Plastic product production operation system and method based on data driving and multi-terminal cooperation

Abstract

The invention provides a plastic product production operation system and method based on data driving and multi-terminal cooperation, and belongs to the technical field of product quality control, and the system comprises a terminal data collection module which collects original terminal data generated by multiple terminals, and the original terminal data comprises original equipment terminal data and original quality inspection terminal data; the original equipment terminal data and the original quality inspection terminal data are associated through a unique identifier; the terminal data access module is used for accessing original terminal data into the system; the protocol conversion module is used for performing protocol analysis and data conversion on the original terminal data to obtain target terminal data in a uniform format; the data processing module is used for processing and analyzing the target terminal data and then determining an adjustment control strategy; and the application module receives the adjustment control strategy to perform operation adjustment on the equipment terminal. According to the method, the operation of the equipment terminal is optimized through real-time monitoring, the quality problem caused by equipment operation fluctuation or process deviation is reduced, and the defective rate is reduced.

Description

Technical Field

[0001] The present invention relates to the field of product quality control technology, and in particular to a plastic product production and operation system and method based on data-driven and multi-terminal collaboration. Background Art

[0002] In the current plastics manufacturing industry, quality control is crucial for ensuring product performance and market competitiveness, and companies are increasingly demanding higher quality. To improve product quality, a Manufacturing Execution System (MES) is introduced for production planning and scheduling. When a new production task is received, it is dispatched through the MES and then produced. For example, CN107831750A discloses an IMES intelligent manufacturing execution system, which includes a basic data module, a production planning module, a production scheduling module, a quality management module, a manufacturing resource management module, a staff management module, a warehouse management module, a comprehensive reporting module, a system management module, and a mobile terminal system. The production planning and scheduling modules organize operations and complete product production. However, in the traditional plastics production process, quality inspection is typically performed after production is complete. When quality issues are discovered during quality inspection of plastic products, the company then investigates the equipment on the production line for any malfunctions. This post-production quality inspection approach presents numerous drawbacks. Because problems cannot be detected and corrected in real time during the production process, a large number of defective products continue to be produced undetected. Later, to identify the causes of the defects, a significant amount of time and effort is required to retrospectively analyze the entire production chain, resulting in a significant waste of production resources such as raw materials, equipment, and manpower, increasing production costs and reducing efficiency. Summary of the Invention

[0003] In response to the above problems, the present invention provides a plastic product production and operation system and method based on data-driven and multi-terminal collaboration, which can overcome the shortcomings of post-quality inspection of plastic products, achieve effective control of the production process, improve the quality of plastic products, and ensure the qualification rate of plastic products.

[0004] The present invention provides a plastic product production and operation system based on data drive and multi-terminal collaboration, the system includes a terminal data acquisition module, a terminal data access module, a protocol conversion module, a data processing module, and an application module; The terminal data collection module is used to collect original terminal data generated by multiple terminals, wherein the original terminal data includes original equipment terminal data for production equipment operation and original quality inspection terminal data for plastic product quality inspection, and the original equipment terminal data and the original quality inspection terminal data are associated with each other through a unique identifier; The terminal data access module is used to access the original terminal data generated by multiple terminals collected by the terminal data collection module into the system; The protocol conversion module is used to perform protocol analysis and data conversion on the original terminal data generated by multiple terminals and collected by the terminal data collection module and accessed by the terminal data access module to obtain target terminal data in a unified format, and transmit the target terminal data, wherein the target terminal data includes target device terminal data and target quality inspection terminal data; The data processing module is configured to receive the target terminal data transmitted by the protocol conversion module and process and analyze the target terminal data to determine an adjustment control strategy; The application module receives the adjustment control strategy determined by the data processing module to adjust the operation of the device terminal.

[0005] Optionally, the original quality inspection terminal data includes quality index data, and the data processing module (40) uses an adaptive PID algorithm to determine the adjustment control strategy:

[0006] Where, For the time to be optimized t The adjustment amount of the device terminal, is the quality indicator error, is the proportional coefficient, according to the quality index error at the current moment The size of the control is proportional to the The error of quality index From time 0 Time t Integration reflects the accumulation of quality indicator errors over a period of time. is the integration coefficient, Quality index error The time derivative, is the differential coefficient.

[0007] Optionally, the terminal data acquisition module includes an equipment terminal data acquisition unit and a quality inspection terminal data acquisition unit, the equipment terminal data acquisition unit is used to collect the original equipment terminal data, and the quality inspection terminal data acquisition unit is used to collect the original quality inspection terminal data, wherein the original equipment terminal data includes the operating status data and process parameter data of the equipment terminal, and the quality inspection terminal data acquisition unit collects the original quality inspection terminal data through manual and / or self-service quality inspection machines.

[0008] Optionally, the self-service quality inspection machine is a multimodal quality inspection machine, which detects appearance defects through high-resolution CMOS and AI algorithms, realizes size measurement through laser triangulation and CCD image measurement, detects surface roughness through confocal microscopy, and detects assembly integrity through deep vision ToF camera.

[0009] Optionally, the original equipment terminal data and the original quality inspection terminal data are associated with each other through a batch number, serial number or timestamp unique identifier to form a complete product quality file.

[0010] Optionally, the data processing module includes a data preprocessing unit and a data analysis unit. The data preprocessing unit cleans, filters and normalizes the target terminal data, obtains the standardized target terminal data and stores it. The data analysis unit analyzes the preprocessed target terminal data and automatically generates an adjustment control strategy for the device terminal based on the data analysis results.

[0011] Optionally, the data processing module uses a correlation analysis method to locate the target terminal device data that is significantly correlated with the product quality index:

[0012] Where, X Indicates device terminal data, Y Indicates quality inspection terminal data, Cov is the covariance, σ is the standard deviation.

