Intelligent processing method and system for process parameters of plastic mold

Abstract

The invention relates to the technical field of plastic production, in particular to an intelligent processing method and system for plastic mold process parameters. The method comprises the following steps: carrying out sensor network deployment on a plastic mold to obtain a mold parameter sensing network; operating parameters of the plastic mold are collected in real time through the mold parameter sensing network, and mold operating state data flow is obtained; performing process state feature extraction on the mold operation state data stream to obtain mold process state feature data; performing spatio-temporal data hierarchical analysis on the mold process state feature data to obtain mold state multi-scale feature data; and performing layered thermal field dynamic modeling on the plastic mold based on the mold state multi-scale characteristic data to obtain a mold thermal field evolution model. The key process parameters of the mold can be identified, and the stability and reliability of the mold performance are enhanced.

Description

Technical Field

[0001] The present invention relates to the technical field of plastic production, and in particular to an intelligent processing method and system for plastic mold process parameters. Background Art

[0002] Plastic molds are indispensable tools for producing various plastic products. Precise control of the process parameters of these molds is essential to ensure product quality and improve production efficiency. However, most of the existing methods for processing plastic mold process parameters rely on the operator's personal experience and adjustment through repeated trials. This method is not only inefficient, but also often difficult to adapt to complex and changing production conditions. Specifically, the traditional method has the following major problems:

[0003] First, due to the lack of real-time monitoring and data analysis of the mold production process, traditional methods are difficult to accurately capture the slight changes in the mold during the production process. If these changes are not identified and adjusted in time, it will lead to a decline in product quality and reduced production efficiency. Secondly, plastic molds are affected by many factors during the production process, such as mold temperature and pressure. The dynamic changes of these factors put higher requirements on the optimization of process parameters, making parameter adjustment more complicated. Traditional processing methods often cannot effectively cope with these dynamic changes, resulting in unstable mold performance, affecting the quality of the final product and the overall efficiency of production. Summary of the invention

[0004] Based on this, it is necessary for the present invention to provide a method and system for intelligently processing plastic mold process parameters to solve at least one of the above technical problems.

[0005] To achieve the above object, a method for intelligently processing process parameters of a plastic mold comprises the following steps:

[0006] Step S1: deploying a sensor network on the plastic mold to obtain a mold parameter sensing network; collecting operation parameters of the plastic mold in real time through the mold parameter sensing network to obtain a mold operation status data stream; extracting process status features from the mold operation status data stream to obtain mold process status feature data;

[0007] Step S2: performing spatiotemporal data hierarchical analysis on the mold process state characteristic data to obtain multi-scale characteristic data of the mold state; performing hierarchical thermal field dynamic modeling on the plastic mold based on the multi-scale characteristic data of the mold state to obtain a mold thermal field evolution model;

[0008] Step S3: collecting historical operation data of the plastic mold to obtain a mold historical operation data set; mining the process parameter correlation of the mold historical operation data set to obtain a mold process parameter correlation network;

[0009] Step S4: Based on the mold thermal field evolution model and the mold process parameter association network, a state prediction model is performed on the plastic mold to obtain a mold state prediction model; based on the mold state prediction model, the process parameters of the plastic mold are optimized in real time according to the mold operation state data stream to obtain a mold process optimization parameter solution;

[0010] Step S5: According to the mold process optimization parameter scheme, the process parameters of the plastic mold are adjusted to obtain a mold control instruction sequence; according to the mold control instruction sequence and the multi-scale feature data of the mold state, a closed-loop control strategy is constructed for the plastic mold to obtain a mold intelligent manufacturing control scheme.

[0011] The present invention collects mold operating parameters in real time through a sensor network, thereby improving the accuracy of process parameter control, thereby improving product quality and production efficiency. Through real-time data analysis and dynamic modeling, the adaptability of the production process to complex and changeable environments is enhanced, especially under the influence of dynamic temperature and pressure factors, the stability of mold performance is maintained. Through intelligent monitoring and control capabilities, the thermal field evolution of the mold can be predicted, and real-time optimization of process parameters can be achieved. Through automated and intelligent process parameter optimization solutions, the reliance on manual experience and trial and error methods is reduced, time and cost are saved, and production efficiency is improved. Through mining and associative network analysis of historical mold operation data, key process parameters can be identified, and the stability and reliability of mold performance can be enhanced. The implementation of a closed-loop control strategy further improves product quality and the controllability of the production process. The present invention also helps to extend the service life of the mold by real-time monitoring and optimization of process parameters.

[0012] Preferably, the present invention further provides an intelligent processing system for plastic mold process parameters, which is used to execute the intelligent processing method for plastic mold process parameters as described above. The intelligent processing system for plastic mold process parameters comprises:

[0013] The sensing and acquisition module is used to deploy a sensor network for the plastic mold to obtain a mold parameter sensing network; to collect the operation parameters of the plastic mold in real time through the mold parameter sensing network to obtain a mold operation status data stream; to extract process status features from the mold operation status data stream to obtain mold process status feature data;

[0014] The feature modeling module is used to perform spatiotemporal data hierarchical analysis on the feature data of the mold process state to obtain multi-scale feature data of the mold state; based on the multi-scale feature data of the mold state, the plastic mold is subjected to hierarchical thermal field dynamic modeling to obtain a mold thermal field evolution model;

[0015] The data mining module is used to collect historical operation data of the plastic mold to obtain a mold historical operation data set; perform process parameter correlation mining on the mold historical operation data set to obtain a mold process parameter correlation network;

[0016] The prediction and optimization module is used to perform state prediction modeling on the plastic mold based on the mold thermal field evolution model and the mold process parameter association network to obtain the mold state prediction model; based on the mold state prediction model, the process parameters of the plastic mold are optimized in real time according to the mold operation state data stream to obtain the mold process optimization parameter solution;

[0017] The control strategy module is used to control the process parameters of the plastic mold according to the mold process optimization parameter plan to obtain the mold control instruction sequence; based on the mold control instruction sequence and the multi-scale feature data of the mold state, a closed-loop control strategy for the plastic mold is constructed to obtain the mold intelligent manufacturing control plan.

[0018] The present invention improves the efficiency and accuracy of data collection through the perception collection module. The feature modeling module enhances the ability to identify the mold process status, and the multi-scale feature analysis improves the understanding and prediction of mold state changes. The dynamic modeling and prediction functions enable the system to scientifically adjust the process parameters, while the data mining module deepens the analysis of historical data and optimizes the process parameters. The prediction and optimization module realizes real-time optimization of process parameters, improving production efficiency and product quality. The control strategy module enhances the stability and reliability of production through intelligent control and closed-loop control. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Other features, objects and advantages of the present invention will become more apparent from the detailed description made with reference to the following drawings:

[0020] Figure 1 A schematic flow chart of the steps of an intelligent processing method for plastic mold process parameters according to an embodiment is shown.

[0021] Figure 2 A detailed flowchart diagram of step S13 of an embodiment is shown.

[0022] Figure 3 A detailed flowchart diagram of step S135 of an embodiment is shown. DETAILED DESCRIPTION

[0023] The technical method of the present invention is described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by technicians in this field without creative work are within the scope of protection of the present invention.

[0024] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor methods and / or microcontroller methods.

[0025] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0026] To achieve this, please refer to Figures 1 to 3 The present invention provides an intelligent processing method for plastic mold process parameters, comprising the following steps:

[0027] Step S1: deploying a sensor network on the plastic mold to obtain a mold parameter sensing network; collecting operation parameters of the plastic mold in real time through the mold parameter sensing network to obtain a mold operation status data stream; extracting process status features from the mold operation status data stream to obtain mold process status feature data;

[0028] Step S2: performing spatiotemporal data hierarchical analysis on the mold process state characteristic data to obtain multi-scale characteristic data of the mold state; performing hierarchical thermal field dynamic modeling on the plastic mold based on the multi-scale characteristic data of the mold state to obtain a mold thermal field evolution model;

[0029] Step S3: collecting historical operation data of the plastic mold to obtain a mold historical operation data set; mining the process parameter correlation of the mold historical operation data set to obtain a mold process parameter correlation network;

[0030] Step S4: Based on the mold thermal field evolution model and the mold process parameter association network, a state prediction model is performed on the plastic mold to obtain a mold state prediction model; based on the mold state prediction model, the process parameters of the plastic mold are optimized in real time according to the mold operation state data stream to obtain a mold process optimization parameter solution;

[0031] Step S5: According to the mold process optimization parameter scheme, the process parameters of the plastic mold are adjusted to obtain a mold control instruction sequence; according to the mold control instruction sequence and the multi-scale feature data of the mold state, a closed-loop control strategy is constructed for the plastic mold to obtain a mold intelligent manufacturing control scheme.

[0032] In this embodiment, a sensor network is first deployed to obtain mold parameters. The mold is accurately modeled using 3D scanning technology, and temperature, pressure and displacement sensors are installed at key locations. These sensors are connected to a central data acquisition system through a wireless communication module. The system is developed based on the LabVIEW software platform and is used to receive and store data from sensors in real time. Using deep learning technology, especially convolutional neural network (CNN), the real-time collected mold operation status data stream is feature extracted to obtain mold process status feature data. These data include key parameters such as temperature distribution, pressure change and physical displacement of the mold. Using spatiotemporal data analysis techniques, such as wavelet transform and time series analysis, the mold process status feature data is hierarchically analyzed to obtain multi-scale feature data of the mold state. These data cover different levels from micro to macro. Based on these multi-scale feature data, finite element analysis (FEA) software, such as ANSYS, is used to perform hierarchical thermal field dynamic modeling on the plastic mold to obtain a mold thermal field evolution model. This model can simulate the thermal behavior of the mold under different production conditions and predict its long-term performance. Collect the historical operation data of the mold, use data mining tools, such as Python's Scikit-learn library, to mine the correlation of process parameters on the data set, and obtain the mold process parameter correlation network. This network reveals the relationship and influence between different process parameters. Combine the mold thermal field evolution model and the process parameter correlation network to build a state prediction model. Use machine learning frameworks, such as TensorFlow, to train the model to predict the state changes of the mold in the future production cycle. Based on the state prediction model, optimize the process parameters in real time to obtain the mold process optimization parameter solution. Use optimization algorithms, such as genetic algorithms, to adjust the process parameters to improve production efficiency and product quality. According to the process optimization parameter solution, adjust the process parameters of the mold to obtain the mold control instruction sequence. Use PLC programming and the Industrial Internet of Things (IIoT) platform to integrate these instructions into the closed-loop control strategy of the mold and build a mold intelligent manufacturing control solution.

