Mould matching system for magnetic material processing

Abstract

The invention discloses a mold matching system for magnetic material processing, which belongs to the technical field of magnetic material processing, and comprises a mold library management module for storing and managing attribute information of molds; the technological parameter input module is used for inputting technological parameters related to each working procedure of magnetic material production; the intelligent matching engine is used for calculating and dynamically adjusting the influence weights of the process parameters on the sizes of the blank in the die width direction, the orientation direction and the pressing direction according to a historical production data cluster; based on the target blank size and the process parameters input in real time, an optimal mold and a matched process scheme are obtained through matching; and the closed-loop feedback optimization module is used for collecting the actual size of the sintered blank, carrying out difference comparison on the actual size and the preset size, and carrying out parameter updating and optimization on the three-dimensional weight dynamic optimization model through a deep reinforcement learning mechanism based on difference data. According to the method, collaborative optimization of the mold and process parameters is realized, and the problem of low efficiency of traditional manual matching is solved.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of magnetic material processing, and particularly relates to a mold matching system for magnetic material processing. BACKGROUND

[0002] In the field of magnetic material manufacturing, mold matching is a key link affecting product quality. The existing technology has the following defects: strong experience dependence: traditional mold matching relies on the experience of engineers and lacks quantitative standards; manual calculation: manual calculation of similar mold sizes is performed by manual calculation, and then the mold is selected; lack of closed-loop feedback: the matching result is disconnected from the actual production effect, and continuous improvement cannot be formed. Therefore, the insufficient mold matching precision leads to a decrease in product qualification rate, a long time for mold replacement, and the need to adjust parameters for new material adaptation, with an average debugging period of 3 days; the mold utilization efficiency is less than 70%, and resources are wasted seriously.

[0003] Therefore, a mold matching system for magnetic material processing is needed to solve the above problems. SUMMARY

[0004] Therefore, the purpose of the present application is to provide a mold matching system for magnetic material processing, which solves the problems of experience-dependent, low efficiency and lack of closed-loop optimization in the matching of magnetic material manufacturing molds in the prior art.

[0005] To achieve the above purpose, the present application provides the following technical scheme: The present application provides a mold matching system for magnetic material processing, comprising: a mold library management module for storing and managing the attribute information of the mold, the attribute information including mold number, mold width direction size, orientation direction size, pressing direction size and state information; A process parameter input module is used for inputting process parameters related to each process of magnetic material production, wherein the processes include melting, hydrogen crushing, airflow grinding, forming and isostatic pressing, and the process parameters include powder composition, additive amount, powder particle size, forming density and isostatic pressing pressure; An intelligent matching engine, the core of which is a three-dimensional weight dynamic optimization model, is used for calculating and dynamically adjusting the influence weight of each process parameter on the size of the blank in the mold width, orientation and pressing directions according to the historical production data cluster; based on the target blank size and the real-time input process parameters, the optimal mold and supporting process scheme are matched; A closed-loop feedback optimization module is used for collecting the actual size of the sintered blank, comparing the actual size with the preset size, and updating and optimizing the three-dimensional weight dynamic optimization model based on the difference data through a deep reinforcement learning mechanism.

[0006] Furthermore, the mold library management module is equipped with a dynamic classification mechanism for mold status and a cloud storage-based retrieval interface. The attribute information includes mold material, service life count, wear coefficient, and maintenance history. The status information includes three levels: available, under maintenance, and retired. The attribute information is associated with a mold usage frequency database. The module is equipped with an intelligent retrieval engine that uses a fuzzy matching-based query algorithm, allowing users to quickly locate molds through multi-dimensional filtering conditions.

[0007] Furthermore, the query algorithm execution steps include: S1, Input the target size parameters, and convert the size into a uniform unit vector using a normalization function; S2, calculate the matching score of the candidate mold using the following formula: In the formula, For the target size vector, For mold dimension vectors, Let be the number of dimensions, where =3, corresponding to mold width, orientation, and pressing direction; S3, combining state information, generates a ranking list based on matching degree; where state information is assigned as a weighting factor. Available status Under repair Retired The mold library management module integrates an early warning subsystem, which automatically triggers maintenance reminders and updates status information when the mold lifespan count exceeds the threshold.

[0008] Furthermore, the process parameter input module includes an automatic parameter verification unit that uses a rule engine for real-time verification. If a parameter exceeds the preset range, an alarm is triggered and a correction value is suggested. A multi-source data integration interface is used to receive melting temperature, hydrogen crushing time, and air jet mill speed parameters collected by the production line sensors.

