A built-in motor direct drive power tool tower adaptive control method and system

Abstract

The application relates to the technical field of control of a built-in motor direct drive power tool turret, in particular to a built-in motor direct drive power tool turret adaptive control method and system. The method comprises the following steps: collecting vibration signals and performing frequency analysis to obtain actual vibration frequency information; acquiring torque instruction signals and performing frequency analysis to obtain torque instruction frequency information; determining the frequency overlap and / or the amplitude value relationship at the corresponding frequency according to the actual vibration frequency information and the torque instruction frequency information, and discriminating whether the current vibration is induced by the torque instruction or caused by external cutting force fluctuation; if the vibration is induced by the torque instruction, adjusting the torque instruction output characteristic to suppress the vibration; and if the vibration is caused by the external cutting force fluctuation, adjusting the torque compensation strategy to stabilize the cutting. The method solves the problems that the traditional built-in motor direct drive power tool turret is frequently and quickly adjusted in torque, high-frequency micro-vibration is excited in the tool turret system, tool wear is accelerated, the machining surface quality is reduced, and the cumulative damage of the tool turret parts is caused.

Description

Technical Field

[0001] This application relates to the field of control technology for a turret with built-in motor direct drive force, and more specifically, to an adaptive control method and system for a turret with built-in motor direct drive force. Background Technology

[0002] In precision CNC machining, the built-in motor-driven direct-drive turret is a key component for achieving high efficiency and precision. However, existing adaptive control systems face significant challenges in machining multi-layered composite materials and other applications where cutting forces fluctuate rapidly and at high frequencies. Composite materials exhibit rapid changes in internal hardness, toughness, and fiber orientation, resulting in high-frequency, irregular fluctuations in cutting forces. When the control system detects drastic changes in the drive current, it frequently adjusts the motor torque to maintain the rotational speed. However, because load fluctuations exceed the response bandwidth of the control algorithm, torque output is prone to lag or overshoot, causing transient deviations in tool speed. Furthermore, the high-frequency, rapidly changing torque commands are equivalent to continuous pulse excitation, potentially exciting the turret's inherent vibration modes and generating high-frequency micro-vibrations. These vibration frequencies often exceed the sampling range or resolution of current and speed sensors, making them difficult for the system to detect and suppress. The result is continuous tool tip vibration, accelerating tool wear, significantly shortening tool life, and causing micro-cracks or structural damage on the workpiece surface, reducing part fatigue strength and reliability. Prolonged exposure to such operating conditions can also cause cumulative damage to the precision components inside the turret, such as damage to the bearing lubrication film, accelerated wear, increased clearance, and encoder signal drift and decreased accuracy. Ultimately, this can lead to a hidden decline in turret rigidity and control accuracy, and even occasional accuracy problems when machining conventional materials.

[0003] To address the aforementioned issues, existing technologies urgently need improvement. Summary of the Invention

[0004] This application discloses an adaptive control method and system for a turret with a built-in motor direct drive force, which aims to solve the problem that in precision CNC machining, when a turret with a built-in motor direct drive force is machining special materials with high frequency and rapid fluctuating cutting forces, the traditional adaptive control system causes frequent and rapid torque adjustment actions, which excites high-frequency micro-vibrations inside the turret system, leading to accelerated tool wear, decreased surface quality, and cumulative damage to turret components.

[0005] The technical solution of this application is as follows: In a first aspect, this application discloses an adaptive control method for a turret with a built-in motor direct drive force, comprising: Vibration signals of the built-in motor direct drive turret during the machining process are collected, and frequency analysis is performed on the vibration signals to obtain actual vibration frequency information characterizing the actual vibration state of the built-in motor direct drive turret. The actual vibration frequency information includes frequency components within a predetermined frequency band and the amplitude and / or energy value corresponding to each frequency component. The torque command signal driving the built-in motor is acquired, and the frequency of the torque command signal is analyzed to obtain the torque command frequency information characterizing the torque command change characteristics. The torque command frequency information includes frequency components within a predetermined frequency band and the amplitude and / or energy value corresponding to each frequency component. Based on the actual vibration frequency information and the torque command frequency information, determine the frequency overlap and / or the amplitude relationship at the corresponding frequency, and determine whether the current vibration of the built-in motor direct drive turret is mainly induced by the torque command or mainly caused by external cutting force fluctuations, and obtain the determination result. The frequency overlap includes at least the overlapping frequency components of the two in the predetermined frequency band, and the amplitude relationship at the corresponding frequency includes at least the amplitude ratio and / or energy ratio of the vibration signal and the torque command signal at the overlapping frequency components. When the judgment result indicates that the current vibration is mainly induced by the torque command, the output characteristics of the torque command are adjusted to reduce the excitation effect of the vibration-inducing frequency component in the torque command on the built-in motor direct drive turret, thereby suppressing the vibration. When the judgment result indicates that the current vibration is mainly caused by external cutting force fluctuations, the torque compensation strategy used to drive the built-in motor is adjusted to enhance the compensation capability for external cutting force fluctuations, thereby suppressing vibration and maintaining the stability of the cutting process.

[0006] Secondly, this application also discloses a built-in motor direct drive force turret adaptive control system, comprising: The vibration frequency information acquisition module is used to collect vibration signals of the built-in motor direct drive turret during the processing, and perform frequency analysis on the vibration signals to obtain actual vibration frequency information that characterizes the actual vibration state of the built-in motor direct drive turret. The actual vibration frequency information includes frequency components within a predetermined frequency band and the amplitude and / or energy value corresponding to each frequency component. The torque command frequency information acquisition module is used to acquire the torque command signal driving the built-in motor, and perform frequency analysis on the torque command signal to obtain torque command frequency information characterizing the torque command change characteristics. The torque command frequency information includes frequency components within a predetermined frequency band and the amplitude and / or energy value corresponding to each frequency component. The vibration source discrimination module is used to determine the frequency overlap and / or amplitude relationship at the corresponding frequency based on the actual vibration frequency information and the torque command frequency information, and to determine whether the current vibration of the built-in motor direct drive turret is mainly induced by the torque command or mainly caused by external cutting force fluctuations, and to obtain the discrimination result. The frequency overlap includes at least the overlapping frequency components of the two in the predetermined frequency band, and the amplitude relationship at the corresponding frequency includes at least the amplitude ratio and / or energy ratio of the vibration signal and the torque command signal at the overlapping frequency components. The torque command output characteristic adjustment module is used to adjust the output characteristics of the torque command when the judgment result indicates that the current vibration is mainly induced by the torque command, so as to reduce the excitation effect of the vibration-inducing frequency component in the torque command on the built-in motor direct drive turret, thereby suppressing vibration; The torque compensation strategy adjustment module is used to adjust the torque compensation strategy used to drive the built-in motor when the judgment result indicates that the current vibration is mainly caused by external cutting force fluctuations, so as to enhance the compensation capability for external cutting force fluctuations, thereby suppressing vibration and maintaining the stability of the cutting process.

[0007] Beneficial Effects: This application effectively solves the problem in existing technologies, especially when machining special materials with high-frequency, rapidly fluctuating cutting forces, where traditional adaptive control systems unintentionally induce high-frequency micro-vibrations in the turret due to frequent and rapid torque adjustments. This application can accurately identify the source of vibration, avoiding the single or inappropriate control strategies adopted by traditional control systems due to their inability to distinguish vibration sources. This effectively suppresses problems such as accelerated tool wear, decreased surface quality, and cumulative damage to turret components. Compared with existing technologies, the control method of this application has higher adaptability and robustness, significantly improving the efficiency, accuracy, and reliability of precision CNC machining, extending the service life of tools and turrets, and ensuring the machining quality of workpieces. Attached Figure Description

[0008] Figure 1 A flowchart illustrating an adaptive control method for a turret with a built-in motor direct drive force provided in this application.

[0009] Figure 2 A flowchart of a built-in motor direct drive force turret adaptive control system provided for this application.

[0010] In the diagram: 1. Vibration frequency information acquisition module; 2. Torque command frequency information acquisition module; 3. Vibration source identification module; 4. Torque command output characteristic adjustment module; 5. Torque compensation strategy adjustment module. Detailed Implementation

[0011] The technical solution of this application will be described below with reference to the accompanying drawings. The embodiments described are only for explaining this application and do not constitute a limitation on the scope of protection.

[0012] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, the terms "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0013] Reference Figure 1This application proposes an adaptive control method for a turret with built-in motor direct drive force, comprising: S1000: Collects vibration signals from the built-in motor direct drive turret during the machining process, performs frequency analysis on the vibration signals, and obtains actual vibration frequency information characterizing the actual vibration state of the built-in motor direct drive turret. The actual vibration frequency information includes the frequency components within the predetermined frequency band and the amplitude and / or energy value corresponding to each frequency component.

[0014] S2000: Acquires the torque command signal driving the built-in motor, performs frequency analysis on the torque command signal, and obtains torque command frequency information characterizing the torque command change characteristics; The torque command frequency information includes frequency components within a predetermined frequency band and the amplitude and / or energy value corresponding to each frequency component.

