CN118839580A - Production method of mold for cold stamping 6G antenna parts - Google Patents

Abstract

The application provides a production method of a mold for cold stamping 6G antenna parts, which comprises the following steps: acquiring cavity surface roughness data and frequency band layout parameters of a mold for cold stamping 6G antenna parts, acquiring related data through a surface morphology measuring instrument and a spectrum analyzer, and establishing a data mapping relation between the cavity surface roughness and the frequency band layout, wherein the mapping relation is used for subsequently analyzing the performance of the mold under the 6G frequency band; according to the obtained cavity surface roughness data and frequency band layout parameters, adopting a finite element analysis method to simulate distribution characteristics of residual stress on the cavity surface under different frequency band layouts, obtaining a residual stress distribution cloud chart, determining a stress concentration area, and visually displaying stress distribution conditions by the residual stress distribution cloud chart, wherein the method is used for identifying potential high-risk areas of die abrasion.

Description

Production method of mold for cold stamping 6G antenna parts

Technical Field

The invention relates to the technical field of production of 6G antenna part dies, in particular to a production method of a die for cold stamping 6G antenna parts.

Background

There are certain errors and residual stresses in the cavity surface roughness frequency band layout of the mold for cold stamping 6G antenna parts, and these factors have an important influence on the mold wear. In actual production, the threshold value of the surface roughness of the cavity and the distribution characteristics of residual stress under different frequency band layouts are difficult to accurately control, so that the abrasion of the die is aggravated, and the service life is reduced. Meanwhile, the improper type selection of the cavity surface roughness frequency band layout can introduce additional stress concentration and accelerate the failure of the die. In addition, the frequency band layout of the surface roughness of the cavity lacks an effective on-line monitoring and feedback control mechanism, so that the frequency band layout parameters are difficult to adjust in real time, the surface quality fluctuation of the cavity is large, and the abrasion state of the die is difficult to predict. The association relation between the frequency band layout defect and the cavity surface roughness is not clear, and a quantitative model of frequency band layout design and surface quality control is difficult to establish, so that the optimization of the die design is difficult. There is a need to develop a defect mechanism study of cavity surface roughness frequency band layout, reveal frequency band layout defects and cavity surface roughness evolution rules, establish a multi-scale correlation model of frequency band layout design, surface quality prediction and mold wear evaluation, and realize full life cycle quality control and intelligent optimization of a mold.

Disclosure of Invention

The invention provides a production method of a mold for cold stamping 6G antenna parts, which mainly comprises the following steps:

Acquiring cavity surface roughness data and frequency band layout parameters of a mold for cold stamping 6G antenna parts, acquiring related data through a surface morphology measuring instrument and a spectrum analyzer, and establishing a data mapping relation between the cavity surface roughness and the frequency band layout, wherein the mapping relation is used for subsequently analyzing the performance of the mold under the 6G frequency band;

According to the obtained cavity surface roughness data and frequency band layout parameters, adopting a finite element analysis method to simulate distribution characteristics of residual stress on the cavity surface under different frequency band layouts, obtaining a residual stress distribution cloud picture, determining a stress concentration area, and visually displaying stress distribution conditions by the residual stress distribution cloud picture, wherein the residual stress distribution cloud picture is used for identifying a potential high risk area of die abrasion;

based on cavity surface roughness data, frequency band layout parameters and residual stress distribution cloud pictures, training and establishing a prediction model between the frequency band layout parameters, the cavity surface roughness and the die wear through a machine learning algorithm such as a support vector machine or a random forest, and judging the wear state and the service life of the die under the current frequency band layout and the surface roughness according to model prediction results;

In the cold stamping forming process, an online monitoring system is adopted to collect surface roughness of a cavity and frequency band layout parameters in real time, the wear state of a die is dynamically estimated through comparison and analysis with a prediction model, when the wear state exceeds a preset threshold value, a feedback control mechanism is triggered, and the frequency band layout parameters, such as the size or the interval of an antenna radiation unit, are automatically adjusted according to the output result of the prediction model so as to slow down the die wear;

Aiming at the association relation between the frequency band layout defects and the cavity surface roughness, the on-line monitoring data and the prediction model result are combined, the evolution rule of the frequency band layout defects and the surface roughness is excavated through a data excavation technology such as association rule analysis, a multi-scale association model for frequency band layout design, surface quality prediction and die wear evaluation is established, and the model realizes the cross-scale comprehensive analysis by integrating microscopic surface roughness, mesoscopic frequency band layout and macroscopic die wear data;

Integrating on-line monitoring data, a prediction model result and a multi-scale association model analysis result into a full life cycle management system of the die, realizing full cycle monitoring and early warning of the die quality through big data analysis and visualization technology, optimizing a die maintenance strategy by utilizing the multi-scale association model, predicting potential faults, and providing data support for die design optimization;

Based on integrated full life cycle quality data of the die and a multi-scale associated model, an intelligent optimization algorithm such as a genetic algorithm is adopted to optimize frequency band layout design parameters, in the optimization process, the surface quality of a cavity, the abrasion state of the die and the antenna performance requirements are considered, and the balance of the die life maximization and the antenna performance optimization is realized by adjusting the geometric parameters and the arrangement mode of antenna radiation units, so that the intellectualization and the high efficiency of the die design, the manufacture, the use and the maintenance are realized.

The invention has the beneficial effects that:

According to the invention, the data mapping relation is established by collecting the surface roughness of the cavity and the frequency band layout parameters, and the residual stress distribution is simulated by utilizing finite element analysis. Based on these data, a machine learning algorithm is used to construct a model of die wear prediction. In the cold stamping process, the state of the die is monitored in real time, and when the abrasion exceeds a threshold value, the frequency band layout parameters are automatically adjusted to slow down the abrasion. Meanwhile, a multi-scale association model is established through a data mining technology, so that cross-scale comprehensive analysis is realized. And finally, optimizing frequency band layout design parameters by utilizing an intelligent optimization algorithm on the basis of considering the surface quality of a cavity, the abrasion state of a die and the performance requirement of an antenna, and realizing the balance of the maximization of the service life of the die and the optimization of the performance of the antenna. The invention remarkably improves the efficiency of mold design, manufacture, use and maintenance, prolongs the service life of the mold and ensures the stability of the antenna performance through multidimensional data analysis and intelligent optimization.

