High-precision five-axis turntable cooperative control method and system based on multi-closed-loop feedback

Abstract

The invention provides a high-precision five-axis rotary table cooperative control method and system based on multi-closed-loop feedback, and the method comprises the steps: obtaining a real-time acceleration signal of a five-axis rotary table, carrying out the spectrum analysis of the acceleration signal, determining a current resonant frequency drift range and a main frequency component, and carrying out the frequency spectrum analysis; dynamically simulating the turntable structure by adopting a pre-established finite element model, and predicting the vibration mode change trend of the structure by comparing the deviation between the modal frequency obtained by simulation and the actually measured signal spectrum; according to the vibration mode change trend, matching similar vibration mode modes from a pre-established rotary table vibration mode database, obtaining corresponding vibration energy transmission path distribution, and determining an energy transmission abnormal region by analyzing the distribution of energy on each transmission path to obtain vibration energy transmission characteristics; according to an optimized motion instruction of a five-axis turntable, the dynamic response of the turntable is estimated, vibration frequency domain characteristics are extracted, and frequency domain compensation parameters are generated in combination with a preset frequency domain error matrix.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and in particular to a high-precision five-axis turntable collaborative control method and system based on multi-closed-loop feedback. Background Art

[0002] Five-axis machining technology holds a core position in fields such as aerospace and automotive manufacturing. Its ability to process complex surfaces with high precision is crucial to advancing industrial progress. However, existing methods have significant shortcomings in addressing vibration control issues under high-speed motion. Traditional control strategies typically rely on a single feedback mechanism, making it difficult to adapt to the rapid changes in vibration characteristics in dynamic environments, resulting in reduced machining accuracy and even equipment damage. Especially when a five-axis rotary table operates at high speeds, vibration becomes a key bottleneck restricting machining quality. Resonant frequency drift is a primary technical challenge. This phenomenon stems from the dynamic changes in structural stiffness during high-speed motion, causing vibration energy to be transferred unintendedly within the system. This frequency drift further induces complex variations in the vibration modes of the structural nodes, making vibration modes difficult to predict and conventional control algorithms unable to capture these dynamic characteristics in real time. These vibration mode variations further complicate the regulation of multi-loop feedback systems, as the coordination of multiple control loops is disrupted under complex vibration modes. This is particularly true in multi-mode operation scenarios, where the system requires frequent switching of control parameters, increasing the complexity of real-time adjustments. Therefore, how to effectively suppress the unexpected transmission of vibration energy caused by the resonant frequency drift and structural vibration mode changes under high-speed motion has become a key issue in improving the stability and accuracy of the five-axis machining system. Summary of the Invention

[0003] The present invention provides a high-precision five-axis turntable coordinated control method based on multi-closed-loop feedback, which mainly includes:

[0004] Acquire the real-time acceleration signal of the five-axis turntable, perform spectrum analysis on the acceleration signal, determine the current resonant frequency drift range and main frequency components, use the pre-established finite element model to dynamically simulate the turntable structure, and predict the change trend of the structural vibration mode by comparing the deviation of the simulated modal frequency with the measured signal spectrum;

[0005] According to the vibration shape change trend, similar vibration shape modes are matched from the pre-established turntable vibration shape database to obtain the corresponding vibration energy transfer path distribution. By analyzing the energy distribution on each transfer path, the energy transfer abnormal area is determined and the vibration energy transfer characteristics are obtained;

[0006] Based on the vibration energy transfer characteristics, the closed-loop feedback control system of the five-axis machining system is optimized to determine whether the optimized parameter set meets the coordination requirements of the turntable in multi-mode operation scenarios. If not, the parameters are adjusted to obtain the adjusted control parameter set. If so, the parameter set is output as a candidate control parameter.

[0007] Based on the candidate control parameters, the motion trajectory of the five-axis turntable is smoothed to obtain a smoothed motion instruction sequence. Simulation is then used to analyze whether the instruction sequence reduces vibration energy transfer. If so, the optimized motion instructions are obtained. If not, the high-order dynamic parameters of the motion trajectory are iteratively adjusted until the requirements for reducing vibration energy transfer are met.

[0008] Optimize motion instructions for the five-axis turntable, estimate the turntable's dynamic response, extract vibration frequency domain characteristics, and generate frequency domain compensation parameters based on the preset frequency domain error matrix;

[0009] Input frequency domain compensation parameters, combine with the real-time status of the five-axis turntable, calculate the dynamic compensation gain coefficient, perform phase correction on the control signal based on the dynamic compensation gain coefficient, perform frequency domain stability analysis on the phase-corrected control signal, and determine whether the frequency domain stability meets the preset frequency domain margin condition. If so, formulate and output the corresponding control strategy based on the final control signal. If not, adjust the weight of the frequency domain error matrix until the frequency domain stability meets the preset frequency domain margin condition.

[0010] Execute the control strategy, extract the eigenvector of the vibration signal during the five-axis turntable machining process, and determine whether the eigenvector deviates from the stability expectation. If the eigenvector does not deviate from the stability expectation, maintain the current control strategy. If the eigenvector deviates from the expectation, obtain the current vibration mode parameters, calculate the adjustment amount based on the historical modal data in the turntable vibration mode database, and verify the convergence of the adjustment amount in the multi-mode operation scenario. Based on the convergence results, update the modal weight coefficient in the turntable vibration mode database and output the control parameter set.

[0011] The present invention provides a high-precision five-axis turntable coordinated control system based on multi-closed-loop feedback, which mainly includes:

[0012] The real-time acceleration signal processing module is used to obtain the real-time acceleration signal of the five-axis turntable, perform spectrum analysis on the acceleration signal, determine the current resonance frequency drift range and main frequency components, and use a pre-established finite element model to dynamically simulate the turntable structure. By comparing the deviation between the simulated modal frequency and the measured signal spectrum, the trend of structural vibration mode change is predicted;

[0013] The spectrum analysis and resonance detection module is used to match similar vibration mode modes from the pre-established turntable vibration mode database according to the vibration mode change trend, obtain the corresponding vibration energy transfer path distribution, and determine the energy transfer abnormal area by analyzing the energy distribution along each transfer path to obtain the vibration energy transfer characteristics;

[0014] The dynamic simulation and vibration mode prediction module is used to optimize the parameters of the closed-loop feedback control system of the five-axis machining system based on the vibration energy transfer characteristics. It determines whether the optimized parameter set meets the coordination requirements of the turntable in multi-mode operation scenarios. If not, the parameters are adjusted to obtain the adjusted control parameter set. If so, the parameter set is output as a candidate control parameter.

[0015] The vibration mode matching and energy analysis module is used to smooth the motion trajectory of the five-axis turntable based on the candidate control parameters, obtain the smoothed motion instruction sequence, and analyze through simulation whether the instruction sequence reduces vibration energy transfer. If so, the optimized motion instruction is obtained. If not, the high-order dynamic parameters of the motion trajectory are iteratively adjusted until the requirements for reducing vibration energy transfer are met.

[0016] The control parameter optimization module is used to optimize the motion instructions of the five-axis turntable, estimate the dynamic response of the turntable, extract the vibration frequency domain characteristics, and generate frequency domain compensation parameters based on the preset frequency domain error matrix;

[0017] The motion trajectory optimization module is used to input frequency domain compensation parameters, calculate the dynamic compensation gain coefficient based on the real-time status of the five-axis turntable, perform phase correction on the control signal based on the dynamic compensation gain coefficient, perform frequency domain stability analysis on the control signal after phase correction, and determine whether the frequency domain stability meets the preset frequency domain margin condition. If so, the corresponding control strategy is formulated and output based on the final control signal. If not, the weight of the frequency domain error matrix is ​​adjusted until the frequency domain stability meets the preset frequency domain margin condition.

[0018] The dynamic compensation and stability analysis module is used to execute the control strategy, extract the characteristic vector of the vibration signal during the five-axis turntable machining process, and determine whether the characteristic vector deviates from the stability expectation. If the characteristic vector does not deviate from the stability expectation, the current control strategy is maintained. If the characteristic vector deviates from the expectation, the current vibration mode parameters are obtained, and the adjustment amount is calculated based on the historical modal data in the turntable vibration mode database. The convergence of the adjustment amount in the multi-mode operation scenario is verified, and the modal weight coefficient in the turntable vibration mode database is updated according to the convergence results, and the control parameter set is output.

