A digital twin, or digital twin, is a perfect digital copy of information that describes a real entity, be it a product, a process or a system. The concept was born out of a 2002 presentation by Michael Grieves, now chief scientist for Advanced Manufacturing at the
Florida Institute of Technology, on the occasion of the establishment of a Product Lifecycle Management (PLM) center. Given that any information in the real space can be contained and reflected in the virtual one, once the “door” of communication and connection of information between the two spaces was found, each product
(or process) would have been formed by two systems interacting with each other , the physical one and the digital one. Two intertwined systems for the entire life cycle of the product, in which the virtual one would have helped the physicist to explore the “virtual subspaces”, or the different scenario possibilities, in all phases of creation, production, operation and disposal.
A dynamic mirroring of information (Information Mirror Model) which is also twinning between systems: since 2009 we are therefore talking about the Digital twin model. Grieves distinguishes between Digital twin Prototype (DTP) and Digital twin Instance (DTI), or the complete objects: both operating in the Digital twin Environment (DTE) which today is, more and more often, the cloud.
Another common classification is between product digital twin (efficient design), production digital twin (manufacturing planning) and performance digital twin (data optimization). However, the concept is the same: the digital twin is used both to predict / prevent the future behavior and performance of the physical product /
process at a change of variables, and to monitor its status in real time and promptly intervene in case of failures or optimize its performance. At the design stage, it is used to infer the best component configuration for design and functionality based on the necessary requirements.
Tests, continuous tests without “acting” on the real product / process, with the possibility of following its entire life cycle “remotely” and saving time, energy and resources. An example? Think of the costs, in the aerospace industry, of a test that ended badly, with the explosion of a rocket on the launch pad. With digital twin, the simulation has real effects, but it’s all virtual.
How the digital twin model works
The digital twin is generally a cloud PLM suite, composed of different software that follow the different phases of the product / process life. The “door” that connects physical space and virtual space is the Internet of Things:
the sensors and transducers (see article) positioned in the product and / or in the plant and / or in the production line that send data flows in time real, reworked and archived thanks to machine learning. But how does the digital twin affect an object or a production line?
In system engineering, for the 3D design phase, libraries are used that generate scenarios in a single model and parameterize the components from the data sheets, so as to simulate the customized configuration required and speed up the development process.