[0033] Preferably, step S1 comprises the following steps:

[0034] Step S11: Perform three-dimensional scanning and modeling on the plastic mold to obtain geometric topological feature data of the mold;

[0035] Specifically, a 3D laser scanner can be used to scan the plastic mold in all directions. This scanner can emit laser pulses and measure the time required for the laser reflected from the mold surface to reach the sensor, thereby calculating the exact position of each point on the mold surface. By performing multiple scans at different positions and angles of the mold, the complete geometric shape and size data of the mold can be obtained. The obtained point cloud data is then imported into professional 3D modeling software, such as Autodesk 3ds Max or SolidWorks. In the software, algorithms are used to process the point cloud data, remove noise and fill in missing parts, and finally generate a complete 3D digital model of the mold. This model contains the mold geometric topological feature data, such as edge, vertex and face information.

[0036] Step S12: Acquire a mold historical operation data set, and construct a historical temperature field for the plastic mold based on the mold historical operation data set to obtain a mold historical temperature field;

[0037] Specifically, historical operation data sets can be exported from the control system of the plastic mold. These data are stored in the database, including various parameter records of the mold during the previous production processes, such as temperature, pressure, and cooling time. Use the SQL query language to retrieve these data from the database and export them as files in CSV format. Use data analysis software such as MATLAB or Python with NumPy and Pandas libraries to process these historical data. First, clean the data to remove outliers and incomplete records, and then use statistical analysis methods to identify the performance change trend of the mold under different production conditions. Use finite element analysis (FEA) software such as ANSYS or ABAQUS to import the cleaned data, and simulate the temperature distribution of the mold during the previous production processes based on the three-dimensional model of the mold and historical temperature data. Through simulation, the historical temperature field of the mold (the temperature field distribution diagram of the mold at different time points) can be obtained.

[0038] Step S13: Adaptively configure the sensor network of the plastic mold according to the geometric topological feature data of the mold and the historical temperature field of the mold to obtain a monitoring sensor layout plan, and deploy the sensor network of the plastic mold according to the monitoring sensor layout plan to obtain a mold parameter perception network;

[0039] Specifically, please refer to the sub-steps of step S13 for the detailed implementation process of this embodiment.

[0040] Step S14: collecting the operation parameters of the plastic mold in real time through the mold parameter sensing network to obtain a mold operation status data stream;

[0041] Specifically, a series of sensors, including temperature sensors, pressure sensors, and displacement sensors, can be deployed at key locations of plastic molds. These sensors are connected to a central data acquisition system through wireless communication modules, which is developed based on the LabVIEW software platform. LabVIEW is a graphical programming software that can receive data from sensors in real time and store it in a database. Thermocouples are used as temperature sensors, which are placed in the mold cavity and cooling channel of the mold to monitor the temperature changes of the mold in real time. Pressure sensors are installed in the injection unit to monitor the injection pressure. Displacement sensors are used to monitor the position and speed of mold opening and closing. The data from these sensors are transmitted to the central data acquisition system through a wireless network, and the system records the data in the form of a time series, forming a data stream of the mold operation status. These data streams include real-time values ​​of parameters such as temperature, pressure, and displacement.

[0042] Step S15: Use a preset deep convolutional neural network to perform feature extraction on the mold operation status data stream to obtain mold process state mapping parameters, and perform transfer learning reconstruction on the mold process state mapping parameters to obtain mold process state feature data.

[0043] Specifically, a pre-trained deep convolutional neural network (DCNN) model can be used to model the mold operation status data stream. This DCNN model is built on a deep learning framework such as TensorFlow or PyTorch, and it has been trained to identify and extract features related to the process status of plastic molds. In the specific implementation, the data stream collected in real time is input into the DCNN model. The first layer of the model is a convolution layer, which can extract local features in the data, such as local changes in temperature and pressure. The subsequent pooling layer is used to reduce the spatial dimension of the features while retaining the most important information. The combination of multiple convolutional layers and pooling layers further extracts deeper features. Finally, the fully connected layer maps these features to the process status parameters of the mold, such as cooling efficiency and filling quality. The DCNN model is reconstructed using the transfer learning method. Specifically, a small amount of field data is used to fine-tune the fully connected layer of the model to better adapt it to the characteristics and production conditions of the current mold. In this way, the characteristic data of the mold process status is finally obtained.

[0044] The present invention obtains the geometric topological feature data of the mold through three-dimensional scanning modeling, provides a physical basis for sensor network configuration and process parameter optimization, and effectively utilizes historical temperature field data to analyze the performance changes of the mold under different production conditions, providing support for performance prediction under new conditions. The accuracy and comprehensiveness of data collection are improved by combining the geometric characteristics and historical temperature field of the mold through adaptive sensor network configuration. The application of real-time data collection and deep convolutional neural networks improves the efficiency and accuracy of feature extraction. By reconstructing the process state mapping parameters through transfer learning, it is possible to continuously learn and adapt to new production environments, optimize process parameters, and improve production efficiency and product quality.

[0045] Preferably, step S13 comprises the following steps:

[0046] Step S131: performing curvature feature recognition on the mold geometric topological feature data to obtain mold surface curvature feature data, and performing sensor initial layout topology construction on the plastic mold according to the mold surface curvature feature data to obtain mold sensor network topology data;

[0047] Specifically, computer-aided design (CAD) software, such as SolidWorks or CATIA, can be used to process the mold geometry topology feature data. The built-in curvature analysis tool is used to identify the curvature features of the mold surface. In the specific operation, the curvature calculation command is executed on the three-dimensional model of the mold. The software calculates the curvature value on each surface and displays it on the model in a color-coded manner, where different colors represent different curvature levels. After obtaining the mold surface curvature feature data, the initial layout of the sensor is planned based on these data. In this process, priority is given to arranging sensors in areas with large or drastic curvature changes, because these areas are usually the key points of mold performance. Use topology optimization algorithms, such as genetic algorithms or simulated annealing algorithms, to determine the optimal location of the sensor. These algorithms can minimize the number of sensors while meeting monitoring requirements, thereby reducing costs. In this way, the initial topology of the mold sensor network is constructed, and the mold sensor network topology data containing the location and number of sensors is generated.

[0048] Step S132: applying the information entropy maximization criterion to the mold sensor network topology data to obtain sensor network redundancy distribution data, and adjusting the sensor density of the mold sensor network topology data according to the sensor network redundancy distribution data to obtain a mold sensor network density distribution map;

[0049] Specifically, an information entropy maximization criterion can be first defined, which aims to maximize the amount of information collected by the sensor network while avoiding data redundancy. Using programming tools such as MATLAB or Python, combined with the entropy calculation formula in information theory, the redundancy of the current sensor network topology data is evaluated. In the specific implementation, the information entropy of the data collected by each sensor is first calculated, and then the redundant areas in the network are identified according to the distribution of entropy values. For example, if the data collected by two sensors are highly correlated, the sum of their information entropies will be lower than the sum of the information entropies of two independent sensors. In this way, the redundancy distribution data of the sensor network can be obtained. Adjust the density of sensors according to the redundancy distribution data. In areas with high redundancy, reduce the number of sensors or adjust their positions to reduce redundancy and optimize resource allocation. In areas with less information, increase sensors or improve their layout. In this way, a density distribution map of the mold sensor network can be obtained, which clearly shows the optimal distribution of sensors.

[0050] Step S133: identifying hotspot areas of the mold's historical temperature field to obtain hotspot distribution characteristic data of the temperature field;

[0051] Specifically, data processing software, such as MATLAB or Python, can be used with NumPy and SciPy libraries to analyze the historical temperature field of the mold. In the specific operation, the historical temperature field data is first imported into the software, and image processing technology is applied to identify the hot spots in the temperature field. A temperature threshold is set, and the area above the threshold is marked as a hot spot. Clustering algorithms, such as K-means or DBSCAN, are used to identify abnormal concentrated areas in the temperature distribution. These algorithms can divide the mold surface into multiple clusters based on the similarity of temperature values, among which the cluster with the highest temperature value is identified as a hot spot. In this way, the temperature field hot spot distribution characteristic data can be obtained, which includes information such as the location, size and temperature value of the hot spot area.

[0052] Step S134: performing Bayesian adaptive optimization adjustment on the mold sensor network density distribution map according to the temperature field hot spot distribution characteristic data to obtain a mold sensor network configuration plan;

[0053] Specifically, a Bayesian optimization model can be first defined, which takes the density distribution of the sensor network as input and the monitoring effect (such as prediction accuracy or monitoring coverage) as output. Using Gaussian process (GP) as the core of Bayesian optimization, it is able to predict the output based on historical data and give the uncertainty of the prediction. In each iteration, a new sensor layout configuration is selected based on the prediction and uncertainty of the Gaussian process, and its performance is evaluated in an actual or simulated environment. Based on the evaluation results, the Gaussian process model is updated, and the iterative optimization process continues until the optimal sensor network configuration scheme is found. In this process, special attention is paid to hot spots and the sensor density in these areas is increased to ensure that critical temperature changes can be accurately monitored. Through Bayesian adaptive optimization adjustment, a mold sensor network configuration scheme that comprehensively considers the hot spot distribution characteristics of the temperature field and the monitoring effect is finally obtained.

