A method and system for optimizing spherical polishing parameters based on polishing combination processing

Abstract

The invention discloses a spherical polishing parameter optimization method and system based on polishing combination processing, which relates to the field of mechanical processing technology, and comprises: reconstructing the polishing trajectory data in time series, determining the embedding dimension and the delay time, reconstructing the one-dimensional polishing trajectory data into a multidimensional phase space, and calculating the first characteristic index; decomposing the vibration signal through empirical mode decomposition, calculating the correlation coefficient between each intrinsic mode function and the self-excited vibration, selecting the intrinsic mode function whose correlation coefficient with the self-excited vibration is greater than the correlation threshold for reconstruction, and calculating the root mean square value RMS through the reconstructed self-excited vibration signal; setting the optimization target to minimize the first characteristic index and the root mean square value, adjusting the spherical polishing parameters through a genetic algorithm, generating a plurality of parameter combinations, and determining the optimal parameter combination through a fitness function, thereby optimizing the polishing parameters and improving the polishing quality and efficiency.

Description

Technical Field

[0001] The invention relates to the technical field of mechanical processing, and in particular to a method and system for optimizing spherical polishing parameters based on polishing combination processing. Background Art

[0002] In the field of precision manufacturing and optical processing, spherical polishing is a crucial process. Its quality directly affects the performance and application range of the product. The main goal of spherical polishing is to obtain a low-roughness surface to meet the high-precision requirements of optical components, precision mechanical parts, etc.

[0003] However, chaotic vibration in the spherical polishing process is a highly complex and unpredictable dynamic phenomenon. Its root lies in the nonlinear factors and the complexity of interactions within the polishing system. These factors together lead to the instability of the moving trajectory of the polishing tool. Chaotic vibration not only makes the polishing trajectory difficult to accurately control, but may also cause uneven polishing marks or spots on the polishing surface, thereby seriously damaging the uniformity and consistency of the polishing quality.

[0004] In addition, self-excited vibration is another issue that needs to be paid special attention to during the spherical polishing process. When the polishing tool contacts the workpiece, self-excited vibration modes may be excited due to the dynamic characteristics and interactions within the system. This vibration will not only cause fluctuations in the polishing rate and affect the stability and repeatability of the polishing process, but will also aggravate the roughness changes of the polished surface, making it difficult for the final spherical surface roughness to meet the requirements of precision manufacturing. Summary of the invention

[0005] 1. Technical issues to be resolved

[0006] In view of the shortcomings of the prior art, the present invention provides a method and system for optimizing spherical polishing parameters based on polishing combination processing. By real-time acquisition and analysis of data in the spherical polishing process, advanced data analysis technologies and optimization algorithms such as time series reconstruction, chaos theory, and empirical mode decomposition are used to achieve effective monitoring and control of chaotic vibration and self-excited vibration, thereby optimizing polishing parameters and improving polishing quality and efficiency.

[0007] (II) Technical solution

[0008] To achieve the above objectives, the present invention is implemented through the following technical solutions: a method for optimizing spherical polishing parameters based on polishing combination processing, comprising:

[0009] Collect the processing data in the spherical polishing process in real time, and construct the spherical polishing data set after aggregation;

[0010] Extracting polishing trajectory data from the spherical polishing data set, reconstructing the polishing trajectory data in time series, determining the embedding dimension and delay time, reconstructing the one-dimensional polishing trajectory data into a multi-dimensional phase space, and calculating the first characteristic index through the distances of all adjacent point pairs in the multi-dimensional phase space at the initial time and at all time steps;

[0011] Extract vibration signal data from the spherical polishing data set, decompose the vibration signal into several intrinsic mode functions and a residual by empirical mode decomposition, calculate the correlation coefficient between each intrinsic mode function and the self-excited vibration, select the intrinsic mode function whose correlation coefficient with the self-excited vibration is greater than the correlation threshold for reconstruction, obtain the reconstructed self-excited vibration signal, and calculate the root mean square value RMS by the reconstructed self-excited vibration signal;

[0012] Initialize the spherical polishing parameters, set the optimization goal to minimize the first characteristic index and the root mean square value, adjust the spherical polishing parameters through the genetic algorithm, generate several parameter combinations, calculate the corresponding first characteristic index and root mean square value for each set of parameter combinations, and determine the optimal parameter combination through the fitness function.