[0019] The technical solution provided by the embodiment of the present invention may have the following beneficial effects:

[0020] The present invention discloses a high-precision five-axis turntable collaborative control method and system based on multi-closed-loop feedback. By real-time acquisition of acceleration signals and spectrum analysis, combined with dynamic simulation of finite element models, the trend of structural vibration mode changes is predicted. According to the vibration mode changes, the vibration energy transfer path is matched, the control system parameters and motion trajectory are optimized, and the vibration energy transfer is reduced. A dynamic compensation algorithm is used to perform phase correction on the control signal, and a frequency domain stability analysis is performed to ensure that the preset margin conditions are met. By real-time monitoring of vibration eigenvectors, the vibration mode database and control model parameters are dynamically updated to achieve adaptive vibration control in multi-mode operation scenarios. The present invention can effectively suppress the vibration of the five-axis turntable in complex machining processes, improve machining accuracy and efficiency, and adapt to the collaborative operation requirements of multiple machining modes. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 This is a flow chart of a high-precision five-axis turntable collaborative control method based on multi-closed-loop feedback of the present invention.

[0022] Figure 2 This is a structural schematic diagram of a high-precision five-axis turntable collaborative control system based on multi-closed-loop feedback of the present invention. DETAILED DESCRIPTION

[0023] To further understand the content of the present invention, the present invention is described in detail with reference to the accompanying drawings and examples. The present application is further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are intended only to illustrate the relevant invention and are not intended to limit the invention. It should also be noted that, for ease of description, only portions relevant to the invention are shown in the accompanying drawings.

[0024] like Figure 1-2 In this embodiment, a high-precision five-axis turntable coordinated control method based on multi-closed-loop feedback may specifically include:

[0025] S101. Acquire the real-time acceleration signal of the five-axis turntable, perform spectrum analysis on the acceleration signal, determine the current resonance frequency drift range and main frequency components, use a pre-established finite element model to dynamically simulate the turntable structure, and predict the trend of structural vibration mode changes by comparing the deviation between the simulated modal frequency and the measured signal spectrum.

[0026] The real-time acceleration signal in the built-in data interaction bus of the five-axis turntable is collected, and the acceleration signal is subjected to anti-aliasing filtering according to the sampling theorem to obtain the first signal data; the spectrum decomposition is performed using the segmented windowed Fourier transform based on the first signal data, and the frequency component whose amplitude ratio exceeds the preset threshold is extracted from the frequency amplitude sequence by the peak detection calculation method as the main frequency component; the upper and lower limits of the frequency change range are determined for the main frequency component by the spectrum envelope analysis method, and the frequency deviation quantification model is obtained by comparing the main frequency component with the reference vibration mode frequency value in the pre-established finite element model of the turntable structure; a displacement response prediction model is established at the geometric feature points of the turntable structure according to the frequency deviation quantification model, and the displacement response prediction model is numerically differentiated to obtain the vibration mode change rate function, and the structural vibration mode change trend is judged by the positive and negative values ​​of the values ​​of each component of the vibration mode change rate function.

[0027] For example, real-time acceleration sensor signals are directly collected from the five-axis turntable's built-in data exchange bus. The collected acceleration signals are processed in parallel using multiple threads. A signal queue storage mechanism is established in the bus read buffer. The acceleration signal acquisition intervals are divided according to the actual physical sampling period. Anti-aliasing filtering is performed on the signal data according to the sampling theorem to obtain first signal data. Based on the first signal data, a piecewise windowed Fourier transform is used to perform rapid spectral decomposition of the signal. The corresponding frequency band signals are intercepted from the original spectrum by setting spectral resolution parameters. The amplitudes of each frequency band are sorted and statistically analyzed using a peak detection calculation method to obtain a frequency amplitude sequence. Frequency components whose amplitude percentage exceeds a preset threshold are extracted from the sequence as the main frequency components. For the main frequency components, spectral envelope analysis is used to determine the upper and lower limits of the frequency variation range. The maximum offset within the frequency variation range is calculated to obtain the resonant frequency drift range. Based on a pre-established finite element model of the turntable structure, the turntable stiffness parameters, mass distribution parameters, and boundary constraint parameters are set in the model. Dynamic response characteristic simulation is performed using a harmonic response calculation method, and the reference vibration mode frequency values ​​are extracted from the simulated response results. By comparing the measured main frequency components with the reference vibration mode frequencies, a frequency deviation quantification model was established. This model includes the time-domain frequency deviation variation function f(t) and the deviation accumulation function g(t), where f(t) represents the instantaneous frequency deviation and g(t) represents the cumulative effect of the frequency deviation over time. Based on the frequency deviation quantification model, a displacement response prediction model was established at the geometric characteristic points of the turntable structure. The displacement response was numerically fitted using the Lagrange interpolation algorithm to obtain the vibration mode variation prediction function h(x,y,z,t), where x, y, and z are the spatial coordinate components and t is the time component. Numerically differentiating the vibration mode variation prediction function h(x,y,z,t) yields the vibration mode variation velocity function v(x,y,z,t). The trend of the structural vibration mode variation is determined by analyzing the positive and negative values ​​of the various components of v(x,y,z,t). In a five-axis machining system, the turntable structure supports the workpiece and performs multi-degree-of-freedom motion, and its dynamic characteristics directly affect machining accuracy. The accelerometer is installed at a key position of the turntable structure, with a sampling frequency of 1024 Hz and a sampling time of 60 seconds to obtain the vibration response signal during the turntable movement. The collected raw signal is processed using a Butterworth low-pass filter with a cutoff frequency of 200 Hz to filter out high-frequency noise components. By performing a fast Fourier transform on the filtered signal, a spectrum with a frequency resolution of 0.5 Hz is obtained, and the top five frequency components with the highest amplitude are identified from the spectrum. When the turntable is in an unloaded state, the main frequency components are distributed between 35 Hz and 75 Hz, with the most significant frequency being 52.5 Hz, corresponding to an amplitude of 0.386 g. When the turntable is loaded with a 10 kg cylindrical workpiece, the significant frequency decreases to 48.7 Hz and the amplitude increases to 0.452 g, indicating that the mass of the workpiece has changed the dynamic characteristics of the turntable structure.When analyzing the frequency drift range, a comparison of the spectral data under no-load and loaded conditions revealed a calculated frequency drift range of 3.8Hz to 5.2Hz. Spectral envelope analysis determined the upper limit of the frequency variation to be 54.3Hz, the lower limit to be 47.1Hz, and the maximum frequency offset to be 7.2Hz. This frequency drift phenomenon reflects the varying stiffness of the turntable structure under different operating conditions. A finite element model containing 20,000 elements was established for the turntable structure. The material properties used were Q235 steel with an elastic modulus of 210GPa and a Poisson's ratio of 0.3. In the model boundary condition settings, the connection surface between the turntable base and the machine column was set as a fixed constraint, and distributed mass elements were applied to the upper surface of the turntable worktable to simulate the workpiece mass. Modal analysis calculated the first six natural frequencies, of which the first natural frequency was 51.8Hz, close to the measured main frequency of 52.5Hz. Based on the comparison of the measured frequencies and modal frequencies, a frequency deviation quantification model was established. During the 60-second test, the mean instantaneous frequency deviation was 0.7 Hz, with a standard deviation of 0.25 Hz. The cumulative frequency deviation function showed a fluctuating upward trend, reaching a cumulative deviation of 1.2 Hz at 60 seconds. Twelve characteristic points were selected on the turntable structure to establish a displacement response prediction model. Analysis of the prediction model revealed that the Z-axis vibration mode change rate of the turntable structure was 0.15 mm / s, indicating that the turntable's dynamic characteristics in the vertical direction were gradually deteriorating. This deterioration trend is closely related to the turntable's load-bearing capacity and positioning accuracy, necessitating timely structural reinforcement or maintenance.