[0054] Step S135: performing multi-objective constraint verification analysis on the mold sensor network configuration scheme to obtain comprehensive sensor performance evaluation data, and fine-tuning the mold sensor network configuration scheme according to the comprehensive sensor performance evaluation data to obtain a monitoring sensor layout scheme;

[0055] Specifically, please refer to the sub-steps of step S135 for the detailed implementation process of this embodiment.

[0056] Step S136: positioning and installing sensors on the plastic mold according to the monitoring sensor arrangement plan to obtain a monitoring sensor physical network, and online calibrating the monitoring sensor physical network to obtain a mold parameter sensing network.

[0057] Specifically, according to the monitoring sensor layout plan, the sensor positioning and installation can be prepared. Select the sensor type suitable for the plastic mold, such as thermocouples, pressure sensors and displacement sensors, which can monitor the temperature, pressure and movement state of the mold in real time. During the installation process, a high-precision laser rangefinder is used to ensure that the installation position of the sensor meets the design requirements and that each sensor can accurately cover the specified hot spot area. In the specific operation, first mark the installation position of the sensor on the mold surface, and then use a special adhesive or mechanical fixture to firmly install the sensor on the mold. For thermocouples, ensure that its probe is in close contact with the mold surface to obtain accurate temperature readings; for pressure sensors, ensure that they are installed near the injection port to monitor the injection pressure in real time; and displacement sensors are installed at the opening and closing parts of the mold to monitor the movement state of the mold. After completing the installation, connect the sensors to the data acquisition system and perform preliminary functional tests to ensure that all sensors are working properly and can transmit data. Use calibration equipment, such as standard temperature and pressure sources, to verify the accuracy of the sensors. Specifically, the thermocouple is placed in a standard temperature source with a known temperature, and the deviation of its reading from the standard value is observed; for the pressure sensor, a gas or liquid source with a known pressure is used for calibration. Through these calibration processes, the difference between the actual output of each sensor and the standard value is recorded, and corresponding compensation adjustments are made in the data acquisition system. After completing the online calibration, the mold parameter perception network is finally constructed.

[0058] The present invention scientifically arranges the sensor network through curvature feature recognition, comprehensively covers the key areas of the mold, and improves the scientific nature of the sensor layout. By applying the information entropy maximization criterion to optimize the redundancy of the sensor network, the data acquisition efficiency is improved and resource waste is avoided. The sensor density is adjusted according to the redundancy distribution data to enhance the adaptability of the network. At the same time, the key areas are accurately located through hot spot area identification to provide a basis for sensor layout. Through Bayesian adaptive optimization adjustment, the sensor layout can be dynamically adjusted according to the temperature field hotspots, enhancing the key area monitoring capability. Through multi-objective constraint verification analysis, the sensor performance is comprehensively evaluated to ensure the optimality of the configuration scheme, and the fine-tuning based on the performance evaluation data further improves the accuracy and reliability of monitoring. The practicality and accuracy of the sensor network are improved by online calibration monitoring of the sensor entity network.

[0059] Preferably, step S135 includes the following steps:

[0060] Step S1351: performing spatial topological density evaluation on the mold sensor network configuration scheme to obtain sensor space coverage evaluation data, and performing temperature field gradient sensitivity calculation on the mold sensor network configuration scheme to obtain mold temperature field capture rate evaluation data;

[0061] Specifically, professional sensor network analysis software can be used to evaluate the spatial topological density of the mold sensor network configuration scheme. The three-dimensional model of the mold and the layout position of the sensor are imported. The software evaluates the spatial coverage of the sensor by calculating the detection range of the sensor and the coverage of the mold surface. In the specific operation, the detection radius of the sensor is set, and the detection area of ​​the sensor is simulated in the software. The coverage is displayed by color coding, and the uncovered area is displayed in red, the partially covered area is displayed in yellow, and the fully covered area is displayed in green. In this way, the sensor spatial coverage evaluation data can be obtained. Next, use finite element analysis (FEA) software such as ANSYS to import the temperature field data of the mold and set the position of the sensor. The software simulates the change of the temperature field and calculates the response of each sensor to the temperature gradient. Special attention is paid to the sensitive areas of the sensor to temperature changes, that is, the areas with the largest temperature gradient, because these areas are critical to the thermal control of the mold. By simulating the response of the sensor under different temperature gradients, the mold temperature field capture rate evaluation data is finally obtained.

[0062] Step S1352: Evaluate the timing response characteristics of the mold sensor network configuration scheme to obtain sensor sampling delay evaluation data, and predict the network life cycle of the mold sensor network configuration scheme to obtain sensor network timeliness evaluation data;

[0063] Specifically, a data acquisition system can be used to collect sensor data in an actual production environment and analyze its response time. In specific operations, a known temperature or pressure change is made on the mold, and the time required for the sensor to detect this change and respond is recorded. Repeat this test for each sensor and record the sampling delay, which is the time difference from the change to the sensor response. In this way, sensor sampling delay evaluation data can be obtained. Then, based on the sensor's technical specifications and historical maintenance data, the service life and replacement cycle of each sensor are predicted. Use reliability engineering software, such as Weibull++, to analyze the failure rate and life distribution of sensors. By inputting the sensor's usage conditions and environmental factors, the software calculates the expected life and replacement schedule of the sensor network. Through these analyses, the sensor network timeliness evaluation data is finally obtained.

[0064] Step S1353: performing a data transmission stability evaluation on the mold sensor network configuration scheme to obtain sensor reliability evaluation data, and performing a dynamic energy consumption evaluation on the mold sensor network configuration scheme to obtain sensor network energy efficiency evaluation data;

[0065] Specifically, a network analyzer can be used to test the wireless communication link between the sensor and the central data acquisition system. In specific operations, data transmission is simulated at different locations of the mold, and the loss rate, delay and error rate of the data packet are recorded in the data acquisition system. Special attention is paid to the impact of electromagnetic interference or physical obstacles on data transmission during the operation of the mold. Through these tests, the sensor reliability evaluation data is finally obtained, including indicators of communication stability and data integrity. Next, energy consumption analysis software, such as EnergyPlus, is used to simulate the energy consumption of the sensor under different working conditions. In specific operations, the technical parameters of the sensor, such as power consumption and working mode, as well as the operating cycle and ambient temperature of the mold, are entered. The software simulates the energy consumption of the sensor in different time periods and calculates the total energy consumption of the entire network. Through these analyses, the energy efficiency evaluation data of the sensor network is finally obtained, including indicators such as peak energy consumption, average energy consumption and energy efficiency ratio.

[0066] Step S1354: Based on the sensor space coverage evaluation data, the mold temperature field capture rate evaluation data, the sensor sampling delay evaluation data, the sensor network timeliness evaluation data, the sensor reliability evaluation data and the sensor network energy efficiency evaluation data, a multi-objective constraint comprehensive evaluation is performed on the mold sensor network configuration scheme to obtain the sensor performance comprehensive evaluation data;

[0067] Specifically, multi-objective optimization software, such as MATLAB's Multi-Objective Optimization Toolbox, can be used to integrate sensor spatial coverage evaluation data, mold temperature field capture rate evaluation data, sensor sampling delay evaluation data, sensor network timeliness evaluation data, sensor reliability evaluation data, and sensor network energy efficiency evaluation data. In the specific operation, multiple evaluation indicators are defined, such as coverage, capture rate, sampling delay, timeliness, reliability, and energy efficiency, and weights are assigned to each indicator. Use the weighted summation method or the analytic hierarchy process (AHP) to integrate these indicators and evaluate the overall performance of different sensor configuration schemes. Through comprehensive evaluation, the comprehensive evaluation data of sensor performance is finally obtained.

[0068] Step S1355: weighting the comprehensive evaluation data of sensor performance to obtain a sensor network index optimization weight coefficient;

[0069] Specifically, all evaluation indicators can be listed first, including sensor spatial coverage, temperature field capture rate, sampling delay, network timeliness, reliability and energy efficiency. These indicators are compared in pairs and scored according to their impact on mold performance. For example, temperature field capture rate and reliability are considered to be the most important. Use AHP software tools, such as ExpertChoice, to input these pairwise comparison matrices and calculate the relative weight of each indicator. In this way, the optimized weight coefficients of sensor network indicators can be obtained, which reflect the relative importance of different evaluation indicators in the comprehensive evaluation.

[0070] Step S1356: performing multi-objective iterative optimization calculation on the sensor network index optimization weight coefficient to obtain a mold sensor network configuration optimization solution;

[0071] Specifically, a genetic algorithm (GA) can be used, which is a search algorithm that simulates the natural selection process and is used to solve multi-objective optimization problems. In the specific operation, the parameters of the sensor network configuration, such as sensor type, quantity, location, and working mode, are first defined. Then, the weight coefficients are optimized according to the sensor network indicators, and a fitness function is constructed, which integrates all evaluation indicators. Genetic algorithm software tools, such as MATLAB's genetic algorithm toolbox or Python's DEAP library, are used to initialize a sensor network configuration population and set the algorithm parameters, such as crossover rate, mutation rate, and number of iterations. In each iteration, the algorithm evaluates the fitness of the current population and selects the best individuals for crossover and mutation to generate a new population. This process is repeated until the stop condition is met, such as reaching a predetermined number of iterations or fitness threshold. In this way, an optimized sensor network configuration solution is finally obtained.

[0072] Step S1357: verify the feasibility of and calibrate the mold sensor network configuration optimization plan to obtain a monitoring sensor layout plan.