[0015] S3000: Based on the actual vibration frequency information and the torque command frequency information, determine the frequency overlap and / or the amplitude relationship at the corresponding frequency, and determine whether the current vibration of the built-in motor direct drive turret is mainly induced by the torque command or mainly caused by external cutting force fluctuations, and obtain the judgment result; The frequency overlap includes at least the overlapping frequency components of the two signals within a predetermined frequency band, and the amplitude relationship at the corresponding frequencies includes at least the amplitude ratio and / or energy ratio of the vibration signal and the torque command signal at the overlapping frequency components.

[0016] S4000: When the judgment result indicates that the current vibration is mainly induced by the torque command, the output characteristics of the torque command are adjusted to reduce the excitation effect of the vibration-inducing frequency component in the torque command on the built-in motor direct drive turret, thereby suppressing vibration; S5000: When the judgment result indicates that the current vibration is mainly caused by external cutting force fluctuations, the torque compensation strategy used to drive the built-in motor is adjusted to enhance the compensation capability for external cutting force fluctuations, thereby suppressing vibration and maintaining the stability of the cutting process.

[0017] Among them, the built-in motor direct-drive turret refers to a structure in which the drive motor is directly integrated inside the turret, directly driving the tool for rotation and feed. Its characteristics include fast response speed, short transmission chain, and high precision. Vibration signal refers to a physical quantity signal reflecting the vibration state of the turret's mechanical structure, collected by sensors (such as accelerometers). Frequency analysis is a technique that converts time-domain signals into frequency-domain signals, for example, by using Fast Fourier Transform (FFT) to obtain the frequency components, amplitude, and / or energy values ​​of the signal. Torque command signal refers to the electrical signal sent by the machine tool's CNC system or servo driver to the built-in motor to control the motor's output torque. Preset frequency band refers to one or more frequency ranges pre-set according to the turret's mechanical characteristics and common vibration modes, used for centralized analysis of the frequency characteristics of relevant vibrations and torque commands. The predetermined frequency band can be determined through trial cutting calibration, system identification, or historical data, and is stored in the control system as a fixed or updatable configuration parameter.

[0018] Specifically, the control method of this application includes the following main features: In one implementation, the vibration signal is acquired by a high-sensitivity accelerometer installed on the turret body. The analog signal is digitized by an analog-to-digital converter (ADC) and then analyzed by a digital signal processor (DSP) or industrial PC using Discrete Fourier Transform (DFT) or Fast Fourier Transform (FFT) to obtain the frequency components and their corresponding amplitudes and / or energy values ​​within a predetermined frequency band (e.g., 50Hz to 2000Hz) (energy values ​​can be characterized by power spectrum or amplitude squared). The torque command signal is read from the servo driver via industrial Ethernet protocols such as EtherCAT and Profinet or an analog output interface and subjected to FFT to obtain the frequency components and their corresponding amplitudes and / or energy values ​​within the same predetermined frequency band, and aligned with the actual vibration frequency information on the frequency axis. The frequency overlap of the two is compared with the corresponding... When considering the amplitude relationship at different frequencies, for example, if both the vibration signal and the torque command signal have significant peaks at 100Hz, frequency overlap is considered to exist. The amplitude ratio and / or energy ratio of the vibration signal and the torque command signal at the overlapping frequency components are calculated. If necessary, cross-spectral density analysis or coherence analysis can be used to verify the causal direction of "torque command first, vibration later". When the discrimination result indicates that the current vibration is mainly induced by the torque command, a notch filter or low-pass filter with a center frequency of 100Hz is introduced into the torque command channel to attenuate the vibration-inducing frequency components. When the discrimination result indicates that the current vibration is mainly caused by external cutting force fluctuations, the servo controller gain is increased or a torque feedforward control algorithm is used, along with amplitude limiting / smoothing constraints, to enhance the compensation capability for external cutting force fluctuations and avoid introducing new vibration-inducing frequency components.

[0019] In another embodiment of this application, S3000 is further proposed to include: S3100: Acquire tool wear condition information, including wear level information used to characterize the degree of wear; S3200: Obtain the mechanical characteristic parameters of the built-in motor direct drive turret. The mechanical characteristic parameters include the set of natural vibration frequencies of the built-in motor direct drive turret structure and / or the set of damping ratios corresponding to the natural vibration frequencies. S3300: Performs time-frequency analysis on vibration signals and torque command signals to obtain the energy distribution of vibration signals and torque command signals at different time points and different frequencies, wherein the energy distribution includes the amplitude spectrum and / or power spectrum of each frequency point within a predetermined time window. S3400: Based on energy distribution, calculate the time-frequency correlation and instantaneous phase difference between the vibration signal and the torque command signal, wherein the time-frequency correlation includes the correlation coefficient and / or cross-correlation peak value within a predetermined frequency band; the instantaneous phase difference includes the phase difference sequence at corresponding frequency points within the predetermined frequency band and / or the time lag amount deduced from the phase difference; S3500: Determine the frequency overlap and / or amplitude relationship at the corresponding frequency based on the actual vibration frequency information and the torque command frequency information, and the amplitude ratio and / or energy ratio are extracted from the amplitude spectrum and / or power spectrum in the energy distribution; S3600: Judgment is made based on a combination of frequency overlap and / or amplitude relationships at corresponding frequencies, time-frequency correlation, instantaneous phase difference, wear condition information, and mechanical characteristic parameters. S3700: Determine the expected time lag range for judgment based on mechanical characteristic parameters; S3800: Adjust the first preset condition and / or the second preset condition used for judgment based on the wear condition information; S3900: When the frequency overlap and / or the amplitude relationship at the corresponding frequency meets the adjusted first preset condition, and the time-frequency correlation meets the adjusted second preset condition, and the time lag falls within the expected time lag range, it is determined that the current vibration is mainly induced by the torque command; otherwise, it is determined that the current vibration is mainly caused by the external cutting force fluctuation, and the judgment result is obtained.

[0020] Specifically, the first step in the discrimination process is to obtain the tool wear status information. The degree of tool wear directly affects the stability and frequency characteristics of the cutting force, thereby altering the turret's vibration response. Wear status information can be understood as data used to quantify the degree of tool wear, such as wear level information, like slight wear, moderate wear, or severe wear. The wear level information can be obtained by mapping a wear characteristic index to a preset wear threshold range. The wear level information is used to adjust the first and / or second preset conditions used for discrimination in subsequent steps, thereby reducing the risk of misjudgment when wear causes spectral feature drift.

[0021] Simultaneously, it is necessary to obtain the mechanical characteristic parameters of the turret with its built-in motor direct drive. These parameters are inherent physical properties of the turret structure, such as the set of natural vibration frequencies and the set of damping ratios corresponding to those natural vibration frequencies. The natural vibration frequency is the frequency at which the system vibrates freely without external force, while the damping ratio reflects the rate at which the system's vibrational energy dissipates. The mechanical characteristic parameters can be obtained through offline modal testing or online micro-excitation identification and are used as prior parameters in the discrimination process. These mechanical characteristic parameters are used to determine the expected time lag range in subsequent steps and can be updated as the set of natural vibration frequencies and the set of damping ratios change.

[0022] To more comprehensively analyze the dynamic characteristics of vibration and torque commands, this application performs time-frequency analysis on the vibration and torque command signals. Time-frequency analysis is a signal processing technique that can simultaneously reveal the changes of a signal in the time and frequency domains, such as using short-time Fourier transform or wavelet transform. Through time-frequency analysis, the energy distribution of the signal at different time points and frequencies can be obtained, where the energy distribution can be specifically represented as the amplitude spectrum and / or power spectrum at each frequency point within a predetermined time window. The length of the predetermined time window can be determined based on the sampling rate and the target frequency resolution to balance transient capture and frequency resolution.

[0023] Based on this, and using energy distribution, the time-frequency correlation and instantaneous phase difference between the vibration signal and the torque command signal are calculated. Time-frequency correlation measures the degree of linear correlation between two signals within a specific time-frequency region; for example, it can be calculated as a correlation coefficient or cross-correlation peak. Instantaneous phase difference reveals the phase lead or lag relationship between two signals at a specific frequency component; for example, it can be represented by a phase difference sequence or a calculated time lag, where the time lag is used to convert the phase difference into a quantitative indicator that can be directly compared to the mechanical response delay.

[0024] Furthermore, the amplitude ratio and / or energy ratio are extracted from the amplitude spectrum and / or power spectrum in the energy distribution. This extraction uses the overlapping frequency components determined by the frequency overlap as an index to calculate the amplitude ratio and / or energy ratio of the vibration signal and the torque command signal within the corresponding time window, so that the "amplitude relationship at the corresponding frequency" can still be extracted and compared consistently according to the time window under time-varying operating conditions.

[0025] Ultimately, this application makes a judgment based on the frequency overlap and / or the amplitude relationship at the corresponding frequency, time-frequency correlation, instantaneous phase difference, wear state information, and mechanical characteristic parameters. Specifically, the expected time lag range for judgment is determined based on the mechanical characteristic parameters; the phase delay range of the system frequency response within this frequency band is determined based on the set of natural vibration frequencies and the set of damping ratios; and the phase difference is converted into a time lag and compared with this range. The first preset condition and / or the second preset condition for judgment are adjusted based on the wear state information, wherein the first preset condition may correspond to the threshold requirements of amplitude ratio and / or energy ratio, and the second preset condition may correspond to the threshold requirements of time-frequency correlation. When the frequency overlap and / or the amplitude relationship at the corresponding frequency meets the adjusted first preset condition, and the time-frequency correlation meets the adjusted second preset condition, and the time lag falls within the expected time lag range, it is determined that the current vibration is mainly induced by torque commands; otherwise, it is determined that the current vibration is mainly caused by external cutting force fluctuations, thus obtaining the final judgment result.