[0013] Furthermore, the process of determining the embedding dimension includes:

[0014] S11: setting the initial value of dimension m, m=2, and calculating the nearest neighbor point of each point in the polishing trajectory data time series, which can be achieved by calculating the Euclidean distance or other suitable distance metrics;

[0015] S12: For each nearest neighbor point, check whether it still maintains the nearest neighbor relationship in different dimensions;

[0016] S13: Increment the dimension, m=m+1. If the nearest neighbor point changes after increasing the dimension, the point is considered to be a pseudo nearest neighbor.

[0017] S14: Repeat the above steps S11 to S13 until the number of pseudo nearest neighbors no longer changes, and use the current dimension as the embedding dimension.

[0018] Furthermore, the process of determining the delay time includes:

[0019] Set several preset delay times, and calculate the autocorrelation function value of the polishing trajectory data at different preset delay times. The calculation formula is: Where R(τ) represents the autocorrelation function value when the preset delay time is τ, x i represents the i-th data point in the polishing trajectory data time series, i = 1, 2, ..., N, N represents the number of data points, Indicates the average value of the data;

[0020] The calculated autocorrelation function value and the corresponding preset delay time are plotted into a curve graph, and the preset delay time corresponding to when the autocorrelation function value in the curve graph reaches the first minimum value is used as the delay time.

[0021] Furthermore, according to the embedding dimension and delay time, the one-dimensional polishing trajectory data is reconstructed into a multi-dimensional phase space, and the reconstruction formula is: i =(x i , x i +τ,…,x i +(m-1)τ), where Y i represents the i-th point in the point set of the m-dimensional phase space, m represents the embedding dimension, and τ represents the delay time.

[0022] Furthermore, in the multidimensional phase space, the nearest neighbor point is found for each point to form an adjacent point pair, and the distance between the adjacent point pairs at the initial moment is calculated. For each time step, the distance between the adjacent point pairs at the time step is calculated. The first characteristic index is calculated by the distance between all adjacent point pairs at the initial moment and at all time steps in the multidimensional phase space: Where λ represents the first characteristic index, h represents the time step, h = 1, 2, ..., H, H represents the total number of time steps, d h It represents the distance between two points at the time step h, and d0 represents the distance between two points at the initial moment.

[0023] Furthermore, the decomposition steps include:

[0024] S31: Find all local maxima and minima of the vibration signal;

[0025] S32: constructing an upper envelope and a lower envelope by cubic spline interpolation;

[0026] S33: Calculate the average values ​​of the upper envelope and the lower envelope to obtain a local mean line;

[0027] S34: subtracting the local mean line from the vibration signal to obtain a candidate component of the first intrinsic mode function;

[0028] S35: Check whether the candidate component meets the IMF condition: the number of local extreme points and zero-crossing points is the same or differs by 1. If the IMF condition is not met, the candidate component is used as a new vibration signal and the above steps are repeated. If the IMF condition is met, the candidate component is used as an intrinsic mode function.

[0029] S36: Repeat the above steps until the residual signal cannot be decomposed into a candidate component that meets the IMF condition. In each iteration, the extracted intrinsic mode function is subtracted from the new vibration signal to obtain the next residual signal to be decomposed.

[0030] Furthermore, each intrinsic mode function is subjected to Fourier transform to obtain a spectrum, and the peak value in the spectrum is extracted. Combined with the self-excited vibration frequency range, the correlation coefficient between each intrinsic mode function and the self-excited vibration is calculated. The calculation formula is as follows:

[0031]

[0032] Among them, r represents the correlation coefficient, a t and b t Respectively represent the values ​​of the intrinsic mode function and the self-excited vibration signal at time point t, and They are respectively a t and b t The average value of , t=1,2,…,T, where T represents the number of time points.

[0033] Furthermore, the reconstruction formula of the self-excited vibration signal is: Where S(t) represents the reconstructed self-excited vibration signal, IMF n Represents the nth selected intrinsic mode function; the root mean square value RMS is calculated by reconstructing the self-excited vibration signal: Wherein, M represents the number of time points in the reconstructed self-excited vibration signal.