[0073] Specifically, the sensors can be arranged at predetermined positions on the mold according to the mold sensor network configuration optimization plan. Use high-precision positioning equipment, such as CMM (coordinate measuring machine) or laser positioning system, to ensure that the sensor installation position is accurate. After installation, the feasibility of the sensor network is verified by connecting the sensor to the data acquisition system and performing a series of functional tests. In the functional test, different production conditions are simulated, including different temperatures, pressures and injection speeds, to verify the performance of the sensor network under various conditions. The output data of the sensor is recorded and compared with the expected results to evaluate the performance of the sensor. If any deviation is found, the sensor is fine-tuned, including repositioning, replacement or calibration, to ensure the accuracy of the data. Real-time production tests can also be performed to use the mold in actual injection molding production and monitor the performance of the sensor network. Special attention is paid to the real-time and accuracy of the data, as well as the responsiveness of the sensor network to changes in the production process. Through these tests, data from actual operation is collected and the sensor network configuration is finally calibrated based on this data. During the calibration process, professional calibration equipment, such as standard thermometers and pressure gauges, is used to accurately correct the sensor readings. The calibration parameters of each sensor are recorded and adjusted accordingly in the data acquisition system to ensure that the sensor network can provide accurate and reliable data. Finally, a verified and calibrated monitoring sensor layout plan is obtained.

[0074] The present invention ensures that the sensor network can comprehensively and effectively monitor the mold status by comprehensively evaluating the sensor layout, while improving the response speed and long-term stability of the sensor network. Through the evaluation of the timing response characteristics, the mold state changes can be captured in time, while the data transmission stability and sensor reliability evaluation ensure the stable transmission of data and reduce the risk of data loss and errors. Through the dynamic evaluation of energy consumption, energy consumption is optimized, energy efficiency is improved, and operating costs are reduced. Through multi-objective constraint comprehensive evaluation and weight adjustment, the performance of the sensor network can be refined according to different production requirements and conditions. Through iterative optimization calculation, the sensor network configuration is continuously optimized to achieve the best performance balance. Finally, through feasibility verification and calibration, the actual application effect of the sensor network configuration scheme is ensured, the accuracy and reliability of monitoring are improved, thereby improving the intelligence and automation level of the entire production process.

[0075] Preferably, step S2 comprises the following steps:

[0076] Step S21: performing time domain decomposition on the mold process state feature data to obtain a mold state time domain feature sequence;

[0077] Specifically, data processing software, such as MATLAB or Python with Pandas library, can be used to perform time domain decomposition on the mold process state characteristic data. First, the data is sorted according to the time series, and then the time-frequency analysis methods such as Fourier transform (FFT) or short-time Fourier transform (STFT) are applied to convert the time domain signal into a frequency domain signal, thereby extracting the periodic and instantaneous characteristics. For example, it is found that the temperature change of the mold during the injection molding process shows periodic fluctuations. Through FFT analysis, the frequency and amplitude of these fluctuations can be obtained to form the time domain characteristic sequence of temperature. Similarly, the time domain characteristic sequences of pressure and displacement can also be obtained. These sequences include not only the basic statistical characteristics of each parameter (such as mean, variance and peak value), but also their dynamic change characteristics (such as rise time, fall time and periodic fluctuations). The obtained characteristic sequences are integrated to finally obtain the mold state time domain characteristic sequence.

[0078] Step S22: performing spatial domain grid division on the mold process state characteristic data to obtain a mold state spatial domain characteristic grid;

[0079] Specifically, a 3D modeling software such as SolidWorks or CATIA can be used to build a 3D model of the mold, and it can be divided into several small areas according to the geometric structure of the mold to form a grid. Each grid point represents a virtual measurement point, and the corresponding process state parameters are extracted from the data acquisition system according to the position of these points. The spatial domain grid division of the mold process state characteristic data is performed using data processing software such as MATLAB or Python with the NumPy library. The data of each grid point is interpolated to estimate the continuous distribution on the mold surface. For example, the discrete measurement point data is converted into a continuous temperature field, pressure field and displacement field of the entire mold surface using methods such as nearest neighbor interpolation, linear interpolation or spline interpolation. In this way, the spatial domain characteristic grids of the mold state are finally obtained, which clearly show the distribution of the process state parameters of the mold in different areas. These spatial domain characteristic grids include not only the average and extreme values ​​of each parameter on the mold, but also their gradients and distribution patterns.

[0080] Step S23: performing tensor fusion on the mold state time domain feature sequence and the mold state space domain feature grid to obtain multi-scale feature data of the mold state;

[0081] Specifically, data science tools can be used, such as Python with TensorFlow or PyTorch libraries, which provide tensor operation functions and allow data of different dimensions to be merged into a multidimensional array (tensor). For example, the time domain feature sequence of the mold state has time domain feature sequences of temperature, pressure and displacement, each of which contains data of hundreds of time points; at the same time, there are corresponding spatial domain feature grids in the mold state spatial domain feature grid, and each grid contains data of different areas of the mold. These data are integrated into a three-dimensional tensor, where the first dimension represents different parameters (temperature, pressure, displacement), the second dimension represents the time point, and the third dimension represents the spatial grid point. In this way, the detailed characteristics of the mold state at different time points and different spatial locations can be captured. Tensor decomposition techniques, such as singular value decomposition (SVD) or tensor network decomposition (TNF), are applied to extract multi-scale features. These techniques can identify key patterns and trends in the data and decompose them into a set of core multi-scale features, and finally obtain multi-scale feature data of the mold state.

[0082] Step S24: performing hierarchical heat conduction modeling on the plastic mold based on the multi-scale feature data of the mold state to obtain a hierarchical heat conduction model of the mold;

[0083] Specifically, finite element analysis (FEA) software, such as ANSYS or ABAQUS, can be used to establish a heat conduction model for the mold. Import the three-dimensional geometric model of the mold, and define the material properties and boundary conditions based on the multi-scale characteristic data of the mold state. For example, set different thermal conductivities based on the temperature distribution data, or set specific boundary conditions based on the location of the cooling channel. Use the time domain and space domain information in the multi-scale characteristic data to set the initial conditions and loading conditions of the model. For example, set the initial temperature distribution of the model based on the time domain characteristic sequence of the temperature, or adjust the boundary conditions of the model based on the multi-scale characteristics of the displacement. In this way, the heat conduction behavior of the mold in the actual production process can be simulated. Run the finite element analysis, and the software will calculate the parameters such as the temperature distribution, heat flux density, and thermal stress of the mold at different time points and in different regions. Adjust the parameters of the model based on these results until the model can accurately predict the actual heat conduction behavior of the mold. Finally, a mold layered heat conduction model is obtained, which can describe the heat conduction characteristics of the mold at different levels.

[0084] Step S25: performing boundary condition constraint analysis on the mold layered heat conduction model to obtain mold thermal field boundary characteristic data, and performing parameter correction on the mold layered heat conduction model according to the mold thermal field boundary characteristic data to obtain a mold thermal field evolution model.

[0085] Specifically, please refer to the sub-steps of step S25 for the detailed implementation process of this embodiment.

[0086] The present invention enhances the feature recognition capability through time domain decomposition and space domain grid division, improves the dimension and depth of feature recognition, and improves the data fusion efficiency by using tensor fusion technology. Based on these data, hierarchical heat conduction modeling can more accurately simulate the heat conduction process of plastic molds, thereby improving the accuracy and practicality of the model. The model parameters are optimized through boundary condition constraint analysis and parameter correction, so that the model is closer to the thermal field evolution in actual production, thereby improving the accuracy of model prediction. By optimizing process parameters and improving the accuracy of model prediction, it helps to reduce energy waste and material loss in production, reduce production costs, and at the same time improve the reliability and stability of the system in the face of production changes, and reduce the risk of production interruption.

[0087] Preferably, step S25 comprises the following steps:

[0088] Step S251: performing boundary temperature field distribution identification on the mold layered heat conduction model to obtain mold surface temperature gradient data, and constructing temperature boundary condition constraints on the mold layered heat conduction model according to the mold surface temperature gradient data to obtain a mold temperature boundary constraint model;

[0089] Specifically, finite element analysis (FEA) software, such as ANSYS, can be used to process the heat conduction model of the mold. First, the three-dimensional geometric model and material properties of the mold are imported, and then applied to set the initial conditions and boundary conditions of the model. Run the heat conduction simulation, and the software calculates the temperature distribution of the mold at different time points. After the simulation is completed, the node temperature data on the mold surface is extracted, and the temperature difference between adjacent nodes is calculated to obtain the temperature gradient. Use the software's post-processing tools to visualize the temperature distribution and gradient, usually in the form of isotherms or temperature cloud maps. By analyzing these charts, it is possible to identify the areas with the most significant temperature changes, that is, the areas with the largest temperature gradients. According to the mold surface temperature gradient data, the temperature boundary condition constraints are constructed for the mold layered heat conduction model. Set boundary conditions, such as fixed temperature or heat flow, especially in the areas with the largest temperature gradients. In this way, the mold temperature boundary constraint model is finally obtained.

[0090] Step S252: performing heat flux density distribution identification on the mold layered heat conduction model to obtain mold heat flow boundary characteristic data, and performing heat flow boundary condition constraint construction on the mold layered heat conduction model according to the mold heat flow boundary characteristic data to obtain a mold heat flow boundary constraint model;

[0091] Specifically, finite element analysis (FEA) software can be used to process the mold heat conduction model. After completing the simulation of the temperature field, extract the heat flux density data, which describes the heat flux intensity and direction at each point on the mold surface. Use the software's post-processing tools to visualize the heat flux density distribution and display it in the form of a heat flow line diagram or a heat flow cloud diagram. By analyzing these charts, it is possible to identify the areas with the highest heat flux density, that is, the hot spots of the mold, and the areas where the heat flux density changes most dramatically. Based on these heat flow boundary characteristic data, the heat flow boundary condition constraints are constructed for the mold layered heat conduction model. Set boundary conditions, such as heat flow input or thermal resistance, especially in areas with the highest heat flux density. In this way, the mold heat flow boundary constraint model is finally obtained.

