Auxiliary material parameter optimization method, device, equipment and storage medium

Abstract

The present application discloses a method, device, equipment, and storage medium for optimizing auxiliary material parameters, relating to the field of semiconductor analysis technology. The method comprises: obtaining auxiliary material parameter data during the cutting process and using it as a first data set; determining a second data set based on the first data set and a preset fitting algorithm; determining a target cutting yield function based on the second data set and a pre-built parameter adjustment model; spatially dividing the first data set according to preset partitioning conditions to obtain multiple polyhedral regions; and determining a target auxiliary material parameter combination based on the target cutting yield function, the preset partitioning conditions, and each polyhedral region. The auxiliary material parameter optimization method provided in the present application optimizes and adjusts various auxiliary material parameters that affect cutting quality to determine the optimal parameter combination, thereby optimizing or predicting multiple cutting auxiliary material parameter combinations and ensuring cutting quality.

Description

Technical Field

[0001] The present application relates to the field of semiconductor analysis technology, and in particular to a method, device, equipment and storage medium for optimizing auxiliary material parameters. Background Art

[0002] In modern manufacturing, cutting processes often involve a variety of cutting accessories. Because the performance of these accessories crucially impacts cutting quality, their parameters need to be optimized or adjusted. However, complex nonlinear relationships and potential interactions exist between different accessory parameters. Existing methods for optimizing these parameters often analyze the impact of a single parameter or factor on cutting quality and are therefore not suitable for optimizing multiple accessory parameters. Summary of the Invention

[0003] The embodiments of the present application provide a method, device, equipment and storage medium for optimizing auxiliary material parameters to optimize or predict a combination of multiple auxiliary material parameters, thereby ensuring cutting quality.

[0004] In order to solve the above technical problems, the embodiments of the present application disclose the following technical solutions:

[0005] In a first aspect, a method for optimizing auxiliary material parameters is provided, which is applied to adjust auxiliary material parameters of a cutting device. The method comprises:

[0006] Acquire auxiliary material parameter data during the cutting process and use it as the first data set;

[0007] determining a second data set according to the first data set and a preset fitting algorithm;

[0008] determining a target cutting yield function based on the second data set and a pre-built parameter adjustment model;

[0009] Performing spatial division on the first data set according to a preset division condition to obtain a plurality of polyhedral regions;

[0010] A target auxiliary material parameter combination is determined according to the target cutting yield function, the preset division condition and each of the polyhedral regions.

[0011] In some embodiments, the preset division condition includes a first preset division condition and a second preset division condition;

[0012] The first preset division condition is: the data dimension of the first data set is smaller than the preset dimension and the data segmentation granularity is larger than the preset segmentation granularity;

[0013] The second preset division condition is: the data dimension of the first data set is greater than or equal to the preset dimension, or the data segmentation granularity of the first data set is smaller than the preset segmentation granularity.

[0014] In some embodiments, the method for determining a target excipient parameter combination comprises:

[0015] When the preset division condition is the first preset division condition, determining the target auxiliary material parameter combination according to the center point and the vertex of each of the polyhedral regions and the target cutting yield function;

[0016] When the preset division condition is the second preset division condition, the target auxiliary material parameter combination is determined using a Bayesian optimization algorithm according to the target cutting yield function and the center points of each of the polyhedral regions.

[0017] In some embodiments, determining the target auxiliary material parameter combination according to the center points and vertices of each of the polyhedral regions and the target cutting yield function includes:

[0018] Determining an average yield of each polyhedral region according to the center point and the vertex of each polyhedral region and the target cutting yield function;

[0019] The target auxiliary material parameter combination is determined according to the average yield of each polyhedral region.

[0020] In some embodiments, determining the target auxiliary material parameter combination using a Bayesian optimization algorithm based on the target cutting yield function and the center points of each polyhedral region includes:

[0021] Constructing a yield prediction agent model according to the target cutting yield function and the center points of each polyhedral region;

[0022] determining an acquisition function according to the yield prediction agent model;

[0023] Optimizing the acquisition function to determine the center point corresponding to the polyhedron region with the maximum expected improvement in the current iteration cycle;

[0024] According to the center point corresponding to the polyhedron region with the largest expected improvement in the current iteration cycle, the operation of updating the yield prediction agent model is returned and repeatedly executed until the target auxiliary material parameter combination is determined when a preset iteration stop condition is met.

[0025] In some embodiments, a method of determining a second data set includes:

[0026] Fitting the first data set according to the preset fitting algorithm to determine the influence function of each auxiliary material parameter on the cutting yield;

[0027] Determine the cutting yield value corresponding to each auxiliary material parameter according to each influence function;

[0028] The second data set is determined according to the cutting yield values ​​corresponding to the various auxiliary material parameters.

