Digital twin simulation and optimization method for large-scale stamping production line

Abstract

The invention relates to a digital twin simulation and optimization method for a large-scale stamping production line, and the method comprises a rapid lightweight modeling technology, a multifunctional fault indicator feedback technology, a three-dimensional dynamic real-time driving simulation technology, and a big data analysis intelligent optimization decision technology. According to the method, virtual-real fusion and a three-dimensional dynamic technology are combined, a three-dimensional visual user interface oriented to the whole manufacturing process is developed based on Unity, the whole stamping production line is simulated in real time, potential problems are predicted and solved through real-time monitoring of the stamping production line, various problems influencing the production efficiency in the production process are optimized, and the manual management and control difficulty is reduced.

Description

Technical Field

[0001] The present invention relates to the field of intelligent manufacturing technology, and in particular to a digital twin simulation and optimization method for a large-scale stamping production line. Background Art

[0002] Amid the global trend toward digitalization and intelligent manufacturing, manufacturing companies are facing the urgent task of improving production efficiency, reducing operating costs, and enhancing product quality. As a key component of the manufacturing industry, the digital transformation of stamping lines has become an irreversible trend. Digital twin technology, by creating virtual replicas of physical entities, enables real-time monitoring, prediction, and optimization of production processes. Introducing this technology in stamping lines allows for precise simulation of production processes, helping companies proactively identify potential issues, reduce downtime, and ultimately improve overall efficiency.

[0003] Building high-quality 3D models is the foundation for digitalizing stamping production lines. However, current large-scale digital stamping production line models suffer from slow construction speeds, low accuracy, insufficient lightweight processing capabilities, and poor simulation performance, making them difficult to effectively apply.

[0004] In intelligent stamping production lines, due to the complexity of the equipment and the high speed of operation, once a fault occurs, it is often difficult to quickly locate and correct it. Traditional methods that rely on manual inspections and empirical judgment are not only inefficient but also lack high accuracy. Therefore, there is an urgent need for a device that can monitor and indicate fault conditions in real time to optimize production line reliability and maintenance efficiency.

[0005] For highly integrated large-scale stamping production lines, core manufacturing resources such as precision presses, automated robotic arms, and customized chemical equipment molds are prone to problems such as reduced precision and increased wear under long-term, uninterrupted, high-load working conditions. Specifically: The press will gradually reduce the accuracy and stability of the hydraulic system, mechanical transmission components and work surface due to continuous high-intensity stamping operations, which directly affects the dimensional accuracy and shape consistency of the stamped parts.

[0006] As a key component for performing high-precision positioning and grasping tasks, the robotic arm may experience a decrease in operational accuracy during frequent operations due to factors such as joint wear and sensor accuracy drift, thereby affecting the efficiency and product quality of the entire production line.

[0007] As tools directly involved in stamping, tooling and die are particularly susceptible to issues such as material fatigue, wear, and thermal stress deformation. These issues not only affect the surface quality, dimensional accuracy, and structural integrity of stamped parts, but can also lead to mold damage, increasing production costs and maintenance frequency.

[0008] To ensure the continuous and stable operation of large-scale stamping lines and consistent product quality, regular equipment maintenance and precision calibration plans must be established, and advanced monitoring technologies must be used to promptly identify and resolve problems. Simultaneously, tooling and die design and material selection should be continuously improved to extend their service life and enhance production capacity. Furthermore, by collecting production line data in real time and feeding it into virtual simulation models, more refined control and predictive maintenance can be achieved over the production process. Adjusting production strategies based on simulation results will further enhance efficiency and quality. Summary of the Invention

[0009] In order to effectively solve the problems of slow speed and low precision of 3D modeling and simulation, high difficulty of manual control, weak business integration capability, slow fault repair, and no autonomous prediction capability in the stamping production lines currently used on a large scale in workshops, the present invention proposes a digital twin simulation and optimization method for large-scale stamping production lines. The digital twin technology is used to quickly construct a virtual mirror image of its entity. Combining the fusion of virtual and real with 3D dynamic technology, Unity is used to develop a 3D visual user interface for the entire manufacturing process. The entire stamping production line is simulated in real time. By real-time monitoring of the stamping production line, potential problems are predicted and solved, various problems affecting production efficiency in the production process are optimized, and the difficulty of manual control is reduced.