[0092] Step S253: extracting comprehensive boundary condition features of the mold layered heat conduction model based on the mold temperature boundary constraint model and the mold heat flow boundary constraint model to obtain mold thermal field boundary feature data;

[0093] Specifically, the mold temperature boundary constraint model and the mold heat flow boundary constraint model can be loaded in the finite element analysis (FEA) software, such as ANSYS. Next, run a new heat conduction simulation, this time considering the influence of both the temperature boundary and the heat flow boundary. After the simulation is completed, extract the boundary feature data in the simulation results, including the temperature distribution on the boundary, heat flux density, heat flow direction, and thermal resistance on the boundary. Use the post-processing tools of the software to visualize these data and display them in the form of charts and graphs, such as temperature cloud maps, heat flow line diagrams, and thermal resistance distribution diagrams. By analyzing these charts, the key boundary features of the mold thermal field can be identified, such as the area with the largest temperature gradient, the area with the highest heat flux density, and the area with the largest thermal resistance, and finally the mold thermal field boundary feature data is obtained.

[0094] Step S254: performing boundary condition parameter sensitivity identification on the mold thermal field boundary feature data to obtain model parameter sensitivity distribution data;

[0095] Specifically, parameter sensitivity analysis tools, such as the parameter study function in FEA software or Python's Salib library, can be used to evaluate the sensitivity of the boundary characteristic data of the mold thermal field to changes in model parameters. Select some key model parameters, such as the thermal conductivity, heat capacity and heat flux density on the boundary of the material, and then set a series of parameter change ranges. For each parameter, run a series of heat conduction simulations, changing only one parameter value in each simulation while keeping the other parameters unchanged. Record the results of each simulation, especially the boundary characteristic data of the mold thermal field, such as changes in temperature distribution and heat flux density. By comparing the simulation results under different parameter changes, it is possible to identify which parameters are most sensitive to the boundary characteristic data of the mold thermal field, record the identified sensitive parameters, and obtain the model parameter sensitivity distribution data. For example, it is found that the thermal conductivity of the material has a significant effect on the temperature distribution of the mold, while the change in heat flux density has little effect on the heat flux density distribution.

[0096] Step S255: constructing a key parameter adjustment strategy for the mold layered heat conduction model according to the model parameter sensitivity distribution data to obtain a model parameter optimization solution;

[0097] Specifically, the parameter sensitivity distribution data can be analyzed to identify the parameters that have the greatest impact on the mold thermal field boundary characteristic data. For example, it is found that the thermal conductivity and heat flux density of the mold material are the key parameters that affect the temperature distribution. Use this data to determine which parameters need to be adjusted, as well as the direction and range of adjustment. Use optimization algorithms, such as genetic algorithms or particle swarm optimization (PSO), to build adjustment strategies in finite element analysis (FEA) software. Define an objective function, such as minimizing the deviation of temperature distribution or maximizing thermal efficiency, and set constraints, such as physical limitations of material properties and operating ranges of process parameters. Use optimization software tools, such as MATLAB's optimization toolbox or Python's SciPy library, to execute the optimization algorithm. At the beginning of the algorithm, an initial set of parameter values ​​is provided, and the algorithm uses iterative search to find the optimal parameter combination that can make the objective function reach the optimal value. In this way, the optimization solution of the model parameters can be obtained, including the adjusted parameter values ​​such as thermal conductivity and heat flux density.

[0098] Step S256: Parameter adjustment calculation is performed on the model parameter optimization scheme according to the mold thermal field boundary characteristic data to obtain model parameter correction result data, and parameters of the mold layered heat conduction model are updated according to the model parameter correction result data to obtain the mold thermal field evolution model.

[0099] Specifically, the model parameter optimization scheme can be applied to the mold layered heat conduction model in the finite element analysis (FEA) software. Update the material properties and boundary conditions of the model to reflect the optimized parameter values. Run the updated model for heat conduction simulation. After the simulation is completed, extract the simulation results, especially the boundary characteristic data of the mold thermal field, such as temperature distribution and heat flux density, and compare them with the actual production data or design goals. Calculate the deviation between the simulation results and the target, and adjust the model parameters based on these deviations. Use parameter estimation techniques, such as least squares or Bayesian updating, to calculate the corrected values ​​of the parameters. Finally, update the parameters of the mold layered heat conduction model based on the parameter correction result data. This means re-entering the corrected parameter values ​​into the model and running the simulation again until the model can accurately predict the actual thermal field behavior of the mold. In this way, a mold thermal field evolution model is finally obtained, which can describe the thermal field changes of the mold at different times and under different conditions.

[0100] The present invention significantly improves the temperature control accuracy by identifying the temperature gradient data of the mold surface and constructing a temperature boundary constraint model. By identifying the heat flux density distribution and constructing a heat flux boundary constraint model, the heat flux management is optimized, energy waste is reduced and production efficiency is improved. The adaptability of the model to different thermal field boundary conditions is enhanced by comprehensive feature extraction, and the applicability and flexibility of the model in a variable production environment are improved. The sensitivity analysis capability of the model parameters is improved by sensitivity identification. The dynamic optimization of the model parameters is achieved by constructing a key parameter adjustment strategy, which improves the prediction accuracy and response speed of the model. Parameter adjustment calculation and parameter correction improve the reliability and stability of the model. Finally, the acquisition of the mold thermal field evolution model enhances the prediction capability of the production process.

[0101] Preferably, step S3 comprises the following steps:

[0102] Step S31: collecting historical operation data of the plastic mold to obtain a mold historical operation data set;

[0103] Specifically, you can access the mold's control system and data loggers, which store detailed parameters for each production run. Use dedicated data export tools, such as OPC UA clients or Modbus protocol tools, to extract data from the control system. These tools can communicate with the PLC (Programmable Logic Controller) to read and export historical data at specified time intervals. The exported data is usually saved in CSV or Excel format, including timestamps, temperature readings, pressure readings and other relevant parameters. Through these steps, you can get a data set containing the mold's historical operation data.

[0104] Step S32: extracting time series features from the mold historical operation data set to obtain a mold operation data time series feature set;

[0105] Specifically, data analysis software, such as Python with Pandas and NumPy libraries, can be used to load and process mold historical operation data sets. Sort the data in chronological order. Next, extract the time series features, including the basic statistical features of each parameter (such as mean, variance, maximum and minimum values), as well as more complex time series features (such as trends, seasonality, and periodicity). For example, it is found that the temperature of the mold shows a clear downward trend during the cooling stage, or the pressure readings show periodic fluctuations within a specific time period. Use time series analysis techniques, such as the autoregressive moving average (ARMA) model or the long short-term memory network (LSTM), to identify these time series features. Through these analyses, the time series feature set of the mold operation data is finally obtained, including the time series change characteristics of parameters such as temperature and pressure.

[0106] Step S33: performing a time series causal test on the mold operation data time series feature set to obtain parameter time series causal relationship data, and constructing a time series causal association graph based on the parameter time series causal relationship data to obtain a process parameter time series causal network;

[0107] Specifically, causal analysis toolkits in statistical software or programming languages ​​(such as R or Python), such as Granger causality test tools, can be used to analyze the timing feature set of mold operation data. Select two key parameters, such as mold temperature and cooling time, and perform Granger causality test. This test can determine whether one parameter can predict the future value of another parameter, thereby revealing the causal relationship between them. For example, it was found that changes in mold temperature statistically significantly predicted changes in cooling time, indicating that temperature is the Granger cause of cooling time. Record all such causal relationships and use graphical tools such as Gephi or Cytoscape to construct a timing causal association graph. In this graph, each node represents a parameter, and the directed edges between nodes represent timing causal relationships. In this way, a comprehensive process parameter timing causal network can be obtained, which shows the dynamic relationship and interaction between mold operation parameters.

[0108] Step S34: performing information entropy transfer identification on the process parameter timing causal network to obtain parameter flow transfer characteristic data;

[0109] Specifically, Python can be used with the NetworkX and Infomap libraries to analyze the information flow and importance of nodes in the network. Information entropy transfer identification can understand the uncertainty and complexity of information flow in the network. Calculate the information entropy of each node in the process parameter timing causal network, which involves measuring the amount of information output by each node and the uncertainty of this information. The edges between nodes are assigned weights to represent the intensity of information flow. Pay special attention to nodes with high information entropy, because these nodes are key hubs in the network and play a decisive role in the information flow of the entire system. In this way, parameter flow transfer feature data are finally obtained, which reveals which parameters play a core role in information flow and how information is propagated in the process parameter timing causal network.

[0110] Step S35: network coupling is performed on the process parameter time series causal network and the parameter flow transfer characteristic data to obtain an initial process parameter association network;

[0111] Specifically, network analysis software, such as UCINET or Python's NetworkX library, can be used to process the process parameter temporal causal network. The parameter flow transfer characteristic data is added as weights to the edges of the network, so that each edge not only represents the causal relationship between the two parameters, but also the intensity of the information flow. By adjusting the weight of the edge to reflect the size of the information entropy, the edge with higher information entropy is given a larger weight, indicating a stronger information flow. This can combine the characteristics of temporal causality and information flow to form a comprehensive network, namely the initial process parameter association network. In this network, the nodes represent the process parameters, and the edges represent the causal relationship between the parameters and the intensity of information flow. Further analysis of this network identifies which parameters have a strong influence on other parameters and how the information is propagated in the network. In this way, the initial process parameter association network is finally obtained.

[0112] Step S36: performing a key path search on the initial process parameter association network to obtain a key path set of the association network; and reconstructing the topological structure of the initial process parameter association network according to the key path set of the association network to obtain a mold process parameter association network.