[0029] In some embodiments, a method for determining a target cutting yield function includes:

[0030] The second data set is input into the parameter adjustment model for adjustment, and the target cutting yield function is determined through regularization constraints; the parameter adjustment model is a Lasso regression model.

[0031] In a second aspect, a device for optimizing auxiliary material parameters is provided, comprising: an acquisition module for acquiring auxiliary material parameter data during a cutting process as a first data set;

[0032] A first determining module, configured to determine a second data set based on the first data set and a preset fitting algorithm;

[0033] a second determining module, configured to determine a target cutting yield function according to the second data set and a pre-built parameter adjustment model;

[0034] A spatial division module, configured to spatially divide the first data set according to a preset division condition to obtain a plurality of polyhedral regions;

[0035] The third determination module is used to determine a target auxiliary material parameter combination according to the target cutting yield function, the preset division condition and each of the polyhedral regions.

[0036] In a third aspect, an electronic device is provided, comprising a processor and a memory; a computer program is stored in the memory, and the processor is used to execute the computer program stored in the memory to implement the auxiliary material parameter optimization method as described in any one of the first aspects.

[0037] In a fourth aspect, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the auxiliary material parameter optimization method as described in any one of the first aspects when executed.

[0038] One of the above technical solutions has the following advantages or beneficial effects:

[0039] Compared with the prior art, the auxiliary material parameter optimization method of the present application includes: obtaining auxiliary material parameter data during the cutting process and using it as a first data set; determining a second data set based on the first data set and a preset fitting algorithm; determining a target cutting yield function based on the second data set and a pre-built parameter adjustment model; spatially dividing the first data set according to preset division conditions to obtain multiple polyhedral regions; and determining a target auxiliary material parameter combination based on the target cutting yield function, the preset division conditions, and each polyhedral region. The auxiliary material parameter optimization method provided in the present application optimizes and adjusts various auxiliary material parameters that affect cutting quality to determine the optimal parameter combination, thereby optimizing or predicting multiple cutting auxiliary material parameter combinations and ensuring cutting quality.

[0040] The auxiliary material parameter optimization device of the present application optimizes and adjusts various auxiliary material parameters that affect cutting quality to determine the optimal parameter combination, thereby optimizing or predicting a variety of cutting auxiliary material parameter combinations, thereby ensuring cutting quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0042] Figure 1 This is a flow chart of a method for optimizing auxiliary material parameters provided in the examples of this application;

[0043] Figure 2 Schematic diagram of the overall process of the excipient parameter optimization method provided in the examples of this application;

[0044] Figure 3 This is a three-dimensional effect diagram after optimizing the parameters of the cutting fluid provided in the embodiment of the present application;

[0045] Figure 4 This is a schematic structural diagram of an auxiliary material parameter optimization device according to an embodiment of the present application;

[0046] Figure 5 A schematic structural diagram of an electronic device according to an embodiment of the present application;

[0047] Reference numerals:

[0048] 101 - acquisition module; 102 - first determination module; 103 - second determination module; 104 - space division module; 105 - third determination module; 100 - auxiliary material parameter optimization device; 501 - memory; 502 - processor. DETAILED DESCRIPTION

[0049] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the embodiments described are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without making creative efforts are within the scope of protection of this application.

[0050] In the description of the present application, it should be understood that the terms "upper", "lower", "front", "back", "left", "right", "top", "bottom", "inside", "outside" and the like indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present application. In addition, the terms "first" and "second" are used for descriptive purposes only, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, the features defined as "first" and "second" may explicitly or implicitly include one or more features. In the description of the present application, "multiple" means two or more, and at least one means one, two or more, unless otherwise clearly and specifically defined.

[0051] Figure 1 This is a flow chart of a method for optimizing auxiliary material parameters provided in an embodiment of the present application. This method can be applied to semiconductor silicon wafer analysis systems to adjust auxiliary material parameters of cutting equipment and optimize or predict multiple combinations of cutting auxiliary material parameters that affect silicon wafer cutting quality. This method can be executed by an auxiliary material parameter optimization device, which can be implemented in software and / or hardware and can be configured in a processor of a semiconductor analysis system. Figure 1 , the method comprises the following steps:

[0052] Step 110: Acquire auxiliary material parameter data during the cutting process and use it as the first data set.