[0010] To achieve the above objectives, the technical solution of the present invention is: a large-scale stamping production line simulation and optimization method based on digital twin, comprising the following steps: S1. Simulate large-scale stamping production lines based on rapid lightweight modeling technology, build scenario models, and optimize business integration; S2. Collect, transmit, and process fault data from production line equipment through multi-function fault indicator feedback technology, and dynamically optimize data transmission protocols; S3. Leverage 3D dynamic real-time drive simulation technology to achieve intelligent matching of virtual and real scenes, data association, and model synchronization, while enabling online monitoring through the digital twin 3D visualization interface. S4. Based on big data analysis and intelligent optimization decision-making methods, we use deep neural network-driven data mining technology, combined with reinforcement learning and transfer learning algorithms to optimize production quality, equipment status and business processes.

[0011] Furthermore, the fast lightweight modeling technology includes: three-dimensional point cloud reverse modeling technology, component modeling technology, and full-process digital modeling technology.

[0012] Specifically, step S1 includes the following steps: S101. By collecting and connecting existing data, modeling is performed using 3D point cloud reverse modeling technology and component modeling technology; S102. Calling component models in a modular component library modeled using component-based modeling technology, matching the collected point cloud data with similar components in the modular component library, and performing intelligent parameter optimization on the matched components; S103. By integrating full-scale, dynamic and static data using full-process digital modeling technology, a digital twin with multi-dimensional integration of "physical-information-business" is formed, virtually simulating the entire manufacturing information and business process. S104. Build and deploy a 3D dynamic simulation model of a large-scale stamping production line based on Unity.

[0013] Specifically, the multifunctional fault indicator feedback technology in step S2 includes: Hardware part: including sensor, signal processing circuit and indicator module; Software part: including sensor signal processing, fault judgment logic and data transmission protocol; The collected multi-source heterogeneous data is transmitted to the digital twin platform through the industrial network, and the protocol incompatibility problem is solved using dynamic optimization algorithms, and then imported into the data middle platform for overall management.

[0014] Specifically, step S3 includes the following steps: S301. Build a digital twin 3D visualization interface based on Unity, and achieve 3D dynamic mapping between physical space and virtual scenes through multi-source heterogeneous data fusion; S302. Based on three-dimensional dynamic real-time drive simulation technology, intelligent matching of physical equipment and virtual scenes, and allowing users to adjust production parameters through the interface; S303. Through distributed computing and storage technology, real-time collection of physical equipment operating data is carried out, and the virtual model is updated synchronously to maintain consistency between the virtual and the real; S304. Use step S2 to monitor equipment status, personnel location, vehicle information, order status, and warehouse status, and feed multi-source heterogeneous data back to the digital twin 3D visualization interface; S305. Through the background simulation engine, the virtual scene is updated according to the real-time data and fault warnings and production optimization instructions are generated.

[0015] Specifically, step S4 includes the following steps: S401. Collect multi-source heterogeneous data from the production process, including product quality data, equipment status data, material information, and production progress data, and integrate cross-departmental data; S402. Perform feature extraction and anomaly detection on the fused data using manufacturing data mining technology driven by deep neural networks; S403. Explore optimal production strategies based on deep reinforcement learning algorithms and adapt strategy generalization capabilities to different production scenarios through deep transfer learning networks; S404. Dynamically verify and iteratively optimize the optimal production strategy using a deep adversarial learning network; S405. Use deep autoencoders to compress and reconstruct the features of redundant, repeated, and missing data in the entire business process, generate a standardized business process diagram, and output optimized decision instructions to the upper-level system.

