A prediction method and prediction system for the topography height distribution parameters in grinding machining

Abstract

The present application discloses a method and a system for predicting the morphological height distribution parameters in grinding processing. This method obtains the grinding processing parameters of a target workpiece; inputs the grinding processing parameters into a morphological height distribution parameter prediction model to obtain a prediction result output by the morphological height distribution parameter prediction model. The morphological height distribution parameter prediction model is trained based on the grinding processing parameters and morphological height distribution parameters of a basic workpiece; the basic workpiece and the target workpiece are workpieces of the same process parameter type. It can perform predictions by constructing a multi-parameter prediction model, comprehensively reflect the microscopic load-bearing characteristics and height characteristics of the functional surface, so as to achieve high-precision and cross-scale correlation prediction between the grinding processing parameters and the morphological height distribution parameters, accurately predict the morphological height distribution parameters, improve the accuracy and reliability of the prediction results, and further improve the quality of grinding processing.

Description

Technical Field

[0001] The present application relates to the technical field of grinding process manufacturing, and in particular, to a method and system for predicting the morphological height distribution parameters of grinding machining. Background Art

[0002] As a key process in precision manufacturing, grinding machining removes materials and forms a high-precision surface through the interaction between abrasive grains and the workpiece, which can improve its surface integrity. Its processing quality directly affects the service performance of components. In the grinding machining process, surface topography is a key indicator for evaluating surface integrity and directly affects key service characteristics such as friction, wear, lubrication, contact stiffness, and fatigue resistance of components. Among them, in the quantitative characterization of surface topography, the height distribution feature plays a dominant role.

[0003] At present, the three-dimensional roughness parameters adopted in the process manufacturing industry are difficult to comprehensively reflect the microscopic load-bearing characteristics of functional surfaces, and are only limited to single-variable prediction, making it difficult to meet the requirements of multi-parameter collaborative optimization in actual industrial scenarios. As a result, the existing evaluation system cannot accurately predict the performance of components under complex working conditions, thereby reducing the comprehensive reliability of the prediction results.

[0004] Therefore, the present application proposes a method and system for predicting the morphological height distribution parameters of grinding machining. Compared with the prior art, it can perform predictions through multiple parameters, achieving comprehensive and high-precision prediction of the height characteristics of surface topography, solving problems such as insufficient coverage of the prediction parameters for the height information of surface topography and correlation interference of the prediction parameters in traditional methods, and can realize multi-processing parameter collaborative optimization, significantly improving the quality control and optimization ability of the grinding process, and then obtaining ideal morphological height distribution parameters, which can not only meet the grinding machining requirements under complex working conditions, but also further improve the grinding machining efficiency and quality. Summary of the Invention

[0005] The following is an overview of the subject matter described in detail herein. This overview is not intended to limit the scope of protection of the claims.

[0006] The main purpose of the embodiments of the present disclosure is to propose a method and system for predicting the morphological height distribution parameters of grinding machining, which can accurately predict the morphological height distribution parameters through high-precision and cross-scale correlation prediction of grinding machining parameters and morphological height distribution parameters, and then improve the grinding machining quality.

[0007] The first aspect of the embodiments of the present application provides a method for predicting the morphological height distribution parameters of grinding machining, which is used for a central controller. The method includes:

[0008] Obtaining the grinding machining parameters of a target workpiece;

[0009] Input the grinding process parameters into the prediction model of the topography height distribution parameters to obtain the prediction result output by the prediction model of the topography height distribution parameters. The prediction model of the topography height distribution parameters is trained based on the grinding process parameters and topography height distribution parameters of the base workpiece; the base workpiece and the target workpiece are workpieces of the same process parameter type.

[0010] Among them, the training process of the prediction model of the topography height distribution parameters includes;

[0011] Obtain a model training set, which includes the grinding process parameters and topography parameters of the base component;

[0012] Calculate the corresponding topography height distribution parameters according to the topography parameters. The topography height distribution parameters at least include the root mean square height, skewness, and kurtosis;

[0013] Preset multiple initial prediction models of the topography height distribution parameters;

[0014] Train multiple initial prediction models of the topography height distribution parameters according to the model training set to obtain multiple root mean square height prediction models, skewness prediction models, and kurtosis models;

[0015] Select the optimal root mean square height prediction model, optimal skewness prediction model, and optimal kurtosis model from the multiple root mean square height prediction models, skewness prediction models, and kurtosis models based on the entropy weight method and the technique for order preference by similarity to ideal solution;

[0016] Use the genetic algorithm to obtain the prediction model of the topography height distribution parameters according to the optimal root mean square height prediction model, the optimal skewness prediction model, and the optimal kurtosis model.

[0017] In some embodiments of the present application, the calculation formula for calculating the corresponding topography height distribution parameters according to the topography parameters includes:

[0018] ;

[0019] ;

[0020] ;

[0021] Among them, is the root mean square height, is the skewness, is the kurtosis, x , y are the spatial coordinates within the integration region, P ( x , y ) is the topography height matrix, μ is the mean value of the topography height matrix,dx , dy is the integration variable x and y is the infinitesimal change amount of A RE is P ( x , y ) the area of the integrated region.

[0022] In some embodiments of the present application, selecting the optimal root mean square height prediction model, the optimal skewness prediction model, and the optimal kurtosis model from the multiple root mean square height prediction models, skewness prediction models, and kurtosis models based on the entropy weight method and the Technique for Order Preference by Similarity to an Ideal Solution (TOPSIS) includes:

[0023] Inputting the grinding process parameters of the base workpiece into multiple root mean square height prediction models to obtain a first prediction result set output by the corresponding root mean square height prediction models;

[0024] Inputting the grinding process parameters of the base workpiece into multiple skewness prediction models to obtain a second prediction result set output by the corresponding skewness prediction models;

[0025] Inputting the grinding process parameters of the base workpiece into multiple kurtosis prediction models to obtain a third prediction result set output by the corresponding kurtosis prediction models;

[0026] Calculating the first performance parameters of the corresponding prediction models according to the first prediction result set, the second prediction result set, the third prediction result set, and the topography height distribution parameters of the base workpiece;

[0027] Using the entropy weight method and the TOPSIS, selecting the optimal root mean square height prediction model, the optimal skewness prediction model, and the optimal kurtosis model from multiple root mean square height prediction models, multiple skewness prediction models, and multiple kurtosis models according to each of the first performance parameters.

[0028] In some embodiments of the present application, using the genetic algorithm to obtain the topography height distribution parameter prediction model according to the optimal root mean square height prediction model, the optimal skewness prediction model, and the optimal kurtosis model includes:

[0029] Using the genetic algorithm to integrate the optimal root mean square height prediction model, the optimal skewness prediction model, and the optimal kurtosis model to obtain multiple initial topography height distribution parameter prediction models, and inversely solving the weight coefficients of the optimal root mean square height prediction model, the optimal skewness prediction model, and the optimal kurtosis model;

[0030] Determine the predicted model of the height distribution of the surface topography from multiple initial predicted models of the height distribution parameters of the surface topography according to the weight coefficients of the optimal root mean square height prediction model, the optimal skewness prediction model, and the optimal kurtosis model.

[0031] In some embodiments of the present application, the step of determining the predicted model of the height distribution of the surface topography from multiple initial predicted models of the height distribution parameters of the surface topography according to the weight coefficients of the optimal root mean square height prediction model, the optimal skewness prediction model, and the optimal kurtosis model includes:

[0032] Extract the objective function from multiple initial predicted models of the height distribution parameters of the surface topography;

[0033] Calculate the second performance parameters of multiple initial predicted models of the height distribution parameters of the surface topography according to the weight coefficients of the optimal root mean square height prediction model, the optimal skewness prediction model, and the optimal kurtosis model;

[0034] Determine the predicted model of the height distribution of the surface topography from multiple initial predicted models of the height distribution parameters of the surface topography according to the second performance parameters.