Drawings

Fig. 1 is a schematic diagram of a method for producing a cold stamping 6G antenna component according to the present invention.

Fig. 2 is a schematic diagram II of a method for producing a cold stamping 6G antenna component according to the present invention.

Detailed Description

The technical solutions in the embodiments of the present invention will be clearly and specifically described below with reference to the drawings in the embodiments of the present invention. The described embodiments are only a few embodiments of the present invention.

As shown in fig. 1-2, the method for producing a mold for cold stamping a 6G antenna component according to this embodiment specifically includes:

Step S101, cavity surface roughness data and frequency band layout parameters of a mold for cold stamping 6G antenna parts are obtained, related data are collected through a surface morphology measuring instrument and a spectrum analyzer, and a data mapping relation between the cavity surface roughness and the frequency band layout is established. The mapping relation is used for subsequent analysis of the performance of the lower die of the 6G frequency band.

According to cavity surface roughness data and frequency band layout parameters of a mold for cold stamping 6G antenna parts, a nonlinear regression model of the cavity surface roughness and the frequency band layout is established by adopting a support vector machine algorithm, and a data mapping relation of the cavity surface roughness and the frequency band layout is obtained;

Classifying and predicting the performance of the die under the 6G frequency band by adopting a decision tree algorithm according to the established data mapping relation between the surface roughness of the die cavity and the frequency band layout, judging that the die can be used for producing 6G antenna parts if the prediction result meets a preset performance threshold value, otherwise judging that the die needs to be subjected to surface treatment; for a mold needing surface treatment, treating the surface of a cavity in a chemical corrosion or mechanical polishing mode, acquiring the surface roughness data of the cavity after treatment, inputting the surface roughness data into an established data mapping model of the surface roughness of the cavity and the frequency band layout, and predicting the corresponding frequency band layout parameters;

according to the predicted frequency band layout parameters, optimizing the structural size of the antenna parts by adopting a genetic algorithm to obtain the structural parameters of the antenna parts meeting the 6G frequency band requirements;

According to the optimized structural parameters of the antenna parts, a corresponding cold stamping die is processed through a numerical control machining center, the surface roughness of a machined die cavity is detected by utilizing a three-coordinate measuring instrument, and detection data are input into a data mapping model of the cavity surface roughness and frequency band layout for verification; if the verification result shows that the surface roughness of the die cavity and the frequency band layout parameters meet the requirements, cold stamping forming of the 6G antenna part is carried out by using the die, and a finished product part is obtained; if the verification result shows that the surface roughness of the die cavity and the frequency band layout parameters do not meet the requirements, returning to process the surface of the die cavity again until the requirements are met.

Specifically, first, cavity surface roughness data and frequency band layout parameters of a mold for cold stamping 6G antenna parts are collected, wherein the cavity surface roughness data comprises Ra and Rz parameters, and the frequency band layout parameters comprise center frequency and bandwidth. Training the collected data by using a support vector machine algorithm, establishing a nonlinear regression model of cavity surface roughness and frequency band layout, and mapping the cavity surface roughness to the frequency band layout space by using a Gaussian kernel function to obtain a corresponding relation between the cavity surface roughness and the frequency band layout space.

On the basis, the performance of the die is classified and predicted by utilizing a decision tree algorithm, the performance threshold is set to be smaller than 5% of frequency band layout deviation, the surface roughness Ra of the cavity is smaller than 8 mu m, if the prediction result meets the threshold requirement, the die can be used for producing antenna parts, and otherwise, the surface of the cavity needs to be treated.

And (3) for the die to be processed, adopting a chemical corrosion or mechanical polishing mode to reduce the surface roughness Ra of the die cavity to below 4 mu m, inputting the processed roughness data into an established mapping model, and predicting the corresponding frequency band layout parameters.

According to the prediction result, optimizing the structural size of the antenna parts by adopting a genetic algorithm, for example, taking the height of an oscillator and the position of a feed point as optimization variables, taking the coverage range of a frequency band and the performance of a directional diagram as optimization targets, and obtaining the optimal structural parameters meeting the requirement of a 6G frequency band after 50 generations of evolution.

Processing a corresponding cold stamping die through a numerical control machining center by utilizing optimized structural parameters, controlling the machining precision within +/-02 mm, detecting the surface roughness of a machined die cavity by adopting a three-coordinate measuring instrument, inputting detection data into a mapping model for verification, and if the surface roughness Ra of the die cavity is smaller than 4 mu m and the frequency band layout deviation is within 5%, performing cold stamping forming by utilizing the die, wherein the stamping pressure is 800kN, the pressure maintaining time is 10s, and obtaining an antenna part finished product with the dimensional precision meeting the requirement; if the verification result does not meet the requirement, the surface of the cavity is treated again, and after 3-5 iterations, the cold stamping die meeting the performance requirement is finally obtained, and the production quality of the antenna parts is ensured.

Step S102, simulating distribution characteristics of residual stress on the surface of the cavity under different frequency band layouts by adopting a finite element analysis method according to the acquired cavity surface roughness data and frequency band layout parameters, obtaining a residual stress distribution cloud picture, and determining a stress concentration area. The residual stress distribution cloud chart visually displays stress distribution conditions and is used for identifying potential high-risk areas of die wear.

And acquiring the roughness data of the cavity surface, establishing a three-dimensional model of the cavity surface, and mapping the roughness data to the model surface. And carrying out frequency band division on the three-dimensional model of the cavity surface according to the frequency band layout parameters to generate finite element models under different frequency band layouts. And setting material properties, boundary conditions and load working conditions for the finite element model under each frequency band layout, and carrying out finite element analysis and calculation.