[0013] Optionally, the data processing module predicts the product quality of the plastic product obtained when the equipment terminal is produced under different operating status data and / or process parameter data through a quality prediction neural network model between the target quality inspection terminal data and the target equipment terminal data obtained through pre-training.

[0014] Optionally, the application module includes an early warning unit, a process feedback optimization unit and a kanban unit. The early warning unit is used to send an alarm. When the data processing module processes and analyzes the target terminal data and finds that there may be hidden dangers that may affect the quality of plastic products, the alarm is sent out. The process feedback optimization unit is used to send the adjustment control strategy determined by the data processing module to the production personnel or directly to the edge controller of the equipment terminal to adjust the operation of the equipment terminal.

[0015] The present invention provides a plastic product production and operation method based on data-driven and multi-terminal collaboration, the method comprising: The terminal data collection module collects original terminal data generated by multiple terminals, wherein the original terminal data includes original equipment terminal data for production equipment operation and original quality inspection terminal data for plastic product quality inspection, and the original equipment terminal data and the original quality inspection terminal data are associated with each other through a unique identifier; The terminal data access module accesses the original terminal data generated by multiple terminals collected by the terminal data collection module into the system; Using a protocol conversion module to perform protocol parsing and data conversion on the original terminal data generated by multiple terminals collected by the terminal data collection module and accessed by the terminal data access module to obtain target terminal data in a unified format, and transmit the target terminal data, wherein the target terminal data includes target device terminal data and target quality inspection terminal data; Receiving the target terminal data transmitted by the protocol conversion module through the data processing module and determining an adjustment control strategy after processing and analyzing the target terminal data; The application module is used to receive the adjustment control strategy determined by the data processing module to adjust the operation of the device terminal.

[0016] The present invention provides a plastic product production and operation system and method based on data-driven and multi-terminal collaboration, which has at least the following beneficial effects: 1. Unified access to equipment terminal data and quality inspection terminal data breaks down "information silos" and enables real-time linkage between production data and quality data. Through a closed-loop process of collection-access-conversion-processing-application, data is transformed into a direct driving force for optimizing plastic product quality. This avoids the lag and subjectivity of traditional manual experience-driven methods, as well as the shortcomings of "post-event quality inspection," thereby improving product quality and ensuring the qualified rate of plastic products. 2. The original equipment terminal data and the original quality inspection terminal data are associated through a unique identifier, and the entire process data is recorded to support rapid tracing of quality issues. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a structural diagram of a plastic product production and operation system based on data drive and multi-terminal collaboration according to an embodiment of the present invention.

[0018] Figure 2 Schematic diagram of the hierarchical architecture of a plastic product production and operation system based on data-driven and multi-terminal collaboration according to an embodiment of the present invention.

[0019] Figure 3 It is a flow chart of a plastic product production and operation method based on data-driven and multi-terminal collaboration according to an embodiment of the present invention.

[0020] Description of the accompanying drawings: 10, terminal data acquisition module; 20, terminal data access module; 30, protocol conversion module; 40, data processing module; 50, application module. DETAILED DESCRIPTION

[0021] The following is a combination of specific embodiments and appendix Figure 1-3 The invention is described in detail so that those skilled in the art can more fully understand the purpose, features and effects of the invention.

[0022] Unless otherwise defined, all technical and scientific terms used herein have the same meanings as commonly understood by those skilled in the art to which the invention belongs. In the event that the definition of a term in the present invention conflicts with the meaning commonly understood by those skilled in the art to which the invention belongs, the definition in the present invention shall prevail.

[0023] Current plastic product production relies primarily on post-production manual quality control and analysis, failing to implement dynamic adjustments and early warnings during the production process. This results in a passive situation where defects are discovered only after the product has left the production line. This invention provides a data-driven, multi-terminal collaborative plastic product production and operations system and method, improving production control accordingly. This system enables real-time control and optimization of product quality, significantly improving product yields.

[0024] Example 1 As a specific embodiment of the present invention, this embodiment provides a plastic product production and operation system based on data drive and multi-terminal collaboration, referring to Figure 1 Combined with Figure 2 , including a terminal data acquisition module 10, a terminal data access module 20, a protocol conversion module 30, a data processing module 40, and an application module 50.

[0025] The terminal data collection module 10 is used to collect original terminal data generated by multiple terminals, wherein the original terminal data includes original equipment terminal data for production equipment operation and original quality inspection terminal data for plastic product quality inspection, and the original equipment terminal data and the original quality inspection terminal data are associated with each other through a unique identifier; The terminal data access module 20 is used to access the original terminal data generated by multiple terminals collected by the terminal data collection module 10 into the system; The protocol conversion module 30 is used to perform protocol analysis and data conversion on the original terminal data generated by multiple terminals and collected by the terminal data collection module 10 and accessed by the terminal data access module 20 to obtain target terminal data in a unified format, and transmit the target terminal data, wherein the target terminal data includes target device terminal data and target quality inspection terminal data; The data processing module 40 is configured to receive the target terminal data transmitted by the protocol conversion module 30 and process and analyze the target terminal data to determine an adjustment control strategy; The application module 50 receives the adjustment control strategy determined by the data processing module 40 to adjust the operation of the device terminal.