[0113] Specifically, algorithms in graph theory, such as the Dijkstra algorithm or the Floyd-Warshall algorithm, can be used to search for critical paths in the initial process parameter association network. These algorithms can find the shortest paths between nodes in the network, which represent the direct and indirect relationships between parameters. The shortest paths between all pairs of nodes are recorded, and those nodes that appear in multiple paths are identified, which are the key nodes in the network. Edges shared in multiple shortest paths can also be identified, which are the key edges in the network. These key nodes and edges constitute the critical path set of the association network. The topological structure of the initial process parameter association network is reconstructed according to the critical path set. The key nodes and edges are retained, while those nodes and edges that have little impact on the connectivity of the network are removed. Finally, a more compact and efficient network is obtained, namely the mold process parameter association network. In this network, the key nodes and edges are highlighted, which represent the core relationships between the mold process parameters.

[0114] The present invention provides data-driven decision support for process parameter adjustment by collecting and analyzing historical mold operation data. The understanding of mold operation time dependency is deepened through time series feature extraction, and the information of mold performance changes is enriched. The causal relationship between parameters is clarified through time series causal verification and network construction, which is crucial for predicting and controlling mold performance. The transfer characteristics of parameter flow are optimized through information entropy transfer identification, which improves information processing efficiency and quality. The system's understanding of the complex relationship between process parameters is enhanced through network coupling. The robustness and predictive ability of the network are enhanced through critical path search and topological structure reconstruction.

[0115] Preferably, step S4 comprises the following steps:

[0116] Step S41: extracting thermal evolution features of the mold thermal field evolution model to obtain a dynamic feature sequence of the mold thermal field;

[0117] Specifically, finite element analysis (FEA) software, such as ANSYS, can be used to simulate the thermal field evolution of the mold during the actual production process. , import the mold thermal field evolution model, input the material properties, boundary conditions and thermal loads of the mold, and then run the thermal analysis module. The software will calculate the temperature distribution of the mold at different time points. After the simulation is completed, the temperature data on the surface and inside of the mold are extracted to form a time series. Use data processing software such as MATLAB or Python with NumPy and SciPy libraries to analyze these time series. Calculate the statistical characteristics of the temperature data, such as the mean, variance, peak and valley values, as well as the rate and acceleration of temperature change. Through these analyses, the dynamic characteristic sequences of the mold thermal field can be obtained. These sequences describe the thermal state changes of the mold during the injection cycle, including temperature changes in the heating, cooling and steady-state stages. In this way, the dynamic characteristic sequence of the mold thermal field is finally obtained.

[0118] Step S42: identifying the process parameter influencing factors of the mold process parameter association network to obtain the process parameter influencing factors;

[0119] Specifically, statistical analysis software, such as the Statsmodels library of R or Python, can be used to analyze the data in the mold process parameter association network. Focus on the key nodes in the network, that is, those key process parameters identified in the network coupling and topological structure reconstruction. Perform regression analysis on these key parameters to identify the extent of their influence on mold performance. For example, pay attention to parameters such as injection temperature, injection speed and cooling time, and their influence on product dimensional accuracy and surface quality. Construct a multivariate linear regression model with product performance as the dependent variable and process parameters as independent variables. Through regression analysis, the coefficients of each process parameter can be obtained, which represent the influencing factors of the process parameters. These influencing factors indicate the relative importance of each parameter change on product performance. In this way, the process parameter influencing factors are finally obtained.

[0120] Step S43: constructing a model based on the dynamic characteristic sequence of the mold thermal field and the influencing factors of the process parameters to obtain an initial state prediction model, and performing transfer learning on the initial state prediction model to obtain a mold state prediction model;

[0121] Specifically, a machine learning framework such as TensorFlow or PyTorch can be used to combine the dynamic feature sequence of the mold thermal field and the influencing factors of the process parameters to build an initial state prediction model. This model is a deep learning model, such as a long short-term memory network (LSTM) or a convolutional neural network (CNN), which is used to capture complex patterns and trends in time series data. The mold thermal field data and corresponding process parameters within a certain period are collected and input into the model as a training set. Through training, the model learns the mapping relationship between thermal field features and process parameters. Use transfer learning technology to optimize this initial model. Transfer learning allows a model pre-trained on a similar task to be adjusted to a specific task. For example, use an LSTM model pre-trained on a general time series prediction task and fine-tune it to the mold state prediction task. The model is adapted to a specific data set by adjusting the model's weights and biases, which involves additional training cycles on small batches of mold data. In this way, an optimized mold state prediction model is finally obtained.

[0122] Step S44: extracting real-time features from the mold operation status data stream to obtain real-time status feature data of the mold, and using the mold status prediction model to perform multi-step prediction on the real-time status feature data of the mold to obtain a mold status prediction result;

[0123] Specifically, real-time analysis algorithms can be applied to extract features from the mold operation status data stream, such as trends, periodicity, and anomaly detection. These real-time status feature data are then input into the mold status prediction model. The model analyzes the input data and outputs a prediction of the mold status in the future. For example, the model can predict the temperature distribution, pressure changes, or potential failure points of the mold in the next few hours. In this way, the prediction results of the mold status can be obtained.

[0124] Step S45: making a process parameter tuning decision for the plastic mold according to the mold state prediction result, obtaining a parameter tuning vector, and solving the parameter tuning vector to obtain an initial value of the process parameter optimization solution;

[0125] Specifically, the process parameters that need to be adjusted can be determined to optimize the mold performance based on the mold state prediction results and combined with the process parameter influencing factors. Multi-objective optimization algorithms, such as genetic algorithms or simulated annealing algorithms, are implemented in Python's SciPy library or MATLAB's optimization toolbox. These algorithms are able to search for the optimal combination of process parameters while considering multiple objectives, such as improving production efficiency, reducing energy consumption, and improving product quality. For example, it is found that the prediction results show that the temperature of the mold in the next production cycle will be higher than the ideal state, which will cause product quality problems. Therefore, the parameters of the cooling system are adjusted to control the mold temperature. This requirement is transformed into an optimization problem, the objective function (such as minimizing temperature deviation) and constraints (such as the operable range of the cooling system) are defined, and then the optimization algorithm is used to solve it. In this way, a parameter tuning vector is finally obtained, which contains the adjusted process parameter values.

[0126] Step S46: verify the constraints of the initial values ​​of the process parameter optimization scheme to obtain the feasibility evaluation results of the scheme, and fine-tune the initial values ​​of the process parameter optimization scheme based on the feasibility evaluation results to obtain the mold process optimization parameter scheme.

[0127] Specifically, the initial values ​​of the process parameter optimization scheme can be used and the impact of these parameter adjustments on mold performance can be simulated in the finite element analysis (FEA) software. Check whether the simulation results meet all process and design constraints, such as the temperature does not exceed the thermal stability limit of the material, and the pressure is within the tolerance range of the mold structure. Use constraint satisfaction analysis tools, such as Python's PuLP library, to evaluate the feasibility of the scheme. This tool can identify which constraints are violated and the extent of the violation. For example, it is found that a certain parameter adjustment causes the temperature of a certain area of ​​the mold to exceed the thermal stability limit of the material, which will affect the quality of the product and the life of the mold. Based on these evaluation results, the initial values ​​of the process parameter optimization scheme are fine-tuned. Increase or decrease the adjustment range of certain parameters, or change the order of parameter adjustment to ensure that all constraints are met. Repeat the simulation and evaluation process until a process parameter scheme that satisfies all constraints and optimizes mold performance is found. In this way, a verified and fine-tuned mold process optimization parameter scheme is finally obtained.

[0128] The present invention significantly improves the accuracy of the mold state prediction model by extracting the dynamic feature sequence of the mold thermal field and identifying the influencing factors of the process parameters, and can more accurately predict the changes in the mold state. By optimizing the initial prediction model using transfer learning, the adaptability of the model to variable production conditions is enhanced. Through real-time feature extraction and multi-step prediction capabilities, it is possible to respond to mold state changes in advance, optimize process parameter adjustments, and provide scientific parameter tuning vectors, thereby improving decision-making efficiency. By verifying the constraints and feasibility of the initial values ​​of the optimization scheme, the practicality of the scheme is ensured. By fine-tuning the process parameter optimization scheme, refined tuning is achieved, which improves production efficiency and product quality, and reduces production defects and waste.

[0129] Preferably, step S5 comprises the following steps:

[0130] Step S51: performing parameter decomposition on the mold process optimization parameter scheme to obtain a process parameter control sequence;

[0131] Specifically, the process parameter optimization scheme is a complex that includes multiple parameter adjustments, such as injection temperature, holding time, and cooling rate adjustments. Use project management software, such as Microsoft Project, or professional process design software, such as Autodesk Moldflow, to decompose these overall schemes. In the software, each optimization parameter is set as a task, and the specific operation steps, required time, and responsible person for each task are defined. For example, raising the injection temperature from 260°C to 270°C is set as a task and is set to be completed before the start of the next production cycle. Extending the holding time from 10 seconds to 12 seconds is another task and is scheduled to be executed after the injection temperature is adjusted. Finally, the complex optimization scheme is decomposed into a series of specific process parameter control sequences. In this way, a detailed process parameter control sequence can be obtained, and each sequence item is a specific operation instruction.

[0132] Step S52: optimizing the timing of the process parameter control sequence to obtain a mold control instruction sequence;

[0133] Specifically, process simulation software, such as Arena from Rockwell Automation, or scheduling algorithms in programming languages, such as the Scheduler library in Python, can be used to simulate the impact of different parameter adjustment sequences on the production process. Input the process parameter control sequence and consider factors such as the capacity limitations of the production equipment, material supply, and worker operation time. Through simulation, evaluate the production efficiency and product quality of different control sequences. For example, it was found that adjusting the injection temperature first and then the holding time would cause production delays, while adjusting the holding time first and then the injection temperature could better utilize the equipment idle time. Optimize the timing of the control sequence based on the simulation results to ensure that each step can be executed at the best time. Convert the optimized control sequence into a mold control instruction sequence. Use a programmable logic controller (PLC) or an industrial Internet of Things (IIoT) platform to program each control step into a specific control instruction. For example, program the adjustment of the injection temperature into a temperature control instruction, and program the adjustment of the holding time into a time control instruction. In this way, a mold control instruction sequence with optimized timing is finally obtained.