And obtaining a distribution cloud picture of residual stress on the surface of the cavity under different frequency band layouts by adopting a finite element analysis result, and visually displaying the stress distribution condition. And identifying a stress concentration area through the stress distribution cloud chart, and judging whether the stress concentration degree exceeds the allowable stress of the material. If the stress concentration exceeds the allowable stress of the material, the area is determined as a potential high risk area of die wear. And (3) combining analysis results of the high-risk areas of the die abrasion under different frequency band layouts, determining an optimal frequency band layout scheme, and guiding the optimization of the surface frequency band layout of the cavity.

Specifically, firstly, a three-coordinate measuring instrument is utilized to scan the surface of a cavity to obtain surface roughness data, the scanning interval is 1mm, and a group of point cloud data containing X, Y, Z coordinates and corresponding roughness Ra values is obtained.

And then, importing the point cloud data by using reverse engineering software Imageware, generating a three-dimensional CAD model of the cavity surface through surface fitting, and mapping a roughness Ra value to the CAD model surface to generate a cavity surface roughness distribution map. According to the structural characteristics of the product, frequency band division is carried out on the three-dimensional model of the surface of the cavity, 6 frequency band areas are divided, and the frequency band areas correspond to different functional areas of the product respectively. And (3) carrying out grid division on each frequency band region by utilizing a grid division tool in CAD software to generate a finite element model of 6 frequency band regions, wherein the size of a grid unit is 1mm.

In the finite element analysis software ANSYS, a cavity material is defined as Cr12MoV die steel, the elastic modulus is 210GPa, the Poisson ratio is 3, and the yield strength is 1000MPa.

And (3) applying injection pressure of 3MPa on the cavity surface, applying ejection force of 3000N on the bottom surface of the cavity, and setting the friction factor between the cavity surface and the plastic part to be 3.

And (3) performing finite element analysis to obtain Von Mises stress distribution cloud pictures of 6 frequency band regions, wherein the stress distribution situation can be seen through the stress cloud pictures. And identifying a stress concentration region, finding that the stress is maximum at the center of the cavity, wherein the maximum stress value is 980MPa, and is close to the material yield strength of 1000MPa, and judging the region as a potential die abrasion high risk region.

The analysis results of the 6 frequency band areas are combined, and the frequency band division scheme 2 is found to have the minimum stress concentration degree, the maximum stress value is 850MPa and is lower than the yield strength of the material, so that the frequency band division scheme 2 is determined to be an optimal scheme, and the die abrasion risk can be effectively reduced.

Step S103, training and establishing a prediction model among the frequency band layout parameters, the surface roughness of the cavity and the abrasion of the die by a machine learning algorithm such as a support vector machine or a random forest based on the cavity surface roughness data, the frequency band layout parameters and the residual stress distribution cloud image, and judging the abrasion state and the service life of the die under the current frequency band layout and the surface roughness according to the model prediction result.

And acquiring cavity surface roughness data, frequency band layout parameters and residual stress distribution cloud image data as a training data set of a machine learning algorithm. Training based on a training data set by adopting a support vector machine and a random forest machine learning algorithm, and establishing a prediction model among frequency band layout parameters, cavity surface roughness and die abrasion.

And according to the established prediction model, inputting the frequency band layout parameters and the surface roughness data of the current cavity, and predicting to obtain the abrasion state of the die. If the predicted wear state exceeds a preset threshold, judging that the die has reached the service life and needs to be maintained or replaced; otherwise, judging that the die can still be used continuously. And acquiring a cavity residual stress distribution cloud chart, extracting residual stress distribution characteristics, and taking the residual stress distribution cloud chart as supplementary input of a prediction model to improve the accuracy of the prediction of the wear state of the die. By analyzing the prediction result, key factors influencing the die abrasion, such as key frequency bands in frequency band layout parameters and critical values of surface roughness, are determined, and basis is provided for optimizing die design.

And according to the prediction result of the wear state of the die and the analysis of key influencing factors, a die maintenance strategy is formulated, such as periodic detection of surface roughness, optimization of frequency band layout, extension of the service life of the die and improvement of production efficiency.

Specifically, in order to predict the wear state of the mold, first, cavity surface roughness data, frequency band layout parameters, and residual stress distribution cloud image data need to be collected. The surface roughness data can be measured by a roughness meter, the frequency band layout parameters can be extracted from a mould design drawing, and the residual stress distribution cloud picture can be obtained by simulation calculation of finite element analysis software.

After sufficient data is collected, it is randomly divided into a training set for training the machine learning model and a test set for evaluating the predictive performance of the model. In the training process, the grid search method can be used for optimizing the hyper-parameters of the model, such as the kernel function type of the support vector machine and the number of decision trees of random forests, so as to improve the prediction accuracy of the model.

By comparing the predicted result of the test set with the actual wear state, the evaluation index, such as the mean square error and the decision coefficient, of the model is calculated, and the predicted performance of the model is evaluated. When the predicted performance of the model reaches the expected target, the model can be applied to actual production.

In the application process, frequency band layout parameters and surface roughness data of a cavity are required to be acquired in real time and are input into a prediction model to obtain a predicted value of the abrasion state of the die. Meanwhile, a cavity residual stress distribution cloud image is required to be obtained, residual stress distribution characteristics such as maximum residual stress and residual stress gradient are extracted, and the residual stress distribution cloud image is used as supplementary input of a prediction model so as to improve prediction accuracy.

According to the prediction result, a corresponding die maintenance strategy can be formulated, for example, when the predicted wear state exceeds a preset threshold value, the die is maintained or replaced in time, so that the problem of product quality degradation caused by excessive wear of the die is avoided. By analyzing the prediction result, key factors influencing the die wear, such as key frequency bands in the frequency band layout parameters and critical values of surface roughness, can be determined. For example, by comparing the abrasion states of the mold under different frequency band layout parameters, it is found that when the energy density of a certain frequency band exceeds 5J/mm, the abrasion rate of the mold is significantly increased, so that the energy density of the frequency band can be controlled below 5J/mm, so as to prolong the service life of the mold.

