Definition
Digital Twin
A digital twin is an exact virtual representation of a real physical object throughout its entire lifecycle.
Such an object can be a single component, a system, or an entire system network. In this context, the representation encompasses all properties and behavioral aspects relevant to the specific application.
Digital shadow
The data layer—the digital shadow—constitutes the primary input information for the digital twin.
The digital shadow represents the totality of all measured variables and other information sources that describe the object’s states in the form of data.
These can be structured data generated from process variables (e.g., forces, temperatures, vibrations) or machine parameters (e.g., axis positions, spindle speeds, motor currents); such data follow a fixed format. Unstructured data (texts, videos, photographs), on the other hand, do not have a fixed format. Nevertheless, these types of data can be important for describing conditions in machining processes (e.g., optical images of cutting-edge wear, visual recordings of temperature fields).
Models
The models form the functional core of the digital twin.
They link the data of the digital shadow with the physical characteristics of the real object, thereby enabling its virtual representation.
Depending on the use case, analytical, numerical, empirical, or AI-based models are employed.
They serve to calculate states and properties, predict system behavior, and virtually analyze, evaluate, and optimize manufacturing processes.
Uncertainties
Uncertainty describes the discrepancy between a physical object and its digital representation. It arises from measurement inaccuracies, model assumptions, and the combination of multiple models, and it affects the validity of simulations and predictions. Quantifying these uncertainties is an essential prerequisite for reliably evaluating Digital Twin results and deriving robust decisions.
A digital twin creates transparency regarding processes and data flows. It thus forms a basis for the structured analysis, evaluation, and optimization of industrial manufacturing systems.