[0053] Auxiliary materials include cutting fluid, lubricants, cleaning agents, and more. The performance of these materials plays a crucial role in cutting quality. For example, cutting fluid, for example, has multiple properties—such as surface tension of the finished product, pH value, temperature, conductivity, turbidity, chemical oxygen demand (COD), concentration, and fluid flow per blade—that interact with the cutting process and directly impact the final cutting yield.

[0054] Among them, the auxiliary material parameter data in the cutting process include various auxiliary material parameter data that affect the cutting quality, such as cutting fluid turbidity, cutting fluid concentration, cutting fluid temperature, cutting fluid pH value, etc., which can be set according to actual conditions and are not specifically limited here.

[0055] Step 120: Determine a second data set based on the first data set and a preset fitting algorithm.

[0056] Specifically, since the first data set is the auxiliary material parameter data in the cutting process, by fitting the first data set through a preset fitting algorithm, it is possible to fit various auxiliary material parameter data that affect the cutting quality, so as to characterize the influence of each auxiliary material parameter data on the cutting quality, and obtain a new data set based on each fitting result, namely the second data set.

[0057] For example, in the technical solution of the embodiment of the present application, the preset fitting algorithm is a piecewise linear function fitting algorithm (PWLF), which will not be described in detail below. In addition, the preset fitting algorithm can also be other fitting algorithms, which can be set according to actual conditions and are not specifically limited here.

[0058] In some embodiments, the method for determining the second data set includes: fitting the first data set according to a preset fitting algorithm to determine the influence function of each auxiliary material parameter on the cutting yield; determining the cutting yield value corresponding to each auxiliary material parameter according to each influence function; and determining the second data set according to the cutting yield values ​​corresponding to each auxiliary material parameter.

[0059] Specifically, by fitting the auxiliary material parameter data during the cutting process using a multi-segment linear function fitting algorithm, we can obtain the influence function of each auxiliary material parameter on the cutting yield. Then, based on each influence function, we can calculate the cutting yield value corresponding to each auxiliary material parameter. The cutting yield value corresponding to each auxiliary material parameter is then used as a new variable data set, the second data set. This fitting algorithm analysis facilitates subsequent rapid prediction of the optimal auxiliary material parameter combination, avoiding the high cost and low efficiency of traditional experiments.

[0060] For example, taking the auxiliary material parameter data as the cutting fluid key point data, the specific implementation process of determining the second data set is as follows: First, collect the cutting fluid key point data through experiments. Then, use a multi-segment linear function to fit each auxiliary material parameter. The relationship between the cutting yield y and the influence function of each auxiliary material parameter on the cutting yield is obtained. ,Right now:

[0061] ;

[0062] in, 、 For the The fitting parameters of the segment; For the The value range of the segment.

[0063] Finally, according to the influence function of each auxiliary material parameter on the cutting yield , calculate the cutting yield value corresponding to each auxiliary material parameter, such as in, is the number of auxiliary material parameters. And the cutting yield value corresponding to each auxiliary material parameter is used as a new variable data set , that is, the second data set.

[0064] Step 130: Determine a target cutting yield function based on the second data set and the pre-built parameter adjustment model.

[0065] Specifically, by fine-tuning the cutting yield values ​​corresponding to various auxiliary material parameters using a pre-built parameter adjustment model, auxiliary material parameters with minimal impact on the cutting yield can be eliminated while retaining those with greater influence. This facilitates subsequent optimization of auxiliary material parameter combinations and improves the applicability and accuracy of the parameter adjustment model. Furthermore, model construction and fitting algorithm analysis facilitate rapid prediction of optimal auxiliary material parameter combinations, such as the optimal cutting fluid ratio, avoiding the high cost and low efficiency of traditional testing. This method is simple and easy to implement, can be integrated into production management systems, and enables real-time optimization.

[0066] In some embodiments, the method for determining the target cutting yield function includes: inputting the second data set into a parameter adjustment model for adjustment, and determining the target cutting yield function through regularization constraints; the parameter adjustment model is a Lasso regression model.

[0067] Among them, the parameter adjustment model is the Lasso regression model.

[0068] Specifically, the cutting yield value corresponding to each auxiliary material parameter, that is, the second data set, is input into the Lasso regression model for adjustment. And by introducing regularization constraints in the Lasso regression model, the target cutting yield function is obtained. Regularization constraints can ensure the sparsity of regression coefficients, thereby effectively eliminating auxiliary material parameters that have little impact on cutting yield. At the same time, all coefficients are restricted to positive numbers to retain the influence trend of a single auxiliary material parameter factor.

[0069] Among them, the calculation formula of the Lasso regression model is:

[0070] ;

[0071] in, .

