A smart control system for precision hardware machining based on CNC machine tools

Abstract

This invention discloses an intelligent control system for precision hardware machining based on CNC machine tools, belonging to the field of CNC machine tool control technology. The system first relies on a feature analysis module to acquire the three-dimensional geometric features and material properties of the workpiece, providing data support for matching machining parameters. Then, through the mapping database and multi-factor weighted algorithm of the machining parameter determination module, it achieves matching between machining parameters and the three-dimensional features of the workpiece, avoiding blind parameter selection. Next, it uses a chatter suppression pre-optimization module to predict chatter risks at different machining stages in advance and optimize parameters, actively avoiding potential chatter. Simultaneously, a chatter judgment module extracts vibration features in real time, identifies chatter and its intensity, and provides real-time feedback for parameter adjustment. Finally, the machining parameter optimization module dynamically determines the compensation amount based on the chatter intensity, solving the problem in existing technologies where it is difficult to match machining parameters based on the three-dimensional features of the workpiece for chatter suppression and to compensate machining parameters based on feedback during the machining process.

Description

Technical Field

[0001] This invention relates to the field of CNC machine tool control technology, specifically to an intelligent control system for precision hardware machining based on CNC machine tools. Background Technology

[0002] In the CNC machining of precision hardware parts, the workpiece's chatter-prone structure and material properties are diverse, the relationship between machining parameters and chatter suppression effects is complex, and the matching logic between chatter feedback and parameter compensation during machining is ambiguous. Existing technologies have the following shortcomings: First, existing technologies focus more on geometric parameters and have not established a relationship between geometric features and chatter attributes, resulting in a lack of comprehensive data support for machining parameter matching. They rely on general parameters or manual experience and cannot achieve targeted adaptation to the chatter-prone characteristics of the workpiece.

[0003] Secondly, existing technologies do not have clearly defined parameters for the weighting of flutter impact, nor have they established a flutter risk prediction mechanism. Flutter suppression is mostly a reactive, post-event approach. At the same time, the compensation direction and magnitude in the compensation process are vague, making it difficult to effectively solve the flutter problem.

[0004] Therefore, there is an urgent need for an intelligent control system that integrates feature analysis, intelligent parameter matching, chatter prediction and pre-optimization, and dynamic compensation to solve the problem that existing technologies cannot match processing parameters based on the three-dimensional features of the workpiece to suppress chatter and compensate processing parameters based on feedback during the processing. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides an intelligent control system for precision hardware parts processing based on CNC machine tools. This system solves the problem that existing technologies struggle to match processing parameters based on the three-dimensional features of the workpiece, suppress chatter, and compensate processing parameters based on feedback during the processing.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solution: an intelligent control system for precision hardware parts processing based on CNC machine tools, comprising: a feature analysis module, used to analyze the three-dimensional geometric features and material properties of the precision hardware parts to be processed, and to determine the three-dimensional feature analysis results of the workpiece.

[0007] The machining parameter determination module is used to construct a mapping database and uses a multi-factor weighted algorithm to match the initial machining parameter set corresponding to the workpiece's three-dimensional feature analysis results from the mapping database.

[0008] The chatter suppression pre-optimization module is used to predict chatter based on the three-dimensional feature analysis results of the workpiece, determine the chatter risk level at different processing stages, generate a chatter suppression pre-optimization strategy adapted to the initial processing parameter set, and output the pre-optimized processing parameter set.

[0009] The chatter detection module is used to send the pre-optimized machining parameter set to the CNC machine tool for machining, extract vibration characteristics from the real-time data of the machining process, and determine whether chatter exists and its intensity.

[0010] The machining parameter optimization module is used to determine the machining parameter compensation amount based on the chatter intensity when chatter is present, and to compensate the pre-optimized machining parameter set; if no chatter is detected, the pre-optimized machining parameter set remains unchanged.

[0011] Compared with existing technologies, this invention has the following advantages: First, it relies on the feature analysis module to obtain the three-dimensional geometric features and material properties of the workpiece, providing data support for matching processing parameters; then, through the mapping database and multi-factor weighted algorithm of the processing parameter determination module, it achieves the matching of processing parameters with the three-dimensional features of the workpiece, avoiding blind parameter selection; next, it uses the chatter suppression pre-optimization module to predict chatter risks at different processing stages in advance and optimize parameters to actively avoid potential chatter; at the same time, it uses the chatter judgment module to extract vibration features in real time, identify chatter and intensity, and provide real-time feedback for parameter adjustment; finally, the processing parameter optimization module dynamically determines the compensation amount based on the chatter intensity, solving the problem that existing technologies are difficult to match processing parameters based on the three-dimensional features of the workpiece, perform chatter suppression, and compensate processing parameters based on feedback during the processing. Attached Figure Description

[0012] Figure 1 This is a schematic diagram of the module connection of the intelligent control system for precision hardware processing based on CNC machine tools according to the present invention.

[0013] Figure 2 This is a flowchart illustrating the construction of a mapping database in the intelligent control system for precision hardware machining based on CNC machine tools, as described in this invention.

