Control method for medical infusion bottle cap production, electronic device, and storage medium

Abstract

The application provides a control method for medical infusion bottle cap production, an electronic device and a storage medium, comprising: collecting data of multiple sensors in an injection molding system in real time; constructing a digital twin mapping system comprising a virtual mapping layer, a data layer, a simulation layer and a simulation and analysis layer; inputting the sensor data into the system to generate dynamic simulation simulation data, constructing a first parameter prediction model based on the two types of data through a data-driven modeling algorithm; or using a transfer learning algorithm to construct a second parameter prediction model; based on the above parameter prediction model, using a fuzzy reasoning algorithm to search and determine the optimal production parameters from an experience case set comprising historical data and model deduction data; outputting the optimal parameters to the injection molding system and updating in real time. Through the fusion of digital twin and sensor data, precise mapping and dynamic simulation are realized, the optimal production parameters are determined by combining the parameter prediction model and the fuzzy reasoning algorithm, the production accuracy, stability and efficiency are improved, and the product quality is guaranteed.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to the field of intelligent control, and particularly relates to an intelligent control method for medical infusion bottle cap production, an electronic device and a storage medium. BACKGROUND

[0002] In the medical field, the infusion bottle cap as a key component of the drug packaging directly affects the safety and stability of the drug. In particular, the infusion bottle cap has a very high cooperation precision requirement with the bottle body and the rubber plug, and the size tolerance usually needs to be controlled within ±0.01 mm, and the surface smoothness and the number of chippings are also strictly required, so the production and processing process of the medical infusion bottle cap needs to be strictly controlled.

[0003] At present, the related technology has equipped a visual detection system (such as a laser range finder and the like) and a corresponding visual detection method to detect the defect problems of each batch of products, but most of the time, the products are transported to the visual detection system for secondary product screening after being demolded, and the parameters cannot be monitored and adjusted in real time during the production process. Moreover, the detection precision and accuracy of the related visual detection system and the equipped method are insufficient, so that the infusion bottle cap with defects cannot be identified and is still transported to the next production process, thereby causing waste of cost and even causing irreparable consequences. SUMMARY

[0004] The embodiments of the present disclosure provide an intelligent control method for medical infusion bottle cap production, an electronic device and a storage medium to at least solve the technical problem that the parameters in the processing and production process cannot be controlled in real time in the related technology, and the accuracy or precision of the related visual detection method is not high.

[0005] In a first aspect of the present disclosure, an intelligent control method for medical infusion bottle cap production is provided, which is applied to a medical infusion bottle cap injection molding system including a plurality of sensors. The intelligent control method includes: acquiring data of the plurality of sensors in the medical infusion bottle cap injection molding system in real time; constructing a digital twin mapping system of the medical infusion bottle cap injection molding system, wherein the digital twin mapping system includes a virtual mapping layer, a data layer, a model layer and a simulation and analysis layer; inputting the plurality of sensor data into the digital twin mapping system to obtain dynamic simulation and analysis data, constructing a first parameter prediction model based on the plurality of sensor data and the dynamic simulation and analysis data by using a data-driven modeling algorithm, or constructing a second parameter prediction model by using a transfer learning algorithm; determining optimal production parameters from an experience case set by using a fuzzy reasoning algorithm based on the first parameter prediction model or the second parameter prediction model, wherein the experience case set includes historical data and model deduction data; and outputting the optimal production parameters to the medical infusion bottle cap injection molding system and updating the optimal production parameters in real time.

[0006] In a second aspect of the present disclosure, an intelligent control device for medical infusion production is provided, which is applied to a medical infusion bottle cap injection molding system including a plurality of sensors. The intelligent control method includes: an acquisition module for acquiring data of the plurality of sensors in the medical infusion bottle cap injection molding system in real time; a parameter prediction model generation module for constructing a digital twin mapping model of the medical infusion bottle cap injection molding system, inputting the plurality of sensor data into the digital twin mapping system to obtain dynamic simulation data, and generating a parameter prediction model based on the plurality of sensor data and the dynamic simulation data; an optimal production parameter determination module for determining optimal production parameters based on the parameter prediction model and using a fuzzy reasoning algorithm to search from an experience case set, wherein the experience case set includes historical data and model deduction data; and an output and update module for outputting the optimal production parameters to the medical infusion bottle cap injection molding system and updating in real time. The intelligent control method realizes accurate mapping and dynamic simulation of the medical infusion bottle cap injection molding system through the fusion of digital twin and sensor data, and efficiently determines the optimal production parameters by combining the parameter prediction model and the fuzzy reasoning algorithm, thereby improving the accuracy, stability and efficiency of the medical infusion bottle cap injection molding production and ensuring the quality of the medical infusion bottle cap.

[0007] In a third aspect of the present disclosure, an electronic device is provided, which includes a processor and a memory for storing executable instructions of the processor, wherein the processor is configured to execute the executable instructions to implement the above method.

[0008] In a fourth aspect of the present disclosure, a computer readable storage medium is provided, which stores a computer program, and the computer program is executed by a processor to implement the above method.

[0009] Compared with related technologies, the embodiments of the present disclosure construct a digital twin mapping system corresponding to the medical infusion bottle cap injection molding system, generate a parameter prediction model by using the plurality of sensor data and the dynamic simulation data generated by the digital twin mapping system, determine optimal production parameters from an experience case set based on the parameter prediction model and the fuzzy reasoning algorithm, output the optimal production parameters to the medical infusion bottle cap injection molding system, and update in real time. The method can predict defects (surface defects and chipping risks) of the product after forming in advance, and apply the optimal production parameters to the medical infusion bottle cap injection molding system in real time, thereby improving the accuracy, stability and efficiency of the medical infusion bottle cap injection molding production and ensuring the production quality of the medical infusion bottle cap.