And step S104, in the cold stamping forming process, acquiring the surface roughness of the cavity and the frequency band layout parameters in real time by adopting an online monitoring system, and dynamically evaluating the abrasion state of the die by comparing the surface roughness with a prediction model. When the abrasion state exceeds a preset threshold value, triggering a feedback control mechanism, and automatically adjusting frequency band layout parameters, such as the size or the interval of antenna radiating units, according to the output result of the prediction model so as to slow down the abrasion of the die.

And acquiring roughness data and frequency band layout parameters of the surface of the die in the cold stamping forming process, inputting the acquired data into a pre-established prediction model for comparison and analysis, and obtaining a wear state evaluation result of the die. Judging whether the evaluation result exceeds a preset threshold value, if so, triggering a feedback control mechanism, and determining frequency band layout parameters to be adjusted, such as the size or the spacing of antenna radiating units, by adopting an optimization algorithm according to the output result of the prediction model. And establishing a correlation model between the frequency band layout parameters and the die wear by a machine learning algorithm, and determining an optimal frequency band layout parameter adjustment scheme by an intelligent optimization algorithm based on the correlation model so as to minimize the die wear. And transmitting the optimized frequency band layout parameters to a numerical control system, controlling a numerical control machine tool to automatically adjust the size and the interval of the antenna radiating units, changing the frequency band layout, and slowing down the abrasion speed of the die. And monitoring the influence of the adjusted frequency band layout parameters on the die abrasion in real time, and returning to optimize the frequency band layout parameters again if the abrasion state still exceeds the threshold value until the die abrasion state is lower than the preset threshold value. And (3) establishing a plurality of prediction models and associated models in advance according to different mold materials and cold stamping process parameters, dynamically switching and selecting an optimal model according to actual production conditions, and improving the accuracy of prediction and control. And (3) adopting a big data analysis technology to perform mining analysis on the acquired mass die wear data and frequency band layout parameter data, extracting an implicit association rule and an evolution trend, and providing data support for optimization of the prediction model and the association model.

Specifically, in the cold stamping forming process, the roughness data of the surface of the die can be measured in real time by adopting a high-precision laser displacement sensor, and meanwhile, the frequency band layout parameters including the size and the spacing of the antenna radiating units are obtained by using a vector network analyzer. And inputting the acquired roughness data and frequency band layout parameters into a prediction model constructed based on a Support Vector Machine (SVM) algorithm, and obtaining a die abrasion state evaluation result through comparison and analysis. If the evaluation result exceeds a preset threshold (e.g., the surface roughness Ra is greater than 8 μm), a feedback control mechanism is triggered. And determining frequency band layout parameters to be adjusted by adopting a Particle Swarm Optimization (PSO) algorithm according to the output result of the SVM prediction model, for example, adjusting the size of the antenna radiating unit from 10mm multiplied by 10mm to 8mm multiplied by 8mm, and adjusting the spacing from 5mm to 8mm. And establishing a correlation model between the frequency band layout parameters and the die wear by a Convolutional Neural Network (CNN) algorithm, and determining an optimal frequency band layout parameter adjustment scheme by adopting a Genetic Algorithm (GA) based on the correlation model so as to minimize the die wear. And transmitting the optimized frequency band layout parameters to a numerical control system, controlling a numerical control machine tool to automatically adjust the size and the interval of the antenna radiating units, changing the frequency band layout, and slowing down the abrasion speed of the die. And monitoring the influence of the adjusted frequency band layout parameters on the die abrasion in real time, and if the abrasion state still exceeds the threshold value, re-optimizing the frequency band layout parameters until the die abrasion state is lower than the preset threshold value. And (3) establishing a plurality of prediction models and associated models in advance according to different die materials (such as high-speed steel and hard alloy) and cold stamping process parameters (such as stamping speed and pressure), dynamically switching and selecting an optimal model according to actual production conditions, and improving the accuracy of prediction and control. And adopting a Hadoop big data platform to perform mining analysis on the acquired mass die wear data and frequency band layout parameter data, extracting implicit association rules and evolution trends, and providing data support for optimization of the prediction model and the association model.

Step S105, aiming at the association relation between the frequency band layout defect and the cavity surface roughness, combining on-line monitoring data and a prediction model result, excavating a frequency band layout defect and surface roughness evolution rule through a data excavation technology such as association rule analysis, and establishing a multi-scale association model of frequency band layout design, surface quality prediction and mold abrasion evaluation. The model realizes the comprehensive analysis of the cross-scale by integrating the micro surface roughness, the mesoscopic frequency band layout and the macroscopic mould abrasion data.

And acquiring frequency band layout design parameters, cavity surface roughness measurement data and mold wear monitoring data, and constructing a multi-scale data set. Preprocessing the multi-scale data, including data cleaning, feature extraction and data integration, to obtain a standardized data representation. And adopting an association rule mining algorithm, such as Apriori or FP-growth, and finding association rules and modes between the frequency band layout defects and the surface roughness from the standardized data. And constructing a prediction model of the frequency band layout defect and the surface roughness evolution, such as a decision tree, a support vector machine or a neural network, according to the mined association rule. And applying the prediction model to on-line monitoring data to predict the evolution trend of the cavity surface quality and the frequency band layout defects in real time. And comprehensively analyzing cross-scale correlation among frequency band layout, surface roughness and die abrasion, establishing a multi-scale correlation model, and quantifying an influence mechanism among different factors. Based on the multi-scale association model, a frequency band layout design scheme is optimized, a die maintenance strategy is formulated, the surface quality of a cavity is improved, and the service life of a die is prolonged.