[0026] The MES systems currently used in the production process of plastic products on the market have obvious deficiencies in quality control: most systems only support "post-quality inspection" and lack real-time data perception capabilities; the quality inspection process is separated from the process, and a closed loop of parameter adjustment cannot be formed. This patent constructs a plastic product production and operation system based on data-driven and multi-terminal collaboration through a closed-loop process of data collection-access-conversion-processing-application, uniformly accessing equipment terminal data and quality inspection terminal data from multiple terminals, breaking the "information island" and realizing real-time linkage between production data and quality data; in the production process of plastic products, through the closed loop of collection-access-conversion-processing-application, data is converted into a direct driving force for quality optimization, avoiding the lag and subjectivity of traditional manual experience-driven control, enabling manufacturers to achieve a shift in plastic product quality from experience-based control to data-based control, and through real-time monitoring and parameter optimization, reducing quality problems caused by equipment operation fluctuations or process deviations, improving the quality of plastic products, and reducing the defective rate.

[0027] Furthermore, the terminal data collection module 10 includes an equipment terminal data collection unit and a quality inspection terminal data collection unit, which collect the raw equipment terminal data and the raw quality inspection terminal data in real time. In this embodiment, equipment terminals refer to production equipment at the plastic product production site, such as injection molding machines and extruders. A plastic product production line may include multiple different equipment terminals. Quality inspection terminals include equipment used to perform product quality inspection during the plastic product production process. In this embodiment, product quality inspection is "in-process quality inspection."

[0028] The device terminal data acquisition unit is configured to collect the raw device terminal data, including the device terminal's operating status data and process parameter data. Specifically, the device terminal data acquisition unit may be various sensors, industrial robots, programmable logic controllers (PLCs), or computer numerical control (CNCs).

[0029] During the production process of plastic products, MES associates the molds used by the current equipment, production products, production personnel, production processes, and production procedures. When the production order is executed, the equipment terminal data acquisition unit collects the equipment terminal production operation-related parameters in real time, such as using sensors to collect temperature, pressure, vibration and other data when the production equipment is running. By connecting various sensors, such as temperature sensors, pressure sensors, displacement sensors, flow sensors, etc., the equipment terminal's operating status data and process parameter data are obtained in real time.

[0030] Specifically, take an injection molding machine as an example. For an injection molding machine, its operating status data includes: Basic status: power on / off status, operating mode (manual / automatic / debugging), equipment alarm information (fault code / alarm time), equipment utilization (idle / operating / downtime); Key component status: motor load current (to determine whether it is overloaded), hydraulic system pressure (whether the oil pressure is normal), oil temperature, water temperature (to prevent overheating damage), and lubrication system status (oil level, oil pressure).

[0031] For injection molding machine equipment, its process parameter data includes: Injection molding stage parameters: injection pressure, speed, position, holding pressure, time, melt temperature (temperature of each section of the barrel), mold temperature (mold temperature controller data); Cycle parameters: cycle time (total cycle, injection / cooling / mold opening and closing time), shot volume, screw position, screw speed.

[0032] The quality inspection terminal data acquisition unit is used to collect the original quality inspection terminal data obtained from the quality inspection of the products during the production process of plastic products, wherein the original quality inspection terminal data includes real-time data and non-real-time data. Preferably, the quality inspection terminal data acquisition unit is a multi-terminal collaborative quality inspection execution unit. Optionally, such as mobile terminals, including PADs, smart phones, industrial touch screens or barcode scanners + small-screen terminals, or other quality inspection data acquisition devices, such as self-service quality inspection machines, the quality inspection terminal data is collected through manual entry, barcode scanning recognition or other methods. Supports the simultaneous execution of quality inspection and task records by multiple terminals to improve the efficiency of front-line execution.

[0033] Specifically, in one feasible embodiment, quality inspections are conducted according to the quality inspection requirements assigned by the MES. To promptly determine whether there are any quality issues with plastic products, indicators closely related to plastic product quality are pre-determined and quality inspections are conducted on the plastic products based on these indicators. These indicators include the plastic product's appearance quality, product dimensions, and mechanical properties. Appearance quality includes surface smoothness, the presence of scratches, and bubbles, while mechanical properties include tensile strength, yield strength, and elongation. Quality inspections are conducted through manual spot checks, shipping inspections, and self-service quality inspection machines, generating a large amount of raw quality inspection terminal data for plastic products. This data includes information such as the product's various quality indicator values, production batches, production times, and production equipment.

[0034] During the production process, corresponding quality inspection documents are generated in sequence for incoming material inspection, initial production inspection, production process inspection, and product warehousing inspection. Work is automatically assigned according to the quality inspection category. Production personnel can view their personal quality inspection assignments through industrial PAD, mobile APP, PC background, and intelligent testing equipment, and upload the quality inspection results to the system.

[0035] Self-service quality inspection machines automatically transmit inspection results to the production system. These machines boast multimodal inspection capabilities, utilizing high-resolution CMOS (Complementary Metal-Oxide Semiconductor Image Sensor) and the ResNet50 AI algorithm to detect cosmetic defects, laser triangulation and CCD (Charge-Coupled Device) imaging for dimensional measurement, confocal microscopy for surface roughness, and deep vision ToF (Time of Flight) cameras for assembly integrity. The self-service quality inspection machines are integrated with the system to transmit inspection results.

[0036] Preferably, the original equipment terminal data and the original quality inspection terminal data are associated with each other through a unique identifier to form a complete product quality file. The unique identifier may be a batch number, a serial number or a timestamp.

[0037] Furthermore, the terminal data access module 20 includes an industrial IoT gateway that connects the terminal data generated by the equipment terminals and the quality inspection terminals to the system. Throughout the plastic product production process, the raw terminal data generated by different terminals may be transmitted using different proprietary communication protocols. Therefore, the terminal data access module 20 supports multi-protocol access, such as Modbus / TCP, OPC-UA, and MQTT.