[0035] In some embodiments of the present application, the first performance parameters include the mean absolute difference, the mean square error, and the mean square error of the residuals, and the calculation formulas of the first performance parameters include:

[0036] ;

[0037] ;

[0038] ;

[0039] Wherein, is the number of samples in the model training set, and are respectively the actual measurement value and the model prediction value of the th sample, is the residual, is the average value of the residuals, is the mean absolute difference, is the mean square error, is the mean square error of the residuals.

[0040] In some embodiments of the present application, before obtaining the model training set, the method further includes:

[0041] Obtain the initial grinding process parameters of the base workpiece;

[0042] Perform polynomial feature transformation on the initial grinding process parameters to obtain the grinding process parameters of the base workpiece;

[0043] The performing polynomial feature transformation on the grinding process parameters includes:

[0044] ;

[0045] Among them, b =( b 1, b 2, … , b n ), b 1 to b n are the initial grinding process parameters of the base workpiece, is the preset order, are the transformed grinding process parameters of the base workpiece.

[0046] To achieve the above object, a second aspect of the embodiments of the present invention provides a prediction system for the profile height distribution parameters of grinding processing, and the system includes:

[0047] An acquisition module, configured to acquire the grinding process parameters of the target workpiece;

[0048] A prediction module, configured to input the grinding process parameters into the profile height distribution parameter prediction model to obtain a prediction result output by the profile height distribution parameter prediction model, and the profile height distribution parameter prediction model is trained according to the grinding process parameters and profile height distribution parameters of the base workpiece; the base workpiece and the target workpiece are workpieces of the same process parameter type;

[0049] Among them, the training process of the profile height distribution parameter prediction model includes;

[0050] Obtain a model training set, and the model training set includes the grinding process parameters and profile parameters of the base component;

[0051] Calculate the corresponding profile height distribution parameters according to the profile parameters, and the profile height distribution parameters at least include root mean square height, skewness, and kurtosis;

[0052] Preset a plurality of initial profile height distribution parameter prediction models;

[0053] Train a plurality of the initial profile height distribution parameter prediction models according to the model training set to obtain a plurality of root mean square height prediction models, skewness prediction models, and kurtosis models;

[0054] Select the optimal root mean square height prediction model, the optimal skewness prediction model, and the optimal kurtosis model from the multiple root mean square height prediction models, skewness prediction models, and kurtosis models based on the entropy weight method and the Technique for Order Preference by Similarity to an Ideal Solution (TOPSIS).

[0055] Use a genetic algorithm to obtain the predicted model of the surface height distribution parameters based on the optimal root mean square height prediction model, the optimal skewness prediction model, and the optimal kurtosis model.

[0056] To achieve the above object, a third aspect of the embodiments of the present invention provides an electronic device, including: at least one control processor and a memory communicatively connected to the at least one control processor; the memory stores instructions executable by the at least one control processor, and when the instructions are executed by the at least one control processor, the at least one control processor is enabled to execute the above-described method for predicting the surface height distribution parameters of a grinding process.

[0057] To achieve the above object, a fourth aspect of the embodiments of the present invention provides a computer-readable storage medium storing computer-executable instructions for causing a computer to execute the above-described method for predicting the surface height distribution parameters of a grinding process.

[0058] An embodiment of the present application provides a method for predicting the surface height distribution parameters of a grinding process. By obtaining the grinding process parameters of a target workpiece; inputting the grinding process parameters into a predicted model of the surface height distribution parameters to obtain a prediction result output by the predicted model of the surface height distribution parameters, where the predicted model of the surface height distribution parameters is trained based on the grinding process parameters and the surface height distribution parameters of a base workpiece; the base workpiece and the target workpiece are workpieces of the same process parameter type, and it is possible to perform predictions by constructing a multi-parameter prediction model, comprehensively reflecting the microscopic load-bearing characteristics and height characteristics of the functional surface, so as to achieve high-precision, cross-scale correlation prediction between the grinding process parameters and the surface height distribution parameters, accurately predict the surface height distribution parameters, improve the accuracy and reliability of the prediction results, and further improve the quality of the grinding process.

[0059] It can be understood that the beneficial effects of the above second aspect to the fourth aspect compared with the related art are the same as those of the above first aspect compared with the related art. For details, reference can be made to the relevant descriptions in the above first aspect, and details will not be repeated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] The above and / or additional aspects and advantages of the present application will become apparent and be readily understood from the following description of the embodiments in conjunction with the accompanying drawings, where:

[0061] Figure 1It is a schematic flowchart of a method for predicting the topography height distribution parameters of a grinding process provided by an embodiment of the present application;

[0062] Figure 2 It is a schematic structural diagram of a training system for predicting the topography height distribution parameters of a grinding process provided by an embodiment of the present application;

[0063] Figure 3 It is a schematic hardware structure diagram of an electronic device provided by an embodiment of the present application. Detailed implementation manners

[0064] The embodiments of the present application will be described in detail below. The examples of the embodiments are shown in the accompanying drawings, where the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present application, and should not be construed as a limitation to the present application.

[0065] In the description of the present application, if the first, second, etc. are described, it is only for the purpose of distinguishing technical features and should not be construed as indicating or implying relative importance or implicitly indicating the quantity of the indicated technical features or implicitly indicating the sequence of the indicated technical features.

[0066] In the description of the present application, it should be understood that the orientation or positional relationship indicated by terms such as up and down, etc., is based on the orientation or positional relationship shown in the accompanying drawings, and is only for the convenience of describing the present application and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus should not be construed as a limitation to the present application.

[0067] In the description of the present application, it should be noted that unless otherwise clearly defined, terms such as setting, installation, connection, etc. should be understood in a broad sense, and those skilled in the art can reasonably determine the specific meanings of the above terms in the present application in combination with the specific content of the technical solution.

[0068] Finishing processes such as grinding are often used as the last processing step for components, which can improve their surface integrity. Among them, surface topography is a key indicator for evaluating surface integrity, and it affects the service performance of components such as friction, wear, lubrication, contact, and fatigue resistance. In the quantitative characterization of surface topography, the height distribution characteristics play a dominant role. Among the 26 topography parameters defined by the ISO 25178 standard, 20 parameters are directly related to height, highlighting its core position. Among them, with the help of the Johnson transformation system, a height matrix with specified height distribution parameters (root mean square height Sq, skewness Ssk, kurtosis Sku) can be generated, and then a connection can be established with the remaining 17 roughness parameters. Among them, the three-dimensional roughness parameter Sa (the two-dimensional form is Ra), as a general evaluation index in the industry, can characterize the macroscopic undulation of the surface, but cannot fully reflect the microscopic load-bearing characteristics of the functional surface. Therefore, establishing a cross-scale correlation model between grinding process parameters and surface topography height distribution parameters can achieve the directional optimization of surface functional topography, which is the core way to break through the bottlenecks of the fatigue life and tribological performance of key components of high-end equipment.

[0069] The integrated learning model has higher robustness and prediction accuracy than a single machine model. The existing integrated learning model framework based on the genetic algorithm has the following deficiencies: (1) In the data preprocessing stage, the existing research has not constructed a multi-index weighted evaluation system, resulting in the feature space arbitrarily falling into a local optimum; (2) The existing research is limited to single-variable prediction, and using the genetic algorithm in the case of multiple optimization indexes will cause the indexes with larger dimensions to be optimized first. Therefore, there is an urgent need to construct an integrated learning model framework for multi-variable prediction under multiple indexes.

[0070] Therefore, the three-dimensional roughness parameters currently used in the process manufacturing industry are difficult to fully reflect the microscopic load-bearing characteristics of the functional surface, and are only limited to single-variable prediction, which is difficult to meet the needs of multi-parameter collaborative optimization in the actual industrial scenario, resulting in the existing evaluation system being unable to accurately predict the performance of components under complex working conditions, thereby reducing the comprehensive reliability of the prediction results.

