CNC machine tool machining parameter self-optimization system based on cross-frequency interference coupling control

Abstract

This invention relates to the field of intelligent control technology for CNC machine tools, specifically to a self-optimization system for machining parameters of CNC machine tools based on cross-frequency interference coupling control. The system includes a cross-frequency interference event determination unit and a dynamic coupling control unit. In this invention, the cross-frequency interference event determination unit calculates the cross-frequency interference intensity factor over M consecutive control cycles by comparing the increase in the characteristic value of the blade frequency band with the decrease in the characteristic value of the tool frequency band, and generates a coordinated control command. Upon receiving the command, the dynamic coupling control unit executes a step-down adjustment of the feed rate, synchronously calculates and injects spindle torque compensation current to stabilize the rotational speed, and simultaneously samples transient jitter through a speed micro-oscillation suppression submodule and generates an anti-phase pulse to cancel it out. When the characteristic value of the blade frequency band fluctuates, the tool frequency band returns to the reference, and the interference intensity factor reaches the safe zone, the control is terminated, thereby improving machining accuracy.

Description

Technical Field

[0001] This invention relates to the field of intelligent control technology for CNC machine tools, and more specifically, to a self-optimization system for machining parameters of CNC machine tools based on cross-frequency interference coupling control. Background Technology

[0002] Intelligent control of CNC machine tools is an important technology. In the field of high-end equipment manufacturing, the machining accuracy and efficiency of CNC machine tools directly determine the quality of parts and production benefits. The optimization of machining parameters is the core link. By adjusting parameters in real time according to the machining conditions, it is possible to avoid tool overload and wear, reduce the scrap rate of parts, and improve machining efficiency while ensuring accuracy. It is a key support for promoting the upgrade of CNC machine tools from fixed parameter machining to adaptive intelligent machining, especially suitable for the high-precision machining needs of thin-walled and easily vibrating parts such as blades.

[0003] However, existing CNC machine tool machining parameter control technology suffers from core problems such as lag in cross-frequency interference response and poor control coordination. This problem stems from three practical limitations:

[0004] First, the lack of a precise cross-frequency interference detection mechanism, relying solely on single-band characteristic value monitoring, fails to promptly identify abnormal correlations between blade vibration frequencies and tool operating frequencies, making it difficult to trigger timely control when interference events occur. Second, parameter control lacks dynamic coupling; spindle speed is not synchronously compensated when adjusting feed rate, easily leading to spindle load fluctuations due to sudden feed rate changes, resulting in speed deviations from the target range. Furthermore, transient speed jitter during feed rate step transitions is not addressed. Third, stable state determination is incomplete, relying solely on a single parameter to determine whether to terminate control, without considering tool status, interference intensity, and other multi-dimensional factors. This easily leads to premature or excessive control. These problems create a chain reaction, meaning that untimely response to cross-frequency interference will exacerbate the problem. Blade vibration and tool wear can lead to dimensional deviations in parts. Rotation speed fluctuations and transient jitter can damage the surface quality of the machined parts, resulting in scratches or precision errors. Inaccurate stability judgment can either allow interference to continue to affect the machining process or lead to over-tuning, reducing production efficiency. It may also shorten the service life of the spindle and tools, increasing equipment maintenance costs. Ultimately, existing technologies cannot meet the requirements of high-precision parts machining for rapid anti-interference, parameter coordinated control, and accurate stability judgment. There is an urgent need for a self-optimization scheme for machining parameters that can accurately identify cross-frequency interference, dynamically couple feed rate and rotation speed, and comprehensively determine the stable state. To solve this technical problem, we provide a self-optimization system for CNC machine tool machining parameters based on cross-frequency interference coupling control. Summary of the Invention

[0005] The purpose of this invention is to provide a self-optimization system for CNC machine tool machining parameters based on cross-frequency interference coupling control, so as to solve the problems mentioned in the background art.

[0006] 1. Due to the inaccurate judgment and delayed response of cross-frequency interference, it is impossible to identify the abnormal correlation between the blade and the tool frequency band in a timely manner. Therefore, this case uses a cross-frequency interference event judgment unit to compare the amplitude difference of the characteristic values ​​of the blade and the tool frequency band, calculate the interference intensity factor, and combine it with the tool wear correction threshold to identify cross-frequency interference in a timely manner and generate coordinated control commands.

[0007] 2. Due to poor coordination of parameter control, adjusting the feed rate causes speed fluctuations and transient jitter. Therefore, this case uses a dynamic coupling control unit to synchronously and stepwise reduce the feed rate and compensate for the spindle torque current. The speed micro-oscillation suppression submodule is added to counteract the jitter and stabilize the spindle speed.

[0008] To achieve the above objectives, a self-optimization system for CNC machine tool machining parameters based on cross-frequency interference coupling control is provided, including:

[0009] The cross-frequency interference event determination unit is used to generate a coordinated control command when the increase in the characteristic value of the blade frequency band is greater than the decrease in the characteristic value of the tool frequency band and the difference exceeds the tolerance threshold. The dynamic coupling control unit receives the coordinated control command and executes the feed rate step reduction and speed compensation mechanism.

[0010] When performing a single-stage feed rate reduction operation, the spindle load change caused by the sudden change in feed rate is captured in real time. Based on the spindle load change, the spindle torque compensation current is dynamically calculated and injected within the same control cycle to compress the actual speed fluctuation to a preset floating range. This is performed synchronously after each feed rate reduction is completed.

[0011] (a) Extract the current blade frequency band feature value fallback rate;

[0012] (b) Dynamically reconstruct subsequent control parameters based on the rate of decline:

[0013] If the fall rate is lower than the first threshold, the next level of feed rate is increased synchronously, the step size is reduced, and the torque compensation current peak is increased proportionally.

[0014] If the rate of decline is higher than the second threshold, the feed rate of the next stage is reduced in step size and the torque compensation current intensity is reduced proportionally. During the transition gap between two adjacent stages of reduction, the speed micro-oscillation suppression submodule is activated to detect the transient speed jitter caused by the step switching of feed rate in real time, and generates a pulse compensation current that is opposite to the jitter waveform to cancel it. When the blade frequency band characteristic value is stable at the safety threshold for a continuous preset control period, the feed rate reduction sequence is terminated and the torque compensation is released.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0016] 1. By using the cross-frequency interference event determination unit, the interference intensity factor is calculated based on the increment / decrease ratio of the blade and tool frequency band characteristic values ​​within M consecutive control cycles. Combined with the tool wear state coefficient, the tolerance threshold is corrected in real time. After multi-cycle determination, the command is generated to achieve the technical effect of accurately and timely identifying cross-frequency interference. This solves the problem of existing single-frequency band monitoring and response lag. It has the advantages of avoiding interference from aggravating blade vibration and tool wear, and reducing part size deviation.

[0017] 2. When the feed rate is stepped down through the dynamic coupling control unit, the spindle torque compensation current injection is calculated synchronously. Combined with the speed micro-oscillation suppression submodule, transient jitter is sampled and an anti-phase pulse is generated to cancel it out. This achieves the technical effect of stabilizing the spindle speed and eliminating transient jitter, solving the problems of poor coordination of existing parameter control, speed fluctuation and jitter damaging the surface quality of the machined surface. It has the advantages of ensuring machining accuracy and avoiding surface scratches or accuracy deviations.

[0018] 3. By using a multi-condition convergence strategy, the fluctuation of blade frequency band, the return range of tool frequency band to the reference range, and the decay of interference intensity factor to the safe zone are monitored simultaneously to trigger the termination control signal, thereby achieving the technical effect of accurately determining the stable state. This solves the problems of existing single parameter judgment, premature or excessive control, and has the advantages of avoiding efficiency reduction due to excessive control or interference due to insufficient control, as well as extending the service life of the spindle and tool. Attached Figure Description

[0019] Figure 1 This is an overall block diagram of the present invention.

[0020] The meanings of the labels in the diagram are as follows:

[0021] 1. Cross-frequency interference event determination unit; 2. Dynamic coupling control unit. Detailed Implementation

[0022] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] This invention provides a self-optimization system for CNC machine tool machining parameters based on cross-frequency interference coupling control. Please refer to [link / reference]. Figure 1 As shown, it includes:

[0024] Cross-frequency interference event determination unit 1 is used to generate a coordinated control command when the increase in the characteristic value of the blade frequency band is greater than the decrease in the characteristic value of the tool frequency band and the difference exceeds the tolerance threshold. Dynamic coupling control unit 2 receives the coordinated control command. When determining the relationship between the increase in the characteristic value of the blade frequency band and the decrease in the characteristic value of the tool frequency band, cross-frequency interference event determination unit 1 establishes a dynamic compensation coefficient calculation mechanism, which specifically includes:

[0025] The ratio of the absolute value of the increment of the blade frequency band characteristic value to the absolute value of the decrease of the tool frequency band characteristic value within M consecutive control cycles is used as the cross-frequency interference intensity factor. When the cross-frequency interference intensity factor exceeds the upper limit of the dynamically adjusted tolerance threshold and continues to reach the preset number of cycles, a coordinated control command is generated. The upper limit of the tolerance threshold is based on the tool wear state coefficient, which is corrected in real time. The tool wear state coefficient is obtained through the correlation analysis of the cutting force harmonic components and the tool fundamental frequency offset. The specific implementation method is as follows:

[0026] When machining thin-walled parts such as blades on CNC machine tools, cross-frequency interference often originates from abnormal coupling between the blade vibration frequency band and the tool working frequency band. Increased blade vibration can easily lead to out-of-tolerance machining surface finish, while increased tool load may accelerate wear. Therefore, the cross-frequency interference event judgment unit 1 needs to accurately identify the changing relationship between the characteristic values ​​of the two frequency bands to avoid misjudgments caused by monitoring a single frequency band. To this end, a dynamic compensation coefficient calculation mechanism has been established. First, the definition of the core parameters needs to be clarified. The blade frequency band characteristic value refers to the amplitude of a specific frequency band signal related to blade machining vibration, which is collected by the spindle vibration sensor. An increase in its value usually means that the blade vibration is intensified, which may cause machining chatter. The tool frequency band characteristic value is the value corresponding to the inherent characteristics of the tool during tool rotation or cutting. The frequency signal amplitude, specifically the increment, refers to the difference between the blade frequency band characteristic value and the value within two consecutive control cycles. The absolute value of the increment is the non-negative value of this difference, directly reflecting the rise of the blade frequency band characteristic value. The decrease is the difference between the previous cycle value and the next cycle value of the tool frequency band characteristic value; the absolute value of the decrease reflects the fall of the tool frequency band characteristic value. The control cycle is the basic time unit for the system to sample, calculate, and adjust machining parameters. This system sets it to 10ms. This duration ensures real-time response to changes in frequency band characteristic values ​​without causing data redundancy due to over-sampling. Based on the above parameters, the system first calculates the cross-frequency interference intensity factor. This factor is the core quantitative indicator for measuring the severity of cross-frequency interference, and its design logic is as follows:

[0027] When cross-frequency interference occurs, the rise in blade frequency band characteristic values ​​and the fall in tool frequency band characteristic values ​​are often correlated. Looking at the change in only one frequency band can easily lead to misjudgment; therefore, it is necessary to establish a correlation through the cumulative changes of both. Specifically, in the calculation, select M consecutive control cycles, where M is 5, i.e., the cumulative change within 50ms, balancing real-time performance and anti-interference capabilities. First, accumulate the absolute value of the increase in blade frequency band characteristic values ​​within these M cycles, and then compare the absolute value of the decrease in tool frequency band characteristic values ​​during the same period. The larger this value, the more significant the increase in blade frequency band value relative to the decrease in tool frequency band value, indicating cross-frequency interference intensity. The more severe the interference, the more necessary it is to determine a dynamically adjusted tolerance threshold. This tolerance threshold is the critical value used to determine whether cross-frequency interference control is triggered, and it has an upper and lower limit. The upper tolerance threshold we are focusing on here is the maximum allowable cross-frequency interference intensity factor. Exceeding this value indicates that the interference may affect machining accuracy. The initial value is calibrated to 1.2 based on a large amount of normal machining data. This means that caution is needed when the blade frequency band rises by more than 20% compared to the tool frequency band fall. However, tool wear will cause the tool frequency band characteristic value to naturally decrease. If a fixed threshold is still used, misjudgment is likely. Therefore, the upper tolerance threshold needs to be adjusted in real time based on the tool wear state coefficient. The tool wear state coefficient is obtained through correlation analysis between the cutting force harmonic components and the tool's fundamental frequency offset. The cutting force harmonic components are the high-frequency components other than the fundamental frequency after the cutting force signal has undergone Fourier transform. As tool wear intensifies, the non-uniformity of the cutting force increases, and the amplitude of the harmonic components significantly increases. The tool's fundamental frequency offset is the difference between the tool's actual natural frequency and its initial natural frequency. A new tool has an initial natural frequency of 1000Hz; after wear, due to changes in mass distribution, the fundamental frequency may shift to 950Hz, an offset of 50Hz. By calculating the Pearson correlation coefficient between the two, this coefficient becomes the tool wear state coefficient. The more severe the wear, the higher the tool wear state coefficient. The stronger the correlation between the cutting force harmonic components and the fundamental frequency offset, the closer the coefficient is to 1. The real-time correction logic for the upper limit of the tolerance threshold is: Corrected upper limit = Initial upper limit × (1 + 0.5 × Tool wear state coefficient). This formula is designed to take into account that the more severe the tool wear, the greater the natural decrease in the tool frequency band characteristic value. Therefore, the upper limit of the threshold needs to be appropriately relaxed to avoid misjudgment. For example, when the tool wear state coefficient is 0.6 (moderate wear), the corrected upper limit = 1.2 × (1 + 0.5 × 0.6) = 1.56. If the coefficient is 0.9 (severe wear), the corrected upper limit = 1.2 × (1 + 0.5 × 0.9) = 1.74. The system updates the tool wear state coefficient every control cycle, synchronously correcting the upper limit of the tolerance threshold. Finally, when the cross-frequency interference intensity factor exceeds the corrected upper limit of the tolerance threshold, and this over-threshold state continues for a preset number of cycles, it is determined that the cross-frequency interference has stabilized, excluding instantaneous fluctuations. At this time, the cross-frequency interference event determination unit 1 generates a coordinated control command, which includes the current intensity factor, tool wear state coefficient, and tolerance threshold correction value, and sends it to the dynamic coupling control unit 2. This provides a precise basis for the interference level for subsequent feed rate step reduction and spindle speed compensation, ensuring that the control action neither lags behind the development of the interference nor is over-controlled due to instantaneous fluctuations.

[0028] It also implements a feed rate step reduction and speed compensation mechanism:

[0029] When performing a single-stage feed rate reduction operation, the spindle load change caused by the sudden feed rate change is captured in real time. The dynamic coupling control unit 2 captures the spindle load change caused by the sudden feed rate change in real time through embedded current loop dynamic sampling, specifically including:

[0030] At the start of the feed rate reduction command, the high-frequency spindle current acquisition module is simultaneously activated. It captures the transient changes at the leading edge of the current waveform at a sampling rate no less than a preset proportion of the control cycle frequency, obtaining the abrupt change in the integral area of ​​the current waveform. Based on this abrupt change in the integral area of ​​the current waveform and a preset spindle load-current conversion model, the spindle load change is calculated. The feed rate abrupt change is defined as a preset percentage threshold where the single-stage feed rate reduction step size exceeds the current feed reference value. Further explanation is needed:

[0031] After the cross-frequency interference event determination unit 1 generates the coordinated control command, the dynamic coupling control unit 2 needs to prioritize capturing the spindle load change caused by parameter mutations during the feed rate step reduction process. This is the core basis for subsequent calculation of the spindle torque compensation current. If the load change is not captured accurately, it will easily lead to a deviation between the compensation current and the actual requirement, which will in turn cause speed fluctuations and damage the machining accuracy. Therefore, the system achieves real-time, high-precision capture through embedded current loop dynamic sampling. The specific implementation method is as follows:

[0032] The definition of feed rate mutation needs to be clarified. It refers to a single-stage feed rate reduction step exceeding a preset percentage threshold of the current feed reference value. The current feed reference value is the feed rate when the system is in stable machining before adjustment. The preset percentage threshold is calibrated to 10% through machining experiments on 100 sets of blades of different materials. That is, when the single-stage reduction step exceeds 80mm / min (800mm / min × 10%), it is judged as a feed rate mutation. This type of mutation will disrupt the original cutting load balance of the spindle, causing a momentary change in the output torque of the spindle motor, which in turn causes current fluctuations. This change needs to be accurately captured by high-frequency sampling. At the start of the feed rate reduction command being sent to the spindle drive module through the system bus, it needs to be synchronized with the command issuance at the microsecond level to avoid sampling lag caused by time difference. The dynamic coupling control unit 2 will synchronously trigger the spindle current high-frequency acquisition module in the embedded current loop. This acquisition module is equipped with a 16-bit high-speed ADC chip, and its sampling rate needs to meet a preset ratio of not less than 5 times the control cycle frequency. The system control cycle is 10ms (corresponding to a frequency of 100Hz). Therefore, the sampling rate is set to 500Hz, meaning one current data point is collected every 2ms. This high sampling rate can fully cover the rapid change in current during the instantaneous feed rate change, avoiding the omission of key transient information due to excessively long sampling intervals. During the acquisition process, it is necessary to focus on capturing the transient change at the leading edge of the current waveform. This term specifically refers to the steep change that occurs in the initial stage of the spindle current transitioning from its original stable value to a new stable value after a sudden feed rate change. For example, when the feed rate suddenly decreases from 800mm / min to 700mm / min (step size 100mm / min, exceeding the 10% threshold), the spindle cutting load decreases instantaneously, and the spindle motor does not need to output its original torque. The current will drop rapidly from a stable 5A to 4.2A. This steep drop segment starting from 5A completes 80% of the current change within the first 20ms, which is the transient change at the leading edge of the current waveform. Its amplitude and speed directly reflect the intensity of the load change and are the core area for extracting load change information. When obtaining the instantaneous change in the integral area of ​​the current waveform from the transient change at the leading edge of the current waveform, the specific steps are as follows:

[0033] Starting from the moment the feed rate reduction command is issued, current waveform data for the following 20ms is captured to ensure complete coverage of the transient changes at the forefront. First, the average spindle current value in the 50ms before the command is issued is taken as the stable current reference value. Then, the actual current value at each sampling point within the analysis window is subtracted from this reference value to obtain the current deviation value at each moment. The average current deviation value of two adjacent sampling points is then multiplied by the sampling time interval to obtain the integral area within that interval. The integral results of all intervals are then summed, and the final value is the abrupt change in the integral area of ​​the current waveform. For example, the integral of the deviation values ​​of 10 sampling points is -1.6 A·ms. The negative sign indicates a decreasing current trend, while the absolute value reflects the cumulative degree of current change. When calculating the spindle load change based on the abrupt change in the integral area of ​​the current waveform and the preset spindle load-current conversion model, the model relationship established in the previous offline calibration must be relied upon.

[0034] This conversion model obtains a linear relationship by loading a known standard load onto the spindle, synchronously collecting the spindle current integral area change under the corresponding load, and fitting the data using the least squares method. The final model expression is ΔT=K×ΔS+ Where ΔT is the spindle load change to be solved, K is the model calibration coefficient, which was experimentally determined to be 0.008 N·m / (A·ms), representing a load change of 0.008 N·m corresponding to a current integral area change of 1 A·ms, and ΔS is the actual captured current waveform integral area change. The correction term, with a value of 0.02 N·m, is used to compensate for errors caused by fixed losses such as spindle bearing friction. Specifically, substituting ΔS = -1.6 A·ms into the model yields ΔT = 0.008 × (-1.6) + 0.02 = -0.108 N·m. The negative sign indicates that a sudden change in feed rate causes a decrease in spindle load. The absolute value of 0.108 N·m represents the actual load change. After calculation, this load change is transmitted in real-time to the torque compensation current calculation module to ensure that a compensation current matching the load change is generated within the same control cycle, preventing spindle speed fluctuations due to sudden load changes. Based on the spindle load change, the spindle torque compensation current is dynamically calculated and injected within the same control cycle. The method of dynamically calculating the spindle torque compensation current within the same control cycle by the dynamic coupling control unit 2 includes time-sequence forced constraints, specifically:

[0035] After obtaining the spindle load change at the beginning of the control cycle, the torque compensation current value is calculated and prepared for output within that cycle. The calculation process adopts the inverse solution path of the load-current conversion model. The spindle load change is input into a preset torque compensation response function, which integrates the spindle electromechanical time constant and the moment of inertia compensation coefficient, outputs the torque compensation current reference value, and simultaneously superimposes the dynamic inertia compensation component determined by the feed rate reduction. It should be further explained that:

[0036] After the dynamic coupling control unit 2 captures the spindle load change through the embedded current loop, in order to prevent the spindle speed from deviating from the target range due to this change, the calculation and output preparation of the spindle torque compensation current must be completed within the same control cycle. The timing constraint of this process is the core. It requires a fixed control cycle of 10ms. After the spindle load change is acquired at the beginning of the cycle (the first 2ms), the compensation current calculation and output parameter preparation must be completed within the remaining 8ms to ensure that the compensation action is synchronized with the load change and to avoid speed fluctuations caused by control lag. The specific implementation is as follows:

[0037] The timing details of calculation and output preparation within the control cycle are clearly defined. The 10ms control cycle is divided into a data reception segment (0-2ms), a calculation segment (2-8ms), and an output preparation segment (8-10ms). Within 0-2ms, the spindle load change is calculated, denoted as ΔT, such as the previously calculated -0.108 N·m. The negative sign indicates the reception and verification of the load decrease. From 2-8ms, the core calculation stage begins, generating a compensation current reference value through the inverse solution path of the load-current conversion model and the torque compensation response function. From 8-10ms, the reference value is converted into a digital signal recognizable by the current loop PID controller and loaded into the register of the PWM modulation module, awaiting injection into the spindle motor at the start of the next control cycle. The entire process has no timing redundancy, ensuring real-time matching between compensation and load changes. During the calculation, the inverse solution path of the load-current conversion model is crucial. The previously established forward model is ΔT = K × ΔS + ΔT is the change in spindle load, K is a calibration coefficient of 0.008 N·m / (A·ms), and ΔS is the sudden change in the current integral area. The correction term is 0.02 N·m, used to infer the load change from the current change. The reverse solution requires deriving the required current reference from the known ΔT, as detailed below:

[0038] First, remove the fixed correction term from the forward model, then calculate the net load change ΔTnet = ΔT - Substituting ΔT = -0.108 N·m, we get ΔTnet = -0.108 - 0.02 = -0.128 N·m. This value reflects the pure load change that needs to be offset by current compensation. Based on the torque constant Kt of the spindle motor, which characterizes the relationship between motor current and output torque, it has been experimentally calibrated to 0.5 N·m / A, meaning 1A of current corresponds to 0.5 N·m of torque. Calculating the basic current compensation component Ibasic = ΔTnet / Kt, and substituting the data, we get Ibasic = -0.128 / 0.5 = -0.256A. The negative sign indicates that the current needs to be reduced to match the load decrease. When the load decreases, the motor can maintain its speed without the original current. Reducing the current can prevent the speed from increasing. Therefore, the basic current compensation component needs to be reduced. The input torque compensation response function is a mathematical model used to convert changes in spindle load into corresponding torque compensation requirements and further map them to a current reference value. Its core is to integrate the mechanical and electrical characteristics of the spindle to ensure that the output current accurately offsets the impact of load changes on the rotational speed, rather than relying solely on a simple current-torque linear relationship. The spindle electromechanical time constant and moment of inertia compensation coefficient integrated into the function need to be clearly defined first: The spindle electromechanical time constant (τ) is a parameter characterizing the response speed of the spindle motor from receiving a current command to outputting the corresponding torque. It is determined by the armature resistance (0.5Ω), inductance (1mH), and moment of inertia of the spindle motor. In this system, it was measured to be 15ms through no-load experiments. The higher this value, the better. The smaller the value, the faster the motor torque response. During compensation, the current change rate needs to be appropriately reduced to avoid overshoot. The moment of inertia compensation coefficient (J) is a coefficient used to correct the influence of the spindle and load (tool, workpiece) moment of inertia on torque demand. Its unit is N·m / (rad / s²), and its value is equal to the total moment of inertia of the spindle: spindle itself 0.015 N·m / (rad / s²) + blade / workpiece 0.003 N·m / (rad / s²) + tool 0.002 N·m / (rad / s²), totaling 0.02 N·m / (rad / s²). The larger the moment of inertia, the more obvious the angular acceleration fluctuation caused by feed rate changes, and the greater the additional torque required for compensation. The specific expression for the torque compensation response function is Ir. ef = Ibasic × (1 - e^(-t / τ)) + J × α / Kt, where t is the calculation time within the current control cycle, taking the midpoint of the calculation segment as 5ms, and α is the change in spindle angular acceleration corresponding to the feed rate reduction. In the specific calculation, first calculate the exponential term e^(-t / τ) = e^(-5 / 15) ≈ 0.716, therefore (1 - e^(-t / τ)) ≈ 0.284. This term is used to correct the motor response delay, ensuring that the current output gradually approaches the target value over time, avoiding torque impact caused by sudden current changes. The second step is to calculate the change in angular acceleration α: the feed rate reduction is (current feed - feed after reduction) / control cycle = (800 - 700) mm / min ÷ 0.01s = 10000mm / min², combined with the spindle pitch (5mm / r), it is converted into angular acceleration α = (feed rate adjustment × 2π) / (60² × pitch) ≈ (10000 × 6.28) / (3600 × 5) ≈ 3.49 rad / s². The third step is to calculate the dynamic inertia compensation component Inertial = J × α / Kt = 0.02 × 3.49 / 0.5 ≈ 0.1396A. This component is used to compensate for changes in feed rate. The inertial force of the generated torque requires the fourth step to superimpose the basic current compensation component and the dynamic inertial compensation component to obtain the torque compensation current reference value Iref=(-0.256×0.284)+0.1396≈-0.0727+0.1396≈0.0669A. The final output positive current reference value indicates that although the load decreases and the current needs to be reduced, the inertial compensation needs to slightly increase the current. After combining, a compensation current of 0.067A needs to be injected to maintain the speed stability. During the 8-10ms output preparation segment of the control cycle, the system converts the calculated Iref (0.067A) into the target input value of the current loop PID controller and configures the carrier frequency (20kHz) and duty cycle range (0%~100%) of the PWM modulation module to ensure that at the start of the next control cycle, the compensation current can be accurately injected into the three-phase winding of the spindle motor through the gate drive circuit, completing the compensation current calculation and output preparation under the entire timing constraint, laying the foundation for subsequent stabilization of the spindle speed, and compressing the actual speed fluctuation to the preset floating range. The method of injecting spindle torque compensation current by the dynamic coupling control unit 2 to compress the actual speed fluctuation to the preset floating range includes a current closed-loop injection mechanism, specifically including:

[0039] The torque compensation current reference value is used to generate a PWM modulation signal via a current loop PID controller, which is then injected into the three-phase windings of the spindle motor through a gate drive circuit. Simultaneously, a speed fluctuation suppression feedback loop is established, and the spindle encoder feedback signal is acquired in real time. When the deviation of the actual speed from the target speed exceeds the floating range, the integral gain coefficient of the PID controller is automatically increased to achieve range compression. Further explanation is needed:

[0040] After completing the calculation and output preparation of the spindle torque compensation current reference value, in order to avoid deviations in the compensation current during transmission or execution due to line losses and motor load fluctuations, which would cause the spindle speed to deviate from the target range, the dynamic coupling control unit 2 constructs a closed-loop control of calculation-execution-feedback-adjustment through a current closed-loop injection mechanism to ensure that the actual speed fluctuation is compressed to the preset floating range. The specific implementation method is as follows:

[0041] The definition of the floating range needs to be clarified. It refers to the maximum allowable fluctuation range of the spindle target speed. This range is calibrated based on the machining accuracy requirements. Blade machining has high requirements for speed stability. Therefore, the floating range is set to ±0.17% of the target speed. That is, when the target speed is 3000 r / min, the floating range is 2995 r / min-3005 r / min. If the actual speed exceeds this range, it will directly affect the surface roughness of the blade. Gain adjustment needs to be initiated immediately. Then, the process of generating a PWM modulation signal from the torque compensation current reference value through the current loop PID controller is executed. The specific process is as follows:

[0042] The current-loop PID controller first receives the torque compensation current reference value generated earlier. Simultaneously, it collects the actual operating current of the motor through a Hall current sensor connected in series in the spindle motor power supply circuit. The deviation between the two is calculated as ΔI = reference value - actual current = 0.067A - 0.06A = 0.007A. A positive deviation indicates insufficient actual current, requiring an increase in output to reach the reference value. The controller performs proportional (P), integral (I), and derivative (D) calculations on the deviation value: the proportional loop is calculated with a preset proportional coefficient Kp = 2.0, and the output P = Kp × ΔI = 2.0 × 0.007A = 0.0 14A is used for rapid response to the current deviation. The integral stage is calculated with an initial integral gain Ki=0.5, and performs an integral operation on the cumulative deviation value of the past 5 sampling periods (2ms per period). The output I=Ki×(cumulative deviation×sampling period)=0.5×(0.034A×0.01s)=0.00017A·s. This is used to eliminate static deviation. The derivative stage is calculated with a derivative coefficient Kd=0.1. Based on the difference between the current deviation and the deviation of the previous period (0.006A) (0.001A), the output D=Kd×(deviation difference / sampling period)=0.1×(0.001A / 0.001A). The first step is to set the amplitude of the PWM modulation control signal to 0.0725A. The second step is to set the amplitude of the PWM signal to 0.0725A. The third step is to set the amplitude of the PWM signal to 0.0725A. The fourth step is to generate the PWM signal using the triangular carrier comparison method: the controller has a built-in 20kHz triangular carrier wave (peak value 0.5A, valley value 0A, period 50μs). The total control quantity is compared with the triangular wave in real time. When the control quantity is greater than the carrier value, the PWM output is high (3.3V), and when it is less, it is low (0V). This generates a PWM modulation signal with a duty cycle of approximately 14.5% (0.0725A / 0.5A). The duration of the high level of this signal determines the current energy injected into the motor per unit cycle. The larger the duty cycle, the stronger the injected current. The generated PWM modulation signal is injected into the three-phase windings of the spindle motor through the gate drive circuit. The gate drive circuit consists of an optocoupler isolation module and a power amplifier module. The isolation module prevents the high-voltage signal from the motor from interfering with the control circuit, and the amplifier module...The 3V signal is amplified to 12V to drive the IGBT switch. By controlling the energizing duration and frequency of the windings, the motor output torque is adjusted, thereby stabilizing the spindle speed. Simultaneously, to monitor speed fluctuations in real time and dynamically adjust the compensation strategy, the system needs to establish a speed fluctuation suppression feedback loop. This feedback loop is a closed-loop system that uses the actual speed deviation as its core and adjusts the PID parameters of the current loop in reverse. Its function is to enhance the system's ability to eliminate static deviations when speed fluctuations exceed the floating range by increasing the PID integral gain coefficient, thus preventing the speed from deviating from the target value for a long period. The establishment of the speed fluctuation suppression feedback loop relies on the coordinated work of three types of core components: The first type is the spindle speed acquisition component, namely an incremental photoelectric encoder installed at the end of the spindle, with a resolution of 1024 lines / revolution and a sampling frequency of 1kHz, ensuring that 1024 pulse signals are acquired per revolution, and the speed calculation accuracy reaches ±1r / min. The encoder outputs a pulse signal proportional to the spindle speed in real time. The second type is signal processing. The first component filters the pulse signal output by the encoder using a 5-point moving average filter to eliminate pulse jitter and speed conversion caused by motor vibration. The actual speed is calculated using the formula: speed = 60 / (pulse period × encoder lines × reduction ratio) (reduction ratio is 1:1, pulse period is 0.204ms). For example, the calculated actual speed is 3008 r / min. The second component is a deviation feedback and parameter adjustment component. This component first calculates the deviation between the actual speed and the target speed (3008 r / min - 3000 r / min = 8 r / min). Then, it compares the deviation with the floating range (2995 r / min - 3005 r / min). If the deviation exceeds the range (e.g., 8 r / min > 5 r / min), the PID integral gain coefficient adjustment module is triggered; otherwise, the original parameters are maintained. When the deviation of the actual speed from the target speed exceeds the floating range, the system automatically increases the integral gain coefficient of the PID controller. The specific adjustment logic is as follows:

[0043] First, set the initial value of the integral gain Ki=0.5. Based on the no-load test calibration, ensure that the speed is stable and the maximum upper limit Kimax=2.0 when there is no load fluctuation, so as to avoid excessive speed overshoot due to excessive gain. Then calculate the deviation excess ratio, that is, (actual deviation value - upper limit of the floating range) / upper limit of the floating range × 100%. For example, if the actual deviation is 8r / min and the upper limit of the floating range is 5r / min, the excess ratio = (8-5) / 5×100%=60%. Then adjust according to the formula of enhanced Ki=initial Ki+(excess ratio×0.3) (0.3 is the gain adjustment coefficient, which is determined by 100 sets of fluctuation tests to balance the adjustment speed and stability). Substituting the data, we get enhanced Ki=0.5+(60%×0.3)=0.5+0.18=0.68. If the enhanced Ki does not exceed Kimax, the new Ki is immediately loaded into the integral link of the current loop PID controller. If it exceeds the upper limit, Kimax is used as the final value. After adjustment, the integral term of the PID controller is more effective at eliminating static deviations. For example, with the original Ki=0.5, it took 10 control cycles (100ms) to eliminate an 8r / min deviation. With the Ki value increased to 0.68, only 6 control cycles (60ms) are needed to bring the speed back below 3005r / min, achieving rapid compression of speed fluctuations. Throughout the entire current closed-loop injection mechanism, the generation of the PWM modulation signal and the monitoring of the speed fluctuation suppression feedback loop are synchronized, with a sampling and feedback interval of 2ms. This ensures that the compensation current is always dynamically matched to the speed deviation. When the speed returns to the floating range, the system gradually reduces the integral gain coefficient back to its initial value, preventing continuous high gain from causing the speed to oscillate around the target value. Ultimately, this achieves long-term stability of the spindle speed within the range of 3000r / min ± 5r / min, meeting the speed stability requirements of high-precision blade machining.

[0044] Each time the feed rate is reduced by one level, execute the following synchronously:

[0045] (a) Extracting the current blade frequency band feature value fall rate: The dynamic coupling control unit 2 uses windowed differential analysis to extract the current blade frequency band feature value fall rate, specifically including:

[0046] After each feed rate reduction operation is completed, a monitoring window of fixed duration is opened. Within this window, the blade frequency band feature value sequence is collected at equal height time intervals. The blade frequency band feature value sequence is then subjected to least squares linear fitting, and the absolute value of the slope of the fitted line is defined as the fall rate. The length of the monitoring window is dynamically adjusted according to the current spindle speed to ensure coverage of at least three complete blade vibration cycles. It should be further noted that:

[0047] After the dynamic coupling control unit 2 compresses the spindle speed fluctuation to the preset floating range through the current closed-loop injection mechanism, in order to determine the effect of feed rate reduction on blade vibration suppression, i.e., whether the blade frequency band characteristic value gradually decreases as the feed rate decreases, the windowed differential analysis method is used to extract the current blade frequency band characteristic value decrease rate. This rate directly determines the subsequent feed rate reduction step size and the adjustment direction of the torque compensation current. If the decrease is slow, the reduction step size needs to be increased; if the decrease is fast, the step size needs to be decreased. This is the core basis for realizing dynamic parameter reconstruction. The specific implementation method is as follows:

[0048] After each feed rate reduction operation is completed, if a single-stage reduction from 800 mm / min to 700 mm / min is achieved and the spindle speed stabilizes at 3000 r / min after compensation, the system immediately opens a monitoring window of a fixed duration. This fixed duration is not absolutely fixed but dynamically adjusted based on the principle of covering at least three complete blade vibration cycles. The blade vibration cycle is the time interval between the reciprocating vibrations of the blade under cutting force, and its value is negatively correlated with the spindle speed. The higher the spindle speed and the higher the cutting frequency, the shorter the forced vibration cycle of the blade. The current actual speed must first be obtained through the spindle encoder, and then the cycle value must be determined by combining it with the speed-blade vibration cycle mapping table calibrated in previous experiments. For example, when the spindle speed is 3000 r / min, the corresponding blade vibration cycle is T = 20 ms (a cycle of vibration peak-valley-peak is completed every 20 ms). Therefore, the monitoring window length is set to 3 × T = 60 ms. If the spindle speed decreases to... 2500 r / min corresponds to a period T = 24 ms, and the window length is adjusted to 72 ms to ensure that the critical process of blade vibration falling back from the current peak can be fully captured within the window. This avoids missing the trend of characteristic value changes due to an excessively short window, or causing data redundancy and delaying subsequent control due to an excessively long window. Next, within the monitoring window, the blade frequency band characteristic value sequence is collected at equal height time intervals. The equal height time interval refers to a fixed collection interval that is much shorter than the blade vibration period, taking 1 / 5 of the period. For example, when the period is 20 ms, the interval is set to 4 ms. This ensures that the data point density is sufficient to reflect the smooth changes of characteristic values ​​without excessively increasing the computational load. During acquisition, the spindle vibration sensor first collects the raw vibration signal, which is then processed by a bandpass filter to obtain the pure blade frequency band characteristic values, represented by voltage values. Higher voltage indicates more intense blade vibration. For example, with a 60 ms window and a 4 ms interval, a total of 16 data points are collected (60 ms ÷ 4 ms + 1 = 16), which are then... =5.2V at 0ms =4.9V at 4ms =4.7V at 8ms =4.5V at 12ms... At 60ms, the voltage is 2.8V. These voltage values ​​arranged in chronological order constitute the blade frequency band characteristic value sequence. Least squares linear fitting is performed on the blade frequency band characteristic value sequence. This process aims to extract the trend of characteristic values ​​changing over time through mathematical modeling, as follows:

[0049] Let the time of the i-th data point within the monitoring window be the independent variable. (i=1,2,…,16) =(i-1)×4ms, if i=1 =0ms, i=2 =4ms, the corresponding blade frequency band characteristic value is the dependent variable. (like correspond =5.2V); The goal of linear fitting is to obtain the optimal straight line y=at+b, where a is the slope of the fitted line, reflecting the rate of change of the eigenvalues; b is the intercept, reflecting the initial eigenvalue level, minimizing the sum of squared residuals from all data points to this line, and the residuals are the actual values. with fitted value To find the difference, first calculate the average value over time t. Substituting the data yields =(0+4+8+…+60) / 16=450ms / 16≈28.125ms, then calculate the average value of the eigenvalue y. Substituting the data yields =(5.2+4.9+4.7+…+2.8) / 16≈3.9V, then calculate the molecule (i from 1 to 16), that is, the sum of the deviations of each time point from the mean multiplied by the deviations of the corresponding characteristic value from the mean, for example... =0 - 28.125 = -28.125ms =5.2-3.9=1.3V, the product is -36.5625ms·V. =4 - 28.125 = -24.125 ms =4.9 - 3.9 = 1.0V, the product is -24.125ms·V... Adding up the 16 products, we finally get Sty≈-185ms·V; at the same time, we calculate the denominator Stt= (i from 1 to 16), that is, the sum of squares of the deviations from the mean at each time point, for example =(-28.125)²≈791.02ms², =(-24.125)²≈581.91ms²…… After summing, Stt≈12187.5ms². According to the least squares formula, the slope a=Sty / Stt. Substituting the data, we get a≈-185ms·V / 12187.5ms²≈-0.0152V / ms. The negative sign indicates that the characteristic value of the blade frequency band decreases over time, that is, the vibration gradually weakens. The intercept b= -a×t≈3.9V-(-0.0152V / ms)×28.125ms≈3.9+0.428≈4.328V. To ensure the validity of the fit, the residual sum of squares Q also needs to be verified. ,like At 16ms, the fitted value y = -0.0152 × 16 + 4.328 ≈ 4.085V, and the actual value... =4.3V, residual =0.215V, residual square ≈0.046V², if the sum of squared residuals Q of all data points is less than 0.5V² (preset fitting accuracy threshold), then the fitting is considered effective. Finally, the absolute value of the slope 'a' of the fitted straight line is defined as the fall rate, i.e., fall rate = |a| ≈ 0.0152V / ms. This value represents the voltage amplitude of the blade frequency band characteristic value decreasing per millisecond. The larger the value, the more significant the suppression effect of the current feed rate reduction on blade vibration, and the subsequent reduction step size can be appropriately reduced. If the fall rate is only 0.005V / ms (below the first threshold of 0.01V / ms), it indicates that the suppression effect is insufficient, and the next stage feed rate reduction step size needs to be increased and the torque compensation current needs to be enhanced, thereby providing accurate quantitative basis for the dynamic coupling control unit 2 to reconstruct subsequent control parameters.