[0016] Specifically, the abnormality detection mentioned in step S402 specifically includes: Use the production schedule anomaly discrimination model to identify schedule deviations; Use production quality defect detection models to locate defect causes; Use equipment fault status identification model to predict equipment failure risk.

[0017] Beneficial effects of the present invention: First, in response to the functional requirements of large-scale stamping production lines, we deeply integrated the core concepts of digital twins to build a highly versatile and customized system architecture. This architecture not only defines the system hierarchy and functional modules between large-scale stamping production lines and digital twin systems, but also integrates the digital twin platform, providing a solid architectural foundation and clear guidance path for the stable construction of the system framework and the implementation of technical details. This significantly optimizes the overall collaborative efficiency and scalability of large-scale stamping production lines and provides a reliable framework for the basic modeling of digital twins in large scenarios. 2. For large-scale stamping production lines, 3D point cloud reverse modeling technology and component modeling technology are used to achieve the optimized configuration of model assets and the extremely fast deployment of large-scale scene modeling, and to the greatest extent possible, build intelligent scenes that match reality at the fastest speed. This lays a solid foundation for the accurate simulation and efficient optimization of large-scale stamping production lines. This solution not only accelerates the construction process of digital twin models, but also ensures the high fidelity and maintainability of the models. From a series of optimization processes such as suppressing subtle structures, processing surface features, and stitching coupling gaps, an integrated simulation from tiny details to the overall structure is achieved, achieving highly simulated millimeter-level modeling and realistically restoring real scenes; 3. Intelligent decision-making optimization technology based on big data analysis: Integrating advanced monitoring and analysis technologies, using big data analysis methods, and conducting in-depth mining of massive data in the production process, it can identify and warn of potential problems in advance, so that countermeasures can be taken quickly. In addition, by continuously optimizing the design scheme and material selection strategy of tooling and molds, the service life of key components has been effectively extended, and the overall production efficiency and product quality have been significantly improved, bringing significant economic benefits and competitiveness to the company. Through the deep integration of deep learning of models and algorithm learning, big data can combine various past and existing data to make the best decision on product expectations, optimize business processes, and sort out the logical relationships and sequences between various designs, R&D, procurement, etc. on the original basis, saving a lot of manpower and time costs; 4. Deploy various sensors on the stamping production line equipment to collect real-time data on the equipment's operating status. This data is transmitted to the digital twin platform via the industrial network and integrated with the virtual production line model, resolving the issues of incomplete data collection and long production lines that make it difficult to detect some problems in a timely manner. At the same time, the system improves the single function of fault indicators and their lack of autonomous deep learning capabilities. Using a visual interface, the status information of the fault indicator is displayed on the production line site or on the visual platform. In addition, fault information is remotely notified to relevant personnel via SMS, email, or mobile applications to ensure timely processing and avoid production line shutdowns. This improves fault detection efficiency, reduces manual inspection and troubleshooting time, and reduces labor costs. At the same time, it enhances production line reliability, enables timely detection and resolution of faults, and ensures stable operation. Maintenance management is also optimized, improving the efficiency of the stamping production line. Distributed computing and storage technologies are used to provide fault data and historical records to support production line maintenance and optimization. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. The drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. For those skilled in the art, other drawings can be obtained based on the drawings without paying any creative work.

[0019] Figure 1 This is a diagram of the system architecture of the present invention; Figure 2 It is a schematic diagram of the operating mode of the present invention. DETAILED DESCRIPTION

[0020] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0021] like Figure 1-2 The simulation and optimization method of a large-scale stamping production line based on digital twins shown in FIG. includes the following steps: S1. Simulate large-scale stamping production lines based on rapid lightweight modeling technology, build digital twin models of physical scenarios, and optimize business integration; S2. Collect, transmit, and process fault data from production line equipment through multi-function fault indicator feedback technology, and dynamically optimize data transmission protocols; S3. Leverage 3D dynamic real-time drive simulation technology to achieve intelligent matching of virtual and real scenes, data association, and model synchronization. Use Unity to build a digital twin 3D visualization interface for online monitoring. S4. Based on big data analysis and intelligent optimization decision-making technology, we use deep neural network-driven data mining technology, combined with reinforcement learning and transfer learning algorithms to optimize production quality, equipment status and business processes.

