Physical to Virtual (P2V) Updating and Enhancement

Physical to Virtual (P2V) Updating and Enhancement

• Measurement as input.
  • Physical measurement to update the digital model. Mainly for Geometric Model
  • Physical system data as input for physics-based analysis/design. Mainly for Physic-based Model.
• Probabilistic model updating
  • State estimation using Bayesian filters
    • methods:
      • Kalman filter
      • Extended Kalman filter
      • Unscented Kalman filter
      • Particle filter
    • problems:
      • filtering (present). Example: fault diagnosis
      • prediction (future). Example: remaining useful life (RUL)
      • smoothing
• ML model updating
• Fault diagnostics and failure prognostics
• Ontology-based reasoning

Measurement as input (straightforward P2V) problems

• It is only applicable to cases where digital states (e.g., position, mobility, traffic flow) can be directly updated using monitoring data.
• However, for many engineering problems, digital states cannot be directly updated and are affected by various uncertainty sources.
• A digital state often cannot be directly measured (or observed) but can be estimated through noisy measurements that depend on the digital state. The digital state of a physical system in operation changes over time and can be estimated as new information about the physical system becomes available.

References