[0009] Furthermore, the three-dimensional weight dynamic optimization model of the intelligent matching engine employs a multi-objective genetic algorithm for weight calculation and dynamic adjustment, wherein the influence weight is defined as... Each weight's initial value is determined based on regression analysis of historical production data clusters and is updated through iterative optimization; the specific steps are as follows: B1, Construct the 3D size influence matrix M, matrix elements Indicates the first The process parameter affects the first The influence coefficients for each dimension direction, and the initial matrix were extracted from historical data through principal component analysis; B2 introduces an adaptive learning rate mechanism to dynamically adjust the weights. The weight update formula is: wherein, is the weight of the th iteration, is the learning rate, is the number of historical data samples, are the actual and expected blank dimensions, respectively, is the process parameter, is the parameter sensitivity calculated by partial derivative; B3, calculate the fitness function to match the optimal solution, use multi-objective genetic algorithm to solve the global optimal solution.

[0010] Further, the deep reinforcement learning mechanism adopts an Actor-Critic architecture based on policy gradient, and the specific process includes: C1, define the state space as difference data (the deviation vector of actual size and expected size), wherein, is the width direction size deviation, is the orientation direction size deviation, is the pressing direction size deviation; define the action space as model parameter update , wherein, is the width weight adjustment amount, is the orientation weight adjustment amount, is the pressing weight adjustment amount; calculate the reward function , wherein, is the stability penalty coefficient; C2, collect the actual size of the blank after sintering, obtain three-dimensional point cloud data through a laser scanner, and convert it into a size vector; C3, calculate the size deviation After that, the Actor network outputs parameter update suggestions, and the Critic network evaluates the state value function , and the update formula is: wherein, is the network parameter, is the learning rate, is the discount factor, is the policy function, and are the current state and the next state, respectively; C4, drive model parameter update and store experience replay cache.

[0011] Further, the magnetic material processing mold matching system further comprises a physical simulation module integrated in the front stage of the intelligent matching engine, for predicting the size deformation of the blank based on finite element analysis, the physical simulation module is configured to: input process parameters and mold geometric data, perform multi-physical field simulation, and output a strain distribution map; the simulation process adopts an explicit dynamic equation: In the formula, is a mass matrix, is a damping matrix, is a stiffness matrix, is a displacement vector, is a pressing force; The simulation result is converted into a size prediction value, and is weighted and fused with the output of the three-dimensional weight dynamic optimization model, and the formula is: In the formula, the weight According to the historical accuracy rate.

[0012] Further, the historical production data cluster is preprocessed by an improved density clustering algorithm to optimize the input quality of the three-dimensional weight dynamic optimization model; the processing steps include: Data acquisition stage: collect multi-dimensional data points, and use a sliding window technology to process time series data; Clustering stage: apply a self-defined distance metric: In the formula, is a parameter initial influence weight, is a data point feature, are two data points in a multi-dimensional space, respectively; After clustering, high-density clusters and noise points are generated, and only high-density clusters are used for model training; at the same time, cluster labels are added to the data for the intelligent matching engine to preferentially match.

[0013] Further, the magnetic material processing mold matching system further comprises a graphical user interface module for providing process parameter input, matching result display and feedback analysis functions; wherein the graphical user interface module is a multi-tab page structure: the first tab page is used for process parameter input, supports drag-and-drop uploading of data files, and embeds real-time verification prompts; the second tab page displays the output of the intelligent matching engine, including a three-dimensional rendering diagram of the optimal mold and a Gantt chart of the matching process scheme, wherein the process scheme displays parameter suggestions for each process in time sequence; the third tab page is used for the closed-loop feedback optimization module, visualizes the difference comparison curve and the model optimization progress bar, and allows the user to manually adjust the parameter weight; in addition, the graphical user interface integrates report export function, generates PDF report, and supports multi-language switching.

[0014] The beneficial effects of this invention are as follows: This invention achieves coordinated optimization of mold and process parameters, solving the inefficiency problem of traditional manual matching; through closed-loop feedback, the model continuously adapts to production fluctuations, improving the blank size qualification rate; and the system response time is less than 3 seconds, making it suitable for high-frequency production environments of magnetic materials.

