Dynamic demolding intelligent regulation and control method and system for packaging plastic uptake mold

Abstract

The invention discloses a dynamic demolding intelligent regulation and control method and system for a packaging blister mold, and relates to the technical field of intelligent control. The method comprises the following steps: constructing a multi-parameter sensing array, and collecting a blister multi-dimensional sensing parameter stream; according to the target packaging plastic uptake mold, a packaging demolding pressure control model library is built, matching is conducted according to the material characteristic information of the target product package and the packaging demolding pressure control model library, and a target demolding pressure control model is determined; performing demolding control analysis on the plastic uptake multi-dimensional sensing parameter flow, and outputting a demolding pressure gradient curve; and a demolding executing mechanism is adopted for conducting dynamic demolding control and feedback optimization compensation on the target packaging blister mold. The technical problems that in the prior art, real-time pressure matching is difficult to achieve according to the mold structure and material difference in the blister package demolding process, demolding is unstable, and product damage is likely to be generated are solved, and the technical effect that the demolding process stability is improved through dynamic pressure regulation and control driven by multi-dimensional perception is achieved.

Description

Technical Field

[0001] This invention relates to the field of intelligent control technology, specifically to a dynamic demolding intelligent control method and system for packaging blister molds. Background Technology

[0002] Vacuum forming is a common process in packaging manufacturing, and its demolding process directly affects the product's forming quality and appearance integrity. Existing vacuum forming molds generally rely on fixed pressure, fixed rhythm, or experience-based adjustments for demolding control, lacking real-time responsiveness to differences in mold cavity structure, changes in material thermoplastic properties, and environmental fluctuations during the forming process. Especially in packaging products with thin-walled, deep-cavity, or multi-curved surface structures, excessive demolding pressure can easily lead to stretching deformation, edge cracks, or surface whitening, while insufficient demolding pressure can cause defects such as sticking and localized sluggish demolding, affecting production cycle time and yield. Traditional processes lack a comprehensive sensing and dynamic adjustment mechanism for key parameters such as air pressure, mold deformation, material cooling state, and environmental disturbances, making it difficult to achieve precise control of pressure distribution during demolding and resulting in insufficient demolding stability. Summary of the Invention

[0003] This application provides a dynamic demolding intelligent control method and system for packaging blister molds, which solves the technical problem in the prior art that it is difficult to achieve real-time pressure matching according to the differences in mold structure and materials during the demolding process of blister packaging, resulting in unstable demolding and easy product damage.

[0004] The first aspect of this application provides a dynamic demolding intelligent control method for packaging blister molds, the method comprising: A multi-parameter sensing array is constructed, comprising a pressure sensor, a deformation monitoring sensor, a temperature sensor, and an environmental monitoring module. This array collects multi-dimensional sensing parameter flows of the target product packaging's blister molding process. Based on the target packaging's blister mold, a packaging demolding pressure control model library is built. The library is matched with the material properties of the target product packaging to determine the target demolding pressure control model. Based on this model, the multi-dimensional sensing parameter flows are analyzed for demolding control, outputting a demolding pressure gradient curve. A demolding actuator is then used to dynamically control and optimize the demolding of the target packaging's blister mold based on the demolding pressure gradient curve.

[0005] A second aspect of this application provides a dynamic demolding intelligent control system for packaging blister molds, the system comprising: Sensing Array Construction Unit: Constructs a multi-parameter sensing array, including a pressure sensor, a deformation monitoring sensor, a temperature sensor, and an environmental monitoring module. This array collects multi-dimensional sensing parameter flows of the target product packaging's blister molding process. Model Library Matching Unit: Based on the target packaging blister mold, a packaging demolding pressure control model library is built. The model is matched with the material properties of the target product packaging to determine the target demolding pressure control model. Parameter Analysis Unit: Based on the target demolding pressure control model, the multi-dimensional sensing parameter flow of the blister molding process is analyzed for demolding control, and a demolding pressure gradient curve is output. Demolding Control Unit: A demolding actuator is used to perform dynamic demolding control and feedback optimization compensation of the target packaging blister mold based on the demolding pressure gradient curve.

