Physical Modeling in Digital Twin

Physical Modeling in Digital Twin

Definition

Physical modeling: modeling using mathematical model (differential equation) to be solved using numerical methods.
Physical modeling:

• differential equation
• block diagram

Stacks

Differential Equation

Block diagram

• Simulink (MATLAB)
• Xcos (Scilab)
• Modelica

According to qiEnablingTechnologiesTools2021

Physical modeling: describes accuracy information, material information, and assembly information. Feature modeling includes interactive feature definition, automatic feature recognition, and feature-based design.

According to thelenComprehensiveReviewDigital2022

• Solid body structural analysis using Finite Element Analysis (FEA). Software: Ansys Mechanical, Abaqus FEA, Altair HyperMesh.
• Thermal and fluid flow analysis using Computational Fluid Dynamics (CFD). Software: Ansys Fluent, Autodesk CFD, SimScale. Free software: OpenFOAM, Stanford University's Unstructured (SU2) package.
• Kinematic and dynamic analysis using multi-body dynamics (MBD). Software: MATLAB and Simulink Simscape Multibody, MSC ADAMS, RecurDyn. Free software: MBDyn.
• Hybrid Multiphysics simulations. Software: COMSOL Multiphysics, MSC Nastran. Free software: Advanced Simulation Library (ASL) and Chrono.

Physics based Modeling: Fidelity

• Level of fidelity (higher >> lower)
  • Fluid Dynamics: CFD >> Eddy Simulation and Navier-Stokes equation
  • Battery modeling: 3D Doyle-fuller Newman (DFN) >> 2+1D DFN >> DFN >> single particle model
• Balance between accuracy and computational cost
  • High fidelity physics modeling is still only an approximation
  • Need to quantify and compensate bias of the model (see thelenComprehensiveReviewDigital2022)
  • high-fidelity models contain larger numbers of parameters that need to be calibrated offline, and during online deployment, a subset of the parameters need to be further tuned

Other Examples of Physic-based Modeling

• Empirical Performance Model (Rassolkin et al., 2022) link
• link
• link