[0038] Optionally, for multiple terminal data acquisition modules 10 that communicate using the same communication protocol, a distributed control method can be adopted to connect the multiple terminal data acquisition modules 10 to the same CAN bus and use the CAN bus protocol for communication. For example, if multiple terminal data acquisition modules 10 all use the Modbus communication protocol, the multiple terminal data acquisition modules 10 are connected to the same CAN bus and use the CAN bus to communicate through the Modbus communication protocol, thereby achieving access to the original terminal data.

[0039] Furthermore, the protocol conversion module 30 parses and converts raw terminal data transmitted using different communication protocols to obtain target terminal data in a unified format, and then transmits the target terminal data. Because the raw terminal data from different terminal data acquisition modules 10 may be transmitted using different proprietary communication protocols, the protocol conversion module 30 supports multi-protocol parsing, analyzing the format, semantics, and communication protocol of the raw terminal data. This protocol conversion module 30 is compatible with multiple industrial communication protocols, reducing dependence on specific equipment brands and improving system versatility.

[0040] In addition, the formats of the raw terminal data collected by different terminal data collection modules 10 vary, resulting in heterogeneity. Therefore, after obtaining the raw terminal data through protocol parsing, it is necessary to convert the raw terminal data into target terminal data in a unified format. Optionally, the raw terminal data can be converted into JSON or XML format to eliminate data format barriers from different terminals.

[0041] Furthermore, the data processing module 40 processes and analyzes the target terminal data in a unified format to determine an adjustment control strategy for the device terminal.

[0042] Optionally, in a feasible embodiment, the data processing module 40 includes a data preprocessing unit and a data analysis unit.

[0043] The data preprocessing unit cleans and filters the target terminal data. Specifically, it removes outliers and fills missing data to improve data quality, such as using interpolation to fill missing data. If units do not conform to the standard, the units are unified, such as from mm to inches, and the target terminal data is standardized through normalization. The standardized data is stored in a local database for subsequent data analysis.

[0044] Preferably, the target device terminal data and the target quality inspection terminal data are stored separately, the target device terminal data is saved to the device database, and the target quality inspection terminal data is saved to the quality inspection database. Optionally, the standardized data is stored in a cloud platform, and remote access and backup are supported.

[0045] Furthermore, when storing data on devices, using a tiered storage architecture allows for flexible adjustment of storage resource allocation based on different business needs and data characteristics, improving the efficiency and performance of the entire storage system. Furthermore, this tiered storage architecture improves data reliability and availability, reducing the risk of data loss by backing up and storing redundant data at different levels.

[0046] The data analysis unit is configured to analyze the pre-processed target terminal data. The pre-processed target terminal data is first analyzed using a quality inspection anomaly algorithm. Based on the data analysis results, an equipment operation adjustment plan or optimization suggestion for the equipment terminal is automatically generated, such as adjusting the injection pressure of the equipment terminal from 80 MPa to 85 MPa.

[0047] For example, when the target quality inspection terminal data is analyzed to be abnormal, such as appearance defects, dimensional errors or surface roughness, or assembly integrity abnormalities, indicating that the plastic product has defects, the target quality inspection data of the defective plastic product is analyzed with the target device terminal data of the equipment terminal in the production process to determine the abnormal data in the target device terminal data that affects the defects of the plastic product, and then an adjustment control strategy for the equipment terminal is given to optimize the operating status data and / or process parameter data of the equipment terminal, so that the production of plastic products can be restored to normal in a timely manner, and the qualified rate of plastic products can be improved, which solves the problem in the prior art that the system cannot be linked with the terminal equipment and there are information islands between equipment parameters and product quality.

[0048] Preferably, to better analyze the relationship between the target device terminal data and the target quality inspection terminal data, and to improve the efficiency of locating abnormal target device terminal data when plastic product defects occur, by comparing product quality data from different batches and under different production conditions, the relative change patterns and trends of the target device terminal data and the target quality inspection terminal data under quality fluctuations are determined. Preferably, a correlation analysis method is used to determine the target terminal device data that is significantly correlated with product quality indicators. This allows the corresponding target terminal device data to be quickly located when a product quality indicator becomes abnormal, thereby optimizing the operating status data and / or process parameter data of the device terminal.

[0049] For example, if analysis reveals unstable product dimensional accuracy, it may be necessary to focus on parameters such as mold wear, injection pressure, and hold time. If significant appearance quality issues, such as bubbles or flow marks, are present, the impact of factors such as injection temperature and injection speed may need to be considered. To address product quality issues caused by unstable injection mold temperature, the data processing module 40 provides a control strategy for adjusting mold temperature control parameters during the injection molding process to maintain the mold temperature within an appropriate range.

[0050] Specifically, the association analysis algorithm used in the correlation analysis is:

[0051] Where, X Indicates the terminal data of the equipment (such as injection pressure, melt temperature), Y Indicates quality inspection terminal data (such as tensile strength, surface roughness), Cov is the covariance, σ is the standard deviation.

[0052] By calculating the process parameters (i.e. X ) and quality indicators (i.e. Y ) to identify the key parameters that have the greatest impact on product quality. For example, analyze the correlation between injection molding parameters (injection pressure, melt temperature) and the mechanical properties of plastic products (tensile strength, surface roughness).

[0053] The random forest algorithm can be used to calculate the importance of features, quantify the influence of each process parameter on the quality index, and identify the parameters such as injection pressure and melt temperature that have the greatest impact on product strength (importance score exceeds 0.7).

[0054] In addition, root cause analysis can be used to locate the key influencing factors of plastic product quality problems through methods such as causal inference and association rule mining.