[0014] Figure 3 This is a flowchart illustrating the compensation process for pre-optimized machining parameter sets in the intelligent control system for precision hardware machining based on CNC machine tools, as described in this invention. Detailed Implementation

[0015] The present invention will now be described in detail with reference to the accompanying drawings and embodiments. Please refer to the accompanying drawings. Figure 1 The present invention provides a technical solution: an intelligent control system for precision hardware processing based on CNC machine tools, comprising: a feature analysis module, used to analyze the three-dimensional geometric features and material properties of the precision hardware to be processed, and to determine the three-dimensional feature analysis results of the workpiece.

[0016] It should be noted that the specific process for determining the three-dimensional feature analysis results of the workpiece is as follows: Based on the CAD design file, obtain the three-dimensional coordinate data of the easily chattering parts of the precision hardware to be processed, and preset the dynamic damping coefficient and preliminary dynamic elastic modulus. It should also be noted that the three-dimensional coordinate data can also come from high-precision three-dimensional scanning data, including the three-dimensional coordinate data of easily chattering parts such as thin walls, slender shafts, deep cavities, and suspended bosses.

[0017] The purpose of setting the dynamic damping coefficient and preliminary dynamic elastic modulus is to provide a benchmark for subsequent deviation correction from the measured values. The actual dynamic damping coefficient and dynamic elastic modulus obtained by dynamic mechanical testing replace the preset values, eliminating the deviation between theoretical parameters and actual material properties, ensuring the accuracy of the flutter attribute set, and providing a reliable material property basis for matching processing parameters.

[0018] The main reason for determining the three-dimensional coordinate data of easily fluttering areas is to focus on the easily fluttering areas, avoid data interference from non-fluttering areas, and provide a precise basis for targeted analysis of the system.

[0019] Since clearly defining the structural type enables targeted extraction of geometric parameters, the type of workpiece prone to chatter is determined based on the workpiece type in the CAD design file. At the same time, the geometric parameters of each chatter-prone part are extracted to form a geometric feature set corresponding to the chatter-prone parts and their geometric parameters.

[0020] It should be noted that the types of workpiece structures prone to chatter include thin-walled chatter structures, slender shaft chatter structures, deep cavity chatter structures, and composite chatter structures. Geometric parameters include, but are not limited to, wall thickness, wall span, shaft length-to-diameter ratio, groove depth-to-width ratio, hole diameter-to-depth ratio, and overhang length of suspended parts, which facilitates direct quantification of the chatter sensitivity characteristics of the chatter-prone structure.

[0021] Dynamic excitation tests are performed on workpiece samples by dynamic mechanical testing to directly obtain the actual dynamic damping coefficient and dynamic elastic modulus of the material. The preset dynamic damping coefficient and preliminary dynamic elastic modulus are compared with the actual dynamic damping coefficient and dynamic elastic modulus, and the deviation data are corrected to form a flutter attribute set.

[0022] In this embodiment, the actual dynamic damping coefficient and dynamic elastic modulus are mechanical properties characterizing the flutter resistance of the material: the dynamic damping coefficient determines the material's ability to attenuate vibration energy; the larger the coefficient, the faster the vibration attenuation and the more difficult it is to maintain flutter; the dynamic elastic modulus directly affects the structural stiffness and vibration response characteristics of the workpiece and is a parameter for deriving the critical condition for flutter. Through actual measurement correction, the deviation between theoretical parameters and actual material properties can be eliminated, ensuring the accuracy of material property data, providing a reliable material performance basis for flutter prediction and processing parameter compensation, and ensuring the effectiveness of the system's flutter suppression strategy.

[0023] The process of obtaining the actual dynamic damping coefficient and dynamic elastic modulus of the material by sampling workpieces through dynamic mechanical testing is as follows: Select representative sample workpieces of the precision hardware to be processed, apply dynamic excitation to the sample workpieces using sinusoidal frequency sweep excitation or pulse excitation, collect the vibration response signal during the excitation process by vibration sensors installed on the easily chattering parts of the sample workpieces, use a dynamic signal analysis system to perform frequency domain conversion and feature extraction on the vibration response signal, and calculate the actual dynamic damping coefficient and dynamic elastic modulus of the sample workpiece material based on the frequency response function.

[0024] It should be noted that the dynamic signal analysis system involved in this process performs frequency domain conversion and feature extraction on the vibration response signal, and the actual dynamic damping coefficient and dynamic elastic modulus of the sampled workpiece material are calculated based on the frequency response function, which are all existing technologies.

[0025] The geometric feature set of each precision hardware part to be processed is associated with the chatter attribute set one-to-one, and redundancy is eliminated and standardized. According to the preset accuracy and unified measurement unit standard data format, the structured three-dimensional feature analysis result of the workpiece is generated, which provides directly usable targeted data for the subsequent processing parameter determination module to match the initial processing parameter group and the chatter suppression pre-optimization module to predict chatter.

