness calculation bandwidth, and optimize the first-order bending mode of the main shaft with a calculation error of <5%.

[0134] A nonlinear adaptive controller, connected to the signal processing and feature extraction module, calculates the preload compensation amount through a nonlinear adaptive model based on the nonlinear mapping relationship between preload and dynamic stiffness and the target stiffness range, integrating the current dynamic stiffness deviation, thermal influence factor, and wear health index.

[0135] A preload actuator includes a housing, which has at least one hydraulic channel penetrating the space inside and outside the housing.

[0136] The preload ring is connected to the inner side of the housing through a two-way sealing structure. It has a flange on its outer side, which abuts against the outer ring of the main shaft bearing. Its inner side corresponds to the output end of the oil pressure channel, forming a closed oil pressure chamber.

[0137] Several elastic elements are disposed between the inner side of the preload ring and the housing to provide initial preload force under zero oil pressure conditions;

[0138] A hydraulic servo system is used to provide controllable hydraulic pressure to the hydraulic channels;

[0139] The oil temperature sensor, integrated into the oil circuit of the housing, is used to monitor the oil temperature;

[0140] The oil pressure compensation module corrects the oil pressure-preload mapping relationship based on the oil temperature sensor signal.

[0141] In the initial state, whether the spindle is static or dynamic, a forward thrust is applied to the inner side of the preload ring 4 through several elastic elements 5. The flange 41 abuts against the outer ring 31 to provide basic preload force to compensate for the error clearance in the bearing 3, thereby optimizing the accuracy of the spindle during normal operation and reducing noise. When the working state is switched and further control of the spindle speed and improvement of working accuracy are required, the medium is input into the hydraulic channel 11 through an external pressurization device. Since the preload ring 4 is sealed to the inner side of the housing 1, the medium cannot flow. Therefore, the pressure inside the hydraulic channel 11 gradually increases. When the pressure reaches a certain value, the medium still cannot break through the sealing structure, but it will push the preload ring 4 to have a tendency to move forward, and this will be converted into force applied to the outer ring 3. Under the pressure of 1, the inner ring 32 is limited and cannot move under this state, and then the outer ring 31 squeezes the ball body 33 tighter to increase the preload. Preferably, the sealing structure is that the housing 1 includes a first space and a second space, with a first step 12 between them. The preload ring 4 includes a first contact part 42 and a second contact part 43, with a second step 44 between them. The first contact part 42 abuts against the side wall of the first space, and a first seal 45 is provided between them. The second contact part 43 abuts against the side wall of the second space, and a second seal 46 is provided between them. From the cross-section, the preload ring 4 has a "Z" shaped structure. With such a structure, both the housing 1 and the preload ring 4 are easier to process.

[0142] It should be noted that multiple hydraulic channels 11 can be set in the housing 1, forming multiple output ends to simultaneously apply thrust to the preload ring 4, or only one hydraulic channel 11 can be set, with only one output end applying pressure to the inner side of the preload ring 4. After the preload ring 4 is pushed from this output end, a gap will be generated between the preload ring 4 and the inner wall of the housing 1, and then gas will fill the gap and apply the same thrust to the inner side of the preload ring 4, so that the preload ring 4 and the inner wall of the housing 1 form a stable pressure chamber. When it is necessary to switch to the initial state, the outer ring 31 will automatically reset after depressurization, and the preload ring 4 will be pushed back. However, under the action of the spring 5, the preload ring 4 will still remain in contact with the outer ring 31.

[0143] In this embodiment, the bearing 3 is a ball bearing, and there are at least two bearings 3 that abut against each other. The abutment referred to here is between the outer rings 31 and between the inner rings 32, and is unrelated to the ball body 33. When the shaft 2 rotates, the two inner rings 32 rotate synchronously. Therefore, the abutting structure will not affect the normal rotation of the shaft 2. Moreover, using more than two bearings is more conducive to the center alignment of the shaft 2. Furthermore, by using the mutual abutting method, when one outer ring 31 is squeezed, the other outer ring 31 will also be subjected to force, thereby increasing the preload of both bearings 3 simultaneously.

