Digital Twin is a virtual representation of a physical object, system or process by using simulation and other computerized digital technologies to interact with the corresponding physical object, system or process. It can be used to analyze and simulate a system or process in real-world conditions to adapt to changes, increase flexibility, optimize operations, improve quality, save time, reduce costs and add value.
The concept of Digital Twins has been around since the early 2000s. It has then received greater attention and stronger momentum to thrive in the last few years, especially since it was named one of Gartner’s Top 10 Strategic Technology Trends for consecutively three years 2017, 2018, 2019.
Many publications, presentations, and demonstrations on the creation of digital twins by using numerical simulation with finite element modeling have been made. But most have focused on product or system designing. It is far more challenging, with still much work to be done, to create digital twins for the dynamic manufacturing processes, such as welding and joining processes. This, however, is essential for the complete digitalization of industry.
In case of welding, the Digital Twins shall be created for the complete welding lifecycle from Designing, Optimizing, Planning, to Producing, and Evaluating the welds. This requires the simulations with virtual models to be highly accurate and consistent with capabilities to make reliable predictions and optimizations.
SORPAS® is now ready to take the challenge for creating the first Welding Digital Twins for the complete lifecycle of resistance spot welding processes.
Digital Twins for Resistance Spot Welding
As illustrated in the figure, five digital twins are created with different functionalities through the complete lifecycle of resistance welding:
- twin 1 for Weld Designing
- twin 2 for Weld Optimizing
- twin 3 for Weld Planning
- twin 4 for Weld Producing
- twin 5 for Weld Evaluating
SORPAS® can be used to build each digital twin with numerical models with interactions to realistic data of the physical welding processes of welding production including the components, electrodes, welding equipment and welding procedures.
To enhance the interactions between the digital twins and the physical welding processes, machine learning and artificial intelligence (AI) will be incorporated too.
The Digital Twin for Weld Designing is built with geometric models and materials data of sheets, electrodes, and welding machine. The model and data of the sheets and electrodes must be realistic to accurately represent the shapes and materials of the real sheets and electrodes. The model and data of the welding machine may be generalized with typical data. No interaction is requested between the welding machine dynamics and the weld designs in the Designing phase. Its main purpose is to decide the best weld design from all possible geometric design options and available material selections.
The Digital Twin for Weld Optimizing is built on the basis of Weld Design with further detailed data and specifications of the welding machine and control equipment. Sensors may be included in the physical welding processes in order to improve the accuracy of numerical simulations by comparisons to the actual measurements in the physical welding tests. Its purpose is to obtain the optimal welding process window and assist optimizing and monitoring the physical welding processes.
The Digital Twin for Weld Planning is built on the basis of Weld Design and Weld Optimization with the purpose to decide the detailed weld schedules of specific welding process parameters for the given weld designs, welding equipment, and the welding sequences. Interactions between the Digital Twins and the physical welding processes may be established. Machine learning and artificial intelligence (AI) can be incorporated with the digital twins to improve autonomous decision-making capabilities.
The Digital Twin for Weld Producing is built on the basis of the Weld Design, Weld Optimization and Weld Planning with the purpose to take responses during the welding production when any welding conditions are out of range or causing problems to weld quality or production stability. The digital twins shall foresee and predict all possible root causes of the welding problems, while machine learning and AI will automatically identify the problems and make decisions to take reactions to solve the problems, correct errors, or indicate faults.
The Digital Twin for Weld Evaluating is built with the purpose to evaluate the weld quality and the weld performance. The weld quality is evaluated by predicting the weld nugget sizes and the material properties affected during and after welding. The performance of the welds is evaluated by predicting the weld strengths and failure modes.
All the Welding Digital Twins are connected by interactions to each other through the complete welding lifecycle from designing, optimizing, planning to producing and evaluating. This makes it possible to evaluate the weld qualities already during the designing phase and through all following phases. The interactions between the digital twins and physical welding processes with the assistance of machine learning and artificial intelligence (AI) can further improve the welding production stability and assure final weld quality.