[0010] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and cannot limit the present disclosure. Attached Figure Description

[0011] To more clearly illustrate the technical solutions in the embodiments of this disclosure, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of this disclosure. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1 The sensor composition structure of a medical infusion bottle cap injection molding system provided in some embodiments of this disclosure is shown;

[0013] Figure 2 A flowchart is shown below illustrating an intelligent control method for the production of medical infusion bottle caps provided in some embodiments of this disclosure;

[0014] Figure 3 A flowchart illustrating the generation of a parameter prediction model in an intelligent control method for the production of medical infusion bottle caps provided in some embodiments of this disclosure is shown.

[0015] Figure 4 The flowchart shown is a process for constructing a second parameter prediction model in an intelligent control method for the production of medical infusion bottle caps provided in some embodiments of this disclosure;

[0016] Figure 5 The structure of an intelligent control device for the production of medical infusion bottle caps provided in some embodiments of this disclosure is shown;

[0017] Figure 6 The structure of an electronic device provided in some embodiments of this disclosure is shown. Detailed Implementation

[0018] To better understand the above technical solutions, the technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0019] It should be noted that any variation of the terms "comprising" and "having" is intended to cover inclusion that does not include other components, such as components A, B, and C, but not necessarily those explicitly listed, but may include other components not explicitly listed.

[0020] In addition, the term "and / or" appearing in this document merely describes an associated relationship, which means that there can be three relationships, for example, A and / or B, which can represent three cases: A exists alone, A and B exist together, and B exists alone. In addition, the character " / " generally represents an "or" relationship between the front and rear associated objects.

[0021] Medical infusion bottle caps are a key component of infusion containers. Their main functions are to seal the liquid medicine, maintain a sterile state, facilitate puncture and drug addition, or connect infusion devices. They play a crucial role in ensuring drug safety, effectiveness, and ease of clinical use. Currently, in the production of medical infusion bottle caps, defect detection is mostly done after demolding through visual inspection to screen out qualified products. However, this method cannot adjust processing parameters in real time during production, and the existing visual inspection method has low detection precision and accuracy, resulting in a small number of defective medical infusion bottle caps being produced and delivered to the next process, leading to problems such as poor sealing, leakage, and unqualified chipping in the subsequent production of infusion bottles, causing waste of production costs and even irreversible consequences. Therefore, it is necessary to strictly control the processing of medical infusion bottle caps.

[0022] To solve the above problems, an intelligent control method for medical infusion bottle cap production is proposed in exemplary embodiments of the present disclosure. This method combines the concepts of the Internet of Things and digital twinning by establishing a mapping relationship between physical entities and virtual systems. It analyzes real-time data and historical data during production and uses a parameter prediction model to predict production conditions. It also references historical experience data and can provide timely feedback to the production control center when production parameters are found to be abnormal (which may cause defects in medical infusion bottle caps). This method realizes real-time adjustment of production parameters during the injection molding process and improves production yield. Defects in medical infusion bottle caps include at least surface defects such as uneven surfaces, bubbles, burrs, shrinkage marks, and flash, as well as insufficient chipping during demolding and puncturing.

[0023] In some embodiments, the intelligent control method for medical infusion bottle cap production can be applied to a medical infusion bottle cap injection molding system. As shown in Figure 1 Figure 1 ​An exemplary structure of a medical infusion bottle cap injection system sensor assembly is shown, which includes various sensors for detecting various production data during the production process, and the sensors are connected to the control center 100. Generally, the medical infusion bottle cap injection molding process includes the following four main stages: filling, pressure maintaining, cooling, and demolding. In the filling stage, the screw moves forward to push the molten material (melt) into the mold cavity; in the pressure maintaining stage, the screw maintains the pressure, and the mold is filled with material; in the cooling stage, the molded parts are sufficiently cooled by the cooling liquid flow; finally, in the demolding stage, the mold is opened, and the molded parts are ejected. In order to cover the detection data of each stage of the injection molding process, in some embodiments, the various sensors can include: a piezoelectric pressure sensor 101 arranged on the nozzle / mold cavity runner for detecting the pressure of the molten material, such as Kistler 6183A; an infrared temperature sensor 102 arranged at the nozzle outlet for detecting the melt temperature, such as Optris CTlaser 3M; an embedded PT100 thermocouple 103 arranged on the mold core and cooling water channel for detecting the mold temperature, such as OMEGA F2020; a strain gauge sensor 104 arranged on the pull rod / moving die plate for measuring the clamping force, such as HBM U9C; a magnetostrictive displacement sensor 105 arranged on the injection cylinder for measuring the screw position, such as MTS Temposonics R series, etc. In addition, the injection molding system can also include an online visual detection device 106 arranged on the post-ejection conveying belt for detecting the size of the bottle cap and other information, such as Keyence CV-X series; and a microscopic high-speed camera and AI image processing device 107 arranged in the embedded observation window in the mold cavity for observing the surface condition and chipping of the bottle cap, such as Phantom TMX 7510.

[0024] In some embodiments, as shown in Figure 2 Figure 2 An intelligent control method for medical infusion bottle cap production is shown, which includes steps S101-S105.

[0025] Step S101, real-time acquisition of various sensor data in the medical infusion bottle cap injection molding system.

[0026] In some embodiments, in this step, the key sensor data in the medical infusion bottle cap injection molding system is acquired in real time, including machine condition data such as melt pressure, melt temperature, mold temperature, clamping force, screw position, etc. measurement values; it can also include real-time results from the automatic visual detection system, such as bottle cap appearance, size, etc., as well as images of bottle cap chipping taken by the microscopic high-speed camera and processing data of the AI image processing device, etc.​

[0027] Step S102, constructing a digital twin mapping system of the medical infusion bottle cap injection molding system, wherein the digital twin mapping system comprises a virtual mapping layer, a data layer, a model layer, and a simulation and analysis layer.

[0028] In some embodiments, the digital twin mapping system will be combined to help obtain the predicted optimized production parameters. It should be noted that the digital twin mapping system is a virtual representation of the entity system, which can simulate each stage of the production process of the entity system, and through real-time data updating, use simulation, machine learning and reasoning to assist decision-making. The core features of the digital twin mapping system include three basic components: physical objects, virtual counterparts, and dynamic connection paths that facilitate continuous interaction between the two, which can include a virtual mapping layer, a data layer, a model layer, and a simulation and analysis layer. As shown in Figure 3 The specific construction steps S1021-S1024 of the digital twin mapping system will be described below.