[0071] Based on this, the embodiments of the present application provide a method and a system for predicting the topography height distribution parameters of grinding processing, aiming to accurately predict the topography height distribution parameters through high-precision, cross-scale correlation prediction of grinding processing parameters and topography height distribution parameters, and then improve the grinding processing quality.

[0072] The method and system for predicting the topography height distribution parameters of grinding processing provided by the embodiments of the present application are specifically described through the following embodiments. First, the method for predicting the topography height distribution parameters of grinding processing in the embodiments of the present application is described.

[0073] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Among them, Artificial Intelligence (AI) is a theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use knowledge to obtain the best results.

[0074] Artificial intelligence basic technologies generally include technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, and mechatronics. Artificial intelligence software technologies mainly include several major directions such as computer vision technology, robotics, biometric technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0075] The method for predicting the topography height distribution parameters of grinding processing provided by the embodiments of the present application relates to the technical field of grinding process manufacturing. The method for predicting the topography height distribution parameters of grinding processing provided by the embodiments of the present application can be applied to a terminal, or to a server side, or can also be software running on a terminal or a server side. In some embodiments, the terminal can be a smart phone, a tablet computer, a laptop computer, a desktop computer, etc.; the server side can be configured as an independent physical server, or can be configured as a server cluster or a distributed system composed of multiple physical servers, or can also be configured as a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms; the software can be an application for implementing the method for predicting the topography height distribution parameters of grinding processing, etc., but is not limited to the above forms.

[0076] The present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet-type devices, multi-processor systems, microprocessor-based systems, set-top boxes, programmable consumer electronic devices, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, and so on. The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application can also be practiced in a distributed computing environment where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0077] It should be noted that in each specific embodiment of the present application, when it comes to performing relevant processing based on data related to the user's identity or characteristics, such as user information, user behavior data, user historical data, and user location information, the user's permission or consent will be obtained first. Moreover, the collection, use, and processing of these data will comply with relevant laws, regulations, and standards. In addition, when the embodiments of the present application need to obtain the user's sensitive personal information, the user's separate permission or separate consent will be obtained through methods such as pop-up windows or redirecting to a confirmation page. After clearly obtaining the user's separate permission or separate consent, the necessary user-related data for the normal operation of the embodiments of the present application will be obtained.

[0078] For this reason, referring to Figure 1 , an embodiment of the present application provides a method for predicting the morphological height distribution parameters of grinding machining. This method is applied to a central controller, and the controller can be a server, an electronic device, a mobile terminal, etc., which is not specifically limited here. The method includes the following steps S110 to S120.

[0079] Step S110: Obtain the grinding machining parameters of the target workpiece.

[0080] In this step, the grinding machining parameters refer to the key process variables that affect the surface morphology during the machining process. In some embodiments, a white light interferometer can be used to measure the existing grinding surface morphology, such as the grinding wheel speed, feed rate, grinding depth, etc. These parameters enhance the model's ability to capture non-linear relationships through polynomial feature transformation, and specifically, second-order or third-order polynomial expansion can be used to achieve this.

[0081] Step S120: Input the grinding machining parameters into the morphological height distribution parameter prediction model to obtain the prediction result output by the morphological height distribution parameter prediction model. The morphological height distribution parameter prediction model is trained based on the grinding machining parameters and morphological height distribution parameters of the base workpiece; the base workpiece and the target workpiece are workpieces of the same process parameter type.

[0082] In this step, the grinding machining parameters of the target workpiece are obtained; the parameters are input into the morphological height distribution parameter prediction model to obtain the prediction result. The model is obtained through training with the machining parameters and morphological parameters of the base workpiece; the training process includes obtaining the model training set, calculating the morphological height distribution parameters, presetting multiple initial models, training to obtain a set of prediction models for root mean square height, skewness, and kurtosis, screening the optimal sub-model based on the entropy weight method and the technique for order preference by similarity to ideal solution (TOPSIS), and integrating using a genetic algorithm to obtain the final prediction model, realizing the collaborative prediction of multi-dimensional parameters of the grinding machining surface morphology and overcoming the defect of limited prediction dimensions of a single model.

[0083] Specifically, the topography height distribution parameter prediction model is a composite model integrated by multiple basic prediction models, which is specifically trained according to the grinding parameters and topography height distribution parameters of the basic workpiece, where the basic workpiece and the target workpiece are workpieces of the same process parameter type.

[0084] The following explains the specific training steps of the topography height distribution parameter prediction model:

[0085] Step S210: Obtain a model training set, which includes the grinding parameters and topography parameters of the basic components;

[0086] Step S220: Calculate the corresponding topography height distribution parameters according to the topography parameters. The topography height distribution parameters at least include the root mean square height, skewness, and kurtosis;

[0087] Step S230: Preset multiple initial topography height distribution parameter prediction models;

[0088] Step S240: Train multiple initial topography height distribution parameter prediction models according to the model training set to obtain multiple root mean square height prediction models, skewness prediction models, and kurtosis models;

[0089] Step S250: Select the optimal root mean square height prediction model, optimal skewness prediction model, and optimal kurtosis model from multiple root mean square height prediction models, skewness prediction models, and kurtosis models based on the entropy weight method and the technique for order preference by similarity to an ideal solution (TOPSIS);

[0090] Step S260: Use the genetic algorithm to obtain the topography height distribution parameter prediction model according to the optimal root mean square height prediction model, optimal skewness prediction model, and optimal kurtosis model.

[0091] In this step, a training set is constructed by collecting the actual processing data of the basic workpiece, and the grinding parameters are input into prediction models with different structures after feature expansion. Each sub-model is specifically trained for specific topography parameters to form independent prediction capabilities for the root mean square height, skewness, and kurtosis. In the model optimization stage, multi-dimensional performance indicators of each candidate model on the validation set are calculated. The entropy weight method is used to determine the objective weights of each indicator, and the TOPSIS is combined to calculate the closeness of each scheme to the ideal solution, and the optimal sub-models under each parameter category are selected. Finally, the three optimal sub-models are used as the base models, and their combined weights are optimized by the genetic algorithm to form a joint prediction model that can synchronously output multiple parameters, realizing a technical route of step-by-step modeling and intelligent integration. On the basis of ensuring the independent prediction accuracy of each parameter, a dynamic correlation model between parameters is established, providing comprehensive data support for process parameter optimization and effectively improving the reliability of component performance prediction under complex working conditions.

[0092] Among them, the prediction model of surface topography height distribution parameters is a composite model integrated by multiple basic prediction models. The root mean square height characterizes the degree of surface fluctuation, the skewness reflects the symmetry of height distribution, and the kurtosis indicates the sharpness of distribution. The three together constitute a complete description of surface topography. The combined application of the entropy weight method and the technique for order preference by similarity to an ideal solution refers to determining the objective weights of indicators by calculating the entropy values of model prediction results and conducting multi-objective decision-making in combination with the distance to the ideal solution. When specifically implemented, an evaluation index system including prediction accuracy, stability, and generalization ability can be constructed. The genetic algorithm integration process includes operations such as chromosome coding, fitness function design, selection, crossover, and mutation, and determines the optimal combination weights of each sub-model through iterative optimization.

[0093] In one embodiment, a white light interferometer is used to measure the existing topography parameters, and after measurement, the corresponding height distribution parameters (root mean square height Sq , skewness Ssk , kurtosis Sku ) are calculated. Among them, the calculation formulas for calculating the corresponding topography height distribution parameters according to the topography parameters include:

[0094] ;

[0095] ;

[0096] ;

[0097] Among them, is the root mean square height, is the skewness, is the kurtosis, x , y are the spatial coordinates within the integration region, P ( x , y ) is the topography height matrix, μ is the mean value of the topography height matrix, dx , dy are the small change amounts of the integration variables x and y , which are used for differential summation of the region, A RE is the defined region, which can be regarded as P ( x , y ) the area of the integrated region.