[0026] For example, the preset accuracy requirements are geometric parameters to 0.001 mm and mechanical parameters to 0.01 GPa.

[0027] The process of determining the three-dimensional feature analysis results of the workpiece involves acquiring the three-dimensional coordinate data of the easily fluttering parts, identifying the easily fluttering structural types, extracting flutter-sensitive geometric parameters, and simultaneously correcting the dynamic damping coefficient and dynamic elastic modulus through dynamic mechanical testing. Then, through correlation integration, redundancy elimination, and standardization, structured analysis results are generated. This achieves targeted focus on the flutter-related features of the workpiece, avoids interference from non-critical data, and ensures the accuracy, consistency, and usability of geometric features and material property data.

[0028] The machining parameter determination module is used to construct a mapping database and uses a multi-factor weighted algorithm to match the initial machining parameter set corresponding to the workpiece's three-dimensional feature analysis results from the mapping database.

[0029] like Figure 2 As shown, the process of constructing the mapping database is as follows: for hardware parts with different types of easily chattering structures, obtain the three-dimensional feature analysis results of the workpiece, and match them with different processing parameter groups. These different processing parameter groups can be determined based on experience, and multiple processing parameter groups with different values ​​are matched to provide sufficient and effective samples.

[0030] The vibration amplitude and flutter occurrence rate during the trial processing were collected, and the effective processing parameter groups that meet the preset suppression threshold were screened. The feature analysis results - effective processing parameter groups - flutter suppression effect were standardized to obtain a ternary sample dataset.

[0031] For example, the preset suppression threshold is determined based on the target machining accuracy level. For example, for precision hardware parts with an accuracy requirement of IT5, the vibration amplitude threshold is set to no more than 0.01 mm and the chatter occurrence rate threshold is set to no more than 1%; for IT6 precision workpieces, the vibration amplitude can be adjusted to no more than 0.02 mm and the chatter occurrence rate to no more than 3%.

[0032] The analytical results of the workpiece's three-dimensional features in the ternary sample dataset are transformed into an independent variable vector. The standardized machining parameter set is transformed into an intermediate variable vector. The standardized vibration amplitude and chatter rate are transformed into a dependent variable vector. Taking the conversion of workpiece 3D feature analysis results into independent variable vectors as an example, all quantified parameters are arranged in a predetermined fixed order, such as a unified sorting rule of geometric parameters first followed by material property parameters, to ensure that the parameter arrangement logic of each sample is consistent. Then, the sorted standardized parameters are used as vector elements to construct a one-dimensional independent variable vector. Similarly, the conversion methods for standardized machining parameter groups, standardized vibration amplitude, and chatter occurrence rate are also the same.

[0033] With the values ​​of the independent variable vector fixed, the partial correlation coefficient between each element in the intermediate variable vector and each element in the dependent variable vector is calculated based on the partial correlation coefficient calculation formula, and the partial correlation coefficient is filled into the correlation coefficient matrix.

[0034] It should be noted that the calculation process of the partial correlation coefficient is as follows: First, residual calculation is performed to eliminate the influence of the independent variable vector. Specifically, for the elements in the intermediate variable vector, a linear regression model is constructed with the independent variable vector as the independent variable and each element in the intermediate variable vector as the dependent variable. The regression equation is solved, and the corresponding residual is obtained by subtracting the actual value of the dependent variable from the regression predicted value of the dependent variable. Similarly, a linear regression model is constructed with the independent variable vector as the independent variable and the standardized vibration amplitude and flutter incidence rate as the dependent variables, and the residuals corresponding to the standardized vibration amplitude and flutter incidence rate are calculated respectively.

[0035] First, the partial correlation coefficient between the residuals of each element in the intermediate variable vector and the residuals of the standardized vibration amplitude is calculated using the Pearson correlation coefficient formula, which is an existing technology. Then, the same method is used to calculate the partial correlation coefficient between the residuals of each element in the intermediate variable vector and the standardized flutter incidence.

[0036] It should be noted that each row of the correlation coefficient matrix represents the distribution of the correlation strength between the corresponding processing parameter and the vibration amplitude and flutter incidence, respectively. The first table represents the distribution of the correlation strength between the vibration amplitude and all processing parameters, and the second table represents the distribution of the correlation strength between the flutter incidence and all processing parameters. The value range of the matrix elements is [-1, 1], where a value greater than 0 indicates that the effective processing parameter is positively correlated with the flutter suppression effect index, and parameter optimization can improve the suppression effect. A value less than 0 indicates a negative correlation, and parameter optimization will reduce the suppression effect. The closer to 1, the stronger the correlation strength, and the closer to 0, the weaker the correlation strength.

[0037] Based on the partial correlation coefficient, a set of processing parameters that meet the set conditions is selected. Then, based on the set of processing parameters, the parameter combination that minimizes the vibration amplitude and the chatter rate is locked. The parameter combination is then mapped to the corresponding workpiece three-dimensional feature analysis results, vibration amplitude and chatter rate to obtain a mapping database.

