Why is the concept of the digital twin gaining in prominence right now?
Willcox: Today’s tremendous computing power, combined with powerful algorithms, makes it possible to build and use digital twins. Machine learning can help to identify meaningful patterns in the large amounts of data we can collect from complex systems, such as an aircraft, but we also need physics-based models to make our digital twin predictive and useful. Another innovation is new hardware architectures that allow us to collect and analyze data efficiently to then incorporate the data into digital twins. One example of this is neuromorphic chips that are lightweight and energy-efficient, and so may be well suited for analyzing data onboard the system itself while it is in operation.
Do simulations have to mirror real processes for a correct result? Or is it enough if they simply predict the correct outcome?
Willcox: The use of a black box model is only OK as long as it works, but the question is how would you know? A critical question is whether you can trust your model – and trust is higher when you understand what is actually happening. That’s why physics-based models are essential.
What role does Siemens play in this field?
Willcox: I have met Siemens experts at workshops and conferences and have had exchanges with them. Siemens is instrumental in advancing these modeling technologies and leading the development and implementation of digital twins for real systems.
Willcox: Computer modeling and digital twins are becoming increasingly important. They improve the performance of a system, extend its life, and help reduce costs. Currently, this is mainly used for expensive and complex machines. In the future, however, the use of digital twins could become part of our daily lives, for example, when you optimize the energy management of your house. I expect that we will see more applications of digital twins in industry. To achieve this, however, will require changes to engineering curricula and training. Next-generation engineers and technicians need to be equipped not only to use these new instruments, but also to understand their potential and limitations.View Article