Specifically, firstly, frequency band layout design parameters are obtained through a high-precision three-dimensional scanner, such as the number of frequency bands is 8, the width of each frequency band is 10mm, cavity surface roughness measurement data are measured through a roughness meter, the surface roughness Ra value is 8 mu m, mold wear monitoring data are collected in real time through an online monitoring system, the mold temperature and pressure parameters are included, and a multi-scale data set is constructed. Preprocessing the multi-scale data, cleaning the data and extracting features by adopting a minimum-maximum standardization method, extracting frequency band layout parameters, surface roughness and die abrasion features, and obtaining standardized data representation through data integration. And adopting an Apriori association rule mining algorithm, setting the minimum support degree to be 05, setting the minimum confidence degree to be 8, and finding association rules between the frequency band layout defects and the surface roughness from the standardized data, such as rules with higher surface roughness as the frequency band width is larger. According to the excavated association rule, a predictive model of frequency band layout defects and surface roughness evolution is constructed by adopting a support vector machine algorithm, and the proportion of a training set to a testing set is 7: and 3, the accuracy of the model reaches 85%. The prediction model is applied to on-line monitoring data to predict the surface quality of a cavity and the evolution trend of the defects of the frequency band layout in real time, and when the surface roughness exceeds 2 mu m, the defects possibly occur in the frequency band layout are predicted. The cross-scale correlation among the frequency band layout, the surface roughness and the die wear is comprehensively analyzed, a multi-scale correlation model is established, the gray correlation between the frequency band width, the surface roughness and the die temperature is calculated by adopting a gray correlation analysis method, the influence mechanism among different factors is quantized, the influence degree of the frequency band width on the surface roughness is 7, and the influence degree of the surface roughness on the die wear is 6. Based on a multi-scale association model, a frequency band layout design scheme is optimized by adopting a genetic algorithm, the population number is set to be 50, the evolution algebra is set to be 100, the crossover probability is set to be 8, the mutation probability is set to be 1, the optimal frequency band layout scheme is obtained, the frequency band number is set to be 6, the width of each frequency band is set to be 12mm, meanwhile, a die maintenance strategy is formulated, die polishing is carried out when the surface roughness exceeds 0 mu m, and the service life of the die is prolonged by 20%.

And S106, integrating the online monitoring data, the prediction model result and the multi-scale correlation model analysis result into a full life cycle management system of the die, and realizing full cycle monitoring and early warning of the quality of the die through big data analysis and visualization technology. And optimizing a die maintenance strategy by utilizing a multi-scale correlation model, predicting potential faults, and providing data support for die design optimization.

Real-time data including temperature, pressure and speed parameters in the production process of the die are acquired through an online monitoring system, and the data are transmitted to a full life cycle management system. Inputting the obtained real-time monitoring data into a prediction model, analyzing and calculating through an algorithm to obtain a performance prediction result of the die in a future period of time, and judging whether the die has potential quality risks. If the prediction result shows that the quality risk exists, triggering an early warning mechanism, automatically generating early warning information by the system, and prompting related personnel to take countermeasures in time through visual interface display. According to the prediction result and the historical data, the relation between each parameter of the die and the performance and service life of the die is analyzed through the association model, so that key factors affecting the quality of the die are obtained, and a basis is provided for the subsequent die design optimization. And excavating and analyzing various data in the whole life cycle of the die by adopting a big data analysis technology, finding out the association rule and pattern among the data, and providing data support for optimizing the maintenance strategy of the die. The design parameters, material properties and process condition data of the mold are input into a correlation model, training is carried out through a machine learning algorithm, and a mold performance prediction model is established and used for guiding mold design optimization. Based on historical data and prediction results in the full life cycle management system, a heuristic algorithm is adopted to optimize the maintenance strategy of the die, a scientific maintenance plan is formulated, the service life of the die is prolonged to the maximum extent, and the maintenance cost is reduced.

Specifically, real-time data in the production process of the die is acquired through an online monitoring system, such as the temperature sensor acquires data once every second, the pressure sensor acquires data once every 5 seconds, and the speed sensor acquires data once every 1 second. And transmitting the acquired data to a database of the full life cycle management system through the industrial Ethernet. The system utilizes a support vector machine algorithm to analyze and calculate the real-time monitoring data, predicts the performance change trend of the die in the future 8 hours by comparing the real-time monitoring data with the historical data, and triggers an early warning mechanism if the prediction result shows that the die has the risk of exceeding 5% of the dimension in the future 2 hours. The system automatically generates early warning information, displays the early warning information in a popup window and voice prompt mode through a visual interface, and simultaneously pushes the early warning information to a mobile terminal APP of related personnel. The system adopts a correlation rule mining algorithm to analyze the correlation among the properties of the die material, design parameters, process conditions, the performance and service life of the die material, and discovers that key factors influencing the quality of the die, such as the hardness of the die material and the service life of the die material are positively correlated, and the correlation coefficient is 85. And establishing a mould performance prediction model by utilizing a decision tree algorithm, and predicting the service life and key quality characteristics of the mould by inputting design parameters, material properties and process conditions of the mould. The maintenance strategy of the die is optimized by adopting a genetic algorithm, and the optimal die maintenance scheme is obtained by setting the die maintenance period and maintenance content parameters and combining the historical data and the prediction result of the die, so that the average service life of the die can be prolonged by 20%, and the annual maintenance cost is reduced by 15%.

Step S107, optimizing frequency band layout design parameters by adopting an intelligent optimization algorithm such as a genetic algorithm based on the integrated full life cycle quality data of the die and the multi-scale correlation model. In the optimization process, the geometric parameters and arrangement modes of the antenna radiating units are adjusted by considering the surface quality of a cavity, the abrasion state of a die and the performance requirements of an antenna, so that the balance of the maximization of the service life of the die and the optimization of the performance of the antenna is realized, and the intellectualization and the high efficiency of the design, the manufacture, the use and the maintenance of the die are realized.