[0028] S102. Based on the vibration shape change trend, similar vibration shape modes are matched from a pre-established turntable vibration shape database to obtain the corresponding vibration energy transfer path distribution. By analyzing the energy distribution on each transfer path, the energy transfer abnormal area is determined to obtain the vibration energy transfer characteristics.

[0029] A dynamic time warping algorithm is used to perform pattern matching on the records stored in the turntable vibration mode database, and a set of similar vibration mode records with distance values ​​less than the vibration mode matching threshold is screened out according to the Euclidean distance value between the vibration mode displacement vectors to obtain a first vibration mode group; for the vibration mode records in the first vibration mode group, a topological diagram of the energy transfer network between the nodes of the turntable structure is constructed according to the principle of minimum energy loss, and the transfer paths and corresponding transfer power values ​​between the nodes are extracted from the topological diagram to obtain a first path distribution table; spatial mapping is performed according to the three-dimensional coordinates of the nodes in the first path distribution table, and a density clustering algorithm is used to divide the transfer power values ​​into regions to obtain a set of boundary point coordinates of the energy transfer abnormal area; a transfer path sequence is obtained for the set of boundary point coordinates of the energy transfer abnormal area, and the vibration energy transfer characteristics are obtained by calculating the path length, the number of corners, the node impedance value and the transfer power attenuation curve.

[0030] Exemplarily, based on the acquired modal change trend data, a dynamic time warping algorithm is used to perform pattern matching on the records stored in the turntable modal database. By calculating the Euclidean distance between the modal displacement vectors, a set of similar modal records with distance values ​​less than a modal matching threshold is selected from the database records to obtain a first modal group. For each record in the first modal group, the corresponding structural node coordinate data is retrieved. A topological diagram of the energy transfer network between the turntable structure nodes is constructed using the minimum energy loss principle. The transfer paths and corresponding transfer power values ​​between each node are extracted from the topological diagram to obtain a first path distribution table. The first path distribution table is spatially mapped according to the three-dimensional coordinates of the nodes. The transfer power values ​​are then divided into regions using a density clustering algorithm. Anomaly indices for each region are calculated based on two discriminant factors: the mean transfer power and node connectivity. This yields a set of coordinates for the boundary points of the energy transfer anomaly regions. From this set of coordinates, all transfer paths within the region are retrieved. The coordinates of the path's starting and ending points, as well as the coordinate sequences of the intermediate transfer nodes, are extracted. Spectral clustering is then used to group the transfer paths to obtain a second path distribution table. For each transfer path in the second path distribution table, three characteristic parameters—path length, number of corners, and node impedance—are calculated. A path transfer power attenuation curve is then generated. Vibration energy transfer characteristic data is derived by combining these characteristic parameters with the attenuation curve morphology. The vibrations generated during machining on a five-axis turntable are transferred between structural components. The vibration energy transfer characteristics reflect the dynamic performance of the structure. The turntable's vibration shape database stores a large amount of historical operational data, including the distribution of vibration shape displacement fields and corresponding energy transfer path information. When a new vibration shape trend is detected, pattern matching is performed by calculating the Euclidean distance between the vibration shape displacement vectors. Records with a distance value less than 0.15 are identified as similar vibration shapes. In a practical application, a turntable structure with 24 measurement points was used. The collected vibration shape displacement data showed a displacement amplitude of 0.42 mm in the Z-axis direction, which is closest to record number MT-756 in the database, with a Euclidean distance of 0.132. The energy transfer network between structural nodes corresponding to this record contains 76 possible paths. Based on the principle of minimum energy loss, energy transfer preferentially selects the path with the lowest impedance. The calculated transfer power distribution table shows that the main transfer paths are concentrated in the area where the turntable column and the workbench are connected. When analyzing the spatial distribution of the transfer power, a search domain with a radius of 120mm was used for density clustering, and nodes with similar transfer powers and concentrated spatial locations were divided into the same area. After statistical analysis, areas with an average transfer power exceeding 85W and a node connectivity greater than 4 were marked as abnormal areas. In one analysis, an abnormality was found at the front connection of the turntable column. This area contained 8 nodes, forming 12 transfer paths. Extraction of the characteristic parameters of the transfer paths within the abnormal area showed that the average path length was 275mm, including 3 corner nodes, and the node impedance values ​​were distributed between 0.8 and 1.2kg / s.The attenuation curve of the transferred power from the starting point to the end point shows an exponential attenuation characteristic, with an attenuation coefficient of 0.382. This attenuation characteristic indicates that the energy is greatly hindered during the transfer process, which is consistent with the increase in contact surface impedance caused by the loosening of the turntable column connection bolts. By comparing the transfer path characteristics under different working conditions, it was found that under normal working conditions, the transfer path of the turntable structure showed a uniform distribution characteristic, the power attenuation curve was flat, and the attenuation coefficient was usually less than 0.2. Under abnormal working conditions, energy transfer will show local concentration, the attenuation curve is steep, and it is often accompanied by the reconstruction of the transfer path, which is manifested as the disappearance of some original paths and the emergence of new transfer channels. This change in transfer characteristics contains important information about the degradation of the dynamic performance of the structure.

[0031] S103. Optimize the parameters of the closed-loop feedback control system of the five-axis machining system based on the vibration energy transfer characteristics, and determine whether the optimized parameter set meets the coordination requirements of the turntable in the multi-mode operation scenario. If not, adjust the parameters to obtain the adjusted control parameter set. If so, output the parameter set as a candidate control parameter.

[0032] According to the vibration energy transfer characteristics, the position closed-loop feedback signal is subjected to recursive least squares adaptive filtering, and the position loop control parameters are obtained by calculating the mean square error between the filtered signal and the turntable standard response signal; the Kalman filter algorithm is used to suppress the noise of the feedback signal after the position loop control parameters are substituted into the speed loop and the current loop, and the compensation coefficient is extracted from the filtered signal to construct a compensation function to obtain a closed-loop control parameter table; the correlation coefficient between the actual output curve and the standard reference curve is calculated according to the closed-loop control parameter table, and if the correlation coefficient is greater than the preset matching threshold, it is recorded as a candidate control parameter set; for operating modes where the parameter matching value is less than the preset matching threshold, the gradient descent method is used to progressively optimize the corresponding control parameters within the preset parameter adjustment range, and the adjusted control parameter set is determined by calculating the parameter matching value after each iteration.