[0055] Furthermore, in this embodiment, the data processing module 40 presets a qualified range for quality indicators and a reasonable range for initial target device terminal data. When a plastic product fails quality inspection but the initial target device terminal data remains within the reasonable range, or when a plastic product passes quality inspection but the initial target device data falls outside the reasonable range, the reasonable range for the target device terminal data is optimized to obtain an optimized reasonable range for the target device terminal data, thereby dynamically adjusting the reasonable range for device terminal data in plastic product production. The reasonable range for the initial target device terminal data can be determined based on the quality indicator requirements of the plastic product through theoretical calculation and simulation, and verified through small-batch trial production to obtain an initial standardized process card or operating instruction for plastic product production.

[0056] In another feasible embodiment, to more accurately determine the relationship between the target quality inspection terminal data and the target device terminal data, a quality prediction neural network model is pre-established based on learning and training based on a large amount of historical data, such as training based on a long short-term memory (LSTM) artificial neural network. An adjustment and control strategy for the device terminal is determined based on the quality prediction neural network model. Specifically, the neural network model predicts the product quality of plastic products obtained when the device terminal is produced under different operating status data and / or process parameter data, thereby providing a reference for production optimization. During plastic product production, analysis of the target device terminal data based on the quality prediction neural network model can identify potential device terminal failures in advance, perform preventive maintenance, and ensure the stable operation of production equipment. Based on real-time data and the quality prediction neural network model, relevant parameters of the device terminal are dynamically adjusted, achieving a transition from "post-detection" to "pre-prevention" and "in-process regulation" in plastic product quality control.

[0057] In another feasible embodiment, cluster analysis is employed to better identify the target device terminal data that may impact specific product quality issues. From the collected target device terminal data, target device terminal data that may potentially impact specific product quality is selected as input variables for the cluster analysis. For example, for the appearance of a plastic product, operating status data and process parameter data of the device terminal that may potentially impact the appearance are selected. For the dimensional accuracy of a plastic product, operating status data and process parameter data related to molding dimensions that may potentially impact the dimensional accuracy are selected.

[0058] We then analyze the characteristics of the data points within each cluster to identify commonalities and differences in product quality issues within the cluster. For example, if products within a cluster all exhibit the same type of cosmetic defect and a process parameter has a similar value range or variation trend, we can preliminarily determine a correlation between that process parameter and the cosmetic defect.

[0059] Furthermore, the data processing module 40 supports comprehensive quality assessment. For example, if the self-service quality inspection machine in the quality inspection terminal data collection unit has multimodal detection capabilities, the results of different detection modalities can be integrated into a final quality score through a weighted fusion algorithm.

[0060]

[0061] Where, Indicates the i The quality score of the detection modality, Represents the corresponding weight, satisfying =1, n Indicates the number of detection modes, weight It can be adjusted dynamically according to the importance of detection.

[0062] The target terminal data obtained from each detection modality of the self-service quality inspection machine is given a quality score and a corresponding weight, and a final quality score is obtained for the multimodal detection results. Results with a final quality score lower than the set quality score are also considered to have product defects. In fact, the target terminal data obtained based on any detection modality may indicate that the plastic product is qualified, but the multimodal accumulation may show that the plastic product has potential quality problems. Therefore, the results obtained by the multimodal quality inspection data fusion algorithm of the present invention can more accurately determine whether there are quality problems with the plastic product, avoid the problem of lack of integration of results from different detection equipment, and significantly reduce the missed detection rate.

[0063] Preferably, the data processing module 40 of the present invention dynamically adjusts equipment process parameters based on quality indicator errors based on PID (Proportional Integral Derivative) control theory. In plastics production, when product quality indicators deviate from target values, the process parameter values ​​that need to be adjusted are automatically calculated, achieving closed-loop control. The optimization algorithm for the dynamic adjustment control strategy is as follows:

[0064] Where, For the time to be optimized t In the plastic product quality control system scenario, the adjustment value is the specific adjustment value for the production equipment operation. For example, adjusting the injection molding machine's injection speed can affect the production process of plastic products by changing the injection speed, thereby affecting product quality. The quality index error is the difference between the target value and the actual value. For example, when measuring a certain quality index of a plastic product (such as tensile strength, size specifications, etc.), the pre-set target quality index value is subtracted from the value actually detected by the quality inspection terminal data acquisition unit. The difference is , which reflects the degree to which the current product quality indicators deviate from expectations; is the proportional coefficient, according to the quality index error at the current moment The size of the control function is proportional to the quality index. When the quality index error is large, Multiplying the error value by a larger control amount adjustment part can make the system respond quickly and adjust in the direction of reducing the error; when the error is small, the corresponding control amount adjustment part is also small; The error of quality index From time 0 Time tIntegration reflects the accumulation of quality indicator errors over a period of time; is the integral coefficient, the integral term From the initial moment 0 To the current moment t Integrate the quality index error, Multiplying this integral value accumulates past errors, even if the current error Although it is very small, as long as there is a certain amount of error accumulation in the past, the integral term will generate a control quantity to eliminate the steady-state error of the system and make the system output closer to the target value; Quality index error The derivative with respect to time reflects the changing trend of the error, describing whether the error is increasing rapidly, decreasing, or changing relatively slowly; is the differential coefficient, Multiply by the differential term , and can make adjustments in advance based on the rate of error change. If the error changes quickly, the differential term will generate a large control variable, allowing the system to react in advance to prevent the error from increasing or decreasing further, thus playing a "predictive" role and enhancing system stability.