[0006] One or more technical solutions provided in this application have at least the following technical effects or advantages: First, a multi-parameter sensing array is constructed, including a pressure sensor, a deformation monitoring sensor, a temperature sensor, and an environmental monitoring module. This array collects multi-dimensional sensing parameter flows of the target product packaging's blister packaging. Next, based on the target packaging blister mold, a packaging demolding pressure control model library is built. The library is matched with the material properties of the target product packaging to determine the target demolding pressure control model. Then, based on the target demolding pressure control model, the multi-dimensional sensing parameter flow of the blister packaging is analyzed for demolding control, outputting a demolding pressure gradient curve. Finally, a demolding actuator is used to dynamically control and optimize the demolding of the target packaging blister mold based on the demolding pressure gradient curve. This solves the technical problem in existing technologies where real-time pressure matching based on mold structure and material differences is difficult, leading to unstable demolding and potential product damage. It achieves the technical effect of improving the stability of the demolding process through multi-dimensional sensing-driven dynamic pressure regulation. Attached Figure Description

[0007] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0008] Figure 1 A schematic diagram of the dynamic demolding intelligent control method for packaging blister molds provided in this application embodiment; Figure 2 This is a schematic diagram of the dynamic demolding intelligent control system for packaging blister molds provided in an embodiment of this application.

[0009] Explanation of reference numerals in the attached figures: 11, sensing array construction unit; 12, model library matching unit; 13, parameter parsing unit; 14, demolding control unit. Detailed Implementation

[0010] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.

[0011] Example 1, as Figure 1 As shown, this application provides a dynamic demolding intelligent control method for packaging blister molds, wherein the method includes: A multi-parameter sensing array is constructed, which includes a pressure sensor, a deformation monitoring sensor, a temperature sensor, and an environmental monitoring module. The multi-dimensional sensing parameter flow of the blister packaging of the target product is collected through the multi-parameter sensing array.

[0012] In this embodiment, based on the key sensing requirements of the target product packaging during the thermoforming and demolding process, the air pressure changes, mold structure deformation, material temperature evolution, and environmental disturbance factors that need to be monitored are determined. Accordingly, air pressure sensors, deformation monitoring sensors (such as strain gauges and micro-displacement sensing elements), temperature sensors, and an environmental monitoring module for collecting background variables such as humidity, air velocity, and dust concentration are selected. Subsequently, based on the cavity size, stress characteristic areas, and thermal sensitive area distribution of the thermoforming mold, spatial topology modeling is performed on the installation positions of various sensors. Optimal placement is determined through finite element stress simulation and thermal field distribution analysis. Monitoring points and deployment density; after determining the points, the specifications of each sensor are matched and fixedly installed. A synchronous acquisition link between the sensors is realized through embedded wiring, and the multi-parameter sensing array is connected to the real-time data acquisition bus of the thermoforming production line. During production operation, the multi-parameter sensing array collects in parallel the changes in air pressure in the mold cavity, the slight deformation of the mold cavity wall, the temperature gradient during material softening and cooling, and environmental disturbance parameters at a set sampling frequency, forming a continuous multi-dimensional sensing parameter stream for thermoforming, providing a real-time input data basis for demolding pressure analysis and subsequent dynamic demolding control.

[0013] Furthermore, constructing a multi-parameter sensing array includes: Based on the product packaging production requirements, the target for monitoring blister demolding is determined; sensor selection analysis is performed on the target to obtain the type and specifications of the demolding monitoring sensors; array topology design is performed for each sensor type to determine the spatial topology parameters; based on the sensor type, specifications, and spatial topology parameters, the sensors are deployed to construct a multi-parameter sensing array.

[0014] First, based on the packaging production requirements of the target product, the potential risks during the thermoforming demolding process, such as mold sticking, localized overpressure, material stretching and whitening, and uneven cooling, are analyzed to determine the monitoring targets for thermoforming demolding and clarify the range of core parameters that need to be monitored in real time. After clarifying the monitoring targets, sensor selection analysis is conducted for monitoring requirements such as air pressure changes, mold wall micro-deformation, material temperature gradient, and environmental disturbances. Considering the cavity structure of the thermoforming mold, heating method, material softening characteristics, and demolding cycle, the type of demolding monitoring sensor that matches the mold is selected, and the specifications such as range, sensitivity, temperature resistance rating, and response speed of each type of sensor are determined. Subsequently, the various... For each type of demolding monitoring sensor, an array topology design was carried out. Through finite element force analysis, thermal field distribution analysis, and airflow path simulation of the mold cavity, the monitoring points that can reflect the key demolding state and the corresponding spatial topology parameters of the demolding monitoring sensors were determined, including the number of points, location layout, sensor spacing, and installation method. Finally, based on the sensor type, sensor specifications, and spatial topology parameters, the layout of the sensors on the mold surface, mold cavity, or peripheral structure was structurally configured. The sensor installation was completed through embedded wiring, high-temperature resistant fasteners, and synchronous acquisition interfaces, constructing a multi-parameter sensing array covering air pressure, deformation, temperature, and environmental disturbances.

[0015] Based on the target packaging blister mold, a packaging demolding pressure control model library is built. The target product packaging material characteristics information is matched with the packaging demolding pressure control model library to determine the target demolding pressure control model.