[0029] In some embodiments, the construction of the digital twin mapping system of the medical infusion bottle cap injection molding system can specifically include:

[0030] Step S1021, based on the geometry and structure, physical properties, behavior and state information of the medical infusion bottle cap injection molding system, a virtual mapping layer corresponding to the medical infusion bottle cap injection molding system is constructed.

[0031] The core of digital twinning is the accurate mapping of physical entities, which needs to completely reproduce the static characteristics and dynamic behavior of physical objects. In some embodiments, the geometry and structure data can include the three-dimensional structure, size, shape, etc. of the injection molding system, for example, it can include the assembly structure of the parts of the industrial equipment, which can usually be constructed by CAD, laser scanning, CT / MRI, etc. The physical property data includes the intrinsic properties such as material characteristics, mechanical parameters, and chemical properties, which provide basic data for simulation analysis, such as the density, elastic modulus, and heat resistance of the melt material. The behavior and state data include the dynamic response law of the injection molding system under different conditions, such as the change curve of the rotational speed and temperature of the injection molding system with the load.

[0032] Step S1022, constructing a data layer, wherein the data layer includes at least one of the data of the plurality of sensors of the medical infusion bottle cap injection molding system, historical data, and environmental data obtained in real time.

[0033] The digital twin needs to rely on real-time input and fusion of massive data, and collects, processes, stores and circulates data between the entity and the virtual model. In some embodiments, the data layer can include: a plurality of sensor data acquired in real time, which can include state data of the injection molding system such as temperature, pressure, image, etc.; historical data can include historical operation records, fault logs, maintenance archives of the medical infusion bottle cap injection molding system, and can also include process data related to the production and processing of the injection molding system (such as production plan, supply chain information, user operation record, etc.), which can be used for model training and trend analysis; environmental data can include external parameters of the environment where the injection molding system is located (such as temperature, humidity, air pressure, etc.), which can be used to compensate for the interference of environmental factors on production parameters; or a combination of one or more of the above data.

[0034] It should be noted that in the data layer, various data obtained need to be cleaned and fused to ensure the accuracy and consistency of the data. The method of cleaning and fusing data can use existing methods, and will not be described in detail here.

[0035] Step S1023, at least one data in the data layer is acquired, and geometry and structure, physical properties, behavior and state information, and rule knowledge are fused to construct a model layer.

[0036] The model layer is the core of the digital twin mapping system, and its essence is the digital image of the physical entity. It needs to comprehensively analyze the physical entity, extract key features and parameters, and provide the basis for modeling. After the model layer is constructed, it can realize the rule abstraction and dynamic deduction of the physical entity.

[0037] In some embodiments, the geometry and structure dimension represents the spatial form of the physical entity. When constructing the model layer, the three-dimensional structure, assembly relationship and key dimensions of the medical infusion bottle cap injection molding system need to be accurately mapped, providing a spatial carrier for subsequent simulation; the physical property dimension represents the physical properties and laws. When constructing the model layer, the material properties of the melt and the action laws of the physical fields such as pressure / temperature / injection speed in the injection molding process are considered as the core engine of simulation analysis; the behavior and state information dimension represents the dynamic running process, which describes the state change of each stage of the injection molding process with time, so that the model layer can realize dynamic twinning; the rule dimension represents the constraints and knowledge, which converts the field knowledge, production specifications and safety limits into rules recognizable by the model, thereby ensuring the rationality of the simulation results. Based on the data in the data layer, the above-mentioned geometry, physics, behavior and rule dimensions are fused to construct the model layer, which can ensure the accuracy of the digital twin mapping system in quality control and production prediction.

[0038] Step S1024, a simulation and analysis layer is constructed to synchronously simulate the current state of the medical infusion bottle cap injection molding system, and to dynamically simulate the state, behavior or future trend of the medical infusion bottle cap injection molding system.

[0039] In some embodiments, in the simulation and analysis layer, on the one hand, the current state of the medical infusion bottle cap injection molding system is accurately simulated in synchronization according to the real-time collected multiple sensor data, and real-time mirroring of the entity system is realized. For example, the corresponding temperature field distribution in the virtual mapping layer can be adjusted in real time according to the temperature data measured by the temperature sensor in real time; or the pressure size and distribution in the dynamic change model can be changed to keep the virtual mapping layer and the running state of the medical infusion bottle cap injection molding system highly consistent; on the other hand, based on the constructed parameter prediction model and the data in the data layer, the new design scheme and / or production parameter can be simulated and verified, and the future trend of the injection molding system can be predicted and simulated. Through simulation operation on different production parameters or their combinations, scientific basis can be provided for production decision-making.

[0040] Step S103, inputting the multiple sensor data into the digital twin mapping system to obtain dynamic simulation data, constructing a first parameter prediction model based on the multiple sensor data and the dynamic simulation data by using a data-driven modeling algorithm, or constructing a second parameter prediction model by using a transfer learning algorithm.

[0041] In some embodiments, by using a data-driven modeling method, the regularity of the medical infusion bottle cap injection molding system is abstracted and dynamically deduced, and based on this, a first parameter prediction model is constructed. More specifically, various scenarios can be simulated in the digital twin mapping system, so that the virtual mapping layer forms a comprehensive representation of the medical infusion bottle cap injection molding process, the production results of different production parameters are observed, and a data-driven modeling method (such as regression analysis) is used to construct a first parameter prediction model, thereby realizing real-time monitoring, fault detection and decision support of the medical infusion bottle cap injection molding system.

[0042] In some embodiments, in order to improve the production qualification rate of the medical infusion bottle cap, the first parameter prediction model at least includes a surface defect prediction model and a chipping risk prediction model. The surface defect prediction model is used to predict the probability of burr, shrinkage, crack, etc. The chipping prediction model is used to predict the chipping risk of demolding, puncture, etc.