[0144] It also includes a limiting structure for fixing the inner ring 32 of the bearing 3; more specifically, the limiting structure includes a first limiting member 61 fixed at the end of the housing 1 and a bushing 62 sleeved on the rotating shaft 2. One end of the bushing 62 abuts against one side of the inner ring 32 of the bearing 3, and the other side of the inner ring 32 of the bearing 3 abuts against the second limiting member 63. Generally, a bearing 3 is also provided at the other end of the housing 1. The bushing 62 is located between the inner rings 32 of the two bearings 3 and mutually limits their contact. It also includes a second limiting member 63. The second limiting member 63 is located on the side away from the preload ring 4 and corresponds to the outer ring 31 of the bearing 3. The second limiting member 63 and the outer ring 31 of the bearing 3 have a movable gap to prevent the outer ring 31 from moving excessively and damaging the bearing 3.

[0145] It also includes a safety monitoring and fault tolerance module, which monitors the status of each component of the system in real time and automatically switches to a safe mode when a fault is detected.

[0146] The security monitoring and fault tolerance module includes:

[0147] The sensor health monitoring unit detects the effectiveness and consistency of signals from each sensor in real time.

[0148] Hydraulic system fault diagnosis unit monitors hydraulic pressure build-up time, pressure stability, and signs of leakage;

[0149] When a serious fault is detected, the degradation strategy execution unit automatically switches to the fixed preload mode and maintains the most recent effective compensation value.

[0150] The nonlinear adaptive controller adopts a model reference adaptive control architecture, and the reference model is:

[0151]

[0152] in, For nominal dynamic stiffness, For transmission delay, , It is a time constant. For the Laplace operator.

[0153] It also includes a predictive scheduling module, which predicts the processing interval length based on the CNC machining program. If the current interval time is insufficient to complete the full compensation process, the compensation is postponed to the next processing interval that meets the time requirements.

[0154] This invention's machining spindle control system integrates five core modules: multimodal sensing, intelligent decision-making, precise execution, safety tolerance, and predictive scheduling. This integration achieves fully automated closed-loop intelligent control of spindle preload. The system not only significantly improves machining accuracy, equipment reliability, and service life, but also achieves a fundamental shift from "passive maintenance" to "proactive health management" through predictive maintenance and intelligent scheduling. This solution boasts high technological maturity, and its modular design facilitates widespread adoption, providing crucial technical support for the transformation and upgrading of intelligent manufacturing equipment.

[0155] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

[0156] Furthermore, it should be understood that although this specification describes embodiments, not every embodiment contains only one independent technical solution. This narrative style of the specification is merely for clarity. Those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

Claims

1. A method for intelligent preload control of a machining spindle, characterized in that: Includes the following steps: S1: Under healthy spindle conditions, apply multiple known preloads via a preload actuator, measure the spindle dynamic stiffness under each preload, establish a preload-dynamic stiffness mapping relationship, and determine the target stiffness range. ; S2: During machining breaks or when the spindle is idle, apply a small excitation signal to the spindle system and simultaneously collect multi-axis vibration acceleration signals, multi-point temperature signals, and acoustic emission signals of the spindle system. S3: Perform adaptive filtering on the acquired signal and calculate the current dynamic stiffness value. Thermal Influence Factors and wear and tear health index ; S4: Based on the deviation between the current dynamic stiffness value and the target stiffness value And integrate thermal influence factors and wear and tear health index The preload compensation amount is calculated using a nonlinear adaptive model. ; S5: Based on the preload compensation amount The graded drive hydraulic preload actuator dynamically adjusts the preload of the spindle bearing, wherein the adjustment range in a single instance does not exceed 10% of the maximum preload. S6: Repeat steps S2-S3 to obtain the compensated spindle dynamic stiffness value. If the deviation from the target stiffness value does not fall within the preset tolerance range, then repeat steps S4-S5 for iterative compensation until the convergence criterion is met or the maximum number of iterations is reached.

2. The control method according to claim 1, characterized in that: It also includes step S7: During long-term operation, monitor the trend change of the spindle dynamic stiffness and the wear health index. When the recalibration trigger condition is met, re-execute step S1 and update the preload-dynamic stiffness mapping relationship. The recalibration trigger condition includes any of the following conditions: The 30-day sliding average change in the dynamic stiffness of the main shaft exceeds 8% of the initial value; The wear and tear health index (Hw) drops below 0.7; The correlation coefficient of the temperature-stiffness relationship curve is less than 0.9; The cumulative running time exceeds the preset scheduled maintenance cycle.