[0065] The traditional PID control parameters are fixed, which makes it difficult to adapt to the nonlinear and time-varying characteristics of the production process. The adjustment process is slow and prone to overshoot. This invention adopts an adaptive PID algorithm to dynamically adjust the 、 、 Parameters can be adjusted to meet the needs of different products and production stages. For example, in injection molding machine injection speed control, the adaptive PID algorithm can shorten the adjustment time by 50% and reduce overshoot by 70%. Dynamic adjustment of the control strategy can improve product quality stability by 25%.

[0066] Optionally, when facing multi-objective optimization, such as optimizing product quality and production efficiency at the same time, weight coefficients can be introduced to balance different objectives:

[0067] Where, is the quality weight, which is adjusted dynamically according to production demand. is the quality error, For production costs.

[0068] Furthermore, the application module 50 includes an early warning unit, a process feedback optimization unit and a kanban unit.

[0069] The early warning unit is configured to issue an alarm. When the data processing module 40 processes and analyzes the target terminal data and discovers potential risks that could affect the quality of plastic products, an alarm is issued. In one feasible embodiment, an alarm is issued when the target quality inspection terminal data exceeds a preset quality acceptance standard, such as when a dimensional deviation exceeds a tolerance or an appearance defect rate exceeds a set threshold, or when data from the target equipment terminal is abnormal; when process parameter data such as injection temperature, pressure, and time exceed normal fluctuations, potentially affecting product quality; or when operational status data from the equipment terminal, such as motor speed or hydraulic system pressure, is abnormal, potentially affecting production stability and product quality.

[0070] The process feedback optimization unit is used to send the adjustment control strategy determined by the data processing module 40 to production personnel, who then further determine whether to adjust the operation of the equipment terminal. Alternatively, the process feedback optimization unit can directly send the strategy to the edge controller of the equipment terminal to automatically adjust the operation of the equipment terminal, thus solving the problem of production personnel having difficulty making accurate decisions based on data. The following uses the operation adjustment of an injection molding machine as an example to illustrate.

[0071] Specifically, in one example, after the data processing module 40 determines the adjustment control strategy, the process feedback optimization unit automatically transmits the new device terminal data (i.e., the recommended optimization parameters) for the injection molding machine to the corresponding edge controller of the injection molding machine to adjust the injection molding machine's operation. For example, based on the adjustment control strategy, the temperature, pressure, and time of the injection molding machine are automatically adjusted to ensure that the injection molding machine operates in accordance with the new process requirements. If a deviation is found between the actual production data and the optimization recommendations of the adjustment control strategy, the system automatically performs fine-tuning. If the injection molding machine edge controller is unable to receive the new device terminal data, the system prompts production personnel to check the device communication connection and provides appropriate maintenance instructions.

[0072] Specifically, in another example, after the data processing module 40 determines the adjustment control strategy, the process feedback optimization unit transmits the adjustment control strategy to production personnel, who then manually adjust the injection molding machine's device terminal data based on the adjustment control strategy, supporting manual intervention decisions. During the adjustment process, detailed information such as the adjustment time, parameter values, and production personnel is recorded. After the adjustment is complete, a quality management or technical staff member reviews the adjusted device terminal data to ensure that it meets the optimization recommendations of the adjustment control strategy. If any issues are found with the new device terminal data settings or if they adversely affect product quality, further adjustments and improvements are made.

[0073] The kanban unit is used to visually display the processing and analysis results of the data processing module 40, thereby realizing visual management of plastic product quality results.

[0074] Furthermore, in one feasible embodiment, the early warning unit has a pre-set early warning mechanism. Using static and dynamic thresholds, thresholds for different parameters are pre-set based on data content. The system monitors in real time within these threshold ranges. A multi-level adaptive intelligent early warning model with customized learning mechanisms is also included. Model judgment dimensions can be continuously refined during the production process. Through in-depth model learning and combined judgment with multiple conditions from other data, a multi-level early warning trigger mechanism is implemented. Fixed trigger mechanisms and corresponding requirements are set for different warning levels. For lower-level warnings, processing time can be limited, while for higher-level warnings, a shutdown requirement can be directly returned, causing the terminal equipment to shut down immediately. An early warning suppression strategy is added to the early warning process, merging repeated warnings within a short period of time while preventing subsequent warnings from being suppressed due to post-trigger warnings. The system also features flexible detection capabilities, automatically reducing detection frequency during non-production periods based on production conditions. Generated warning information is communicated via the workshop's ANDON light system, internal platform messaging, and text messages. This data-based early warning mechanism prevents quality issues with plastic products from occurring early, allowing for early detection and timely adjustments to equipment operations, thereby increasing the yield of qualified plastic products.

[0075] For example, in a feasible embodiment, for key equipment such as injection molding machines, parameters such as motor load current and hydraulic system pressure are monitored in real time, and an early warning is triggered when the parameters deviate from the normal range:

[0076] Where, The terminal data of the current target device. and are the mean and standard deviation of historical data, To prevent small constants from being zero in the denominator, AnomalyScore is an anomaly score. When AnomalyScore > Threshold, it is considered an anomaly. A statistical model is built based on historical device terminal operation data, and anomaly detection is performed by calculating the degree of deviation between current parameters and historical distributions.

[0077] Adopt dynamic threshold mechanism and update in real time through sliding window and , adapt to changes in the operating status of the adaptive equipment and reduce the false alarm rate.

[0078] Furthermore, to ensure the practicality of the early warning mechanism, the system will analyze the early warning accuracy (TP / FP) every month, review the effectiveness of the rules quarterly, and update the algorithm model annually.

[0079] Optionally, when the system receives data on unqualified plastic products detected by the self-service quality inspection machine, the system performs exception processing on the unqualified products, removes the products from the qualified products by adjusting the product conveying direction, distinguishes qualified products from unqualified products, and triggers an exception alarm by the system.