[0034] Furthermore, the parameter combinations are sorted according to the fitness values, and the first several parameter combinations are selected according to the sorting results to enter the next round of iteration; check whether the optimization process meets the convergence conditions: the preset number of iterations is reached or the first characteristic index and the root mean square value reach the optimization target. If the convergence conditions are met, the iteration is stopped and the optimal polishing parameter combination is output, otherwise, the iteration continues.

[0035] A spherical polishing parameter optimization system based on polishing combined processing, comprising:

[0036] The data acquisition module collects the processing data in the spherical polishing process in real time and constructs the spherical polishing data set after aggregation;

[0037] The polishing trajectory analysis module extracts the polishing trajectory data from the spherical polishing data set, reconstructs the polishing trajectory data in time series, determines the embedding dimension and delay time, reconstructs the one-dimensional polishing trajectory data into a multi-dimensional phase space, and calculates the first characteristic index through the distances of all adjacent point pairs in the multi-dimensional phase space at the initial time and at all time steps;

[0038] The vibration signal analysis module extracts vibration signal data from the spherical polishing data set, decomposes the vibration signal into several intrinsic mode functions and a residual through empirical mode decomposition, calculates the correlation coefficient between each intrinsic mode function and the self-excited vibration, selects the intrinsic mode function whose correlation coefficient with the self-excited vibration is greater than the correlation threshold for reconstruction, obtains the reconstructed self-excited vibration signal, and calculates the root mean square value RMS through the reconstructed self-excited vibration signal;

[0039] The parameter optimization module initializes the spherical polishing parameters, sets the optimization goal to minimize the first characteristic index and the root mean square value, adjusts the spherical polishing parameters through the genetic algorithm, generates several parameter combinations, calculates the corresponding first characteristic index and root mean square value for each set of parameter combinations, and determines the optimal parameter combination through the fitness function.

[0040] (III) Beneficial effects

[0041] The present invention provides a method and system for optimizing spherical polishing parameters based on polishing combined processing, which has the following beneficial effects:

[0042] (1) Through phase space reconstruction, the one-dimensional polishing trajectory data is converted into a point set in the multidimensional phase space, which can more comprehensively capture the dynamic behavior of the polishing process. The first characteristic index is calculated as an important indicator of chaotic characteristics, which can accurately evaluate whether there is chaotic vibration in the polishing process.

[0043] (2) By calculating the first characteristic index and extracting vibration signal data from the spherical polishing data set based on the index, this method can accurately identify the self-excited vibration existing in the polishing process. By using empirical mode decomposition, the vibration signal is decomposed into several intrinsic mode functions (IMFs) and a residual, which can capture the fluctuation characteristics of the signal in different frequency ranges. By selecting IMFs with high correlation with self-excited vibrations for reconstruction, the characteristics of self-excited vibrations are further highlighted. By calculating the root mean square value of the reconstructed self-excited vibration signal, the influence of self-excited vibration on the polishing process can be quantitatively evaluated. The size of the RMS value directly reflects the intensity of the self-excited vibration.

[0044] (3) By minimizing the first characteristic index, it helps to reduce the chaotic behavior of the polishing trajectory and make the polishing trajectory more stable, thereby improving the uniformity of the polishing surface. At the same time, by minimizing the root mean square value (RMS) of the self-excited vibration signal, it can reduce the self-excited vibration during the polishing process, reduce the fluctuation of the polishing rate and surface roughness, and improve the stability and consistency of the polishing quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 It is a schematic diagram of the steps of the method for optimizing spherical polishing parameters based on polishing combination processing of the present invention;

[0046] Figure 2 This is a schematic diagram of the empirical mode decomposition process of the present invention;

[0047] Figure 3 It is a schematic diagram of the structure of the spherical polishing parameter optimization system based on the polishing combination processing of the present invention. DETAILED DESCRIPTION

[0048] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0049] See also Figure 1-Figure 2 The present invention provides a method for optimizing spherical polishing parameters based on polishing combined processing, comprising the following steps:

[0050] Step 1: Real-time collection of processing data during the spherical surface polishing process, including but not limited to movement trajectory, vibration signal, contact pressure, polishing rate and surface roughness after polishing, and then summarizing and constructing a spherical surface polishing data set;

[0051] The step 1 includes the following contents:

[0052] Step 101: Install a high-precision displacement sensor or laser rangefinder on the polishing machine to capture the moving trajectory of the polishing tool in real time, install a vibration sensor on the polishing machine to capture the vibration signal during the polishing process in real time, install a pressure sensor at the contact point between the polishing tool and the workpiece to accurately measure the contact pressure during the polishing process, install a rotation speed sensor or use the built-in monitoring system of the polishing machine to record the rotation rate of the polishing tool or the overall rate of the polishing process in real time, and use a portable or integrated surface roughness measuring instrument to detect the roughness of the workpiece surface before and after polishing;

[0053] Step 102: Remove abnormal values ​​or invalid data caused by sensor failure, data transmission error, etc., calibrate the sensor regularly to ensure that the collected data is accurate, and for surface roughness data, perform multiple measurements and take the average value to improve accuracy;

[0054] Step 103: Summarize the data from different sensors according to the timestamp to construct a spherical polishing data set, including but not limited to the moving trajectory of the polishing tool, the contact pressure, the polishing rate, and the surface roughness after polishing, and transmit the integrated data to the subsequent processing module;

[0055] When using, combine the contents of step 101 to step 103:

[0056] The comprehensiveness of the data is ensured by real-time collection of processing data including movement trajectory, contact pressure, polishing rate, and surface roughness after polishing. These data cover the key parameters and results of the polishing process, providing a solid foundation for subsequent analysis and optimization.

[0057] Step 2: extract polishing trajectory data from the spherical polishing data set, reconstruct the polishing trajectory data in time series, determine the embedding dimension and delay time, reconstruct the one-dimensional polishing trajectory data into a multidimensional phase space, and calculate the first characteristic index through the distances of all adjacent point pairs in the multidimensional phase space at the initial time and at all time steps;

[0058] The step 2 includes the following contents:

[0059] Step 201: extracting polishing trajectory data from the spherical polishing data set, reconstructing the polishing trajectory data in time series, sorting the polishing trajectory data in time order, and determining the embedding dimension and delay time. The process of determining the embedding dimension includes:

[0060] S11: setting the initial value of the dimension, such as m=2, and calculating the nearest neighbor point of each point in the polishing trajectory data time series, which can be achieved by calculating the Euclidean distance or other suitable distance metrics;

[0061] S12: For each nearest neighbor point, check whether it still maintains the nearest neighbor relationship in different dimensions;

[0062] S13: Increasing the dimension, for example, m=m+1. If the nearest neighbor point changes (i.e., it is no longer the nearest neighbor) after increasing the dimension, the point is considered to be a pseudo nearest neighbor.

[0063] S14: Repeat the above steps S11 to S13 until the number of pseudo nearest neighbors no longer changes, and use the current dimension as the embedding dimension;

[0064] The process of determining the delay time includes:

[0065] S15: setting a number of preset delay times, and calculating the autocorrelation function value of the polishing trajectory data at different preset delay times, and the calculation formula is: Where R(τ) represents the autocorrelation function value when the preset delay time is τ, x i represents the i-th data point in the polishing trajectory data time series, i = 1, 2, ..., N, N represents the number of data points, Indicates the average value of the data;

[0066] S16: Plotting the calculated autocorrelation function value and the corresponding preset delay time into a curve graph, and finding the preset delay time corresponding to when the autocorrelation function value reaches the first minimum value in the curve graph as the delay time;

[0067] It should be noted that the embedding dimension is a key parameter in the phase space reconstruction, which determines the complexity of the phase space. If the embedding dimension is too low, the dynamic behavior may not be fully captured; if the embedding dimension is too high, it may increase the computational complexity and noise interference. The delay time is another important parameter, which determines the time interval between adjacent data points in the time series. Choosing a suitable delay time can ensure that the points in the phase space can fully reflect the dynamic behavior.

[0068] Step 202: Reconstruct the one-dimensional polishing trajectory data into a multi-dimensional phase space (i.e., m-dimensional phase space) according to the embedding dimension and the delay time. The reconstruction formula is: i =(x i , x i +τ,…,x i +(m-1)τ), where Y i represents the i-th point in the point set of the m-dimensional phase space, m represents the embedding dimension, and τ represents the delay time;