And acquiring quality data of the full life cycle of the die, wherein the quality data comprises cavity surface quality data and die wear state data, constructing a multi-scale association model, and establishing an association relation between the quality data and the die life and between the quality data and the antenna performance. Judging the current use state and the residual service life of the die according to the surface quality data of the cavity and the wear state data of the die, and triggering an antenna frequency band layout optimization flow if the residual service life of the die is lower than a preset threshold value. And adopting a genetic algorithm intelligent optimization algorithm, taking geometric parameters and arrangement modes of antenna radiating units as optimization variables, taking the service life of a die and the performance of an antenna as optimization targets, and carrying out multi-target optimization solution. In the optimization process, the surface quality of the cavity and the abrasion state of the die are correlated with the optimization variable through a multi-scale correlation model of the die, so that the optimization process can consider the quality factor of the die. And obtaining a group of optimal antenna layout design parameters meeting the requirements of the service life of the die and the performance of the antenna through iterative optimization, wherein the optimal antenna layout design parameters comprise the geometric parameters and the arrangement modes of the antenna radiating units. And according to the optimized antenna layout design parameters, intelligently modifying and redesigning the die, generating a new die processing program and technological parameters, and guiding the processing and manufacturing of the die. And continuously collecting quality data of the die in the use and maintenance stage of the die, evaluating the state of the die in real time, and triggering the optimization flow again when the quality of the die is reduced below a threshold value, upgrading and updating the die, so as to realize intelligent management of the full life cycle of the die.

Specifically, a sensor array is arranged on the surface of a die, surface morphology data and abrasion data of a die cavity are collected in real time, a wavelet analysis algorithm is utilized to conduct feature extraction and fusion on the collected multi-scale quality data, and a nonlinear mapping model of the quality data, the die service life and the antenna performance is constructed. And when the model predicts that the residual service life of the mold is lower than 1000 injection molding cycles, automatically triggering an optimization flow. The method comprises the steps of adopting a multi-objective genetic algorithm, taking geometric parameters of the length and width of an antenna radiating unit and array arrangement density as optimization variables, taking an injection molding cycle with the service life of a mold being longer than 5000 times and an antenna gain being longer than 15dB as optimization targets, taking the quality characteristic parameters of the mold into constraint conditions, and searching for optimal antenna design parameter combinations. Based on the optimization result, the intelligent CAD/CAE system is utilized to automatically modify and redesign the die structure, and a digital processing program is generated to guide the die processing. Meanwhile, the quality state of the die is continuously monitored, and when the surface roughness of the die cavity exceeds 6um or the abrasion loss of the die exceeds 2mm, the optimization is restarted, so that the intelligent upgrading of the die is realized.

The foregoing is merely illustrative of some preferred embodiments of the present invention, but the invention is not limited thereto and many modifications and variations are possible. Any modifications or variations which are based on the basic principles of the present invention should be considered as falling within the scope of the present invention.

Claims (8)

1. The production method of the mold for the cold stamping 6G antenna part is characterized by comprising the following steps of:

Acquiring cavity surface roughness data and frequency band layout parameters of a mold for cold stamping 6G antenna parts, acquiring related data through a surface morphology measuring instrument and a spectrum analyzer, and establishing a data mapping relation between the cavity surface roughness and the frequency band layout, wherein the mapping relation is used for subsequently analyzing the performance of the mold under the 6G frequency band;

According to the obtained cavity surface roughness data and frequency band layout parameters, adopting a finite element analysis method to simulate distribution characteristics of residual stress on the cavity surface under different frequency band layouts, obtaining a residual stress distribution cloud picture, determining a stress concentration area, and visually displaying stress distribution conditions by the residual stress distribution cloud picture, wherein the residual stress distribution cloud picture is used for identifying a potential high risk area of die abrasion;

Training and establishing a prediction model among the frequency band layout parameters, the surface roughness of the cavity and the abrasion of the die by a machine learning algorithm based on the cavity surface roughness data, the frequency band layout parameters and the residual stress distribution cloud image, and judging the abrasion state and the service life of the die under the current frequency band layout and the surface roughness according to the model prediction result;

In the cold stamping forming process, an online monitoring system is adopted to collect surface roughness of a cavity and frequency band layout parameters in real time, the wear state of a die is dynamically estimated through comparison and analysis with a prediction model, when the wear state exceeds a preset threshold value, a feedback control mechanism is triggered, and the frequency band layout parameters are automatically adjusted according to the output result of the prediction model so as to slow down the die wear;

Aiming at the association relation between the frequency band layout defects and the cavity surface roughness, the on-line monitoring data and the prediction model result are combined, the evolution rule of the frequency band layout defects and the surface roughness is excavated through a data excavation technology, a multi-scale association model of frequency band layout design, surface quality prediction and die wear evaluation is established, and the model realizes the cross-scale comprehensive analysis by integrating microscopic surface roughness, mesoscopic frequency band layout and macroscopic die wear data;

Integrating on-line monitoring data, a prediction model result and a multi-scale association model analysis result into a full life cycle management system of the die, realizing full cycle monitoring and early warning of the die quality through big data analysis and visualization technology, optimizing a die maintenance strategy by utilizing the multi-scale association model, predicting potential faults, and providing data support for die design optimization;

Based on integrated full life cycle quality data of the die and a multi-scale associated model, an intelligent optimization algorithm is adopted to optimize frequency band layout design parameters, in the optimization process, the surface quality of a cavity, the abrasion state of the die and the performance requirements of an antenna are considered, and the balance of the maximization of the service life of the die and the optimization of the performance of the antenna is realized by adjusting the geometric parameters and the arrangement mode of the antenna radiation units, so that the intellectualization and the high efficiency of the design, the manufacture, the use and the maintenance of the die are realized.