[0080] By collecting data in real time and analyzing the data by the data processing module 40, production anomalies, such as equipment parameter exceeding the limit, can be discovered in time, and then an early warning mechanism can be used to issue an early warning, thereby reducing the defective rate of plastic products.

[0081] Furthermore, the application module 50 also includes a quality traceability unit. Because the quality inspection terminal data for plastic products includes information such as the values ​​of various quality indicators, production batches, production times, and production equipment, a detailed quality history file is established for each product, recording all quality-related information during the production process. Furthermore, the equipment terminal data is linked to the original quality inspection terminal data via a unique identifier. Therefore, the quality traceability unit enables the traceability of plastic products. When quality issues arise with plastic products, the production batch, production date, production equipment, and other information of the problematic product, as well as the operating status data and process parameter data of the equipment terminal, can be quickly located, achieving complete quality traceability of the plastic product through a "one item, one code" system.

[0082] In the plastics manufacturing industry, product traceability and documentation are essential for meeting regulatory requirements. For example, in the injection molding of automotive parts, regulations require companies to establish comprehensive product traceability systems to enable timely product recalls in the event of quality issues. The creation and management of documentation can help companies meet regulatory requirements and avoid the risks and losses associated with non-compliance. This invention improves delivery reliability by outputting a complete product quality history.

[0083] Optionally, the data-driven, multi-terminal collaborative plastics product production and operations system also includes a data interaction module for data exchange with MES, equipment management software, and ERP. This module transmits raw terminal data collected by the terminal data collection module 10 to the MES, equipment management software, and ERP. Alternatively, in another feasible embodiment, it transmits data pre-processed by the data processing module 40 to the MES, equipment management software, and ERP. To ensure data real-time, reliability, and security, data retransmission is achieved through local cache queues (SQLite / RocksDB), data tagging (sequence_id + checksum), and windowed retransmission (TCP-like mechanism). Apache Kafka, a messaging middleware, is used to support high-throughput data flows, and RabbitMQ is used for reliable transaction messaging.

[0084] Optionally, the data processing module 40 is connected to the application module 50 via an API interface unit. Specifically, the API interface is divided into four types: real-time data, historical query, file transfer, and command issuance, corresponding to WebSocket, REST, SFTP, and MQTT+REST protocols respectively.

[0085] Optionally, the data-driven, multi-terminal collaborative plastic product production and operations system also includes an effectiveness evaluation module for evaluating the effects of adjustments to the operation of the equipment terminals made according to the adjustment control strategy determined by the data processing module 40. Specifically, evaluation indicators are set, and effectiveness evaluation is performed based on changes in these indicators before and after the adjustments, such as product qualification rate, production cycle, and equipment failure rate. By comparing the product qualification rate, production cycle, and equipment failure rate of plastic products before and after the adjustments, the effectiveness of process optimization of the equipment terminals using the adjustment control strategy is evaluated. Based on the effectiveness evaluation results, the operation of the equipment terminals is continuously optimized.

[0086] Comparing and analyzing various indicators after optimization based on the adjusted control strategy with those before optimization provides a more intuitive understanding of the effects of the optimization measures. For example, a comparison reveals that the product qualification rate increased from 80% to 90% and the production cycle was shortened from 30 seconds to 25 seconds, demonstrating that process optimization based on the adjusted control strategy has achieved certain results in improving product quality and production efficiency. This invention improves the stability of plastic product quality by intervening in the production process through a data-driven control system.

[0087] Example 2 As another specific embodiment of the present invention, the present invention provides a plastic product production and operation method based on data drive and multi-terminal collaboration, based on the system implementation of embodiment 1, referring to Figure 3 , the specific steps are as follows: S100. Collecting original terminal data generated by multiple terminals through a terminal data collection module, wherein the original terminal data includes original equipment terminal data for production equipment operation and original quality inspection terminal data for plastic product quality inspection, and the original equipment terminal data and the original quality inspection terminal data are associated with each other through a unique identifier; S200, the terminal data access module accesses the original terminal data generated by multiple terminals collected by the terminal data collection module into the system; S300, using a protocol conversion module to perform protocol parsing and data conversion on the original terminal data generated by the multiple terminals and collected by the terminal data collection module and accessed by the terminal data access module to obtain target terminal data in a unified format, and transmit the target terminal data, wherein the target terminal data includes target device terminal data and target quality inspection terminal data; S400, receiving the target terminal data transmitted by the protocol conversion module through the data processing module and processing and analyzing the target terminal data to determine an adjustment control strategy; S500: Utilize an application module to receive the adjustment control strategy determined by the data processing module to adjust the operation of the device terminal.

[0088] The plastic product production and operation method based on data-driven and multi-terminal collaboration of the present invention associates the original equipment terminal data for production equipment operation and the original quality inspection terminal data for plastic product quality inspection collected in real time by multiple terminals, and constructs a data-based process optimization feedback mechanism. It can timely adjust the operation status of equipment terminals during the production process of plastic products, reduce the number of unqualified products, improve product quality, and thus reduce resource waste; determine adjustment control strategies through data analysis, reduce manual dependence, and provide support for production personnel's decision-making; and make adjustment control strategies determined based on data more accurate, thereby reducing repeated debugging and rework and improving production efficiency.

[0089] The above description is merely a preferred embodiment of the present invention and does not constitute any other form of limitation to the present invention. Any modification or equivalent variation based on the technical essence of the present invention shall still fall within the scope of protection claimed by the present invention.