[0069] Step 203: In the multidimensional phase space, find the nearest neighbor point for each point to form an adjacent point pair, calculate the distance between the adjacent point pairs at the initial time (h = 0), and for each time step (h = 1, 2, ..., H), calculate the distance between the adjacent point pairs at time h. Calculate the first characteristic index through the distances between all adjacent point pairs in the multidimensional phase space at the initial time and time h. The calculation formula is: Where λ represents the first characteristic index, h represents the time step, h = 1, 2, ..., H, H represents the total number of time steps, d h represents the distance between two points at time step h, d0 represents the distance between two points at the initial moment, and the distance is obtained by calculating the Euclidean distance or other appropriate distance metric;

[0070] It should be noted that when λ is greater than zero, chaotic behavior exists. The larger the value of λ, the higher the degree of chaos. Conversely, the smaller the value of λ, the lower the degree of chaos. Chaotic behavior is a complex, nonlinear dynamic behavior that manifests itself as the extreme sensitivity of the system state to the initial conditions, that is, a small initial difference will lead to a huge difference in long-term behavior.

[0071] When used, combine the contents of step 201 to step 203:

[0072] Through phase space reconstruction, the one-dimensional polishing trajectory data is converted into a point set in the multidimensional phase space, which can more comprehensively capture the dynamic behavior of the polishing process. The calculation of the first characteristic index as an important indicator of chaotic characteristics can accurately evaluate whether there is chaotic vibration in the polishing process.

[0073] Step 3: Extract vibration signal data from the spherical polishing data set, decompose the vibration signal into several intrinsic mode functions and a residual through empirical mode decomposition, calculate the correlation coefficient between each intrinsic mode function and the self-excited vibration, select the intrinsic mode function whose correlation coefficient with the self-excited vibration is greater than the correlation threshold for reconstruction, obtain the reconstructed self-excited vibration signal, and calculate the root mean square value RMS through the reconstructed self-excited vibration signal;

[0074] The step three includes the following contents:

[0075] Step 301: extract vibration signal data from the spherical polishing data set, and decompose the vibration signal into several intrinsic mode functions and a residual by empirical mode decomposition. The decomposition step includes:

[0076] S31: Find all local maxima and minima of the vibration signal;

[0077] S32: constructing an upper envelope and a lower envelope by cubic spline interpolation;

[0078] S33: Calculate the average values ​​of the upper envelope and the lower envelope to obtain a local mean line;

[0079] S34: subtracting the local mean line from the vibration signal to obtain a candidate component of the first intrinsic mode function;

[0080] S35: Check whether the candidate component meets the IMF condition: the number of local extreme points and zero-crossing points is the same or differs by 1. If the IMF condition is not met, the candidate component is used as a new vibration signal and the above steps S31 to S34 are repeated. If the IMF condition is met, the candidate component is used as an intrinsic mode function.

[0081] S36: repeating the above steps S31 to S35 until the residual signal cannot be decomposed into a candidate component that meets the IMF condition. In each iteration, the extracted intrinsic mode function is subtracted from the new vibration signal to obtain the next residual signal to be decomposed;

[0082] Step 302: Perform Fourier transform on each eigenmode function to obtain a spectrum: Among them, G(f) represents the spectrum of the signal g(t), f represents frequency, t represents time, and j represents an imaginary unit. The peak value in the spectrum is extracted. The peak value corresponds to the significant frequency component in the intrinsic mode function. Combined with the self-excited vibration frequency range, the correlation coefficient between each intrinsic mode function and the self-excited vibration is calculated. The calculation formula is as follows:

[0083]

[0084] Among them, r represents the correlation coefficient, a t and b t Respectively represent the values ​​of the intrinsic mode function and the self-excited vibration signal at time point t, and They are respectively a t and b t The average value of , t=1,2,…,T, T represents the number of time points;

[0085] Step 303: pre-set a correlation threshold, select an intrinsic mode function whose correlation coefficient with the self-excited vibration is greater than the correlation threshold for reconstruction, and obtain a reconstructed self-excited vibration signal. The reconstruction formula is: Where S(t) represents the reconstructed self-excited vibration signal, IMF n represents the nth selected intrinsic mode function;

[0086] It should be noted that the setting of the correlation threshold is to perform statistical distribution analysis on the correlation coefficients between multiple IMFs and self-excited vibrations, find out their distribution characteristics (such as mean, standard deviation, etc.), and set a reasonable threshold range based on these characteristics;

[0087] Step 304: Calculate the root mean square value RMS through the reconstructed self-excited vibration signal: Where M represents the number of time points in the reconstructed self-excited vibration signal;