[0072] Among them, the target cutting yield function for:

[0073] ;

[0074] in, is the regression coefficient; is the regularization parameter; is the number of samples; and Is a positive integer.

[0075] Step 140: spatially divide the first data set according to a preset division condition to obtain a plurality of polyhedral regions.

[0076] The shape of the polyhedron region may be a cube, a regular tetrahedron, etc., and may be set according to actual conditions, and is not specifically limited here.

[0077] Specifically, the first data set is spatially divided according to a preset division condition to obtain a plurality of polyhedral regions, whereby the auxiliary material parameter data is distributed in the plurality of polyhedrons. The amount of auxiliary material parameters included in each polyhedron may be the same or different. The auxiliary material parameter data may be distributed at the center and / or vertex positions of the corresponding polyhedron.

[0078] In some embodiments, the preset partitioning condition includes a first preset partitioning condition and a second preset partitioning condition; wherein, the first preset partitioning condition is: the data dimension of the first data set is smaller than the preset dimension and the data segmentation granularity is larger than the preset segmentation granularity; the second preset partitioning condition is: the data dimension of the first data set is greater than or equal to the preset dimension, or, the data segmentation granularity of the first data set is smaller than the preset segmentation granularity.

[0079] The data dimension of the first data set refers to the number of auxiliary material parameter data items contained in the first data set. The data segmentation granularity of the first data set refers to the size of the data intervals of the auxiliary material parameter data in the first data set. The smaller the data intervals, the finer the data segmentation granularity; conversely, the coarser the data segmentation granularity.

[0080] Specifically, when the data dimension of the first data set is smaller than the preset dimension and the data segmentation granularity is larger than the preset segmentation granularity, it indicates that the data dimension of the first data set is small and the segmentation granularity is coarse. When the data dimension of the first data set is greater than or equal to the preset dimension, it indicates that the data dimension of the first data set is high. When the data segmentation granularity of the first data set is smaller than the preset segmentation granularity, it indicates that the segmentation granularity of the first data set is fine.

[0081] The preset dimension can be a value such as 5000, and the specific value can be set according to the actual situation of the auxiliary material parameters, and no specific limitation is made here. The preset segmentation granularity can be divided into 4 or 5 intervals, and the specific value can be set according to the actual situation, and no specific limitation is made here.

[0082] Step 150 : determining a target auxiliary material parameter combination according to the target cutting yield function, the preset division conditions, and each polyhedron region.

[0083] Specifically, the second data set is adjusted through the parameter adjustment model to obtain the target cutting yield function, and the target auxiliary material parameter combination is further optimized based on the fine-tuned target cutting yield function, the various polyhedral areas obtained by spatial division, and the preset division conditions, so as to achieve optimization or prediction of multiple auxiliary material parameter combinations, thereby ensuring cutting quality.

[0084] In some embodiments, the method for determining the target auxiliary material parameter combination includes: when the preset division condition is the first preset division condition, determining the target auxiliary material parameter combination based on the center points and vertices of each polyhedral area and the target cutting yield function; when the preset division condition is the second preset division condition, determining the target auxiliary material parameter combination based on the target cutting yield function and the center points of each polyhedral area using a Bayesian optimization algorithm.

[0085] Specifically, when the data dimension of the first data set is small and the segmentation granularity is coarse, the target auxiliary material parameter combination is determined based on the center points and vertices of each polyhedral region and the target cutting yield function. When the data dimension of the first data set is high or the segmentation granularity is fine, the target auxiliary material parameter combination is determined using a Bayesian optimization algorithm based on the target cutting yield function and the center points of each polyhedral region. Thus, by selecting a suitable auxiliary material parameter combination optimization method based on the actual dimension and / or segmentation granularity of the auxiliary material parameter data during the cutting process to determine the optimal auxiliary material parameter combination, it is possible to optimize multiple auxiliary material parameter combinations to ensure cutting quality while expanding the application scenarios of the parameter adjustment method, improving the applicability of the method, and further improving the accuracy of the auxiliary material parameter combination optimization.

[0086] In some embodiments, a target auxiliary material parameter combination is determined based on the center points and vertices of each polyhedral region and the target cutting yield function, including: determining the average yield of each polyhedral region based on the center points and vertices of each polyhedral region and the target cutting yield function; and determining the target auxiliary material parameter combination based on the average yield of each polyhedral region.

[0087] Specifically, the first data set, i.e., the data space of various auxiliary material parameters during the cutting process, is divided into multiple polyhedral regions. Each polyhedral region includes a center point and multiple vertices. The auxiliary material parameter data is distributed at the center point and / or vertices of each polyhedral region. Thus, based on the center point and vertices of each polyhedral region and the target cutting yield function, the average yield of each polyhedral region can be determined.