[0134] Step S53: reconstructing the multi-scale feature data of the mold state to obtain a mold state evaluation index system, and constructing a mold state feedback mechanism based on the mold state evaluation index system to obtain mold state feedback mechanism data;

[0135] Specifically, data analysis software, such as Python with Pandas and Scikit-learn libraries, can be used to process and analyze multi-scale characteristic data of mold status. These data include information in multiple dimensions such as temperature, pressure, and displacement. Through dimensionality reduction techniques such as principal component analysis (PCA) or factor analysis, these multi-dimensional data are reconstructed into a set of representative state evaluation indicators. For example, the data of multiple temperature sensors are reconstructed into three evaluation indicators: "average temperature", "temperature uniformity" and "hot spot area". We define the calculation method and threshold of each indicator to form a mold status evaluation indicator system. This system can fully reflect the key state characteristics of the mold and provide a basis for real-time monitoring and evaluation of the mold status. Based on this evaluation indicator system, a feedback mechanism for the mold status is constructed. Using the PID (proportional-integral-differential) controller in control theory or a more advanced adaptive control algorithm, feedback control strategies are designed in MATLAB / Simulink or Python's Control Systems Toolbox. These strategies can automatically adjust the process parameters according to the deviation between the actual value and the set value of the mold status evaluation indicator to maintain the stability of the mold status. In this way, the mold status feedback mechanism data is finally obtained, which includes the parameter settings and adjustment logic of the feedback control strategy.

[0136] Step S54: constructing a feedforward control strategy for the mold control instruction sequence to obtain a mold state feedforward control model;

[0137] Specifically, the mold control instruction sequence can be analyzed first to determine which instructions are key control points, such as setting the injection temperature, adjusting the holding time, etc. Use the feedforward control theory, combined with the mold state prediction model, to design the feedforward control strategy. Use MATLAB / Simulink or Python's Control Systems Toolbox to simulate the feedforward control strategy. In the simulation environment, the predicted data of the mold state is input, and the feedforward control model adjusts the control instructions in advance based on these data to reduce or eliminate the impact of the disturbance on the mold state. For example, if the prediction model predicts that the mold temperature will rise in the next production cycle, the feedforward control model will automatically adjust the parameters of the cooling system in advance to prevent the temperature from being too high. Through simulation and adjustment, the parameters of the feedforward control strategy are optimized to ensure that the control instructions can accurately and timely respond to changes in the mold state. In this way, a feedforward control model of the mold state can be obtained, which can adjust the control instructions in advance based on the predicted data.

[0138] Step S55: constructing a feedback control strategy for the mold state feedforward control model according to the mold state feedback mechanism data to obtain a mold hybrid control strategy;

[0139] Specifically, a hybrid control strategy can be designed using mold state feedback mechanism data and mold state feedforward control models in control system design software, such as Siemens TIA Portal or Rockwell Automation Studio 5000. This strategy combines the predictive nature of feedforward control with the responsiveness of feedback control to achieve more precise control. For example, there is a feedforward control model that predicts mold temperature changes due to raw material batch changes and automatically adjusts the output of the heating element. Now, a feedback control strategy needs to be added that uses real-time temperature data to fine-tune the heating output to ensure that the mold temperature remains at the set point. Configure the PID controller and set the proportional (P), integral (I), and derivative (D) parameters so that the heating output can be automatically adjusted when there is a deviation between the actual temperature and the set temperature. In this way, a hybrid control strategy for the mold is finally obtained, which can simultaneously use predicted data and real-time feedback data to optimize mold state control.

[0140] Step S56: Integrate the closed-loop control system of the mold hybrid control strategy to obtain an initial control solution, and construct a real-time correction mechanism for the initial control solution to obtain a mold intelligent manufacturing control solution.

[0141] Specifically, the mold hybrid control strategy can be integrated into the closed-loop control system of the mold. Use industrial automation software platforms, such as Allen-Bradley FactoryTalk or Siemens WinCC, to achieve this integration. These platforms provide data acquisition, monitoring and control (SCADA) functions, allowing real-time monitoring of mold status and adjustment of process parameters according to the control strategy. Configure the control system and program the feedforward and feedback control logic in the hybrid control strategy into the PLC. Then, perform system testing to verify whether the control strategy can be executed correctly and adjust the parameters to achieve optimal performance. During the test, if it is found that the control response is not sensitive enough or over-adjusted in some cases, the control parameters need to be fine-tuned. In addition, a real-time correction mechanism is built so that the control strategy can be quickly adjusted when the control system detects unexpected disturbances or model prediction errors. Use machine learning algorithms, such as adaptive control or reinforcement learning, to continuously optimize the control strategy to adapt it to changes in the production process. In this way, the mold intelligent manufacturing control solution is finally obtained.

[0142] The present invention realizes refined process parameter control through parameter decomposition, thereby improving the controllability of the production process. The rationality of the process parameter control sequence is ensured through time sequence optimization sorting, thereby improving production efficiency. The mold state evaluation capability is enhanced through the establishment of feature reconstruction and state evaluation index system. The real-time feedback mechanism can quickly respond to mold state changes and improve the response speed. The control accuracy and stability are improved through the combination of feedforward and feedback control. The flexibility of the system under different production conditions is improved through the implementation of hybrid control strategy. The mold intelligent manufacturing control solution is made more complete and reliable through the integration of closed-loop control system.

[0143] Preferably, the present invention further provides an intelligent processing system for plastic mold process parameters, which is used to execute the intelligent processing method for plastic mold process parameters as described above. The intelligent processing system for plastic mold process parameters comprises:

[0144] The sensing and acquisition module is used to deploy a sensor network for the plastic mold to obtain a mold parameter sensing network; to collect the operation parameters of the plastic mold in real time through the mold parameter sensing network to obtain a mold operation status data stream; to extract process status features from the mold operation status data stream to obtain mold process status feature data;

[0145] The feature modeling module is used to perform spatiotemporal data hierarchical analysis on the feature data of the mold process state to obtain multi-scale feature data of the mold state; based on the multi-scale feature data of the mold state, the plastic mold is subjected to hierarchical thermal field dynamic modeling to obtain a mold thermal field evolution model;

[0146] The data mining module is used to collect historical operation data of the plastic mold to obtain a mold historical operation data set; perform process parameter correlation mining on the mold historical operation data set to obtain a mold process parameter correlation network;

[0147] The prediction and optimization module is used to perform state prediction modeling on the plastic mold based on the mold thermal field evolution model and the mold process parameter association network to obtain the mold state prediction model; based on the mold state prediction model, the process parameters of the plastic mold are optimized in real time according to the mold operation state data stream to obtain the mold process optimization parameter solution;

[0148] The control strategy module is used to control the process parameters of the plastic mold according to the mold process optimization parameter plan to obtain the mold control instruction sequence; based on the mold control instruction sequence and the multi-scale feature data of the mold state, a closed-loop control strategy for the plastic mold is constructed to obtain the mold intelligent manufacturing control plan.

[0149] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0150] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. An intelligent processing method for plastic mold process parameters, characterized in that: The following steps are involved: Step S1: deploying a sensor network on the plastic mold to obtain a mold parameter sensing network; The operation parameters of the plastic mold are collected in real time through the mold parameter perception network to obtain the mold operation status data stream; Extracting process state characteristics from the mold operation state data stream to obtain mold process state characteristic data; Step S2: performing spatiotemporal data hierarchical analysis on the mold process state feature data to obtain multi-scale feature data of the mold state; Based on the multi-scale characteristic data of the mold state, the plastic mold is subjected to hierarchical thermal field dynamic modeling to obtain the mold thermal field evolution model; Step S3: collecting historical operation data of the plastic mold to obtain a mold historical operation data set; mining the process parameter correlation of the mold historical operation data set to obtain a mold process parameter correlation network; Step S4: Based on the mold thermal field evolution model and the mold process parameter association network, a state prediction model is performed on the plastic mold to obtain a mold state prediction model; based on the mold state prediction model, the process parameters of the plastic mold are optimized in real time according to the mold operation state data stream to obtain a mold process optimization parameter solution; Step S5: According to the mold process optimization parameter scheme, the process parameters of the plastic mold are adjusted to obtain a mold control instruction sequence; according to the mold control instruction sequence and the multi-scale feature data of the mold state, a closed-loop control strategy is constructed for the plastic mold to obtain a mold intelligent manufacturing control scheme.

2. The intelligent processing method of plastic mold process parameters according to claim 1, characterized in that: Step S1 includes the following steps: Step S11: Perform three-dimensional scanning and modeling on the plastic mold to obtain geometric topological feature data of the mold; Step S12: Acquire a mold historical operation data set, and construct a historical temperature field for the plastic mold based on the mold historical operation data set to obtain a mold historical temperature field; Step S13: Adaptively configure the sensor network of the plastic mold according to the geometric topological feature data of the mold and the historical temperature field of the mold to obtain a monitoring sensor layout plan, and deploy the sensor network of the plastic mold according to the monitoring sensor layout plan to obtain a mold parameter perception network; Step S14: collecting the operation parameters of the plastic mold in real time through the mold parameter sensing network to obtain a mold operation status data stream; Step S15: Use a preset deep convolutional neural network to perform feature extraction on the mold operation status data stream to obtain mold process state mapping parameters, and perform transfer learning reconstruction on the mold process state mapping parameters to obtain mold process state feature data.