[0088] When used, combine the contents of step 301 to step 304:

[0089] By calculating the first characteristic index and extracting vibration signal data from the spherical polishing data set based on the index, this method can accurately identify the self-excited vibration existing in the polishing process. By using empirical mode decomposition, the vibration signal is decomposed into several intrinsic mode functions (IMFs) and a residual, which can capture the fluctuation characteristics of the signal in different frequency ranges. By selecting IMFs with high correlation with self-excited vibrations for reconstruction, the characteristics of self-excited vibrations are further highlighted. By calculating the root mean square value of the reconstructed self-excited vibration signal, the influence of self-excited vibration on the polishing process can be quantitatively evaluated. The size of the RMS value directly reflects the intensity of the self-excited vibration.

[0090] Step 4: Initialize the spherical polishing parameters, set the optimization goal to minimize the first characteristic index and the root mean square value, adjust the spherical polishing parameters through the genetic algorithm, generate several parameter combinations, calculate the corresponding first characteristic index and root mean square value for each set of parameter combinations, and select the optimal parameter combination according to the fitness function for the next round of iteration until the convergence conditions are met.

[0091] The step 4 includes the following contents:

[0092] Step 401: obtaining the first characteristic index and the root mean square value RMS of the self-excited vibration signal, and initializing the spherical polishing parameters, including the rotation speed, feed speed, contact pressure of the polishing tool, and the type and concentration of the polishing liquid;

[0093] Step 402: setting the optimization goal to minimize the first characteristic index and the root mean square value, adjusting the spherical polishing parameters through the genetic algorithm, generating several parameter combinations, repeating steps 1 to 3 for each group of parameter combinations, calculating the corresponding first characteristic index and root mean square value, sorting the parameter combinations according to the fitness value, and then selecting the first several parameter combinations according to the sorting result to enter the next round of iteration, the fitness function: Fi = ω1*λ+ω2*RMS, where Fi represents the fitness value, ω1 and ω2 represent weight coefficients, which are used to balance the influence of the first characteristic index and the root mean square value in the fitness function, and the weight coefficients can be adjusted according to the actual situation;

[0094] Step 403: Check whether the optimization process meets the convergence condition: the preset number of iterations is reached or the first characteristic index and the root mean square value reach the optimization target. If the convergence condition is met, the iteration is stopped and the optimal polishing parameter combination is output. Otherwise, the iteration is continued.

[0095] It should be noted that the iterative process of the genetic algorithm is as follows: first, a set of initial polishing parameter combinations (i.e., initial population) is randomly generated, and then, for each set of parameter combinations, the corresponding first characteristic index and root mean square value are calculated according to the steps 1 to 3 described above, and then the fitness function value is calculated. Next, according to the fitness function value, a part of excellent parameter combinations are selected as the parent generation, and new offspring parameter combinations are generated through crossover operation. At the same time, in order to maintain the diversity of the population, mutation operations are performed on some offspring parameter combinations to obtain a new generation of parameter combination population;

[0096] When used, combine the contents of step 401 to step 403:

[0097] By minimizing the first characteristic index, it helps to reduce the chaotic behavior of the polishing trajectory and make the polishing trajectory more stable, thereby improving the uniformity of the polishing surface. At the same time, by minimizing the root mean square value (RMS) of the self-excited vibration signal, it can reduce the self-excited vibration in the polishing process, reduce the fluctuation of the polishing rate and surface roughness, and improve the stability and consistency of the polishing quality.

[0098] See also Figure 3 The present invention also provides a spherical polishing parameter optimization system based on polishing combination processing, including: a data acquisition module, a polishing trajectory analysis module, a vibration signal analysis module and a parameter optimization module; wherein,

[0099] The data acquisition module collects the processing data in the spherical polishing process in real time and constructs the spherical polishing data set after aggregation;

[0100] The polishing trajectory analysis module extracts the polishing trajectory data from the spherical polishing data set, reconstructs the polishing trajectory data in time series, determines the embedding dimension and delay time, reconstructs the one-dimensional polishing trajectory data into a multi-dimensional phase space, and calculates the first characteristic index through the distances of all adjacent point pairs in the multi-dimensional phase space at the initial time and at all time steps;