[0015] Other advantages, objectives, and features of the invention will be set forth in the following description and will be apparent to those skilled in the art in some respects, or may be learned by practice of the invention. The objectives and other advantages of the invention can be realized and obtained through the following description. Attached Figure Description

[0016] To make the objectives, technical solutions, and beneficial effects of this invention clearer, the following figures are provided for illustration: Figure 1 This is a flowchart of the mold matching system for magnetic material processing according to the present invention. Detailed Implementation

[0017] like Figure 1 As shown, the present invention provides a mold matching system for processing magnetic materials, comprising: The mold library management module is used to store and manage the attribute information of molds, including mold number, mold width dimension, orientation dimension, pressing dimension, and status information. The process parameter input module is used to input process parameters related to each step of the magnetic material production process, including melting, hydrogen crushing, air jet milling, molding and isostatic pressing. The process parameters include powder composition, dosage, powder particle size, molding density and isostatic pressing pressure. The intelligent matching engine, at its core, is a three-dimensional weighted dynamic optimization model. It is used to calculate and dynamically adjust the influence weight of each process parameter on the dimensions of the blank in the three directions of mold width, orientation, and pressing based on historical production data clusters. Based on the target blank size and the real-time input process parameters, it matches the optimal mold and supporting process scheme. The closed-loop feedback optimization module is used to collect the actual size of the sintered blank, compare the actual size with the preset size, and update and optimize the parameters of the three-dimensional weighted dynamic optimization model based on the difference data through a deep reinforcement learning mechanism.

[0018] This solution achieves collaborative optimization of mold and process parameters, solving the inefficiency of traditional manual matching; through closed-loop feedback, the model continuously adapts to production fluctuations (such as changes in raw materials), improving the blank size qualification rate; and the system response time is less than 3 seconds, making it suitable for high-frequency production environments of magnetic materials.

[0019] In an embodiment of the present application, the mold library management module is provided with a mold state dynamic classification mechanism and a cloud storage-based retrieval interface, wherein the attribute information further includes mold material, service life count, wear coefficient and maintenance history; the state information includes three levels of states of available, under repair and retired, and the attribute information is associated to a mold use frequency database; the module is configured with an intelligent retrieval engine, which adopts a fuzzy matching-based query algorithm to allow a user to quickly locate a mold through multi-dimensional screening conditions including mold width direction size range, orientation direction size tolerance ±0.5 mm and pressing direction size precision threshold; the query algorithm includes the following steps: S1, input target size parameters, and convert the size to a unified unit vector through a standardization function; S2, calculate the matching degree score of the candidate mold through the following formula: In the formula, is a target size vector, is a mold size vector, is the number of dimensions (n=3, corresponding to mold width, orientation and pressing direction); S3, generate a matching degree sorting list in combination with state information (assign a weight factor , wherein the available state , under repair , and retired ); wherein the mold library management module integrates a warning subsystem, which automatically triggers a maintenance reminder and updates the state information when the mold service life count exceeds a threshold value (such as 10,000 times), to ensure real-time data. In the present scheme, the attributes such as mold material and wear coefficient are added, and the use frequency database is associated through state classification (available / under repair / retired); the intelligent retrieval engine calculates the matching degree of the candidate mold after standardizing the input target size into a vector (such as mold width, orientation and pressing) and superimposes the state weight (available state ×1.0, under repair ×0.3) to generate a sorting list; through the enhancement of attribute details and intelligent retrieval, the mold utilization rate is improved by more than 10%, and manual screening errors are avoided.

[0020] In an embodiment of the present application, the process parameter input module includes a parameter automatic verification unit, which adopts a rule engine for instant verification, and if the parameter exceeds the preset range, an alarm is triggered and a correction value is suggested; a multi-source data integration interface is used to receive smelting temperature, hydrogen fragmentation time and air flow mill speed parameters collected by the production line sensors;

[0021] In the present application, the process parameters are input in the following ways: ​A1, user manual input interface, configured with data range verification rules (for example, powder particle size range 0.5-50 , additive amount error tolerance ±0.1%); A2, real-time acquisition of smelting temperature, hydrogen fragmentation time and air flow mill speed parameters; the parameter automatic verification unit adopts a rule engine for instant verification, if the parameters exceed the preset range, an alarm is triggered and a correction value is suggested; at the same time, based on an external database (such as a material science library), the parameters (such as the trace element ratio of powder composition) are supplemented by API calling and then normalized data processing is carried out, and the processing formula is: In the formula, is the original parameter value, are the minimum and maximum values of the historical data respectively; A3, input the normalized parameters to the intelligent matching engine.