[0033] For example, based on vibration energy transfer characteristic data, a recursive least-squares adaptive filter is used to process the position closed-loop feedback signal. By calculating the mean square error between the filter output signal and the turntable's standard response signal, the position loop proportional gain, integral gain, and differential gain are iteratively optimized to obtain the position loop control parameters. The position loop control parameters are substituted into the velocity and current loop controllers. A Kalman filter algorithm is used to suppress noise in the feedback signal. The velocity and current compensation coefficients are extracted from the filtered signal, and a compensation function is constructed using a polynomial fitting method to obtain a closed-loop control parameter table. Standard reference curves for the three machining modes of milling, drilling, and tapping are read from the turntable operation database. The degree of match between the actual output curve corresponding to the closed-loop control parameter table and the standard reference curve is compared using a correlation coefficient calculation method to obtain the parameter matching value for each mode. If the parameter matching value for all modes is greater than a preset matching threshold, the closed-loop control parameter table is recorded as a candidate control parameter set. If any operating mode has a parameter matching value less than the preset matching threshold, the parameter optimization process is initiated. For operating modes where the parameter matching value is less than a preset matching threshold, a gradient descent method is used to incrementally optimize the corresponding control parameters within a preset parameter adjustment range. The adjusted control parameter set is determined by calculating the parameter matching value after each iteration and used as the candidate control parameter set. A turntable motion state predictor is established based on the candidate control parameter set. The stability of the predictor's output response curve is verified using the Lyapunov stability criterion. The coordinated operation determination result is obtained by calculating the steady-state error of the response curve under each operating mode. The position, velocity, and current multi-loop control in a five-axis machining system plays a key role in turntable motion accuracy. When optimizing control parameters based on vibration energy transfer characteristics, the position closed-loop feedback signal is first adaptively filtered. In practice, the position loop proportional gain is initially set to 280. Using a recursive least squares algorithm, the proportional gain is adjusted to 325, the integral gain is adjusted from 0.15 to 0.22, and the differential gain is adjusted from 42 to 56, reducing the position tracking error to 0.002 mm. During the optimization of the velocity loop and current loop parameters, a Kalman filter is used to eliminate high-frequency noise in the feedback signal. The velocity loop feedforward compensation coefficient is obtained through polynomial fitting, and the compensation function is in the form of a quadratic function with coefficients of 0.85, 0.12, and 0.03, respectively. The current loop compensation uses a piecewise linear function, setting different compensation coefficients in different speed ranges: 1.2 for the low speed range, 1.0 for the medium speed range, and 0.8 for the high speed range. When the turntable operates in different processing modes, the control parameters need to meet the coordination requirements. Taking a certain model of turntable as an example, the standard reference curve in the milling mode requires the position response overshoot to be less than 5% and the steady-state error to be less than 0.005mm; the drilling mode requires the position response time to be less than 80ms and the overshoot to be less than 8%; the tapping mode focuses more on speed tracking accuracy and requires the speed fluctuation amplitude to be less than 2%.The parameter matching degree in each mode was calculated by correlation coefficient, which was 0.92 for milling mode, 0.88 for drilling mode, and 0.85 for tapping mode. When the parameter matching degree fell below the preset threshold of 0.90, the parameter optimization process began. The optimization process used the gradient descent method, and the parameter adjustment interval was set to ±20% of the initial value. In one optimization, the adjustment range of the position loop proportional gain was 260 to 340. After 50 iterative calculations, the optimal value was finally determined to be 315. At the same time, the velocity loop and current loop parameters were also adjusted accordingly to form a complete parameter combination. Based on the optimized parameter combination, a turntable motion state predictor was established. The predictor uses a state-space equation to describe the turntable motion characteristics, which includes three state variables: position, velocity, and acceleration. The system stability was verified using the Lyapunov stability criterion, and the calculated real parts of the characteristic roots were all less than -0.5, meeting the stability requirements. During the cyclic switching process of milling, drilling and tapping modes, the steady-state error of the position response is maintained within 0.003mm, the speed fluctuation amplitude does not exceed 1.5%, and the current fluctuation coefficient is less than 0.08, indicating that the optimized control parameters have good mode synergy.

[0034] S104. Smooth the motion trajectory of the five-axis turntable based on the candidate control parameters to obtain a smoothed motion instruction sequence. Use simulation to analyze whether the instruction sequence reduces vibration energy transfer. If so, obtain optimized motion instructions. If not, iteratively adjust the high-order dynamic parameters of the motion trajectory until the requirement for reducing vibration energy transfer is met.

[0035] The trajectory sampling points are connected using a cubic spline interpolation algorithm, and the trajectory is locally corrected according to the maximum curvature change rate to obtain a first instruction sequence; for the first instruction sequence, the least squares method is used to fit and calculate the acceleration change curve and the jump change curve, and the second instruction sequence is obtained through smoothing and integration operations; according to the second instruction sequence, the Runge-Kutta numerical integration method is used to calculate the force and displacement response curves of each joint of the turntable, and the vibration energy attenuation ratio is obtained by wavelet transform calculation; if the vibration energy attenuation ratio is less than the preset specification requirement value, the second instruction sequence is segmented according to the dynamic constraint boundary, and the Bezier curve is used to perform local optimization in the section with drastic acceleration changes to obtain a third instruction sequence; the third instruction sequence is verified for acceleration continuity, and the trajectory smoothness is judged by calculating the acceleration difference between adjacent instruction points. If the difference is less than the preset smoothness threshold, it is output as an optimized motion instruction sequence.

[0036] For example, a five-axis turntable trajectory description function is constructed based on candidate control parameters. The trajectory sampling points are connected using a cubic spline interpolation algorithm. The trajectory is locally corrected by setting a maximum curvature change rate. The corrected curve is uniformly sampled according to the turntable control period to obtain a first instruction sequence. For the first instruction sequence, the least squares method is used to fit the acceleration and jerk curves. These acceleration and jerk curves are piecewise smoothed using a sliding window whose length is the turntable response time. The smoothed curves are then integrated to obtain a second instruction sequence. The second instruction sequence is then substituted into the turntable's virtual motion calculation unit. The force and displacement response curves of each turntable joint are calculated using the Runge-Kutta numerical integration method. The vibration displacement amplitude is extracted from the displacement response curve, and the vibration energy attenuation ratio is calculated using a wavelet transform. If the vibration energy attenuation ratio exceeds the preset specification requirement, the second instruction sequence is marked as optimized. If the vibration energy attenuation ratio is less than the preset specification requirement, a particle swarm optimization algorithm is used to adjust the acceleration upper limit and jerk limit. New dynamic constraint boundaries are obtained through iterative calculation. The second command sequence is segmented according to the new dynamic constraint boundaries. Bezier curves are used for local optimization in sections with dramatic acceleration changes, while the original cubic spline curve remains unchanged in sections with smooth motion. The third command sequence is then extracted from the combined curve according to the control cycle. The third command sequence is verified for jerk continuity, and the trajectory smoothness is determined by calculating the jerk difference between adjacent command points. If the difference is less than a preset smoothness threshold, the optimized command sequence is output. Smoothing the trajectory of a five-axis turntable is crucial for suppressing vibration energy transfer. In practical applications, the turntable control cycle is typically set to 1ms, and the trajectory is uniformly sampled based on this sampling period. For a 30-degree turntable rotation, the initial sampled trajectory contains 300 discrete points, which are then connected into a continuous curve using cubic spline interpolation. Local corrections are performed at locations where the curvature change rate exceeds 0.15 / mm to achieve smoother curve transitions. After smoothing, the trajectory's dynamic characteristics need to be analyzed. Taking a certain type of turntable as an example, the dynamic analysis of the first instruction sequence found that the acceleration changes dramatically during the start and stop phases of rotation, with the maximum acceleration reaching 2.5 rad / s 2 , the jump peak reaches 12rad / s 3 The acceleration curve is smoothed using a sliding window of 50ms in length, reducing the peak acceleration to 1.8 rad / s. 2 , the peak jump is reduced to 8rad / s 3. In the virtual motion calculation, the motion equations of each joint of the turntable are solved using the fourth-order Runge-Kutta method, and the time step is set to 0.1ms. The calculation results show that the vibration displacement amplitude of the turntable B axis under the original trajectory is 0.086mm. The wavelet transform analysis shows that the main vibration energy is concentrated in the frequency band of 35Hz to 65Hz. After trajectory optimization, the vibration displacement amplitude is reduced to 0.052mm, and the vibration energy attenuation ratio reaches 0.62. When the vibration energy attenuation does not meet the requirements, the dynamic parameters need to be adjusted. Through the particle swarm optimization algorithm, the population size is set to 50, the number of iterations is set to 100, and the acceleration and jump limits are optimized. The optimization results show that reducing the acceleration upper limit to 1.5rad / s 2 , the jump limit is set to 6rad / s 3 The vibration suppression effect is the best when the acceleration changes more than 0.8rad / s. 2 The segment is reconstructed using a 5th-order Bezier curve. The control points of the Bezier curve are determined according to the acceleration continuity constraint to ensure that the acceleration and jerk at the connection of the curve are continuous. The optimized trajectory significantly improves the dynamic performance while maintaining the original motion time. The maximum difference in jerk between adjacent command points is reduced from the original 0.86rad / s 4 Reduced to 0.32 rad / s 4 , significantly improving the smoothness of the trajectory. This multi-level optimization-based trajectory smoothing method not only reduces the intensity of vibration energy transmission but also ensures the dynamic continuity of the trajectory. By properly setting the optimization parameters, a smooth transition of the turntable motion is achieved while maintaining machining accuracy. Practice has demonstrated that this method exhibits good adaptability in various machining conditions, such as high-speed cutting and precision drilling.