[0026] The solution presented in this application significantly improves the accuracy of vibration source identification by introducing multi-dimensional information and more complex discrimination logic. In some preferred embodiments, a specific example is provided below: Suppose that during a CNC turning process, a turret with a built-in motor direct drive exhibits significant vibration. First, the system acquires information about the current tool wear status, such as through visual inspection or current signal analysis, determining that the tool is at a moderate wear level. Simultaneously, the system has pre-acquired the mechanical characteristic parameters of the turret, including its first-order natural vibration frequency of 500Hz and corresponding damping ratio of 0.05. Next, the system performs time-frequency analysis on the acquired vibration signal and the torque command signal driving the built-in motor, obtaining their energy distribution at different time points and frequencies. For example, within a certain time window, the vibration signal exhibits a significant energy peak at 480Hz, and the torque command signal also shows a corresponding energy component at 480Hz.

[0027] Based on these energy distributions, the system further calculated the time-frequency correlation between the vibration signal and the torque command signal at 480Hz, finding a correlation coefficient as high as 0.85, indicating a strong linear relationship between the two. Simultaneously, the calculated instantaneous phase difference showed that the torque command signal led the vibration signal by approximately 15 degrees. Based on the system's mechanical characteristic parameters, this 15-degree phase difference falls precisely within the expected time lag range (e.g., 10 to 20 degrees) for the turret at 480Hz. At this point, the system synthesizes the information for judgment. Since the tool is under moderate wear, the system fine-tuned the first preset condition (frequency overlap and amplitude relationship) and the second preset condition (time-frequency correlation). Under the adjusted conditions, the frequency overlap and amplitude relationship at 480Hz satisfy the first preset condition, the time-frequency correlation of 0.85 satisfies the second preset condition, and the 15-degree instantaneous phase difference falls within the expected time lag range. Therefore, the system determines that the current vibration is mainly induced by the torque command. Based on this judgment result, the torque command output characteristic adjustment module will adjust the torque command, for example by introducing a notch filter at 480Hz to attenuate the frequency component, thereby effectively suppressing vibration.

[0028] Conversely, if in another scenario, the vibration signal exhibits an energy peak at 300Hz, but the torque command signal has negligible energy at that frequency, or low time-frequency correlation (e.g., 0.3), or the instantaneous phase difference far exceeds the expected time lag range (e.g., the torque command lags behind the vibration), the system will determine that the current vibration is primarily caused by external cutting force fluctuations. In this case, the torque compensation strategy adjustment module will enhance its ability to compensate for external cutting force fluctuations to suppress vibration and maintain a stable cutting process.

[0029] In another embodiment of this application, it is further proposed that obtaining tool wear status information may include the following steps: S3110: Obtain the motor current signal fed back by the built-in motor servo driver; S3120: Perform frequency analysis on the motor current signal to obtain the motor current frequency spectrum, wherein the motor current frequency spectrum includes at least the amplitude spectrum and / or power spectrum of each frequency point within a predetermined frequency range. S3130: Obtain cutting parameter information during the current machining process, including feed rate, spindle speed and / or depth of cut; S3140: Based on the cutting parameter information, determine the cutting force frequency characteristic reference for comparison. The cutting force frequency characteristic reference shall include at least the reference frequency spectrum corresponding to the cutting parameter information and the frequency band division information of the reference frequency spectrum. S3150: Compare the motor current frequency spectrum with the cutting force frequency characteristic reference to obtain the comparison result, wherein the comparison result includes at least the difference measure between the motor current frequency spectrum and the reference frequency spectrum, and the difference measure is Euclidean distance, correlation coefficient and / or spectral energy difference. S3160: Calculate the wear characteristic index to characterize the degree of wear based on the difference measure; S3170: Based on the correspondence between the wear characteristic index and the preset wear threshold range, determine the wear level information to obtain wear status information.

[0030] Specifically, obtaining the motor current signal from the built-in motor servo driver refers to acquiring the current data of the built-in motor in real time during operation through the current sensor inside the built-in motor servo driver or an external current transformer. This current signal directly reflects the load changes borne by the motor, including load fluctuations caused by cutting forces.

[0031] Frequency analysis of motor current signals yields the motor current frequency spectrum. This can be understood as converting the time-domain motor current signal into frequency-domain spectral data using Fast Fourier Transform (FFT) or other frequency analysis methods. The motor current frequency spectrum includes at least all frequency components within a predetermined frequency range, along with their corresponding amplitude and / or power spectra. This spectrum reveals periodic or aperiodic components within the current signal. The predetermined frequency range can be set based on the common excitation frequency, servo bandwidth, and sampling rate of the machine tool / tool ​​system to ensure coverage of wear-related characteristic frequency bands and to maintain consistency with the reference frequency spectrum used for subsequent comparisons on the frequency axis to avoid distortion in difference measurement.

[0032] In practical applications, acquiring cutting parameter information during the current machining process specifically refers to obtaining process parameters such as current feed rate, spindle speed, and / or depth of cut from the machine tool's CNC system or sensors. These parameters are key factors affecting cutting force and tool wear. Furthermore, based on the cutting parameter information, a cutting force frequency characteristic reference is determined for comparison. The purpose is to establish a cutting force frequency characteristic benchmark under ideal or known wear conditions. This cutting force frequency characteristic reference includes at least a reference frequency spectrum corresponding to the current cutting parameter information and frequency band division information for that reference frequency spectrum. The frequency band division information is used to indicate the boundaries of the frequency bands that need to be compared and the weight or priority of each frequency band (optional), so as to differentiate between different wear-sensitive frequency bands. The reference frequency spectrum can be established through prior experiments, simulations, or empirical data. It reflects the typical cutting force frequency response of a healthy tool or a tool with a specific degree of wear under specific cutting parameters. It can be collected once when replacing a new tool or machining the same type of workpiece for the first time as a reference to reduce the error caused by parameter drift. The reference frequency spectrum corresponding to "replacing a new tool or machining the same type of workpiece for the first time" can be used as a wear-free reference so that the wear characteristic index changes monotonically as wear intensifies.

[0033] The comparison involves quantifying the difference between the actual motor current frequency spectrum and the theoretical or reference cutting force frequency characteristic. The comparison results include at least a measure of the difference between the motor current frequency spectrum and the reference frequency spectrum, such as Euclidean distance, correlation coefficient, and / or spectral energy difference. Euclidean distance measures the overall difference in the shapes of the two spectra, the correlation coefficient measures the similarity between the two spectra, and the spectral energy difference reflects the increase or decrease in energy within a specific frequency band. Furthermore, the difference measures can be weighted and summarized by frequency band (optional) to highlight wear-sensitive frequency bands, taking into account frequency band division information.

[0034] Therefore, the wear characteristic index, which is used to characterize the degree of wear, is calculated based on the difference measurement value. The purpose is to transform the complex spectral differences into a simple numerical index. The wear characteristic index can be directly calculated from the difference measurement value or obtained by combining multiple difference measurement values ​​according to a preset rule, so as to ensure that the index monotonically reflects the wear aggravation trend. Furthermore, the wear characteristic index under different cutting parameters can be made comparable through normalization.

[0035] Finally, based on the correspondence between the wear characteristic index and the preset wear threshold range, the wear level information is determined to obtain the wear state information. The preset wear threshold range is set based on experimental data or experience. For example, when the wear characteristic index falls into a certain range, the tool is judged as slightly worn; when it falls into another range, it is judged as moderately or severely worn. When the wear characteristic index falls exactly on the threshold boundary, it can be assigned to one of the adjacent ranges according to preset rules (e.g., assigned to a higher wear level) to avoid judgment gaps and maintain consistency in level judgment. In this way, the wear state of the tool can be quantified, providing key input for subsequent vibration discrimination, and used to adjust the first preset condition and / or the second preset condition based on the wear state information.

[0036] In another embodiment of this application, after obtaining the motor current signal fed back by the built-in motor servo driver, the process includes: S3111: Filter the motor current signal to suppress electrical noise and / or transient interference, obtaining a filtered motor current signal. The filtering process includes: S3112: Obtain motor load information, which includes one or more of the following: load current, estimated load torque, and / or load power; S3113: Obtain information about the surrounding electrical environment, including electromagnetic interference intensity, power supply voltage fluctuation amplitude and / or power supply ripple amplitude; S3114: Identify the frequency components and intensity of industrial noise in the motor current signal based on motor load information, cutting parameter information, and surrounding electrical environment information; S3115: Adjust the filter parameters according to the frequency components and intensity of industrial noise. The filter parameters include cutoff frequency, center frequency, bandwidth, filter order and / or attenuation coefficient. S3116: Based on the adjusted filter parameters, the motor current signal is filtered to obtain the filtered motor current signal, which is used for subsequent frequency analysis to obtain the motor current frequency spectrum for comparison.