[0101] The vibration signal analysis module extracts vibration signal data from the spherical polishing data set, decomposes the vibration signal into several intrinsic mode functions and a residual through empirical mode decomposition, calculates the correlation coefficient between each intrinsic mode function and the self-excited vibration, selects the intrinsic mode function whose correlation coefficient with the self-excited vibration is greater than the correlation threshold for reconstruction, obtains the reconstructed self-excited vibration signal, and calculates the root mean square value RMS through the reconstructed self-excited vibration signal;

[0102] The parameter optimization module initializes the spherical polishing parameters, sets the optimization goal to minimize the first characteristic index and the root mean square value, adjusts the spherical polishing parameters through the genetic algorithm, generates several parameter combinations, calculates the corresponding first characteristic index and root mean square value for each set of parameter combinations, and determines the optimal parameter combination through the fitness function.

[0103] In the application, the several formulas involved are all calculated by removing dimensions and taking their numerical values. The formula is a formula of the most recent real situation obtained by collecting a large amount of data and performing software simulation. The coefficients in the formula are set by technicians in this field according to actual conditions.

[0104] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination thereof. When implemented by software, the above embodiments may be implemented in whole or in part in the form of a computer program product. A person of ordinary skill in the art may appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein may be implemented in combination with electronic hardware, computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution.

[0105] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, and may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0106] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

Claims

1. A method for optimizing spherical polishing parameters based on polishing combination processing, characterized in that: include: Collect the processing data in the spherical polishing process in real time, and construct the spherical polishing data set after aggregation; Polishing trajectory data is extracted from the spherical polishing data set, the time series of the polishing trajectory data is reconstructed, and the embedding dimension and delay time are determined. The process of determining the embedding dimension includes: S11: setting the initial value of dimension m, m=2, and calculating the nearest neighbor point of each point in the polishing trajectory data time series; S12: For each nearest neighbor point, check whether it still maintains the nearest neighbor relationship in different dimensions; S13: Increment the dimension, m=m+1. If the nearest neighbor point changes after increasing the dimension, the point is considered to be a pseudo nearest neighbor. S14: Repeat the above steps S11 to S13 until the number of pseudo nearest neighbors no longer changes, and use the current dimension as the embedding dimension; The process of determining the delay time includes: Set several preset delay times, and calculate the autocorrelation function value of the polishing trajectory data at different preset delay times. The calculation formula is: Where R(τ) represents the autocorrelation function value when the preset delay time is τ, x i represents the i-th data point in the polishing trajectory data time series, i = 1, 2, ..., N, N represents the number of data points, Indicates the average value of the data; The calculated autocorrelation function value and the corresponding preset delay time are plotted into a curve graph, and the preset delay time corresponding to the first minimum value of the autocorrelation function value in the curve graph is taken as the delay time; The one-dimensional polishing trajectory data is reconstructed into a multidimensional phase space, and the first characteristic index is calculated through the distances of all adjacent point pairs in the multidimensional phase space at the initial moment and at all time steps, including: In the multidimensional phase space, find the nearest neighbor point for each point to form a pair of adjacent points, calculate the distance between the adjacent point pairs at the initial moment, and for each time step, calculate the distance between the adjacent point pairs at that time step. Calculate the first characteristic index through the distances of all adjacent point pairs in the multidimensional phase space at the initial moment and at all time steps: Where λ represents the first characteristic index, H represents the total number of time steps, and d h It represents the distance between two points at the time step h, where h represents the time step, h=1,2,…,H, and d0 represents the distance between two points at the initial moment; Extract vibration signal data from the spherical polishing data set, decompose the vibration signal into several intrinsic mode functions and a residual by empirical mode decomposition, calculate the correlation coefficient between each intrinsic mode function and the self-excited vibration, select the intrinsic mode function whose correlation coefficient with the self-excited vibration is greater than the correlation threshold for reconstruction, obtain the reconstructed self-excited vibration signal, and calculate the root mean square value RMS by the reconstructed self-excited vibration signal; Initialize the spherical polishing parameters, set the optimization goal to minimize the first characteristic index and the root mean square value, adjust the spherical polishing parameters through the genetic algorithm, generate several parameter combinations, calculate the corresponding first characteristic index and root mean square value for each set of parameter combinations, and determine the optimal parameter combination through the fitness function.

