Design of a digital twin system for micro-lens array embossing manufacturing

Micro-lens arrays, due to their unique structure and outstanding optical performance, have been widely applied in advanced domains such as display technology, optical imaging, and laser processing. Among the manufacturing techniques, embossing molding is considered an efficient and precise method for fabricating micro-lens arrays. However, given the intricate structural characteristics of these arrays, precise control over the process becomes paramount. This study introduces a digital twin system for micro-lens array embossing manufacturing, grounded on a process knowledge base. The system relies on the rich process knowledge within the backend database combined with real-time sensor data, providing a powerful drive for the digital twin, thereby ensuring the precision and stability of the embossing manufacturing process. Additionally, the digital twin technology not only offers real-time monitoring and feedback for the production process of the micro-lens arrays but also aids researchers in optimizing process parameters, further enhancing product quality and consistency.


Introduction
Micro-lens arrays, due to their unique structure at the microscopic level and superior optical performance, have become key components in the modern optoelectronic domain, being widely applied in advanced technological fields like display technology [1], optical imaging [2], and laser processing [3].The fabrication of such microscopic components, given their high precision requirements, makes the selection and optimization of the production process a pivotal factor limiting their large-scale application.
Embossing molding, as an efficient method to produce micro-lens arrays, has been extensively researched and employed [4][5][6].This method, by accurately controlling key parameters such as material, temperature, pressure, and time, can yield micro-lens arrays with very high precision.However, ensuring the stability and reliability of this process in large-scale production is crucial.Traditional manufacturing monitoring often relies on post-production quality checks, which not only increases production costs but also makes it difficult to intervene timely in case of anomalies during the production process.
In recent years, digital twin technology, as a novel approach capable of digitally simulating physical entities, has demonstrated its vast potential across various application domains [7].Merging physical modeling, sensor technology, data analysis, and simulation techniques, digital twin provides a revolutionary tool for modern manufacturing in terms of production monitoring and optimization.For the embossing manufacture of micro-lens arrays, utilizing digital twin technology enables real-time monitoring of the production process, facilitating the prompt identification and rectification of deviations, thus ensuring the high quality and consistency of products.
This paper aims to explore the integration of digital twin technology with the embossing manufacturing process of micro-lens arrays, proposing a novel and efficient method for monitoring and optimizing the production of micro-lens arrays.

System introduction
The digital twin system framework for the micro-lens array embossing manufacturing center is illustrated in Figure 1, consisting of the physical system module of the embossing machine, the virtual twin system module, the backend knowledge database module, and the communication module.Firstly, the physical entity module of the embossing machine comprises the motion mechanism of the embossing machine, the heating module, and various sensors.Secondly, the virtual twin system module includes the twin model of the embossing machine and a monitoring interface.The backend process knowledge database for the embossing machine is built based on SQLServer [8], storing assembly process knowledge.Finally, the communication module employs the Modbus TCP protocol to achieve bidirectional communication between the virtual and physical modules of the embossing machine.

Physical system module of the embossing machine
Figure 2 illustrates the physical system structure of the embossing machine.To meet the full-process requirements of the micro-lens embossing technique, this system mainly consists of actuating components and sensors.The molding actuator achieves the molding of preforms through a high-power servo electric cylinder.Both the end of the servo cylinder and the mold placement platform are equipped with heating modules, ensuring that the preform is heated uniformly from both directions, enhancing heat conduction efficiency and allowing the material to quickly reach the desired temperature and state.To ensure the accuracy of pressure during the molding process, the servo cylinder is fitted with a precise pressure sensor for realtime monitoring and feedback.Furthermore, a vacuum environment is necessary during the molding process to prevent bubbles that affect the quality of micro-lenses, hence a complete vacuum system is employed, ensuring the sealing of the molding environment and precise control of the vacuum level.

Virtual twin system module
In the construction of the digital twin, accurately simulating the motion relationships of the virtual model is of paramount importance.To achieve this, it is first necessary to deeply adjust and optimize the hierarchical structure of the 3D CAD model.
The output shaft of the servo actuator is a key dynamic component in the system.Together with the serial components, they form a whole with relative motion, and their motion relationships should be strictly consistent.To this end, the output shaft of the servo actuator and its related components are defined as an independent sub-assembly, ensuring that they are presented as a unified motion module in the overall model.
The gate valve is a critical device for ensuring the sealing integrity of the embossing system, and its actions directly affect the system's sealing efficiency.By simulating the gate valve, its opening and closing actions can be translated into the upward and downward movement of the gate, achieving a realistic simulation effect in a virtual environment.
To achieve high-quality visualization effects in the digital twin system, the Unity engine was chosen as the rendering platform [9].For the original CAD model, to better integrate with the Unity engine, the 3DSMAX software was used to convert it into an fbx format.This format not only retains the hierarchical structure and animation information of the model but also ensures efficient rendering effects in Unity.
Through the above steps, the digital twin system of the embossing machine was successfully constructed, ensuring its performance in the virtual environment remains highly consistent with the actual physical system.

