Self-Calibrating Mechanical Digital Twins for Multi-Physics Fault Localization in Industrial Gear Transmission Systems: A Systematic Review
Sciences du Nord Nature Science and Technology · 31.12.2025 · 2 (02) · 117-131
DOI: 10.71238/snnst.v2i02.232 ↗Kurz gefasst
Systematische Übersicht in der Zeitschrift Sciences du Nord Nature Science and Technology, online erschienen am 31.12.2025. Verfasst von Rupesh Ramesharo Dhore. Der Volltext ist frei zugänglich (Open Access, CC BY-SA 4.0).
Abstract (Originalsprache)
The loading applied to an industrial gear transmission system is non-stationary, the nature of the signatures of any faults is complex and difficult to localise due to the simultaneous meshes, time varying transmission paths and progressive deterioration. A digital twin can be used to continuously diagnose the machine in real-time, but if the twin is calibrated at commissioning, it will be inaccurate if any part of the physical system changes, for example due to wear, pitting or cracking. This systematic review aims to summarize the information from 30 peer-reviewed publications spanning over the last 14 years (2011-2024) and serve as a basis for laying the foundations of a self-calibrating mechanical digital twin for multi-physics fault localization. The themes of the corpus were themed, and the formulations governing the corpus were incorporated into a general framework of Bayesian state updating, hybrid physics–data decomposition, modelling of mesh stiffness, modelling of vibration signal on planets, degradation laws, and discrepancy minimisation for calibration. The synthesis shows that 2021 or later, 56.7% of the corpus was available, and 23.3% of the sources in the corpus were from a single journal: physics-informed learning resulted in relative errors less than 6.7 × 10−4 with only 100 samples taken from the boundaries. This representative example of a planetary stage demonstrates small difference in fault frequency between planet and ring (1.0Hz) to enable a frequency only location even though the use of transfer path analysis and calibrated virtual data is encouraged. There was a mean bias of 30% from 27 samples of a self-calibrating twin and a mean bias of 32.7% from 27 samples of a static twin. Main challenges are hybrid, continuously calibrated twins, and computational cost, parameter identifiability, cyber security and validation requirements are the most balanced solution.
Zitieren
APA
Dhore, R. R. (2025). Self-Calibrating Mechanical Digital Twins for Multi-Physics Fault Localization in Industrial Gear Transmission Systems: A Systematic Review. Sciences du Nord Nature Science and Technology, 2(02), 117-131. https://doi.org/10.71238/snnst.v2i02.232
Vancouver
Dhore RR. Self-Calibrating Mechanical Digital Twins for Multi-Physics Fault Localization in Industrial Gear Transmission Systems: A Systematic Review. Sciences du Nord Nature Science and Technology. 2025;2(02):117-131. doi:10.71238/snnst.v2i02.232
BibTeX
@article{Dhore2025,
author = {Dhore, Rupesh Ramesharo},
title = {Self-Calibrating Mechanical Digital Twins for Multi-Physics Fault Localization in Industrial Gear Transmission Systems: A Systematic Review},
journal = {Sciences du Nord Nature Science and Technology},
year = {2025},
volume = {2},
number = {02},
pages = {117--131},
doi = {10.71238/snnst.v2i02.232}
}
EUVIMED