Digital twin

A digital representation of a specific physical asset or process that is updated with data from its real-world counterpart. It can support monitoring, simulation, and decision-making across the counterpart’s life cycle.

Internet of Things: A network of physical devices that sense or act on their surroundings and exchange data. Connected sensors can stream the counterpart’s operating data into its twin.
Physics-based modeling: Modeling that represents a system through equations derived from physical laws. Physical equations can encode how an asset behaves beyond its measured operating points.
Predictive maintenance: Maintenance planning that uses data and models to estimate when equipment is likely to fail or need service. A twin can combine condition data and simulated behavior to estimate maintenance needs.
Computer simulation: The use of computation to imitate the behavior of a real or hypothetical system. A simulation need not represent a specific counterpart or receive its ongoing data.
Data assimilation: The process of combining observations with a model to estimate the state of a system. It incorporates live measurements into the twin’s evolving state.
Finite element method: A numerical method that approximates solutions to engineering equations by dividing a domain into smaller elements. It often supplies detailed structural or thermal models for an asset twin.
Building information modeling: The creation and management of structured digital information about a building across its life cycle. A building twin can extend this model with operational data from the completed structure.
Digital model: A digital description of an object or system that is not necessarily connected to its physical counterpart. A static model lacks the continuing data connection that defines many twins.
System identification: The use of observed inputs and outputs to build or refine a model of a system. It helps construct models that reproduce the counterpart’s behavior.
Reduced-order modeling: The construction of a simpler model that approximates a more computationally expensive system. A compact model can make repeated twin simulations practical.

Linked from 21 pages