[0050] b. Dynamically reconstruct subsequent control parameters based on the rate of decline:

[0051] If the retracement rate is lower than the first threshold, the next level feed rate is synchronously increased by adjusting the step size and the torque compensation current peak value is proportionally increased. The method of the dynamic coupling control unit 2 to synchronously increase the next level feed rate by adjusting the step size and proportionally increase the torque compensation current peak value includes a linkage scaling mechanism, specifically including:

[0052] When the fallback rate is lower than the first threshold, the step size increment coefficient is determined based on the difference between the current fallback rate and the first threshold. This causes the next feed rate to be adjusted downwards by the increment coefficient proportionally. Simultaneously, based on the preset multiple relationship between the torque compensation current reference value and the feed rate adjustment step size, the peak output limit of the torque compensation current is increased proportionally. The increase in peak current value and the increase in step size maintain a linear relationship. It should be further noted that:

[0053] After the dynamic coupling control unit 2 extracts the current blade frequency band characteristic value fall rate using windowed differential analysis, it needs to compare this rate with a preset first threshold. The first threshold is a critical value for measuring the effect of feed rate reduction on blade vibration suppression. Based on the blade machining accuracy requirements and calibrated from 100 sets of different feed rate control experiments, it is set to 0.01V / ms. If the fall rate is lower than this value, it indicates that the current feed rate reduction is insufficient and the blade vibration suppression effect is not as expected. It is necessary to synchronously expand the next level feed rate reduction step size and enhance the peak value of torque compensation current through a linkage scaling mechanism to avoid continuous inefficient control leading to machining accuracy deviation. The specific implementation method is as follows:

[0054] When the retracement rate falls below the first threshold, the system first uses a data comparison module to confirm the relationship between the current retracement rate and the first threshold. The real-time retracement rate obtained from the windowed differential analysis module is then compared with the first threshold stored in the control parameter library. If the real-time value is less than the threshold and remains so for two consecutive control cycles (20ms), eliminating the possibility of misjudgment due to instantaneous fluctuations, the linkage scaling mechanism is triggered. Simultaneously, the current feed rate and torque compensation current parameters are locked as the reference values ​​for the next level of control. Then, based on the current retracement rate and the first threshold... The step size increment coefficient is determined by the difference in values. The difference in values ​​refers to the absolute difference between the first threshold and the current fall rate. The calculation formula is Δv = first threshold - current fall rate. Substituting the data, we get Δv = 0.01V / ms - 0.006V / ms = 0.004V / ms. The step size increment coefficient is a proportional coefficient used to increase the step size of the next feed rate. Its value is positively correlated with the difference in values. The design logic is: the larger the difference in values, the more significant the gap between the current control effect and the expectation, and the more the step size needs to be increased to accelerate the suppression of blade vibration. The specific calculation uses a linear mapping formula: step size increment coefficient k = 1 + (Δv / first threshold) × adjustment gain, where the adjustment gain was experimentally determined to be 1.5 to balance the step size expansion speed and load stability, avoiding a sudden increase in spindle load caused by an excessively large coefficient. Substituting the data, we get k = 1 + (0.004 / 0.01) × 1.5 = 1 + 0.6 = 1.6. Simultaneously, the upper limit of the step size increment coefficient is set to 2.0 to prevent excessive step size expansion, such as a sudden increase from 100mm / min to over 200mm / min, exceeding the spindle load tolerance. If the calculated k exceeds 2.0, then 2.0 is used as the final coefficient, causing the next level feed rate reduction step size value to increase proportionally according to the increment coefficient. The system first retrieves the current feed rate reduction step size reference value from the control parameter library, i.e., the value executed by the previous level. The step size, such as 100 mm / min, is based on the initial feed rate of 800 mm / min and the initial adjustment range set according to the machining requirements. The next step size is obtained through multiplication: next step size = current step size reference value × step size increment coefficient. Substituting the data, the next step size = 100 mm / min × 1.6 = 160 mm / min. To further ensure the stability of the spindle load, the upper limit of the calculated next step size needs to be checked. The preset upper limit of the step size is 20% of the current feed rate. For example, if the current feed rate is 700 mm / min, the upper limit is 140 mm / min. If the calculated 160 mm / min exceeds the upper limit, then 140 mm / min is used as the final next step size to avoid the spindle load change exceeding the decoupling capability due to the step size being too large. For example, if the load change exceeds 0...The feed rate was 2 N·m, exceeding the adjustment range of the torque compensation current. Therefore, the target value for the next feed rate reduction was determined to be 700 mm / min - 140 mm / min = 560 mm / min. This target value was simultaneously stored in the register of the feed rate control module, awaiting execution in the next control cycle. At the same time, based on the preset multiple relationship between the torque compensation current reference value and the feed rate reduction step size, the peak output limit of the torque compensation current was increased proportionally. The preset multiple relationship was a fixed ratio established through previous spindle load-current calibration experiments. Its physical meaning is the torque compensation current reference value required to match the unit feed rate reduction step size, which was experimentally determined to be 0.00. 0.00067A / (mm / min), meaning that for every 1mm / min decrease in feed rate, a step size adjustment requires 0.00067A of torque compensation current to stabilize the spindle speed. This relationship is stored in the torque compensation parameter library and can be dynamically updated according to changes in spindle model and machining material. The system first calculates the new torque compensation current reference value based on the next feed rate step size adjustment (140mm / min) and the preset multiple relationship: New reference value = next step size × preset multiple relationship = 140mm / min × 0.00067A / (mm / min) ≈ 0.0938A; The peak output limit value of the torque compensation current is the safety limit for the maximum output value of the torque compensation current. The full threshold is initially set to 1.2 times the current reference value, with a 20% adjustment margin to prevent excessive current from burning out the spindle motor. For example, if the current reference value of 0.067A corresponds to a limit value of 0.0804A, when increasing proportionally, the limit value remains at the ratio of new reference value × 1.2. Substituting the data, the new limit value is approximately 0.0938A × 1.2 ≈ 0.1126A. This limit value is then simultaneously applied to the output limit module of the current loop PID controller to ensure that the compensation current does not exceed the safe range. Finally, the peak current enhancement amplitude and the step size expansion amplitude maintain a linear correspondence. This linear correspondence is ensured through quantitative calculation: first, the step size expansion amplitude is calculated separately... The magnitude of the increase in peak current is 40%, and the step size increase is calculated as follows: Step size increase = (next step size - current step size) / current step size × 100% = (140 - 100) / 100 × 100% = 40%. Peak current increase is calculated as follows: (new limit value - current limit value) / current limit value × 100% = (0.1126 - 0.0804) / 0.0804 × 100% ≈ 40%. The magnitudes of the two are completely consistent, achieving a linear correspondence. This correspondence means that the feed rate decrease and the step size increase means that the spindle load change increases. For example, if the step size increases from 100 mm / min to 140 mm / min, the load change increases from 0.108 N·m to 0.At 151 N·m, the peak torque compensation current needs to be proportionally increased to ensure that the compensation capability matches the load variation, preventing speed fluctuations from exceeding the preset floating range (2995 r / min - 3005 r / min) due to insufficient current. Simultaneously, the linear correspondence is achieved through the proportional latch function of the hardware register. Each time the step increment coefficient is updated, the peak current adjustment module automatically reads the coefficient and synchronously adjusts the limit value, eliminating the need for additional calculations and ensuring real-time and coordinated control. Ultimately, this achieves precise linkage between feed rate reduction and torque compensation, accelerating the suppression of blade vibration while stabilizing the spindle speed.

[0055] The dynamic coupling control unit 2 synchronously contracts the next stage feed rate, lowers the step size, and proportionally reduces the torque compensation current intensity using a reverse adjustment mechanism, specifically including:

[0056] When the fallback rate exceeds the second threshold, a step size contraction factor is generated based on the proportion by which the fallback rate exceeds the second threshold. This causes the next stage feed rate to be reduced by the step size value proportionally to the contraction factor. Simultaneously, based on the output limit range of the current loop PID controller, the maximum allowable output value of the torque compensation current intensity is reduced synchronously according to the step size contraction ratio. The degree of current intensity reduction is positively correlated with the degree of step size contraction. It should be further noted that:

[0057] Unlike when the fall rate is below the first threshold, requiring a larger adjustment range to enhance blade vibration suppression, when the fall rate of the blade frequency band feature value extracted by the dynamic coupling control unit 2 is above the second threshold, it indicates that the current feed rate reduction has exceeded expectations in suppressing vibration. The second threshold is a critical value calibrated based on the balance between vibration suppression and machining efficiency. If the fall rate is too fast during blade machining, it will lead to excessive feed rate reduction, such as a sudden drop from 700 mm / min to below 500 mm / min, which will reduce production efficiency and may also cause spindle runaway due to excessively light load. Therefore, the second threshold is set to 0.02 V / ms, determined through 50 sets of high-efficiency machining experiments. This value ensures that vibration suppression meets the standard without excessively sacrificing efficiency. At this time, it is necessary to synchronously reduce the feed rate reduction step size and weaken the torque compensation current intensity through a reverse adjustment mechanism to avoid excessive control that disrupts the machining balance. The specific implementation method is as follows:

[0058] When the fallback rate exceeds the second threshold, the system first activates a dual-cycle verification mechanism. It obtains the real-time fallback rate (e.g., 0.026V / ms) from the windowed differential analysis module and compares it with the second threshold (0.02V / ms) stored in the control parameter library. If the real-time value exceeds the second threshold for two consecutive control cycles (20ms), excluding misjudgments caused by sensor transient noise, the reverse adjustment mechanism is triggered. Simultaneously, it retrieves the current feed rate reduction step size reference value from the feed rate control module and the current loop control module, such as the previous stage's 140mm / min and the maximum allowable output value of the torque compensation current (e.g., 0.1126A), as the reference parameters for reverse adjustment. To ensure clear reference for subsequent adjustments, a step-size contraction factor is generated based on the proportion by which the fallback rate exceeds the second threshold. The excess proportion quantifies the degree to which the fallback rate exceeds the threshold, and the calculation formula is: Excess Proportion = (Current Fallback Rate - Second Threshold) / Second Threshold × 100%. Substituting the real-time fallback rate of 0.026V / ms and the second threshold of 0.02V / ms, we get the excess proportion = (0.026 - 0.02) / 0.02 × 100% = 30%. The larger this value, the more excessive the current vibration suppression effect, requiring a more significant reduction in the step size. The step-size contraction factor is a proportional coefficient used to reduce the next step size, and its value is negatively correlated with the excess proportion. The design logic is as follows:

[0059] The higher the excess ratio, the smaller the shrinkage factor and the greater the step size reduction. However, it is necessary to avoid excessive shrinkage leading to an excessively small step size (e.g., below 50 mm / min), which would affect processing efficiency. Therefore, a linear constraint formula is adopted: Step size shrinkage factor k = 1 - (Excess ratio × Adjustment coefficient), where the adjustment coefficient is experimentally calibrated to 0.5 (to balance the shrinkage amplitude and efficiency; if the adjustment coefficient is too large, a 30% excess ratio will cause the shrinkage factor to drop below 0.85, resulting in excessive step size reduction; if it is too small, over-adjustment cannot be effectively avoided). Substituting the data, we get k = 1 - (30% × 0.5) = 1 - 0.15 = 0.85. At the same time, the lower limit of the step size shrinkage factor is set to 0.5, that is, the step size is reduced to a minimum of 50% of the original step size. If the excess ratio is too large (e.g., 8...), the shrinkage factor will be reduced to a minimum of 50% of the original step size. 0%), the calculated k=1-(80%×0.5)=0.6, is still higher than the lower limit, ensuring that the step size will not be excessively reduced, so that the next level feed rate adjustment step size is reduced proportionally according to the shrinkage factor. The system first confirms the current feed rate adjustment step size base value (140mm / min), and then obtains the next level step size value through multiplication: next level step size = current step size base value × step size shrinkage factor, substituting the data, we get the next level step size = 140mm / min×0.85=119mm / min; to further ensure processing efficiency, the calculated next level step size needs to be checked for a lower limit. The preset feed rate adjustment step size lower limit is 50mm / min. This value is the minimum step size to ensure a balance between efficiency and accuracy in blade processing. Below This value will extend the processing cycle by more than 30%. If the calculated step size is less than 50 mm / min, then 50 mm / min will be used as the final next step size. The final target value for adjusting the next feed rate is determined to be the current feed rate (560 mm / min) - the next step size (119 mm / min) = 441 mm / min. This target value is simultaneously written into the execution register of the feed rate control module, waiting for the next control cycle to trigger execution. At the same time, based on the output limit range of the current loop PID controller, the maximum allowable output value of the torque compensation current intensity is synchronously reduced according to the step size contraction ratio. First, the step size contraction ratio, i.e., the magnitude of the step size reduction, is calculated. The formula is: Step size contraction ratio = (current step size reference value) / ... -Next step size value) / Current step size reference value × 100%, substituting the data, we get the step size reduction ratio = (140-119) / 140 × 100% = 15%. This ratio directly determines the reduction in current intensity. The output limit range of the current loop PID controller is the maximum current value allowed by the controller. In the blade processing scenario, it is preset to 0-0.5A to avoid excessive current burning out the spindle motor. The system first retrieves the current maximum allowable output value of torque compensation current (0.1126A), and then calculates it according to the formula: new current limit value = current current limit value × (1 - step size reduction ratio). Substituting the data, we get the new current limit value = 0.1126A × (1%~15%) = 0.1126A × 0.85 ≈ 0.0957A. Then, verify that the new current limit value is within the PID output limit range (0-0.5A). After confirming that it is correct, load the new limit value into the output limit module of the current loop PID controller to ensure that the compensation current output does not exceed this value, avoiding excessive current causing the spindle speed to exceed the target value (e.g., exceeding 3005 r / min). Finally, the degree of current intensity reduction is positively correlated with the degree of step size contraction. This positive correlation is achieved through quantization matching. For every 1% increase in the step size contraction ratio, the maximum allowable current output value decreases by 1% synchronously. For example, if the step size contraction ratio increases from 15% to 20%, the current limit value decreases from 0.0957A to 0.1126A × (1%~20%) = 0.0901A. The changes in both are completely consistent. The core logic of this positive correlation design is as follows: Lowering the feed rate and reducing the step size means a smaller change in spindle load (the load change corresponding to a step size of 119 mm / min is approximately 0.13 N·m, a 14% reduction from the original 0.151 N·m at 140 mm / min). This eliminates the need for the original high-current compensation to maintain stable spindle speed. However, if the reduction in current intensity lags behind the reduction in step size, it will lead to excessive compensation current, causing spindle speed fluctuations (e.g., the speed increases from 3000 r / min to 3008 r / min, exceeding the fluctuation range). Conversely, insufficient compensation will result in the speed falling below the target value. Therefore, positive correlation matching ensures that the feed rate and current parameters are adjusted in tandem, avoiding over-adjustment while maintaining a balance between spindle speed stability and machining efficiency.

[0060] If the rate of fallback exceeds the second threshold, the feed rate of the next stage is simultaneously reduced by a smaller step size, and the torque compensation current intensity is reduced proportionally. During the transition gap between adjacent stages of reduction, the speed micro-oscillation suppression submodule is activated to detect transient speed jitter caused by the step switching of the feed rate in real time, and generates a pulse compensation current that is inversely phase to the jitter waveform to cancel it out. The method by which the speed micro-oscillation suppression submodule generates the inverse pulse compensation current includes waveform mirror reconstruction technology, specifically including:

[0061] The complete waveform of transient speed jitter is captured by a high-speed ADC sampling unit. The amplitude envelope and phase characteristics of the waveform are extracted. Based on the phase characteristics, a reference sinusoidal carrier with a preset phase offset angle is generated. Then, the amplitude envelope and the reference sinusoidal carrier are modulated and synthesized to generate a fully compensated current pulse that is inversely phase to the jitter waveform. The pulse width adaptively matches the transition gap duration of the feed rate switching. It should be further explained that:

[0062] When the dynamic coupling control unit 2 synchronously contracts the next feed rate step size and weakens the torque compensation current intensity through the reverse adjustment mechanism, the transition gap between adjacent feed rate stages, such as the process of switching from a step size of 140 mm / min to a step size of 119 mm / min, usually lasts within 20-30 ms. Even with torque compensation current support, the sudden change in feed rate may still cause instantaneous fluctuations in the spindle load, leading to transient speed jitter. That is, the spindle speed deviates from the target value for a short period of time, such as jumping from 3000 r / min to 3008 r / min instantly and then quickly falling back. Although this type of jitter is short in duration, it will directly affect the surface smoothness of the blade machining. Therefore, it is necessary to generate an inverse pulse compensation current through the waveform mirror reconstruction technology of the speed micro-oscillation suppression submodule to accurately counteract the jitter. The specific implementation method is as follows:

[0063] The complete waveform of transient speed jitter is captured by a high-speed ADC sampling unit. This high-speed ADC sampling unit is an analog-to-digital converter module with 16-bit resolution and a 1MHz sampling rate. The sampling rate is much higher than the frequency of transient jitter, which is typically 500-1000Hz. The 1MHz sampling rate ensures that more than 1000 data points are collected per cycle, completely restoring waveform details. Its triggering timing is synchronized with the feed rate switching command. When the feed rate control module issues a switching command, it immediately sends a start signal to the high-speed ADC sampling unit to avoid missing the initial stage of jitter due to delay. Transient speed jitter refers to the jitter that occurs during feed rate switching. The shaft speed generates short-term, high-frequency fluctuations around the target value, with amplitudes typically within ±8 r / min. The duration is positively correlated with the feed rate switching speed. During sampling, the ADC unit uses the differential signal from the spindle encoder (1024 lines / revolution resolution, outputting 1024 pulses per revolution, resulting in 51200 pulses per second at 3000 r / min) to acquire speed data in real time. After converting the analog pulse signal into a digital value, it is quickly transferred to the data buffer via the DMA (Direct Memory Access) channel to avoid delays caused by CPU intervention. The sampling duration is set to 1 / 3 of the feed rate switching transition gap.To ensure complete coverage of the entire jitter phase—from rise to peak to fall—a 2x multiplier is used. Simultaneously, the 3σ criterion is employed to eliminate pulse interference in the sampled data, ultimately yielding a smooth transient speed jitter waveform. For example, within the time axis (0-24ms), the speed increases from 3000 r / min to 3008 r / min and then falls back to 3000 r / min. The amplitude envelope and phase characteristics of this waveform are extracted. The amplitude envelope, formed by connecting the peak points at various moments in the transient speed jitter waveform, reflects the change in jitter intensity over time. The extraction is performed using a sliding window maximum value method: 10 sampling points form a window, corresponding to 10 μs, much smaller than the jitter period. The entire jitter waveform is traversed, and the maximum value within each window is taken. The maximum rotational speed is used as a data point in the envelope. These data points are then fitted into a continuous amplitude envelope using cubic spline interpolation, such as an envelope value of 3000 r / min at 0 ms, 3003 r / min at 5 ms, 3008 r / min at 10 ms, 3005 r / min at 15 ms, and 3000 r / min at 24 ms. Phase characteristics refer to the phase attributes of the jitter waveform on the time axis, usually based on the moment the waveform peak occurs, expressed as a phase angle (0-360°). First, the peak moment of the jitter waveform is located using a peak detection algorithm, and the corresponding phase angle is set to 90° (sine wave peak phase). Then, other phase angles are calculated based on the sampling time interval (1 μs). The phase angle at a given moment ultimately yields a complete phase distribution curve. This characteristic determines the phase reference for subsequent anti-phase compensation and is the core basis for ensuring that the compensation waveform is out of phase with the jitter waveform. Based on the phase characteristic, a reference sinusoidal carrier with a preset phase offset angle is generated. The reference sinusoidal carrier is a sinusoidal signal with a frequency consistent with the fundamental frequency of the transient speed jitter. Its fundamental frequency is extracted by performing a Fast Fourier Transform (FFT) on the jitter waveform. For example, if the FFT analysis shows that the main frequency component (fundamental frequency) of the jitter waveform is 800Hz, then the frequency of the reference sinusoidal carrier is set to 800Hz to ensure that the carrier can perfectly match the jitter waveform in frequency. The preset angle is 180° because a compensation signal out of phase with the jitter waveform needs to be generated, with a phase difference of 180°. When two waveforms are superimposed, their amplitudes cancel each other out. Specifically, a DDS (Direct Digital Frequency Synthesis) signal generator is used: first, the sampling period of the carrier is calculated based on the base frequency of 800Hz (1 / 800Hz = 1.25ms). Then, combined with the peak moment in the phase characteristics (10ms), the phase angle of the carrier at 10ms is set to 90° + 180° = 270°, which is opposite to the peak phase of the jitter waveform. Subsequently, discrete carrier data points are generated according to a sine function (e.g., the carrier value is -1V at 10ms, 0V at 10.625ms, and 1V at 11.25ms, corresponding to the out-of-phase phase of the 800Hz sine wave). Simultaneously, the maximum amplitude of the carrier is set to 1 / 800Hz, equal to the maximum amplitude of the jitter waveform.For example, if the maximum jitter amplitude is 8 r / min, the corresponding maximum carrier amplitude is converted to 0.1V using the current-speed conversion relationship to ensure sufficient compensation to offset the jitter. Then, the amplitude envelope is modulated and synthesized with the reference sinusoidal carrier to generate a fully compensated current pulse that is inversely phase to the jitter waveform. Amplitude modulation is used for this synthesis. The core logic is to ensure that the amplitude variation of the reference sinusoidal carrier is consistent with the jitter waveform, while maintaining the opposite phase. This ensures that the synthesized compensation waveform maintains an equal amplitude and opposite phase relationship with the jitter waveform at every moment, thus accurately offsetting the jitter. In practice, the speed value of the amplitude envelope is first converted using the speed-current conversion model, as previously specified as 0.0125 A / r·min. That is, every 1 r / min of speed deviation corresponds to a current compensation requirement of 0.0125A, which is converted into a current amplitude. For example, an envelope of 3008 r / min corresponds to a current amplitude of 0.1A. This current amplitude envelope is then multiplied by a reference sinusoidal carrier wave. For instance, at 10 ms, the current amplitude envelope of 0.1A is multiplied by the carrier value of -1V to obtain an instantaneous current value of -0.1A; at 11 ms, the current amplitude envelope of 0.0625A is multiplied by the carrier value of 0.5V to obtain an instantaneous current value of -0.03125A. Finally, the multiplication result is input into a low-pass filter with a cutoff frequency of 1 kHz to filter out high-frequency noise generated during modulation, resulting in a fully compensated current pulse in the time domain. The pulse waveform is completely inversely related to the transient speed jitter waveform. The compensation current decreases as the jitter increases and reaches its trough at the jitter peak. Furthermore, the amplitude variation follows the same pattern as the jitter, directly acting on the spindle motor to counteract it. Finally, the pulse width adaptively matches the transition gap duration of the feed rate switching. This transition gap duration refers to the complete time for the feed rate to switch from the current step size to the next step size. For example, switching from 140mm / min to 119mm / min requires 22ms, determined by the servo response speed of the feed drive module. This duration is transmitted in real-time to the pulse width adjustment unit of the speed micro-oscillation suppression submodule. The core of adaptive matching is to ensure that the duration of the compensation current pulse completely covers the entire transition gap. To prevent residual jitter caused by premature termination of compensation, the transition gap is carefully managed. Specifically, the module first reads the transition gap duration from the feed rate control module (e.g., 22ms), and then uses this duration as the reference value for the pulse width. Simultaneously, it samples the rotational speed jitter within the transition gap in real time. If the jitter ends 1ms prematurely, the pulse width is shortened by 1ms to ensure the pulse width always matches the jitter duration. For example, if the transition gap is 22ms, the width of the compensation current pulse is set to 22ms, output synchronously from the start of the feed rate switch until the switch ends and the jitter completely disappears. Ultimately, through this precise coverage and anti-phase cancellation method, the amplitude of transient rotational speed jitter is compressed to within ±1r / min, completely eliminating its impact on machining accuracy.

