The use of a physics-based “Digital Twin” is getting more and more traction to develop the equipment virtually. Digital twins allow engineers to find the optimal design before the unit goes into production. However, these digital twins can’t be deployed at the operational level because they can be complex or too slow to react at the speed of operation.
The complexity arises from multiple length and time scales involved. One needs to deploy more than one simulation method and tool in a coordinated manner in order to make this feasible.
In this talk we will explore the scope of the length scales involved: multi-physics simulation, reduced order models (ROM), AI/ML integration, process simulation, fluid dynamics, thermal analysis, and structural integrity assessment. We will then consider how reduced order models (ROM) based on the concept of parameter-space exploration combined with ML can help us take a step towards executable Digital Twins.
It examines how these digital twins can accurately replicate physical processes, enabling real-time monitoring, predictive maintenance, and optimization, demonstrating their effectiveness in creating robust and agile digital representations.
The discussion highlights case studies from BASF, GSK, and Vale, where multi-physics detailed and/or system simulation solutions have been successfully implemented, illustrating the tangible benefits brought forth by digital twins, such as reduced downtime, increased productivity, and enhanced decision-making capabilities. Furthermore, the paper addresses challenges and best practices in adopting digital twin technology, offering insights into overcoming integration hurdles and maximizing the potential of digital simulations.