[0037] Specifically, filtering aims to separate and remove unwanted noise components, particularly electrical noise and transient interference, from the raw motor current signal. Electrical noise typically refers to high-frequency or broadband random signals superimposed on the useful signal, generated by the power supply, driver, or other electrical equipment; transient interference may be caused by switching actions, load changes, etc., and manifests as short-duration, high-amplitude pulses. Effective filtering ensures that the motor current signal used in subsequent analysis is cleaner, more accurately reflects the changes in cutting force caused by tool wear, and reduces the interference of noise on the motor current frequency spectrum and the calculation of wear characteristic indices.

[0038] Motor load information can be understood as parameters reflecting the current operating state of the motor. For example, the load current directly indicates the power demand of the motor output, the estimated load torque reflects the torque required for the motor to overcome external resistance, and the load power integrates current and voltage information. Motor load information can be directly output by the servo driver or calculated from the motor current and voltage signals (if applicable), and is used to characterize the harmonic noise and amplitude variation trends that may appear in the motor current signal under different operating conditions. This information helps to evaluate the motor's operating characteristics under different conditions, thereby indirectly inferring the possible types and intensities of noise.

[0039] In practical applications, ambient electrical environment information specifically refers to external electromagnetic factors that affect the quality of motor current signals. Electromagnetic interference intensity can quantify the degree of electromagnetic wave coupling to signals in the environment; power supply voltage fluctuations and power supply ripple amplitude directly affect the output stability of the motor driver, thus leading to corresponding noise components in the current signal. By acquiring this information, we can gain a more comprehensive understanding of the sources and characteristics of noise and provide prior constraints (such as possible interference frequency bands or ripple-related frequency ranges) for identifying "noise frequency components and intensity."

[0040] Based on motor load information, cutting parameter information, and surrounding electrical environment information, the frequency components and intensity of industrial noise in the motor current signal can be identified. For example, under heavy load or high-speed cutting, the motor load information indicates higher current and torque, which may be accompanied by harmonic noise of specific frequencies. Simultaneously, if strong electromagnetic interference exists in the surrounding electrical environment, broadband noise will be introduced into the current signal. By comprehensively analyzing this information, a noise model can be established or machine learning methods can be used to accurately identify the main frequency components and their intensity under the current operating conditions. The intensity of industrial noise can be characterized by amplitude, power spectral density, or dB scaling, and corresponds one-to-one with frequency components within a predetermined frequency range.

[0041] Furthermore, based on the identified frequency components and intensity of the industrial noise, the filter parameters can be adaptively adjusted. For example, if the main noise is identified as being concentrated in a specific frequency range, the filter's center frequency or cutoff frequency can be set near that range, and the bandwidth adjusted to precisely cover the noise band. The filter order and attenuation coefficient are used to control the filter's steepness and suppression effect. Through this adaptive adjustment, the filter can be optimized for the noise characteristics under current operating conditions, achieving efficient and accurate noise suppression, and synchronously updating the filter parameters to maintain the filtering effect when the noise frequency drifts or its intensity changes.

[0042] Finally, based on the adjusted filter parameters, the motor current signal is filtered to obtain the filtered motor current signal. This filtered signal has a higher signal-to-noise ratio and can more accurately reflect the changes in cutting force between the tool and the workpiece, thus providing a reliable data foundation for subsequent frequency analysis.

[0043] In another embodiment of this application, S4000 specifically includes: S4100: When the machine tool is in the non-cutting stage, it applies a frequency sweeping small torque excitation with a preset amplitude to the built-in motor and simultaneously collects the vibration signal at this time; S4200: Based on the frequency response relationship between the sweep frequency micro-torque excitation and the vibration signal at this time, the inherent vibration frequency of the built-in motor direct drive turret is identified; S4300: Adaptively updates the filter parameters used to suppress vibration-induced frequency components according to the natural vibration frequency, so that the frequency components of the torque command at the natural vibration frequency are attenuated to suppress vibration.

[0044] Specifically, when the machine tool is in a non-cutting phase, i.e., without external cutting force, a frequency-sweeping micro-torque excitation with a preset amplitude is applied to the built-in motor. This frequency-sweeping micro-torque excitation refers to a torque signal with continuously or stepwise varying frequency components within a certain frequency range. Its amplitude is set sufficiently small to avoid damaging the turret that directly drives the built-in motor, yet sufficient to induce a vibration response. The preset amplitude can be selected from 1% to 5% of the motor's rated torque, and the frequency sweep range can be set according to a predetermined frequency band or a possible natural vibration frequency range. During this process, the vibration signal is simultaneously acquired, and the reference signal of the frequency-sweeping micro-torque excitation is simultaneously recorded to ensure that the subsequent frequency response calculation has an input-output correspondence.

[0045] Based on the applied sweep-frequency micro-torque excitation and the synchronously acquired vibration signal, the frequency response of the built-in motor direct-drive turret can be calculated. The frequency response characterizes the system's response characteristics under different frequency excitations, typically represented by the frequency response function (FRF). The FRF can be estimated from the frequency domain ratio of the vibration signal to the sweep-frequency micro-torque excitation or its transfer function. By analyzing this frequency response, the inherent vibration frequencies of the built-in motor direct-drive turret can be identified. These frequencies are usually represented by peak points in the frequency response function, indicating the system's tendency to resonate at specific frequencies. Furthermore, the corresponding damping characteristics can be estimated based on the changing trends of the peak bandwidth and peak amplitude, aiding in setting the bandwidth and attenuation depth of the filtering parameters.

[0046] Furthermore, once the inherent vibration frequency of the built-in motor direct-drive turret is identified, the filter parameters used to suppress vibration-inducing frequency components can be adaptively updated. These filter parameters may include, but are not limited to, the center frequency, bandwidth, and attenuation depth of the filter (e.g., a notch filter). By precisely setting the center frequency or attenuation frequency of the filter at the identified inherent vibration frequency, the frequency components in the torque command that coincide with these inherent vibration frequencies can be selectively attenuated. The filter parameters are updated to apply directly to the digital filter or equivalent filter element in the torque command channel, ensuring that the updated parameters directly affect subsequent torque command signals and form a closed-loop "identification-update-suppression" chain. This attenuation aims to reduce the excitation effect of the torque command on the built-in motor direct-drive turret, thereby effectively suppressing vibration.

[0047] In another embodiment of this application, the step of adjusting filter parameters according to the frequency components and intensity of industrial noise includes: S3115-1: Based on the frequency components and intensity of industrial noise, determine the target values ​​of filter parameters for suppressing industrial noise, including center frequency, bandwidth and / or cutoff frequency; S3115-2: Calculate the instantaneous rate of change of the motor current signal based on the amplitude spectrum and / or power spectrum of the motor current signal within adjacent predetermined time windows; S3115-3: Update the frequency components and intensity of industrial noise within adjacent predetermined time windows, and calculate the noise change rate based on the update results of adjacent predetermined time windows. The noise change rate includes the noise center frequency change rate, the noise energy change rate, and / or the noise bandwidth change rate. S3115-4: Based on the comparison results between the instantaneous rate of change and / or the noise change rate and the preset noise change threshold, determine whether the industrial noise has changed rapidly; S3115-5: In response to the judgment result indicating a rapid change in industrial noise, increase the adjustment step size of the filter parameters from the current value to the target value of the filter parameters and shorten the update cycle of the filter parameters. S3115-6: In response to the judgment result indicating that the industrial noise has not changed rapidly, the filter parameters are made to approach the target value of the filter parameters from the current value by a preset adjustment step size and a preset update cycle; S3115-7: Based on cutting parameter information and motor load information, predict the frequency components and intensity of industrial noise in the next predetermined time window to obtain the predicted value of noise change trend. Before the arrival of the next predetermined time window, preset the filter parameters to the predicted parameter values ​​corresponding to the predicted noise change trend values. The predicted parameter values ​​are used as the initial values ​​of the filter parameters in the next predetermined time window. The filter parameters are then updated to the target values ​​of the filter parameters in the next predetermined time window based on the frequency components and intensity of industrial noise identified in real time.

[0048] To achieve adaptive adjustment of dynamic noise, it is necessary to calculate the instantaneous rate of change of the motor current signal and the noise change rate. The instantaneous rate of change of the motor current signal can be obtained by comparing the amplitude spectrum and / or power spectrum within adjacent predetermined time windows, for example, by calculating the change in the Euclidean distance or correlation coefficient between spectral lines, and normalizing this change to a rate of change per unit time for consistent comparison with a preset noise change threshold. The noise change rate is obtained by calculating the rate of change of the noise center frequency, noise energy, and / or noise bandwidth after updating the frequency components and intensity of industrial noise within adjacent predetermined time windows. These rates of change reflect the dynamic evolution trend of noise characteristics, where the rate of change of the noise center frequency, the rate of change of noise energy, and / or the rate of change of noise bandwidth can be used as components of the noise change rate.

[0049] In practical applications, by comparing the instantaneous rate of change and / or the rate of noise change with a preset noise change threshold, it can be determined whether industrial noise has changed rapidly. For example, if the rate of change of the noise center frequency exceeds the preset threshold, the noise is considered to be changing rapidly. In response to the judgment indicating a rapid change in industrial noise, to quickly adapt to the new noise characteristics, the adjustment step size of the filter parameters as they approach the target value is increased, while the update cycle of the filter parameters is shortened, thereby accelerating the filter's response speed. Conversely, if the judgment indicates that the industrial noise has not changed rapidly, parameter approximation is performed with a preset adjustment step size and update cycle to maintain filter stability. The values ​​of the adjustment step size and update cycle can respectively limit the maximum single change and the maximum update time interval of the filter parameters to avoid parameter oscillation.