[0098] In this step, before obtaining the model training set, it also includes:

[0099] Obtaining the initial grinding process parameters of the basic workpiece;

[0100] Perform polynomial feature transformation on the initial grinding process parameters to obtain the grinding process parameters of the base workpiece;

[0101] Performing polynomial feature transformation on the grinding process parameters includes:

[0102] ;

[0103] Among them, b =( b 1, b 2, … , b n ), b 1 to b n are the initial grinding process parameters of the base workpiece, is the preset order, is the grinding process parameter of the base workpiece after transformation.

[0104] In this step, perform polynomial feature transformation on the grinding process parameters to obtain the grinding process parameters of the base workpiece, where the polynomial feature transformation is through the formula:

[0105] ;

[0106] Among them, b =( b 1, b 2, … , b n ), b 1 to b n are the initial grinding process parameters of the base workpiece, is the preset order, is the grinding process parameter of the base workpiece after transformation.

[0107] Specifically, the initial grinding process parameters refer to the original input variables that directly affect the surface topography in the grinding process, such as numerical parameters like the grinding wheel linear speed, feed rate, grinding depth, etc., which can be collected by sensors or obtained from process parameter tables. Polynomial feature transformation refers to expanding the original parameters into a high-order feature vector containing quadratic terms and interaction terms through mathematical methods. For example, expanding the single parameter into , and other combined terms to enhance the model's ability to capture non-linear relationships. The preset order refers to the highest degree of polynomial expansion, which can be set to 2nd order or 3rd order, and preferably the final value is determined by balancing the computational complexity and model accuracy.

[0108] In some embodiments, the preset order is selected through an optimal order selection strategy. Among them, the optimal order selection strategy is applied to multivariate prediction, which can more comprehensively consider the problem of optimal order selection in polynomial feature transformation. Specifically, as follows:

[0109] First, normalize the obtained initial dataset, where the initial dataset includes grinding process parameters (grinding wheel speed, depth of cut, feed rate) and surface topography height distribution parameters ( Sq , Ssk , Sku ), and the formula is as follows:

[0110] (1);

[0111] Apply formula (1) successively to normalize the grinding process parameters (grinding wheel speed, depth of cut, feed rate) and surface topography height distribution parameters ( Sq , Ssk , Sku ) in the initial dataset.

[0112] Taking the grinding wheel speed as an example, is the number of samples, represents the grinding wheel speed value of the -th sample, is the minimum grinding wheel speed value among all samples, is the maximum grinding wheel speed value among all samples.

[0113] Furthermore, after randomly shuffling all the normalized data three times, specify the training set and test set ratios as 89% and 11%, and then divide the data into training sets and test sets of 80% and 20% in the way of five-fold cross-validation, that is, divide the data five times in total. Among them, the training set is used to train 8 machine learning models, and the test set is used to evaluate 8 machine learning models.

[0114] Furthermore, transform the independent variables (i.e., grinding process parameters) of all the divided data into new features according to polynomial features of order 1 to 5. Among them, the polynomial feature transformation is defined as:

[0115] ;

[0116] In the formula, and represent the features to be transformed, specifically the grinding wheel speed, depth of cut and feed rate. is a positive integer and is the preset order.

[0117] For example, when , ,

[0118] ;

[0119] The input features have changed from 3 features to 10 features.

[0120] Furthermore, roughness parameters to be predicted are sequentially selected Sq , Ssk , Sku , and a large number of data sets are formed through orthogonality in the above steps. Each data set trains 8 machine learning models in sequence, and each model is in the form of multivariate prediction of a single variable. Then, the MAE (mean absolute difference), RMSE (mean square error), STDR (mean square error of residuals) of all the trained machine learning models are calculated. Among them, the data set for error calculation is the pre-divided test set.

[0121] Specifically, the calculation formulas for each error are as follows.

[0122] ;

[0123] ;

[0124] ;

[0125] In the formula, M is the number of samples, and are respectively the measured value and the predicted value of the th sample, is the residual, is the average value of the residuals, mean absolute difference, is the mean square error, is the mean square error of the residuals.

[0126] Furthermore, the MAE , RMSE , STDR values calculated after prediction according to the test set by integrating each machine learning model trained from all the above data sets are used as the targets, that is, it is expected that these values are as small as possible, so as to calculate the weights of these three parameters using the entropy weight method and calculate the scores of all models using the comprehensive evaluation method (Technique for Order Preference by Similarity to an Ideal Solution, TOPSIS).

[0127] Furthermore, according to the roughness parameters ( Sq , Ssk , Sku), separate all machine learning models with polynomial feature transformation orders (1st to 5th order), that is, there are 15 classifications in total. Calculate the corresponding TOPSIS average score for each classification to target each parameter ( Sq , Ssk , Sku) respectively select the polynomial feature transformation order corresponding to the highest score, so as to expand the original feature space through polynomial feature transformation (1st to 5th order), enhance the model's ability to capture non-linear relationships, and moreover, the influence of dimension can be eliminated through data normalization, the robustness of evaluation can be ensured through multiple data partitions and five-fold cross-validation, the weights of evaluation parameters can be objectively assigned through the entropy weight method, the performance of the model can be comprehensively scored through the comprehensive evaluation method and the contributions of different error indicators can be balanced, so as to select the optimal order based on the score and ensure the balance between prediction accuracy and model complexity.

[0128] In this step, the transformed high-dimensional features are used as the input of the model training set to jointly construct a non-linear mapping relationship with the topography height distribution parameters, so that the subsequent established prediction model can more accurately characterize the complex correlation between grinding parameters and surface topography, thereby improving the adaptability of the model to complex process condition changes and reducing the prediction deviation caused by missing feature information.

[0129] In some embodiments, the specific training of the topography height distribution parameter prediction model includes: obtaining an initial model training set, and performing normalization, polynomial feature transformation and data partition processing on the initial model training set to obtain a model training set for model training.

[0130] Specifically, construct 8 machine learning mapping models for grinding process parameters (after polynomial feature transformation) and topography matrix height distribution parameters (root mean square height Sq , skewness Ssk , kurtosis Sku ), including Random Forest (RF), Extreme Gradient Boosting (Xgboost), Support Vector Regression (SVR), Least Absolute Shrinkage and Selection Operator (LASSO), Ridge Regression (RR), Elastic Net Regression (ENR), Extreme Gradient Boosting based on Newton-Raphson Optimizer (NRBO-Xgboost), Particle Swarm Optimization-based BP Neural Network (PSO-BP). Among them, 3 variables ( Sq , Ssk , Sku ) are predicted separately, that is, 24 machine learning models are established, and the hyperparameters in each model are determined by five-fold cross-validation to prevent model underfitting or overfitting.

[0131] Further, calculate the topography height distribution parameters predicted by 24 machine learning models Sq , Ssk ,Sku The error term, preferably calculated for all machine learning models MAE (Mean Absolute Deviation), RMSE (Mean Squared Error), STDR (Mean Squared Error of Residuals). Among them, the calculation formulas for each error are as follows:

[0132] ;

[0133] ;

[0134] ;

[0135] In the formula, is the number of samples, and are respectively the measured value and the predicted value of the th sample, is the residual, is the average value of the residuals.

[0136] Furthermore, by integrating the MAE , RMSE , STDR values of all machine learning models, according to the metrics MAE (Mean Absolute Deviation), RMSE (Mean Squared Error), STDR (Mean Squared Error of Residuals), combined with the entropy weight method and the TOPSIS method, the optimal 4 machine learning models for predicting each topography height distribution parameter are determined, that is, 12 machine learning models are selected.

[0137] Furthermore, based on the regularization term, the optimal coefficients ~ corresponding to the 12 machine learning models are inversely solved using the non-dominated genetic algorithm (NSGA-Ⅲ), thereby effectively reducing the "over-contribution" of a single machine learning model, ensuring the collaborative robustness of multiple models, and suppressing model overfitting and preventing over-sensitivity to noise data. Among them, the design of the objective function under multiple metrics is included.