Knowledge base module
In the digital twin system, the backend knowledge base is a crucial component, primarily encompassing action knowledge and process knowledge.These insights not only offer an in-depth understanding of the equipment for the entire system but also lay the foundation for precise control and real-time monitoring.
Action knowledge primarily covers the actions and operations the equipment carries out according to the process.It serves as a digital description of the movements of the equipment's executing mechanisms, offering a reference for judging any abnormal actions.On the other hand, process knowledge is central to the equipment manufacturing process.Integrating various process parameters of the device, it offers a reference for controlling the quantities in the process.
The core idea behind building the process knowledge base is to comprehensively describe the equipment's state at every moment.To this end, the knowledge base was constructed following the principle of "process step -execution time -action -temperature -vacuum level -pressurestatus."This detailed categorization ensures every detail of the production process is captured and recorded:  The process step identifies every step or stage of the manufacturing process.


Execution time records the start and end times for each process step. Action details the specific operations the equipment performs in a particular process step. Temperature, vacuum level, and pressure are the key process parameters.All parameters in the knowledge base make up the entire process, influencing the quality and performance of the product.
 Status provides the specific conditions of the equipment at a given moment, such as whether it's running normally, needs maintenance, etc.
After integrating this detailed data, it is organized in tabular form and stored in the SQL Server database.This not only facilitates real-time monitoring and control but also ensures data security and integrity.This structured knowledge base backs the digital twin system robustly, guaranteeing the accuracy and stability of equipment operations.

Communication module
As illustrated in Figure 3, the communication architecture of the digital twin system for the molding machine is layered and organized, which can be divided into two main parts.The first part involves the communication between the actuator and sensors using the high-speed EtherCAT bus with the PLC, achieving real-time control commands and position feedback, providing low latency and high reliability in communication.Key sensors, such as pressure sensors, laser displacement sensors, and vacuum gauges, transmit various physical parameters to the PLC via the 485 communication protocol.The PLC then analyzes and processes this information, forming a unified data stream for the supervisory computer.
The second part pertains to the communication of integrated information in the PLC to the digital twin system supervisory computer using the Modbus TCP communication protocol [10].To ensure stable communication, detailed parameter configurations were performed on the PLC.The supervisory computer has a specialized communication interface program that can read data from the PLC and perform type conversion and analysis to adapt to the Unity environment.Simultaneously, a timer ensures periodic data reading from the PLC and updates the device model in the supervisory computer, guaranteeing the real-time accuracy and precision of the digital twin system.
In summary, this communication approach offers an efficient, stable, and precise data transmission and processing solution for the digital twin system of the molding machine, ensuring the authenticity and real-time nature of the digital twin.

Knowledge-based driven digital twin system
To ensure that the data in the digital twin system precisely matches the data from the actual production site and, at the same time, integrates with the backend process knowledge base, a digital twin detection strategy was adopted.
Firstly, frame-by-frame detection is conducted within the twin system.The system continuously inspects every frame of data, comparing in real-time whether the motion of the physical entity matches the simulation of the digital twin.The objective of this phase is to accurately capture any potential minor discrepancies or deviations.
Next, parameters that deviate from the process knowledge base data undergo error correction.Once the system identifies any inconsistencies during the frame-by-frame detection, it immediately initiates an error correction mechanism.This mechanism adjusts the digital twin model based on data from the actual production site, ensuring synchronization and consistency between the two.
Lastly, during the production process control, ensuring the proper sequential operation of the execution mechanisms is crucial for the success of the entire process flow.To achieve this precise synchronized simulation and control, the concept of a state machine is employed.The state machine plays a pivotal role throughout the process flow, capable of transitioning states based on real-time sensor data, continuously tracking, and simulating every process step.Moreover, the process knowledge base, being the system's core, not only retains the current process sequence but also provides the necessary state parameters for the digital twin environment.This means that when the state of the physical production line changes, the digital twin system can simulate and compare in real time, ensuring production efficiency and quality stability.

Experiments and conclusions
Based on the existing mechanical structure of the molding machine and in conjunction with the methods proposed and described in this paper, a digital twin system of the microlens array molding machine was successfully constructed, as shown in Figure 4.This digital twin system is not just a simple virtual model, but a highly integrated system that incorporates an advanced process knowledge database.Through this system, the manufacturing process of the microlens array can be monitored in real-time and effectively.Such real-time monitoring not only greatly improves production efficiency but also significantly reduces the error rate, providing a strong guarantee for ensuring product quality and performance.

Figure 2 .
Figure 2. Physical system structure of the embossing machine.

Figure 3 .
Figure 3. Communication architecture of the digital twin system for the molding machine.