[0016] In this embodiment, information such as the structural parameters, cavity dimensions, cavity curvature variations, vent distribution, and molding cooling method of the target packaging blister mold is obtained. Combined with the material characteristics of the corresponding product packaging, including key attributes such as material thickness, thermoplastic properties, softening temperature range, cooling rate, and yield stress, the potential stress differences and pressure-sensitive areas during demolding are analyzed. Based on this, demolding process perception data, demolding pressure control data, and demolding effect data collected from historical blister production lines are used as sample sources to conduct training data mining on demolding behavior under different materials, mold structures, and molding conditions, constructing... A packaging demolding pressure control model library is established, where each model corresponds to a specific set of material property parameters and mold structural features. When demolding control is actually executed, the material property information of the target product packaging is matched with the material property identifiers and structural identifiers in the model library using multi-dimensional feature matching. The most suitable demolding pressure control model is retrieved through methods such as material similarity calculation, structural mapping, and pressure response curve fitting. The final determined target demolding pressure control model can output a pressure control strategy that matches the structural characteristics and material mechanical behavior of the target thermoforming mold, which is used to drive the subsequent demolding pressure gradient curve analysis and dynamic demolding control execution.

[0017] Furthermore, a packaging demolding pressure control model library is built, including: Historical data mining is performed on the target packaging blister mold to obtain a blister mold demolding database. The blister mold demolding database includes demolding process perception data, demolding pressure control data, and corresponding demolding effect data for each product packaging type. The blister mold demolding database is then optimized according to the corresponding demolding effect data to obtain a usable blister mold demolding dataset. A packaging material characteristic database is constructed, and demolding pressure is trained on the usable blister mold demolding dataset based on the packaging material characteristic database to build a packaging demolding pressure control model library.

[0018] First, based on the operational records of the target packaging blister mold in the actual production line, historical data mining was performed on its demolding process to collect and organize multi-source data that constitutes the blister mold demolding database. This database includes demolding process perception data such as air pressure, deformation, temperature, and environmental disturbances collected during the demolding process for different product packaging types, as well as corresponding demolding pressure control command data and demolding effect data of the formed product. Subsequently, the finished product quality, demolding integrity, and surface defect status of different demolding samples in the blister mold demolding database were optimized and screened to eliminate those with unstable pressure control. Invalid samples with obvious interference from abnormal working conditions or molding defects are collected to obtain a usable blister mold demolding dataset that can be effectively used for training. Based on this, a packaging material characteristic database is constructed, which includes thermal parameters, mechanical properties, thickness distribution, softening temperature range and cooling shrinkage characteristics of various packaging materials. The usable blister mold demolding dataset is aligned and associated according to material characteristics. The demolding behavior of different material types is modeled by deep learning training methods to generate a model parameter set that can characterize the demolding pressure response law under multiple materials, multiple structures and multiple working conditions, thereby building a packaging demolding pressure control model library.

[0019] Furthermore, based on the packaging material property database, the demolding pressure of the available blister mold demolding dataset is trained to build a packaging demolding pressure control model library, including: Clustering and labeling are performed on the data of each packaging material in the packaging material characteristic database to obtain packaging material characteristic labeling parameters; the available blister mold demolding dataset is classified and integrated according to the packaging material characteristic labeling parameters to obtain blister mold demolding calibration dataset; and a deep neural network is used to train the demolding pressure on the blister mold demolding calibration dataset to build a packaging demolding pressure control model library.

[0020] First, clustering and calibration were performed on various packaging material data in the packaging material property database. Based on key characteristics such as material thickness, thermoplastic range, softening curve, elastic modulus, cooling shrinkage coefficient, and temperature resistance stability, a clustering algorithm was used to divide the feature space of material types, obtaining packaging material property calibration parameters to distinguish the mechanical and thermal behaviors of different materials. After obtaining the material property calibration parameters, the demolding dataset of usable blister molds was classified and integrated according to calibration parameters such as material category, material softening characteristics, pressure sensitivity, and cooling behavior. Demolding process perception data, pressure control data, and demolding effect data corresponding to different materials were then further processed. The process is reorganized to form a thermoforming mold demolding calibration dataset that reflects the demolding process patterns of various material types. Subsequently, based on the demolding calibration dataset of each type of thermoforming mold, a deep neural network training framework is constructed. The perceptual parameter sequence, mold structural features, and historical demolding pressure curves are used as training inputs, and the pressure response strategy with the optimal demolding effect is used as the learning objective. Demolding pressure is trained on calibration datasets of different material categories to obtain a pressure control model that can characterize material differences and changes in the demolding stage. Finally, the trained model is indexed and managed according to the material characteristic calibration parameters to complete the construction of the packaging demolding pressure control model library.