[0088] The calculation formula for the average yield of each polyhedron area is:

[0089] ;

[0090] in, For the Average yield of polyhedral regions; For the The set of vertices and center points of a polyhedral region; The target cutting yield function at any point The value at .

[0091] The specific implementation method for determining the target auxiliary material parameter combination based on the average yield of each polyhedral region is to sort the average yields of each polyhedral region by value (from largest to smallest or from smallest to largest), and select the auxiliary material parameter combinations corresponding to the top N polyhedral regions with the largest average yield values ​​as the target auxiliary material parameter combination. The value of N can be 10 or other values, and can be set according to actual conditions, and is not specifically limited here.

[0092] In some embodiments, a Bayesian optimization algorithm is used to determine a target auxiliary material parameter combination based on a target cutting yield function and the center points of each polyhedral region, including: constructing a yield prediction proxy model based on the target cutting yield function and the center points of each polyhedral region; determining an acquisition function based on the yield prediction proxy model; optimizing the acquisition function to determine the center point corresponding to the polyhedral region with the largest expected improvement in the current iteration cycle; returning to and repeatedly executing the operation of updating the yield prediction proxy model based on the center point corresponding to the polyhedral region with the largest expected improvement in the current iteration cycle, until the target auxiliary material parameter combination is determined when a preset iteration stop condition is met.

[0093] Specifically, according to the target cutting yield function and the center points of each polyhedron area, the specific implementation process of using the Bayesian optimization algorithm to determine the target auxiliary material parameter combination includes the following steps:

[0094] Step 1: Define the objective function. The objective function is the target cutting yield function that needs to be optimized, which is used to describe the performance of the center point of each polyhedron area. Bayesian optimization samples the yield data of the currently known polyhedron area. , constructing a yield prediction proxy model. The currently known polyhedron regions refer to those sampled from all polyhedron regions obtained by spatial partitioning, for example, 5,000 polyhedron regions are sampled. The specific number of samples can be set based on actual conditions and is not specifically limited here.

[0095] Step 2: Build a yield prediction agent model. The specific construction process is: Use the Gaussian process (GP) regression model to fit the center point of the polyhedron area and yield The Gaussian process is a nonparametric model that can estimate the output of unknown points and provide an uncertainty measure (i.e., variance) of the output.

[0096] Among them, the Gaussian process regression model is:

[0097] ;

[0098] in, is the mean function, which means that for the input predictions; is the covariance function (kernel function), which measures the similarity between sample points.

[0099] Step 3: Determine the acquisition function. Select expected improvement (EI) as the acquisition function. Expected improvement measures the yield improvement that selecting a new point may bring.

[0100] The formula of the acquisition function is:

[0101] ;

[0102] in, The yield predicted by the yield prediction agent model; The currently known best yield is the best yield calculated from all polyhedral regions sampled in step 1.

[0103] in, is the expected improvement value, which means at point The expected improvement at . The goal of the expected improvement is to select those solutions that may improve the current optimal solution. area.

[0104] Step 4: Optimize the acquisition function , find the center point of the polyhedron area that maximizes the expected improvement :

[0105] ;

[0106] Among them, the center point of the polyhedron area that maximizes the expected improvement It is predicted by the yield prediction agent model and can maximize the current best yield point.

[0107] Step 5: Calculate the actual yield. After selecting the center point of the polyhedron area that maximizes the expected improvement Afterwards, the actual yield of the corresponding polyhedron area is calculated Then, this real yield data Add it to the yield data of the currently known polyhedron area in step 1 to update the yield prediction agent model.

[0108] Step 6: Update the yield prediction agent model using new sample points Updates a Gaussian process regression model so that it can more accurately predict points that have not yet been evaluated.

[0109] Step 7: Repeat the iteration. This means returning to and repeatedly updating the yield prediction proxy model until the target auxiliary material parameter combination is determined when the preset iteration stopping condition is met. The preset iteration stopping condition generally refers to reaching the maximum number of iterations or converging to the optimal solution.