[0037] S105 , for the optimized motion instructions of the five-axis turntable, the dynamic response of the turntable is estimated, the vibration frequency domain characteristics are extracted, and the frequency domain compensation parameters are generated in combination with the preset frequency domain error matrix.

[0038] The optimized motion instructions of the five-axis turntable are obtained and analyzed to obtain the predicted characteristic data of the dynamic response. The frequency domain decomposition method is used for analysis to determine the distribution information of the vibration frequency domain. The main vibration components of the vibration frequency domain characteristics are extracted from the distribution information of the vibration frequency domain. The main vibration components are matched and analyzed by a preset error matrix to obtain the deviation distribution in the frequency domain. It is determined whether the deviation distribution exceeds the preset threshold range. If so, the specific interval of the deviation distribution is recorded. The specific interval of the deviation distribution is processed by the convolution operation method. In combination with the error matrix, the corresponding compensation parameters are calculated to obtain the frequency domain compensation parameters.

[0039] For example, by analyzing the optimization instructions of the five-axis turntable, the predicted characteristic data of the dynamic response is obtained, and the predicted characteristic data is preprocessed using a signal processing method to obtain preliminary vibration characteristic information. Based on the preliminary vibration characteristic information, the data is analyzed using frequency domain decomposition technology to extract the distribution information of the vibration frequency domain and determine the main vibration components in the frequency domain. For the distribution information of the vibration frequency domain, a matching analysis is performed using a preset error matrix to obtain the deviation distribution in the frequency domain, and determine whether the deviation distribution exceeds the preset threshold range. If it exceeds, the specific interval of the deviation distribution is recorded. The specific interval of the deviation distribution is processed using a convolution operation method, and the corresponding compensation parameters are calculated in combination with the data of the error matrix to obtain the adjustment value of the frequency domain compensation. For the optimized motion instructions of the five-axis turntable, the dynamic response of the turntable is first estimated through a dynamic simulation model. The specific method is to input the optimized motion instruction sequence into the finite element analysis model of the turntable, set the turntable material parameters to high-strength steel, the elastic modulus to 210GPa, and the Poisson's ratio to 0.3, and use the time-domain finite difference method to calculate the response data of the turntable under instruction drive. The time step is set to 0.05ms, and the total simulation time is 2s. The predicted vibration displacement curves of each joint of the turntable are obtained, and it is found that the peak vibration amplitude of the B-axis under specific working conditions is 0.045mm. Subsequently, the predicted vibration characteristics are decomposed in the frequency domain, and the fast Fourier transform algorithm is used to convert the time domain vibration signal into a frequency domain signal. The sampling frequency is set to 20kHz. The spectrum analysis shows that the main vibration energy is concentrated in the range of 40Hz to 80Hz, with a peak frequency of 55Hz and a corresponding power spectrum density of 0.0023mm 2 / Hz. Next, a convolution operation is performed based on the vibration frequency domain characteristics and the preset frequency domain error matrix. The frequency domain error matrix is ​​constructed based on historical processing data. The matrix dimension is 100×100, and the element value range is between -0.5 and 0.5, representing the error weight at different frequencies. The frequency domain deviation is extracted through the convolution operation to generate the frequency domain compensation parameters. The calculation results show that the compensation parameter value at 55Hz is 0.38, indicating that the vibration at this frequency needs to be targeted and suppressed.

[0040] S106: Input frequency domain compensation parameters and calculate the dynamic compensation gain coefficient based on the real-time status of the five-axis turntable. Phase correction is performed on the control signal based on the dynamic compensation gain coefficient. Frequency domain stability analysis is performed on the phase-corrected control signal to determine whether the frequency domain stability meets the preset frequency domain margin condition. If so, a corresponding control strategy is formulated and output based on the final control signal. If not, the weights of the frequency domain error matrix are adjusted until the frequency domain stability meets the preset frequency domain margin condition.

[0041] The sliding Fourier transform is used to perform frequency domain decomposition on the state signal collected by the five-axis turntable position sensor. The amplitude characteristics and phase characteristics of the spectral components are obtained based on the frequency domain decomposition, and the dynamic compensation matrix is ​​calculated in combination with the frequency domain compensation parameters; the turntable motion response data is online identified based on the dynamic compensation matrix using the recursive least squares method, and the dynamic compensation gain coefficient is obtained by performing singular value decomposition on the online identification data; the closed-loop transfer function is calculated for the dynamic compensation gain coefficient using the Nyquist stability criterion, if the amplitude margin of the closed-loop transfer function is greater than the preset amplitude threshold and the phase margin is greater than the preset phase threshold, the compensation control signal is marked as the final control signal, and the control strategy is determined; if the amplitude margin and phase margin do not meet the preset margin requirements, the gradient descent method is used to iteratively adjust the weights in the frequency domain error matrix, and the frequency domain compensator is constructed by the pole configuration method to obtain the control strategy.

[0042] Exemplarily, based on the real-time status signals collected by the five-axis turntable's position, velocity, and acceleration sensors, a sliding Fourier transform is used to perform frequency domain decomposition of the signals. Amplitude and phase characteristics are extracted from the spectral components, and then combined with frequency domain compensation parameters to calculate a dynamic compensation matrix. For the dynamic compensation matrix, a recursive least squares method is used to perform online identification of the turntable's motion response data. A frequency domain error matrix is ​​constructed based on the position error, velocity error, and acceleration error. Dynamic compensation gain coefficients are obtained through singular value decomposition, and phase compensation is performed on the original control signal to obtain a compensated control signal. The amplitude-frequency and phase-frequency characteristics are extracted from the compensated control signal. The amplitude margin and phase margin of the closed-loop transfer function are calculated using the Nyquist stability criterion. Frequency domain stability is determined by comparing the amplitude margin threshold with the phase margin threshold to obtain a stability determination result. If the stability determination result indicates that the amplitude margin is greater than a preset amplitude threshold and the phase margin is greater than a preset phase threshold, the compensated control signal is marked as the final control signal. If the stability determination result does not meet the preset margin requirements, the weight optimization process is initiated. In the weight optimization process, the gradient descent method is used to iteratively adjust the position, velocity, and acceleration error weights in the frequency-domain error matrix. The compensation gain coefficients are recalculated based on the adjusted weights to obtain the optimized control signal. A frequency-domain compensator is constructed using the pole placement method to optimize the control signal. By setting the compensator's dominant pole position, the resonant peak of the closed-loop response is suppressed. The control parameters are extracted from the compensator's transfer function to obtain the final control command. Dynamic characteristic compensation is crucial for improving machining accuracy during high-speed machining of a five-axis rotary table. For a certain model of rotary table, position, velocity, and acceleration signals are synchronously acquired at a 1024Hz sampling frequency. The signals are then subjected to real-time frequency-domain decomposition using a 256-point sliding Fourier transform. Analysis results show a significant resonant peak in the 45Hz to 85Hz frequency range, with an amplitude of 0.42mm and a phase lag of 76 degrees. To account for the frequency-domain characteristics, a frequency weighting function is introduced when constructing the dynamic compensation matrix. Near the resonant frequency band, the position error weight is set to 1.5, the velocity error weight to 0.8, and the acceleration error weight to 0.4. Using recursive least squares for online identification, the calculated compensation gain exhibits a peak characteristic in the resonant frequency band, with a maximum compensation gain of 2.8 and a phase lead compensation angle of 65 degrees. Frequency-domain stability analysis shows that the closed-loop transfer function of the compensated control signal exhibits a good stability margin. The original control signal had an amplitude margin of 4.2 dB and a phase margin of 28 degrees, both of which fell short of the preset requirements. After phase compensation, the amplitude margin of the closed-loop transfer function increased to 8.6 dB and the phase margin to 42 degrees, meeting the design requirements of an amplitude margin greater than 6 dB and a phase margin greater than 35 degrees. If the compensation effect is unsatisfactory, the error weights need to be optimized. The weight parameters are adjusted using the gradient descent method, with a learning rate of 0.05 used for iteration.During the optimization process, it was found that the closed-loop response characteristics were optimal when the position error weight was adjusted to 1.8, the velocity error weight to 0.6, and the acceleration error weight to 0.3. In frequency-domain compensator design, the location of the dominant pole directly affects the compensation effect. Using a pole placement method, the real part of the compensator's dominant pole was set at -12°, and the imaginary part at ±75°, forming a conjugate pole pair. This configuration creates a notch characteristic in the compensator's resonant frequency band, with an attenuation ratio of 0.45, effectively suppressing system resonance. The dynamically compensated turntable control signal exhibits excellent frequency-domain characteristics. Under high-speed cutting conditions, turntable positioning accuracy was improved to 0.002 degrees, and velocity fluctuation was reduced to 1.2% of the rated value. In actual machining, workpiece surface roughness improved from Ra1.8μm to Ra0.8μm, and roundness error was reduced from 0.012mm to 0.005mm. This dynamic compensation method based on frequency-domain analysis not only improves control accuracy but also ensures system stability, achieving significant improvements in workpiece dimensional accuracy and surface quality.