[0022] Wherein, the step S1 includes the following steps: S101. By collecting existing data and connecting them together, modeling is carried out using 3D point cloud reverse modeling technology and component modeling technology; S102. Calling component models in a modular component library modeled using component-based modeling technology, matching the collected point cloud data with similar components in the modular component library, and performing intelligent parameter optimization on the matched components; S103. By integrating full-scale, dynamic and static data using full-process digital modeling technology, a digital twin with multi-dimensional integration of "physical-information-business" is formed, virtually simulating the entire manufacturing information and business process. S104. Build and deploy a 3D dynamic simulation model of a large-scale stamping production line based on Unity.

[0023] Wherein, the step S3 includes the following steps: S301. Build a digital twin 3D visualization interface based on Unity. Implement 3D dynamic mapping between physical space and virtual scenes through multi-source heterogeneous data fusion, enabling dynamic model loading and real-time rendering. S302. Based on model-driven technology, intelligently match physical equipment with virtual scenes and allow users to adjust production parameters through the interface; S303. Through distributed computing and storage technology, real-time collection of physical equipment operating data is carried out, and the virtual model is updated synchronously to maintain consistency between the virtual and the real; S304. Use sensors and fault indicators to monitor equipment status, personnel location, vehicle information, order status, and warehouse status, and feed the data back to the digital twin 3D visualization interface; S305. Through the background simulation engine, the virtual scene is updated according to the real-time data and fault warnings and production optimization instructions are generated.

[0024] The step S4 comprises the following steps: S401. Collect multi-source heterogeneous data from the production process, including product quality data, equipment status data, material information, and production progress data, and integrate cross-departmental data; S402. Perform feature extraction and anomaly detection on the fused data using manufacturing data mining technology driven by deep neural networks; S403. Explore optimal production strategies based on deep reinforcement learning algorithms and adapt strategy generalization capabilities to different production scenarios through deep transfer learning networks; S404. Dynamically verify and iteratively optimize the optimal production strategy using a deep adversarial learning network; S405. Use deep autoencoders to compress and reconstruct the features of redundant, repeated, and missing data in the entire business process, generate a standardized business process diagram, and output optimized decision instructions to the upper-level system.

[0025] The abnormality detection mentioned in step S402 specifically includes: Use the production schedule anomaly discrimination model to identify schedule deviations; Use production quality defect detection models to locate defect causes; Use equipment fault status identification model to predict equipment failure risk.

[0026] Specifically, such as Figure 1 As shown, the fast lightweight modeling technology includes: three-dimensional point cloud reverse modeling technology, component modeling technology, and full-process digital modeling technology.

[0027] 3D point cloud reverse modeling technology: Scan production line equipment and generate point cloud data to build a digital twin 3D model. Specifically, 3D point cloud reverse modeling technology based on large-scale point cloud fitting technology breaks through key technologies such as point cloud data slicing, optimization, and fitting, enabling rapid, high-precision, and lightweight modeling of digital twin factories and forming a modular component library. Through technical means such as point cloud data acquisition, preprocessing, surface reconstruction, scene rendering processing, and repeated feature matching, a digital twin 3D model is constructed. The relative position of equipment and the real scene can be changed at any time, optimizing the placement of different positions of equipment in the same or similar factories, laying the foundation for large-scale simulation. At the same time, it solves problems such as the lack of suppression of fine structure simulation, incomplete surface feature processing, and incomplete suture of coupling gaps in previous modeling.