[0110] It is understandable that the embodiment of the present application provides an auxiliary material parameter optimization method, including: obtaining auxiliary material parameter data during the cutting process and using it as a first data set; determining a second data set based on the first data set and a preset fitting algorithm; determining a target cutting yield function based on the second data set and a pre-built parameter adjustment model; spatially dividing the first data set according to a preset partitioning condition to obtain multiple polyhedral regions; and determining a target auxiliary material parameter combination based on the target cutting yield function, the preset partitioning condition, and each polyhedral region. The auxiliary material parameter optimization method provided in the present application optimizes and adjusts various auxiliary material parameters that affect the cutting quality to determine the optimal parameter combination, thereby optimizing or predicting a variety of cutting auxiliary material parameter combinations, thereby ensuring the cutting quality. In addition, the method can also select a suitable auxiliary material parameter combination optimization method to determine the optimal auxiliary material parameter combination based on the actual dimension and / or segmentation granularity of the auxiliary material parameter data during the cutting process. While optimizing a variety of auxiliary material parameter combinations to ensure cutting quality, it can also expand the application scenarios of the parameter adjustment method, improve the applicability of the method, and further improve the accuracy of the auxiliary material parameter combination optimization.

[0111] Figure 2 This is a schematic diagram of the overall process of the auxiliary material parameter optimization method provided in the embodiment of this application. For example, taking the auxiliary material parameter as the key data of the cutting fluid as an example, please refer to Figure 2The overall process of this auxiliary material parameter optimization method includes: first, collecting key cutting fluid data, namely the first dataset. Then, multi-segment function fitting is performed on the key cutting fluid data to generate a new variable dataset, namely the second dataset. Secondly, Lasso regression is used to fine-tune the new variable dataset. The first dataset is spatially divided into multiple polyhedral regions, and the average yield of each polyhedral region is calculated. Finally, the parameter combination with the highest average yield is selected as the target auxiliary material parameter combination. It can be seen that this method has the following advantages: First, Lasso regression is used to eliminate auxiliary material parameters with little impact on yield, significantly improving the applicability and accuracy of the model. Second, by combining mathematical modeling with experimental data, the optimal parameter combination in production can be accurately predicted, thereby improving production efficiency and product quality. Third, the method is simple and easy to implement and can be integrated into production management systems for real-time optimization. Fourth, the method can scientifically analyze the complex relationship between multiple auxiliary material parameters and cutting yield, quickly predict the optimal auxiliary material parameter combination, and verify its production applicability, showing significant theoretical value and practical application prospects. This method can be widely used in fields such as cutting fluid ratio optimization and production process parameter adjustment.

[0112] Figure 3 This is a three-dimensional rendering of the optimized parameters of the cutting fluid provided in the embodiments of the present application. For example, taking the three auxiliary material parameters of the cutting fluid (ionic liquid addition ratio, replacement ratio, and stock solution addition ratio) as examples, these three auxiliary material parameters are optimized according to the auxiliary material parameter optimization method provided in the embodiments of the present application. The specific optimization process is as follows: using a multi-segment linear function fitting to obtain the influence function of the ionic liquid addition ratio on the cutting yield, the influence function of the replacement ratio on the cutting yield, and the influence function of the stock solution addition ratio on the cutting yield. Then, based on the influence function of the ionic liquid addition ratio on the cutting yield, the cutting yield value corresponding to the ionic liquid addition ratio is calculated; based on the influence function of the replacement ratio on the cutting yield, the cutting yield value corresponding to the replacement ratio is calculated; and based on the influence function of the stock solution addition ratio on the cutting yield, the cutting yield value corresponding to the stock solution addition ratio is calculated. The cutting yield values ​​corresponding to the ionic liquid addition ratio, the cutting yield values ​​corresponding to the replacement ratio, and the cutting yield values ​​corresponding to the stock solution addition ratio are then used as a new variable data set, i.e., the second data set. Next, the second data set is fine-tuned using a parameter adjustment model, such as a Lasso regression model, to obtain a target cutting yield function. And the ionic liquid addition ratio, replacement ratio, and stock solution addition ratio are spatially divided according to the preset division conditions to obtain multiple polyhedral regions. Assuming that the data dimensions of the ionic liquid addition ratio, replacement ratio, and stock solution addition ratio are small and the segmentation granularity is coarse, the average yield value of each polyhedral region is calculated, and the average yield value of each polyhedral region is sorted in descending order. The auxiliary material parameter combinations corresponding to the top 10 polyhedral regions with the largest average yield value are selected as the target auxiliary material parameter combination, and the three-dimensional effect diagram of the optimal parameter combination is obtained as shown below. Figure 3 shown. Figure 3 Each circular point in the figure represents the average yield value of each polyhedron area. The smaller the diameter of the circular point, the larger the average yield value of the corresponding polyhedron area, indicating that the cutting yield of the corresponding auxiliary material parameters is higher.