2. The method for producing a mold for cold stamping 6G antenna parts according to claim 1, wherein the step of obtaining the cavity surface roughness data and the frequency band layout parameters of the mold for cold stamping 6G antenna parts, collecting the related data by a surface topography measuring instrument and a spectrum analyzer, and establishing a data mapping relationship between the cavity surface roughness and the frequency band layout, wherein the mapping relationship is used for the subsequent analysis of the performance of the mold under the 6G frequency band, comprises the steps of:

According to cavity surface roughness data and frequency band layout parameters of a mold for cold stamping 6G antenna parts, a nonlinear regression model of the cavity surface roughness and the frequency band layout is established by adopting a support vector machine algorithm, and a data mapping relation of the cavity surface roughness and the frequency band layout is obtained;

Classifying and predicting the performance of the die under the 6G frequency band by adopting a decision tree algorithm according to the established data mapping relation between the surface roughness of the die cavity and the frequency band layout, judging that the die can be used for producing 6G antenna parts if the prediction result meets a preset performance threshold value, otherwise judging that the die needs to be subjected to surface treatment;

For a mold needing surface treatment, treating the surface of a cavity in a chemical corrosion or mechanical polishing mode, acquiring the surface roughness data of the cavity after treatment, inputting the surface roughness data into an established data mapping model of the surface roughness of the cavity and the frequency band layout, and predicting the corresponding frequency band layout parameters;

according to the predicted frequency band layout parameters, optimizing the structural size of the antenna parts by adopting a genetic algorithm to obtain the structural parameters of the antenna parts meeting the 6G frequency band requirements;

According to the optimized structural parameters of the antenna parts, a corresponding cold stamping die is processed through a numerical control machining center, the surface roughness of a machined die cavity is detected by utilizing a three-coordinate measuring instrument, and detection data are input into a data mapping model of the cavity surface roughness and frequency band layout for verification;

If the verification result shows that the surface roughness of the die cavity and the frequency band layout parameters meet the requirements, cold stamping forming of the 6G antenna part is carried out by using the die, and a finished product part is obtained;

If the verification result shows that the surface roughness of the die cavity and the frequency band layout parameters do not meet the requirements, returning to process the surface of the die cavity again until the requirements are met.

3. The method for producing a mold for cold stamping 6G antenna parts according to claim 1, wherein the step of simulating distribution characteristics of residual stress on a cavity surface under different frequency band layouts by using a finite element analysis method according to the acquired cavity surface roughness data and frequency band layout parameters to obtain a residual stress distribution cloud image, determining a stress concentration area, and visually displaying stress distribution conditions of the residual stress distribution cloud image for identifying a potential mold abrasion high risk area comprises the steps of:

Acquiring cavity surface roughness data, establishing a cavity surface three-dimensional model, and mapping the roughness data to the model surface;

According to the frequency band layout parameters, carrying out frequency band division on the three-dimensional model on the surface of the cavity to generate finite element models under different frequency band layouts;

setting material properties, boundary conditions and load working conditions for the finite element model under each frequency band layout, and carrying out finite element analysis calculation;

adopting a finite element analysis result to obtain a distribution cloud picture of residual stress on the surface of the cavity under different frequency band layouts, and visually displaying the stress distribution situation;

Identifying a stress concentration area through a stress distribution cloud chart, and judging whether the stress concentration degree exceeds the allowable stress of the material;

If the stress concentration exceeds the allowable stress of the material, determining the area as a potential high-risk area of die wear;

and (3) combining analysis results of the high-risk areas of the die abrasion under different frequency band layouts, determining an optimal frequency band layout scheme, and guiding the optimization of the surface frequency band layout of the cavity.

4. The method for producing a mold for cold stamping 6G antenna parts according to claim 1, wherein the training to build a predictive model between the frequency band layout parameters, the surface roughness of the mold and the wear of the mold by a machine learning algorithm based on the cavity surface roughness data, the frequency band layout parameters and the residual stress distribution cloud image, and judging the wear state and the service life of the mold under the current frequency band layout and the surface roughness according to the model predictive result comprises:

Acquiring cavity surface roughness data, frequency band layout parameters and residual stress distribution cloud image data as a training data set of a machine learning algorithm;

Training based on a training data set by adopting a support vector machine and a random forest machine learning algorithm, and establishing a frequency band layout parameter and a prediction model between cavity surface roughness and die abrasion;

according to the established prediction model, inputting the frequency band layout parameters and the surface roughness data of the current cavity, and predicting to obtain the abrasion state of the die;

if the predicted wear state exceeds a preset threshold, judging that the die has reached the service life and needs to be maintained or replaced;

Otherwise, judging that the die can still be used continuously;

Acquiring a cavity residual stress distribution cloud picture, extracting residual stress distribution characteristics, and taking the residual stress distribution cloud picture as supplementary input of a prediction model to improve the accuracy of the prediction of the wear state of the die;

By analyzing the prediction result, determining key factors influencing the die abrasion, such as key frequency bands in frequency band layout parameters and critical values of surface roughness, and providing basis for optimizing die design;

and according to the prediction result of the wear state of the die and the analysis of key influencing factors, a die maintenance strategy is formulated, such as periodic detection of surface roughness and optimization of frequency band layout, so that the service life of the die is prolonged, and the production efficiency is improved.

5. The method for producing a mold for a cold stamping 6G antenna part according to claim 1, wherein in the cold stamping forming process, an on-line monitoring system is adopted to collect surface roughness of a cavity and frequency band layout parameters in real time, the mold wear state is dynamically evaluated through comparison analysis with a prediction model, when the wear state exceeds a preset threshold value, a feedback control mechanism is triggered, and the frequency band layout parameters are automatically adjusted according to an output result of the prediction model to slow down the mold wear, and the method comprises:

acquiring roughness data and frequency band layout parameters of the surface of a die in the cold stamping forming process, inputting the acquired data into a pre-established prediction model for comparison and analysis, and obtaining a wear state evaluation result of the die;

Judging whether the evaluation result exceeds a preset threshold value, if so, triggering a feedback control mechanism, and determining frequency band layout parameters to be adjusted, such as the size or the spacing of antenna radiating units, by adopting an optimization algorithm according to the output result of the prediction model;