[0138] Specifically, the regularized ensemble learning model is embedded as the target model in the NSGA-Ⅲ algorithm, and the output results are as follows each time:

[0139] ;

[0140] In the formula, ~ are the optimal coefficients to be solved, is the optimal machine learning model for the current topography height distribution parameter.

[0141] Furthermore, according to the metricsMAE (Mean Absolute Deviation), RMSE (Mean Squared Error), STDR (Mean Squared Error of Residuals), the following objective function is established:

[0142] ;

[0143] In the formula, the first three terms of each objective ~ respectively represent the MAE , RMSE , STDR cumulative errors corresponding to the predicted height distribution parameter values of the ensemble learning model, and the last term is the regularization error term of the optimal coefficients ~ to be solved, and the regularization parameter β takes 0.01.

[0144] Furthermore, the entropy weight method is used to calculate the weights of the objectives ~ , where the objectives ~ are to be minimized, and then the comprehensive evaluation method is used to select the optimal solution from the Pareto solution set.

[0145] Furthermore, the initial model training set is randomly divided 5 times, and the above training process is repeated for each group of data, and then the averages of all results are calculated to obtain the optimal machine learning model for predicting the topography height distribution parameters.

[0146] In some embodiments, in step S250, the optimal root mean square height prediction model, the optimal skewness prediction model, and the optimal kurtosis model are selected from multiple root mean square height prediction models, skewness prediction models, and kurtosis models based on the entropy weight method and the Technique for Order of Preference by Similarity to Ideal Solution, including the following steps S310 to S350:

[0147] Step S310: Input the grinding process parameters of the base workpiece into multiple root mean square height prediction models to obtain a first prediction result set output by the corresponding root mean square height prediction models;

[0148] Step S320: Input the grinding process parameters of the base workpiece into multiple skewness prediction models to obtain a second prediction result set output by the corresponding skewness prediction models;

[0149] Step S330: Input the grinding process parameters of the base workpiece into multiple kurtosis prediction models to obtain a third prediction result set output by the corresponding kurtosis prediction models;

[0150] Step S340: Calculate the first performance parameter of the corresponding prediction model according to the first prediction result set, the second prediction result set, the third prediction result set, and the topography height distribution parameters of the base workpiece;

[0151] Step S350: Use the entropy weight method and the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) method to select the optimal root mean square height prediction model, the optimal skewness prediction model, and the optimal kurtosis model from multiple root mean square height prediction models, multiple skewness prediction models, and multiple kurtosis models according to the first performance parameters.

[0152] In this embodiment, the grinding process parameters of the base workpiece are input into multiple root mean square height prediction models to obtain the first prediction result set, input into multiple skewness prediction models to obtain the second prediction result set, and input into multiple kurtosis prediction models to obtain the third prediction result set, so as to calculate the performance parameters of the corresponding prediction models according to the prediction result sets and the topography height distribution parameters of the base workpiece, and then use the entropy weight method and the TOPSIS method to select the optimal models in each model category according to the performance parameters.

[0153] Among them, the entropy weight method is a mathematical method for calculating the objective weights of each performance index based on the information entropy theory. It is preferably implemented by calculating the information entropy value of the index and normalizing it to obtain the weight, which is used to eliminate the deviation of subjective weighting; the TOPSIS method is a multi-attribute decision-making method for ranking by calculating the Euclidean distances between the alternatives and the ideal optimal solution and the worst solution. It is preferably implemented by constructing the positive and negative ideal solution matrices and calculating the relative closeness degree, which is used to quantify the comprehensive performance of the model; the performance parameters include the mean absolute difference, the mean square error, and the mean square error of the residuals. It is preferably implemented by using a preset formula to calculate the deviation index between the actual measured value and the predicted value, which is used to objectively evaluate the prediction accuracy of the model.

[0154] Specifically, by inputting the grinding process parameters of the base workpiece into multiple root mean square height, skewness, and kurtosis prediction models respectively, the prediction result sets of each model for the same workpiece can be obtained. Then, based on the actually measured topography height distribution parameters, the performance parameters of each prediction model on the training set are calculated using the mean absolute difference, the mean square error, and the mean square error of the residuals formulas.

[0155] Among them, the first performance parameters include the mean absolute difference, the mean square error, and the mean square error of the residuals. The calculation formulas for the first performance parameters include:

[0156] ;

[0157] ;

[0158] ;

[0159] Among them, is the number of samples in the model training set, and are respectively the actual measured value and the model predicted value of the th sample, is the residual, is the average value of the residuals, the mean absolute difference, is the mean square error, is the mean square error of the residuals.

[0160] Furthermore, the entropy weight method is used to assign weights to each performance parameter, and the closeness degree between each model and the ideal solution is calculated by the technique for order preference by similarity to an ideal solution (TOPSIS) method. Finally, the model with the highest closeness degree is selected as the optimal model for the corresponding category. Thus, the entropy weight method is used to objectively determine the multi-index weights, and the TOPSIS method is combined to comprehensively evaluate the closeness degree between the model and the ideal solution, which can more comprehensively reflect the comprehensive performance of the model in terms of prediction accuracy, stability and deviation control, realizes the quantitative comparison of the prediction capabilities of multiple models, avoids the limitations of a single evaluation index, and provides an objective and quantifiable model screening basis for the collaborative prediction of multi-parameters of the grinding surface topography.

[0161] In some embodiments, in step S260, the genetic algorithm is used to obtain the topography height distribution parameter prediction model according to the optimal root mean square height prediction model, the optimal skewness prediction model and the optimal kurtosis model, including the following steps S410 to S420:

[0162] Step S410: Use the genetic algorithm to integrate the optimal root mean square height prediction model, the optimal skewness prediction model and the optimal kurtosis model to obtain multiple initial topography height distribution parameter prediction models, and inversely solve the weight coefficients of the optimal root mean square height prediction model, the optimal skewness prediction model and the optimal kurtosis model;

[0163] Step S420: Determine the topography height distribution parameter prediction model from multiple initial topography height distribution parameter prediction models according to the weight coefficients of the optimal root mean square height prediction model, the optimal skewness prediction model and the optimal kurtosis model.

[0164] In this embodiment, the genetic algorithm is used to integrate the optimal root mean square height prediction model, the optimal skewness prediction model and the optimal kurtosis model to obtain multiple initial topography height distribution parameter prediction models, and inversely solve the weight coefficients of the optimal root mean square height prediction model, the optimal skewness prediction model and the optimal kurtosis model. Furthermore, according to the weight coefficients of the optimal root mean square height prediction model, the optimal skewness prediction model and the optimal kurtosis model, the topography height distribution parameter prediction model is determined from multiple initial topography height distribution parameter prediction models.

[0165] Among them, the genetic algorithm refers to an optimization algorithm that simulates the process of natural selection, preferably implemented by chromosome coding, fitness function design, and crossover and mutation operations, and the optimal weight combination is selected through iteration; integration refers to the fusion of the prediction results of multiple independent models, preferably implemented by linear superposition or non-linear combination, and the overall performance is improved by integrating the advantages of different models; the weight coefficient refers to the contribution ratio of each model in the integration, preferably solved by error backpropagation or an optimization algorithm, and is used to quantify the influence degree of different sub-models on the final prediction result; the inverse solution of the weight coefficient refers to the reverse derivation of the optimal weight of each model according to the goal of minimizing the prediction error, preferably calculated by gradient descent or genetic algorithm iteration to ensure that the comprehensive performance of the integrated model reaches the optimal.