[0043] S107. Execute the control strategy, extract the eigenvector of the vibration signal during the five-axis turntable machining process, and determine whether the eigenvector deviates from the stability expectation. If the eigenvector does not deviate from the stability expectation, maintain the current control strategy. If the eigenvector deviates from the expectation, obtain the current vibration mode parameters, calculate the adjustment amount based on the historical modal data in the turntable vibration mode database, and verify the convergence of the adjustment amount in the multi-mode operation scenario. Update the modal weight coefficient in the turntable vibration mode database according to the convergence result, and output the control parameter set.

[0044] A wavelet packet decomposition algorithm is used to perform multi-scale analysis on the vibration signal of a five-axis turntable, and the vibration eigenvector is obtained by calculating the energy distribution and relative phase of each frequency band. The vibration eigenvector is compared with the vibration stability boundary, and when the vibration eigenvector exceeds the vibration stability boundary, the modal displacement field is read from the finite element database. An optimization objective function including displacement error term, velocity error term and acceleration error term is constructed for the modal displacement field, and the turntable control parameters are obtained by calculating the genetic algorithm. The modal contribution of each order is calculated based on the modal displacement field to generate a modal weight matrix. The turntable control parameters are adaptively updated according to the modal weight matrix to generate a control parameter set.

[0045] For example, vibration sensor signals are collected during the machining process of a five-axis turntable. A wavelet packet decomposition algorithm is used to perform multi-scale analysis of the signals. The energy distribution and phase characteristics of each frequency band are calculated to extract vibration eigenvectors. These eigenvectors are then compared with the vibration stability boundary to determine the vibration state and obtain a state assessment result. Based on the state assessment result, if the eigenvectors are within the vibration stability boundary, the current control parameters are recorded and vibration signals are continuously collected. If the eigenvectors are outside the vibration stability boundary, the current modal displacement field is read from the finite element database and, combined with historical modal data, the eigendecomposition method is used to generate a structural parameter update matrix. Based on the structural parameter update matrix, an optimization objective function is constructed that includes displacement error terms, velocity error terms, and acceleration error terms. A genetic algorithm is used to optimize the turntable control parameters. A parameter adjustment solution is obtained by calculating the convergence of the parameters under multi-mode operating conditions. This parameter adjustment solution is input into a recursive neural network. Based on the modal displacement field data, the contribution of each mode under different operating conditions is calculated to generate a modal weight matrix. This modal weight matrix is ​​then used to adaptively update the control parameters to obtain the optimized control parameters. A response surface methodology was used to establish a mapping relationship between optimized control parameters and different machining modes. A cubic polynomial fitting function was constructed to achieve continuous parameter variation during mode switching. Control parameters were extracted from the fitting function to obtain an adaptive parameter sequence. This adaptive parameter sequence was input into an online monitoring unit, and a vibration feature extraction mechanism based on a sliding time window was established to continuously monitor and evaluate the turntable's vibration status, ensuring real-time updating of control parameters. Vibration status monitoring and adaptive parameter updating for a five-axis turntable are critical for ensuring machining accuracy. In practical applications, four high-precision vibration sensors were placed at key locations on the turntable, with a sampling frequency of 2048 Hz. The vibration signal was decomposed into 64 frequency bands using a six-layer wavelet packet decomposition. The calculated eigenvectors consisted of frequency band energy ratios and phase characteristics. Vibration stability boundaries were determined through statistical analysis, with the energy ratio upper limit set at 0.35 and the phase deviation limited to within 45 degrees. When an abnormal vibration eigenvector was detected, the control parameters were promptly updated. Taking a certain model of turntable as an example, during high-speed milling, the energy ratio of the eigenvector exceeded 0.42, and the phase deviation reached 58 degrees, exceeding the stability boundary. At this time, the vibration mode displacement field, containing displacement data for 24 nodes, was extracted from the finite element database. Eigendecomposition was performed in combination with historical modal data, and the resulting structural parameter update matrix reflected the trend of change in turntable stiffness. During the parameter optimization process, the constructed objective function comprehensively considered three aspects: displacement error, velocity error, and acceleration error. The displacement error weight was set to 0.5, the velocity error weight was 0.3, and the acceleration error weight was 0.2. Optimization was performed using a genetic algorithm with a population size of 60 and a number of iterations of 100. The optimization results showed that the position loop gain should be reduced by 15% and the velocity loop feedforward coefficient should be increased by 8% to achieve parameter convergence under multi-mode conditions.When a recurrent neural network is used to calculate modal weights, the input layer contains eight nodes, corresponding to the eight main vibration modes. Network training uses a batch size of 32 data sets, with a learning rate set to 0.01. Calculation results show that in high-speed milling mode, the contributions of the first three modes are 0.42, 0.35, and 0.15, respectively. These values ​​directly affect the adaptive update of control parameters. The parameter mapping relationship established by the response surface methodology adopts the form of a cubic polynomial. When switching between milling, drilling, and tapping modes, the control parameters exhibit continuous variation characteristics.

[0046]

[0047] , P i (m) represents the response surface function of the i-th control parameter with respect to the machining mode m, where m is the mode variable (which can be the coded value of milling, drilling, or tapping), and b ij is the preset fitting coefficient, and the formula realizes the mapping of parameters to processing modes.

[0048] For example, the position loop gain varies from 280 to 360 during mode switching, with a smooth curve that avoids the impact of sudden parameter changes. The online monitoring unit uses a sliding time window of 2 seconds with a window overlap rate of 75%. By calculating the vibration characteristics within the window in real time, the monitoring results show that the optimized control parameters reduce the turntable vibration amplitude by 46%, the spectral energy is more concentrated, and the phase characteristics are more stable. During the continuous 8-hour processing process, the positioning accuracy of the turntable was maintained within 0.002 degrees, and the roundness error was controlled within 0.006mm, proving the effectiveness of the parameter adaptive update mechanism.

[0049] The present invention provides a high-precision five-axis turntable coordinated control system based on multi-closed-loop feedback, which mainly includes:

[0050] The real-time acceleration signal processing module is used to obtain the real-time acceleration signal of the five-axis turntable, perform spectrum analysis on the acceleration signal, determine the current resonance frequency drift range and main frequency components, and use a pre-established finite element model to dynamically simulate the turntable structure. By comparing the deviation between the simulated modal frequency and the measured signal spectrum, the trend of structural vibration mode change is predicted;

[0051] The spectrum analysis and resonance detection module is used to match similar vibration mode modes from the pre-established turntable vibration mode database according to the vibration mode change trend, obtain the corresponding vibration energy transfer path distribution, and determine the energy transfer abnormal area by analyzing the energy distribution along each transfer path to obtain the vibration energy transfer characteristics;

[0052] The dynamic simulation and vibration mode prediction module is used to optimize the parameters of the closed-loop feedback control system of the five-axis machining system based on the vibration energy transfer characteristics. It determines whether the optimized parameter set meets the coordination requirements of the turntable in multi-mode operation scenarios. If not, the parameters are adjusted to obtain the adjusted control parameter set. If so, the parameter set is output as a candidate control parameter.

