Reproduce “State-of-the-Art Review: The Use of Digital Twins to Support Artificial Intelligence-Guided Predictive Maintenance” 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

Sizhe Ma, Katherine A. Flanigan, Mario Bergés

A requirements-based review organizes research on digital twins for AI-guided predictive maintenance across six informational and eight functional capabilities. It finds that existing systems support monitoring, diagnosis, prognosis, and recommendations in specific domains, but none yet provides a fully standardized, explainable, secure, adaptive, and transferable end-to-end framework.

Key insights

  • Digital twins can bring sensor data, simulations, engineering knowledge, degradation models, maintenance histories, and operational decisions into a shared, bidirectional information structure.
  • Physics-based, data-driven, Bayesian, fuzzy, surrogate, and reduced-order models each address different requirements, but their trade-offs in fidelity, speed, scalability, interpretability, and uncertainty handling remain unresolved.
  • Most current twins are tailored to individual assets or industries and provide only partial requirement coverage; a digital twin alone does not guarantee better maintenance accuracy or lower cost.

Summary by alphaXiv.

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