[0064] When the blade frequency band characteristic value remains stable at a safe threshold for a continuous preset control period, the feed rate reduction sequence is terminated synchronously and torque compensation is released. The determination mechanism for the continuous stability of the blade frequency band characteristic value at the safe threshold adopts a multi-condition convergence strategy, specifically including:

[0065] Within a continuous preset control cycle, when the three conditions are simultaneously met—the fluctuation amplitude of the blade frequency band characteristic value is lower than the stability threshold, the tool frequency band characteristic value returns to the reference fluctuation range, and the cross-frequency interference intensity factor decays to the safe zone—a synchronization termination signal is triggered. Simultaneously, the termination signal shuts down the feed rate reduction sequence generator and cuts off the compensation output channel of the current loop PID controller, causing the system to return to the reference control mode. It should be further noted that:

[0066] After the speed micro-oscillation suppression submodule eliminates the transient jitter caused by feed rate switching through the inverse pulse compensation current, the dynamic coupling control unit 2 needs to further determine whether the blade frequency band characteristic value has been continuously stabilized at the safety threshold. This is the core step in deciding whether to terminate the feed rate reduction and torque compensation. If only a single condition is used for judgment, it is easy to trigger the termination due to instantaneous stability, resulting in residual cross-frequency interference. Therefore, a multi-condition convergence strategy is adopted to comprehensively confirm from three dimensions: blade vibration, tool condition, and interference intensity, to ensure long-term stability of the machining state after the control is terminated. The specific implementation method is as follows:

[0067] Within a continuous preset control cycle, where the preset control cycle is the minimum time window for the system to determine a stable state, its duration is calibrated based on the hysteresis characteristics of vibration stability during blade processing. Blade vibration requires a certain amount of time to stabilize from being suppressed. A cycle that is too short may misjudge instantaneous fluctuations as stable, while a cycle that is too long will prolong the control time and reduce efficiency. Through stability experiments on 50 sets of blades made of different materials, five control cycles were determined, each lasting 10ms, for a total of 50ms. This means that all conditions must be met synchronously within 50ms for the system to be considered truly stable. The system will start a cycle counter, counting from the first detection that all three conditions are met. If any condition is not met in any intermediate cycle, the counter is immediately reset and counted again. This continuous verification mechanism effectively avoids misjudgments caused by instantaneous sensor noise or accidental load fluctuations. Then, it synchronously ensures that the fluctuation amplitude of the blade frequency band characteristic value is below the stability threshold. The fluctuation amplitude of the blade frequency band characteristic value refers to the difference between the maximum and minimum values ​​of the blade frequency band characteristic value within a continuous preset control cycle, reflecting the stability of the vibration. The stability threshold is calibrated to 0.3V based on the blade machining accuracy requirements; that is, when the fluctuation amplitude is ≤0.3V, the vibration will not affect the quality of the machined surface. In specific implementation, the system collects the blade frequency band characteristic value every 10ms, collecting 5 data points within 5 control cycles (50ms). The maximum value Vmax and minimum value Vmin of these 5 data points are first calculated to obtain the fluctuation amplitude Δ. V = Vmax - Vmin. Simultaneously, it's necessary to verify whether the average of these five data points is within the safety threshold range. The safety threshold is 2.0-3.0V. A value below 2.0V indicates weak vibration, potentially accompanied by insufficient cutting; a value above 3.0V still poses a potential chatter risk. This condition is only satisfied when ΔV ≤ 0.3V and the average value is within the 2.0-3.0V range. This avoids misjudgments caused by only considering the fluctuation amplitude and ignoring static deviations. Then, the tool frequency band characteristic value returns to the reference fluctuation range. The tool frequency band characteristic value is the vibration signal voltage value corresponding to the tool's natural frequency (800-1200Hz), and its magnitude is positively correlated with the tool load and wear condition. The reference fluctuation range is the characteristic value of a new tool under normal machining conditions. The range (e.g., 4.0-5.0V) will be dynamically corrected based on the tool wear state coefficient calculated by the cross-frequency interference event judgment unit 1. The more severe the wear, the more significant the decrease in the tool frequency band characteristic value, and the reference range needs to be adjusted accordingly. For example, when the tool wear state coefficient is 0.6 (moderate wear), the reference range is corrected to 3.5-4.5V, and when the coefficient is 0.9 (severe wear), it is corrected to 3.0-4.0V to ensure that the judgment standard matches the actual tool state. In specific implementation, the system also collects the tool frequency band characteristic value every 10ms and calculates the mean value Vtoolavg within 5 control cycles. If Vtoolavg falls within the corrected reference fluctuation range, and the deviation between the characteristic value and the mean value in each cycle is ≤0.2V, to avoid instantaneous deviation from the reference range, this condition is met. This step ensures that the tool condition also returns to normal, avoiding instability in subsequent machining due to abnormal tool load. Then, the cross-frequency interference intensity factor decays to the safe zone. The cross-frequency interference intensity factor is an indicator defined above, measuring the degree of correlation between the blade and tool frequency band anomalies. It is the ratio of the absolute value of the increment of the blade frequency band characteristic value to the absolute value of the decrease of the tool frequency band characteristic value within M consecutive control cycles. The safe zone is calibrated to 0.3-0.8 based on the factor range when there is no interference during normal machining. Below 0.3 indicates that the interference is too weak and no adjustment is needed, while above 0.8 indicates that interference still exists. In cases of significant cross-frequency interference risk, continued regulation is necessary. Specifically, the system calculates the cross-frequency interference intensity factor in real time using the method described above. If the factor remains within the range of 0.3-0.8 for five consecutive control cycles, and the fluctuation range is ≤0.1 (avoiding drastic factor changes), this condition is met. It should be noted that when the blade frequency band characteristic value has stabilized (no significant increment), the factor is set to 0.5 (the median value of the safe zone) to prevent misjudgment that the safe zone has not been reached due to a factor of 0 caused by no blade increment, ensuring the integrity of the judgment logic. When the above three conditions are simultaneously met within five consecutive control cycles, a synchronization termination signal is triggered. The multi-condition convergence determination module generates a high-level synchronous termination signal (3.3V) lasting 10ms to ensure stable signal reception. This signal is simultaneously transmitted to the feed rate control module and current loop control module of the dynamic coupling control unit 2 via the system's internal high-speed bus (transmission delay ≤1μs). To avoid signal loss, dual-path redundant transmission is used; subsequent operations can only be executed when both signals are received, further improving reliability. The termination signal simultaneously shuts down the feed rate down-adjustment sequence generator and cuts off the compensation output channel of the current loop PID controller. The feed rate down-adjustment sequence generator stores and generates... The module that adjusts the step size of subsequent feed rates has its internal registers containing the step size values ​​to be executed at the next level. After receiving the termination signal, the feed rate control module immediately sends a clear and shut-off command to the sequence generator to clear the step size data in the register. At the same time, it locks the current feed rate as the final machining feed rate and sends a control command to the feed drive module to maintain the current feed rate, stopping all subsequent feed rate adjustment operations. The compensation output channel of the current loop PID controller is the path for injecting torque compensation current into the spindle motor. After receiving the termination signal, the current loop control module first sets the compensation current reference value of the PID controller to 0.The rate ramp from 0.1A / ms is reduced to 0 to prevent sudden current drops from causing spindle torque fluctuations. Then, the compensation output channel is cut off, and the PWM modulation module outputs a signal with a duty cycle of 0, completely stopping the compensation current injection and ensuring the spindle motor returns to normal current drive. Finally, the system returns to the baseline control mode. The baseline control mode is the initial machining parameter control mode, based on the target speed and feed rate set in the work order. It uses a conventional PID control algorithm to maintain stable machining without dynamic torque compensation or feed rate adjustment. In practice, the system disables the sub-modules related to dynamic control, switches spindle speed control to conventional speed PID, and feed rate control to conventional position PID. Simultaneously, key data from this control process (feed rate adjustment records, compensation current change curves, and stability judgment results) are stored in the machining log database (using a MySQL cluster, supporting data backtracking) to provide parameter references for subsequent machining of similar workpieces. This completes the closed-loop transition from dynamic parameter optimization to stable machining, ensuring a balance between high precision and high efficiency in blade machining.