[0022] In the present scheme, the process parameter input module is provided with a batch processing function, allowing batch uploading of historical process data sets (format supporting CSV / JSON), and generating parameter trend reports (such as correlation analysis of isostatic pressure and molding density). The present design reduces the input error rate to below 1% through automatic verification and multi-source integration, improving the system input reliability.

[0023] In an embodiment of the present application, the three-dimensional weight dynamic optimization model of the intelligent matching engine adopts a multi-objective genetic algorithm (MOGA) for weight calculation and dynamic adjustment, wherein the influence weight is defined as , and the initial value of each weight is determined based on the regression analysis of the historical production data set, and is updated through iterative optimization; the following steps are specifically executed: B1, construct a three-dimensional size influence matrix M, the matrix elements represent the influence coefficient of the th process parameter (such as powder composition) on the th size direction (mold width, orientation, pressing), and the initial matrix is extracted from the historical data by principal component analysis (PCA); B2, introduce an adaptive learning rate mechanism to dynamically adjust the weight, and the weight update formula is: In the formula, is the weight of the th iteration, is the learning rate, is the number of historical data samples, are the actual and expected blank sizes respectively, for process parameters, for calculating parameter sensitivity by partial derivative; B3, calculating fitness function To match the optimal solution, MOGA (Multi-Objective Genetic Algorithm) is used to solve the global optimal solution.

[0024] In this scheme, the combination of MOGA and dynamic sensitivity solves the multi-parameter conflict problem (such as the trade-off between powder particle size and isostatic pressure).

[0025] In an embodiment of the present application, the deep reinforcement learning mechanism adopts an Actor-Critic architecture based on policy gradient, and the specific process includes: C1, defining the state space as difference data (the deviation vector of actual size and expected size), In the formula, is the width direction size deviation, is the orientation direction size deviation, is the pressing direction size deviation; defining the action space as model parameter update , In the formula, is the width weight adjustment amount, is the orientation weight adjustment amount, is the pressing weight adjustment amount; calculate the reward function In the formula, is the stability penalty coefficient; C2, collect the actual size of the sintered blank, obtain three-dimensional point cloud data through a laser scanner, and convert it into a size vector; C3, calculate the size deviation After that, the Actor network outputs parameter update suggestions, and the Critic network evaluates the state value function , the update formula is: In the formula, is the network parameter, is the learning rate, is the discount factor, is the policy function, and are the current state and the next state respectively; C4, drive model parameter update, and store experience replay cache to improve learning efficiency.

[0026] In the scheme, the deep reinforcement learning mechanism runs once every 100 production iterations, ensuring real-time optimization of the model, and the deviation reduction rate is as high as 20%; the scheme realizes self-adaptive optimization through the reinforcement learning framework, reducing the dependence on historical data.

[0027] In an embodiment of the present application, the magnetic material processing mold matching system further comprises a physical simulation module integrated in the front stage of the intelligent matching engine, which is used to predict the size deformation of the blank based on finite element analysis (FEA), and the module is configured to: input process parameters (such as forming density, isostatic pressing pressure) and mold geometric data, perform multi-physical field simulation (including thermal-mechanical coupling), and output strain distribution map; the simulation process adopts an explicit dynamic equation: In the formula, is a mass matrix, is a damping matrix, is a stiffness matrix, is a displacement vector, is a pressing force; The simulation result is converted into a size prediction value and is weighted and fused with the output of the three-dimensional weight dynamic optimization model, and the formula is: In the formula, the weight is dynamically adjusted according to the historical accuracy rate; at the same time, the module supports parameter sensitivity analysis (such as the influence curve of pressure change on size), guiding process optimization. The physical simulation module combines physical models and data-driven AI, solves the failure problem of pure data models under extreme process conditions (such as high-pressure isostatic pressing), improves the system robustness, and expands the applicability of the matching scheme.