[0021] Furthermore, a deep neural network is used to train the demolding pressure on the demolding calibration dataset of the blister mold, and a packaging demolding pressure control model library is built, including: The thermoforming mold demolding calibration dataset is labeled in stages to obtain a thermoforming stage demolding calibration sample set; a deep neural network is used to train the demolding pressure on the thermoforming stage demolding calibration sample set to obtain a multi-calibration parameter demolding pressure control model set; the multi-calibration parameter demolding pressure control model set is identified and stored according to the packaging material characteristic calibration parameters to build a packaging demolding pressure control model library.

[0022] First, the thermoforming mold demolding calibration dataset was annotated in stages. Based on the material softening degree, mold cavity pressure change trend, and structural stress differences during the thermoforming demolding process, the demolding process was divided into a pre-demolding stage, a main demolding stage, and a post-demolding stage. The corresponding stage's sensing data sequences, historical pressure control data, and effect data were simultaneously annotated to form a thermoforming staged demolding calibration sample set that reflects the characteristics of different demolding stages. Then, a deep neural network training framework was constructed, and the calibration sample sets for each stage were input into the training model. The model learned about air pressure changes, deformation response, temperature gradients, and environmental disturbances. The nonlinear relationship between demolding and release effects is investigated, and key pressure control features at each stage are adaptively extracted. This process trains a set of multi-calibrated parameter demolding pressure control models that can reflect the differences in material properties, demolding stages, and mold structures. After obtaining the multi-calibrated parameter model set, the models are classified and indexed by the calibration parameters of packaging material properties. Each model is organized and stored according to calibration tags such as material category, softening range, thickness parameters, and pressure sensitivity, thus constructing a structured packaging demolding pressure control model library. This enables the system to quickly match and call the corresponding demolding pressure control model based on the target material and mold characteristics.

[0023] Furthermore, the sample set for staged demolding calibration of thermoforming was obtained, including: The demolding process of packaging blister molds is divided into stages to obtain a multi-stage blister demolding control system, which includes a pre-demolding stage, a main demolding stage, and a post-demolding stage. Demolding pressure control analysis is performed on the blister mold demolding calibration dataset according to the multi-stage blister demolding control system to obtain a set of staged calibration demolding pressure control curves. Based on the set of staged calibration demolding pressure control curves, sample annotation is performed on the blister mold demolding calibration dataset to obtain a blister staged demolding calibration sample set.

[0024] First, the demolding process of the packaging blister mold was analyzed for its technological characteristics. Based on the stress change law of the material from softening to cooling and solidification, the pressure release rate of the mold cavity, and the stress-sensitive areas of the product structure during demolding, the overall demolding process was divided into stages, resulting in a multi-stage blister demolding control system including a pre-demolding stage, a main demolding stage, and a post-demolding stage. Subsequently, based on the above stage division, a stage correspondence analysis was performed on the perceived data sequence, historical pressure control curves, and demolding effect records in the blister mold demolding calibration dataset. This was achieved by calculating indicators such as the pressure change slope, structural deformation response, temperature decay rate, and environmental disturbance sensitivity at each stage. The differential features of pressure regulation at different stages are extracted to form a set of staged calibration demolding pressure control curves that can reflect the pressure behavior patterns at each stage. After obtaining the calibration pressure control curves for each stage, the characteristic curves of that stage are used as labels to assign stage labels and delineate sample boundaries for all samples in the thermoforming mold demolding calibration dataset. The perception data, pressure control data, and effect data of different stages are structured, mapped, and labeled to form a set of thermoforming staged demolding calibration samples corresponding to the pre-demolding stage, main demolding stage, and post-demolding stage. This provides a clear staged training data foundation for subsequent demolding pressure training based on deep neural networks.

[0025] Furthermore, a set of multi-calibration parameter demolding pressure control models is obtained, including: The demolding pressure is trained on the staged demolding calibration sample set of the thermoforming using a deep neural network to obtain an initial demolding pressure control model set; loss verification evaluation and iterative parameter tuning are performed on the initial demolding pressure control model set to obtain a multi-calibration parameter demolding pressure control model set.

[0026] First, using a phased demolding calibration sample set for thermoforming as training input, deep neural network training structures were constructed for the pre-demolding, main demolding, and post-demolding stages. The sensory parameter sequences, mold structural features, material property calibration parameters, and corresponding stage target pressure response behaviors for each stage were input into the neural network. By learning the nonlinear mapping relationship between air pressure changes, deformation response, material cooling rate, and environmental disturbance parameters, an initial demolding pressure control model set capable of characterizing the demolding pressure features of different stages was obtained. Subsequently, the initial model set was loss-validated and evaluated based on the validation set. This was achieved by analyzing the predicted pressure curve and... The error distribution, convergence stability, and stage switching smoothness among the calibration pressure curves were analyzed to identify the sources of deviation for each model in specific scenarios. A gradient optimization strategy was then used to iteratively fine-tune the network weights, learning rate, and regularization parameters, enabling the models to more accurately reflect the degree of material softening, structural stress changes, and dynamic characteristics of pressure-sensitive areas. After multiple rounds of iterative training and evaluation, a multi-calibration parameter demolding pressure control model set was finally obtained, covering calibration parameters for different material properties, different mold structural features, and different demolding stages. This provides a reliable model foundation for subsequent indexing and storage based on material properties and generation of dynamic demolding pressure gradient curves.