[0043] Specifically, based on the dynamic characteristics of the simulation injection process and the material properties in the digital twin system, the severity of the surface defect can be calculated by the following formula:

[0044] (1)

[0045] In the above formula (1), D represents the severity of surface defects, is the holding pressure fluctuation value, T1 is the melt temperature, T2 is the mold temperature, t1 is the cooling time, V is the injection speed, and MFR represents the melt flow rate; is the holding pressure fluctuation reference value, is the temperature change rate reference value, is the injection speed reference value, is the melt flow rate reference value, and the reference values of the above parameters can be obtained through design of experiments (DOE) or historical production database. k0-k4 are model parameters, which can be obtained through dynamic simulation data fitting.

[0046] It can be understood that, when is larger, the surface of the cap after demolding is more prone to shrink marks, and if the holding pressure is insufficient, regional depressions are more likely to occur; when the melt temperature and the mold temperature differ too much, the surface is prone to stress shrinkage due to sudden temperature drop and rapid cooling; when the cooling time is too short, the surface hardening is uneven, resulting in depressions; and when the injection speed is too high, the melt is prone to turbulent flow, thereby causing scratches or weld marks on the surface; when the MFR is too high, the melt is prone to accumulate in the dead corners of the mold to produce flash, and if it is too low, the filling is insufficient, resulting in surface depressions. Therefore, various production parameter scenarios can be simulated in the digital twin mapping system to obtain dynamic simulation data, and the above model parameters can be obtained through least squares fitting. In some embodiments, the value range of k0 can be -2.8-0.3, k1=0.3, k2=0.04, k3=0.05, and k4=0.02.

[0047] In addition, due to the particularity of medical infusion bottle caps, the risk of debris, impurities or shedders occurring on the surface of the bottle cap after demolding also needs to be considered. The chipping risk prediction mainly predicts the possible chipping of the medical infusion bottle cap material during the injection molding production process, as well as in the subsequent storage, transportation, use (such as puncture and drug injection, infusion operation) and other links. Specifically, the chipping includes: during injection molding, burrs or fine cracks are generated on the edge, sealing surface and other parts of the cap due to unreasonable mold structure or improper process parameter setting (such as temperature, pressure, holding time, etc.), and then chipping occurs during demolding or subsequent processing; or during storage and transportation, chipping may be caused by defects in the material of the cap itself (such as excessive brittleness), extrusion or friction of external force; or during clinical use, chipping may be caused by inappropriate puncture angle and force, or problems in the material and structure design of the cap, and the chipping may mix into the drug solution, affecting the safety of drug use.

[0048] Specifically, the chipping risk index can be calculated by the following formula:

[0049] (2)

[0050] In the above formula (2), F represents the chipping risk index, P 熔体 is the melt pressure, t2 represents the time of the injection molding process, is the rate of change of the melt pressure with time, reflecting the speed of the melt pressure rising or falling in the injection molding process; V 剪切速率 is the melt shear rate, which represents the shear deformation rate generated by the velocity gradient when the melt flows in the screw or the flow channel, which can be calculated by the differential of the screw displacement or simulated by the injection molding flow using simulation software such as Moldflow to output the shear rate field; is the temperature difference of the melt in the mold, reflecting the unevenness of the mold temperature distribution; is the reference value of the rate of change of the melt pressure with time, is the reference value of the melt shear rate, is the reference value of the temperature difference of the melt in the mold, and the reference values of the above parameters can be obtained by design of experiments (DOE) or historical production database. α, β, γ are model parameters, which can be determined by dynamic simulation data in combination with the properties of the melt material.

[0051] It can be understood that when changes rapidly, the stress in the melt is released or concentrated instantaneously, which can cause local material fracture, peeling, and thus produce chippings; when V 剪切速率 is too high, the friction between the melt and the surface of the screw and the mold will be intensified, which can cause the material to overheat and decompose or the surface to wear, thus causing the cap to produce chippings; when is too large, the cooling speed of the melt in the mold will be uneven, generating internal stress, which can cause the material on the surface or inside of the cap to peel off due to stress concentration and form chippings. Similarly, various scenarios of different production parameters can be simulated in the digital twin mapping system, and the model parameters can be determined by multiple linear regression analysis in combination with the properties of the melt material. In some embodiments, taking medical-grade polypropylene as an example, = 1.23, = 0.089, = 0.56.

[0052] In other embodiments, in order to improve the accuracy of the prediction model, the prediction model can also be input into an artificial neural network for training to obtain an optimized prediction model. It can be understood that a large amount of sample data is required in the process of training the artificial neural network model, but usually it takes a lot of manpower and time cost to construct these data, and in order to solve this problem, the algorithm of transfer learning can be introduced, which enhances the effect and performance of model training by means of the knowledge of related tasks.

[0053] The transfer learning technique is used to transfer knowledge between different component data sets, while source data sets from software simulation and experimental testing are used. For example, in order to obtain surface defect related data from simulation, it is necessary to correlate simulation output variables with experimental observations. Transfer learning can reduce the amount of data required for training the parameter prediction model, and software simulation data can be used as a low-cost alternative to experimental data. The prediction index value obtained by transfer learning from simulation is comparable to that of an untrained network, but the amount of data required for the target data set is greatly reduced.

[0054] It can be understood that the second parameter prediction model constructed by the transfer learning algorithm also includes at least a surface defect prediction model and a spalling risk prediction model, which are determined by selecting the relationship between different production parameters and quality results. Specifically, as shown in Figure 4 The second parameter prediction model constructed by the transfer learning algorithm mainly includes the following steps S1031-S1033:

[0055] Step S1031: The process of injection molding of medical infusion bottle caps is simulated by Moldex3D software to create a source data set, wherein the source data set includes an experimental data set and a software simulation data set, and the software simulation data set is generated by Latin hypercube sampling experimental simulation.

[0056] In this step, the CAD model of the medical infusion bottle cap is imported into the Moldex3D software, and at the same time, in order to obtain simulation results within an acceptable time range and further improve the accuracy of the results, some modifications can be made to each model, such as removing some unnecessary features in the medical infusion bottle cap that increase the grid burden, prolong the calculation time but do not improve the accuracy of the results.