[0166] Specifically, in the genetic algorithm, first, chromosome coding is performed on the three optimal sub-models of root mean square height, skewness, and kurtosis. For example, the weight coefficients of each model are encoded as gene sequences, and then multiple groups of weight combinations are generated through crossover and mutation operations to form different initial morphology height distribution parameter prediction models. Among them, each group of weight combinations corresponds to an integrated model, and its prediction result is obtained by weighting the outputs of each sub-model. Furthermore, through the fitness function, based on the prediction error of the validation set, for example, the mean square error is used as the evaluation index, the weight combination with the smallest error is selected. Thus, after multiple generations of iteration, the individual with the highest fitness is retained as the final weight coefficient to determine the optimal integrated model structure, so as to realize the adaptive optimization of weights through the genetic algorithm, be able to dynamically adjust the contribution ratio of each sub-model according to the specific data distribution, and effectively improve the generalization ability of the integrated model under complex working conditions.

[0167] In some embodiments, in step S350, selecting the optimal root mean square height prediction model, the optimal skewness prediction model, and the optimal kurtosis model from multiple root mean square height prediction models, multiple skewness prediction models, and multiple kurtosis models according to each performance parameter includes the following steps S510 to step S530:

[0168] Step S510: Extract the objective function from multiple initial morphology height distribution parameter prediction models;

[0169] Step S520: Calculate the second performance parameter of multiple initial morphology height distribution parameter prediction models according to the weight coefficients of the optimal root mean square height prediction model, the optimal skewness prediction model, and the optimal kurtosis model;

[0170] Step S530: Determine the morphology height distribution parameter prediction model from multiple initial morphology height distribution parameter prediction models according to the second performance parameter.

[0171] In this embodiment, after extracting the objective function from multiple initial topography height distribution parameter prediction models, according to the weight coefficients of the optimal root mean square height prediction model, the optimal skewness prediction model, and the optimal kurtosis model, calculate the second performance parameter of multiple initial topography height distribution parameter prediction models, and then determine the topography height distribution parameter prediction model from multiple initial topography height distribution parameter prediction models according to the second performance parameter.

[0172] Among them, the objective function refers to a mathematical expression used to evaluate the comprehensive performance of the prediction model, preferably constructed using indicators such as mean square error or mean absolute error. Through the objective function, the deviation degree between the prediction result and the actual measurement value can be quantified; the weight coefficient refers to the contribution ratio of each sub-model during the integration process, preferably obtained by inverse solution using a genetic algorithm, and is used to reflect the influence weight of different sub-models on the final prediction result; the second performance parameter refers to the comprehensive evaluation index of the integrated model, preferably using a multi-objective optimization method to combine the weight coefficient with the error index of each sub-model, so as to screen out the optimal integrated model.

[0173] Specifically, when determining the topography height distribution parameter prediction model, it is first necessary to extract a preset objective function from the candidate initial models, for example, minimizing the prediction error or maximizing the stability as the objective.

[0174] Furthermore, use the optimal weight coefficients obtained by inverse solution using the genetic algorithm to perform weighted calculation on the error index of each initial model to obtain the second performance parameter. Then, by comparing the second performance parameters of different initial models, select the model with the best comprehensive performance as the final topography height distribution parameter prediction model. Thus, by integrating multiple optimal sub-models through the weight coefficient and using the second performance parameter for multi-objective optimization, the correlation between different parameters can be effectively balanced, making the prediction result more in line with the actual processing conditions, realizing the advantages of integrating multiple sub-models, improving the robustness and applicability of the prediction model, and meeting the requirements for comprehensive evaluation of multi-parameters of the surface topography in complex industrial scenarios.

[0175] In some embodiments, test the prediction effect of the topography height distribution parameter prediction model in actual application. Grind a flat part (8×8×20 mm ) made of high-strength gear material 9310, and the size of the machined surface is 8×8 mm. Among them, the tool is a 120-mesh CBN grinding wheel with a diameter of 100 mm and a width of 10 mm, equipped with a coolant (Castrol Syntilo 2000) with a flow rate of 120 l / min.

[0176] Furthermore, use a white light interferometer Wyko NT910 to measure the surface topography of the 9310 flat ground part. Among them, the grinding process parameters of the numerical control machine tool are shown in Table 1.

[0177] Table 1

[0178]

[0179] An orthogonality set is formed by the processing parameters in Table 1, and the surface topography is measured twice under each set of processing parameters. The size of the topography matrix is 599×599, and the spacing is 1.986 , that is, there are 96 groups of original topography data, and then the height distribution parameters corresponding to each topography are calculated Sq 、 Ssk 、 Sku 。

[0180] Furthermore, polynomial feature transformation of order 1 - 5 is performed on the grinding processing parameters (wheel speed, depth of cut, feed rate) in sequence. The TOPSIS comprehensive scores under each order transformation are calculated using the optimal order, and the calculation results are shown in Table 2

[0181] Table 2

[0182]

[0183] It can be seen from Table 2 that the optimal polynomial transformation orders of the topography height distribution parameters Sq 、 Ssk 、 Sku are 1, 2, and 2 orders respectively. Among them, the value of the parameter Sku is relatively Sa 、 Ssk large, because all results are integrated and scored through the optimal order selection strategy, which leads to a relatively low TOPSIS score corresponding to the parameter Sku

[0184] Furthermore, the data set is randomly divided 5 times. By inputting each group of data into 8 preset machine learning models (RF, XGboost, SVM, LASSO, RR, ENR, NRBO - XGboost, and PSO - BP) and the topography height distribution parameter prediction model, the corresponding evaluation index results are calculated. Among them, the evaluation index results of the 8 machine learning models are shown in Table 3

[0185] Table 3

[0186]

[0187] Among them, the evaluation index results of the topography height distribution parameter prediction model are shown in Table 4

[0188] Table 4

[0189]

[0190] ​Comparing Table 3 with Table 4, among the 8 commonly used machine learning models, the height distribution parameters of the surface topography height distribution parameter prediction model Sq , Ssk , Sku have the smallest errors under the evaluation metrics MSE , RMSE , STDR , achieving more accurate prediction of height distribution parameters.

[0191] As Figure 2 shown, some embodiments of the present application provide a system for predicting the surface topography height distribution parameters of grinding machining. The system includes an acquisition module 210 and a prediction module 220. Specifically:

[0192] The acquisition module 210 is configured to acquire the grinding machining parameters of the target workpiece.

[0193] The prediction module 220 is configured to input the grinding machining parameters into the surface topography height distribution parameter prediction model to obtain the prediction result output by the surface topography height distribution parameter prediction model. The surface topography height distribution parameter prediction model is trained based on the grinding machining parameters and surface topography height distribution parameters of the base workpiece; the base workpiece and the target workpiece are workpieces of the same process parameter type.

[0194] In some embodiments, the prediction module 220 may include:

[0195] ;

[0196] ;

[0197] ;

[0198] wherein, is the root mean square height, is the skewness, is the kurtosis, x , y are the spatial coordinates within the integration region, P ( x , y ) is the surface topography height matrix, μ is the mean of the surface topography height matrix, dx , dy are the infinitesimal changes of the integration variables x and y , A RE is P ( x , y ) the area of the integration region.

[0199] In some embodiments, the prediction module 220 may include: inputting the grinding process parameters of the base workpiece into a plurality of root mean square height prediction models to obtain a first prediction result set output by the corresponding root mean square height prediction models.

[0200] In some embodiments, the prediction module 220 may include: inputting the grinding process parameters of the base workpiece into a plurality of skewness prediction models to obtain a second prediction result set output by the corresponding skewness prediction models.

[0201] In some embodiments, the prediction module 220 may include: inputting the grinding process parameters of the base workpiece into a plurality of kurtosis prediction models to obtain a third prediction result set output by the corresponding kurtosis prediction models.

[0202] In some embodiments, the prediction module 220 may include: calculating a first performance parameter of the corresponding prediction models according to the first prediction result set, the second prediction result set, the third prediction result set, and the profile height distribution parameters of the base workpiece.

[0203] In some embodiments, the prediction module 220 may include: using the entropy weight method and the technique for order preference by similarity to ideal solution (TOPSIS) to select the optimal root mean square height prediction model, the optimal skewness prediction model, and the optimal kurtosis model from a plurality of root mean square height prediction models, a plurality of skewness prediction models, and a plurality of kurtosis models according to the respective first performance parameters.