Claims

1. A plastic product production and operation system based on data drive and multi-terminal collaboration, characterized by: The system comprises a terminal data acquisition module (10), a terminal data access module (20), a protocol conversion module (30), a data processing module (40), and an application module (50); The terminal data acquisition module (10) is used to collect original terminal data generated by multiple terminals, wherein the original terminal data includes original equipment terminal data for production equipment operation and original quality inspection terminal data for plastic product quality inspection, and the original equipment terminal data and the original quality inspection terminal data are associated through a unique identifier; The terminal data access module (20) is used to access the original terminal data generated by multiple terminals collected by the terminal data collection module (10) into the system; The protocol conversion module (30) is used to perform protocol analysis and data conversion on the original terminal data generated by multiple terminals and collected by the terminal data collection module (10) and accessed by the terminal data access module (20) to obtain target terminal data in a unified format, and transmit the target terminal data, wherein the target terminal data includes target device terminal data and target quality inspection terminal data; The data processing module (40) is used to receive the target terminal data transmitted by the protocol conversion module (30) and process and analyze the target terminal data to determine an adjustment control strategy; The application module (50) receives the adjustment control strategy determined by the data processing module (40) to adjust the operation of the device terminal.

2. The plastic product production and operation system based on data drive and multi-terminal collaboration according to claim 1 is characterized in that: The original quality inspection terminal data includes quality index data, and the data processing module (40) uses an adaptive PID algorithm to determine the adjustment control strategy: Where, For the time to be optimized t The adjustment amount of the device terminal, is the quality indicator error, is the proportional coefficient, according to the quality index error at the current moment The size of the control is proportional to the The error of quality index From time 0 Time t Integration reflects the accumulation of quality indicator errors over a period of time. is the integration coefficient, Quality index error The time derivative, is the differential coefficient.

3. The plastic product production and operation system based on data drive and multi-terminal collaboration according to claim 2 is characterized in that: The terminal data acquisition module (10) comprises an equipment terminal data acquisition unit and a quality inspection terminal data acquisition unit, wherein the equipment terminal data acquisition unit is used to acquire the original equipment terminal data, and the quality inspection terminal data acquisition unit is used to acquire the original quality inspection terminal data, wherein the original equipment terminal data includes the operation status data and process parameter data of the equipment terminal, and the quality inspection terminal data acquisition unit acquires the original quality inspection terminal data through manual and / or self-service quality inspection machines.

4. The plastic product production and operation system based on data drive and multi-terminal collaboration according to claim 3 is characterized in that: The self-service quality inspection machine is a multimodal quality inspection machine that uses high-resolution CMOS and AI algorithms to detect appearance defects, achieves size measurement through laser triangulation and CCD image measurement, detects surface roughness through confocal microscopy, and detects assembly integrity through deep vision ToF camera.

5. The plastic product production and operation system based on data drive and multi-terminal collaboration according to claim 3 is characterized in that: The original equipment terminal data and the original quality inspection terminal data are associated with each other through a batch number, serial number or timestamp unique identifier to form a complete product quality file.

6. The plastic product production and operation system based on data drive and multi-terminal collaboration according to claim 3 is characterized in that: The data processing module (40) includes a data preprocessing unit and a data analysis unit. The data preprocessing unit cleans, filters, and normalizes the target terminal data to obtain the standardized target terminal data and then stores it. The data analysis unit analyzes the preprocessed target terminal data and automatically generates an adjustment control strategy for the device terminal based on the data analysis result.

7. The plastic product production and operation system based on data drive and multi-terminal collaboration according to claim 6 is characterized in that: The data processing module (40) uses a correlation analysis method to locate the target terminal device data that is significantly correlated with the product quality index: Where, X Indicates device terminal data, Y Indicates quality inspection terminal data, Cov is the covariance, σ is the standard deviation.

8. The plastic product production and operation system based on data drive and multi-terminal collaboration according to claim 6 is characterized in that: The data processing module (40) predicts the product quality of the plastic product obtained when the equipment terminal is produced under different operating state data and / or process parameter data using a quality prediction neural network model between the target quality inspection terminal data and the target equipment terminal data obtained through pre-training.

9. The plastic product production and operation system based on data drive and multi-terminal collaboration according to claim 6 is characterized in that: The application module (50) includes an early warning unit, a process feedback optimization unit and a kanban unit. The early warning unit is used to send an alarm externally. When the data processing module (40) processes and analyzes the target terminal data and finds that there may be hidden dangers that affect the quality of plastic products, an alarm is issued; The process feedback optimization unit is used to send the adjustment control strategy determined by the data processing module (40) to the production personnel or directly to the edge controller of the equipment terminal to adjust the operation of the equipment terminal.

10. A plastic product production and operation method based on data drive and multi-terminal collaboration, implemented based on the system according to any one of claims 1 to 9, characterized in that: The method comprises: The original terminal data generated by multiple terminals are collected by a terminal data collection module (10), wherein the original terminal data includes original equipment terminal data for production equipment operation and original quality inspection terminal data for plastic product quality inspection, and the original equipment terminal data and the original quality inspection terminal data are associated with each other through a unique identifier; The terminal data access module (20) accesses the original terminal data generated by multiple terminals collected by the terminal data collection module (10) into the system; The protocol conversion module (30) performs protocol parsing and data conversion on the original terminal data generated by the multiple terminals collected by the terminal data collection module (10) and accessed by the terminal data access module (20) to obtain target terminal data in a unified format, and transmits the target terminal data, wherein the target terminal data includes target device terminal data and target quality inspection terminal data; Receiving the target terminal data transmitted by the protocol conversion module (30) through the data processing module (40) and determining an adjustment control strategy after processing and analyzing the target terminal data; The application module (50) is used to receive the adjustment control strategy determined by the data processing module (40) to adjust the operation of the device terminal.

Images