[0057] It can be understood that the experimental data set is the training data from an existing, data sufficient or sufficiently learned task, which is a specific sample and related information collected by experiments, observations, etc. For example, 1000 effective injection records of medical grade polypropylene injection infusion bottle caps, including production parameters and defect labels.

[0058] For the construction of the software simulation dataset, a grid structure with 4 boundary layers can be selected for the medical infusion bottle cap to improve the simulation accuracy near the surface of the medical infusion bottle cap. In addition, in the simulation, considering that the Latin hypercube sampling experiment almost randomly distributes sampling points in the domain of each variable, it can maximize the reduction of overlap between experimental conditions, improve sampling efficiency, and reduce the correlation between multi-dimensional variables. Therefore, the Latin hypercube sampling experiment is used to systematically change the production parameters, wherein the production parameters are several key production parameters that the operator often adjusts during the iterative machine setting process, such as mold temperature, injection speed, holding pressure, etc. The simulation results corresponding to each key production parameter or combination thereof are obtained, which represent the quality / performance indicators (labels) of the medical infusion bottle cap, such as the shrinkage rate of the medical infusion bottle cap. Finally, 100 data points are formed, each of which is simulated individually and contains production parameters and corresponding labels.

[0059] It should be noted that since the simulation results do not directly give the experimental observation results about, for example, “whether the surface defect is visible” and “whether there is chipping”, a classification feature needs to be determined. In some embodiments, the frozen layer ratio is used as a classification feature for evaluating the visibility of defects. In addition, by randomly changing the production parameters for simulation, the FLR values obtained from these simulations and the plurality of molded medical infusion bottle caps determine an effective FLR threshold value for distinguishing between defective and non-defective cases, and then the FLR threshold value is compared with the FLR value related to each simulation. If the sampled FLR exceeds the threshold value, it is considered that there is a defect on the surface, and the simulation result is classified as class 1. On the contrary, if the FLR value is equal to or lower than the threshold value, it is considered that the defect is not visible, and the simulation result is classified as class 0. For example, the threshold value of the FLR is 5%. This threshold value is used to classify the simulation results into class 0 and class 1.

[0060] Step S1032: Pre-training the second parameter prediction model based on the source dataset, and obtaining the initial weights of the second parameter prediction model by back propagation optimization.

[0061] For the defect prediction of medical infusion bottle cap, the convolutional layer of CNN can extract local features through sliding window, which can automatically learn the hidden correlation between multiple production parameters (such as temperature-pressure coupling effect), and can significantly improve the recognition accuracy of defect causes. In addition, the LSTM neural network can effectively capture long-term temporal dependencies through the gating mechanism (forget gate, input gate, output gate), and can remember the parameter changes at key time nodes (such as pressure peak at the end of filling and temperature drop point in the cooling stage), thereby accurately associating the causal relationship between the time sequence trajectory and the defect, which is important for the prediction of surface defects of medical infusion bottle cap (such as micro-cracks at the puncture site). Therefore, the second parameter prediction model adopts CNN-LSTM neural network, which has the dual ability of spatial correlation capture and temporal dynamic modeling, and can perfectly adapt to the complexity of the medical infusion bottle cap injection molding process, efficiently utilizing limited medical data and accurately identifying subtle defects, providing more reliable prediction support for the quality control of medical infusion bottle cap.

[0062] In some embodiments, in order to optimize the performance of the artificial neural network, the hyperparameters of the artificial neural network need to be determined before applying transfer learning. First, the structure of the artificial neural network can be determined by generating loss function curves and accuracy curves, so as to select a network structure that minimizes the risk of overfitting or underfitting. Then, the remaining hyperparameters can be determined by grid search. For example, the CNN-LSTM neural network can be an artificial neural network including 4 hidden layers, and each hidden layer has 4 neurons. It should be noted that early stopping callback can also be added to the artificial neural network to further reduce the risk of overfitting. Then, the CNN-LSTM neural network is trained using the source dataset by backpropagation optimization to obtain the initial weights.

[0063] Step S1033, performing secondary training on the target dataset, selecting a soft start or random initialization strategy to fine-tune the initial weights, and determining the second parameter prediction model after transfer learning.

[0064] After obtaining the initial weights, the weights of the second parameter prediction model are fine-tuned, the purpose of which is to adapt the model to the surface defect prediction or chipping risk prediction of the medical infusion bottle cap. The target dataset only includes experimental result data, which can be the production parameters for performing the current task and the corresponding label data.

[0065] In some embodiments, a soft start SS or random initialization strategy can be selected to fine-tune the weights of the second parameter prediction model. When the target data is less and the source task is highly similar to the target task, the soft start strategy is preferred, which can avoid the destruction of the initial weights obtained by pre-training and smoothly migrate to the target task. When the target task and the source task are quite different (such as different defect definitions), the random initialization strategy can be selected, which can eliminate negative transfer and make the model more flexible to learn new rules.

[0066] It should be noted that after obtaining the second parameter prediction model, a control network related to the target data set can be trained to verify the parameter prediction model, which serves as a reference for evaluating the effect of transfer learning. In some embodiments, the training process can be repeated by gradually increasing the number of data points of the target data set in knowledge transfer. In each iteration, the network performance is evaluated using three indicators: accuracy, AUC, and recall, and five-fold cross-validation is performed to evaluate the repeatability of the results, and then the standard deviation of various indicators is calculated. When the accuracy, AUC, recall, and standard deviation are all better than the aforementioned control network, the prediction model can be put into the parameter prediction of the injection molding process.

[0067] The introduction of the transfer learning algorithm to train the second parameter prediction model in the embodiments of the present disclosure has the following advantages: transfer learning reduces the need for large data collection and model training for the target task, enhances the robustness and generalization ability of the machine learning model, reduces the risk of overfitting, and ensures reliable operation in different production scenarios. Simplifying the training of artificial neural networks using transfer learning can be used to predict medical infusion bottle cap surface defects and chipping risk.

[0068] Step S104, based on the first parameter prediction model or the second parameter prediction model, the optimal production parameters are determined by searching in the experience case set using a fuzzy inference algorithm, wherein the experience case set includes historical data and model-derived data.