[0050] Furthermore, this application introduces a prediction mechanism. Based on current cutting parameter information and motor load information, the frequency components and intensity of industrial noise within the next predetermined time window can be predicted, thereby obtaining a predicted value for noise change trends. For example, by establishing a mapping model between operating conditions and noise characteristics, future noise conditions can be predicted based on the current machining conditions. Based on this prediction value, before the arrival of the next predetermined time window, the filter parameters are pre-set to the predicted parameter values ​​corresponding to the predicted values. This predicted parameter value will serve as the initial value of the filter parameters within the next predetermined time window, allowing the filter to have a starting point closer to the optimal state when entering a new operating condition. Subsequently, within this predetermined time window, the filter parameters will be fine-tuned based on the frequency components and intensity of the industrial noise identified in real time, updating towards the final target value of the filter parameters to ensure the accuracy of filtering. Thus, the "predicted parameter value - target value of filter parameters" corresponds in time to the two-stage update logic of "initial value of the next predetermined time window - convergence target value within this time window," forming a closed-loop update link of "prediction preset - real-time correction - target value convergence."

[0051] In some preferred embodiments, the following specific example illustrates the situation: Suppose that during heavy cutting on a CNC machine tool, the frequent changes in cutting depth and feed rate cause the motor load of the built-in motor direct drive turret to exhibit periodic fluctuations, which in turn induces rapid changes in the frequency components and intensity of industrial noise.

[0052] First, the system continuously collects motor current signals and performs time-frequency analysis to obtain the amplitude and power spectra within adjacent predetermined time windows. For example, in the first time window, industrial noise is identified as mainly concentrated at 500Hz, with an energy intensity of X. In the second time window, due to changes in cutting conditions, the noise center frequency rapidly shifts to 600Hz, and the energy intensity changes to Y.

[0053] At this point, the system calculates the instantaneous rate of change of the motor current signal and the rate of change of noise. For example, by comparing the spectra of two time windows, it is found that both the rate of change of the noise center frequency and the rate of change of noise energy exceed the preset noise change threshold, thus determining that the industrial noise is changing rapidly.

[0054] In response to this assessment, the system immediately increases the adjustment step size of the filter parameters (such as center frequency and bandwidth) as they approach the target value (e.g., new filter parameters centered at 600Hz), and shortens the filter parameter update period from the usual 100ms to 20ms. This allows the filter to quickly adjust its center frequency and bandwidth to match the new noise characteristics, thereby rapidly and effectively suppressing noise near 600Hz.

[0055] Meanwhile, based on current cutting parameters (such as the feed rate and depth of cut for the upcoming finishing stage) and motor load information, the system predicts that industrial noise will mainly concentrate at 450Hz and its intensity will decrease in the next time window. Therefore, before the next time window arrives, the system will preset the filter parameters to predicted parameter values ​​centered at 450Hz with a narrow bandwidth. When the next time window arrives, the filter will start working with this predicted parameter value as the initial value and will be fine-tuned based on the real-time identified industrial noise.

[0056] In another embodiment of this application, the step of predicting the frequency components and intensity of industrial noise within the next predetermined time window based on cutting parameter information and motor load information, and obtaining a predicted value of noise change trend, includes: S3115-71: Obtain a condition feature vector composed of cutting parameter information and motor load information of the current machining condition; S3115-72: Compare the working condition feature vector with the preset working condition feature library to obtain the first comparison result. The working condition feature library includes multiple working condition modes. Each working condition mode includes a working condition feature vector template, the industrial noise frequency component corresponding to the working condition feature vector template, the industrial noise intensity, and its trend parameters over time. S3115-73: Based on the first comparison result, the working condition mode with the highest matching degree is determined as the matching mode, and the noise change trend prediction value is generated based on the trend parameter of the matching mode. S3115-74: In response to the first comparison result indicating that the matching degree of the matching mode is lower than the preset matching threshold, the adaptive adjustment mode is activated: Based on the time-frequency analysis results of the motor current signal, the instantaneous frequency components and intensity of industrial noise are monitored, and the difference between the instantaneous frequency components and intensity and the noise change trend prediction value is calculated. Based on the difference calculation results, the noise change trend prediction value is corrected to obtain the corrected noise change trend prediction value, which is used to determine and preset the prediction parameter value. S3115-75: Furthermore, record noise change data under operating conditions where the matching degree is lower than the preset matching threshold. The noise change data shall include at least the instantaneous frequency components, intensity and corresponding timestamps, for updating the operating condition feature library.

[0057] Specifically, the operating condition feature vector can be understood as a quantitative description of the current machining state. It is composed of cutting parameter information (such as feed rate, spindle speed, depth of cut, etc.) and motor load information (such as load current, estimated load torque, load power, etc.). Each component can be normalized or weighted according to preset dimensions to avoid comparison bias caused by different unit magnitudes. This information can comprehensively reflect the machine tool's operating status and external interference environment.

[0058] The operating condition feature library is a pre-built database that stores various typical processing operating condition modes. Each operating condition mode not only contains an operating condition feature vector template for comparison with real-time operating condition feature vectors, but also associates it with typical frequency components of industrial noise under that operating condition, industrial noise intensity, and trend parameters of these noises changing over time. The trend parameters can be mathematical models or empirical curves of the changes in noise center frequency, bandwidth, or energy intensity over time, and can be mapped to a prediction output format for the "next predetermined time window" (such as predicted values ​​or rates of change of center frequency / bandwidth / intensity) to directly generate predicted values ​​of noise change trends.

[0059] In practical applications, the first comparison result is obtained by comparing the real-time acquired operating condition feature vector with all operating condition patterns in the operating condition feature database, for example, by calculating the similarity using methods such as Euclidean distance and cosine similarity. Based on this comparison result, the operating condition pattern that best matches the current operating condition can be determined as the matching pattern. The matching degree can be obtained by normalizing the similarity or directly expressed by the similarity index to ensure that the "matching degree threshold" is consistent with the comparison index. Once the matching pattern is determined, the preset trend parameters in the pattern can be used to generate the predicted value of the noise change trend in the next predetermined time window. The predicted value of the noise change trend includes at least the predicted values ​​of the frequency components and intensity of industrial noise in the next predetermined time window.

[0060] Furthermore, when the matching degree of the first comparison result indicates that the matching mode is lower than the preset matching threshold, it suggests that the current operating condition may be a new, insufficiently learned condition, or that the noise characteristics have changed significantly. At this point, the system will activate the adaptive adjustment mode, which is used to correct the generated noise change trend prediction value online. In this mode, based on the time-frequency analysis results of the motor current signal, the instantaneous frequency components and intensity of industrial noise are monitored in real time. Subsequently, these instantaneous monitoring values ​​are compared with the previously generated noise change trend prediction value, for example, calculating the absolute or relative error between the two. Based on the difference calculation results, the noise change trend prediction value is corrected to make it closer to the actual noise situation, thus obtaining the corrected noise change trend prediction value. This corrected prediction value will be used to determine and preset the filter parameters for the next predetermined time window, and will serve as the initial value of the filter parameters for the next predetermined time window to reduce filter parameter update lag.

[0061] Furthermore, to continuously optimize predictive capabilities, when the matching degree falls below a preset matching threshold, the system also records current noise change data, including instantaneous frequency components, intensity, and corresponding timestamps. This data will be used to subsequently update the operating condition feature library, enabling the system to learn new operating condition patterns or update the trend parameters of existing patterns, improving its adaptability to unknown or complex operating conditions. The recorded noise change data can be stored together with the operating condition feature vector at the trigger time to support the subsequent formation of new operating condition feature vector templates or the correction of existing templates.

[0062] The solution in this application constructs a working condition feature vector and compares it with a preset working condition feature library, which can quickly identify the current processing working condition and generate a preliminary noise change trend prediction value based on the trend parameters of the matched working condition pattern. In some preferred embodiments: Suppose that during a certain machining process, the machine tool is performing high-speed milling, with cutting parameters including a spindle speed of 10,000 rpm, a feed rate of 5,000 mm / min, and a depth of cut of 2 mm. Simultaneously, the motor load information shows a load current of 15 A. The system combines this information into a working condition feature vector. This feature vector is then input into a preset working condition feature library for comparison. Assume that a high-speed milling mode exists in the feature library, and its feature vector template matches the current working condition feature vector with a 95% match rate, exceeding a preset matching threshold (e.g., 80%). This high-speed milling mode records the typical frequency components (e.g., 500 Hz, 1000 Hz) and their intensities of industrial noise under this working condition, as well as the trend parameters of these noises over time (e.g., the energy intensity of 500 Hz noise increases slightly at the beginning of machining and then stabilizes). Based on these trend parameters, the system generates a predicted value for the noise change trend within the next predetermined time window. For example, it predicts that the energy intensity of 500 Hz noise will remain at a stable level within the next time window. Based on this prediction, filter parameters (such as center frequency and bandwidth) are preset before the next time window arrives, so that the filter can be prepared in advance and effectively suppress the upcoming industrial noise.