[0113] See accordingly Figure 4 , Figure 4 Schematic diagram of the structure of the auxiliary material parameter optimization device of the embodiment of the present application. The auxiliary material parameter optimization device 100 provided in the embodiment of the present application includes: an acquisition module 101, which is used to acquire auxiliary material parameter data during the cutting process and use it as a first data set; a first determination module 102, which is used to determine a second data set based on the first data set and a preset fitting algorithm; a second determination module 103, which is used to determine a target cutting yield function based on the second data set and a pre-built parameter adjustment model; a spatial division module 104, which is used to spatially divide the first data set according to preset division conditions to obtain multiple polyhedral regions; a third determination module 105, which is used to determine a target auxiliary material parameter combination based on the target cutting yield function, the preset division conditions and each polyhedral region.

[0114] It is understood that the auxiliary material parameter optimization device provided in the embodiment of the present application optimizes and adjusts various auxiliary material parameters that affect cutting quality to determine the optimal parameter combination, thereby optimizing or predicting multiple auxiliary material parameter combinations for cutting, thereby ensuring cutting quality. In addition, the device can also select an appropriate auxiliary material parameter combination optimization method based on the actual dimensions and / or segmentation granularity of the auxiliary material parameter data during the cutting process to determine the optimal auxiliary material parameter combination. While optimizing multiple auxiliary material parameter combinations to ensure cutting quality, it can also expand the application scenarios of the parameter adjustment method, improve the applicability of the method, and further improve the accuracy of the auxiliary material parameter combination optimization.

[0115] In some embodiments, the preset partitioning condition includes a first preset partitioning condition and a second preset partitioning condition; wherein, the first preset partitioning condition is: the data dimension of the first data set is smaller than the preset dimension and the data segmentation granularity is larger than the preset segmentation granularity; the second preset partitioning condition is: the data dimension of the first data set is greater than or equal to the preset dimension, or, the data segmentation granularity of the first data set is smaller than the preset segmentation granularity.

[0116] In some embodiments, the third determination module 105 also includes: a first determination unit, which is used to determine the target auxiliary material parameter combination according to the center points and vertices of each of the polyhedral regions and the target cutting yield function when the preset division condition is the first preset division condition; a second determination unit, which is used to determine the target auxiliary material parameter combination using a Bayesian optimization algorithm according to the target cutting yield function and the center points of each of the polyhedral regions when the preset division condition is the second preset division condition.

[0117] In some embodiments, the first determination unit is further used to: determine the average yield of each polyhedral region based on the center point and vertex of each polyhedral region and the target cutting yield function; determine the target auxiliary material parameter combination based on the average yield of each polyhedral region.

[0118] In some embodiments, the second determination unit is also used to: construct a yield prediction proxy model based on the target cutting yield function and the center points of each polyhedron area; determine the acquisition function based on the yield prediction proxy model; optimize the acquisition function to determine the center point corresponding to the polyhedron area with the largest expected improvement in the current iteration cycle; return and repeat the operation of updating the yield prediction proxy model based on the center point corresponding to the polyhedron area with the largest expected improvement in the current iteration cycle until the target auxiliary material parameter combination is determined when the preset iteration stop condition is met.

[0119] In some embodiments, the first determination module 102 is further used to: fit the first data set according to a preset fitting algorithm to determine the influence function of each auxiliary material parameter on the cutting yield; determine the cutting yield value corresponding to each auxiliary material parameter according to each influence function; and determine the second data set according to the cutting yield value corresponding to each auxiliary material parameter.

[0120] In some embodiments, the second determination module 103 is further configured to: input the second data set into a parameter adjustment model for adjustment, and determine a target cutting yield function through regularization constraints; the parameter adjustment model is a Lasso regression model.

[0121] See accordingly Figure 5 , Figure 5 This is a schematic diagram of the structure of an electronic device according to an embodiment of the present application. An electronic device provided in an embodiment of the present application includes a memory 501 and a processor 502. Memory 501 is used to store computer programs. Processor 502 is used to execute the computer program stored in memory 501. When the computer program stored in memory 501 is executed, processor 502 executes the auxiliary material parameter optimization method according to the aforementioned embodiment of the present application.

[0122] Correspondingly, an embodiment of the present application further provides a computer-readable storage medium, which stores computer instructions, and the computer instructions are used to enable a processor to implement the auxiliary material parameter optimization method as described in the aforementioned embodiment of the present application when executed.

[0123] The above is a detailed introduction to the auxiliary material parameter optimization method, device, equipment and storage medium provided in the embodiments of the present application. Specific examples are used in this article to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the technical solution and core idea of ​​the present application; ordinary technicians in this field should understand that they can still modify the technical solutions recorded in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein; and these modifications or replacements do not cause the essence of the corresponding technical solution to deviate from the scope of the technical solution of the embodiments of the present application.