2. The method for optimizing spherical polishing parameters based on polishing combined processing according to claim 1, characterized in that: According to the embedding dimension and delay time, the one-dimensional polishing trajectory data is reconstructed into the multi-dimensional phase space. The reconstruction formula is: i =(x i , x i +τ,…,x i +(m-1)τ), where Y i represents the i-th point in the point set of the m-dimensional phase space, m represents the embedding dimension, and τ represents the delay time.

3. The method for optimizing spherical polishing parameters based on polishing combined processing according to claim 1, characterized in that: The decomposition steps include: S31: Find all local maxima and minima of the vibration signal; S32: constructing an upper envelope and a lower envelope by cubic spline interpolation; S33: Calculate the average values ​​of the upper envelope and the lower envelope to obtain a local mean line; S34: subtracting the local mean line from the vibration signal to obtain a candidate component of the first intrinsic mode function; S35: Check whether the candidate component meets the intrinsic mode function condition: the number of local extreme points and zero-crossing points is the same or differs by 1. If the intrinsic mode function condition is not met, the candidate component is used as a new vibration signal and the above steps S31 to S34 are repeated. If the intrinsic mode function condition is met, the candidate component is used as the intrinsic mode function. S36: Repeat the above steps S31 to S35 until the residual signal cannot be decomposed into candidate components that meet the conditions of the intrinsic mode function. In each iteration, the extracted intrinsic mode function is subtracted from the new vibration signal to obtain the next residual signal to be decomposed.

4. The method for optimizing spherical polishing parameters based on polishing combined processing according to claim 3 is characterized in that: Perform Fourier transform on each intrinsic mode function to obtain the spectrum, extract the peak value in the spectrum, and calculate the correlation coefficient between each intrinsic mode function and the self-excited vibration in combination with the self-excited vibration frequency range. The calculation formula is as follows: Among them, r represents the correlation coefficient, a t and b t Respectively represent the values ​​of the intrinsic mode function and the self-excited vibration signal at time point t, and They are respectively t and b t The average value of , t=1,2,…,T, where T represents the number of time points.

5. The method for optimizing spherical polishing parameters based on polishing combined processing according to claim 4 is characterized in that: The reconstruction formula of the self-excited vibration signal is: Where S(t) represents the reconstructed self-excited vibration signal, IMF n Represents the nth selected intrinsic mode function; the root mean square value RMS is calculated by reconstructing the self-excited vibration signal: Wherein, M represents the number of time points in the reconstructed self-excited vibration signal.

6. The method for optimizing spherical polishing parameters based on polishing combined processing according to claim 1, characterized in that: Sort the parameter combinations according to the fitness values, and select the first several parameter combinations according to the sorting results to enter the next round of iteration; check whether the optimization process meets the convergence conditions: the preset number of iterations is reached or the first characteristic index and the root mean square value reach the optimization target. If the convergence conditions are met, stop the iteration and output the optimal polishing parameter combination, otherwise, continue the iteration.

7. A spherical polishing parameter optimization system based on polishing combination processing, used to implement the method described in any one of claims 1 to 6, characterized in that: include: The data acquisition module collects the processing data in the spherical polishing process in real time and constructs the spherical polishing data set after aggregation; The polishing trajectory analysis module extracts the polishing trajectory data from the spherical polishing data set, reconstructs the polishing trajectory data in time series, determines the embedding dimension and delay time, reconstructs the one-dimensional polishing trajectory data into a multi-dimensional phase space, and calculates the first characteristic index through the distances of all adjacent point pairs in the multi-dimensional phase space at the initial time and at all time steps; The vibration signal analysis module extracts vibration signal data from the spherical polishing data set, decomposes the vibration signal into several intrinsic mode functions and a residual through empirical mode decomposition, calculates the correlation coefficient between each intrinsic mode function and the self-excited vibration, selects the intrinsic mode function whose correlation coefficient with the self-excited vibration is greater than the correlation threshold for reconstruction, obtains the reconstructed self-excited vibration signal, and calculates the root mean square value RMS through the reconstructed self-excited vibration signal; The parameter optimization module initializes the spherical polishing parameters, sets the optimization goal to minimize the first characteristic index and the root mean square value, adjusts the spherical polishing parameters through the genetic algorithm, generates several parameter combinations, calculates the corresponding first characteristic index and root mean square value for each set of parameter combinations, and determines the optimal parameter combination through the fitness function.

Images