[0069] Generally, the historical data can include production plans, equipment states, process parameters, quality detection, and abnormal handling, etc., wherein the abnormal handling includes situations deviating from normal conditions in the production process and countermeasures. For example, the abnormal phenomenon of frequent flash of products is recorded, and the reasons and how to handle it are analyzed. In addition, virtual data can be generated using other simulation models to supplement the deficiencies of historical data, such as simulation of unoccurring conditions, extreme scenarios, or potential optimization space, to obtain model-derived data. Therefore, in order to improve the efficiency of the digital twin mapping system in accurately predicting production parameters, the experience case set will be constructed by combining historical data and model-derived data in this step.

[0070] For the creation of the experience case library, in some embodiments, a plurality of possible improper molding conditions are created, after defining all the features and parameters of each defect, the significant weight of each defect is calculated, and the experience case library is created based on all the faults, their causal rules, attributes and their weights.

[0071] In some embodiments, the possible defect problems can be predicted by analyzing the relationship between different production parameters and injection result features in the injection molding process, and this relationship can be determined by fuzzy reasoning, which is used to classify the influence and relationship between production parameters and quality results. In order to determine the relationship between production parameters and quality results, a fuzzy reasoning algorithm can be used, which takes production parameters as input and generates membership functions representing the degree of relationship between production parameters and quality results as output.

[0072] For example, after obtaining the predicted surface defect degree and the predicted flaking risk index of the bottle cap by formula (1) and formula (2), the probability value of the bottle cap appearing surface defect or flaking risk is calculated by formula (3), which is in the range of 0-1.

[0073] (3)

[0074] Where Y represents the probability value of appearing surface defect or flaking risk, and Z is the surface defect degree D or the flaking risk index P mentioned above.

[0075] In some embodiments, the production parameters used in the surface defect prediction model and the flaking risk prediction model can be classified, for example, they can be divided into: temperature, pressure, time, speed. In addition, 5 fuzzy subsets are defined, for example, very weak, weak, medium, strong, very strong, to represent the degree of relationship and influence between the above production parameters and quality results, which can be represented by membership functions.

[0076] Specifically, the membership function is:

[0077] (4)

[0078] Where the membership degree to express the degree of relationship between production parameters and quality results; a, b, c are three key parameters belonging to the domain U, the specific values of the three parameters are determined according to the relationship between the type of parameters (such as temperature, pressure, time, etc.) simulated in the simulation and analysis layer and the fuzzy set (such as very weak, weak) obtained by simulating the prediction model, and the domain U is the entire range of values that the characteristic parameters can take in the actual injection molding process (such as the actual measurement range of temperature, pressure, etc. parameters in the injection molding process); wherein, a corresponds to the starting point (left boundary) at which the membership degree starts to rise from 0, b corresponds to the top point (center point) at which the membership degree reaches the maximum value 1, and c corresponds to the end point (right boundary) at which the membership degree falls from 1 to 0; x represents the value of the actually measured production parameter in the injection molding process. In some embodiments, in addition to needing to combine the physical range of actual process parameters (domain U), critical values in production experience (such as threshold values for defect generation), and optimal operating points (such as the parameter values with the highest yield), a, b, and c can also be determined by inputting the production parameter x into the parameter prediction model or the parameter prediction model and determining the probability value obtained from equation (3).

[0079] Specifically, taking the surface defect prediction model as an example, assuming that it is necessary to determine whether the relationship between the melt temperature and the occurrence of surface defects belongs to strong or medium, first, in the surface defect prediction model, set the other three values except the melt temperature to be normal values in the normal processing process, then simulate different scenarios in the simulation and analysis layer, calculate the probability of occurrence of surface defects under the simulation conditions, and then determine the values of a, b, and c in the membership function according to the definitions of a, b, and c. For example, when the melt temperature is less than or equal to 200℃, the material is not fully plasticized, and the surface defect probability is close to 1; when the melt temperature is equal to 230℃, the product yield is the highest, and the surface defect probability is close to 0; when the melt temperature is greater than or equal to 250℃, the material will degrade and produce scorch marks, and the surface defect probability is also close to 1, and according to the simulation results, when the temperature is between 200-230℃, the surface defect probability decreases (corresponding to the membership degree rising from 0 to 1) as the temperature increases; when the temperature is between 230-250℃, the surface defect probability increases (corresponding to the membership degree falling from 1 to 0) as the temperature increases, then it can be determined that a is 200℃, b is 230℃, and c is 250℃. If the monitored melt temperature is 220℃ at this time, then the calculation by the membership function obtains 0.67, indicating that there is a certain probability of producing slight defects at this temperature.

[0080] Then, the fuzzy membership is converted into a clear value by defuzzification formula (5), and the clear value of each production parameter is calculated in the same way, and finally the fuzzy weight of each production parameter is calculated by using the weighted average method, and the fuzzy weight is calculated by formula (6).

[0081] (5)

[0082] wherein C() represents deblurring, represents a blur amount, which is used to represent a less clear, more blurred relationship (for example, the relationship between melt temperature and the occurrence of surface defects).

[0083] (6)

[0084] wherein W fi is the blur weight of the i-th feature, FO fi is the occurrence weight of the i-th feature, which represents the probability that a larger occurrence of surface defects is caused by melt temperature abnormalities in historical failures, and is determined according to production data, is the clear value of the j-th parameter and the related quality result i relationship, PR ji refers to the correlation term between the j-th parameter and the related quality result i.

[0085] After the above process, the production parameter with the highest weight in the occurrence of surface defects or the risk of spalling is determined as the characteristic parameter, and the characteristic parameter is used to search in the experience case set to output the optimal production parameter in the case of partial similarity or complete similarity. If no similar case is found in the experience case set, the predicted parameter output by the first parameter prediction model or the second parameter prediction model is used as the optimal production parameter.

[0086] Step S104, output the optimal production parameter to the medical infusion bottle cap injection molding system and update it in real time.