Claims

1. A method for optimizing auxiliary material parameters, characterized in that: Applied to adjusting auxiliary material parameters of cutting equipment, the method includes: Acquire auxiliary material parameter data during the cutting process and use it as the first data set; determining a second data set according to the first data set and a preset fitting algorithm; determining a target cutting yield function based on the second data set and a pre-built parameter adjustment model; Performing spatial division on the first data set according to a preset division condition to obtain a plurality of polyhedral regions; Determine a target auxiliary material parameter combination according to the target cutting yield function, the preset division condition, and each of the polyhedron regions; the preset division condition includes a first preset division condition and a second preset division condition; and the method for determining the target auxiliary material parameter combination includes: When the preset division condition is the first preset division condition, determining the target auxiliary material parameter combination according to the center point and the vertex of each of the polyhedral regions and the target cutting yield function; When the preset division condition is the second preset division condition, the target auxiliary material parameter combination is determined using a Bayesian optimization algorithm according to the target cutting yield function and the center points of each of the polyhedral regions.

2. The auxiliary material parameter optimization method according to claim 1, characterized in that, The first preset division condition is: the data dimension of the first data set is smaller than the preset dimension and the data segmentation granularity is larger than the preset segmentation granularity; The second preset division condition is: the data dimension of the first data set is greater than or equal to the preset dimension, or the data segmentation granularity of the first data set is smaller than the preset segmentation granularity.

3. The auxiliary material parameter optimization method according to claim 2, characterized in that, Determining the target auxiliary material parameter combination according to the center points and vertices of each polyhedral region and the target cutting yield function includes: Determining an average yield of each polyhedral region according to the center point and the vertex of each polyhedral region and the target cutting yield function; The target auxiliary material parameter combination is determined according to the average yield of each polyhedral region.

4. The auxiliary material parameter optimization method according to claim 2, characterized in that, Determining the target auxiliary material parameter combination using a Bayesian optimization algorithm based on the target cutting yield function and the center points of each polyhedral region includes: Constructing a yield prediction agent model according to the target cutting yield function and the center points of each polyhedral region; determining an acquisition function according to the yield prediction agent model; Optimizing the acquisition function to determine the center point corresponding to the polyhedron region with the maximum expected improvement in the current iteration cycle; According to the center point corresponding to the polyhedron region with the largest expected improvement in the current iteration cycle, the operation of updating the yield prediction agent model is returned and repeatedly executed until the target auxiliary material parameter combination is determined when a preset iteration stop condition is met.

5. The auxiliary material parameter optimization method according to claim 1, characterized in that, The method for determining the second data set comprises: Fitting the first data set according to the preset fitting algorithm to determine the influence function of each auxiliary material parameter on the cutting yield; Determine the cutting yield value corresponding to each auxiliary material parameter according to each influence function; The second data set is determined according to the cutting yield values ​​corresponding to the various auxiliary material parameters.

6. The auxiliary material parameter optimization method according to claim 1, characterized in that, Methods for determining the target cutting yield function include: The second data set is input into the parameter adjustment model for adjustment, and the target cutting yield function is determined through regularization constraints; the parameter adjustment model is a Lasso regression model.

7. A device for optimizing auxiliary material parameters, characterized in that: include: An acquisition module is used to acquire auxiliary material parameter data during the cutting process and use it as a first data set; A first determining module, configured to determine a second data set based on the first data set and a preset fitting algorithm; a second determining module, configured to determine a target cutting yield function according to the second data set and a pre-built parameter adjustment model; a spatial division module, configured to spatially divide the first data set according to a preset division condition to obtain a plurality of polyhedral regions; the preset division condition includes a first preset division condition and a second preset division condition; A third determination module is configured to determine a target auxiliary material parameter combination according to the target cutting yield function, the preset division condition, and each of the polyhedral regions; The third determining module is further configured to: when the preset division condition is the first preset division condition, determine the target auxiliary material parameter combination according to the center points and vertices of each of the polyhedral regions and the target cutting yield function; When the preset division condition is the second preset division condition, the target auxiliary material parameter combination is determined using a Bayesian optimization algorithm according to the target cutting yield function and the center points of each of the polyhedral regions.

8. An electronic device, characterized in that: The electronic device includes a processor and a memory; a computer program is stored in the memory, and the processor is used to execute the computer program stored in the memory to implement the auxiliary material parameter optimization method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the auxiliary material parameter optimization method according to any one of claims 1 to 6 when executed.

Images