[0027] Based on the target demolding pressure control model, the demolding control analysis of the multi-dimensional sensing parameter flow of the thermoforming is performed, and the demolding pressure gradient curve is output.

[0028] In this embodiment, the multi-dimensional sensing parameter stream of vacuum forming, such as air pressure changes, mold wall micro-deformation, material temperature gradient, and environmental disturbances, collected in real time by a multi-parameter sensing array, is input into a target demolding pressure control model that matches the current material properties and mold structure. The model performs feature extraction and time-series correlation analysis on the multi-dimensional parameter stream to identify the degree of material softening, the stress state of the mold cavity, cooling and shrinkage behavior, and dynamic changes in pressure-sensitive areas during demolding. Based on this, the model analyzes the real-time sensing parameters according to the learned pressure response law, calculates the optimal demolding pressure change trend under the current working conditions, and generates continuous pressure control results covering the pre-demolding, main demolding, and post-demolding stages. The above control results are fitted and smoothed according to the time series, and finally output a demolding pressure gradient curve that can characterize the pressure distribution, pressure change rate, and stage control differences during the demolding process, providing a target pressure trajectory and real-time adjustment basis for the demolding actuator to implement dynamic demolding control.

[0029] A demolding actuator is used to perform dynamic demolding control and feedback optimization compensation for the target packaging blister mold based on the demolding pressure gradient curve.

[0030] In this embodiment, the demolding pressure gradient curve is sent to the demolding actuator as the target pressure trajectory. The demolding actuator applies a phased, adjustable demolding driving force to the mold cavity based on the pressure rise rate, peak pressure range, and pressure release rhythm at different stages of the curve. This allows for real-time control of the air pressure distribution and structural support within the mold cavity during demolding, gradually bringing the actual demolding pressure closer to the target pressure gradient trajectory. Simultaneously with the demolding operation, the system collects feedback parameters on the thermoforming demolding status, including air pressure feedback, mold deformation feedback, and material surface detachment status, and compares these parameters with the target pressure. The gradient curves are dynamically compared to determine whether there are deviations such as pressure lag, overshoot, abnormal local stress, or uneven demolding. If the above deviations are detected, the target pressure gradient curve is corrected in real time according to the demolding pressure control model and feedback control rules. A compensated pressure gradient curve is generated by adjusting the pressure rise rate, reducing local peak pressure, or extending the stage transition time. The demolding actuator is then re-driven to implement optimized control, so that the demolding process can maintain stable stress and uniform demolding effect in a continuous cycle of monitoring, correction, and adjustment, thereby improving the stability of the demolding process and the molding quality of the target packaged product.

[0031] Furthermore, the demolding actuator performs dynamic demolding control and feedback optimization compensation on the target packaging blister mold based on the demolding pressure gradient curve, including: A demolding actuator is used to dynamically control the demolding of the target packaging blister mold based on the demolding pressure gradient curve, and the blister demolding status feedback parameters are monitored. The demolding pressure gradient curve is optimized and compensated based on the blister demolding status feedback parameters, and the demolding optimization control is performed through the compensated demolding pressure gradient curve.

[0032] First, the demolding pressure gradient curve is input as the target control trajectory to the demolding actuator. The actuator adjusts the air pressure or mechanical release force applied to the vacuum forming mold cavity in real time based on the pressure change rate, peak pressure, and pressure release rhythm at different stages of the pressure gradient curve, achieving dynamic demolding control during the pre-demolding, main demolding, and post-demolding stages. During dynamic control, the system synchronously collects real-time feedback parameters of the vacuum forming demolding status, including instantaneous air pressure within the mold cavity, micro-deformation of the mold wall, demolding contact state of the material surface, and local cooling shrinkage changes. Based on these feedback parameters, the system determines the actual demolding status. The system detects the deviation between the actual demolding process and the target pressure gradient trajectory. When pressure lag, pressure overshoot, uneven demolding, or abnormal local stress are detected, the system optimizes and compensates the original demolding pressure gradient curve based on feedback parameters. By adjusting the pressure rise rate, correcting local pressure peaks, or changing the stage switching time, a compensated demolding pressure gradient curve is generated. The demolding actuator then executes the compensated curve, realizing demolding optimization control based on real-time feedback. This ensures that the demolding process maintains uniform pressure distribution and reasonable material stress through continuous correction and dynamic adjustment, thereby improving the demolding stability and molding quality of the target packaging blister product.