[0087] In this step, the determined optimal production parameter is returned to the medical infusion bottle cap injection molding system, and the production parameter is adjusted in time to correct the predicted abnormal parameter in time to avoid the output of defective bottle caps, so as to reduce the production cost. It can be understood that the optimal production parameter is updated in real time, and the production process is adjusted in time.

[0088] The intelligent control method for medical infusion bottle cap production proposed in the embodiments of the present disclosure is based on the digital twin of the injection molding system, which aims to improve production efficiency and product quality by integrating artificial intelligence driven detection methods. The digital twin can predict and optimize product and production performance in real time using data from the injection molding process, simulate and predict using the twin mapping system, find the production parameters that need to be adjusted to improve production efficiency and product quality, and continuously optimize and adjust. In addition, the intelligent control method is also based on case-based reasoning and historical production data reference based on fuzzy weight, and the previous experience is used for the production parameter adjustment of the current task, so that the accuracy in predicting and preventing production defects is further improved.

[0089] Based on the same inventive concept, the disclosure also provides an intelligent control device for medical infusion bottle cap production, as shown in Figure 5 Figure 5 An intelligent control device 200 for medical infusion bottle cap production in some embodiments of the disclosure is shown, which includes: an acquisition module 201 for acquiring data of the plurality of sensors in the medical infusion bottle cap injection molding system in real time; a parameter prediction model generation module 202 for constructing a digital twin mapping model of the medical infusion bottle cap injection molding system, inputting the plurality of sensor data into the digital twin mapping system to obtain dynamic simulation data, and generating a parameter prediction model based on the plurality of sensor data and the dynamic simulation data; an optimal production parameter determination module 203 for determining optimal production parameters based on the parameter prediction model and using a fuzzy reasoning algorithm to search from an experience case set, wherein the experience case set includes historical data and model deduction data; and an output and update module 204 for outputting the optimal production parameters to the medical infusion bottle cap injection molding system and updating in real time.

[0090] The specific functions of each functional module in the above device have been described in detail in the intelligent control method for medical infusion bottle cap production provided by some embodiments of the disclosure, and will not be described in detail here.

[0091] Based on the same inventive concept, the electronic device 300 in the embodiments of the disclosure is also in the form of a general computing device. As shown in Figure 6 The electronic device includes a processor 301 and a memory 302 for storing executable instructions of the processor 301, wherein the processor 301 is configured to execute the executable instructions to implement the intelligent control method for medical infusion bottle cap production described in any of the above embodiments.

[0092] In addition, the electronic device can also communicate through a bus interface. As shown in Figure 6 The bus interface 303 provides an interface between the bus and the receiver 304 and the transmitter 305. In some embodiments, the receiver 304 and the transmitter 305 can also be the same element, i.e., a transceiver, to provide a unit for communicating with various other devices on a transmission medium. The processor 301 is responsible for managing the bus generally processing, while the memory 302 can be used to store data used by the processor 301 in performing operations.

[0093] ​In the example embodiments of the present disclosure, based on the same inventive concept, the present disclosure also provides a computer readable storage medium having stored thereon a computer program, which, when executed by a processor, implements the intelligent control method for medical infusion bottle cap production described in any of the above embodiments. In some possible embodiments, various aspects of the present disclosure can also be implemented in the form of a program product, which includes program code for causing a terminal device to perform the steps described in the above “example method” section according to various example embodiments of the present disclosure when the program product is run on the terminal device.

[0094] The program product for implementing the above method according to the embodiments of the present disclosure can store program code in a portable compact disc read-only memory (CD-ROM) and can be run on a terminal device, such as a personal computer. Of course, the program product of the present disclosure is not limited to this, and in the embodiments of the present disclosure, the readable storage medium can be any tangible medium containing or storing a program, which can be used by or in conjunction with an instruction execution system, apparatus or device. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only memory (Rom), a random access memory (RAM), a flash memory, a hard disk drive (HDD) or a solid state disk (SSD), etc., and the storage medium can also include a combination of the above types of memories.

[0095] The program code for performing the operations of the present disclosure can be written in any combination of one or more programming languages, including an object-oriented programming language such as Java, C++, etc., and a conventional procedural programming language such as “C” language or similar programming languages. The program code can be executed entirely on a user computing device, partially on a user device, as a separate software package, partially on a user computing device and partially on a remote computing device, or entirely on a remote computing device or server. In the case involving a remote computing device, the remote computing device can be connected to the user computing device through any kind of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computing device (for example, connected to the Internet through an Internet service provider).

[0096] The computer program instructions can also be loaded onto a computer, other programmable data processing apparatus, or other processing device to cause a series of operational steps to be performed on the computer, other programmable apparatus or other processing device to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide processes for implementing the functions specified in the flowchart block or blocks. Figure 1 The flowchart and / or block diagram in the Figures illustrate the architecture, functionality, and operation of possible implementations of apparatuses and computer program products according to various embodiments of the present disclosure. In this regard, each flowchart and / or block diagram can represent a method, apparatus, and / or computer program product according to an embodiment of the present disclosure. In some embodiments, each flowchart and / or block diagram can represent a portion of a method, apparatus, and / or computer program product according to an embodiment of the present disclosure. The computer program instructions can also be loaded onto a computer, other programmable data processing apparatus, or other processing device to cause a series of operational steps to be performed on the computer, other programmable apparatus or other processing device to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide processes for implementing the functions specified in the flowchart block or blocks. Figure 1 The flowchart and / or block diagram in the Figures illustrate the architecture, functionality, and operation of possible implementations of apparatuses and computer program products according to various embodiments of the present disclosure. In this regard, each flowchart and / or block diagram can represent a method, apparatus, and / or computer program product according to an embodiment of the present disclosure. In some embodiments, each flowchart and / or block diagram can represent a portion of a method, apparatus, and / or computer program product according to an embodiment of the present disclosure. The computer program instructions can also be loaded onto a computer, other programmable data processing apparatus, or other processing device to cause a series of operational steps to be performed on the computer, other programmable apparatus or other processing device to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide processes for implementing the functions specified in the flowchart block or blocks.