[0038] It should be noted that the set condition is that the effective correlation coefficient between each element in all intermediate variable vectors of a certain correlation coefficient matrix and the vibration amplitude is not less than 0.8, and the effective correlation coefficient between each element and the flutter frequency is not less than 0.8.

[0039] Locking the parameter combination that minimizes vibration amplitude and chatter rate involves weighted summation of vibration amplitude and chatter rate corresponding to each processing parameter group to determine the parameter combination corresponding to the minimum weighted value.

[0040] This embodiment collects the three-dimensional feature analysis results of hardware parts with different chatter-prone structures, matches multiple sets of processing parameters, and collects data on chatter suppression effects during trial processing. After standardization, a ternary sample dataset is formed. Significant and effective correlations between processing parameters and chatter suppression effects are screened out. Finally, the optimal combination of processing parameters is locked and a targeted mapping is established, ensuring the accuracy and reliability of the correlation between processing parameters and workpiece chatter-prone characteristics and chatter suppression effects. This provides a basis for the subsequent processing parameter determination module to quickly match and adapt the initial processing parameter set.

[0041] Furthermore, considering the varying degrees of influence of different parameters in the workpiece's 3D feature analysis results on chatter suppression, and the different sensitivities and contributions of different parameters to chatter, with some parameters having a more critical impact on chatter generation and suppression while others have a relatively weak impact, a multi-factor weighted algorithm is used to match the initial processing parameter set corresponding to the workpiece's 3D feature analysis results from the mapping database. Specifically, based on the workpiece's 3D feature analysis results, vibration amplitude, and chatter occurrence rate in the ternary sample dataset, grey relational analysis is used to determine the influence weights of each parameter in the workpiece's 3D feature analysis results on vibration amplitude and chatter occurrence rate. Specifically, grey relational analysis is used to determine the correlation degree of any parameter relative to vibration amplitude and chatter occurrence rate; then, the two correlation degrees are weighted and summed to obtain the comprehensive correlation degree corresponding to this parameter; finally, the comprehensive correlation degree of each parameter is calculated as the ratio of the sum of all comprehensive correlation degrees, and this ratio is used as the influence weight of each parameter.

[0042] It should be noted that the process of obtaining the correlation degree through grey relational calculation is as follows: the vibration amplitude and flutter occurrence rate are used as reference sequences, and any parameter in the three-dimensional feature analysis result of the workpiece is used as the comparison sequence.

[0043] Then, the reference sequence and the comparison sequence are standardized to eliminate the difference in dimensions. The absolute difference between the corresponding elements of the standardized comparison sequence and each reference sequence is then calculated to form the absolute difference sequence.

[0044] Next, the maximum and minimum values ​​are determined from all absolute differences. The resolution coefficient is set to 0.5. The correlation coefficient between the comparison sequence and each reference sequence at each sample point is calculated according to the rule of adding the resolution coefficient to the minimum value and multiplying it by the maximum value as the numerator, and adding the resolution coefficient to the current absolute difference and multiplying it by the maximum value as the denominator.

[0045] Finally, the arithmetic mean of the correlation coefficients of all sample points corresponding to each reference sequence is taken to obtain the correlation degree of the workpiece parameter with respect to vibration amplitude and flutter rate, thus completing the grey relational analysis of a single parameter. The remaining parameters are analyzed one by one according to the same process.

[0046] The three-dimensional feature analysis results of the current precision hardware part to be processed are compared with the three-dimensional feature analysis results of each workpiece in the mapping database to calculate the comprehensive similarity value by multi-factor weighted similarity.

[0047] Specifically, for the current workpiece and a single sample in the database, the corresponding parameters are aligned one by one, the absolute difference of each parameter is calculated, and then the absolute difference is multiplied by the influence weight of the corresponding parameter to obtain the weighted difference of each parameter.

[0048] Then, the weighted differences of all parameters are summed to obtain the total weighted difference between the current workpiece and the sample. The smaller the total difference, the closer the characteristics of the two are.

[0049] Finally, normalization is used to convert the total weighted difference into a comprehensive similarity value, which is usually in the range of [0,1]. The conversion logic is that the comprehensive similarity is equal to 1 minus the ratio of the total weighted difference to the maximum value of the total weighted difference. This ensures that the smaller the total weighted difference, the larger the comprehensive similarity value, which intuitively reflects the degree of feature matching between the two.

[0050] The overall similarity values ​​are sorted in descending order, and the parameter combination corresponding to the highest overall similarity value is selected as the initial processing parameter group.

[0051] This embodiment uses grey relational analysis to quantify the influence weight of each parameter in the workpiece's three-dimensional feature analysis results on the chatter suppression effect, enabling multi-factor weighted similarity calculation to focus on chatter-sensitive parameters and avoid matching deviations caused by treating all parameters equally. At the same time, relying on the effective processing parameter group samples that have been screened and verified in the mapping database, the initial processing parameter group that best matches the current workpiece features is quickly locked through weighted matching of structured features, ensuring the relevance and reliability of the initial parameters and the workpiece's chatter-prone characteristics, and avoiding the blind selection of processing parameters.