[0028] In an embodiment of the present application, the historical production data cluster is preprocessed by an improved density clustering algorithm (DBSCAN variant) to optimize the input quality of the three-dimensional weight dynamic optimization model; the specific mechanism includes: in the data collection stage, multi-dimensional data points (each point containing a process parameter vector and a corresponding blank size vector) are collected, and sliding window technology is used to process time series data; in the clustering stage, a self-defined distance metric In the formula, is the initial influence weight of the parameter, is the data point feature, are two data points in a multi-dimensional space, respectively, and the DBSCAN variant introduces a dynamic neighborhood radius , the data density is the number of points per unit space; after clustering, high-density clusters (representing stable process areas) and noise points (representing abnormal data) are generated, only the high-density clusters are used for model training; at the same time, the mechanism adds cluster labels to the data for the intelligent matching engine to preferentially match. The scheme effectively solves the high-dimensional data sparsity problem by using the weight-weighted distance metric and the adaptive , effectively handles the high-dimensional data sparsity problem, improves the model training efficiency by 30%, and reduces the risk of overfitting.

[0029] In an embodiment of the present application, the system further comprises a graphical user interface (GUI) module for providing process parameter input, matching result display and feedback analysis functions; the graphical user interface (GUI) module is a multi-tab structure: the first tab page is used for process parameter input, supports drag-and-drop uploading of data files, and embeds real-time verification prompts (such as highlighting when the powder particle size is out of limit); the second tab page displays the output of the intelligent matching engine, including a three-dimensional rendering of the optimal mold (based on mold attribute information) and a Gantt chart of the supporting process plan, wherein the process plan displays parameter suggestions for each process (melting, hydrogen crushing, etc.) in time sequence; the third tab page is used for the closed-loop feedback optimization module, visualizes the difference comparison curve (actual size vs. expected size) and model optimization progress bar, and allows users to manually adjust parameter weights; in addition, the GUI integrates report export function to generate PDF reports (including matching degree statistics and optimization suggestions), and supports multi-language switching (Chinese and English).

[0030] The scheme reduces the operation threshold through a friendly interface, improves user adoption rate, and is especially suitable for on-site engineers to make quick decisions.

[0031] Finally, it should be pointed out that the above preferred embodiments are only used to illustrate the technical solutions of the present application and are not limiting, although the present application has been described in detail through the above preferred embodiments, those skilled in the art should understand that various changes can be made in form and details without departing from the scope defined by the claims of the present application.

Claims

1. A mold matching system for processing magnetic materials, characterized in that, include: The mold library management module is used to store and manage the attribute information of molds, including mold number, mold width dimension, orientation dimension, pressing dimension, and status information. The process parameter input module is used to input process parameters related to each step of the magnetic material production process. The steps include melting, hydrogen crushing, air jet milling, molding and isostatic pressing. The process parameters include powder composition, dosage, powder particle size, molding density and isostatic pressing pressure. The intelligent matching engine, with a core three-dimensional weighted dynamic optimization model, is used to calculate and dynamically adjust the influence weight of each process parameter on the dimensions of the blank in the three directions of mold width, orientation, and pressing based on historical production data clusters; and to match the optimal mold and supporting process scheme based on the target blank size and the real-time input process parameters. The closed-loop feedback optimization module is used to collect the actual size of the sintered blank, compare the actual size with the preset size, and update and optimize the parameters of the three-dimensional weighted dynamic optimization model based on the difference data through a deep reinforcement learning mechanism.

2. The mold matching system for magnetic material processing according to claim 1, characterized in that: The mold library management module is equipped with a dynamic classification mechanism for mold status and a cloud storage-based retrieval interface. The attribute information includes mold material, service life count, wear coefficient, and maintenance history. The status information includes three levels: available, under maintenance, and retired. The attribute information is associated with a mold usage frequency database. The module is equipped with an intelligent retrieval engine that uses a fuzzy matching-based query algorithm, allowing users to quickly locate molds through multi-dimensional filtering conditions.

3. The mold matching system for magnetic material processing according to claim 2, characterized in that: The query algorithm execution steps include: S1, Input the target size parameters, and convert the size into a uniform unit vector using a normalization function; S2, calculate the matching score of the candidate mold using the following formula: In the formula, For the target size vector, For mold dimension vectors, Let be the number of dimensions, where =3, corresponding to mold width, orientation, and pressing direction; S3, combining state information, generates a ranking list based on matching degree; where state information is assigned as a weighting factor. Available status Under repair Retired The mold library management module integrates an early warning subsystem, which automatically triggers maintenance reminders and updates status information when the mold lifespan count exceeds the threshold.