[0204] In some embodiments, the prediction module 220 may include: using a genetic algorithm to integrate the optimal root mean square height prediction model, the optimal skewness prediction model, and the optimal kurtosis model to obtain a plurality of initial profile height distribution parameter prediction models, and inversely solving the weight coefficients of the optimal root mean square height prediction model, the optimal skewness prediction model, and the optimal kurtosis model.

[0205] In some embodiments, the prediction module 220 may include: determining a profile height distribution parameter prediction model from a plurality of initial profile height distribution parameter prediction models according to the weight coefficients of the optimal root mean square height prediction model, the optimal skewness prediction model, and the optimal kurtosis model.

[0206] In some embodiments, the prediction module 220 may include: extracting an objective function from a plurality of initial profile height distribution parameter prediction models.

[0207] In some embodiments, the prediction module 220 may include: calculating a second performance parameter of a plurality of initial profile height distribution parameter prediction models according to the weight coefficients of the optimal root mean square height prediction model, the optimal skewness prediction model, and the optimal kurtosis model.

[0208] In some embodiments, the prediction module 220 may include: determining a predicted model of the topography height distribution parameter from a plurality of initial topography height distribution parameter prediction models according to the second performance parameter.

[0209] In some embodiments, the prediction module 220 may include:

[0210] ;

[0211] ;

[0212] ;

[0213] Wherein, is the number of samples in the model training set, and are respectively the actual measurement value and the model prediction value of the th sample, is the residual, is the average value of the residuals, the mean absolute difference, is the mean square error, is the mean square error of the residuals.

[0214] In some embodiments, the prediction module 220 may include: obtaining the initial grinding process parameters of the base workpiece.

[0215] In some embodiments, the prediction module 220 may include: performing polynomial feature transformation on the grinding process parameters to obtain the grinding process parameters of the base workpiece.

[0216] In some embodiments, the prediction module 220 may include:

[0217] ;

[0218] Wherein, b =( b 1, b 2, … , b n ), b 1 to b n are the initial grinding process parameters of the base workpiece, is the preset order, are the grinding process parameters of the base workpiece.

[0219] It should be noted that the prediction system for the profile height distribution parameters of grinding machining provided in this embodiment and the above-mentioned prediction method for the profile height distribution parameters of grinding machining are based on the same inventive concept. Therefore, the relevant content of the above-mentioned prediction method for the profile height distribution parameters of grinding machining also applies to the content of the prediction system for the profile height distribution parameters of grinding machining. Therefore, it will not be elaborated here.

[0220] For this purpose, the system obtains the grinding machining parameters of the target workpiece; inputs the grinding machining parameters into the profile height distribution parameter prediction model to obtain the prediction result output by the profile height distribution parameter prediction model. The profile height distribution parameter prediction model is trained according to the grinding machining parameters and profile height distribution parameters of the base workpiece; the base workpiece and the target workpiece are workpieces of the same process parameter type. In this way, high-precision and cross-scale correlation prediction of the grinding machining parameters and the profile height distribution parameters can be realized, accurately predicting the profile height distribution parameters, and thus improving the quality of grinding machining.

[0221] An embodiment of the present application also provides an electronic device. The electronic device includes a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, it implements the above-mentioned prediction method for the profile height distribution parameters of grinding machining.

[0222] As Figure 3 , Figure 3 is a schematic diagram of the hardware structure of the electronic device provided in an embodiment of the present application. The electronic device includes:

[0223] At least one battery;

[0224] At least one memory;

[0225] At least one processor;

[0226] At least one program;

[0227] The program is stored in the memory, and the processor executes at least one program to implement the above-mentioned prediction method for the profile height distribution parameters of grinding machining in the present disclosure.

[0228] The electronic device can be any intelligent terminal including a mobile phone, a tablet computer, a personal digital assistant (PDA), a vehicle-mounted computer, etc.

[0229] The electronic device in the embodiment of the present application will be introduced in detail below.

[0230] The processor 1600 can be implemented in the form of a general-purpose central processing unit (CPU), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, etc., and is used to execute relevant programs to implement the technical solutions provided in the embodiments of the present disclosure;

[0231] The memory 1700 can be implemented in the form of a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM), etc. The memory 1700 can store an operating system and other application programs. When implementing the technical solutions provided in the embodiments of this specification through software or firmware, the relevant program codes are stored in the memory 1700 and are called by the processor 1600 to execute a method for predicting the topography height distribution parameters of a grinding process in the embodiments of the present disclosure.

[0232] The input / output interface 1800 is used to implement information input and output;

[0233] The communication interface 1900 is used to implement communication interaction between this device and other devices, and can achieve communication through wired means (such as USB, network cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.);

[0234] The bus 2000 transmits information between various components of the device (such as the processor 1600, the memory 1700, the input / output interface 1800, and the communication interface 1900);

[0235] Among them, the processor 1600, the memory 1700, the input / output interface 1800, and the communication interface 1900 achieve communication connections with each other inside the device through the bus 2000.

[0236] The embodiments of the present disclosure also provide a storage medium, which is a computer-readable storage medium. The computer-readable storage medium stores computer-executable instructions, and the computer-executable instructions are used to cause a computer to execute the above-mentioned method for predicting the topography height distribution parameters of a grinding process.

[0237] As a non-transitory computer-readable storage medium, the memory can be used to store non-transitory software programs and non-transitory computer-executable programs. In addition, the memory may include high-speed random access memory, and may also include non-transitory memory, such as at least one magnetic disk storage device, a flash memory device, or other non-transitory solid-state storage devices. In some embodiments, the memory may optionally include a memory remotely disposed relative to the processor, and these remote memories may be connected to the processor through a network. Examples of the above networks include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0238] The embodiments described in the embodiments of the present disclosure are for more clearly illustrating the technical solutions of the embodiments of the present disclosure, and do not constitute a limitation on the technical solutions provided by the embodiments of the present disclosure. Those skilled in the art will know that with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by the embodiments of the present disclosure are equally applicable to similar technical problems.

[0239] Those skilled in the art can understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of the present disclosure, and may include more or fewer steps than those shown in the figures, or combine certain steps, or different steps.

[0240] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, that is, they may be located in one place, or may be distributed to multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0241] Those of ordinary skill in the art can understand that all or some of the steps in the methods disclosed above, and the functional modules / units in the systems and devices, can be implemented as software, firmware, hardware, and appropriate combinations thereof.

[0242] The terms "first", "second", "third", "fourth", etc. (if any) in the specification of this application and the above figures are used to distinguish similar objects, and do not have to be used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances, so that the embodiments of the present application described here can be implemented in an order other than those illustrated or described here. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units does not have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products, or devices.

[0243] It should be understood that in this application, "at least one (item)" means one or more, and "a plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that there can be three relationships. For example, "A and / or B" can mean: only A exists, only B exists, and both A and B exist at the same time. Here, A and B can be singular or plural. The character " / " generally represents an "or" relationship between the associated objects before and after. "At least one (item) of the following" or its similar expression refers to any combination of these items, including any combination of single items or plural items. For example, at least one (item) of a, b, or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, and c can be single or multiple.

[0244] In several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of units is only a logical function division. In actual implementation, there can be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed couplings or direct couplings or communication connections to each other can be through some interfaces. The indirect couplings or communication connections of devices or units can be in electrical, mechanical, or other forms.

[0245] The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they can be located in one place, or they can be distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0246] In addition, in each embodiment of this application, the functional units can be integrated into one processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit. The above integrated units can be implemented in the form of hardware or in the form of software functional units.

[0247] When an integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes multiple instructions for causing an electronic device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods in various embodiments of the present application. The aforementioned storage medium includes: various media that can store programs, such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs.

[0248] The above has specifically described the preferred implementation of the embodiments of the present application, but the embodiments of the present application are not limited to the above implementation manners. Those skilled in the art can also make various equivalent deformations or substitutions without departing from the spirit of the embodiments of the present application, and these equivalent deformations or substitutions are all included within the scope defined by the claims of the embodiments of the present application.