[0052] The chatter suppression pre-optimization module is used to predict chatter based on the three-dimensional feature analysis results of the workpiece, determine the chatter risk level at different processing stages, generate a chatter suppression pre-optimization strategy adapted to the initial processing parameter set, and output the pre-optimized processing parameter set.

[0053] Furthermore, the process of determining the chatter risk level at different processing stages is as follows: query the mapping database, select the N processing parameter combinations that match the top N of the workpiece's three-dimensional feature analysis results at the current processing stage, and extract the corresponding vibration amplitude and chatter occurrence rate.

[0054] It should be noted that the processing parameter combination of the top N matching scores is the same as the processing parameter combination corresponding to the comprehensive similarity of the top N in the screening ranking. For example, N is 5.

[0055] The extracted vibration amplitude and flutter incidence were averaged to obtain the predicted vibration amplitude and predicted flutter incidence.

[0056] By selecting the top N highly compatible processing parameter combinations in the mapping database that match the workpiece features at the current processing stage, the corresponding vibration amplitude and chatter rate are extracted and averaged. This not only utilizes the effective data of highly correlated samples, but also smooths out random errors to improve the stability and representativeness of the predicted values.

[0057] The flutter risk quantification value is obtained by weighted summing the ratio of the predicted vibration amplitude to the vibration amplitude threshold and the ratio of the predicted flutter incidence rate to the flutter incidence rate threshold. The flutter risk level is then determined based on the flutter risk quantification value.

[0058] It should be noted that a mapping relationship between flutter risk quantification value and flutter risk level has been pre-built in the database. The larger the flutter risk quantification value, the higher the flutter risk level, which indicates a greater flutter risk.

[0059] Furthermore, the process of generating a pre-optimized flutter suppression strategy adapted to the initial processing parameter set and outputting the pre-optimized processing parameter set is as follows: Clustering is performed using the absolute values ​​of the partial correlation coefficients between each processing parameter in the initial processing parameter set and the vibration amplitude and flutter incidence as the clustering benchmark value, resulting in multiple clusters equal to the number of flutter risk levels. The clustering benchmark value is the mean of the absolute values ​​of the partial correlation coefficients between the vibration amplitude and the flutter incidence rate.

[0060] The mean value of the clustering baseline value corresponding to each cluster is calculated, and the clusters are sorted in descending order based on the mean calculation results. They are then associated with the flutter risk level one by one. The higher the ranking of the cluster, the more significant the impact of the processing parameters contained therein on flutter suppression.

[0061] Based on the determined flutter risk level, the processing parameters contained in the corresponding clusters are extracted. Priority is given to adjusting the processing parameters corresponding to the largest cluster baseline value, with the adjustment direction determined by the sign of the partial correlation coefficient. When the partial correlation coefficient is positive, increasing the processing parameter can reduce vibration amplitude and flutter incidence; when the partial correlation coefficient is negative, decreasing the processing parameter can improve flutter suppression. By prioritizing adjustments based on the cluster baseline values, targeted reduction of flutter risk is achieved.

[0062] It should be noted that, considering that each processing parameter corresponds to two partial correlation coefficients, namely vibration amplitude and chatter rate, if both partial correlation coefficients are not greater than 0, the processing parameter is adjusted downward based on the set adjustment amount; if both partial correlation coefficients are not less than 0, the processing parameter is adjusted upward based on the set adjustment amount; however, if the two partial correlation coefficients are opposite in sign, the adjustment is performed by first adjusting upward and then downward.

[0063] If the adjusted flutter risk quantification value decreases by more than the set amplitude threshold, the adjustment is retained. If the adjustment does not meet the requirements, the processing parameters corresponding to the absolute value of the second largest partial correlation coefficient are adjusted until all parameters are adjusted to obtain the pre-optimized processing parameter set.

[0064] This embodiment uses the mean absolute values ​​of two partial correlation coefficients as the clustering benchmark. Clustering is used to form clusters corresponding to risk levels and sorted by their significance of impact. This ensures that parameter adjustments match the current flutter risk level. Furthermore, the adjustment rules for different combinations of partial correlation coefficients are clearly defined, avoiding ambiguity in adjustment direction. Additionally, parameters with significant impact are prioritized for optimization based on the clustering benchmark value. Combined with the logic of retaining parameters that meet the criteria and iteratively adjusting all parameters that do not, the targeting and operability of parameter adjustments are improved, ensuring that the pre-optimized parameter set effectively reduces flutter risk.

[0065] It should be noted that the process of outputting the pre-optimized machining parameter set is set in the initial machining stage of the workpiece and is used to test and adjust the machine tool. In the formal machining stage, it is controlled according to the subsequent machining parameter optimization module.

[0066] The chatter detection module is used to send the pre-optimized machining parameter set to the CNC machine tool for machining, extract vibration characteristics from the real-time data of the machining process, and determine whether chatter exists and its intensity.