[0063] Furthermore, in another scenario, suppose the machine tool suddenly switches to a new material processing method, or severe tool wear causes a drastic change in cutting force, resulting in the current operating condition feature vector matching all patterns in the operating condition feature library below a preset matching threshold. In this case, the system will activate an adaptive adjustment mode. In this mode, the system continuously performs time-frequency analysis on the motor current signal, monitoring the instantaneous frequency components and intensity of industrial noise in real time. For example, if a new 1200 Hz noise component is found in addition to the expected 500 Hz and 1000 Hz noise, and its intensity is much higher than the predicted value, the system will calculate the difference between these instantaneous monitoring results and the preliminary noise change trend prediction, and correct the prediction value based on the calculation results. The corrected prediction value will include the new 1200 Hz noise component and its intensity, and will be used to redetermine and preset filter parameters. Simultaneously, the system will record the noise change data under the current operating condition (including the new 1200 Hz noise component, its intensity, and the corresponding timestamp) for subsequent updates to the operating condition feature library, enabling the system to learn and adapt to this new processing condition.

[0064] In another embodiment of this application, updating the operating condition feature library is further proposed, including: A1: Preprocess the noise variation data to remove duplicate and outlier data, and obtain the preprocessed noise variation data; A2: Extract key features from the preprocessed noise variation data. Key features include the frequency range, energy intensity, and duration of the noise. A3: Calculate the feature similarity between the preprocessed noise change data based on key features, and merge the data with feature similarity higher than the preset similarity threshold to form merged noise change data; A4: Compare the merged noise variation data with the existing operating condition patterns in the operating condition feature library to obtain the second comparison result: A5: In response to the second comparison result indicating that the similarity between the merged noise change data and the existing operating condition mode is lower than the preset mode threshold, the merged noise change data and its corresponding operating condition feature vector are written together as a new mode into the operating condition feature library. A6: In response to the second comparison result indicating that the similarity is not lower than the preset mode threshold, the trend parameters of the existing operating mode are updated based on the merged noise change data. The trend parameters include at least the noise center frequency, noise bandwidth and / or the rate of change of noise energy intensity.

[0065] Specifically, preprocessing noise-varying data refers to the process of identifying and removing redundant or erroneous data points within the recorded noise-varying data, which may be caused by sensor errors, transmission interference, or system transient fluctuations. Preprocessing aims to use data cleaning techniques, such as deduplication algorithms (e.g., uniqueness checks based on timestamps and feature values) and outlier detection algorithms (e.g., those based on statistical methods or machine learning models), to ensure the data quality and accuracy of subsequent analyses.

[0066] Extracting key features from preprocessed noise variation data involves abstracting and summarizing the data to capture its core attributes. The frequency range of noise can be understood as the frequency band where noise energy is mainly concentrated; energy intensity can be understood as the total energy or average amplitude within that frequency band; and duration represents the length of time from the start to the end of the noise event. These features can be extracted by performing time-frequency analysis (such as short-time Fourier transform or wavelet transform) on the preprocessed noise variation data, combined with threshold judgment or pattern recognition algorithms. The frequency range can be determined by continuous frequency bands where the energy exceeds a preset energy threshold; the energy intensity can be characterized by the power spectral density integral or peak / mean amplitude within that frequency band; and the duration can be determined by the length of a continuous time window that meets the threshold condition. The aim is to simplify data representation and highlight the essential characteristics of noise.

[0067] Furthermore, based on key features, the feature similarity between the preprocessed noise variation data is calculated, and data with feature similarity higher than a preset similarity threshold are merged to form merged noise variation data. Feature similarity is used to quantify the degree of similarity between different noise variation data, and can be achieved through various similarity measurement methods, such as Euclidean distance, cosine similarity, or Pearson correlation coefficient. When using Euclidean distance, the distance can first be mapped to similarity (e.g., through normalization or exponential mapping) to align with the judgment criterion of "higher than the preset similarity threshold." When the calculated similarity is higher than the preset similarity threshold, it indicates that these data may originate from the same noise event or have similar noise characteristics, and therefore can be merged to reduce data redundancy and form a more representative noise pattern. The merged data can be an average, weighted average, or aggregation of feature vectors of the original data, and the corresponding timestamp intervals can be merged simultaneously to retain the occurrence time information of the noise pattern. The purpose is to compress the data volume and improve data representativeness.

[0068] The process involves comparing the merged noise variation data with existing operating condition patterns in the operating condition feature library to obtain a second comparison result. This comparison can also employ similarity measurement methods, such as calculating the similarity between the merged noise variation data and the noise characteristic parameters (e.g., frequency range, energy intensity, duration, and trend parameters) associated with each existing operating condition pattern. This avoids the problem of "lack of noise pattern determination criteria" that can result from comparing only with the operating condition feature vector template. The aim is to determine whether the new data belongs to the category of known patterns.

[0069] In response to the second comparison result indicating that the similarity between the merged noise variation data and the existing operating condition pattern is lower than a preset pattern threshold, the merged noise variation data and its corresponding operating condition feature vector are written into the operating condition feature library as a new pattern. This means that the data represents a new noise pattern not covered by the existing library. At this time, the merged noise variation data and its corresponding operating condition feature vector (i.e., the processing operating condition parameters that cause this noise pattern) are added to the operating condition feature library as a completely new pattern. The new pattern includes at least: a new operating condition feature vector template and the initial values ​​of the associated noise frequency components, noise intensity, and trend parameters. The purpose is to expand the coverage of the operating condition feature library, enabling it to identify more diverse noise conditions.

[0070] In practical applications, in response to the second comparison result indicating that the similarity is not lower than a preset mode threshold, the trend parameters of the existing operating condition mode are updated based on the merged noise change data. When the comparison result indicates that the similarity between the merged noise change data and the existing operating condition mode is not lower than the preset mode threshold, it indicates that the data is highly similar to the existing mode. At this time, instead of creating a new mode, the trend parameters of the existing operating condition mode are updated using the new data. Trend parameters may include noise center frequency, noise bandwidth, and / or the rate of change of noise energy intensity, which reflect the dynamic law of noise characteristics changing with time or operating conditions. The update can be implemented through weighted averaging, exponential smoothing, or other adaptive algorithms to make the trend parameters of the existing mode more accurately reflect the latest noise change law. During the update, higher weights can be assigned to the new data to improve the tracking ability of recent operating condition changes. The purpose is to make the trend parameters of the existing mode more accurately reflect the latest noise change law and improve prediction accuracy.

[0071] In another embodiment of this application, obtaining motor load information may include the following steps: B1: Synchronous sampling to acquire the motor current signal and motor voltage signal fed back by the built-in motor servo driver; B2: Real-time power calculation based on motor current and motor voltage signals to obtain instantaneous motor power data; B3: Perform frequency analysis on instantaneous motor power data to obtain the power frequency spectrum; B4: Based on the power frequency spectrum, specific frequency components related to instantaneous impact loads and / or periodic harmonic loads are identified, and the influence of specific frequency components is suppressed in the power frequency spectrum to obtain a power frequency spectrum that reflects the actual cutting load. B5: Calculate motor load information based on the power frequency spectrum that reflects the actual cutting load.

[0072] Specifically, to accurately obtain motor load information, it is first necessary to synchronously sample the motor current and voltage signals fed back by the built-in motor servo driver. Synchronous sampling means acquiring these two signals at the same time point or in a strictly synchronized manner to ensure the accurate time correspondence between them. Furthermore, based on the acquired motor current and voltage signals, real-time power calculation can be performed to obtain instantaneous motor power data. This real-time power calculation is based on the synchronously sampled motor current and voltage signals. Real-time power calculation is typically achieved by multiplying the instantaneous current value by the instantaneous voltage value, which reflects the actual power consumption of the motor at each sampling moment and serves as one of the fundamental quantities for motor load information.

[0073] Subsequently, frequency analysis is performed on the instantaneous motor power data to obtain the power frequency spectrum. Frequency analysis can employ methods such as Fast Fourier Transform (FFT) to convert the instantaneous power data in the time domain into a power frequency spectrum in the frequency domain, thereby revealing the various frequency components contained in the power signal and their corresponding amplitudes or energies.

[0074] In actual machining processes, the motor load, in addition to the actual cutting load, may also be affected by transient impact loads and / or periodic harmonic loads. Therefore, it is necessary to identify specific frequency components related to these non-cutting loads based on the power frequency spectrum. A specific frequency component is not a single frequency value, but rather refers to one or more discrete frequency points and / or continuous frequency bands in the power frequency spectrum corresponding to transient impact loads and / or periodic harmonic loads, used to characterize the disturbing frequency features unrelated to the actual cutting load. For example, a transient impact load may manifest as a broadband energy surge in the power frequency spectrum, while a periodic harmonic load may manifest as a spike at a specific discrete frequency.

[0075] A specific frequency component must meet at least one or more of the following criteria: it exhibits a spectral peak (discrete frequency point) above a preset threshold and / or a sudden increase in spectral energy (continuous frequency band) in the power frequency spectrum. Specifically, for periodic harmonic loads, the specific frequency component may include spectral peaks located at a fundamental frequency and its integer multiples. The fundamental frequency can be determined by the main spectral peak within a preset harmonic search band in the power frequency spectrum, and / or calculated from the motor speed, pulse frequency, or power supply frequency output by the servo drive / CNC system, and the spectral peaks at their integer multiples are searched using this fundamental frequency as an index. For transient impulsive loads, the specific frequency component may include a broadband spectral energy surge within a preset impulsive frequency band, which can be obtained through trial-and-error calibration or statistical analysis of historical power frequency spectra.