3. The intelligent processing method of plastic mold process parameters according to claim 2, characterized in that: Step S13 includes the following steps: Step S131: performing curvature feature recognition on the mold geometric topological feature data to obtain mold surface curvature feature data, and performing sensor initial layout topology construction on the plastic mold according to the mold surface curvature feature data to obtain mold sensor network topology data; Step S132: applying the information entropy maximization criterion to the mold sensor network topology data to obtain sensor network redundancy distribution data, and adjusting the sensor density of the mold sensor network topology data according to the sensor network redundancy distribution data to obtain a mold sensor network density distribution map; Step S133: identifying hotspot areas of the mold's historical temperature field to obtain hotspot distribution characteristic data of the temperature field; Step S134: performing Bayesian adaptive optimization adjustment on the mold sensor network density distribution map according to the temperature field hot spot distribution characteristic data to obtain a mold sensor network configuration plan; Step S135: performing multi-objective constraint verification analysis on the mold sensor network configuration scheme to obtain comprehensive sensor performance evaluation data, and fine-tuning the mold sensor network configuration scheme according to the comprehensive sensor performance evaluation data to obtain a monitoring sensor layout scheme; Step S136: positioning and installing sensors on the plastic mold according to the monitoring sensor arrangement plan to obtain a monitoring sensor physical network, and online calibrating the monitoring sensor physical network to obtain a mold parameter sensing network.

4. The intelligent processing method of plastic mold process parameters according to claim 3 is characterized in that: Step S135 includes the following steps: Step S1351: performing spatial topological density evaluation on the mold sensor network configuration scheme to obtain sensor space coverage evaluation data, and performing temperature field gradient sensitivity calculation on the mold sensor network configuration scheme to obtain mold temperature field capture rate evaluation data; Step S1352: Evaluate the timing response characteristics of the mold sensor network configuration scheme to obtain sensor sampling delay evaluation data, and predict the network life cycle of the mold sensor network configuration scheme to obtain sensor network timeliness evaluation data; Step S1353: performing a data transmission stability evaluation on the mold sensor network configuration scheme to obtain sensor reliability evaluation data, and performing a dynamic energy consumption evaluation on the mold sensor network configuration scheme to obtain sensor network energy efficiency evaluation data; Step S1354: Based on the sensor space coverage evaluation data, the mold temperature field capture rate evaluation data, the sensor sampling delay evaluation data, the sensor network timeliness evaluation data, the sensor reliability evaluation data and the sensor network energy efficiency evaluation data, a multi-objective constraint comprehensive evaluation is performed on the mold sensor network configuration scheme to obtain the sensor performance comprehensive evaluation data; Step S1355: weighting the comprehensive evaluation data of sensor performance to obtain a sensor network index optimization weight coefficient; Step S1356: performing multi-objective iterative optimization calculation on the sensor network index optimization weight coefficient to obtain a mold sensor network configuration optimization solution; Step S1357: verify the feasibility of and calibrate the mold sensor network configuration optimization plan to obtain a monitoring sensor layout plan.

5. The intelligent processing method of plastic mold process parameters according to claim 1, characterized in that: Step S2 includes the following steps: Step S21: performing time domain decomposition on the mold process state feature data to obtain a mold state time domain feature sequence; Step S22: performing spatial domain grid division on the mold process state characteristic data to obtain a mold state spatial domain characteristic grid; Step S23: performing tensor fusion on the mold state time domain feature sequence and the mold state space domain feature grid to obtain multi-scale feature data of the mold state; Step S24: performing hierarchical heat conduction modeling on the plastic mold based on the multi-scale feature data of the mold state to obtain a hierarchical heat conduction model of the mold; Step S25: performing boundary condition constraint analysis on the mold layered heat conduction model to obtain mold thermal field boundary characteristic data, and performing parameter correction on the mold layered heat conduction model according to the mold thermal field boundary characteristic data to obtain a mold thermal field evolution model.

6. The intelligent processing method of plastic mold process parameters according to claim 5, characterized in that: Step S25 includes the following steps: Step S251: performing boundary temperature field distribution identification on the mold layered heat conduction model to obtain mold surface temperature gradient data, and constructing temperature boundary condition constraints on the mold layered heat conduction model according to the mold surface temperature gradient data to obtain a mold temperature boundary constraint model; Step S252: performing heat flux density distribution identification on the mold layered heat conduction model to obtain mold heat flow boundary characteristic data, and performing heat flow boundary condition constraint construction on the mold layered heat conduction model according to the mold heat flow boundary characteristic data to obtain a mold heat flow boundary constraint model; Step S253: extracting comprehensive boundary condition features of the mold layered heat conduction model based on the mold temperature boundary constraint model and the mold heat flow boundary constraint model to obtain mold thermal field boundary feature data; Step S254: performing boundary condition parameter sensitivity identification on the mold thermal field boundary feature data to obtain model parameter sensitivity distribution data; Step S255: constructing a key parameter adjustment strategy for the mold layered heat conduction model according to the model parameter sensitivity distribution data to obtain a model parameter optimization solution; Step S256: Parameter adjustment calculation is performed on the model parameter optimization scheme according to the mold thermal field boundary characteristic data to obtain model parameter correction result data, and parameters of the mold layered heat conduction model are updated according to the model parameter correction result data to obtain the mold thermal field evolution model.

7. The intelligent processing method of plastic mold process parameters according to claim 1, characterized in that: Step S3 includes the following steps: Step S31: collecting historical operation data of the plastic mold to obtain a mold historical operation data set; Step S32: extracting time series features from the mold historical operation data set to obtain a mold operation data time series feature set; Step S33: performing a time series causal test on the mold operation data time series feature set to obtain parameter time series causal relationship data, and constructing a time series causal association graph based on the parameter time series causal relationship data to obtain a process parameter time series causal network; Step S34: performing information entropy transfer identification on the process parameter timing causal network to obtain parameter flow transfer characteristic data; Step S35: network coupling is performed on the process parameter time series causal network and the parameter flow transfer characteristic data to obtain an initial process parameter association network; Step S36: performing a key path search on the initial process parameter association network to obtain a key path set of the association network; and reconstructing the topological structure of the initial process parameter association network according to the key path set of the association network to obtain a mold process parameter association network.

8. The intelligent processing method of plastic mold process parameters according to claim 1, characterized in that: Step S4 includes the following steps: Step S41: extracting thermal evolution features of the mold thermal field evolution model to obtain a dynamic feature sequence of the mold thermal field; Step S42: identifying the process parameter influencing factors of the mold process parameter association network to obtain the process parameter influencing factors; Step S43: constructing a model based on the dynamic characteristic sequence of the mold thermal field and the influencing factors of the process parameters to obtain an initial state prediction model, and performing transfer learning on the initial state prediction model to obtain a mold state prediction model; Step S44: extracting real-time features from the mold operation status data stream to obtain real-time status feature data of the mold, and using the mold status prediction model to perform multi-step prediction on the real-time status feature data of the mold to obtain a mold status prediction result; Step S45: making a process parameter tuning decision for the plastic mold according to the mold state prediction result, obtaining a parameter tuning vector, and solving the parameter tuning vector to obtain an initial value of the process parameter optimization solution; Step S46: verify the constraints of the initial values ​​of the process parameter optimization scheme to obtain the feasibility evaluation results of the scheme, and fine-tune the initial values ​​of the process parameter optimization scheme based on the feasibility evaluation results to obtain the mold process optimization parameter scheme.

9. The intelligent processing method of plastic mold process parameters according to claim 1, characterized in that: Step S5 includes the following steps: Step S51: performing parameter decomposition on the mold process optimization parameter scheme to obtain a process parameter control sequence; Step S52: optimizing the timing of the process parameter control sequence to obtain a mold control instruction sequence; Step S53: reconstructing the multi-scale feature data of the mold state to obtain a mold state evaluation index system, and constructing a mold state feedback mechanism based on the mold state evaluation index system to obtain mold state feedback mechanism data; Step S54: constructing a feedforward control strategy for the mold control instruction sequence to obtain a mold state feedforward control model; Step S55: constructing a feedback control strategy for the mold state feedforward control model according to the mold state feedback mechanism data to obtain a mold hybrid control strategy; Step S56: Integrate the closed-loop control system of the mold hybrid control strategy to obtain an initial control solution, and construct a real-time correction mechanism for the initial control solution to obtain a mold intelligent manufacturing control solution.

10. An intelligent processing system for plastic mold process parameters, characterized in that: The method for intelligently processing the process parameters of a plastic mold according to claim 1, wherein the intelligent processing system for the process parameters of a plastic mold comprises: The sensing and acquisition module is used to deploy a sensor network for the plastic mold to obtain a mold parameter sensing network; to collect the operation parameters of the plastic mold in real time through the mold parameter sensing network to obtain a mold operation status data stream; to extract process status features from the mold operation status data stream to obtain mold process status feature data; The feature modeling module is used to perform spatiotemporal data hierarchical analysis on the feature data of the mold process state to obtain multi-scale feature data of the mold state; based on the multi-scale feature data of the mold state, the plastic mold is subjected to hierarchical thermal field dynamic modeling to obtain a mold thermal field evolution model; The data mining module is used to collect historical operation data of the plastic mold to obtain a mold historical operation data set; perform process parameter correlation mining on the mold historical operation data set to obtain a mold process parameter correlation network; The prediction and optimization module is used to perform state prediction modeling on the plastic mold based on the mold thermal field evolution model and the mold process parameter association network to obtain the mold state prediction model; based on the mold state prediction model, the process parameters of the plastic mold are optimized in real time according to the mold operation state data stream to obtain the mold process optimization parameter solution; The control strategy module is used to control the process parameters of the plastic mold according to the mold process optimization parameter plan to obtain the mold control instruction sequence; based on the mold control instruction sequence and the multi-scale feature data of the mold state, a closed-loop control strategy for the plastic mold is constructed to obtain the mold intelligent manufacturing control plan.

Images