[0067] It should be noted that the process of determining the presence and intensity of flutter is as follows: real-time vibration data is collected at a preset sampling frequency, and noise reduction and standardization are performed to obtain a standardized vibration signal.

[0068] Based on standardized vibration signals, time-domain features and frequency-domain features are extracted to form a multidimensional flutter feature vector. The extraction of time-domain features and frequency-domain features are both existing technologies. The extracted time-domain features include peak vibration amplitude, peak factor and kurtosis value, and the frequency-domain features include vibration energy ratio and characteristic frequency peak value.

[0069] It should be noted that peak amplitude can directly quantify vibration intensity, and flutter will significantly exceed the normal range when it occurs; peak factor and kurtosis value are sensitive to the impact component and nonlinear distortion in vibration, and can effectively identify sudden impact vibrations caused by flutter.

[0070] The proportion of vibration energy can reflect the energy concentration in the natural frequency range of the machine tool-workpiece-tool system (energy will be concentrated in a specific resonant frequency band during chatter); the peak value of the characteristic frequency can lock the abnormal resonant frequency corresponding to chatter, avoiding confusion with normal cutting vibration.

[0071] Combining time-domain and frequency-domain features not only covers the time-domain performance of flutter in terms of vibration intensity and impact characteristics, but also captures its frequency-domain anomalies in terms of energy distribution and frequency response, reducing missed or false detections.

[0072] If any feature in the multidimensional chatter feature vector does not meet the set safety range, then chatter is determined to exist in the processing process.

[0073] The number of features in the multidimensional flutter feature vector that do not meet the set safety range is taken as the flutter intensity.

[0074] This embodiment extracts multidimensional flutter feature vectors, enabling comprehensive analysis of flutter signals. The rule of determining flutter by any feature exceeding the safe range improves the sensitivity of flutter identification. Furthermore, quantifying flutter intensity using individual features that do not meet the safe range is intuitive, simple, and easy to operate, allowing for rapid determination of flutter severity without complex calculations.

[0075] The machining parameter optimization module is used to determine the machining parameter compensation amount based on the chatter intensity when chatter is present, and to compensate the pre-optimized machining parameter set; if no chatter is detected, the pre-optimized machining parameter set remains unchanged.

[0076] like Figure 3 As shown, the specific process of compensating the pre-optimized processing parameter group is as follows: obtain the average value of each flutter feature from historical effective compensation cases, and calculate the deviation of each flutter feature based on the average value and the boundary of the corresponding safety range.

[0077] It should be noted that if the safety range has an upper and lower boundary, the deviation is calculated using the average of the upper and lower boundaries. The deviation calculation involves calculating the deviation rate between the average and the boundary, and dividing the absolute value of the difference between the two by the boundary value.

[0078] The flutter comprehensive deviation index is obtained by weighting and summing the deviations based on the influence weights of each flutter feature.

[0079] For a given processing parameter, calculate its partial correlation coefficient with each chatter feature, and sum the absolute values ​​of the partial correlation coefficients to obtain the comprehensive partial correlation coefficient corresponding to that processing parameter.

[0080] The ratio of the comprehensive partial correlation coefficient to the sum of the comprehensive partial correlation coefficients of all processing parameters is used as the comprehensive sensitivity weight, so that each processing parameter corresponds to a comprehensive sensitivity weight.

[0081] By statistically analyzing the minimum and maximum values ​​of each processing parameter compensation in historical effective compensation cases, the range of processing parameter compensation amplitude is obtained. It should be noted that the processing parameter compensation values ​​are signed; in historical effective compensation cases, when the processing parameter is compensated downwards, the processing parameter compensation value is negative; otherwise, it is positive or 0.

[0082] A three-dimensional mapping table is established with the flutter comprehensive deviation index as the horizontal axis, the comprehensive sensitivity weight as the vertical axis, and the processing parameter compensation range as the cells.

[0083] Substitute the flutter comprehensive deviation index and comprehensive sensitivity weight calculated when flutter is present into the mapping relationship table to match the corresponding processing parameter compensation range. The processing parameter compensation amount is taken as the median value of the processing parameter compensation range.

[0084] It should be noted that the matching process can use Euclidean distance to calculate the flutter comprehensive deviation index and the similarity between the comprehensive sensitivity weight and each combination in the mapping relationship table, and determine the minimum similarity as the corresponding processing parameter compensation range.

[0085] The pre-optimized baseline value (the value taken in the pre-optimized processing parameter group) of each processing parameter is superimposed with the processing parameter compensation amount to compensate the pre-optimized processing parameter group.

[0086] This embodiment calculates the chatter feature deviation degree and combines it with its influence weight to obtain the comprehensive chatter deviation index. At the same time, it determines the comprehensive sensitivity weight based on the comprehensive partial correlation coefficient between the processing parameters and the chatter features. Then, it accurately matches the compensation range of the processing parameters through a three-dimensional mapping relationship table and takes the median value as the compensation amount. This not only utilizes historical effective experience to ensure the reliability of the compensation basis, but also achieves the adaptation of the compensation amount with the chatter intensity and parameter sensitivity through weight quantification and three-dimensional matching. It can quickly respond to chatter problems in the processing process and effectively correct the deviation of the pre-optimized parameter set.