[0249] The above has described the embodiments of the present application in detail with reference to the accompanying drawings, but the present application is not limited to the above embodiments. Various changes can be made without departing from the purpose of the present application within the scope of knowledge possessed by those of ordinary skill in the art.

Claims

1. A method for predicting the topography height distribution parameters of grinding machining, characterized in that The method includes: Obtaining the grinding process parameters of the target workpiece; Inputting the grinding process parameters into a topography height distribution parameter prediction model to obtain a prediction result output by the topography height distribution parameter prediction model, where the topography height distribution parameter prediction model is trained according to the grinding process parameters and topography height distribution parameters of a base workpiece; the base workpiece and the target workpiece are workpieces of the same process parameter type; Among them, the training process of the topography height distribution parameter prediction model includes; Obtaining a model training set, where the model training set includes the grinding process parameters and topography parameters of a base component; Calculating the corresponding topography height distribution parameters according to the topography parameters, where the topography height distribution parameters at least include root mean square height, skewness, and kurtosis; Presetting a plurality of initial topography height distribution parameter prediction models; Training the plurality of initial topography height distribution parameter prediction models according to the model training set to obtain a plurality of root mean square height prediction models, skewness prediction models, and kurtosis models; Selecting the optimal root mean square height prediction model, optimal skewness prediction model, and optimal kurtosis model from the plurality of root mean square height prediction models, skewness prediction models, and kurtosis models based on the entropy weight method and the technique for order preference by similarity to ideal solution (TOPSIS); Using a genetic algorithm to obtain the topography height distribution parameter prediction model according to the optimal root mean square height prediction model, the optimal skewness prediction model, and the optimal kurtosis model.

2. The method for predicting the morphological height distribution parameters of grinding machining according to claim 1, wherein, The calculation formula for calculating the corresponding topography height distribution parameters according to the topography parameters includes: ； ； ； Among them, is the root mean square height, is the skewness, is the kurtosis, x , y are the spatial coordinates within the integration region, P ( x , y ) is the surface height matrix, μ is the mean value of the surface height matrix, dx , dy are the integration variables x and y are the small change amounts, A RE is P ( x , y ) is the area of the integration region.

3. The method for predicting the topography height distribution parameters of grinding machining according to claim 2, characterized in that The selecting the optimal root mean square height prediction model, optimal skewness prediction model, and optimal kurtosis model from the plurality of root mean square height prediction models, skewness prediction models, and kurtosis models based on the entropy weight method and the technique for order preference by similarity to ideal solution (TOPSIS) includes: Inputting the grinding process parameters of the base workpiece into the plurality of root mean square height prediction models to obtain a first prediction result set output by the corresponding root mean square height prediction models; Inputting the grinding process parameters of the base workpiece into the plurality of skewness prediction models to obtain a second prediction result set output by the corresponding skewness prediction models; Inputting the grinding process parameters of the base workpiece into the plurality of kurtosis prediction models to obtain a third prediction result set output by the corresponding kurtosis prediction models; Calculating the first performance parameters of the corresponding prediction models according to the first prediction result set, the second prediction result set, the third prediction result set, and the topography height distribution parameters of the base workpiece; Using the entropy weight method and the technique for order preference by similarity to ideal solution (TOPSIS), selecting the optimal root mean square height prediction model, optimal skewness prediction model, and optimal kurtosis model from the plurality of root mean square height prediction models, the plurality of skewness prediction models, and the plurality of kurtosis models according to each of the first performance parameters.

4. The method for predicting the morphological height distribution parameters of grinding machining according to claim 1, characterized in that The using a genetic algorithm to obtain the topography height distribution parameter prediction model according to the optimal root mean square height prediction model, the optimal skewness prediction model, and the optimal kurtosis model includes: Integrate the optimal root mean square height prediction model, the optimal skewness prediction model, and the optimal kurtosis model using the genetic algorithm to obtain multiple initial prediction models for the surface height distribution parameters, and inversely solve the weight coefficients of the optimal root mean square height prediction model, the optimal skewness prediction model, and the optimal kurtosis model; Determine the prediction model for the surface height distribution parameters from the multiple initial prediction models for the surface height distribution parameters according to the weight coefficients of the optimal root mean square height prediction model, the optimal skewness prediction model, and the optimal kurtosis model.

5. The method for predicting the topography height distribution parameters of grinding machining according to claim 4, characterized in that The determining the prediction model for the surface height distribution parameters from the multiple initial prediction models for the surface height distribution parameters according to the weight coefficients of the optimal root mean square height prediction model, the optimal skewness prediction model, and the optimal kurtosis model includes: Extract the objective function from the multiple initial prediction models for the surface height distribution parameters; Calculate the second performance parameters of the multiple initial prediction models for the surface height distribution parameters according to the weight coefficients of the optimal root mean square height prediction model, the optimal skewness prediction model, and the optimal kurtosis model; Determine the prediction model for the surface height distribution parameters from the multiple initial prediction models for the surface height distribution parameters according to the second performance parameters.

6. The method for predicting the topography height distribution parameters of grinding machining according to claim 3, characterized in that The first performance parameters include the mean absolute difference, the mean square error, and the mean square error of the residuals. The calculation formula for the first performance parameters includes: ； ； ； wherein, is the number of samples in the model training set, and are respectively the actual measured value and the model predicted value of the th sample, is the residual, is the average value of the residuals, is the mean absolute difference, is the mean square error, is the mean square error of the residuals.

7. The method for predicting the topography height distribution parameters of grinding machining according to claim 1, wherein, Before obtaining the model training set, the method further includes: Obtain the initial grinding process parameters of the base workpiece; Perform polynomial feature transformation on the initial grinding process parameters to obtain the grinding process parameters of the base workpiece; The performing polynomial feature transformation on the grinding process parameters includes: ； Among them, b =( b 1, b 2, … , b n ), b 1 to b n are the initial grinding parameters of the base workpiece, is the preset order, are the grinding parameters of the base workpiece after conversion.

8. A prediction system for the morphological height distribution parameters of grinding machining, characterized in that, The system includes: An acquisition module for acquiring the grinding process parameters of the target workpiece; A prediction module for inputting the grinding process parameters into the prediction model for the surface height distribution parameters to obtain the prediction result output by the prediction model for the surface height distribution parameters. The prediction model for the surface height distribution parameters is trained according to the grinding process parameters and the surface height distribution parameters of the base workpiece; the base workpiece and the target workpiece are workpieces of the same process parameter type; Among them, the training process of the prediction model for the surface height distribution parameters includes; Obtain a model training set, where the model training set includes the grinding process parameters and the surface parameters of the base component; Calculate the corresponding surface height distribution parameters according to the surface parameters. The surface height distribution parameters at least include the root mean square height, skewness, and kurtosis; Preset multiple initial prediction models for the surface height distribution parameters; Train the multiple initial prediction models for the surface height distribution parameters according to the model training set to obtain multiple root mean square height prediction models, skewness prediction models, and kurtosis models; Select the optimal root mean square height prediction model, the optimal skewness prediction model, and the optimal kurtosis model from the multiple root mean square height prediction models, skewness prediction models, and kurtosis models based on the entropy weight method and the technique for order preference by similarity to ideal solution; Using a genetic algorithm, the predicted model of the topography height distribution parameters is obtained according to the optimal root mean square height prediction model, the optimal skewness prediction model, and the optimal kurtosis model.

9. An electronic device, characterized in that, It includes at least one control processor and a memory for communicatingly connecting with the at least one control processor; the memory stores instructions executable by the at least one control processor, and the instructions are executed by the at least one control processor so that the at least one control processor can execute a method for predicting the topography height distribution parameters of a grinding process according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions for causing a computer to execute a method for predicting the topography height distribution parameters of a grinding process according to any one of claims 1 to 7.

Images