[0097] While the preferred embodiments of the present disclosure have been described, additional variations and modifications can be made to the embodiments by those of ordinary skill in the art once they have the benefit of the foregoing description. Therefore, the appended claims are intended to cover all such variations and modifications that fall within the scope of the present disclosure. Thus, the true scope of the present disclosure is not intended to be limited to the preferred embodiments described herein but is to be accorded the widest scope consistent with the principles and the novel features disclosed herein.

[0098] It will be apparent to those skilled in the art that other embodiments of the present disclosure can be developed and made without departing from the scope and spirit of the present disclosure. The application is intended to embrace all known or

[0099] The above only describes the embodiments of the present disclosure and is not intended to limit the present disclosure. Those skilled in the art can make various changes and modifications to the present disclosure without departing from the scope of the description of the above embodiments. It should be understood that various and modifications, equivalent replacements, improvements, etc. can be made to the present disclosure without departing from the scope of the above embodiments, and all should be included in the scope of the claims of the present disclosure.

Claims

1. A control method for medical infusion bottle cap production, characterized by, The application is applied to a medical infusion bottle cap injection molding system, the medical infusion bottle cap injection molding system comprises a plurality of sensors, and the control method comprises: Real-time acquisition of data of the plurality of sensors in the medical infusion bottle cap injection molding system; A digital twin mapping system of the medical infusion bottle cap injection molding system is constructed, wherein the digital twin mapping system comprises a virtual mapping layer, a data layer, a model layer and a simulation and analysis layer; The data of the plurality of sensors are input into the digital twin mapping system to obtain dynamic simulation simulation data, a first parameter prediction model is constructed based on the data of the plurality of sensors and the dynamic simulation simulation data by using a data-driven modeling algorithm, or a second parameter prediction model is constructed by using a transfer learning algorithm; Based on the first parameter prediction model or the second parameter prediction model, the optimal production parameters are determined by searching in an experience case set by using a fuzzy reasoning algorithm, wherein the experience case set at least comprises historical data and model deduction data; The optimal production parameters are output to the medical infusion bottle cap injection molding system and are updated in real time; The first parameter prediction model at least comprises a surface defect prediction model and a chipping risk prediction model; The surface defect prediction model is used to predict the severity of surface defects of the medical infusion bottle cap, and the severity is calculated by the following formula: ； wherein D represents the severity of surface defects, is the holding pressure fluctuation value, T1 is the melt temperature, T2 is the mold temperature, t1 is the cooling time, V is the injection speed, MFR represents the melt flow rate; is the holding pressure fluctuation value, is the reference value of the temperature change rate, is the reference value of the injection speed, is the reference value of the melt flow rate; k0 to k4 are model parameters; The chipping risk prediction model is used to predict the chipping risk index of the medical infusion bottle cap when demolding or puncturing, wherein the chipping risk index is calculated by the following formula: ； wherein F represents a flaking risk index, P 熔体 is the melt pressure, t2 represents the time of the injection molding process, is the rate of change of the melt pressure with respect to time; V 剪切速率 is the melt shear rate; is the temperature difference of the melt within the mold; is the reference value of the rate of change of the melt pressure with respect to time; is the reference value of the melt shear rate; is the reference value of the temperature difference of the melt within the mold; is the model parameter.

2. The control method according to claim 1, characterized by, The construction of the digital twin mapping system of the medical infusion bottle cap injection molding system comprises: Based on the geometry and structure, physical properties, behavior and state information of the medical infusion bottle cap injection molding system, the virtual mapping layer corresponding to the medical infusion bottle cap injection molding system is constructed; The data layer is constructed, wherein the data layer comprises at least one of the data of the plurality of sensors of the medical infusion bottle cap injection molding system acquired in real time, the historical data and the environmental data; The at least one data in the data layer is acquired, the geometry and structure, the physical properties, the behavior and state information and the rule knowledge are fused, and the model layer is constructed; The simulation and analysis layer is constructed, and the current state of the medical infusion bottle cap injection molding system is simulated synchronously, and the state, behavior or future trend of the medical infusion bottle cap injection molding system is dynamically simulated.

3. The control method according to claim 1, characterized by, The construction of the second parameter prediction model by using the transfer learning algorithm comprises: The injection molding process of the medical infusion bottle cap is simulated by using Moldex3D software to create a source data set, wherein the source data set comprises an experimental data set and a software simulation data set, and the software simulation data set is generated by simulating experiments by Latin hypercube sampling; The second parameter prediction model is trained based on the source data set, and the initial weight of the second parameter prediction model is obtained by back propagation optimization, wherein the second parameter prediction model is a CNN-LSTM neural network; The second parameter prediction model is trained on the target data set, and the initial weights are fine-tuned by selecting a soft start or random initialization strategy to determine the second parameter prediction model after migration learning.

4. The control method according to claim 1, characterized by, Based on the first parameter prediction model or the second parameter prediction model, the optimal production parameters are determined by searching in the experience case set using a fuzzy inference algorithm, including: The degree of relationship between each production parameter and the quality result is determined by a membership function, and the production parameter with the highest weight in the risk of surface defects or spalling is determined as the characteristic parameter by defuzzification and weighted average method; The characteristic parameter is used to search in the experience case set, and the corresponding production parameters when partial or complete similarity occurs are used to determine the optimal production parameters.

5. The control method according to claim 4, characterized by The membership function is: ； membership degree to indicate the degree of relationship between the production parameters and the quality results; a corresponds to the beginning of the rise of the membership degree from 0, b corresponds to the vertex of the maximum value of the membership degree of 1, c corresponds to the end of the decrease of the membership degree from 1 to 0; x represents the value of the production parameter measured in the injection molding process of the medical infusion bottle cap.

6. An electronic device, comprising: The processor is configured to execute the executable instructions to implement the control method for medical infusion bottle cap production according to any one of claims 1-5.

7. A computer readable storage medium characterized in that, The computer program is stored on the computer readable storage medium, and the computer program is executed by the processor to implement the control method for medical infusion bottle cap production according to any one of claims 1-5.

Images