4. The mold matching system for magnetic material processing according to claim 3, characterized in that: The process parameter input module includes an automatic parameter verification unit that uses a rule engine for real-time verification. If a parameter exceeds the preset range, an alarm is triggered and a correction value is suggested. It also includes a multi-source data integration interface for receiving melting temperature, hydrogen crushing time, and air jet mill speed parameters collected by sensors on the production line.

5. The mold matching system for magnetic material processing according to claim 4, characterized in that: The intelligent matching engine's three-dimensional weight dynamic optimization model employs a multi-objective genetic algorithm for weight calculation and dynamic adjustment, where the influence weight is defined as... Each weight's initial value is determined based on regression analysis of historical production data clusters and is updated through iterative optimization; the specific steps are as follows: B1, Construct the 3D size influence matrix M, matrix elements Indicates the first The process parameter affects the first The influence coefficients for each dimension direction, and the initial matrix were extracted from historical data through principal component analysis; B2 introduces an adaptive learning rate mechanism to dynamically adjust the weights. The weight update formula is: In the formula, For the first Weights for the next iteration For learning rate, The number of historical data samples. These are the actual and expected blank dimensions, respectively. For process parameters, To calculate parameter sensitivity using partial derivatives; B3, Calculate the fitness function To find the optimal solution, a multi-objective genetic algorithm is used to solve for the global optimal solution.

6. The mold matching system for magnetic material processing according to claim 5, characterized in that: The deep reinforcement learning mechanism adopts an Actor-Critic architecture based on policy gradients, and the specific process includes: C1 defines the state space as differential data. (The vector of deviation between actual size and expected size), where, This refers to the dimensional deviation in the width direction. For the orientation direction dimension deviation, For dimensional deviations in the pressing direction; Define the action space as the model parameter update. In the formula, This is the width weight adjustment amount. This is the amount of adjustment for the orientation weight. To suppress the amount of weight adjustment; Calculate the reward function In the formula, This is a stability penalty coefficient; C2, collect the actual dimensions of the sintered blank, obtain three-dimensional point cloud data through a laser scanner, and convert it into a dimension vector; C3, Calculate dimensional deviations Then, by updating the parameters through the Actor network, the Critic network evaluates the state-value function. The updated formula is: In the formula, For network parameters, For learning rate, As a discount factor, For the policy function, and These represent the current state and the next state, respectively. C4 drives model parameter updates and stores experience replay cache.

7. The mold matching system for magnetic material processing according to claim 6, characterized in that: The magnetic material processing mold matching system also includes a physical simulation module integrated into the front end of the intelligent matching engine, which is used to predict the blank size deformation based on finite element analysis. The physical simulation module is configured to: input process parameters and mold geometric data, perform multi-physics simulation, and output strain distribution map. The simulation process uses explicit dynamic equations: In the formula, For the quality matrix, Here is the damping matrix. Here is the stiffness matrix. For displacement vectors, For suppressive force; The simulation results are converted into predicted dimensions and then weighted and fused with the output of the 3D weighted dynamic optimization model, as shown in the formula: In the formula, the weights Adjust dynamically based on historical accuracy.

8. The mold matching system for magnetic material processing according to claim 7, characterized in that: The historical production data cluster is preprocessed using an improved density clustering algorithm to optimize the input quality of the three-dimensional weighted dynamic optimization model. The processing steps include: Data acquisition phase: Collect multidimensional data points and process time-series data using sliding window technology; Clustering phase: Applying a custom distance metric: In the formula, The initial influence weights are parameters. For data point features, These are two data points in a multidimensional space; After clustering, high-density clusters and noise points are generated, with only high-density clusters used for model training; at the same time, cluster labels are added to the data for the intelligent matching engine to prioritize matching.

9. The mold matching system for magnetic material processing according to claim 8, characterized in that: The mold matching system for magnetic material processing also includes a graphical user interface (GUI) module, which provides functions for process parameter input, matching result display, and feedback analysis. The GUI module has a multi-tab structure: the first tab is for process parameter input, supports drag-and-drop data file upload, and embeds real-time verification prompts; the second tab displays the output of the intelligent matching engine, including a 3D rendering of the optimal mold and a Gantt chart of the matching process scheme, where the process scheme displays parameter suggestions for each process in a time series; the third tab is for the closed-loop feedback optimization module, visualizing difference comparison curves and model optimization progress bars, and allowing users to manually adjust parameter weights; in addition, the GUI integrates report export functionality, generating PDF reports and supporting multi-language switching.

Images