[0028] Component-based modeling technology: Through predefined component libraries, parametric configuration and automated assembly, complex three-dimensional models can be quickly generated to improve efficiency and consistency. Specifically, component-based rapid modeling technology can split complex systems into multiple small, independent and reusable components, and use these components to quickly build and iterate models or applications. Componentization is flexible and reusable, achieving standardization and modularization of the development process, ensuring unified component interfaces and smooth data interaction. With the help of componentization and modularization, various scanned models can be identified and separated in the existing model library, and the model library can be used to build equipment models that are identical to reality, so that small parts of components can be aggregated into specific equipment, and stamping production lines can be quickly and intelligently simulated. At the same time, the scanning results will be classified and updated in real time in the component library to ensure that similar models are not repeated or omitted, which optimizes the original component library to a certain extent.

[0029] Full-process digital modeling technology: Based on the integration of full, dynamic, and static data at multiple scales, a digital twin is constructed that integrates the multidimensional "physical, informational, and business" aspects of the production process, enabling virtual simulation of manufacturing information and business processes. Specifically, while rapidly and meticulously modeling large-scale stamping production line equipment scenarios, this process is simulated within a visual interface, creating a digitally adjustable interface. This simulated business process eliminates issues such as disconnected communication and poor coordination between departments and components, integrating various business operations.

[0030] like Figure 1 As shown, the multifunctional fault indicator feedback technology includes: Hardware part: including sensor, signal processing circuit and indicator module; Software part: including sensor signal processing, fault judgment logic and data transmission protocol; By designing and deploying a fault indicator consisting of sensors, signal processing circuits, and indicator modules, seamless data interaction is achieved using the industrial network, and the collected multi-source heterogeneous data is transmitted to the digital twin platform in real time. Once a fault occurs, the system can quickly and instantly notify the relevant maintenance personnel of the detailed fault information remotely through various means such as text messages, emails, or mobile applications. This solves the problem of data transmission delay after a fault occurs. By integrating an external fault indicator with advanced algorithms, autonomous learning and intelligent judgment of production line faults are achieved. This ensures that the system can automatically issue an alarm and notify relevant personnel to carry out timely repairs as soon as a fault occurs. In addition, dynamic optimization and protocol compatibility issues in the data transmission process have been optimized. With the powerful coordination capabilities of digital twin technology, data is effectively integrated into the data center, realizing simulation and optimization of the entire data transmission process, thereby significantly improving the operating efficiency and reliability of the production line.

[0031] like Figure 2 As shown, the 3D dynamic real-time drive simulation technology simulates virtual and real scenes driven by 3D models through a Unity-based digital twin 3D visualization interface. Details of any link can be adjusted and deployed within the same interface. As the stamping line operates, the visualization interface is updated in real time, and real-time simulation of data and equipment is performed in the background. Through fault indicators and various sensors, a comprehensive simulation of multi-source heterogeneous data, including the entire site status, equipment status, personnel location, vehicle information, order status, and warehouse status, is performed. This allows operators to fully grasp all aspects of information, reducing time and labor costs.

[0032] like Figure 2 As shown, the big data intelligent optimization decision-making method proposes deep neural network-driven manufacturing data mining technology. Using big data analysis models for identifying production schedule anomalies, detecting production quality defects, and identifying equipment fault conditions, it explores optimal strategies through deep reinforcement learning algorithms, achieving intelligent production that self-discovers and self-solves problems during the production process. After deep learning, intelligent judgment is used to address challenges such as difficult data feature analysis, ambiguous classification and recognition, and low integration. By integrating duplication, redundancy, and omissions across various departments, such as R&D and procurement, a most appropriate business process diagram is created. This optimized business process also provides usable and effective data support for higher-level business systems.

[0033] The above description is merely a preferred embodiment of the present invention and does not constitute any form of limitation to the present invention. Although the present invention has been disclosed as above in terms of a preferred embodiment, it is not intended to limit the present invention. Any person skilled in the art can, without departing from the scope of the technical solution of the present invention, make some changes or modifications to equivalent embodiments using the technical contents disclosed above. However, any brief modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention are still within the scope of the technical solution of the present invention.