[0076] Furthermore, when identifying specific frequency components, the amplitude and / or energy of each frequency point are traversed or peaked based on the power frequency spectrum to obtain a candidate peak set. The amplitude and / or energy of the candidate peaks are compared with a preset threshold to determine the frequency points that meet the threshold conditions as part of the specific frequency component. Moreover, when a candidate peak continuously meets the threshold conditions at adjacent frequency points, the continuous frequency band is determined as the frequency band corresponding to the specific frequency component. To avoid misclassifying low-frequency components related to the actual cutting load as specific frequency components, the search for the candidate peak set is limited to a preset frequency range, which is jointly determined by the sampling rate, servo bandwidth, and historically statistically analyzed interference frequency bands.

[0077] The preset threshold is used to distinguish between power changes caused by actual cutting loads and interference changes caused by instantaneous impact loads and / or periodic harmonic loads. It can be determined in the following ways: by determining the threshold based on the background energy level of the historical power frequency spectrum, and / or by determining the threshold based on a preset noise / interference upper limit, and / or by obtaining the threshold through trial cutting calibration. The background energy level can be taken as the mean / median of the power spectral density within the target frequency band and its multiples, or the mean ± k times the standard deviation, where k is a preset coefficient. This allows those skilled in the art to reproduce the determination process for specific frequency components.

[0078] After identifying these specific frequency components, their influence needs to be suppressed in the power frequency spectrum. Suppressing the influence of specific frequency components in the power frequency spectrum includes: amplitude reduction or zeroing of the frequency points corresponding to the specific frequency components, and / or band-stop filtering of the frequency bands corresponding to the specific frequency components, and / or interpolation of adjacent frequency points to the suppressed frequency points / bands to restore spectral continuity. This yields a power frequency spectrum reflecting the actual cutting load, and motor load information is calculated based on this power frequency spectrum. Motor load information may include load current, estimated load torque, or load power, where load power can be calculated from the energy of the power frequency spectrum reflecting the actual cutting load within a predetermined frequency band, and the estimated load torque can be obtained by converting load power and motor speed, with the motor speed obtained from feedback by the servo driver.

[0079] Reference Figure 2 The specific embodiments of this application also disclose a built-in motor direct drive force turret adaptive control system, including: Vibration frequency information acquisition module 1 is used to collect vibration signals of the built-in motor direct drive turret during the processing, and perform frequency analysis on the vibration signals to obtain actual vibration frequency information characterizing the actual vibration state of the built-in motor direct drive turret. The actual vibration frequency information includes frequency components within a predetermined frequency band and the amplitude and / or energy value corresponding to each frequency component. The torque command frequency information acquisition module 2 is used to acquire the torque command signal driving the built-in motor, and perform frequency analysis on the torque command signal to obtain torque command frequency information characterizing the torque command change characteristics. The torque command frequency information includes frequency components within a predetermined frequency band and the amplitude and / or energy value corresponding to each frequency component. The vibration source discrimination module 3 is used to determine the frequency overlap and / or amplitude relationship at the corresponding frequency based on the actual vibration frequency information and the torque command frequency information, and to determine whether the current vibration of the built-in motor direct drive turret is mainly induced by the torque command or mainly caused by external cutting force fluctuations, and to obtain the discrimination result. The frequency overlap includes at least the overlapping frequency components of the two in the predetermined frequency band, and the amplitude relationship at the corresponding frequency includes at least the amplitude ratio and / or energy ratio of the vibration signal and the torque command signal at the overlapping frequency components. The torque command output characteristic adjustment module 4 is used to adjust the output characteristics of the torque command when the judgment result indicates that the current vibration is mainly induced by the torque command, so as to reduce the excitation effect of the vibration-inducing frequency component in the torque command on the built-in motor direct drive turret, thereby suppressing vibration; The torque compensation strategy adjustment module 5 is used to adjust the torque compensation strategy used to drive the built-in motor when the judgment result indicates that the current vibration is mainly caused by external cutting force fluctuations, so as to enhance the compensation capability for external cutting force fluctuations, thereby suppressing vibration and maintaining the stability of the cutting process.

[0080] This system aims to solve the problem of turret vibration induced by improper torque adjustment when facing high-frequency, rapidly fluctuating cutting forces, and the difficulty in identifying the source of vibration in traditional control systems. By intelligently analyzing vibration signals and torque command signals, the system can accurately identify the source of vibration and adjust the torque command output characteristics or torque compensation strategy accordingly. This effectively suppresses the vibration of the built-in motor direct-drive turret under complex working conditions, significantly improves machining accuracy and tool life, and maintains a stable cutting process.

[0081] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. An adaptive control method for a turret with built-in motor direct drive force, characterized in that, include: Vibration signals of the built-in motor direct drive turret during the processing are collected, and frequency analysis is performed on the vibration signals to obtain actual vibration frequency information characterizing the actual vibration state of the built-in motor direct drive turret. The actual vibration frequency information includes frequency components within a predetermined frequency band and the amplitude and / or energy value corresponding to each frequency component. The torque command signal driving the built-in motor is acquired, and the frequency of the torque command signal is analyzed to obtain torque command frequency information characterizing the torque command change characteristics. The torque command frequency information includes frequency components within a predetermined frequency band and the amplitude and / or energy value corresponding to each frequency component. Based on the actual vibration frequency information and the torque command frequency information, determine the frequency overlap and / or the amplitude relationship at the corresponding frequency, and determine whether the current vibration of the built-in motor direct drive turret is mainly induced by the torque command or mainly caused by external cutting force fluctuations, and obtain the determination result. The frequency overlap includes at least the overlapping frequency components of the two in a predetermined frequency band, and the amplitude relationship at the corresponding frequency includes at least the amplitude ratio and / or energy ratio of the vibration signal and the torque command signal at the overlapping frequency components. When the discrimination result indicates that the current vibration is mainly induced by the torque command, the output characteristics of the torque command are adjusted to reduce the excitation effect of the vibration-inducing frequency component in the torque command on the built-in motor direct drive turret, thereby suppressing the vibration; When the discrimination result indicates that the current vibration is mainly caused by external cutting force fluctuations, the torque compensation strategy used to drive the built-in motor is adjusted to enhance the compensation capability for external cutting force fluctuations, thereby suppressing vibrations and maintaining the stability of the cutting process. The step of determining the frequency overlap and / or amplitude relationship at corresponding frequencies based on the actual vibration frequency information and the torque command frequency information, and determining whether the current vibration of the built-in motor direct drive turret is mainly induced by the torque command or mainly caused by external cutting force fluctuations, to obtain a determination result includes: Obtain tool wear status information, including wear level information used to characterize the degree of wear; Obtain the mechanical characteristic parameters of the built-in motor direct drive turret, the mechanical characteristic parameters including the set of natural vibration frequencies of the built-in motor direct drive turret structure and / or the set of damping ratios corresponding to the natural vibration frequencies; Time-frequency analysis is performed on the vibration signal and the torque command signal to obtain the energy distribution of the vibration signal and the torque command signal at different time points and different frequencies, wherein the energy distribution includes the amplitude spectrum and / or power spectrum at each frequency point within a predetermined time window; Based on the energy distribution, the time-frequency correlation and instantaneous phase difference between the vibration signal and the torque command signal are calculated, wherein the time-frequency correlation includes the correlation coefficient and / or cross-correlation peak value within a predetermined frequency band; the instantaneous phase difference includes the phase difference sequence at corresponding frequency points within the predetermined frequency band and / or the time lag amount calculated from the phase difference. The frequency overlap and / or amplitude relationship at the corresponding frequency are determined based on the actual vibration frequency information and the torque command frequency information, and the amplitude ratio and / or energy ratio are extracted from the amplitude spectrum and / or power spectrum in the energy distribution; The determination is made by combining the frequency overlap and / or the amplitude relationship at the corresponding frequency, the time-frequency correlation, the instantaneous phase difference, the wear state information, and the mechanical characteristic parameters: The expected time lag range for judgment is determined based on the aforementioned mechanical characteristic parameters; The first and / or second preset conditions used for discrimination are adjusted based on the wear status information. When the frequency overlap and / or the amplitude relationship at the corresponding frequency meets the adjusted first preset condition, and the time-frequency correlation meets the adjusted second preset condition, and the time lag falls within the expected time lag range, it is determined that the current vibration is mainly induced by the torque command; otherwise, it is determined that the current vibration is mainly caused by external cutting force fluctuations, and a judgment result is obtained.

2. The adaptive control method for the built-in motor direct drive turret according to claim 1, characterized in that, The acquisition of tool wear status information includes: Obtain the motor current signal fed back by the built-in motor servo driver; Frequency analysis is performed on the motor current signal to obtain the motor current frequency spectrum, wherein the motor current frequency spectrum includes at least the amplitude spectrum and / or power spectrum of each frequency point within a predetermined frequency range; Obtain cutting parameter information during the current machining process, including feed rate, spindle speed and / or depth of cut; Based on the cutting parameter information, a cutting force frequency feature reference for comparison is determined. The cutting force frequency feature reference includes at least a reference frequency spectrum corresponding to the cutting parameter information and frequency band division information of the reference frequency spectrum. The motor current frequency spectrum is compared with the cutting force frequency characteristic reference to obtain a comparison result. The comparison result includes at least the difference measure between the motor current frequency spectrum and the reference frequency spectrum. The difference measure is Euclidean distance, correlation coefficient and / or spectral energy difference. The wear characteristic index, used to characterize the degree of wear, is calculated based on the difference metric value; Based on the correspondence between the wear characteristic index and the preset wear threshold range, the wear level information is determined to obtain the wear state information.