[0033] Furthermore, the demolding pressure gradient curve is optimized and compensated based on the vacuum forming demolding state feedback parameters, including: Based on the packaging blister demolding feedback control logic, a demolding pressure control rule library is constructed; based on the demolding pressure control rule library, the demolding pressure gradient curve is matched and optimized according to the blister demolding state feedback parameters.

[0034] First, based on the typical stress patterns, mechanical response characteristics during material softening and cooling stages, and common demolding anomalies (including pressure hysteresis, pressure overshoot, localized sticking, uneven demolding, and abnormal structural deformation) during the thermoforming demolding process, a thermoforming demolding feedback control logic is established to determine demolding state deviations and corresponding adjustment strategies. On this logic, a demolding pressure control rule base is constructed, and the mapping relationships between different combinations of feedback parameters and pressure correction strategies are structured. During actual demolding control, the system continuously receives thermoforming demolding state feedback parameters and matches them with feature conditions in the rule base. The system identifies deviation patterns reflected in feedback parameters and determines corresponding pressure correction strategies, including adjusting the pressure rise slope, correcting local pressure peaks, extending or shortening stage switching time, suppressing excessively rapid pressure release, or enhancing local support pressure. Subsequently, the original demolding pressure gradient curve is optimized and compensated according to the matched control rules to generate a real-time corrected target pressure gradient curve. This compensated curve drives the demolding actuator to implement pressure regulation, enabling the demolding process to achieve adaptive correction based on real-time feedback, thereby improving the stability and uniformity of the demolding process and the molding quality of the final packaged product.

[0035] In summary, the embodiments of this application have at least the following technical effects: First, a multi-parameter sensing array is constructed, including a pressure sensor, a deformation monitoring sensor, a temperature sensor, and an environmental monitoring module. This array collects multi-dimensional sensing parameter flows of the target product packaging's blister packaging. Next, based on the target packaging blister mold, a packaging demolding pressure control model library is built. The library is matched with the material properties of the target product packaging to determine the target demolding pressure control model. Then, based on the target demolding pressure control model, the multi-dimensional sensing parameter flow of the blister packaging is analyzed for demolding control, outputting a demolding pressure gradient curve. Finally, a demolding actuator is used to dynamically control and optimize the demolding of the target packaging blister mold based on the demolding pressure gradient curve. This solves the technical problem in existing technologies where real-time pressure matching based on mold structure and material differences is difficult, leading to unstable demolding and potential product damage. It achieves the technical effect of improving the stability of the demolding process through multi-dimensional sensing-driven dynamic pressure regulation.

[0036] Example 2, based on the same inventive concept as the dynamic demolding intelligent control method for packaging blister molds in the foregoing examples, such as... Figure 2 As shown, this application provides a dynamic demolding intelligent control system for packaging blister molds, wherein the system includes: Sensing Array Construction Unit 11: Constructs a multi-parameter sensing array, which includes a pressure sensor, a deformation monitoring sensor, a temperature sensor, and an environmental monitoring module. The multi-parameter sensing array collects multi-dimensional sensing parameter flows of the target product packaging's blister molding process. Model Library Matching Unit 12: Based on the target packaging blister mold, builds a packaging demolding pressure control model library. It matches the material properties of the target product packaging with the packaging demolding pressure control model library to determine the target demolding pressure control model. Parameter Analysis Unit 13: Based on the target demolding pressure control model, it performs demolding control analysis on the multi-dimensional sensing parameter flow of the blister molding process and outputs a demolding pressure gradient curve. Demolding Control Unit 14: Employs a demolding actuator based on the demolding pressure gradient curve to perform dynamic demolding control and feedback optimization compensation on the target packaging blister mold.

[0037] Furthermore, the sensing array construction unit 11 is used to perform the following method: Based on the product packaging production requirements, the target for monitoring blister demolding is determined; sensor selection analysis is performed on the target to obtain the type and specifications of the demolding monitoring sensors; array topology design is performed for each sensor type to determine the spatial topology parameters; based on the sensor type, specifications, and spatial topology parameters, the sensors are deployed to construct a multi-parameter sensing array.