[0068] In this invention, the cross-frequency interference event determination unit 1 calculates the cross-frequency interference intensity factor within M consecutive control cycles by comparing the rise of the blade frequency band characteristic value with the fall of the tool frequency band characteristic value. It then combines the tool wear state coefficient obtained from the analysis of the cutting force harmonic components and the tool fundamental frequency offset to correct the tolerance threshold and generate a coordinated control command. After receiving the command, the dynamic coupling control unit 2 executes a step-down adjustment of the feed rate, synchronously calculates and injects the spindle torque compensation current to stabilize the speed. At the same time, it samples transient jitter through the speed micro-oscillation suppression submodule and generates an anti-phase pulse to cancel it. When the blade frequency band characteristic value fluctuates, the tool frequency band returns to the reference, and the interference intensity factor reaches the safe zone, the control is terminated to improve machining accuracy.

[0069] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples and are not intended to limit the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A numerical control machine tool machining parameter self-optimization system based on cross-frequency interference coupling regulation, characterized in that, The method comprises a cross-frequency interference event determination unit (1) for generating a cooperative control instruction when the amplitude difference between the rising amplitude of the blade frequency band characteristic value and the falling amplitude of the tool frequency band characteristic value exceeds a tolerance threshold, and a dynamic coupling control unit (2) receiving the cooperative control instruction and executing a feed quantity step-down and a speed compensation mechanism: When a single-stage feed quantity down-regulation operation is performed, the spindle load change caused by the sudden change of the feed quantity is captured in real time, and based on the spindle load change, the spindle torque compensation current is dynamically calculated and injected in the same control cycle to compress the actual speed fluctuation to a preset floating interval. After each stage of feed quantity down-regulation is completed, a synchronous execution is performed. (a) Extracting the current blade frequency band characteristic value falling rate; (b) Dynamically reconstructing subsequent control parameters according to the falling rate: If the falling rate is lower than the first threshold, the next stage of feed quantity down-regulation step is expanded and the torque compensation current peak value is increased proportionally. If the falling rate is higher than the second threshold, the next stage of feed quantity down-regulation step is contracted and the torque compensation current intensity is reduced proportionally. In the transition gap between the adjacent two stages of down-regulation, a speed micro-oscillation suppression sub-module is started to detect the transient speed jitter caused by the step switching of the feed quantity, and a pulse compensation current opposite to the jitter waveform is generated to offset it. When the blade frequency band characteristic value is stable at the safety threshold for a continuous preset control cycle, the feed quantity down-regulation sequence is terminated and the torque compensation is removed.

2. The CNC machine tool processing parameter self-optimization system based on cross-frequency interference coupling regulation of claim 1, wherein: The cross-frequency interference event determination unit (1) establishes a dynamic compensation coefficient calculation mechanism when determining the relationship between the rising amplitude of the blade frequency band characteristic value and the falling amplitude of the tool frequency band characteristic value, which specifically includes: The ratio of the absolute value of the blade frequency band characteristic value increment to the absolute value of the tool frequency band characteristic value decrement in the continuous M control cycles is taken as the cross-frequency interference intensity factor. When the cross-frequency interference intensity factor exceeds the upper limit of the dynamically adjusted tolerance threshold and lasts for a preset number of periods, a cooperative control instruction is generated. The upper limit of the tolerance threshold is corrected in real time based on the tool wear state coefficient, which is obtained by analyzing the correlation between the cutting force harmonic component and the tool fundamental frequency offset.

3. The CNC machine tool processing parameter self-optimization system based on cross-frequency interference coupling regulation of claim 1, wherein: The dynamic coupling control unit (2) dynamically samples the spindle load change caused by the sudden change of the feed quantity through the embedded current loop, which specifically includes: At the starting moment of the feed quantity down-regulation instruction execution, the spindle current high-frequency acquisition module is started to capture the transient change of the current waveform front at a sampling rate not less than a preset proportion of the control cycle frequency, to obtain the current waveform integral area mutation, and based on the current waveform integral area mutation and a preset spindle load-current conversion model, the spindle load change is calculated. The feed quantity mutation is defined as the single-stage feed quantity down-regulation step value exceeding a preset percentage threshold of the current feed reference value.

4. The CNC machine tool processing parameter self-optimization system based on cross-frequency interference coupling regulation of claim 3, wherein: The method for dynamically calculating the spindle torque compensation current by the dynamic coupling control unit (2) in the same control cycle contains timing constraints, which specifically includes: After obtaining the spindle load change amount at the control cycle start stage, the calculation and output preparation of the torque compensation current value are completed within the control cycle. The calculation process adopts the inverse solving path of the load-current conversion model. The spindle load change amount is input into the preset torque compensation response function. The function integrates the spindle electromechanical time constant and the rotational inertia compensation coefficient. The torque compensation current reference value is output. At the same time, the dynamic inertia compensation component determined by the feed rate reduction rate is superimposed.

5. The CNC machine tool processing parameter self-optimization system based on cross-frequency interference coupling regulation of claim 4, wherein: The method for the dynamic coupling regulation unit (2) to inject the spindle torque compensation current to compress the actual speed fluctuation to the preset floating interval includes a current closed-loop injection mechanism, specifically including: The torque compensation current reference value is generated through the current loop PID controller to generate the PWM modulation signal. The PWM modulation signal is injected into the three-phase winding of the spindle motor through the gate drive circuit. At the same time, a speed fluctuation suppression feedback loop is established. The spindle encoder feedback signal is collected in real time. When the deviation of the detected actual speed from the target speed exceeds the floating interval, the integral gain coefficient of the PID controller is automatically increased to realize interval compression.

6. The CNC machine tool processing parameter self-optimization system based on cross-frequency interference coupling regulation of claim 1, wherein: The dynamic coupling regulation unit (2) uses a windowed differential analysis method to extract the current blade frequency band characteristic value falling rate, specifically including: After each level of feed rate reduction operation is completed, a fixed length monitoring window is opened. The blade frequency band characteristic value sequence is collected at equal time intervals within the monitoring window. The blade frequency band characteristic value sequence is subjected to least squares linear fitting. The absolute value of the fitting straight line slope is defined as the falling rate. The monitoring window length is dynamically adjusted according to the current spindle speed to ensure that at least three complete blade vibration periods are covered.

7. The CNC machine tool machining parameter self-optimization system based on cross-frequency interference coupling regulation of claim 6, wherein: The method for the dynamic coupling regulation unit (2) to simultaneously expand the next level of feed rate reduction step size and proportionally increase the torque compensation current peak value includes a linkage scaling mechanism, specifically including: When the falling rate is lower than the first threshold value, the step size increment coefficient is determined based on the difference between the current falling rate and the first threshold value. The next level of feed rate reduction step size value is proportionally expanded by the increment coefficient. At the same time, according to the preset multiple relationship between the torque compensation current reference value and the feed rate reduction step size, the torque compensation current peak value output limit value is proportionally increased. The current peak value enhancement amplitude and the step size expansion amplitude maintain a linear corresponding relationship.

8. The CNC machine tool processing parameter self-optimization system based on cross-frequency interference coupling regulation of claim 6, wherein: The dynamic coupling regulation unit (2) simultaneously contracts the next level of feed rate reduction step size and proportionally weakens the torque compensation current intensity using a reverse adjustment mechanism, specifically including: When the falling rate is higher than the second threshold value, the step size contraction factor is generated according to the exceeding proportion of the falling rate exceeding the second threshold value. The next level of feed rate reduction step size value is proportionally reduced by the contraction factor. At the same time, according to the output limit range of the current loop PID controller, the maximum allowed output value of the torque compensation current intensity is simultaneously reduced by the step size contraction proportion. The current intensity weakening degree and the step size contraction degree maintain a positive correlation.

9. The CNC machine tool processing parameter self-optimization system based on cross-frequency interference coupling regulation of claim 1, wherein: The method for the speed micro-oscillation suppression sub-module to generate a reverse pulse compensation current includes a waveform mirror reconstruction technology, specifically including: The complete waveform of the transient speed jitter is captured by a high-speed ADC sampling unit, the amplitude envelope and phase characteristics of the waveform are extracted, a reference sinusoidal carrier with a preset phase offset angle is generated based on the phase characteristics, the amplitude envelope and the reference sinusoidal carrier are modulated and synthesized to generate a full compensation current pulse which is opposite to the jitter waveform, and the pulse width is adaptively matched with the transition gap length of the feed rate switching.

10. The CNC machine tool processing parameter self-optimization system based on cross-frequency interference coupling regulation of claim 1, wherein: The determination mechanism of the blade frequency band characteristic value continuously and stably reaching the safety threshold adopts a multi-condition convergence strategy, specifically including: Within a continuous preset control period, when three conditions are simultaneously met, namely, the fluctuation amplitude of the blade frequency band characteristic value is lower than the stability threshold, the tool frequency band characteristic value returns to the reference fluctuation range, and the cross-frequency interference intensity factor decays to the safety zone, a synchronous termination signal is triggered, the termination signal simultaneously closes the feed rate down sequence generator and cuts off the compensation output channel of the current loop PID controller, and the system returns to the reference control mode.

Images