[0087] The parameters involved in the above formula are all dimensionless and calculated numerically. The formula is a formula obtained from the most recent real situation by collecting a large amount of data and simulating it with software. The preset parameters in the formula are set by those skilled in the art according to the actual situation.

[0088] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0089] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0090] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0091] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0092] Finally, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An intelligent control system for precision hardware machining based on CNC machine tools, characterized in that, include: The feature analysis module is used to analyze the three-dimensional geometric features and material properties of the precision hardware parts to be processed, and to determine the three-dimensional feature analysis results of the workpiece. The machining parameter determination module is used to construct a mapping database and uses a multi-factor weighted algorithm to match the initial machining parameter set corresponding to the workpiece 3D feature analysis results from the mapping database. The chatter suppression pre-optimization module is used to predict chatter based on the three-dimensional feature analysis results of the workpiece, determine the chatter risk level at different processing stages, generate a chatter suppression pre-optimization strategy adapted to the initial processing parameter set, and output the pre-optimized processing parameter set. The chatter detection module is used to send the pre-optimized machining parameter set to the CNC machine tool for machining, extract vibration characteristics from the real-time data of the machining process, and determine whether chatter exists and its intensity. The machining parameter optimization module is used to determine the compensation amount of machining parameters based on the chatter intensity when chatter is present, and to compensate the pre-optimized machining parameter set; if no chatter is detected, the pre-optimized machining parameter set remains unchanged. The process of generating a chatter suppression pre-optimization strategy adapted to the initial machining parameter set and outputting the pre-optimized machining parameter set is as follows: Clustering was performed using the absolute values ​​of the partial correlation coefficients between each processing parameter in the initial processing parameter group and the vibration amplitude and flutter occurrence rate as the clustering benchmark value, resulting in multiple clusters with the same number of flutter risk levels. The mean value of the clustering baseline value corresponding to each cluster is calculated, and the clusters are sorted in descending order based on the mean calculation results, and then associated with the flutter risk level one by one. Based on the determined flutter risk level, the processing parameters contained in the corresponding clusters are extracted, and the processing parameters corresponding to the largest cluster benchmark value are adjusted first. The adjustment direction is determined according to the sign of the partial correlation coefficient. If the adjusted flutter risk quantification value decreases by more than the set amplitude threshold, the adjustment is retained, and the processing parameters corresponding to the second largest cluster benchmark value are further adjusted to obtain the pre-optimized processing parameter set.

2. The intelligent control system for precision hardware machining based on CNC machine tools according to claim 1, characterized in that, The process of analyzing the three-dimensional geometric features and material properties of precision hardware parts to be machined, and determining the results of the three-dimensional feature analysis of the workpiece, is as follows: Based on the CAD design file, obtain the three-dimensional coordinate data of the easily chattering parts of the precision hardware parts to be processed, and preset the dynamic damping coefficient and the preliminary dynamic elastic modulus; Based on the workpiece type in the CAD design file, determine the type of workpiece structure prone to chatter, extract the geometric parameters of each chatter-prone part, and form a geometric feature set; Dynamic excitation tests are performed on workpiece samples by dynamic mechanical testing to directly obtain the actual dynamic damping coefficient and dynamic elastic modulus of the material. The results are compared with the preset dynamic damping coefficient and preliminary dynamic elastic modulus to correct the deviation data and form a flutter attribute set. The geometric feature set and the chatter attribute set are associated one-to-one, and the three-dimensional feature analysis results of the workpiece are generated according to the preset accuracy and the standardized data format of the unified measurement unit.

3. The intelligent control system for precision hardware machining based on CNC machine tools according to claim 1, characterized in that, The process of building a mapping database is as follows: For hardware parts with different types of easily chattering structures, the three-dimensional feature analysis results of the workpiece are obtained, and different processing parameter groups are matched accordingly. The vibration amplitude and chattering rate during the trial processing are collected, and after screening, they are standardized to obtain a ternary sample dataset. The standardized three-dimensional feature analysis results of the workpiece in the ternary sample dataset are transformed into independent variable vectors, the standardized processing parameter set is transformed into intermediate variable vectors, and the standardized vibration amplitude and chatter occurrence rate are transformed into dependent variable vectors. With the values ​​of the independent variable vector fixed, the partial correlation coefficient between each element in the intermediate variable vector and each element in the dependent variable vector is calculated based on the partial correlation coefficient calculation formula, and the partial correlation coefficient is filled into the correlation coefficient matrix. Based on the partial correlation coefficient, a set of processing parameters that meet the set conditions is selected. Then, based on the set of processing parameters, the parameter combination that minimizes the vibration amplitude and the chatter rate is locked. The parameter combination is then mapped to the corresponding workpiece three-dimensional feature analysis results, vibration amplitude and chatter rate to obtain a mapping database.