3. The adaptive control method for the built-in motor direct drive turret according to claim 2, characterized in that, After acquiring the motor current signal fed back by the built-in motor servo driver, the process includes: The motor current signal is filtered to suppress electrical noise and / or transient interference, resulting in a filtered motor current signal. The filtering process includes: Obtain motor load information, which includes one or more of load current, estimated load torque, and / or load power; Acquire surrounding electrical environment information, including electromagnetic interference intensity, power supply voltage fluctuation amplitude, and / or power supply ripple amplitude; Based on the motor load information, the cutting parameter information, and the surrounding electrical environment information, the frequency components and intensity of industrial noise in the motor current signal are identified; Based on the frequency components and intensity of the industrial noise, the filter parameters are adjusted, including cutoff frequency, center frequency, bandwidth, filter order, and / or attenuation coefficient. The motor current signal is filtered according to the adjusted filter parameters to obtain the filtered motor current signal, which is used for subsequent frequency analysis to obtain the motor current frequency spectrum for comparison.

4. The adaptive control method for a built-in motor direct drive turret according to claim 1, characterized in that, When the discrimination result indicates that the current vibration is mainly induced by the torque command, adjusting the output characteristics of the torque command includes: When the machine tool is in the non-cutting stage, a frequency sweeping small torque excitation of preset amplitude is applied to the built-in motor, and the vibration signal at this time is collected simultaneously; Based on the frequency response relationship between the frequency sweeping micro-torque excitation and the vibration signal at this time, the inherent vibration frequency of the built-in motor direct drive turret is identified; The filter parameters used to suppress the vibration frequency components are adaptively updated according to the natural vibration frequency, so that the frequency components of the torque command at the natural vibration frequency are attenuated to suppress vibration.

5. The adaptive control method for a built-in motor direct drive turret according to claim 3, characterized in that, The step of adjusting the filter parameters based on the frequency components and intensity of the industrial noise includes: Based on the frequency components and intensity of the industrial noise, target values ​​for filter parameters used to suppress the industrial noise are determined, including center frequency, bandwidth, and / or cutoff frequency. The instantaneous rate of change of the motor current signal is calculated based on the amplitude spectrum and / or power spectrum of the motor current signal within adjacent predetermined time windows. The frequency components and intensity of the industrial noise are updated within adjacent predetermined time windows, and the noise change rate is calculated based on the update results of the adjacent predetermined time windows. The noise change rate includes the noise center frequency change rate, the noise energy change rate, and / or the noise bandwidth change rate. Based on the comparison results between the instantaneous rate of change and / or the noise change rate and a preset noise change threshold, it is determined whether the industrial noise has changed rapidly; In response to the judgment result indicating that the industrial noise has changed rapidly, the adjustment step size of the filter parameters from the current value to the target value of the filter parameters is increased and the update cycle of the filter parameters is shortened. In response to the judgment result indicating that the industrial noise has not changed rapidly, the filter parameters are made to approach the target value of the filter parameters from the current value using a preset adjustment step size and a preset update cycle; Based on the cutting parameter information and the motor load information, the frequency components and intensity of the industrial noise in the next predetermined time window are predicted to obtain a noise change trend prediction value. Before the arrival of the next predetermined time window, the filter parameters are preset to the prediction parameter values ​​corresponding to the noise change trend prediction value. The prediction parameter values ​​are used as the initial values ​​of the filter parameters in the next predetermined time window. The filter parameters are then updated to the target values ​​of the filter parameters in the next predetermined time window based on the frequency components and intensity of the industrial noise identified in real time.

6. The adaptive control method for a built-in motor direct drive turret according to claim 5, characterized in that, The step of predicting the frequency components and intensity of the industrial noise within the next predetermined time window based on the cutting parameter information and the motor load information, and obtaining a predicted value for the noise change trend, includes: Obtain a working condition feature vector composed of the cutting parameter information and the motor load information of the current machining condition; The operating condition feature vector is compared with a preset operating condition feature library to obtain a first comparison result. The operating condition feature library includes multiple operating condition modes. Each operating condition mode includes an operating condition feature vector template, industrial noise frequency components corresponding to the operating condition feature vector template, industrial noise intensity, and its trend parameters over time. Based on the first comparison result, the working condition mode with the highest matching degree is determined as the matching mode, and the noise change trend prediction value is generated based on the trend parameter of the matching mode. In response to the first comparison result indicating that the matching degree of the matching mode is lower than the preset matching threshold, an adaptive adjustment mode is activated: based on the time-frequency analysis result of the motor current signal, the instantaneous frequency components and intensity of industrial noise are monitored, and the difference between the instantaneous frequency components and intensity and the noise change trend prediction value is calculated. Based on the difference calculation result, the noise change trend prediction value is corrected to obtain the corrected noise change trend prediction value, which is used for the determination and preset of prediction parameter values. Furthermore, noise change data under operating conditions where the matching degree is lower than a preset matching threshold is recorded. The noise change data includes at least instantaneous frequency components, intensity, and corresponding timestamps, and is used to update the operating condition feature library.

7. The adaptive control method for a built-in motor direct drive turret according to claim 6, characterized in that, Updating the operating condition feature library includes: The noise variation data is preprocessed to remove duplicate and abnormal data, resulting in preprocessed noise variation data. Key features are extracted from the preprocessed noise variation data, including the frequency range, energy intensity, and duration of the noise. Based on the key features, the feature similarity between the preprocessed noise change data is calculated, and the data with feature similarity higher than a preset similarity threshold are merged to form merged noise change data. The merged noise variation data is compared with the existing operating condition patterns in the operating condition feature library to obtain a second comparison result: In response to the second comparison result indicating that the similarity between the merged noise change data and the existing operating condition mode is lower than a preset mode threshold, the merged noise change data and its corresponding operating condition feature vector are written together as a new mode into the operating condition feature library. In response to the second comparison result indicating that the similarity is not lower than a preset mode threshold, the trend parameters of the existing operating mode are updated based on the merged noise change data. The trend parameters include at least the noise center frequency, noise bandwidth and / or the rate of change of noise energy intensity.

8. The adaptive control method for a built-in motor direct drive turret according to claim 3, characterized in that, The acquisition of motor load information includes: Synchronous sampling acquires the motor current signal and motor voltage signal fed back by the built-in motor servo driver; Real-time power calculation is performed based on the motor current signal and the motor voltage signal to obtain instantaneous motor power data; Frequency analysis was performed on the instantaneous motor power data to obtain the power frequency spectrum; Based on the power frequency spectrum, specific frequency components related to instantaneous impact loads and / or periodic harmonic loads are identified, and the influence of the specific frequency components is suppressed in the power frequency spectrum to obtain a power frequency spectrum that reflects the actual cutting load. The motor load information is calculated based on the power frequency spectrum that reflects the actual cutting load.

9. A built-in motor direct drive force turret adaptive control system, characterized in that, In conjunction with the method of claim 1, the system comprises: The vibration frequency information acquisition module is used to collect vibration signals of the built-in motor direct drive turret during the processing, and to perform frequency analysis on the vibration signals to obtain actual vibration frequency information characterizing the actual vibration state of the built-in motor direct drive turret. The actual vibration frequency information includes frequency components within a predetermined frequency band and the amplitude and / or energy value corresponding to each frequency component. The torque command frequency information acquisition module is used to acquire the torque command signal driving the built-in motor, and perform frequency analysis on the torque command signal to obtain torque command frequency information characterizing the torque command change characteristics. The torque command frequency information includes frequency components within a predetermined frequency band and the amplitude and / or energy value corresponding to each frequency component. The vibration source discrimination module is used to determine the frequency overlap and / or amplitude relationship at the corresponding frequency based on the actual vibration frequency information and the torque command frequency information, and to determine whether the current vibration of the built-in motor direct drive turret is mainly induced by the torque command or mainly caused by external cutting force fluctuations, and to obtain the discrimination result. The frequency overlap includes at least the overlapping frequency components of the two in a predetermined frequency band, and the amplitude relationship at the corresponding frequency includes at least the amplitude ratio and / or energy ratio of the vibration signal and the torque command signal at the overlapping frequency components. The torque command output characteristic adjustment module is used to adjust the output characteristics of the torque command when the discrimination result indicates that the current vibration is mainly induced by the torque command, so as to reduce the excitation effect of the vibration-inducing frequency component in the torque command on the built-in motor direct drive turret, thereby suppressing vibration; The torque compensation strategy adjustment module is used to adjust the torque compensation strategy used to drive the built-in motor when the discrimination result indicates that the current vibration is mainly caused by external cutting force fluctuations, so as to enhance the compensation capability for external cutting force fluctuations, thereby suppressing vibration and maintaining the stability of the cutting process.

Images