Establishing a correlation model between the frequency band layout parameters and the die wear by a machine learning algorithm, and determining an optimal frequency band layout parameter adjustment scheme by an intelligent optimization algorithm based on the correlation model so as to minimize the die wear;

Transmitting the optimized frequency band layout parameters to a numerical control system, controlling a numerical control machine tool to automatically adjust the size and the interval of antenna radiating units, changing the frequency band layout, and slowing down the abrasion speed of the die;

Monitoring the influence of the adjusted frequency band layout parameters on the abrasion of the die in real time, and if the abrasion state still exceeds the threshold value, re-optimizing the frequency band layout parameters until the abrasion state of the die is lower than the preset threshold value;

A plurality of prediction models and associated models are established in advance aiming at different die materials and cold stamping process parameters, and the optimal models are dynamically switched and selected according to actual production conditions, so that the accuracy of prediction and control is improved;

And (3) adopting a big data analysis technology to perform mining analysis on the acquired mass die wear data and frequency band layout parameter data, extracting an implicit association rule and an evolution trend, and providing data support for optimization of the prediction model and the association model.

6. The method for producing a mold for cold stamping 6G antenna parts according to claim 1, wherein the cross-scale integrated analysis comprises:

acquiring frequency band layout design parameters, cavity surface roughness measurement data and mold wear monitoring data, and constructing a multi-scale data set;

Preprocessing the multi-scale data, including data cleaning, feature extraction and data integration, to obtain standardized data representation;

adopting an association rule mining algorithm, such as Apriori or FP-growth, and finding association rules and modes between the frequency band layout defects and the surface roughness from the standardized data;

Constructing a prediction model of frequency band layout defects and surface roughness evolution, such as a decision tree, a support vector machine or a neural network, according to the excavated association rule;

applying the prediction model to on-line monitoring data to predict the evolution trend of cavity surface quality and frequency band layout defects in real time;

comprehensively analyzing cross-scale correlation among frequency band layout, surface roughness and die abrasion, establishing a multi-scale correlation model, and quantifying an influence mechanism among different factors;

Based on the multi-scale association model, a frequency band layout design scheme is optimized, a die maintenance strategy is formulated, the surface quality of a cavity is improved, and the service life of a die is prolonged.

7. The method for producing a mold for cold stamping 6G antenna parts according to claim 1, wherein the integrating on-line monitoring data, prediction model results, and multi-scale correlation model analysis results into a full life cycle management system of the mold, realizing full cycle monitoring and early warning of mold quality by big data analysis and visualization technology, optimizing a mold maintenance strategy by using the multi-scale correlation model, predicting potential faults, and providing data support for mold design optimization comprises:

Acquiring real-time data including temperature, pressure and speed parameters in the production process of the mold through an online monitoring system, and transmitting the data to a full life cycle management system;

Inputting the acquired real-time monitoring data into a prediction model, analyzing and calculating through an algorithm to obtain a performance prediction result of the die in a future period of time, and judging whether the die has potential quality risks or not;

if the prediction result shows that the quality risk exists, triggering an early warning mechanism, automatically generating early warning information by a system, and prompting related personnel to take countermeasures in time through visual interface display;

According to the prediction result and the historical data, analyzing the relation between each parameter of the die and the performance and service life of the die through the association model to obtain key factors influencing the quality of the die, and providing basis for the subsequent die design optimization;

Mining and analyzing various data in the whole life cycle of the die by adopting a big data analysis technology, finding out the association rule and mode among the data, and providing data support for optimizing the maintenance strategy of the die;

inputting design parameters, material properties and process condition data of the mold into a correlation model, training through a machine learning algorithm, and establishing a mold performance prediction model for guiding mold design optimization;

Based on historical data and prediction results in the full life cycle management system, a heuristic algorithm is adopted to optimize the maintenance strategy of the die, a scientific maintenance plan is formulated, the service life of the die is prolonged to the maximum extent, and the maintenance cost is reduced.

8. The method for producing a mold for cold stamping 6G antenna parts according to claim 1, wherein the integrated mold full life cycle quality data and multi-scale correlation model are used to optimize frequency band layout design parameters by using an intelligent optimization algorithm-genetic algorithm, and in the optimization process, the balance of maximizing the mold life and optimizing the antenna performance is achieved by adjusting the geometric parameters and arrangement of antenna radiating elements in consideration of the cavity surface quality, the mold wear state and the antenna performance requirements, thereby achieving the intellectualization and high efficiency of mold design, manufacturing, use and maintenance, comprising:

acquiring quality data of a full life cycle of a die, including cavity surface quality data and die wear state data, constructing a multi-scale association model, and establishing an association relationship between the quality data and the die life and antenna performance;

Judging the current use state and the residual service life of the die according to the surface quality data of the cavity and the wear state data of the die, and triggering an antenna frequency band layout optimization flow if the residual service life of the die is lower than a preset threshold value;

Adopting a genetic algorithm intelligent optimization algorithm, taking geometric parameters and arrangement modes of antenna radiating units as optimization variables, taking the service life of a die and the performance of an antenna as optimization targets, and carrying out multi-target optimization solution;

In the optimization process, the surface quality of a cavity and the abrasion state of the die are related to the optimization variable through a multi-scale related model of the die, so that the optimization process can consider the quality factor of the die;

obtaining a group of optimal antenna layout design parameters meeting the requirements of the service life of the die and the performance of the antenna through iterative optimization, wherein the optimal antenna layout design parameters comprise the geometric parameters and the arrangement modes of the antenna radiating units;

according to the optimized antenna layout design parameters, the die is intelligently modified and redesigned to generate a new die processing program and technological parameters, and the processing and manufacturing of the die are guided;

And continuously collecting quality data of the die in the use and maintenance stage of the die, evaluating the state of the die in real time, and triggering the optimization flow again when the quality of the die is reduced below a threshold value, upgrading and updating the die, so as to realize intelligent management of the full life cycle of the die.