Claims

1. A large-scale stamping production line simulation and optimization method based on digital twin, characterized in that: The following steps are involved: S1. Simulate large-scale stamping production lines based on rapid lightweight modeling technology, build digital twin models of physical scenarios, and optimize business integration; S2. Collect, transmit, and process fault data from production line equipment through multi-function fault indicator feedback technology, and dynamically optimize data transmission protocols; S3. Leverage 3D dynamic real-time drive simulation technology to achieve intelligent matching of virtual and real scenes, data association, and model synchronization. Use Unity to build a digital twin 3D visualization interface for online monitoring. S4. Based on big data analysis and intelligent optimization decision-making technology, we use deep neural network-driven data mining technology, combined with reinforcement learning and transfer learning algorithms to optimize production quality, equipment status and business processes.

2. The large-scale stamping production line simulation and optimization method according to claim 1, characterized in that: The fast lightweight modeling technology includes three-dimensional point cloud reverse modeling technology, component modeling technology, and full-process digital modeling technology.

3. The large-scale stamping production line simulation and optimization method according to claim 2, characterized in that: The step S1 comprises the following steps: S101. By collecting existing data and connecting them together, modeling is carried out using 3D point cloud reverse modeling technology and component modeling technology; S102. Calling component models in a modular component library modeled using component-based modeling technology, matching the collected point cloud data with similar components in the modular component library, and performing intelligent parameter optimization on the matched components; S103. By integrating multi-scale, dynamic and static data using full-process digital modeling technology, a digital twin with multi-dimensional integration of "physical-information-business" is formed, virtually simulating the entire manufacturing information and business process. S104. Build and deploy a 3D dynamic simulation model of a large-scale stamping production line based on Unity.

4. The large-scale stamping production line simulation and optimization method according to claim 1, characterized in that: The multifunctional fault indicator feedback technology in step S2 includes: Hardware part: including sensor, signal processing circuit and indicator module; Software part: including sensor signal processing, fault judgment logic and data transmission protocol; The collected multi-source heterogeneous data is transmitted to the digital twin platform through the industrial network, and the protocol incompatibility problem is solved using dynamic optimization algorithms, and then imported into the data middle platform for overall management.

5. The large-scale stamping production line simulation and optimization method according to claim 1, characterized in that: The step S3 comprises the following steps: S301. Build a digital twin 3D visualization interface based on Unity, and achieve 3D dynamic mapping between physical space and virtual scenes through multi-source heterogeneous data fusion; S302. Based on three-dimensional dynamic real-time drive simulation technology, intelligent matching of physical equipment and virtual scenes, and allowing users to adjust production parameters through the interface; S303. Through distributed computing and storage technology, real-time collection of physical equipment operating data is carried out, and the virtual model is updated synchronously to maintain consistency between the virtual and the real; S304. Use step S2 to monitor equipment status, personnel location, vehicle information, order status, and warehouse status, and feed multi-source heterogeneous data back to the digital twin 3D visualization interface; S305. Through the background simulation engine, the virtual scene is updated according to the real-time data and fault warnings and production optimization instructions are generated.

6. The large-scale stamping production line simulation and optimization method according to claim 1, characterized in that: The step S4 comprises the following steps: S401. Collect multi-source heterogeneous data from the production process, including product quality data, equipment status data, material information, and production progress data, and integrate cross-departmental data; S402. Perform feature extraction and anomaly detection on the fused data using manufacturing data mining technology driven by deep neural networks; S403. Explore optimal production strategies based on deep reinforcement learning algorithms and adapt strategy generalization capabilities to different production scenarios through deep transfer learning networks; S404. Dynamically verify and iteratively optimize the optimal production strategy using a deep adversarial learning network; S405. Use deep autoencoders to compress and reconstruct the features of redundant, repeated, and missing data in the entire business process, generate a standardized business process diagram, and output optimized decision instructions to the upper-level system.

7. The large-scale stamping production line simulation and optimization method according to claim 6, characterized in that: The abnormality detection mentioned in step S402 specifically includes: Use the production schedule anomaly discrimination model to identify schedule deviations; Use production quality defect detection models to locate defect causes; Use equipment fault status identification model to predict equipment failure risk.

Images