[0038] Furthermore, the model library matching unit 12 is used to perform the following method: Historical data mining is performed on the target packaging blister mold to obtain a blister mold demolding database. The blister mold demolding database includes demolding process perception data, demolding pressure control data, and corresponding demolding effect data for each product packaging type. The blister mold demolding database is then optimized according to the corresponding demolding effect data to obtain a usable blister mold demolding dataset. A packaging material characteristic database is constructed, and demolding pressure is trained on the usable blister mold demolding dataset based on the packaging material characteristic database to build a packaging demolding pressure control model library.

[0039] Furthermore, the model library matching unit 12 is used to perform the following method: Clustering and labeling are performed on the data of each packaging material in the packaging material characteristic database to obtain packaging material characteristic labeling parameters; the available blister mold demolding dataset is classified and integrated according to the packaging material characteristic labeling parameters to obtain blister mold demolding calibration dataset; and a deep neural network is used to train the demolding pressure on the blister mold demolding calibration dataset to build a packaging demolding pressure control model library.

[0040] Furthermore, the model library matching unit 12 is used to perform the following method: The thermoforming mold demolding calibration dataset is labeled in stages to obtain a thermoforming stage demolding calibration sample set; a deep neural network is used to train the demolding pressure on the thermoforming stage demolding calibration sample set to obtain a multi-calibration parameter demolding pressure control model set; the multi-calibration parameter demolding pressure control model set is identified and stored according to the packaging material characteristic calibration parameters to build a packaging demolding pressure control model library.

[0041] Furthermore, the model library matching unit 12 is used to perform the following method: The demolding process of packaging blister molds is divided into stages to obtain a multi-stage blister demolding control system, which includes a pre-demolding stage, a main demolding stage, and a post-demolding stage. Demolding pressure control analysis is performed on the blister mold demolding calibration dataset according to the multi-stage blister demolding control system to obtain a set of staged calibration demolding pressure control curves. Based on the set of staged calibration demolding pressure control curves, sample annotation is performed on the blister mold demolding calibration dataset to obtain a blister staged demolding calibration sample set.

[0042] Furthermore, the model library matching unit 12 is used to perform the following method: The demolding pressure is trained on the staged demolding calibration sample set of the thermoforming using a deep neural network to obtain an initial demolding pressure control model set; loss verification evaluation and iterative parameter tuning are performed on the initial demolding pressure control model set to obtain a multi-calibration parameter demolding pressure control model set.

[0043] Furthermore, the demolding control unit 14 is used to perform the following method: A demolding actuator is used to dynamically control the demolding of the target packaging blister mold based on the demolding pressure gradient curve, and the blister demolding status feedback parameters are monitored. The demolding pressure gradient curve is optimized and compensated based on the blister demolding status feedback parameters, and the demolding optimization control is performed through the compensated demolding pressure gradient curve.

[0044] Furthermore, the demolding control unit 14 is used to perform the following method: Based on the packaging blister demolding feedback control logic, a demolding pressure control rule library is constructed; based on the demolding pressure control rule library, the demolding pressure gradient curve is matched and optimized according to the blister demolding state feedback parameters.

[0045] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A dynamic demolding intelligent control method for packaging blister molds, characterized in that, The method includes: A multi-parameter sensing array is constructed, which includes a pressure sensor, a deformation monitoring sensor, a temperature sensor, and an environmental monitoring module. The multi-dimensional sensing parameter flow of the blister packaging of the target product is collected through the multi-parameter sensing array. Based on the target packaging blister mold, a packaging demolding pressure control model library is built. The target product packaging material characteristics information is matched with the packaging demolding pressure control model library to determine the target demolding pressure control model. Based on the target demolding pressure control model, the demolding control analysis is performed on the multi-dimensional sensing parameter flow of the thermoforming process, and the demolding pressure gradient curve is output. A demolding actuator is used to perform dynamic demolding control and feedback optimization compensation for the target packaging blister mold based on the demolding pressure gradient curve.

2. The dynamic demolding intelligent control method for packaging blister molds as described in claim 1, characterized in that, Constructing a multi-parameter sensing array includes: Based on the product packaging production requirements, determine the monitoring targets for blister demolding; Sensor selection analysis was performed on the target of the vacuum forming demolding monitoring to obtain the type and specifications of the demolding monitoring sensor; Array topology design is performed for each sensor type in the demolding monitoring sensor type to determine the spatial topology parameters of the demolding monitoring sensor; Based on the type and specifications of the demolding monitoring sensor, as well as the spatial topology parameters of the demolding monitoring sensor, the sensor is deployed to construct a multi-parameter sensing array.

3. The dynamic demolding intelligent control method for packaging blister molds as described in claim 1, characterized in that, Build a packaging demolding pressure control model library, including: Historical data mining is performed on the target packaging blister mold to obtain a blister mold demolding database. The blister mold demolding database includes demolding process perception data, demolding pressure control data and corresponding demolding effect data for each product packaging type. The vacuum forming mold demolding database is optimized based on the corresponding demolding effect data to obtain a usable vacuum forming mold demolding dataset. A packaging material property database is constructed, and the demolding pressure is trained on the available blister mold demolding dataset based on the packaging material property database to build a packaging demolding pressure control model library.