4. The intelligent control system for precision hardware machining based on CNC machine tools according to claim 3, characterized in that, The process of matching the initial machining parameter set corresponding to the workpiece 3D feature analysis result from the mapping database using a multi-factor weighted algorithm is as follows: Based on the analysis results of the three-dimensional features of the workpiece in the ternary sample dataset, the vibration amplitude and chattering rate are analyzed by grey relational analysis to determine the influence weight of each parameter in the analysis results of the three-dimensional features of the workpiece on the vibration amplitude and chattering rate. The three-dimensional feature analysis results of the current precision hardware part to be processed are compared with the three-dimensional feature analysis results of each workpiece in the mapping database to calculate the comprehensive similarity value by multi-factor weighted similarity. The overall similarity values ​​are sorted in descending order, and the parameter combination corresponding to the highest overall similarity value is selected as the initial processing parameter group.

5. The intelligent control system for precision hardware machining based on CNC machine tools according to claim 4, characterized in that, The process of weighting the influence of each parameter on vibration amplitude and flutter rate in the three-dimensional feature analysis results of the workpiece through grey relational analysis is as follows: The correlation between the parameters and the vibration amplitude and flutter incidence rate was analyzed by grey relational analysis. The combined relevance of the parameter is obtained by weighted summation of the two relevance degrees. The ratio of the overall correlation of each parameter to the sum of all overall correlations is used as the influence weight of that parameter.

6. The intelligent control system for precision hardware machining based on CNC machine tools according to claim 1, characterized in that, Based on the analysis results of the workpiece's three-dimensional features, the process of predicting chatter and determining the chatter risk level at different processing stages is as follows: The mapping database is queried to select the N processing parameter combinations that match the workpiece's three-dimensional feature analysis results at the current processing stage, and the corresponding vibration amplitude and chatter occurrence rate are extracted. The extracted vibration amplitude and flutter incidence were averaged to obtain the predicted vibration amplitude and predicted flutter incidence. The flutter risk quantification value is obtained by weighted summing the ratio of the predicted vibration amplitude to the vibration amplitude threshold and the ratio of the predicted flutter incidence rate to the flutter incidence rate threshold. The flutter risk level is then determined based on the flutter risk quantification value.

7. The intelligent control system for precision hardware machining based on CNC machine tools according to claim 1, characterized in that, The process of sending the pre-optimized machining parameter set to the CNC machine tool for machining, extracting vibration characteristics from the real-time data of the machining process, and determining whether chatter exists and its intensity is as follows: Real-time vibration data is collected at a preset sampling frequency, and noise reduction and standardization are performed to obtain a standardized vibration signal. Based on standardized vibration signals, time-domain and frequency-domain features are extracted respectively to form a multidimensional flutter feature vector; If any feature in the multidimensional chatter feature vector does not meet the set safety range, then chatter is determined to exist in the processing process; The number of features in the multidimensional flutter feature vector that do not meet the set safety range is taken as the flutter intensity.

8. The intelligent control system for precision hardware machining based on CNC machine tools according to claim 1, characterized in that, When chatter is present, the process of determining the compensation amount of machining parameters based on the chatter intensity and compensating the pre-optimized machining parameter set is as follows: Based on historical effective compensation cases, calculate the flutter comprehensive deviation index, comprehensive sensitivity weight, and processing parameter compensation range; A three-dimensional mapping table is established with flutter comprehensive deviation index as the horizontal axis, comprehensive sensitivity weight as the vertical axis, and processing parameter compensation range as the cell. Substitute the flutter comprehensive deviation index and comprehensive sensitivity weight calculated when flutter is present into the mapping relationship table to match the corresponding processing parameter compensation range. The compensation direction is determined according to the correlation direction between the corresponding processing parameter and the flutter intensity, and the processing parameter compensation amount is taken as the median value of the processing parameter compensation range. The pre-optimized baseline values ​​of each processing parameter are superimposed with the processing parameter compensation amounts to compensate the pre-optimized processing parameter set.

9. The intelligent control system for precision hardware machining based on CNC machine tools according to claim 8, characterized in that, The process of calculating the flutter comprehensive deviation index, comprehensive sensitivity weight, and processing parameter compensation range based on historical effective compensation cases is as follows: The average value of each flutter feature is obtained from historical effective compensation cases, and the deviation of each flutter feature is calculated based on the average value and the boundary of the corresponding safety range. The flutter comprehensive deviation index is obtained by weighting and summing the deviations based on the influence weights of each flutter feature. For a certain processing parameter, calculate its partial correlation coefficient with each chatter feature, and sum the absolute values ​​of the partial correlation coefficients to obtain the comprehensive partial correlation coefficient corresponding to the processing parameter. The ratio of the comprehensive partial correlation coefficient to the sum of the comprehensive partial correlation coefficients of all processing parameters is used as the comprehensive sensitivity weight. By statistically analyzing the minimum and maximum values ​​of compensation for each processing parameter in historical effective compensation cases, the range of compensation for processing parameters can be obtained.

Images