Data management system enabling digital twin for resistance spot welding

In today's rapidly evolving manufacturing processes, the digital twin is revolutionizing industrial processes through real-time data-driven insights. Resistance Spot Welding (RSW), a commonly used joining technique in the automotive and aerospace industries, is a key area where digital twin technology holds great promises for improving quality, efficiency, and predictive maintenance. However, the integration of digital twin models with RSW requires robust data management systems, capable of handling large volumes of data. This thesis presents the design and implementation of a data management system (DMS) that enables the development and operation of a digital twin for Resistance Spot Welding. The proposed DMS addresses the critical challenges in data collection, processing, storage, and analysis by incorporating modern data architecture principles with advanced analytics. A scalable and flexible architecture is developed, ensuring that the system can handle and process the data generated by sensors and monitoring equipment during the RSW process. The system's data collection framework integrates a wide range of personalizations that capture key parameters, such as welding current, electrode force, temperature, and results of the welding. This real-time data are then processed using modern techniques, like edge computing, to minimize latency and ensure the data's availability for nearly immediate feedback in the digital twin environment. The processed data are then stored in a microservices-based infrastructure, allowing for centralized management, long-term storage, and easy access for future analysis. One of the key point of this thesis is the unlock of the possibility to integrate a predictive model in the framework within the digital twin that uses historical data to forecast potential defects in the welding process. A Machine learning algorithm is actually employed to predict outcomes such as weld quality, joint strength, and the likelihood of equipment failure. These predictions enable proactive adjustments to the welding process, reducing defects, optimizing production, and extending equipment life. The system also incorporates a visualization tool, providing operators with a modern user-friendly interface to monitor welding performances and historical trends of the welding process. By integrating this tool into the system, it facilitates real-time decision-making and supports predictive maintenance strategies, leading to improved production efficiency and reduced downtime due to machine failures. The results of this thesis demonstrate the effectiveness of the proposed data management system in enabling a functional digital twin for Resistance Spot Welding. Furthermore, the scalability of the system is validated by its ability to handle increasing data loads as production environments become more complex and sensor networks expand. This thesis contributes to the growing body of knowledge in digital twin technology by offering a comprehensive data management solution tailored for high-frequency, high-volume industrial processes such as RSW. Starting from these points, future works can focus on refining the predictive models, integrating additional processing methods, and exploring the system's application to other welding techniques and manufacturing processes. Ultimately, the data management system developed in this thesis paves the way for the broader adoption of digital twins in advanced manufacturing, contributing to smarter and more efficient production lines.