[0053] The vibration mode matching and energy analysis module is used to smooth the motion trajectory of the five-axis turntable based on the candidate control parameters, obtain the smoothed motion instruction sequence, and analyze through simulation whether the instruction sequence reduces vibration energy transfer. If so, the optimized motion instruction is obtained. If not, the high-order dynamic parameters of the motion trajectory are iteratively adjusted until the requirements for reducing vibration energy transfer are met.

[0054] The control parameter optimization module is used to optimize the motion instructions of the five-axis turntable, estimate the dynamic response of the turntable, extract the vibration frequency domain characteristics, and generate frequency domain compensation parameters based on the preset frequency domain error matrix;

[0055] The motion trajectory optimization module is used to input frequency domain compensation parameters, calculate the dynamic compensation gain coefficient based on the real-time status of the five-axis turntable, perform phase correction on the control signal based on the dynamic compensation gain coefficient, perform frequency domain stability analysis on the control signal after phase correction, and determine whether the frequency domain stability meets the preset frequency domain margin condition. If so, the corresponding control strategy is formulated and output based on the final control signal. If not, the weight of the frequency domain error matrix is ​​adjusted until the frequency domain stability meets the preset frequency domain margin condition.

[0056] The dynamic compensation and stability analysis module is used to execute the control strategy, extract the characteristic vector of the vibration signal during the five-axis turntable machining process, and determine whether the characteristic vector deviates from the stability expectation. If the characteristic vector does not deviate from the stability expectation, the current control strategy is maintained. If the characteristic vector deviates from the expectation, the current vibration mode parameters are obtained, and the adjustment amount is calculated based on the historical modal data in the turntable vibration mode database. The convergence of the adjustment amount in the multi-mode operation scenario is verified, and the modal weight coefficient in the turntable vibration mode database is updated according to the convergence results, and the control parameter set is output.

[0057] The above description is merely a preferred embodiment of the present application and an illustration of the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in this application is not limited to the technical solutions formed by the specific combination of the above-mentioned technical features, but also encompasses other technical solutions formed by any combination of the above-mentioned technical features or their equivalents without departing from the concept of this application. For example, a technical solution formed by replacing the above-mentioned features with (but not limited to) technical features with similar functions disclosed in this application.

Claims

1. A high-precision five-axis turntable collaborative control method based on multi-closed-loop feedback, characterized in that: The method comprises: Acquire the real-time acceleration signal of the five-axis turntable, perform spectrum analysis on the acceleration signal, determine the current resonant frequency drift range and main frequency components, use the pre-established finite element model to dynamically simulate the turntable structure, and predict the change trend of the structural vibration mode by comparing the deviation of the simulated modal frequency with the measured signal spectrum; According to the vibration shape change trend, similar vibration shape modes are matched from the pre-established turntable vibration shape database, the corresponding vibration energy transfer path distribution is obtained and analyzed, the energy transfer abnormal area is determined, and the vibration energy transfer characteristics are obtained; Based on the vibration energy transfer characteristics, the closed-loop feedback control system of the five-axis machining system is optimized. Candidate control parameters are output based on the judgment result of whether the optimized parameter set meets the requirements of the turntable's coordination in multi-mode operation scenarios; Based on the candidate control parameters, the motion trajectory of the five-axis turntable is smoothed to obtain a smoothed motion instruction sequence. Simulation is then used to analyze whether the instruction sequence reduces vibration energy transfer. If so, the optimized motion instructions are obtained. If not, the dynamic parameters of the motion trajectory are iteratively adjusted until the requirements for reducing vibration energy transfer are met. According to the optimized motion instructions, the dynamic response of the turntable is estimated, the vibration frequency domain characteristics are extracted, and the frequency domain compensation parameters are generated by combining the preset frequency domain error matrix; The control signal is compensated according to the frequency domain compensation parameter, the frequency domain stability analysis is performed on the compensated control signal, and a control strategy is generated according to the frequency domain stability analysis result.

2. The high-precision five-axis turntable coordinated control method based on multi-closed-loop feedback according to claim 1 is characterized in that: The method includes obtaining a real-time acceleration signal of the five-axis turntable, performing spectrum analysis on the acceleration signal, determining the current resonance frequency drift range and main frequency components, using a pre-established finite element model to dynamically simulate the turntable structure, and predicting the trend of structural vibration mode changes by comparing the deviation between the simulated modal frequency and the measured signal spectrum. The method includes: Collecting a real-time acceleration signal from a built-in data interaction bus of the five-axis turntable, and performing anti-aliasing filtering on the acceleration signal according to a sampling theorem to obtain first signal data; Performing spectrum decomposition using a piecewise windowed Fourier transform on the first signal data, and extracting frequency components whose amplitudes exceed a preset threshold from the frequency amplitude sequence as main frequency components using a peak detection calculation method; The frequency spectrum envelope analysis method is used to determine the upper and lower limits of the frequency variation range for the main frequency component, and the main frequency component is compared with the reference vibration mode frequency value in the pre-established turntable structure finite element model to obtain a frequency deviation quantification model; A displacement response prediction model is established at the geometric feature points of the turntable structure based on the frequency deviation quantization model. A numerical differentiation operation is performed on the displacement response prediction model to obtain a vibration mode change rate function. The positive and negative values ​​of the values ​​of each component of the vibration mode change rate function are used to determine the vibration mode change trend of the structure.

3. The high-precision five-axis turntable coordinated control method based on multi-closed-loop feedback according to claim 1 is characterized in that: According to the vibration shape change trend, similar vibration shape modes are matched from a pre-established turntable vibration shape database, the corresponding vibration energy transfer path distribution is obtained and analyzed, the energy transfer abnormal area is determined, and the vibration energy transfer characteristics are obtained, including: The dynamic time warping algorithm is used to perform pattern matching on the records stored in the turntable vibration shape database. Based on the Euclidean distance between the vibration shape displacement vectors, a set of similar vibration shape records with a distance value less than the vibration shape matching threshold is selected to obtain the first vibration shape group. Based on the vibration mode records in the first vibration mode group, a topological diagram of the energy transfer network between nodes of the turntable structure is constructed according to the principle of minimum energy loss. The transmission paths between nodes and the corresponding transmission power values ​​are extracted from the topological diagram to obtain a first path distribution table; Performing spatial mapping based on the three-dimensional coordinates of the nodes in the first path distribution table, and using a density clustering algorithm to divide the transferred power values ​​into regions to obtain a coordinate set of boundary points of energy transfer abnormal regions; A transmission path sequence is obtained for the coordinate set of boundary points in the energy transmission abnormality area, and the vibration energy transmission characteristics are obtained by calculating the path length, the number of corners, the node impedance value and the transmission power attenuation curve.

4. The high-precision five-axis turntable coordinated control method based on multi-closed-loop feedback according to claim 1 is characterized in that: The closed-loop feedback control system of the five-axis machining system is optimized for the vibration energy transfer characteristics, and candidate control parameters are output based on the judgment result of whether the optimized parameter set meets the coordination requirements of the turntable in the multi-mode operation scenario, including: Performing recursive least squares adaptive filtering on the position closed-loop feedback signal according to the vibration energy transfer characteristics, and obtaining the position loop control parameters by calculating the mean square error between the filtered signal and the turntable standard response signal; A Kalman filter algorithm is used to suppress noise in the feedback signal after the position loop control parameters are substituted into the speed loop and the current loop, and a compensation coefficient is extracted from the filtered signal to construct a compensation function to obtain a closed-loop control parameter table; Calculating the correlation coefficient between the actual output curve and the standard reference curve according to the closed-loop control parameter table, and recording it as a candidate control parameter set if the correlation coefficient is greater than a preset matching threshold; For operating modes where the parameter matching value is less than the preset matching threshold, the gradient descent method is used to progressively optimize the corresponding control parameters within the preset parameter adjustment range, and the adjusted control parameter set is determined by calculating the parameter matching value after each iteration.