3. The control method according to claim 1, characterized in that: The micro-amplitude excitation signal mentioned in step S2 is generated by a piezoelectric ceramic vibrator, an electromagnetic vibrator, or a controlled spindle motor torque ripple integrated into the spindle system. The excitation frequency range is 0.1-2kHz, and the amplitude is less than 5μm. The adaptive filtering process described in step S3 uses the least mean square algorithm. The main sensor signal is the acceleration signal close to the excitation point, and the reference sensor signal is the acceleration signal far from the excitation point. The adaptive weight update step size μ is set to 0.01-0.

05.

4. The control method according to claim 1, characterized in that: The thermal influence factor mentioned in step S3 The calculation formula is: in, The average temperature rise of the bearing, For the axial temperature gradient of the bearing, Integrating over time, , , These are the coefficients calibrated experimentally.

5. The control method according to claim 1, characterized in that: The wear and health index mentioned in step S3 The calculation formula is: in, This represents the ratio of the bearing wear characteristic frequency band energy to the baseline energy. This is the ratio of the current acoustic emission intensity to the baseline intensity. For equivalent runtime, , , , These are the weighting coefficient and the attenuation coefficient.

6. The control method according to claim 1, characterized in that: The nonlinear adaptive model mentioned in step S4 is: in, This represents the normalized deviation between the current dynamic stiffness value and the target stiffness value. The preload from the previous moment, , , , , As adaptive parameters, the initial values ​​are obtained from calibration, and the online update rate is limited by: .

7. The control method according to claim 1, characterized in that: The hierarchical driving described in step S5 includes: Fast compensation loop with a response time of less than 100 milliseconds, based on the current dynamic stiffness deviation. K is adjusted using proportional-integral-derivative methods; Slow optimization loop with a response time of 1-10 seconds, based on thermal influence factor. and wear and tear health index Perform feedforward compensation and feedback optimization; The protective constraint module ensures that the preload adjustment rate does not exceed 50 N / s and the cumulative adjustment amount does not exceed 80% of the bearing's ultimate load.

8. The control method according to claim 1, characterized in that: The convergence criterion in step S6 includes any of the following conditions: <1% of the target stiffness value; Two consecutive preload adjustments <5N; The number of iterations reached the preset maximum of 5.

9. A machining spindle control system, characterized in that: It includes a multimodal sensing unit, comprising at least three triaxial accelerometers, at least three temperature sensors, and at least one broadband acoustic emission sensor arranged at the bearing housing; The signal processing and feature extraction module is connected to the multimodal sensing unit and is used to adaptively filter the acquired signal and calculate the current dynamic stiffness value, thermal influence factor and wear health index. A nonlinear adaptive controller, connected to the signal processing and feature extraction module, calculates the preload compensation amount through a nonlinear adaptive model based on the nonlinear mapping relationship between preload and dynamic stiffness and the target stiffness range, integrating the current dynamic stiffness deviation, thermal influence factor, and wear health index. A preload actuator includes a housing, which has at least one hydraulic channel penetrating the space inside and outside the housing. The preload ring is connected to the inner side of the housing through a two-way sealing structure. It has a flange on its outer side, which abuts against the outer ring of the main shaft bearing. Its inner side corresponds to the output end of the oil pressure channel, forming a closed oil pressure chamber. Several elastic elements are disposed between the inner side of the preload ring and the housing to provide initial preload force under zero oil pressure conditions; A hydraulic servo system is used to provide controllable hydraulic pressure to the hydraulic channels; The oil temperature sensor, integrated into the oil circuit of the housing, is used to monitor the oil temperature; The oil pressure compensation module corrects the oil pressure-preload mapping relationship based on the oil temperature sensor signal.

10. The control system according to claim 9, characterized in that: It also includes a safety monitoring and fault tolerance module, which monitors the status of each component of the system in real time and automatically switches to a safe mode when a fault is detected. The safety monitoring and fault tolerance module includes: The sensor health monitoring unit detects the effectiveness and consistency of signals from each sensor in real time. Hydraulic system fault diagnosis unit monitors hydraulic pressure build-up time, pressure stability, and signs of leakage; When a serious fault is detected, the degradation strategy execution unit automatically switches to the fixed preload mode and maintains the most recent effective compensation value. It also includes a predictive scheduling module, which predicts the processing interval length based on the CNC machining program. If the current interval time is insufficient to complete the full compensation process, the compensation is postponed to the next processing interval that meets the time requirements.

Images