Reproduce “Digital Twins in Industrial Applications: Concepts, Mathematical Modeling, and Use Cases” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Ali Mohammad-Djafari

An accessible synthesis of industrial digital twins connects their physical assets, virtual models, sensor interfaces, and synchronization mechanisms to mathematical tools including differential equations, machine learning, PINNs, stochastic degradation models, Bayesian diagnosis, control, and model reduction. Across maintenance, process control, fault diagnosis, supply chains, and lifecycle management, the paper presents hybrid modeling as a way to balance physical consistency with data-driven adaptability, while emphasizing unresolved issues in scalability, data quality, interoperability, and real-time deployment.

Key insights

  • Hybrid models such as PINNs can impose governing equations while learning from sensor observations, offering a compromise between the interpretability of physics-based models and the flexibility of data-driven methods.
  • A digital twin is useful only when its virtual representation is kept synchronized with the physical system; a model without current data becomes a static simulation, while data without a model offers limited explanation or prediction.
  • No single modeling strategy fits all applications: maintenance, control, diagnosis, logistics, and lifecycle management require different observations, objectives, uncertainty treatments, and computational speeds.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.