5. The high-precision five-axis turntable coordinated control method based on multi-closed-loop feedback according to claim 1 is characterized in that: The motion trajectory of the five-axis turntable is smoothed according to the candidate control parameters to obtain a smoothed motion instruction sequence, and a simulation is performed to analyze whether the instruction sequence reduces vibration energy transfer. If so, an optimized motion instruction is obtained. If not, the dynamic parameters of the motion trajectory are iteratively adjusted until the requirement for reducing vibration energy transfer is met, including: The trajectory sampling points are connected using a cubic spline interpolation algorithm, and the trajectory is locally corrected according to the maximum curvature change rate to obtain the first instruction sequence; For the first instruction sequence, the least squares method is used to fit and calculate the acceleration change curve and the jerk change curve, and the second instruction sequence is obtained through smoothing and integration operation; According to the second instruction sequence, the force and displacement response curves of each joint of the turntable are calculated using the Runge-Kutta numerical integration method, and the vibration energy attenuation ratio is calculated by wavelet transform; If the vibration energy attenuation ratio is less than the preset specification requirement, the second instruction sequence is segmented according to the dynamic constraint boundary, and a Bezier curve is used to perform local optimization in the section with drastic acceleration changes to obtain a third instruction sequence; The acceleration continuity of the third instruction sequence is verified, and the trajectory smoothness is determined by calculating the jerk difference between adjacent instruction points. If the difference is less than the preset smoothness threshold, the optimized motion instruction sequence is output.

6. The high-precision five-axis turntable coordinated control method based on multi-closed-loop feedback according to claim 1 is characterized in that: The method of estimating the dynamic response of the turntable according to the optimized motion instructions, extracting the vibration frequency domain characteristics, and generating the frequency domain compensation parameters in combination with the preset frequency domain error matrix includes: Obtain and analyze the optimized motion instructions of the five-axis turntable to obtain the predicted characteristic data of the dynamic response. Use the frequency domain decomposition method to analyze and determine the distribution information of the vibration frequency domain. Extracting the main vibration components of the vibration frequency domain characteristics from the distribution information of the vibration frequency domain, performing matching analysis on the main vibration components through a preset error matrix to obtain the deviation distribution in the frequency domain; Determine whether the deviation distribution exceeds a preset threshold range, and if so, record the specific interval of the deviation distribution; A convolution operation method is used to process the specific interval of the deviation distribution, and the corresponding compensation parameters are calculated in combination with the error matrix to obtain frequency domain compensation parameters.

7. The high-precision five-axis turntable coordinated control method based on multi-closed-loop feedback according to claim 1 is characterized in that: The control signal is compensated according to the frequency domain compensation parameter, the frequency domain stability analysis is performed on the compensated control signal, and a control strategy is generated according to the frequency domain stability analysis result, including: inputting the frequency domain compensation parameter, combining the real-time state of the five-axis turntable, and calculating the dynamic compensation gain coefficient; performing phase correction on the control signal based on the dynamic compensation gain coefficient, performing frequency domain stability analysis on the control signal after the phase correction, and judging whether the frequency domain stability meets the preset frequency domain margin condition; if so, formulating and outputting the corresponding control strategy based on the final control signal; if not, adjusting the weight of the frequency domain error matrix until the frequency domain stability meets the preset frequency domain margin condition.

8. The high-precision five-axis turntable coordinated control method based on multi-closed-loop feedback according to claim 7 is characterized in that: The input frequency domain compensation parameter is combined with the real-time status of the five-axis turntable to calculate the dynamic compensation gain coefficient; based on the dynamic compensation gain coefficient, the control signal is phase corrected, and the frequency domain stability analysis is performed on the control signal after the phase correction to determine whether the frequency domain stability meets the preset frequency domain margin condition. If so, a corresponding control strategy is formulated and output based on the final control signal; if not, the weight of the frequency domain error matrix is ​​adjusted until the frequency domain stability meets the preset frequency domain margin condition, including: The sliding Fourier transform is used to perform frequency domain decomposition on the state signal collected by the five-axis turntable position sensor. The amplitude and phase characteristics of the spectral components are obtained based on the frequency domain decomposition, and the dynamic compensation matrix is ​​calculated by combining the frequency domain compensation parameters. Performing online identification on the turntable motion response data using a recursive least squares method according to the dynamic compensation matrix, and obtaining a dynamic compensation gain coefficient by performing singular value decomposition on the online identification data; calculating a closed-loop transfer function based on the dynamic compensation gain coefficient using the Nyquist stability criterion, and if an amplitude margin of the closed-loop transfer function is greater than a preset amplitude threshold and a phase margin is greater than a preset phase threshold, marking the compensation control signal as a final control signal and determining a control strategy; If the amplitude margin and the phase margin do not meet the preset margin requirements, the weights in the frequency domain error matrix are iteratively adjusted using the gradient descent method, and a frequency domain compensator is constructed using the pole placement method to obtain a control strategy.

9. The high-precision five-axis turntable coordinated control method based on multi-closed-loop feedback according to claim 1 is characterized in that: The execution control strategy extracts the characteristic vector of the vibration signal during the five-axis turntable machining process, determines whether the characteristic vector deviates from the stability expectation, maintains the current control strategy if the characteristic vector does not deviate from the stability expectation, and obtains the current vibration mode parameters if the characteristic vector deviates from the expectation, calculates the adjustment amount based on the historical modal data in the turntable vibration mode database, verifies the convergence of the adjustment amount in the multi-mode operation scenario, updates the modal weight coefficient in the turntable vibration mode database based on the convergence result, and outputs the control parameter set, including: The wavelet packet decomposition algorithm is used to perform multi-scale analysis on the vibration signal of the five-axis turntable, and the vibration eigenvector is obtained by calculating the energy distribution and relative phase of each frequency band. Comparing and judging the vibration eigenvector with the vibration stability boundary, and reading the mode displacement field from the finite element database when the vibration eigenvector exceeds the vibration stability boundary; An optimization objective function including displacement error term, velocity error term and acceleration error term is constructed for the vibration mode displacement field, and the turntable control parameters are obtained by genetic algorithm calculation; The modal contribution of each order is calculated based on the vibration mode displacement field to generate a modal weight matrix, and the turntable control parameters are adaptively updated according to the modal weight matrix to generate a control parameter set.

10. A high-precision five-axis turntable coordinated control system based on multi-closed-loop feedback, characterized in that: The system comprises: The real-time acceleration signal processing module is used to obtain the real-time acceleration signal of the five-axis turntable, perform spectrum analysis on the acceleration signal, determine the current resonance frequency drift range and main frequency components, and use a pre-established finite element model to dynamically simulate the turntable structure. By comparing the deviation between the simulated modal frequency and the measured signal spectrum, the trend of structural vibration mode change is predicted; The spectrum analysis and resonance detection module is used to match similar vibration mode modes from the pre-established turntable vibration mode database according to the vibration mode change trend, obtain the corresponding vibration energy transfer path distribution, and determine the energy transfer abnormal area by analyzing the energy distribution along each transfer path to obtain the vibration energy transfer characteristics; The dynamic simulation and vibration mode prediction module is used to optimize the parameters of the closed-loop feedback control system of the five-axis machining system based on the vibration energy transfer characteristics. It determines whether the optimized parameter set meets the coordination requirements of the turntable in multi-mode operation scenarios. If not, the parameters are adjusted to obtain the adjusted control parameter set. If so, the parameter set is output as a candidate control parameter. The vibration mode matching and energy analysis module is used to smooth the motion trajectory of the five-axis turntable based on the candidate control parameters, obtain the smoothed motion instruction sequence, and analyze through simulation whether the instruction sequence reduces vibration energy transfer. If so, the optimized motion instruction is obtained. If not, the high-order dynamic parameters of the motion trajectory are iteratively adjusted until the requirements for reducing vibration energy transfer are met. The control parameter optimization module is used to optimize the motion instructions of the five-axis turntable, estimate the dynamic response of the turntable, extract the vibration frequency domain characteristics, and generate frequency domain compensation parameters based on the preset frequency domain error matrix; The motion trajectory optimization module is used to compensate the control signal according to the frequency domain compensation parameters, perform frequency domain stability analysis on the compensated control signal, and generate a control strategy based on the frequency domain stability analysis results.

Images