4. The dynamic demolding intelligent control method for packaging blister molds as described in claim 3, characterized in that, Based on the packaging material property database, the demolding pressure of the available blister mold demolding dataset is trained to build a packaging demolding pressure control model library, including: Clustering and calibration are performed on the data of each packaging material in the packaging material property database to obtain packaging material property calibration parameters; The available blister mold demolding dataset is classified and integrated according to the packaging material characteristic calibration parameters to obtain the blister mold demolding calibration dataset. Deep neural networks were used to train the demolding pressure on the demolding calibration dataset of the blister mold, and a packaging demolding pressure control model library was built.

5. The dynamic demolding intelligent control method for packaging blister molds as described in claim 4, characterized in that, Deep neural networks were used to train the demolding pressure on the demolding calibration dataset of the blister molds, and a packaging demolding pressure control model library was built, including: The vacuum forming mold demolding calibration dataset is subjected to staged demolding annotation to obtain a vacuum forming staged demolding calibration sample set; Deep neural networks were used to train the demolding pressure on the staged demolding calibration sample set of the thermoforming process to obtain a set of demolding pressure control models with multiple calibration parameters. The multi-calibration parameter demolding pressure control model set is identified and stored according to the calibration parameters of the packaging material characteristics, and a packaging demolding pressure control model library is built.

6. The dynamic demolding intelligent control method for packaging blister molds as described in claim 5, characterized in that, The sample set for staged demolding calibration of thermoforming was obtained, including: The demolding process of packaging blister molds is divided into stages to obtain a multi-stage blister demolding control, which includes a pre-demolding stage, a main demolding stage, and a post-demolding stage. According to the multi-stage thermoforming demolding control, the demolding pressure control analysis is performed on the thermoforming mold demolding calibration dataset to obtain a set of staged calibration demolding pressure control curves. Based on the set of staged demolding pressure control curves, the demolding calibration dataset of the thermoforming mold is labeled to obtain the staged demolding calibration sample set of thermoforming.

7. The dynamic demolding intelligent control method for packaging blister molds as described in claim 5, characterized in that, Obtain a set of multi-calibration parameter demolding pressure control models, including: Deep neural networks were used to train the demolding pressure on the phased demolding calibration sample set of the thermoforming process to obtain an initial demolding pressure control model set. The initial demolding pressure control model set is subjected to loss verification evaluation and iterative parameter tuning to obtain a multi-calibration parameter demolding pressure control model set.

8. The dynamic demolding intelligent control method for packaging blister molds as described in claim 1, characterized in that, A demolding actuator is used to perform dynamic demolding control and feedback optimization compensation for the target packaging blister mold based on the demolding pressure gradient curve, including: A demolding actuator is used to dynamically control the demolding of the target packaging blister mold based on the demolding pressure gradient curve, and the blister demolding status feedback parameters are monitored and obtained. The demolding pressure gradient curve is optimized and compensated based on the vacuum forming demolding state feedback parameters, and the demolding optimization control is performed using the compensated demolding pressure gradient curve.

9. The dynamic demolding intelligent control method for packaging blister molds as described in claim 8, characterized in that, The demolding pressure gradient curve is optimized and compensated based on the vacuum forming demolding state feedback parameters, including: Based on the feedback control logic of blister packaging demolding, a demolding pressure control rule library is constructed; Based on the demolding pressure control rule base, the demolding pressure gradient curve is matched and optimized according to the thermoforming demolding state feedback parameters.

10. A dynamic demolding intelligent control system for packaging blister molds, characterized in that, The system is used to implement the dynamic demolding intelligent control method for packaging blister molds according to any one of claims 1-9, the system comprising: Sensing array construction unit: Constructs a multi-parameter sensing array, which includes a pressure sensor, a deformation monitoring sensor, a temperature sensor, and an environmental monitoring module. The multi-parameter sensing array is used to collect multi-dimensional sensing parameter streams of the blister packaging of the target product. Model library matching unit: Based on the target packaging blister mold, build a packaging demolding pressure control model library, and match the target product packaging material characteristics information with the packaging demolding pressure control model library to determine the target demolding pressure control model; Parameter analysis unit: Based on the target demolding pressure control model, the demolding control analysis is performed on the multi-dimensional sensing parameter flow of the vacuum forming, and the demolding pressure gradient curve is output; Demolding control unit: The demolding actuator performs dynamic demolding control and feedback optimization compensation on the target packaging blister mold based on the demolding pressure gradient curve.

Images